diff options
Diffstat (limited to 'src/text')
27 files changed, 12267 insertions, 0 deletions
diff --git a/src/text/scanner/example_test.go b/src/text/scanner/example_test.go new file mode 100644 index 0000000..5e8c3fb --- /dev/null +++ b/src/text/scanner/example_test.go @@ -0,0 +1,140 @@ +// Copyright 2018 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package scanner_test + +import ( + "fmt" + "strings" + "text/scanner" + "unicode" +) + +func Example() { + const src = ` +// This is scanned code. +if a > 10 { + someParsable = text +}` + + var s scanner.Scanner + s.Init(strings.NewReader(src)) + s.Filename = "example" + for tok := s.Scan(); tok != scanner.EOF; tok = s.Scan() { + fmt.Printf("%s: %s\n", s.Position, s.TokenText()) + } + + // Output: + // example:3:1: if + // example:3:4: a + // example:3:6: > + // example:3:8: 10 + // example:3:11: { + // example:4:2: someParsable + // example:4:15: = + // example:4:17: text + // example:5:1: } +} + +func Example_isIdentRune() { + const src = "%var1 var2%" + + var s scanner.Scanner + s.Init(strings.NewReader(src)) + s.Filename = "default" + + for tok := s.Scan(); tok != scanner.EOF; tok = s.Scan() { + fmt.Printf("%s: %s\n", s.Position, s.TokenText()) + } + + fmt.Println() + s.Init(strings.NewReader(src)) + s.Filename = "percent" + + // treat leading '%' as part of an identifier + s.IsIdentRune = func(ch rune, i int) bool { + return ch == '%' && i == 0 || unicode.IsLetter(ch) || unicode.IsDigit(ch) && i > 0 + } + + for tok := s.Scan(); tok != scanner.EOF; tok = s.Scan() { + fmt.Printf("%s: %s\n", s.Position, s.TokenText()) + } + + // Output: + // default:1:1: % + // default:1:2: var1 + // default:1:7: var2 + // default:1:11: % + // + // percent:1:1: %var1 + // percent:1:7: var2 + // percent:1:11: % +} + +func Example_mode() { + const src = ` + // Comment begins at column 5. + +This line should not be included in the output. + +/* +This multiline comment +should be extracted in +its entirety. +*/ +` + + var s scanner.Scanner + s.Init(strings.NewReader(src)) + s.Filename = "comments" + s.Mode ^= scanner.SkipComments // don't skip comments + + for tok := s.Scan(); tok != scanner.EOF; tok = s.Scan() { + txt := s.TokenText() + if strings.HasPrefix(txt, "//") || strings.HasPrefix(txt, "/*") { + fmt.Printf("%s: %s\n", s.Position, txt) + } + } + + // Output: + // comments:2:5: // Comment begins at column 5. + // comments:6:1: /* + // This multiline comment + // should be extracted in + // its entirety. + // */ +} + +func Example_whitespace() { + // tab-separated values + const src = `aa ab ac ad +ba bb bc bd +ca cb cc cd +da db dc dd` + + var ( + col, row int + s scanner.Scanner + tsv [4][4]string // large enough for example above + ) + s.Init(strings.NewReader(src)) + s.Whitespace ^= 1<<'\t' | 1<<'\n' // don't skip tabs and new lines + + for tok := s.Scan(); tok != scanner.EOF; tok = s.Scan() { + switch tok { + case '\n': + row++ + col = 0 + case '\t': + col++ + default: + tsv[row][col] = s.TokenText() + } + } + + fmt.Print(tsv) + + // Output: + // [[aa ab ac ad] [ba bb bc bd] [ca cb cc cd] [da db dc dd]] +} diff --git a/src/text/scanner/scanner.go b/src/text/scanner/scanner.go new file mode 100644 index 0000000..c5fc4ff --- /dev/null +++ b/src/text/scanner/scanner.go @@ -0,0 +1,793 @@ +// Copyright 2009 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +// Package scanner provides a scanner and tokenizer for UTF-8-encoded text. +// It takes an io.Reader providing the source, which then can be tokenized +// through repeated calls to the Scan function. For compatibility with +// existing tools, the NUL character is not allowed. If the first character +// in the source is a UTF-8 encoded byte order mark (BOM), it is discarded. +// +// By default, a Scanner skips white space and Go comments and recognizes all +// literals as defined by the Go language specification. It may be +// customized to recognize only a subset of those literals and to recognize +// different identifier and white space characters. +package scanner + +import ( + "bytes" + "fmt" + "io" + "os" + "unicode" + "unicode/utf8" +) + +// Position is a value that represents a source position. +// A position is valid if Line > 0. +type Position struct { + Filename string // filename, if any + Offset int // byte offset, starting at 0 + Line int // line number, starting at 1 + Column int // column number, starting at 1 (character count per line) +} + +// IsValid reports whether the position is valid. +func (pos *Position) IsValid() bool { return pos.Line > 0 } + +func (pos Position) String() string { + s := pos.Filename + if s == "" { + s = "<input>" + } + if pos.IsValid() { + s += fmt.Sprintf(":%d:%d", pos.Line, pos.Column) + } + return s +} + +// Predefined mode bits to control recognition of tokens. For instance, +// to configure a Scanner such that it only recognizes (Go) identifiers, +// integers, and skips comments, set the Scanner's Mode field to: +// +// ScanIdents | ScanInts | SkipComments +// +// With the exceptions of comments, which are skipped if SkipComments is +// set, unrecognized tokens are not ignored. Instead, the scanner simply +// returns the respective individual characters (or possibly sub-tokens). +// For instance, if the mode is ScanIdents (not ScanStrings), the string +// "foo" is scanned as the token sequence '"' Ident '"'. +// +// Use GoTokens to configure the Scanner such that it accepts all Go +// literal tokens including Go identifiers. Comments will be skipped. +// +const ( + ScanIdents = 1 << -Ident + ScanInts = 1 << -Int + ScanFloats = 1 << -Float // includes Ints and hexadecimal floats + ScanChars = 1 << -Char + ScanStrings = 1 << -String + ScanRawStrings = 1 << -RawString + ScanComments = 1 << -Comment + SkipComments = 1 << -skipComment // if set with ScanComments, comments become white space + GoTokens = ScanIdents | ScanFloats | ScanChars | ScanStrings | ScanRawStrings | ScanComments | SkipComments +) + +// The result of Scan is one of these tokens or a Unicode character. +const ( + EOF = -(iota + 1) + Ident + Int + Float + Char + String + RawString + Comment + + // internal use only + skipComment +) + +var tokenString = map[rune]string{ + EOF: "EOF", + Ident: "Ident", + Int: "Int", + Float: "Float", + Char: "Char", + String: "String", + RawString: "RawString", + Comment: "Comment", +} + +// TokenString returns a printable string for a token or Unicode character. +func TokenString(tok rune) string { + if s, found := tokenString[tok]; found { + return s + } + return fmt.Sprintf("%q", string(tok)) +} + +// GoWhitespace is the default value for the Scanner's Whitespace field. +// Its value selects Go's white space characters. +const GoWhitespace = 1<<'\t' | 1<<'\n' | 1<<'\r' | 1<<' ' + +const bufLen = 1024 // at least utf8.UTFMax + +// A Scanner implements reading of Unicode characters and tokens from an io.Reader. +type Scanner struct { + // Input + src io.Reader + + // Source buffer + srcBuf [bufLen + 1]byte // +1 for sentinel for common case of s.next() + srcPos int // reading position (srcBuf index) + srcEnd int // source end (srcBuf index) + + // Source position + srcBufOffset int // byte offset of srcBuf[0] in source + line int // line count + column int // character count + lastLineLen int // length of last line in characters (for correct column reporting) + lastCharLen int // length of last character in bytes + + // Token text buffer + // Typically, token text is stored completely in srcBuf, but in general + // the token text's head may be buffered in tokBuf while the token text's + // tail is stored in srcBuf. + tokBuf bytes.Buffer // token text head that is not in srcBuf anymore + tokPos int // token text tail position (srcBuf index); valid if >= 0 + tokEnd int // token text tail end (srcBuf index) + + // One character look-ahead + ch rune // character before current srcPos + + // Error is called for each error encountered. If no Error + // function is set, the error is reported to os.Stderr. + Error func(s *Scanner, msg string) + + // ErrorCount is incremented by one for each error encountered. + ErrorCount int + + // The Mode field controls which tokens are recognized. For instance, + // to recognize Ints, set the ScanInts bit in Mode. The field may be + // changed at any time. + Mode uint + + // The Whitespace field controls which characters are recognized + // as white space. To recognize a character ch <= ' ' as white space, + // set the ch'th bit in Whitespace (the Scanner's behavior is undefined + // for values ch > ' '). The field may be changed at any time. + Whitespace uint64 + + // IsIdentRune is a predicate controlling the characters accepted + // as the ith rune in an identifier. The set of valid characters + // must not intersect with the set of white space characters. + // If no IsIdentRune function is set, regular Go identifiers are + // accepted instead. The field may be changed at any time. + IsIdentRune func(ch rune, i int) bool + + // Start position of most recently scanned token; set by Scan. + // Calling Init or Next invalidates the position (Line == 0). + // The Filename field is always left untouched by the Scanner. + // If an error is reported (via Error) and Position is invalid, + // the scanner is not inside a token. Call Pos to obtain an error + // position in that case, or to obtain the position immediately + // after the most recently scanned token. + Position +} + +// Init initializes a Scanner with a new source and returns s. +// Error is set to nil, ErrorCount is set to 0, Mode is set to GoTokens, +// and Whitespace is set to GoWhitespace. +func (s *Scanner) Init(src io.Reader) *Scanner { + s.src = src + + // initialize source buffer + // (the first call to next() will fill it by calling src.Read) + s.srcBuf[0] = utf8.RuneSelf // sentinel + s.srcPos = 0 + s.srcEnd = 0 + + // initialize source position + s.srcBufOffset = 0 + s.line = 1 + s.column = 0 + s.lastLineLen = 0 + s.lastCharLen = 0 + + // initialize token text buffer + // (required for first call to next()). + s.tokPos = -1 + + // initialize one character look-ahead + s.ch = -2 // no char read yet, not EOF + + // initialize public fields + s.Error = nil + s.ErrorCount = 0 + s.Mode = GoTokens + s.Whitespace = GoWhitespace + s.Line = 0 // invalidate token position + + return s +} + +// next reads and returns the next Unicode character. It is designed such +// that only a minimal amount of work needs to be done in the common ASCII +// case (one test to check for both ASCII and end-of-buffer, and one test +// to check for newlines). +func (s *Scanner) next() rune { + ch, width := rune(s.srcBuf[s.srcPos]), 1 + + if ch >= utf8.RuneSelf { + // uncommon case: not ASCII or not enough bytes + for s.srcPos+utf8.UTFMax > s.srcEnd && !utf8.FullRune(s.srcBuf[s.srcPos:s.srcEnd]) { + // not enough bytes: read some more, but first + // save away token text if any + if s.tokPos >= 0 { + s.tokBuf.Write(s.srcBuf[s.tokPos:s.srcPos]) + s.tokPos = 0 + // s.tokEnd is set by Scan() + } + // move unread bytes to beginning of buffer + copy(s.srcBuf[0:], s.srcBuf[s.srcPos:s.srcEnd]) + s.srcBufOffset += s.srcPos + // read more bytes + // (an io.Reader must return io.EOF when it reaches + // the end of what it is reading - simply returning + // n == 0 will make this loop retry forever; but the + // error is in the reader implementation in that case) + i := s.srcEnd - s.srcPos + n, err := s.src.Read(s.srcBuf[i:bufLen]) + s.srcPos = 0 + s.srcEnd = i + n + s.srcBuf[s.srcEnd] = utf8.RuneSelf // sentinel + if err != nil { + if err != io.EOF { + s.error(err.Error()) + } + if s.srcEnd == 0 { + if s.lastCharLen > 0 { + // previous character was not EOF + s.column++ + } + s.lastCharLen = 0 + return EOF + } + // If err == EOF, we won't be getting more + // bytes; break to avoid infinite loop. If + // err is something else, we don't know if + // we can get more bytes; thus also break. + break + } + } + // at least one byte + ch = rune(s.srcBuf[s.srcPos]) + if ch >= utf8.RuneSelf { + // uncommon case: not ASCII + ch, width = utf8.DecodeRune(s.srcBuf[s.srcPos:s.srcEnd]) + if ch == utf8.RuneError && width == 1 { + // advance for correct error position + s.srcPos += width + s.lastCharLen = width + s.column++ + s.error("invalid UTF-8 encoding") + return ch + } + } + } + + // advance + s.srcPos += width + s.lastCharLen = width + s.column++ + + // special situations + switch ch { + case 0: + // for compatibility with other tools + s.error("invalid character NUL") + case '\n': + s.line++ + s.lastLineLen = s.column + s.column = 0 + } + + return ch +} + +// Next reads and returns the next Unicode character. +// It returns EOF at the end of the source. It reports +// a read error by calling s.Error, if not nil; otherwise +// it prints an error message to os.Stderr. Next does not +// update the Scanner's Position field; use Pos() to +// get the current position. +func (s *Scanner) Next() rune { + s.tokPos = -1 // don't collect token text + s.Line = 0 // invalidate token position + ch := s.Peek() + if ch != EOF { + s.ch = s.next() + } + return ch +} + +// Peek returns the next Unicode character in the source without advancing +// the scanner. It returns EOF if the scanner's position is at the last +// character of the source. +func (s *Scanner) Peek() rune { + if s.ch == -2 { + // this code is only run for the very first character + s.ch = s.next() + if s.ch == '\uFEFF' { + s.ch = s.next() // ignore BOM + } + } + return s.ch +} + +func (s *Scanner) error(msg string) { + s.tokEnd = s.srcPos - s.lastCharLen // make sure token text is terminated + s.ErrorCount++ + if s.Error != nil { + s.Error(s, msg) + return + } + pos := s.Position + if !pos.IsValid() { + pos = s.Pos() + } + fmt.Fprintf(os.Stderr, "%s: %s\n", pos, msg) +} + +func (s *Scanner) errorf(format string, args ...interface{}) { + s.error(fmt.Sprintf(format, args...)) +} + +func (s *Scanner) isIdentRune(ch rune, i int) bool { + if s.IsIdentRune != nil { + return s.IsIdentRune(ch, i) + } + return ch == '_' || unicode.IsLetter(ch) || unicode.IsDigit(ch) && i > 0 +} + +func (s *Scanner) scanIdentifier() rune { + // we know the zero'th rune is OK; start scanning at the next one + ch := s.next() + for i := 1; s.isIdentRune(ch, i); i++ { + ch = s.next() + } + return ch +} + +func lower(ch rune) rune { return ('a' - 'A') | ch } // returns lower-case ch iff ch is ASCII letter +func isDecimal(ch rune) bool { return '0' <= ch && ch <= '9' } +func isHex(ch rune) bool { return '0' <= ch && ch <= '9' || 'a' <= lower(ch) && lower(ch) <= 'f' } + +// digits accepts the sequence { digit | '_' } starting with ch0. +// If base <= 10, digits accepts any decimal digit but records +// the first invalid digit >= base in *invalid if *invalid == 0. +// digits returns the first rune that is not part of the sequence +// anymore, and a bitset describing whether the sequence contained +// digits (bit 0 is set), or separators '_' (bit 1 is set). +func (s *Scanner) digits(ch0 rune, base int, invalid *rune) (ch rune, digsep int) { + ch = ch0 + if base <= 10 { + max := rune('0' + base) + for isDecimal(ch) || ch == '_' { + ds := 1 + if ch == '_' { + ds = 2 + } else if ch >= max && *invalid == 0 { + *invalid = ch + } + digsep |= ds + ch = s.next() + } + } else { + for isHex(ch) || ch == '_' { + ds := 1 + if ch == '_' { + ds = 2 + } + digsep |= ds + ch = s.next() + } + } + return +} + +func (s *Scanner) scanNumber(ch rune, seenDot bool) (rune, rune) { + base := 10 // number base + prefix := rune(0) // one of 0 (decimal), '0' (0-octal), 'x', 'o', or 'b' + digsep := 0 // bit 0: digit present, bit 1: '_' present + invalid := rune(0) // invalid digit in literal, or 0 + + // integer part + var tok rune + var ds int + if !seenDot { + tok = Int + if ch == '0' { + ch = s.next() + switch lower(ch) { + case 'x': + ch = s.next() + base, prefix = 16, 'x' + case 'o': + ch = s.next() + base, prefix = 8, 'o' + case 'b': + ch = s.next() + base, prefix = 2, 'b' + default: + base, prefix = 8, '0' + digsep = 1 // leading 0 + } + } + ch, ds = s.digits(ch, base, &invalid) + digsep |= ds + if ch == '.' && s.Mode&ScanFloats != 0 { + ch = s.next() + seenDot = true + } + } + + // fractional part + if seenDot { + tok = Float + if prefix == 'o' || prefix == 'b' { + s.error("invalid radix point in " + litname(prefix)) + } + ch, ds = s.digits(ch, base, &invalid) + digsep |= ds + } + + if digsep&1 == 0 { + s.error(litname(prefix) + " has no digits") + } + + // exponent + if e := lower(ch); (e == 'e' || e == 'p') && s.Mode&ScanFloats != 0 { + switch { + case e == 'e' && prefix != 0 && prefix != '0': + s.errorf("%q exponent requires decimal mantissa", ch) + case e == 'p' && prefix != 'x': + s.errorf("%q exponent requires hexadecimal mantissa", ch) + } + ch = s.next() + tok = Float + if ch == '+' || ch == '-' { + ch = s.next() + } + ch, ds = s.digits(ch, 10, nil) + digsep |= ds + if ds&1 == 0 { + s.error("exponent has no digits") + } + } else if prefix == 'x' && tok == Float { + s.error("hexadecimal mantissa requires a 'p' exponent") + } + + if tok == Int && invalid != 0 { + s.errorf("invalid digit %q in %s", invalid, litname(prefix)) + } + + if digsep&2 != 0 { + s.tokEnd = s.srcPos - s.lastCharLen // make sure token text is terminated + if i := invalidSep(s.TokenText()); i >= 0 { + s.error("'_' must separate successive digits") + } + } + + return tok, ch +} + +func litname(prefix rune) string { + switch prefix { + default: + return "decimal literal" + case 'x': + return "hexadecimal literal" + case 'o', '0': + return "octal literal" + case 'b': + return "binary literal" + } +} + +// invalidSep returns the index of the first invalid separator in x, or -1. +func invalidSep(x string) int { + x1 := ' ' // prefix char, we only care if it's 'x' + d := '.' // digit, one of '_', '0' (a digit), or '.' (anything else) + i := 0 + + // a prefix counts as a digit + if len(x) >= 2 && x[0] == '0' { + x1 = lower(rune(x[1])) + if x1 == 'x' || x1 == 'o' || x1 == 'b' { + d = '0' + i = 2 + } + } + + // mantissa and exponent + for ; i < len(x); i++ { + p := d // previous digit + d = rune(x[i]) + switch { + case d == '_': + if p != '0' { + return i + } + case isDecimal(d) || x1 == 'x' && isHex(d): + d = '0' + default: + if p == '_' { + return i - 1 + } + d = '.' + } + } + if d == '_' { + return len(x) - 1 + } + + return -1 +} + +func digitVal(ch rune) int { + switch { + case '0' <= ch && ch <= '9': + return int(ch - '0') + case 'a' <= lower(ch) && lower(ch) <= 'f': + return int(lower(ch) - 'a' + 10) + } + return 16 // larger than any legal digit val +} + +func (s *Scanner) scanDigits(ch rune, base, n int) rune { + for n > 0 && digitVal(ch) < base { + ch = s.next() + n-- + } + if n > 0 { + s.error("invalid char escape") + } + return ch +} + +func (s *Scanner) scanEscape(quote rune) rune { + ch := s.next() // read character after '/' + switch ch { + case 'a', 'b', 'f', 'n', 'r', 't', 'v', '\\', quote: + // nothing to do + ch = s.next() + case '0', '1', '2', '3', '4', '5', '6', '7': + ch = s.scanDigits(ch, 8, 3) + case 'x': + ch = s.scanDigits(s.next(), 16, 2) + case 'u': + ch = s.scanDigits(s.next(), 16, 4) + case 'U': + ch = s.scanDigits(s.next(), 16, 8) + default: + s.error("invalid char escape") + } + return ch +} + +func (s *Scanner) scanString(quote rune) (n int) { + ch := s.next() // read character after quote + for ch != quote { + if ch == '\n' || ch < 0 { + s.error("literal not terminated") + return + } + if ch == '\\' { + ch = s.scanEscape(quote) + } else { + ch = s.next() + } + n++ + } + return +} + +func (s *Scanner) scanRawString() { + ch := s.next() // read character after '`' + for ch != '`' { + if ch < 0 { + s.error("literal not terminated") + return + } + ch = s.next() + } +} + +func (s *Scanner) scanChar() { + if s.scanString('\'') != 1 { + s.error("invalid char literal") + } +} + +func (s *Scanner) scanComment(ch rune) rune { + // ch == '/' || ch == '*' + if ch == '/' { + // line comment + ch = s.next() // read character after "//" + for ch != '\n' && ch >= 0 { + ch = s.next() + } + return ch + } + + // general comment + ch = s.next() // read character after "/*" + for { + if ch < 0 { + s.error("comment not terminated") + break + } + ch0 := ch + ch = s.next() + if ch0 == '*' && ch == '/' { + ch = s.next() + break + } + } + return ch +} + +// Scan reads the next token or Unicode character from source and returns it. +// It only recognizes tokens t for which the respective Mode bit (1<<-t) is set. +// It returns EOF at the end of the source. It reports scanner errors (read and +// token errors) by calling s.Error, if not nil; otherwise it prints an error +// message to os.Stderr. +func (s *Scanner) Scan() rune { + ch := s.Peek() + + // reset token text position + s.tokPos = -1 + s.Line = 0 + +redo: + // skip white space + for s.Whitespace&(1<<uint(ch)) != 0 { + ch = s.next() + } + + // start collecting token text + s.tokBuf.Reset() + s.tokPos = s.srcPos - s.lastCharLen + + // set token position + // (this is a slightly optimized version of the code in Pos()) + s.Offset = s.srcBufOffset + s.tokPos + if s.column > 0 { + // common case: last character was not a '\n' + s.Line = s.line + s.Column = s.column + } else { + // last character was a '\n' + // (we cannot be at the beginning of the source + // since we have called next() at least once) + s.Line = s.line - 1 + s.Column = s.lastLineLen + } + + // determine token value + tok := ch + switch { + case s.isIdentRune(ch, 0): + if s.Mode&ScanIdents != 0 { + tok = Ident + ch = s.scanIdentifier() + } else { + ch = s.next() + } + case isDecimal(ch): + if s.Mode&(ScanInts|ScanFloats) != 0 { + tok, ch = s.scanNumber(ch, false) + } else { + ch = s.next() + } + default: + switch ch { + case EOF: + break + case '"': + if s.Mode&ScanStrings != 0 { + s.scanString('"') + tok = String + } + ch = s.next() + case '\'': + if s.Mode&ScanChars != 0 { + s.scanChar() + tok = Char + } + ch = s.next() + case '.': + ch = s.next() + if isDecimal(ch) && s.Mode&ScanFloats != 0 { + tok, ch = s.scanNumber(ch, true) + } + case '/': + ch = s.next() + if (ch == '/' || ch == '*') && s.Mode&ScanComments != 0 { + if s.Mode&SkipComments != 0 { + s.tokPos = -1 // don't collect token text + ch = s.scanComment(ch) + goto redo + } + ch = s.scanComment(ch) + tok = Comment + } + case '`': + if s.Mode&ScanRawStrings != 0 { + s.scanRawString() + tok = RawString + } + ch = s.next() + default: + ch = s.next() + } + } + + // end of token text + s.tokEnd = s.srcPos - s.lastCharLen + + s.ch = ch + return tok +} + +// Pos returns the position of the character immediately after +// the character or token returned by the last call to Next or Scan. +// Use the Scanner's Position field for the start position of the most +// recently scanned token. +func (s *Scanner) Pos() (pos Position) { + pos.Filename = s.Filename + pos.Offset = s.srcBufOffset + s.srcPos - s.lastCharLen + switch { + case s.column > 0: + // common case: last character was not a '\n' + pos.Line = s.line + pos.Column = s.column + case s.lastLineLen > 0: + // last character was a '\n' + pos.Line = s.line - 1 + pos.Column = s.lastLineLen + default: + // at the beginning of the source + pos.Line = 1 + pos.Column = 1 + } + return +} + +// TokenText returns the string corresponding to the most recently scanned token. +// Valid after calling Scan and in calls of Scanner.Error. +func (s *Scanner) TokenText() string { + if s.tokPos < 0 { + // no token text + return "" + } + + if s.tokEnd < s.tokPos { + // if EOF was reached, s.tokEnd is set to -1 (s.srcPos == 0) + s.tokEnd = s.tokPos + } + // s.tokEnd >= s.tokPos + + if s.tokBuf.Len() == 0 { + // common case: the entire token text is still in srcBuf + return string(s.srcBuf[s.tokPos:s.tokEnd]) + } + + // part of the token text was saved in tokBuf: save the rest in + // tokBuf as well and return its content + s.tokBuf.Write(s.srcBuf[s.tokPos:s.tokEnd]) + s.tokPos = s.tokEnd // ensure idempotency of TokenText() call + return s.tokBuf.String() +} diff --git a/src/text/scanner/scanner_test.go b/src/text/scanner/scanner_test.go new file mode 100644 index 0000000..fe39d30 --- /dev/null +++ b/src/text/scanner/scanner_test.go @@ -0,0 +1,915 @@ +// Copyright 2009 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package scanner + +import ( + "bytes" + "fmt" + "io" + "strings" + "testing" + "unicode/utf8" +) + +// A StringReader delivers its data one string segment at a time via Read. +type StringReader struct { + data []string + step int +} + +func (r *StringReader) Read(p []byte) (n int, err error) { + if r.step < len(r.data) { + s := r.data[r.step] + n = copy(p, s) + r.step++ + } else { + err = io.EOF + } + return +} + +func readRuneSegments(t *testing.T, segments []string) { + got := "" + want := strings.Join(segments, "") + s := new(Scanner).Init(&StringReader{data: segments}) + for { + ch := s.Next() + if ch == EOF { + break + } + got += string(ch) + } + if got != want { + t.Errorf("segments=%v got=%s want=%s", segments, got, want) + } +} + +var segmentList = [][]string{ + {}, + {""}, + {"日", "本語"}, + {"\u65e5", "\u672c", "\u8a9e"}, + {"\U000065e5", " ", "\U0000672c", "\U00008a9e"}, + {"\xe6", "\x97\xa5\xe6", "\x9c\xac\xe8\xaa\x9e"}, + {"Hello", ", ", "World", "!"}, + {"Hello", ", ", "", "World", "!"}, +} + +func TestNext(t *testing.T) { + for _, s := range segmentList { + readRuneSegments(t, s) + } +} + +type token struct { + tok rune + text string +} + +var f100 = "ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff" + +var tokenList = []token{ + {Comment, "// line comments"}, + {Comment, "//"}, + {Comment, "////"}, + {Comment, "// comment"}, + {Comment, "// /* comment */"}, + {Comment, "// // comment //"}, + {Comment, "//" + f100}, + + {Comment, "// general comments"}, + {Comment, "/**/"}, + {Comment, "/***/"}, + {Comment, "/* comment */"}, + {Comment, "/* // comment */"}, + {Comment, "/* /* comment */"}, + {Comment, "/*\n comment\n*/"}, + {Comment, "/*" + f100 + "*/"}, + + {Comment, "// identifiers"}, + {Ident, "a"}, + {Ident, "a0"}, + {Ident, "foobar"}, + {Ident, "abc123"}, + {Ident, "LGTM"}, + {Ident, "_"}, + {Ident, "_abc123"}, + {Ident, "abc123_"}, + {Ident, "_abc_123_"}, + {Ident, "_äöü"}, + {Ident, "_本"}, + {Ident, "äöü"}, + {Ident, "本"}, + {Ident, "a۰۱۸"}, + {Ident, "foo६४"}, + {Ident, "bar９８７６"}, + {Ident, f100}, + + {Comment, "// decimal ints"}, + {Int, "0"}, + {Int, "1"}, + {Int, "9"}, + {Int, "42"}, + {Int, "1234567890"}, + + {Comment, "// octal ints"}, + {Int, "00"}, + {Int, "01"}, + {Int, "07"}, + {Int, "042"}, + {Int, "01234567"}, + + {Comment, "// hexadecimal ints"}, + {Int, "0x0"}, + {Int, "0x1"}, + {Int, "0xf"}, + {Int, "0x42"}, + {Int, "0x123456789abcDEF"}, + {Int, "0x" + f100}, + {Int, "0X0"}, + {Int, "0X1"}, + {Int, "0XF"}, + {Int, "0X42"}, + {Int, "0X123456789abcDEF"}, + {Int, "0X" + f100}, + + {Comment, "// floats"}, + {Float, "0."}, + {Float, "1."}, + {Float, "42."}, + {Float, "01234567890."}, + {Float, ".0"}, + {Float, ".1"}, + {Float, ".42"}, + {Float, ".0123456789"}, + {Float, "0.0"}, + {Float, "1.0"}, + {Float, "42.0"}, + {Float, "01234567890.0"}, + {Float, "0e0"}, + {Float, "1e0"}, + {Float, "42e0"}, + {Float, "01234567890e0"}, + {Float, "0E0"}, + {Float, "1E0"}, + {Float, "42E0"}, + {Float, "01234567890E0"}, + {Float, "0e+10"}, + {Float, "1e-10"}, + {Float, "42e+10"}, + {Float, "01234567890e-10"}, + {Float, "0E+10"}, + {Float, "1E-10"}, + {Float, "42E+10"}, + {Float, "01234567890E-10"}, + + {Comment, "// chars"}, + {Char, `' '`}, + {Char, `'a'`}, + {Char, `'本'`}, + {Char, `'\a'`}, + {Char, `'\b'`}, + {Char, `'\f'`}, + {Char, `'\n'`}, + {Char, `'\r'`}, + {Char, `'\t'`}, + {Char, `'\v'`}, + {Char, `'\''`}, + {Char, `'\000'`}, + {Char, `'\777'`}, + {Char, `'\x00'`}, + {Char, `'\xff'`}, + {Char, `'\u0000'`}, + {Char, `'\ufA16'`}, + {Char, `'\U00000000'`}, + {Char, `'\U0000ffAB'`}, + + {Comment, "// strings"}, + {String, `" "`}, + {String, `"a"`}, + {String, `"本"`}, + {String, `"\a"`}, + {String, `"\b"`}, + {String, `"\f"`}, + {String, `"\n"`}, + {String, `"\r"`}, + {String, `"\t"`}, + {String, `"\v"`}, + {String, `"\""`}, + {String, `"\000"`}, + {String, `"\777"`}, + {String, `"\x00"`}, + {String, `"\xff"`}, + {String, `"\u0000"`}, + {String, `"\ufA16"`}, + {String, `"\U00000000"`}, + {String, `"\U0000ffAB"`}, + {String, `"` + f100 + `"`}, + + {Comment, "// raw strings"}, + {RawString, "``"}, + {RawString, "`\\`"}, + {RawString, "`" + "\n\n/* foobar */\n\n" + "`"}, + {RawString, "`" + f100 + "`"}, + + {Comment, "// individual characters"}, + // NUL character is not allowed + {'\x01', "\x01"}, + {' ' - 1, string(' ' - 1)}, + {'+', "+"}, + {'/', "/"}, + {'.', "."}, + {'~', "~"}, + {'(', "("}, +} + +func makeSource(pattern string) *bytes.Buffer { + var buf bytes.Buffer + for _, k := range tokenList { + fmt.Fprintf(&buf, pattern, k.text) + } + return &buf +} + +func checkTok(t *testing.T, s *Scanner, line int, got, want rune, text string) { + if got != want { + t.Fatalf("tok = %s, want %s for %q", TokenString(got), TokenString(want), text) + } + if s.Line != line { + t.Errorf("line = %d, want %d for %q", s.Line, line, text) + } + stext := s.TokenText() + if stext != text { + t.Errorf("text = %q, want %q", stext, text) + } else { + // check idempotency of TokenText() call + stext = s.TokenText() + if stext != text { + t.Errorf("text = %q, want %q (idempotency check)", stext, text) + } + } +} + +func checkTokErr(t *testing.T, s *Scanner, line int, want rune, text string) { + prevCount := s.ErrorCount + checkTok(t, s, line, s.Scan(), want, text) + if s.ErrorCount != prevCount+1 { + t.Fatalf("want error for %q", text) + } +} + +func countNewlines(s string) int { + n := 0 + for _, ch := range s { + if ch == '\n' { + n++ + } + } + return n +} + +func testScan(t *testing.T, mode uint) { + s := new(Scanner).Init(makeSource(" \t%s\n")) + s.Mode = mode + tok := s.Scan() + line := 1 + for _, k := range tokenList { + if mode&SkipComments == 0 || k.tok != Comment { + checkTok(t, s, line, tok, k.tok, k.text) + tok = s.Scan() + } + line += countNewlines(k.text) + 1 // each token is on a new line + } + checkTok(t, s, line, tok, EOF, "") +} + +func TestScan(t *testing.T) { + testScan(t, GoTokens) + testScan(t, GoTokens&^SkipComments) +} + +func TestInvalidExponent(t *testing.T) { + const src = "1.5e 1.5E 1e+ 1e- 1.5z" + s := new(Scanner).Init(strings.NewReader(src)) + s.Error = func(s *Scanner, msg string) { + const want = "exponent has no digits" + if msg != want { + t.Errorf("%s: got error %q; want %q", s.TokenText(), msg, want) + } + } + checkTokErr(t, s, 1, Float, "1.5e") + checkTokErr(t, s, 1, Float, "1.5E") + checkTokErr(t, s, 1, Float, "1e+") + checkTokErr(t, s, 1, Float, "1e-") + checkTok(t, s, 1, s.Scan(), Float, "1.5") + checkTok(t, s, 1, s.Scan(), Ident, "z") + checkTok(t, s, 1, s.Scan(), EOF, "") + if s.ErrorCount != 4 { + t.Errorf("%d errors, want 4", s.ErrorCount) + } +} + +func TestPosition(t *testing.T) { + src := makeSource("\t\t\t\t%s\n") + s := new(Scanner).Init(src) + s.Mode = GoTokens &^ SkipComments + s.Scan() + pos := Position{"", 4, 1, 5} + for _, k := range tokenList { + if s.Offset != pos.Offset { + t.Errorf("offset = %d, want %d for %q", s.Offset, pos.Offset, k.text) + } + if s.Line != pos.Line { + t.Errorf("line = %d, want %d for %q", s.Line, pos.Line, k.text) + } + if s.Column != pos.Column { + t.Errorf("column = %d, want %d for %q", s.Column, pos.Column, k.text) + } + pos.Offset += 4 + len(k.text) + 1 // 4 tabs + token bytes + newline + pos.Line += countNewlines(k.text) + 1 // each token is on a new line + s.Scan() + } + // make sure there were no token-internal errors reported by scanner + if s.ErrorCount != 0 { + t.Errorf("%d errors", s.ErrorCount) + } +} + +func TestScanZeroMode(t *testing.T) { + src := makeSource("%s\n") + str := src.String() + s := new(Scanner).Init(src) + s.Mode = 0 // don't recognize any token classes + s.Whitespace = 0 // don't skip any whitespace + tok := s.Scan() + for i, ch := range str { + if tok != ch { + t.Fatalf("%d. tok = %s, want %s", i, TokenString(tok), TokenString(ch)) + } + tok = s.Scan() + } + if tok != EOF { + t.Fatalf("tok = %s, want EOF", TokenString(tok)) + } + if s.ErrorCount != 0 { + t.Errorf("%d errors", s.ErrorCount) + } +} + +func testScanSelectedMode(t *testing.T, mode uint, class rune) { + src := makeSource("%s\n") + s := new(Scanner).Init(src) + s.Mode = mode + tok := s.Scan() + for tok != EOF { + if tok < 0 && tok != class { + t.Fatalf("tok = %s, want %s", TokenString(tok), TokenString(class)) + } + tok = s.Scan() + } + if s.ErrorCount != 0 { + t.Errorf("%d errors", s.ErrorCount) + } +} + +func TestScanSelectedMask(t *testing.T) { + testScanSelectedMode(t, 0, 0) + testScanSelectedMode(t, ScanIdents, Ident) + // Don't test ScanInts and ScanNumbers since some parts of + // the floats in the source look like (invalid) octal ints + // and ScanNumbers may return either Int or Float. + testScanSelectedMode(t, ScanChars, Char) + testScanSelectedMode(t, ScanStrings, String) + testScanSelectedMode(t, SkipComments, 0) + testScanSelectedMode(t, ScanComments, Comment) +} + +func TestScanCustomIdent(t *testing.T) { + const src = "faab12345 a12b123 a12 3b" + s := new(Scanner).Init(strings.NewReader(src)) + // ident = ( 'a' | 'b' ) { digit } . + // digit = '0' .. '3' . + // with a maximum length of 4 + s.IsIdentRune = func(ch rune, i int) bool { + return i == 0 && (ch == 'a' || ch == 'b') || 0 < i && i < 4 && '0' <= ch && ch <= '3' + } + checkTok(t, s, 1, s.Scan(), 'f', "f") + checkTok(t, s, 1, s.Scan(), Ident, "a") + checkTok(t, s, 1, s.Scan(), Ident, "a") + checkTok(t, s, 1, s.Scan(), Ident, "b123") + checkTok(t, s, 1, s.Scan(), Int, "45") + checkTok(t, s, 1, s.Scan(), Ident, "a12") + checkTok(t, s, 1, s.Scan(), Ident, "b123") + checkTok(t, s, 1, s.Scan(), Ident, "a12") + checkTok(t, s, 1, s.Scan(), Int, "3") + checkTok(t, s, 1, s.Scan(), Ident, "b") + checkTok(t, s, 1, s.Scan(), EOF, "") +} + +func TestScanNext(t *testing.T) { + const BOM = '\uFEFF' + BOMs := string(BOM) + s := new(Scanner).Init(strings.NewReader(BOMs + "if a == bcd /* com" + BOMs + "ment */ {\n\ta += c\n}" + BOMs + "// line comment ending in eof")) + checkTok(t, s, 1, s.Scan(), Ident, "if") // the first BOM is ignored + checkTok(t, s, 1, s.Scan(), Ident, "a") + checkTok(t, s, 1, s.Scan(), '=', "=") + checkTok(t, s, 0, s.Next(), '=', "") + checkTok(t, s, 0, s.Next(), ' ', "") + checkTok(t, s, 0, s.Next(), 'b', "") + checkTok(t, s, 1, s.Scan(), Ident, "cd") + checkTok(t, s, 1, s.Scan(), '{', "{") + checkTok(t, s, 2, s.Scan(), Ident, "a") + checkTok(t, s, 2, s.Scan(), '+', "+") + checkTok(t, s, 0, s.Next(), '=', "") + checkTok(t, s, 2, s.Scan(), Ident, "c") + checkTok(t, s, 3, s.Scan(), '}', "}") + checkTok(t, s, 3, s.Scan(), BOM, BOMs) + checkTok(t, s, 3, s.Scan(), -1, "") + if s.ErrorCount != 0 { + t.Errorf("%d errors", s.ErrorCount) + } +} + +func TestScanWhitespace(t *testing.T) { + var buf bytes.Buffer + var ws uint64 + // start at 1, NUL character is not allowed + for ch := byte(1); ch < ' '; ch++ { + buf.WriteByte(ch) + ws |= 1 << ch + } + const orig = 'x' + buf.WriteByte(orig) + + s := new(Scanner).Init(&buf) + s.Mode = 0 + s.Whitespace = ws + tok := s.Scan() + if tok != orig { + t.Errorf("tok = %s, want %s", TokenString(tok), TokenString(orig)) + } +} + +func testError(t *testing.T, src, pos, msg string, tok rune) { + s := new(Scanner).Init(strings.NewReader(src)) + errorCalled := false + s.Error = func(s *Scanner, m string) { + if !errorCalled { + // only look at first error + if p := s.Pos().String(); p != pos { + t.Errorf("pos = %q, want %q for %q", p, pos, src) + } + if m != msg { + t.Errorf("msg = %q, want %q for %q", m, msg, src) + } + errorCalled = true + } + } + tk := s.Scan() + if tk != tok { + t.Errorf("tok = %s, want %s for %q", TokenString(tk), TokenString(tok), src) + } + if !errorCalled { + t.Errorf("error handler not called for %q", src) + } + if s.ErrorCount == 0 { + t.Errorf("count = %d, want > 0 for %q", s.ErrorCount, src) + } +} + +func TestError(t *testing.T) { + testError(t, "\x00", "<input>:1:1", "invalid character NUL", 0) + testError(t, "\x80", "<input>:1:1", "invalid UTF-8 encoding", utf8.RuneError) + testError(t, "\xff", "<input>:1:1", "invalid UTF-8 encoding", utf8.RuneError) + + testError(t, "a\x00", "<input>:1:2", "invalid character NUL", Ident) + testError(t, "ab\x80", "<input>:1:3", "invalid UTF-8 encoding", Ident) + testError(t, "abc\xff", "<input>:1:4", "invalid UTF-8 encoding", Ident) + + testError(t, `"a`+"\x00", "<input>:1:3", "invalid character NUL", String) + testError(t, `"ab`+"\x80", "<input>:1:4", "invalid UTF-8 encoding", String) + testError(t, `"abc`+"\xff", "<input>:1:5", "invalid UTF-8 encoding", String) + + testError(t, "`a"+"\x00", "<input>:1:3", "invalid character NUL", RawString) + testError(t, "`ab"+"\x80", "<input>:1:4", "invalid UTF-8 encoding", RawString) + testError(t, "`abc"+"\xff", "<input>:1:5", "invalid UTF-8 encoding", RawString) + + testError(t, `'\"'`, "<input>:1:3", "invalid char escape", Char) + testError(t, `"\'"`, "<input>:1:3", "invalid char escape", String) + + testError(t, `01238`, "<input>:1:6", "invalid digit '8' in octal literal", Int) + testError(t, `01238123`, "<input>:1:9", "invalid digit '8' in octal literal", Int) + testError(t, `0x`, "<input>:1:3", "hexadecimal literal has no digits", Int) + testError(t, `0xg`, "<input>:1:3", "hexadecimal literal has no digits", Int) + testError(t, `'aa'`, "<input>:1:4", "invalid char literal", Char) + testError(t, `1.5e`, "<input>:1:5", "exponent has no digits", Float) + testError(t, `1.5E`, "<input>:1:5", "exponent has no digits", Float) + testError(t, `1.5e+`, "<input>:1:6", "exponent has no digits", Float) + testError(t, `1.5e-`, "<input>:1:6", "exponent has no digits", Float) + + testError(t, `'`, "<input>:1:2", "literal not terminated", Char) + testError(t, `'`+"\n", "<input>:1:2", "literal not terminated", Char) + testError(t, `"abc`, "<input>:1:5", "literal not terminated", String) + testError(t, `"abc`+"\n", "<input>:1:5", "literal not terminated", String) + testError(t, "`abc\n", "<input>:2:1", "literal not terminated", RawString) + testError(t, `/*/`, "<input>:1:4", "comment not terminated", EOF) +} + +// An errReader returns (0, err) where err is not io.EOF. +type errReader struct{} + +func (errReader) Read(b []byte) (int, error) { + return 0, io.ErrNoProgress // some error that is not io.EOF +} + +func TestIOError(t *testing.T) { + s := new(Scanner).Init(errReader{}) + errorCalled := false + s.Error = func(s *Scanner, msg string) { + if !errorCalled { + if want := io.ErrNoProgress.Error(); msg != want { + t.Errorf("msg = %q, want %q", msg, want) + } + errorCalled = true + } + } + tok := s.Scan() + if tok != EOF { + t.Errorf("tok = %s, want EOF", TokenString(tok)) + } + if !errorCalled { + t.Errorf("error handler not called") + } +} + +func checkPos(t *testing.T, got, want Position) { + if got.Offset != want.Offset || got.Line != want.Line || got.Column != want.Column { + t.Errorf("got offset, line, column = %d, %d, %d; want %d, %d, %d", + got.Offset, got.Line, got.Column, want.Offset, want.Line, want.Column) + } +} + +func checkNextPos(t *testing.T, s *Scanner, offset, line, column int, char rune) { + if ch := s.Next(); ch != char { + t.Errorf("ch = %s, want %s", TokenString(ch), TokenString(char)) + } + want := Position{Offset: offset, Line: line, Column: column} + checkPos(t, s.Pos(), want) +} + +func checkScanPos(t *testing.T, s *Scanner, offset, line, column int, char rune) { + want := Position{Offset: offset, Line: line, Column: column} + checkPos(t, s.Pos(), want) + if ch := s.Scan(); ch != char { + t.Errorf("ch = %s, want %s", TokenString(ch), TokenString(char)) + if string(ch) != s.TokenText() { + t.Errorf("tok = %q, want %q", s.TokenText(), string(ch)) + } + } + checkPos(t, s.Position, want) +} + +func TestPos(t *testing.T) { + // corner case: empty source + s := new(Scanner).Init(strings.NewReader("")) + checkPos(t, s.Pos(), Position{Offset: 0, Line: 1, Column: 1}) + s.Peek() // peek doesn't affect the position + checkPos(t, s.Pos(), Position{Offset: 0, Line: 1, Column: 1}) + + // corner case: source with only a newline + s = new(Scanner).Init(strings.NewReader("\n")) + checkPos(t, s.Pos(), Position{Offset: 0, Line: 1, Column: 1}) + checkNextPos(t, s, 1, 2, 1, '\n') + // after EOF position doesn't change + for i := 10; i > 0; i-- { + checkScanPos(t, s, 1, 2, 1, EOF) + } + if s.ErrorCount != 0 { + t.Errorf("%d errors", s.ErrorCount) + } + + // corner case: source with only a single character + s = new(Scanner).Init(strings.NewReader("本")) + checkPos(t, s.Pos(), Position{Offset: 0, Line: 1, Column: 1}) + checkNextPos(t, s, 3, 1, 2, '本') + // after EOF position doesn't change + for i := 10; i > 0; i-- { + checkScanPos(t, s, 3, 1, 2, EOF) + } + if s.ErrorCount != 0 { + t.Errorf("%d errors", s.ErrorCount) + } + + // positions after calling Next + s = new(Scanner).Init(strings.NewReader(" foo६४ \n\n本語\n")) + checkNextPos(t, s, 1, 1, 2, ' ') + s.Peek() // peek doesn't affect the position + checkNextPos(t, s, 2, 1, 3, ' ') + checkNextPos(t, s, 3, 1, 4, 'f') + checkNextPos(t, s, 4, 1, 5, 'o') + checkNextPos(t, s, 5, 1, 6, 'o') + checkNextPos(t, s, 8, 1, 7, '६') + checkNextPos(t, s, 11, 1, 8, '४') + checkNextPos(t, s, 12, 1, 9, ' ') + checkNextPos(t, s, 13, 1, 10, ' ') + checkNextPos(t, s, 14, 2, 1, '\n') + checkNextPos(t, s, 15, 3, 1, '\n') + checkNextPos(t, s, 18, 3, 2, '本') + checkNextPos(t, s, 21, 3, 3, '語') + checkNextPos(t, s, 22, 4, 1, '\n') + // after EOF position doesn't change + for i := 10; i > 0; i-- { + checkScanPos(t, s, 22, 4, 1, EOF) + } + if s.ErrorCount != 0 { + t.Errorf("%d errors", s.ErrorCount) + } + + // positions after calling Scan + s = new(Scanner).Init(strings.NewReader("abc\n本語\n\nx")) + s.Mode = 0 + s.Whitespace = 0 + checkScanPos(t, s, 0, 1, 1, 'a') + s.Peek() // peek doesn't affect the position + checkScanPos(t, s, 1, 1, 2, 'b') + checkScanPos(t, s, 2, 1, 3, 'c') + checkScanPos(t, s, 3, 1, 4, '\n') + checkScanPos(t, s, 4, 2, 1, '本') + checkScanPos(t, s, 7, 2, 2, '語') + checkScanPos(t, s, 10, 2, 3, '\n') + checkScanPos(t, s, 11, 3, 1, '\n') + checkScanPos(t, s, 12, 4, 1, 'x') + // after EOF position doesn't change + for i := 10; i > 0; i-- { + checkScanPos(t, s, 13, 4, 2, EOF) + } + if s.ErrorCount != 0 { + t.Errorf("%d errors", s.ErrorCount) + } +} + +type countReader int + +func (r *countReader) Read([]byte) (int, error) { + *r++ + return 0, io.EOF +} + +func TestNextEOFHandling(t *testing.T) { + var r countReader + + // corner case: empty source + s := new(Scanner).Init(&r) + + tok := s.Next() + if tok != EOF { + t.Error("1) EOF not reported") + } + + tok = s.Peek() + if tok != EOF { + t.Error("2) EOF not reported") + } + + if r != 1 { + t.Errorf("scanner called Read %d times, not once", r) + } +} + +func TestScanEOFHandling(t *testing.T) { + var r countReader + + // corner case: empty source + s := new(Scanner).Init(&r) + + tok := s.Scan() + if tok != EOF { + t.Error("1) EOF not reported") + } + + tok = s.Peek() + if tok != EOF { + t.Error("2) EOF not reported") + } + + if r != 1 { + t.Errorf("scanner called Read %d times, not once", r) + } +} + +func TestIssue29723(t *testing.T) { + s := new(Scanner).Init(strings.NewReader(`x "`)) + s.Error = func(s *Scanner, _ string) { + got := s.TokenText() // this call shouldn't panic + const want = `"` + if got != want { + t.Errorf("got %q; want %q", got, want) + } + } + for r := s.Scan(); r != EOF; r = s.Scan() { + } +} + +func TestNumbers(t *testing.T) { + for _, test := range []struct { + tok rune + src, tokens, err string + }{ + // binaries + {Int, "0b0", "0b0", ""}, + {Int, "0b1010", "0b1010", ""}, + {Int, "0B1110", "0B1110", ""}, + + {Int, "0b", "0b", "binary literal has no digits"}, + {Int, "0b0190", "0b0190", "invalid digit '9' in binary literal"}, + {Int, "0b01a0", "0b01 a0", ""}, // only accept 0-9 + + // binary floats (invalid) + {Float, "0b.", "0b.", "invalid radix point in binary literal"}, + {Float, "0b.1", "0b.1", "invalid radix point in binary literal"}, + {Float, "0b1.0", "0b1.0", "invalid radix point in binary literal"}, + {Float, "0b1e10", "0b1e10", "'e' exponent requires decimal mantissa"}, + {Float, "0b1P-1", "0b1P-1", "'P' exponent requires hexadecimal mantissa"}, + + // octals + {Int, "0o0", "0o0", ""}, + {Int, "0o1234", "0o1234", ""}, + {Int, "0O1234", "0O1234", ""}, + + {Int, "0o", "0o", "octal literal has no digits"}, + {Int, "0o8123", "0o8123", "invalid digit '8' in octal literal"}, + {Int, "0o1293", "0o1293", "invalid digit '9' in octal literal"}, + {Int, "0o12a3", "0o12 a3", ""}, // only accept 0-9 + + // octal floats (invalid) + {Float, "0o.", "0o.", "invalid radix point in octal literal"}, + {Float, "0o.2", "0o.2", "invalid radix point in octal literal"}, + {Float, "0o1.2", "0o1.2", "invalid radix point in octal literal"}, + {Float, "0o1E+2", "0o1E+2", "'E' exponent requires decimal mantissa"}, + {Float, "0o1p10", "0o1p10", "'p' exponent requires hexadecimal mantissa"}, + + // 0-octals + {Int, "0", "0", ""}, + {Int, "0123", "0123", ""}, + + {Int, "08123", "08123", "invalid digit '8' in octal literal"}, + {Int, "01293", "01293", "invalid digit '9' in octal literal"}, + {Int, "0F.", "0 F .", ""}, // only accept 0-9 + {Int, "0123F.", "0123 F .", ""}, + {Int, "0123456x", "0123456 x", ""}, + + // decimals + {Int, "1", "1", ""}, + {Int, "1234", "1234", ""}, + + {Int, "1f", "1 f", ""}, // only accept 0-9 + + // decimal floats + {Float, "0.", "0.", ""}, + {Float, "123.", "123.", ""}, + {Float, "0123.", "0123.", ""}, + + {Float, ".0", ".0", ""}, + {Float, ".123", ".123", ""}, + {Float, ".0123", ".0123", ""}, + + {Float, "0.0", "0.0", ""}, + {Float, "123.123", "123.123", ""}, + {Float, "0123.0123", "0123.0123", ""}, + + {Float, "0e0", "0e0", ""}, + {Float, "123e+0", "123e+0", ""}, + {Float, "0123E-1", "0123E-1", ""}, + + {Float, "0.e+1", "0.e+1", ""}, + {Float, "123.E-10", "123.E-10", ""}, + {Float, "0123.e123", "0123.e123", ""}, + + {Float, ".0e-1", ".0e-1", ""}, + {Float, ".123E+10", ".123E+10", ""}, + {Float, ".0123E123", ".0123E123", ""}, + + {Float, "0.0e1", "0.0e1", ""}, + {Float, "123.123E-10", "123.123E-10", ""}, + {Float, "0123.0123e+456", "0123.0123e+456", ""}, + + {Float, "0e", "0e", "exponent has no digits"}, + {Float, "0E+", "0E+", "exponent has no digits"}, + {Float, "1e+f", "1e+ f", "exponent has no digits"}, + {Float, "0p0", "0p0", "'p' exponent requires hexadecimal mantissa"}, + {Float, "1.0P-1", "1.0P-1", "'P' exponent requires hexadecimal mantissa"}, + + // hexadecimals + {Int, "0x0", "0x0", ""}, + {Int, "0x1234", "0x1234", ""}, + {Int, "0xcafef00d", "0xcafef00d", ""}, + {Int, "0XCAFEF00D", "0XCAFEF00D", ""}, + + {Int, "0x", "0x", "hexadecimal literal has no digits"}, + {Int, "0x1g", "0x1 g", ""}, + + // hexadecimal floats + {Float, "0x0p0", "0x0p0", ""}, + {Float, "0x12efp-123", "0x12efp-123", ""}, + {Float, "0xABCD.p+0", "0xABCD.p+0", ""}, + {Float, "0x.0189P-0", "0x.0189P-0", ""}, + {Float, "0x1.ffffp+1023", "0x1.ffffp+1023", ""}, + + {Float, "0x.", "0x.", "hexadecimal literal has no digits"}, + {Float, "0x0.", "0x0.", "hexadecimal mantissa requires a 'p' exponent"}, + {Float, "0x.0", "0x.0", "hexadecimal mantissa requires a 'p' exponent"}, + {Float, "0x1.1", "0x1.1", "hexadecimal mantissa requires a 'p' exponent"}, + {Float, "0x1.1e0", "0x1.1e0", "hexadecimal mantissa requires a 'p' exponent"}, + {Float, "0x1.2gp1a", "0x1.2 gp1a", "hexadecimal mantissa requires a 'p' exponent"}, + {Float, "0x0p", "0x0p", "exponent has no digits"}, + {Float, "0xeP-", "0xeP-", "exponent has no digits"}, + {Float, "0x1234PAB", "0x1234P AB", "exponent has no digits"}, + {Float, "0x1.2p1a", "0x1.2p1 a", ""}, + + // separators + {Int, "0b_1000_0001", "0b_1000_0001", ""}, + {Int, "0o_600", "0o_600", ""}, + {Int, "0_466", "0_466", ""}, + {Int, "1_000", "1_000", ""}, + {Float, "1_000.000_1", "1_000.000_1", ""}, + {Int, "0x_f00d", "0x_f00d", ""}, + {Float, "0x_f00d.0p1_2", "0x_f00d.0p1_2", ""}, + + {Int, "0b__1000", "0b__1000", "'_' must separate successive digits"}, + {Int, "0o60___0", "0o60___0", "'_' must separate successive digits"}, + {Int, "0466_", "0466_", "'_' must separate successive digits"}, + {Float, "1_.", "1_.", "'_' must separate successive digits"}, + {Float, "0._1", "0._1", "'_' must separate successive digits"}, + {Float, "2.7_e0", "2.7_e0", "'_' must separate successive digits"}, + {Int, "0x___0", "0x___0", "'_' must separate successive digits"}, + {Float, "0x1.0_p0", "0x1.0_p0", "'_' must separate successive digits"}, + } { + s := new(Scanner).Init(strings.NewReader(test.src)) + var err string + s.Error = func(s *Scanner, msg string) { + if err == "" { + err = msg + } + } + + for i, want := range strings.Split(test.tokens, " ") { + err = "" + tok := s.Scan() + lit := s.TokenText() + if i == 0 { + if tok != test.tok { + t.Errorf("%q: got token %s; want %s", test.src, TokenString(tok), TokenString(test.tok)) + } + if err != test.err { + t.Errorf("%q: got error %q; want %q", test.src, err, test.err) + } + } + if lit != want { + t.Errorf("%q: got literal %q (%s); want %s", test.src, lit, TokenString(tok), want) + } + } + + // make sure we read all + if tok := s.Scan(); tok != EOF { + t.Errorf("%q: got %s; want EOF", test.src, TokenString(tok)) + } + } +} + +func TestIssue30320(t *testing.T) { + for _, test := range []struct { + in, want string + mode uint + }{ + {"foo01.bar31.xx-0-1-1-0", "01 31 0 1 1 0", ScanInts}, + {"foo0/12/0/5.67", "0 12 0 5 67", ScanInts}, + {"xxx1e0yyy", "1 0", ScanInts}, + {"1_2", "1_2", ScanInts}, + {"xxx1.0yyy2e3ee", "1 0 2 3", ScanInts}, + {"xxx1.0yyy2e3ee", "1.0 2e3", ScanFloats}, + } { + got := extractInts(test.in, test.mode) + if got != test.want { + t.Errorf("%q: got %q; want %q", test.in, got, test.want) + } + } +} + +func extractInts(t string, mode uint) (res string) { + var s Scanner + s.Init(strings.NewReader(t)) + s.Mode = mode + for { + switch tok := s.Scan(); tok { + case Int, Float: + if len(res) > 0 { + res += " " + } + res += s.TokenText() + case EOF: + return + } + } +} diff --git a/src/text/tabwriter/example_test.go b/src/text/tabwriter/example_test.go new file mode 100644 index 0000000..422ec11 --- /dev/null +++ b/src/text/tabwriter/example_test.go @@ -0,0 +1,73 @@ +// Copyright 2012 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package tabwriter_test + +import ( + "fmt" + "os" + "text/tabwriter" +) + +func ExampleWriter_Init() { + w := new(tabwriter.Writer) + + // Format in tab-separated columns with a tab stop of 8. + w.Init(os.Stdout, 0, 8, 0, '\t', 0) + fmt.Fprintln(w, "a\tb\tc\td\t.") + fmt.Fprintln(w, "123\t12345\t1234567\t123456789\t.") + fmt.Fprintln(w) + w.Flush() + + // Format right-aligned in space-separated columns of minimal width 5 + // and at least one blank of padding (so wider column entries do not + // touch each other). + w.Init(os.Stdout, 5, 0, 1, ' ', tabwriter.AlignRight) + fmt.Fprintln(w, "a\tb\tc\td\t.") + fmt.Fprintln(w, "123\t12345\t1234567\t123456789\t.") + fmt.Fprintln(w) + w.Flush() + + // output: + // a b c d . + // 123 12345 1234567 123456789 . + // + // a b c d. + // 123 12345 1234567 123456789. +} + +func Example_elastic() { + // Observe how the b's and the d's, despite appearing in the + // second cell of each line, belong to different columns. + w := tabwriter.NewWriter(os.Stdout, 0, 0, 1, '.', tabwriter.AlignRight|tabwriter.Debug) + fmt.Fprintln(w, "a\tb\tc") + fmt.Fprintln(w, "aa\tbb\tcc") + fmt.Fprintln(w, "aaa\t") // trailing tab + fmt.Fprintln(w, "aaaa\tdddd\teeee") + w.Flush() + + // output: + // ....a|..b|c + // ...aa|.bb|cc + // ..aaa| + // .aaaa|.dddd|eeee +} + +func Example_trailingTab() { + // Observe that the third line has no trailing tab, + // so its final cell is not part of an aligned column. + const padding = 3 + w := tabwriter.NewWriter(os.Stdout, 0, 0, padding, '-', tabwriter.AlignRight|tabwriter.Debug) + fmt.Fprintln(w, "a\tb\taligned\t") + fmt.Fprintln(w, "aa\tbb\taligned\t") + fmt.Fprintln(w, "aaa\tbbb\tunaligned") // no trailing tab + fmt.Fprintln(w, "aaaa\tbbbb\taligned\t") + w.Flush() + + // output: + // ------a|------b|---aligned| + // -----aa|-----bb|---aligned| + // ----aaa|----bbb|unaligned + // ---aaaa|---bbbb|---aligned| +} diff --git a/src/text/tabwriter/tabwriter.go b/src/text/tabwriter/tabwriter.go new file mode 100644 index 0000000..76dec7b --- /dev/null +++ b/src/text/tabwriter/tabwriter.go @@ -0,0 +1,609 @@ +// Copyright 2009 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +// Package tabwriter implements a write filter (tabwriter.Writer) that +// translates tabbed columns in input into properly aligned text. +// +// The package is using the Elastic Tabstops algorithm described at +// http://nickgravgaard.com/elastictabstops/index.html. +// +// The text/tabwriter package is frozen and is not accepting new features. +package tabwriter + +import ( + "io" + "unicode/utf8" +) + +// ---------------------------------------------------------------------------- +// Filter implementation + +// A cell represents a segment of text terminated by tabs or line breaks. +// The text itself is stored in a separate buffer; cell only describes the +// segment's size in bytes, its width in runes, and whether it's an htab +// ('\t') terminated cell. +// +type cell struct { + size int // cell size in bytes + width int // cell width in runes + htab bool // true if the cell is terminated by an htab ('\t') +} + +// A Writer is a filter that inserts padding around tab-delimited +// columns in its input to align them in the output. +// +// The Writer treats incoming bytes as UTF-8-encoded text consisting +// of cells terminated by horizontal ('\t') or vertical ('\v') tabs, +// and newline ('\n') or formfeed ('\f') characters; both newline and +// formfeed act as line breaks. +// +// Tab-terminated cells in contiguous lines constitute a column. The +// Writer inserts padding as needed to make all cells in a column have +// the same width, effectively aligning the columns. It assumes that +// all characters have the same width, except for tabs for which a +// tabwidth must be specified. Column cells must be tab-terminated, not +// tab-separated: non-tab terminated trailing text at the end of a line +// forms a cell but that cell is not part of an aligned column. +// For instance, in this example (where | stands for a horizontal tab): +// +// aaaa|bbb|d +// aa |b |dd +// a | +// aa |cccc|eee +// +// the b and c are in distinct columns (the b column is not contiguous +// all the way). The d and e are not in a column at all (there's no +// terminating tab, nor would the column be contiguous). +// +// The Writer assumes that all Unicode code points have the same width; +// this may not be true in some fonts or if the string contains combining +// characters. +// +// If DiscardEmptyColumns is set, empty columns that are terminated +// entirely by vertical (or "soft") tabs are discarded. Columns +// terminated by horizontal (or "hard") tabs are not affected by +// this flag. +// +// If a Writer is configured to filter HTML, HTML tags and entities +// are passed through. The widths of tags and entities are +// assumed to be zero (tags) and one (entities) for formatting purposes. +// +// A segment of text may be escaped by bracketing it with Escape +// characters. The tabwriter passes escaped text segments through +// unchanged. In particular, it does not interpret any tabs or line +// breaks within the segment. If the StripEscape flag is set, the +// Escape characters are stripped from the output; otherwise they +// are passed through as well. For the purpose of formatting, the +// width of the escaped text is always computed excluding the Escape +// characters. +// +// The formfeed character acts like a newline but it also terminates +// all columns in the current line (effectively calling Flush). Tab- +// terminated cells in the next line start new columns. Unless found +// inside an HTML tag or inside an escaped text segment, formfeed +// characters appear as newlines in the output. +// +// The Writer must buffer input internally, because proper spacing +// of one line may depend on the cells in future lines. Clients must +// call Flush when done calling Write. +// +type Writer struct { + // configuration + output io.Writer + minwidth int + tabwidth int + padding int + padbytes [8]byte + flags uint + + // current state + buf []byte // collected text excluding tabs or line breaks + pos int // buffer position up to which cell.width of incomplete cell has been computed + cell cell // current incomplete cell; cell.width is up to buf[pos] excluding ignored sections + endChar byte // terminating char of escaped sequence (Escape for escapes, '>', ';' for HTML tags/entities, or 0) + lines [][]cell // list of lines; each line is a list of cells + widths []int // list of column widths in runes - re-used during formatting +} + +// addLine adds a new line. +// flushed is a hint indicating whether the underlying writer was just flushed. +// If so, the previous line is not likely to be a good indicator of the new line's cells. +func (b *Writer) addLine(flushed bool) { + // Grow slice instead of appending, + // as that gives us an opportunity + // to re-use an existing []cell. + if n := len(b.lines) + 1; n <= cap(b.lines) { + b.lines = b.lines[:n] + b.lines[n-1] = b.lines[n-1][:0] + } else { + b.lines = append(b.lines, nil) + } + + if !flushed { + // The previous line is probably a good indicator + // of how many cells the current line will have. + // If the current line's capacity is smaller than that, + // abandon it and make a new one. + if n := len(b.lines); n >= 2 { + if prev := len(b.lines[n-2]); prev > cap(b.lines[n-1]) { + b.lines[n-1] = make([]cell, 0, prev) + } + } + } +} + +// Reset the current state. +func (b *Writer) reset() { + b.buf = b.buf[:0] + b.pos = 0 + b.cell = cell{} + b.endChar = 0 + b.lines = b.lines[0:0] + b.widths = b.widths[0:0] + b.addLine(true) +} + +// Internal representation (current state): +// +// - all text written is appended to buf; tabs and line breaks are stripped away +// - at any given time there is a (possibly empty) incomplete cell at the end +// (the cell starts after a tab or line break) +// - cell.size is the number of bytes belonging to the cell so far +// - cell.width is text width in runes of that cell from the start of the cell to +// position pos; html tags and entities are excluded from this width if html +// filtering is enabled +// - the sizes and widths of processed text are kept in the lines list +// which contains a list of cells for each line +// - the widths list is a temporary list with current widths used during +// formatting; it is kept in Writer because it's re-used +// +// |<---------- size ---------->| +// | | +// |<- width ->|<- ignored ->| | +// | | | | +// [---processed---tab------------<tag>...</tag>...] +// ^ ^ ^ +// | | | +// buf start of incomplete cell pos + +// Formatting can be controlled with these flags. +const ( + // Ignore html tags and treat entities (starting with '&' + // and ending in ';') as single characters (width = 1). + FilterHTML uint = 1 << iota + + // Strip Escape characters bracketing escaped text segments + // instead of passing them through unchanged with the text. + StripEscape + + // Force right-alignment of cell content. + // Default is left-alignment. + AlignRight + + // Handle empty columns as if they were not present in + // the input in the first place. + DiscardEmptyColumns + + // Always use tabs for indentation columns (i.e., padding of + // leading empty cells on the left) independent of padchar. + TabIndent + + // Print a vertical bar ('|') between columns (after formatting). + // Discarded columns appear as zero-width columns ("||"). + Debug +) + +// A Writer must be initialized with a call to Init. The first parameter (output) +// specifies the filter output. The remaining parameters control the formatting: +// +// minwidth minimal cell width including any padding +// tabwidth width of tab characters (equivalent number of spaces) +// padding padding added to a cell before computing its width +// padchar ASCII char used for padding +// if padchar == '\t', the Writer will assume that the +// width of a '\t' in the formatted output is tabwidth, +// and cells are left-aligned independent of align_left +// (for correct-looking results, tabwidth must correspond +// to the tab width in the viewer displaying the result) +// flags formatting control +// +func (b *Writer) Init(output io.Writer, minwidth, tabwidth, padding int, padchar byte, flags uint) *Writer { + if minwidth < 0 || tabwidth < 0 || padding < 0 { + panic("negative minwidth, tabwidth, or padding") + } + b.output = output + b.minwidth = minwidth + b.tabwidth = tabwidth + b.padding = padding + for i := range b.padbytes { + b.padbytes[i] = padchar + } + if padchar == '\t' { + // tab padding enforces left-alignment + flags &^= AlignRight + } + b.flags = flags + + b.reset() + + return b +} + +// debugging support (keep code around) +func (b *Writer) dump() { + pos := 0 + for i, line := range b.lines { + print("(", i, ") ") + for _, c := range line { + print("[", string(b.buf[pos:pos+c.size]), "]") + pos += c.size + } + print("\n") + } + print("\n") +} + +// local error wrapper so we can distinguish errors we want to return +// as errors from genuine panics (which we don't want to return as errors) +type osError struct { + err error +} + +func (b *Writer) write0(buf []byte) { + n, err := b.output.Write(buf) + if n != len(buf) && err == nil { + err = io.ErrShortWrite + } + if err != nil { + panic(osError{err}) + } +} + +func (b *Writer) writeN(src []byte, n int) { + for n > len(src) { + b.write0(src) + n -= len(src) + } + b.write0(src[0:n]) +} + +var ( + newline = []byte{'\n'} + tabs = []byte("\t\t\t\t\t\t\t\t") +) + +func (b *Writer) writePadding(textw, cellw int, useTabs bool) { + if b.padbytes[0] == '\t' || useTabs { + // padding is done with tabs + if b.tabwidth == 0 { + return // tabs have no width - can't do any padding + } + // make cellw the smallest multiple of b.tabwidth + cellw = (cellw + b.tabwidth - 1) / b.tabwidth * b.tabwidth + n := cellw - textw // amount of padding + if n < 0 { + panic("internal error") + } + b.writeN(tabs, (n+b.tabwidth-1)/b.tabwidth) + return + } + + // padding is done with non-tab characters + b.writeN(b.padbytes[0:], cellw-textw) +} + +var vbar = []byte{'|'} + +func (b *Writer) writeLines(pos0 int, line0, line1 int) (pos int) { + pos = pos0 + for i := line0; i < line1; i++ { + line := b.lines[i] + + // if TabIndent is set, use tabs to pad leading empty cells + useTabs := b.flags&TabIndent != 0 + + for j, c := range line { + if j > 0 && b.flags&Debug != 0 { + // indicate column break + b.write0(vbar) + } + + if c.size == 0 { + // empty cell + if j < len(b.widths) { + b.writePadding(c.width, b.widths[j], useTabs) + } + } else { + // non-empty cell + useTabs = false + if b.flags&AlignRight == 0 { // align left + b.write0(b.buf[pos : pos+c.size]) + pos += c.size + if j < len(b.widths) { + b.writePadding(c.width, b.widths[j], false) + } + } else { // align right + if j < len(b.widths) { + b.writePadding(c.width, b.widths[j], false) + } + b.write0(b.buf[pos : pos+c.size]) + pos += c.size + } + } + } + + if i+1 == len(b.lines) { + // last buffered line - we don't have a newline, so just write + // any outstanding buffered data + b.write0(b.buf[pos : pos+b.cell.size]) + pos += b.cell.size + } else { + // not the last line - write newline + b.write0(newline) + } + } + return +} + +// Format the text between line0 and line1 (excluding line1); pos +// is the buffer position corresponding to the beginning of line0. +// Returns the buffer position corresponding to the beginning of +// line1 and an error, if any. +// +func (b *Writer) format(pos0 int, line0, line1 int) (pos int) { + pos = pos0 + column := len(b.widths) + for this := line0; this < line1; this++ { + line := b.lines[this] + + if column >= len(line)-1 { + continue + } + // cell exists in this column => this line + // has more cells than the previous line + // (the last cell per line is ignored because cells are + // tab-terminated; the last cell per line describes the + // text before the newline/formfeed and does not belong + // to a column) + + // print unprinted lines until beginning of block + pos = b.writeLines(pos, line0, this) + line0 = this + + // column block begin + width := b.minwidth // minimal column width + discardable := true // true if all cells in this column are empty and "soft" + for ; this < line1; this++ { + line = b.lines[this] + if column >= len(line)-1 { + break + } + // cell exists in this column + c := line[column] + // update width + if w := c.width + b.padding; w > width { + width = w + } + // update discardable + if c.width > 0 || c.htab { + discardable = false + } + } + // column block end + + // discard empty columns if necessary + if discardable && b.flags&DiscardEmptyColumns != 0 { + width = 0 + } + + // format and print all columns to the right of this column + // (we know the widths of this column and all columns to the left) + b.widths = append(b.widths, width) // push width + pos = b.format(pos, line0, this) + b.widths = b.widths[0 : len(b.widths)-1] // pop width + line0 = this + } + + // print unprinted lines until end + return b.writeLines(pos, line0, line1) +} + +// Append text to current cell. +func (b *Writer) append(text []byte) { + b.buf = append(b.buf, text...) + b.cell.size += len(text) +} + +// Update the cell width. +func (b *Writer) updateWidth() { + b.cell.width += utf8.RuneCount(b.buf[b.pos:]) + b.pos = len(b.buf) +} + +// To escape a text segment, bracket it with Escape characters. +// For instance, the tab in this string "Ignore this tab: \xff\t\xff" +// does not terminate a cell and constitutes a single character of +// width one for formatting purposes. +// +// The value 0xff was chosen because it cannot appear in a valid UTF-8 sequence. +// +const Escape = '\xff' + +// Start escaped mode. +func (b *Writer) startEscape(ch byte) { + switch ch { + case Escape: + b.endChar = Escape + case '<': + b.endChar = '>' + case '&': + b.endChar = ';' + } +} + +// Terminate escaped mode. If the escaped text was an HTML tag, its width +// is assumed to be zero for formatting purposes; if it was an HTML entity, +// its width is assumed to be one. In all other cases, the width is the +// unicode width of the text. +// +func (b *Writer) endEscape() { + switch b.endChar { + case Escape: + b.updateWidth() + if b.flags&StripEscape == 0 { + b.cell.width -= 2 // don't count the Escape chars + } + case '>': // tag of zero width + case ';': + b.cell.width++ // entity, count as one rune + } + b.pos = len(b.buf) + b.endChar = 0 +} + +// Terminate the current cell by adding it to the list of cells of the +// current line. Returns the number of cells in that line. +// +func (b *Writer) terminateCell(htab bool) int { + b.cell.htab = htab + line := &b.lines[len(b.lines)-1] + *line = append(*line, b.cell) + b.cell = cell{} + return len(*line) +} + +func (b *Writer) handlePanic(err *error, op string) { + if e := recover(); e != nil { + if op == "Flush" { + // If Flush ran into a panic, we still need to reset. + b.reset() + } + if nerr, ok := e.(osError); ok { + *err = nerr.err + return + } + panic("tabwriter: panic during " + op) + } +} + +// Flush should be called after the last call to Write to ensure +// that any data buffered in the Writer is written to output. Any +// incomplete escape sequence at the end is considered +// complete for formatting purposes. +func (b *Writer) Flush() error { + return b.flush() +} + +// flush is the internal version of Flush, with a named return value which we +// don't want to expose. +func (b *Writer) flush() (err error) { + defer b.handlePanic(&err, "Flush") + b.flushNoDefers() + return nil +} + +// flushNoDefers is like flush, but without a deferred handlePanic call. This +// can be called from other methods which already have their own deferred +// handlePanic calls, such as Write, and avoid the extra defer work. +func (b *Writer) flushNoDefers() { + // add current cell if not empty + if b.cell.size > 0 { + if b.endChar != 0 { + // inside escape - terminate it even if incomplete + b.endEscape() + } + b.terminateCell(false) + } + + // format contents of buffer + b.format(0, 0, len(b.lines)) + b.reset() +} + +var hbar = []byte("---\n") + +// Write writes buf to the writer b. +// The only errors returned are ones encountered +// while writing to the underlying output stream. +// +func (b *Writer) Write(buf []byte) (n int, err error) { + defer b.handlePanic(&err, "Write") + + // split text into cells + n = 0 + for i, ch := range buf { + if b.endChar == 0 { + // outside escape + switch ch { + case '\t', '\v', '\n', '\f': + // end of cell + b.append(buf[n:i]) + b.updateWidth() + n = i + 1 // ch consumed + ncells := b.terminateCell(ch == '\t') + if ch == '\n' || ch == '\f' { + // terminate line + b.addLine(ch == '\f') + if ch == '\f' || ncells == 1 { + // A '\f' always forces a flush. Otherwise, if the previous + // line has only one cell which does not have an impact on + // the formatting of the following lines (the last cell per + // line is ignored by format()), thus we can flush the + // Writer contents. + b.flushNoDefers() + if ch == '\f' && b.flags&Debug != 0 { + // indicate section break + b.write0(hbar) + } + } + } + + case Escape: + // start of escaped sequence + b.append(buf[n:i]) + b.updateWidth() + n = i + if b.flags&StripEscape != 0 { + n++ // strip Escape + } + b.startEscape(Escape) + + case '<', '&': + // possibly an html tag/entity + if b.flags&FilterHTML != 0 { + // begin of tag/entity + b.append(buf[n:i]) + b.updateWidth() + n = i + b.startEscape(ch) + } + } + + } else { + // inside escape + if ch == b.endChar { + // end of tag/entity + j := i + 1 + if ch == Escape && b.flags&StripEscape != 0 { + j = i // strip Escape + } + b.append(buf[n:j]) + n = i + 1 // ch consumed + b.endEscape() + } + } + } + + // append leftover text + b.append(buf[n:]) + n = len(buf) + return +} + +// NewWriter allocates and initializes a new tabwriter.Writer. +// The parameters are the same as for the Init function. +// +func NewWriter(output io.Writer, minwidth, tabwidth, padding int, padchar byte, flags uint) *Writer { + return new(Writer).Init(output, minwidth, tabwidth, padding, padchar, flags) +} diff --git a/src/text/tabwriter/tabwriter_test.go b/src/text/tabwriter/tabwriter_test.go new file mode 100644 index 0000000..a51358d --- /dev/null +++ b/src/text/tabwriter/tabwriter_test.go @@ -0,0 +1,754 @@ +// Copyright 2009 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package tabwriter_test + +import ( + "bytes" + "fmt" + "io" + "testing" + . "text/tabwriter" +) + +type buffer struct { + a []byte +} + +func (b *buffer) init(n int) { b.a = make([]byte, 0, n) } + +func (b *buffer) clear() { b.a = b.a[0:0] } + +func (b *buffer) Write(buf []byte) (written int, err error) { + n := len(b.a) + m := len(buf) + if n+m <= cap(b.a) { + b.a = b.a[0 : n+m] + for i := 0; i < m; i++ { + b.a[n+i] = buf[i] + } + } else { + panic("buffer.Write: buffer too small") + } + return len(buf), nil +} + +func (b *buffer) String() string { return string(b.a) } + +func write(t *testing.T, testname string, w *Writer, src string) { + written, err := io.WriteString(w, src) + if err != nil { + t.Errorf("--- test: %s\n--- src:\n%q\n--- write error: %v\n", testname, src, err) + } + if written != len(src) { + t.Errorf("--- test: %s\n--- src:\n%q\n--- written = %d, len(src) = %d\n", testname, src, written, len(src)) + } +} + +func verify(t *testing.T, testname string, w *Writer, b *buffer, src, expected string) { + err := w.Flush() + if err != nil { + t.Errorf("--- test: %s\n--- src:\n%q\n--- flush error: %v\n", testname, src, err) + } + + res := b.String() + if res != expected { + t.Errorf("--- test: %s\n--- src:\n%q\n--- found:\n%q\n--- expected:\n%q\n", testname, src, res, expected) + } +} + +func check(t *testing.T, testname string, minwidth, tabwidth, padding int, padchar byte, flags uint, src, expected string) { + var b buffer + b.init(1000) + + var w Writer + w.Init(&b, minwidth, tabwidth, padding, padchar, flags) + + // write all at once + title := testname + " (written all at once)" + b.clear() + write(t, title, &w, src) + verify(t, title, &w, &b, src, expected) + + // write byte-by-byte + title = testname + " (written byte-by-byte)" + b.clear() + for i := 0; i < len(src); i++ { + write(t, title, &w, src[i:i+1]) + } + verify(t, title, &w, &b, src, expected) + + // write using Fibonacci slice sizes + title = testname + " (written in fibonacci slices)" + b.clear() + for i, d := 0, 0; i < len(src); { + write(t, title, &w, src[i:i+d]) + i, d = i+d, d+1 + if i+d > len(src) { + d = len(src) - i + } + } + verify(t, title, &w, &b, src, expected) +} + +var tests = []struct { + testname string + minwidth, tabwidth, padding int + padchar byte + flags uint + src, expected string +}{ + { + "1a", + 8, 0, 1, '.', 0, + "", + "", + }, + + { + "1a debug", + 8, 0, 1, '.', Debug, + "", + "", + }, + + { + "1b esc stripped", + 8, 0, 1, '.', StripEscape, + "\xff\xff", + "", + }, + + { + "1b esc", + 8, 0, 1, '.', 0, + "\xff\xff", + "\xff\xff", + }, + + { + "1c esc stripped", + 8, 0, 1, '.', StripEscape, + "\xff\t\xff", + "\t", + }, + + { + "1c esc", + 8, 0, 1, '.', 0, + "\xff\t\xff", + "\xff\t\xff", + }, + + { + "1d esc stripped", + 8, 0, 1, '.', StripEscape, + "\xff\"foo\t\n\tbar\"\xff", + "\"foo\t\n\tbar\"", + }, + + { + "1d esc", + 8, 0, 1, '.', 0, + "\xff\"foo\t\n\tbar\"\xff", + "\xff\"foo\t\n\tbar\"\xff", + }, + + { + "1e esc stripped", + 8, 0, 1, '.', StripEscape, + "abc\xff\tdef", // unterminated escape + "abc\tdef", + }, + + { + "1e esc", + 8, 0, 1, '.', 0, + "abc\xff\tdef", // unterminated escape + "abc\xff\tdef", + }, + + { + "2", + 8, 0, 1, '.', 0, + "\n\n\n", + "\n\n\n", + }, + + { + "3", + 8, 0, 1, '.', 0, + "a\nb\nc", + "a\nb\nc", + }, + + { + "4a", + 8, 0, 1, '.', 0, + "\t", // '\t' terminates an empty cell on last line - nothing to print + "", + }, + + { + "4b", + 8, 0, 1, '.', AlignRight, + "\t", // '\t' terminates an empty cell on last line - nothing to print + "", + }, + + { + "5", + 8, 0, 1, '.', 0, + "*\t*", + "*.......*", + }, + + { + "5b", + 8, 0, 1, '.', 0, + "*\t*\n", + "*.......*\n", + }, + + { + "5c", + 8, 0, 1, '.', 0, + "*\t*\t", + "*.......*", + }, + + { + "5c debug", + 8, 0, 1, '.', Debug, + "*\t*\t", + "*.......|*", + }, + + { + "5d", + 8, 0, 1, '.', AlignRight, + "*\t*\t", + ".......**", + }, + + { + "6", + 8, 0, 1, '.', 0, + "\t\n", + "........\n", + }, + + { + "7a", + 8, 0, 1, '.', 0, + "a) foo", + "a) foo", + }, + + { + "7b", + 8, 0, 1, ' ', 0, + "b) foo\tbar", + "b) foo bar", + }, + + { + "7c", + 8, 0, 1, '.', 0, + "c) foo\tbar\t", + "c) foo..bar", + }, + + { + "7d", + 8, 0, 1, '.', 0, + "d) foo\tbar\n", + "d) foo..bar\n", + }, + + { + "7e", + 8, 0, 1, '.', 0, + "e) foo\tbar\t\n", + "e) foo..bar.....\n", + }, + + { + "7f", + 8, 0, 1, '.', FilterHTML, + "f) f<o\t<b>bar</b>\t\n", + "f) f<o..<b>bar</b>.....\n", + }, + + { + "7g", + 8, 0, 1, '.', FilterHTML, + "g) f<o\t<b>bar</b>\t non-terminated entity &", + "g) f<o..<b>bar</b>..... non-terminated entity &", + }, + + { + "7g debug", + 8, 0, 1, '.', FilterHTML | Debug, + "g) f<o\t<b>bar</b>\t non-terminated entity &", + "g) f<o..|<b>bar</b>.....| non-terminated entity &", + }, + + { + "8", + 8, 0, 1, '*', 0, + "Hello, world!\n", + "Hello, world!\n", + }, + + { + "9a", + 1, 0, 0, '.', 0, + "1\t2\t3\t4\n" + + "11\t222\t3333\t44444\n", + + "1.2..3...4\n" + + "11222333344444\n", + }, + + { + "9b", + 1, 0, 0, '.', FilterHTML, + "1\t2<!---\f--->\t3\t4\n" + // \f inside HTML is ignored + "11\t222\t3333\t44444\n", + + "1.2<!---\f--->..3...4\n" + + "11222333344444\n", + }, + + { + "9c", + 1, 0, 0, '.', 0, + "1\t2\t3\t4\f" + // \f causes a newline and flush + "11\t222\t3333\t44444\n", + + "1234\n" + + "11222333344444\n", + }, + + { + "9c debug", + 1, 0, 0, '.', Debug, + "1\t2\t3\t4\f" + // \f causes a newline and flush + "11\t222\t3333\t44444\n", + + "1|2|3|4\n" + + "---\n" + + "11|222|3333|44444\n", + }, + + { + "10a", + 5, 0, 0, '.', 0, + "1\t2\t3\t4\n", + "1....2....3....4\n", + }, + + { + "10b", + 5, 0, 0, '.', 0, + "1\t2\t3\t4\t\n", + "1....2....3....4....\n", + }, + + { + "11", + 8, 0, 1, '.', 0, + "本\tb\tc\n" + + "aa\t\u672c\u672c\u672c\tcccc\tddddd\n" + + "aaa\tbbbb\n", + + "本.......b.......c\n" + + "aa......本本本.....cccc....ddddd\n" + + "aaa.....bbbb\n", + }, + + { + "12a", + 8, 0, 1, ' ', AlignRight, + "a\tè\tc\t\n" + + "aa\tèèè\tcccc\tddddd\t\n" + + "aaa\tèèèè\t\n", + + " a è c\n" + + " aa èèè cccc ddddd\n" + + " aaa èèèè\n", + }, + + { + "12b", + 2, 0, 0, ' ', 0, + "a\tb\tc\n" + + "aa\tbbb\tcccc\n" + + "aaa\tbbbb\n", + + "a b c\n" + + "aa bbbcccc\n" + + "aaabbbb\n", + }, + + { + "12c", + 8, 0, 1, '_', 0, + "a\tb\tc\n" + + "aa\tbbb\tcccc\n" + + "aaa\tbbbb\n", + + "a_______b_______c\n" + + "aa______bbb_____cccc\n" + + "aaa_____bbbb\n", + }, + + { + "13a", + 4, 0, 1, '-', 0, + "4444\t日本語\t22\t1\t333\n" + + "999999999\t22\n" + + "7\t22\n" + + "\t\t\t88888888\n" + + "\n" + + "666666\t666666\t666666\t4444\n" + + "1\t1\t999999999\t0000000000\n", + + "4444------日本語-22--1---333\n" + + "999999999-22\n" + + "7---------22\n" + + "------------------88888888\n" + + "\n" + + "666666-666666-666666----4444\n" + + "1------1------999999999-0000000000\n", + }, + + { + "13b", + 4, 0, 3, '.', 0, + "4444\t333\t22\t1\t333\n" + + "999999999\t22\n" + + "7\t22\n" + + "\t\t\t88888888\n" + + "\n" + + "666666\t666666\t666666\t4444\n" + + "1\t1\t999999999\t0000000000\n", + + "4444........333...22...1...333\n" + + "999999999...22\n" + + "7...........22\n" + + "....................88888888\n" + + "\n" + + "666666...666666...666666......4444\n" + + "1........1........999999999...0000000000\n", + }, + + { + "13c", + 8, 8, 1, '\t', FilterHTML, + "4444\t333\t22\t1\t333\n" + + "999999999\t22\n" + + "7\t22\n" + + "\t\t\t88888888\n" + + "\n" + + "666666\t666666\t666666\t4444\n" + + "1\t1\t<font color=red attr=日本語>999999999</font>\t0000000000\n", + + "4444\t\t333\t22\t1\t333\n" + + "999999999\t22\n" + + "7\t\t22\n" + + "\t\t\t\t88888888\n" + + "\n" + + "666666\t666666\t666666\t\t4444\n" + + "1\t1\t<font color=red attr=日本語>999999999</font>\t0000000000\n", + }, + + { + "14", + 1, 0, 2, ' ', AlignRight, + ".0\t.3\t2.4\t-5.1\t\n" + + "23.0\t12345678.9\t2.4\t-989.4\t\n" + + "5.1\t12.0\t2.4\t-7.0\t\n" + + ".0\t0.0\t332.0\t8908.0\t\n" + + ".0\t-.3\t456.4\t22.1\t\n" + + ".0\t1.2\t44.4\t-13.3\t\t", + + " .0 .3 2.4 -5.1\n" + + " 23.0 12345678.9 2.4 -989.4\n" + + " 5.1 12.0 2.4 -7.0\n" + + " .0 0.0 332.0 8908.0\n" + + " .0 -.3 456.4 22.1\n" + + " .0 1.2 44.4 -13.3", + }, + + { + "14 debug", + 1, 0, 2, ' ', AlignRight | Debug, + ".0\t.3\t2.4\t-5.1\t\n" + + "23.0\t12345678.9\t2.4\t-989.4\t\n" + + "5.1\t12.0\t2.4\t-7.0\t\n" + + ".0\t0.0\t332.0\t8908.0\t\n" + + ".0\t-.3\t456.4\t22.1\t\n" + + ".0\t1.2\t44.4\t-13.3\t\t", + + " .0| .3| 2.4| -5.1|\n" + + " 23.0| 12345678.9| 2.4| -989.4|\n" + + " 5.1| 12.0| 2.4| -7.0|\n" + + " .0| 0.0| 332.0| 8908.0|\n" + + " .0| -.3| 456.4| 22.1|\n" + + " .0| 1.2| 44.4| -13.3|", + }, + + { + "15a", + 4, 0, 0, '.', 0, + "a\t\tb", + "a.......b", + }, + + { + "15b", + 4, 0, 0, '.', DiscardEmptyColumns, + "a\t\tb", // htabs - do not discard column + "a.......b", + }, + + { + "15c", + 4, 0, 0, '.', DiscardEmptyColumns, + "a\v\vb", + "a...b", + }, + + { + "15d", + 4, 0, 0, '.', AlignRight | DiscardEmptyColumns, + "a\v\vb", + "...ab", + }, + + { + "16a", + 100, 100, 0, '\t', 0, + "a\tb\t\td\n" + + "a\tb\t\td\te\n" + + "a\n" + + "a\tb\tc\td\n" + + "a\tb\tc\td\te\n", + + "a\tb\t\td\n" + + "a\tb\t\td\te\n" + + "a\n" + + "a\tb\tc\td\n" + + "a\tb\tc\td\te\n", + }, + + { + "16b", + 100, 100, 0, '\t', DiscardEmptyColumns, + "a\vb\v\vd\n" + + "a\vb\v\vd\ve\n" + + "a\n" + + "a\vb\vc\vd\n" + + "a\vb\vc\vd\ve\n", + + "a\tb\td\n" + + "a\tb\td\te\n" + + "a\n" + + "a\tb\tc\td\n" + + "a\tb\tc\td\te\n", + }, + + { + "16b debug", + 100, 100, 0, '\t', DiscardEmptyColumns | Debug, + "a\vb\v\vd\n" + + "a\vb\v\vd\ve\n" + + "a\n" + + "a\vb\vc\vd\n" + + "a\vb\vc\vd\ve\n", + + "a\t|b\t||d\n" + + "a\t|b\t||d\t|e\n" + + "a\n" + + "a\t|b\t|c\t|d\n" + + "a\t|b\t|c\t|d\t|e\n", + }, + + { + "16c", + 100, 100, 0, '\t', DiscardEmptyColumns, + "a\tb\t\td\n" + // hard tabs - do not discard column + "a\tb\t\td\te\n" + + "a\n" + + "a\tb\tc\td\n" + + "a\tb\tc\td\te\n", + + "a\tb\t\td\n" + + "a\tb\t\td\te\n" + + "a\n" + + "a\tb\tc\td\n" + + "a\tb\tc\td\te\n", + }, + + { + "16c debug", + 100, 100, 0, '\t', DiscardEmptyColumns | Debug, + "a\tb\t\td\n" + // hard tabs - do not discard column + "a\tb\t\td\te\n" + + "a\n" + + "a\tb\tc\td\n" + + "a\tb\tc\td\te\n", + + "a\t|b\t|\t|d\n" + + "a\t|b\t|\t|d\t|e\n" + + "a\n" + + "a\t|b\t|c\t|d\n" + + "a\t|b\t|c\t|d\t|e\n", + }, +} + +func Test(t *testing.T) { + for _, e := range tests { + check(t, e.testname, e.minwidth, e.tabwidth, e.padding, e.padchar, e.flags, e.src, e.expected) + } +} + +type panicWriter struct{} + +func (panicWriter) Write([]byte) (int, error) { + panic("cannot write") +} + +func wantPanicString(t *testing.T, want string) { + if e := recover(); e != nil { + got, ok := e.(string) + switch { + case !ok: + t.Errorf("got %v (%T), want panic string", e, e) + case got != want: + t.Errorf("wrong panic message: got %q, want %q", got, want) + } + } +} + +func TestPanicDuringFlush(t *testing.T) { + defer wantPanicString(t, "tabwriter: panic during Flush") + var p panicWriter + w := new(Writer) + w.Init(p, 0, 0, 5, ' ', 0) + io.WriteString(w, "a") + w.Flush() + t.Errorf("failed to panic during Flush") +} + +func TestPanicDuringWrite(t *testing.T) { + defer wantPanicString(t, "tabwriter: panic during Write") + var p panicWriter + w := new(Writer) + w.Init(p, 0, 0, 5, ' ', 0) + io.WriteString(w, "a\n\n") // the second \n triggers a call to w.Write and thus a panic + t.Errorf("failed to panic during Write") +} + +func BenchmarkTable(b *testing.B) { + for _, w := range [...]int{1, 10, 100} { + // Build a line with w cells. + line := bytes.Repeat([]byte("a\t"), w) + line = append(line, '\n') + for _, h := range [...]int{10, 1000, 100000} { + b.Run(fmt.Sprintf("%dx%d", w, h), func(b *testing.B) { + b.Run("new", func(b *testing.B) { + b.ReportAllocs() + for i := 0; i < b.N; i++ { + w := NewWriter(io.Discard, 4, 4, 1, ' ', 0) // no particular reason for these settings + // Write the line h times. + for j := 0; j < h; j++ { + w.Write(line) + } + w.Flush() + } + }) + + b.Run("reuse", func(b *testing.B) { + b.ReportAllocs() + w := NewWriter(io.Discard, 4, 4, 1, ' ', 0) // no particular reason for these settings + for i := 0; i < b.N; i++ { + // Write the line h times. + for j := 0; j < h; j++ { + w.Write(line) + } + w.Flush() + } + }) + }) + } + } +} + +func BenchmarkPyramid(b *testing.B) { + for _, x := range [...]int{10, 100, 1000} { + // Build a line with x cells. + line := bytes.Repeat([]byte("a\t"), x) + b.Run(fmt.Sprintf("%d", x), func(b *testing.B) { + b.ReportAllocs() + for i := 0; i < b.N; i++ { + w := NewWriter(io.Discard, 4, 4, 1, ' ', 0) // no particular reason for these settings + // Write increasing prefixes of that line. + for j := 0; j < x; j++ { + w.Write(line[:j*2]) + w.Write([]byte{'\n'}) + } + w.Flush() + } + }) + } +} + +func BenchmarkRagged(b *testing.B) { + var lines [8][]byte + for i, w := range [8]int{6, 2, 9, 5, 5, 7, 3, 8} { + // Build a line with w cells. + lines[i] = bytes.Repeat([]byte("a\t"), w) + } + for _, h := range [...]int{10, 100, 1000} { + b.Run(fmt.Sprintf("%d", h), func(b *testing.B) { + b.ReportAllocs() + for i := 0; i < b.N; i++ { + w := NewWriter(io.Discard, 4, 4, 1, ' ', 0) // no particular reason for these settings + // Write the lines in turn h times. + for j := 0; j < h; j++ { + w.Write(lines[j%len(lines)]) + w.Write([]byte{'\n'}) + } + w.Flush() + } + }) + } +} + +const codeSnippet = ` +some command + +foo # aligned +barbaz # comments + +but +mostly +single +cell +lines +` + +func BenchmarkCode(b *testing.B) { + b.ReportAllocs() + for i := 0; i < b.N; i++ { + w := NewWriter(io.Discard, 4, 4, 1, ' ', 0) // no particular reason for these settings + // The code is small, so it's reasonable for the tabwriter user + // to write it all at once, or buffer the writes. + w.Write([]byte(codeSnippet)) + w.Flush() + } +} diff --git a/src/text/template/doc.go b/src/text/template/doc.go new file mode 100644 index 0000000..7b30294 --- /dev/null +++ b/src/text/template/doc.go @@ -0,0 +1,455 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +/* +Package template implements data-driven templates for generating textual output. + +To generate HTML output, see package html/template, which has the same interface +as this package but automatically secures HTML output against certain attacks. + +Templates are executed by applying them to a data structure. Annotations in the +template refer to elements of the data structure (typically a field of a struct +or a key in a map) to control execution and derive values to be displayed. +Execution of the template walks the structure and sets the cursor, represented +by a period '.' and called "dot", to the value at the current location in the +structure as execution proceeds. + +The input text for a template is UTF-8-encoded text in any format. +"Actions"--data evaluations or control structures--are delimited by +"{{" and "}}"; all text outside actions is copied to the output unchanged. +Except for raw strings, actions may not span newlines, although comments can. + +Once parsed, a template may be executed safely in parallel, although if parallel +executions share a Writer the output may be interleaved. + +Here is a trivial example that prints "17 items are made of wool". + + type Inventory struct { + Material string + Count uint + } + sweaters := Inventory{"wool", 17} + tmpl, err := template.New("test").Parse("{{.Count}} items are made of {{.Material}}") + if err != nil { panic(err) } + err = tmpl.Execute(os.Stdout, sweaters) + if err != nil { panic(err) } + +More intricate examples appear below. + +Text and spaces + +By default, all text between actions is copied verbatim when the template is +executed. For example, the string " items are made of " in the example above +appears on standard output when the program is run. + +However, to aid in formatting template source code, if an action's left +delimiter (by default "{{") is followed immediately by a minus sign and white +space, all trailing white space is trimmed from the immediately preceding text. +Similarly, if the right delimiter ("}}") is preceded by white space and a minus +sign, all leading white space is trimmed from the immediately following text. +In these trim markers, the white space must be present: +"{{- 3}}" is like "{{3}}" but trims the immediately preceding text, while +"{{-3}}" parses as an action containing the number -3. + +For instance, when executing the template whose source is + + "{{23 -}} < {{- 45}}" + +the generated output would be + + "23<45" + +For this trimming, the definition of white space characters is the same as in Go: +space, horizontal tab, carriage return, and newline. + +Actions + +Here is the list of actions. "Arguments" and "pipelines" are evaluations of +data, defined in detail in the corresponding sections that follow. + +*/ +// {{/* a comment */}} +// {{- /* a comment with white space trimmed from preceding and following text */ -}} +// A comment; discarded. May contain newlines. +// Comments do not nest and must start and end at the +// delimiters, as shown here. +/* + + {{pipeline}} + The default textual representation (the same as would be + printed by fmt.Print) of the value of the pipeline is copied + to the output. + + {{if pipeline}} T1 {{end}} + If the value of the pipeline is empty, no output is generated; + otherwise, T1 is executed. The empty values are false, 0, any + nil pointer or interface value, and any array, slice, map, or + string of length zero. + Dot is unaffected. + + {{if pipeline}} T1 {{else}} T0 {{end}} + If the value of the pipeline is empty, T0 is executed; + otherwise, T1 is executed. Dot is unaffected. + + {{if pipeline}} T1 {{else if pipeline}} T0 {{end}} + To simplify the appearance of if-else chains, the else action + of an if may include another if directly; the effect is exactly + the same as writing + {{if pipeline}} T1 {{else}}{{if pipeline}} T0 {{end}}{{end}} + + {{range pipeline}} T1 {{end}} + The value of the pipeline must be an array, slice, map, or channel. + If the value of the pipeline has length zero, nothing is output; + otherwise, dot is set to the successive elements of the array, + slice, or map and T1 is executed. If the value is a map and the + keys are of basic type with a defined order, the elements will be + visited in sorted key order. + + {{range pipeline}} T1 {{else}} T0 {{end}} + The value of the pipeline must be an array, slice, map, or channel. + If the value of the pipeline has length zero, dot is unaffected and + T0 is executed; otherwise, dot is set to the successive elements + of the array, slice, or map and T1 is executed. + + {{template "name"}} + The template with the specified name is executed with nil data. + + {{template "name" pipeline}} + The template with the specified name is executed with dot set + to the value of the pipeline. + + {{block "name" pipeline}} T1 {{end}} + A block is shorthand for defining a template + {{define "name"}} T1 {{end}} + and then executing it in place + {{template "name" pipeline}} + The typical use is to define a set of root templates that are + then customized by redefining the block templates within. + + {{with pipeline}} T1 {{end}} + If the value of the pipeline is empty, no output is generated; + otherwise, dot is set to the value of the pipeline and T1 is + executed. + + {{with pipeline}} T1 {{else}} T0 {{end}} + If the value of the pipeline is empty, dot is unaffected and T0 + is executed; otherwise, dot is set to the value of the pipeline + and T1 is executed. + +Arguments + +An argument is a simple value, denoted by one of the following. + + - A boolean, string, character, integer, floating-point, imaginary + or complex constant in Go syntax. These behave like Go's untyped + constants. Note that, as in Go, whether a large integer constant + overflows when assigned or passed to a function can depend on whether + the host machine's ints are 32 or 64 bits. + - The keyword nil, representing an untyped Go nil. + - The character '.' (period): + . + The result is the value of dot. + - A variable name, which is a (possibly empty) alphanumeric string + preceded by a dollar sign, such as + $piOver2 + or + $ + The result is the value of the variable. + Variables are described below. + - The name of a field of the data, which must be a struct, preceded + by a period, such as + .Field + The result is the value of the field. Field invocations may be + chained: + .Field1.Field2 + Fields can also be evaluated on variables, including chaining: + $x.Field1.Field2 + - The name of a key of the data, which must be a map, preceded + by a period, such as + .Key + The result is the map element value indexed by the key. + Key invocations may be chained and combined with fields to any + depth: + .Field1.Key1.Field2.Key2 + Although the key must be an alphanumeric identifier, unlike with + field names they do not need to start with an upper case letter. + Keys can also be evaluated on variables, including chaining: + $x.key1.key2 + - The name of a niladic method of the data, preceded by a period, + such as + .Method + The result is the value of invoking the method with dot as the + receiver, dot.Method(). Such a method must have one return value (of + any type) or two return values, the second of which is an error. + If it has two and the returned error is non-nil, execution terminates + and an error is returned to the caller as the value of Execute. + Method invocations may be chained and combined with fields and keys + to any depth: + .Field1.Key1.Method1.Field2.Key2.Method2 + Methods can also be evaluated on variables, including chaining: + $x.Method1.Field + - The name of a niladic function, such as + fun + The result is the value of invoking the function, fun(). The return + types and values behave as in methods. Functions and function + names are described below. + - A parenthesized instance of one the above, for grouping. The result + may be accessed by a field or map key invocation. + print (.F1 arg1) (.F2 arg2) + (.StructValuedMethod "arg").Field + +Arguments may evaluate to any type; if they are pointers the implementation +automatically indirects to the base type when required. +If an evaluation yields a function value, such as a function-valued +field of a struct, the function is not invoked automatically, but it +can be used as a truth value for an if action and the like. To invoke +it, use the call function, defined below. + +Pipelines + +A pipeline is a possibly chained sequence of "commands". A command is a simple +value (argument) or a function or method call, possibly with multiple arguments: + + Argument + The result is the value of evaluating the argument. + .Method [Argument...] + The method can be alone or the last element of a chain but, + unlike methods in the middle of a chain, it can take arguments. + The result is the value of calling the method with the + arguments: + dot.Method(Argument1, etc.) + functionName [Argument...] + The result is the value of calling the function associated + with the name: + function(Argument1, etc.) + Functions and function names are described below. + +A pipeline may be "chained" by separating a sequence of commands with pipeline +characters '|'. In a chained pipeline, the result of each command is +passed as the last argument of the following command. The output of the final +command in the pipeline is the value of the pipeline. + +The output of a command will be either one value or two values, the second of +which has type error. If that second value is present and evaluates to +non-nil, execution terminates and the error is returned to the caller of +Execute. + +Variables + +A pipeline inside an action may initialize a variable to capture the result. +The initialization has syntax + + $variable := pipeline + +where $variable is the name of the variable. An action that declares a +variable produces no output. + +Variables previously declared can also be assigned, using the syntax + + $variable = pipeline + +If a "range" action initializes a variable, the variable is set to the +successive elements of the iteration. Also, a "range" may declare two +variables, separated by a comma: + + range $index, $element := pipeline + +in which case $index and $element are set to the successive values of the +array/slice index or map key and element, respectively. Note that if there is +only one variable, it is assigned the element; this is opposite to the +convention in Go range clauses. + +A variable's scope extends to the "end" action of the control structure ("if", +"with", or "range") in which it is declared, or to the end of the template if +there is no such control structure. A template invocation does not inherit +variables from the point of its invocation. + +When execution begins, $ is set to the data argument passed to Execute, that is, +to the starting value of dot. + +Examples + +Here are some example one-line templates demonstrating pipelines and variables. +All produce the quoted word "output": + + {{"\"output\""}} + A string constant. + {{`"output"`}} + A raw string constant. + {{printf "%q" "output"}} + A function call. + {{"output" | printf "%q"}} + A function call whose final argument comes from the previous + command. + {{printf "%q" (print "out" "put")}} + A parenthesized argument. + {{"put" | printf "%s%s" "out" | printf "%q"}} + A more elaborate call. + {{"output" | printf "%s" | printf "%q"}} + A longer chain. + {{with "output"}}{{printf "%q" .}}{{end}} + A with action using dot. + {{with $x := "output" | printf "%q"}}{{$x}}{{end}} + A with action that creates and uses a variable. + {{with $x := "output"}}{{printf "%q" $x}}{{end}} + A with action that uses the variable in another action. + {{with $x := "output"}}{{$x | printf "%q"}}{{end}} + The same, but pipelined. + +Functions + +During execution functions are found in two function maps: first in the +template, then in the global function map. By default, no functions are defined +in the template but the Funcs method can be used to add them. + +Predefined global functions are named as follows. + + and + Returns the boolean AND of its arguments by returning the + first empty argument or the last argument, that is, + "and x y" behaves as "if x then y else x". All the + arguments are evaluated. + call + Returns the result of calling the first argument, which + must be a function, with the remaining arguments as parameters. + Thus "call .X.Y 1 2" is, in Go notation, dot.X.Y(1, 2) where + Y is a func-valued field, map entry, or the like. + The first argument must be the result of an evaluation + that yields a value of function type (as distinct from + a predefined function such as print). The function must + return either one or two result values, the second of which + is of type error. If the arguments don't match the function + or the returned error value is non-nil, execution stops. + html + Returns the escaped HTML equivalent of the textual + representation of its arguments. This function is unavailable + in html/template, with a few exceptions. + index + Returns the result of indexing its first argument by the + following arguments. Thus "index x 1 2 3" is, in Go syntax, + x[1][2][3]. Each indexed item must be a map, slice, or array. + slice + slice returns the result of slicing its first argument by the + remaining arguments. Thus "slice x 1 2" is, in Go syntax, x[1:2], + while "slice x" is x[:], "slice x 1" is x[1:], and "slice x 1 2 3" + is x[1:2:3]. The first argument must be a string, slice, or array. + js + Returns the escaped JavaScript equivalent of the textual + representation of its arguments. + len + Returns the integer length of its argument. + not + Returns the boolean negation of its single argument. + or + Returns the boolean OR of its arguments by returning the + first non-empty argument or the last argument, that is, + "or x y" behaves as "if x then x else y". All the + arguments are evaluated. + print + An alias for fmt.Sprint + printf + An alias for fmt.Sprintf + println + An alias for fmt.Sprintln + urlquery + Returns the escaped value of the textual representation of + its arguments in a form suitable for embedding in a URL query. + This function is unavailable in html/template, with a few + exceptions. + +The boolean functions take any zero value to be false and a non-zero +value to be true. + +There is also a set of binary comparison operators defined as +functions: + + eq + Returns the boolean truth of arg1 == arg2 + ne + Returns the boolean truth of arg1 != arg2 + lt + Returns the boolean truth of arg1 < arg2 + le + Returns the boolean truth of arg1 <= arg2 + gt + Returns the boolean truth of arg1 > arg2 + ge + Returns the boolean truth of arg1 >= arg2 + +For simpler multi-way equality tests, eq (only) accepts two or more +arguments and compares the second and subsequent to the first, +returning in effect + + arg1==arg2 || arg1==arg3 || arg1==arg4 ... + +(Unlike with || in Go, however, eq is a function call and all the +arguments will be evaluated.) + +The comparison functions work on any values whose type Go defines as +comparable. For basic types such as integers, the rules are relaxed: +size and exact type are ignored, so any integer value, signed or unsigned, +may be compared with any other integer value. (The arithmetic value is compared, +not the bit pattern, so all negative integers are less than all unsigned integers.) +However, as usual, one may not compare an int with a float32 and so on. + +Associated templates + +Each template is named by a string specified when it is created. Also, each +template is associated with zero or more other templates that it may invoke by +name; such associations are transitive and form a name space of templates. + +A template may use a template invocation to instantiate another associated +template; see the explanation of the "template" action above. The name must be +that of a template associated with the template that contains the invocation. + +Nested template definitions + +When parsing a template, another template may be defined and associated with the +template being parsed. Template definitions must appear at the top level of the +template, much like global variables in a Go program. + +The syntax of such definitions is to surround each template declaration with a +"define" and "end" action. + +The define action names the template being created by providing a string +constant. Here is a simple example: + + `{{define "T1"}}ONE{{end}} + {{define "T2"}}TWO{{end}} + {{define "T3"}}{{template "T1"}} {{template "T2"}}{{end}} + {{template "T3"}}` + +This defines two templates, T1 and T2, and a third T3 that invokes the other two +when it is executed. Finally it invokes T3. If executed this template will +produce the text + + ONE TWO + +By construction, a template may reside in only one association. If it's +necessary to have a template addressable from multiple associations, the +template definition must be parsed multiple times to create distinct *Template +values, or must be copied with the Clone or AddParseTree method. + +Parse may be called multiple times to assemble the various associated templates; +see the ParseFiles and ParseGlob functions and methods for simple ways to parse +related templates stored in files. + +A template may be executed directly or through ExecuteTemplate, which executes +an associated template identified by name. To invoke our example above, we +might write, + + err := tmpl.Execute(os.Stdout, "no data needed") + if err != nil { + log.Fatalf("execution failed: %s", err) + } + +or to invoke a particular template explicitly by name, + + err := tmpl.ExecuteTemplate(os.Stdout, "T2", "no data needed") + if err != nil { + log.Fatalf("execution failed: %s", err) + } + +*/ +package template diff --git a/src/text/template/example_test.go b/src/text/template/example_test.go new file mode 100644 index 0000000..9cab2e8 --- /dev/null +++ b/src/text/template/example_test.go @@ -0,0 +1,110 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package template_test + +import ( + "log" + "os" + "strings" + "text/template" +) + +func ExampleTemplate() { + // Define a template. + const letter = ` +Dear {{.Name}}, +{{if .Attended}} +It was a pleasure to see you at the wedding. +{{- else}} +It is a shame you couldn't make it to the wedding. +{{- end}} +{{with .Gift -}} +Thank you for the lovely {{.}}. +{{end}} +Best wishes, +Josie +` + + // Prepare some data to insert into the template. + type Recipient struct { + Name, Gift string + Attended bool + } + var recipients = []Recipient{ + {"Aunt Mildred", "bone china tea set", true}, + {"Uncle John", "moleskin pants", false}, + {"Cousin Rodney", "", false}, + } + + // Create a new template and parse the letter into it. + t := template.Must(template.New("letter").Parse(letter)) + + // Execute the template for each recipient. + for _, r := range recipients { + err := t.Execute(os.Stdout, r) + if err != nil { + log.Println("executing template:", err) + } + } + + // Output: + // Dear Aunt Mildred, + // + // It was a pleasure to see you at the wedding. + // Thank you for the lovely bone china tea set. + // + // Best wishes, + // Josie + // + // Dear Uncle John, + // + // It is a shame you couldn't make it to the wedding. + // Thank you for the lovely moleskin pants. + // + // Best wishes, + // Josie + // + // Dear Cousin Rodney, + // + // It is a shame you couldn't make it to the wedding. + // + // Best wishes, + // Josie +} + +// The following example is duplicated in html/template; keep them in sync. + +func ExampleTemplate_block() { + const ( + master = `Names:{{block "list" .}}{{"\n"}}{{range .}}{{println "-" .}}{{end}}{{end}}` + overlay = `{{define "list"}} {{join . ", "}}{{end}} ` + ) + var ( + funcs = template.FuncMap{"join": strings.Join} + guardians = []string{"Gamora", "Groot", "Nebula", "Rocket", "Star-Lord"} + ) + masterTmpl, err := template.New("master").Funcs(funcs).Parse(master) + if err != nil { + log.Fatal(err) + } + overlayTmpl, err := template.Must(masterTmpl.Clone()).Parse(overlay) + if err != nil { + log.Fatal(err) + } + if err := masterTmpl.Execute(os.Stdout, guardians); err != nil { + log.Fatal(err) + } + if err := overlayTmpl.Execute(os.Stdout, guardians); err != nil { + log.Fatal(err) + } + // Output: + // Names: + // - Gamora + // - Groot + // - Nebula + // - Rocket + // - Star-Lord + // Names: Gamora, Groot, Nebula, Rocket, Star-Lord +} diff --git a/src/text/template/examplefiles_test.go b/src/text/template/examplefiles_test.go new file mode 100644 index 0000000..6534ee3 --- /dev/null +++ b/src/text/template/examplefiles_test.go @@ -0,0 +1,181 @@ +// Copyright 2012 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package template_test + +import ( + "io" + "log" + "os" + "path/filepath" + "text/template" +) + +// templateFile defines the contents of a template to be stored in a file, for testing. +type templateFile struct { + name string + contents string +} + +func createTestDir(files []templateFile) string { + dir, err := os.MkdirTemp("", "template") + if err != nil { + log.Fatal(err) + } + for _, file := range files { + f, err := os.Create(filepath.Join(dir, file.name)) + if err != nil { + log.Fatal(err) + } + defer f.Close() + _, err = io.WriteString(f, file.contents) + if err != nil { + log.Fatal(err) + } + } + return dir +} + +// Here we demonstrate loading a set of templates from a directory. +func ExampleTemplate_glob() { + // Here we create a temporary directory and populate it with our sample + // template definition files; usually the template files would already + // exist in some location known to the program. + dir := createTestDir([]templateFile{ + // T0.tmpl is a plain template file that just invokes T1. + {"T0.tmpl", `T0 invokes T1: ({{template "T1"}})`}, + // T1.tmpl defines a template, T1 that invokes T2. + {"T1.tmpl", `{{define "T1"}}T1 invokes T2: ({{template "T2"}}){{end}}`}, + // T2.tmpl defines a template T2. + {"T2.tmpl", `{{define "T2"}}This is T2{{end}}`}, + }) + // Clean up after the test; another quirk of running as an example. + defer os.RemoveAll(dir) + + // pattern is the glob pattern used to find all the template files. + pattern := filepath.Join(dir, "*.tmpl") + + // Here starts the example proper. + // T0.tmpl is the first name matched, so it becomes the starting template, + // the value returned by ParseGlob. + tmpl := template.Must(template.ParseGlob(pattern)) + + err := tmpl.Execute(os.Stdout, nil) + if err != nil { + log.Fatalf("template execution: %s", err) + } + // Output: + // T0 invokes T1: (T1 invokes T2: (This is T2)) +} + +// This example demonstrates one way to share some templates +// and use them in different contexts. In this variant we add multiple driver +// templates by hand to an existing bundle of templates. +func ExampleTemplate_helpers() { + // Here we create a temporary directory and populate it with our sample + // template definition files; usually the template files would already + // exist in some location known to the program. + dir := createTestDir([]templateFile{ + // T1.tmpl defines a template, T1 that invokes T2. + {"T1.tmpl", `{{define "T1"}}T1 invokes T2: ({{template "T2"}}){{end}}`}, + // T2.tmpl defines a template T2. + {"T2.tmpl", `{{define "T2"}}This is T2{{end}}`}, + }) + // Clean up after the test; another quirk of running as an example. + defer os.RemoveAll(dir) + + // pattern is the glob pattern used to find all the template files. + pattern := filepath.Join(dir, "*.tmpl") + + // Here starts the example proper. + // Load the helpers. + templates := template.Must(template.ParseGlob(pattern)) + // Add one driver template to the bunch; we do this with an explicit template definition. + _, err := templates.Parse("{{define `driver1`}}Driver 1 calls T1: ({{template `T1`}})\n{{end}}") + if err != nil { + log.Fatal("parsing driver1: ", err) + } + // Add another driver template. + _, err = templates.Parse("{{define `driver2`}}Driver 2 calls T2: ({{template `T2`}})\n{{end}}") + if err != nil { + log.Fatal("parsing driver2: ", err) + } + // We load all the templates before execution. This package does not require + // that behavior but html/template's escaping does, so it's a good habit. + err = templates.ExecuteTemplate(os.Stdout, "driver1", nil) + if err != nil { + log.Fatalf("driver1 execution: %s", err) + } + err = templates.ExecuteTemplate(os.Stdout, "driver2", nil) + if err != nil { + log.Fatalf("driver2 execution: %s", err) + } + // Output: + // Driver 1 calls T1: (T1 invokes T2: (This is T2)) + // Driver 2 calls T2: (This is T2) +} + +// This example demonstrates how to use one group of driver +// templates with distinct sets of helper templates. +func ExampleTemplate_share() { + // Here we create a temporary directory and populate it with our sample + // template definition files; usually the template files would already + // exist in some location known to the program. + dir := createTestDir([]templateFile{ + // T0.tmpl is a plain template file that just invokes T1. + {"T0.tmpl", "T0 ({{.}} version) invokes T1: ({{template `T1`}})\n"}, + // T1.tmpl defines a template, T1 that invokes T2. Note T2 is not defined + {"T1.tmpl", `{{define "T1"}}T1 invokes T2: ({{template "T2"}}){{end}}`}, + }) + // Clean up after the test; another quirk of running as an example. + defer os.RemoveAll(dir) + + // pattern is the glob pattern used to find all the template files. + pattern := filepath.Join(dir, "*.tmpl") + + // Here starts the example proper. + // Load the drivers. + drivers := template.Must(template.ParseGlob(pattern)) + + // We must define an implementation of the T2 template. First we clone + // the drivers, then add a definition of T2 to the template name space. + + // 1. Clone the helper set to create a new name space from which to run them. + first, err := drivers.Clone() + if err != nil { + log.Fatal("cloning helpers: ", err) + } + // 2. Define T2, version A, and parse it. + _, err = first.Parse("{{define `T2`}}T2, version A{{end}}") + if err != nil { + log.Fatal("parsing T2: ", err) + } + + // Now repeat the whole thing, using a different version of T2. + // 1. Clone the drivers. + second, err := drivers.Clone() + if err != nil { + log.Fatal("cloning drivers: ", err) + } + // 2. Define T2, version B, and parse it. + _, err = second.Parse("{{define `T2`}}T2, version B{{end}}") + if err != nil { + log.Fatal("parsing T2: ", err) + } + + // Execute the templates in the reverse order to verify the + // first is unaffected by the second. + err = second.ExecuteTemplate(os.Stdout, "T0.tmpl", "second") + if err != nil { + log.Fatalf("second execution: %s", err) + } + err = first.ExecuteTemplate(os.Stdout, "T0.tmpl", "first") + if err != nil { + log.Fatalf("first: execution: %s", err) + } + + // Output: + // T0 (second version) invokes T1: (T1 invokes T2: (T2, version B)) + // T0 (first version) invokes T1: (T1 invokes T2: (T2, version A)) +} diff --git a/src/text/template/examplefunc_test.go b/src/text/template/examplefunc_test.go new file mode 100644 index 0000000..080b5e3 --- /dev/null +++ b/src/text/template/examplefunc_test.go @@ -0,0 +1,54 @@ +// Copyright 2012 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package template_test + +import ( + "log" + "os" + "strings" + "text/template" +) + +// This example demonstrates a custom function to process template text. +// It installs the strings.Title function and uses it to +// Make Title Text Look Good In Our Template's Output. +func ExampleTemplate_func() { + // First we create a FuncMap with which to register the function. + funcMap := template.FuncMap{ + // The name "title" is what the function will be called in the template text. + "title": strings.Title, + } + + // A simple template definition to test our function. + // We print the input text several ways: + // - the original + // - title-cased + // - title-cased and then printed with %q + // - printed with %q and then title-cased. + const templateText = ` +Input: {{printf "%q" .}} +Output 0: {{title .}} +Output 1: {{title . | printf "%q"}} +Output 2: {{printf "%q" . | title}} +` + + // Create a template, add the function map, and parse the text. + tmpl, err := template.New("titleTest").Funcs(funcMap).Parse(templateText) + if err != nil { + log.Fatalf("parsing: %s", err) + } + + // Run the template to verify the output. + err = tmpl.Execute(os.Stdout, "the go programming language") + if err != nil { + log.Fatalf("execution: %s", err) + } + + // Output: + // Input: "the go programming language" + // Output 0: The Go Programming Language + // Output 1: "The Go Programming Language" + // Output 2: "The Go Programming Language" +} diff --git a/src/text/template/exec.go b/src/text/template/exec.go new file mode 100644 index 0000000..5ad3b4e --- /dev/null +++ b/src/text/template/exec.go @@ -0,0 +1,991 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package template + +import ( + "fmt" + "internal/fmtsort" + "io" + "reflect" + "runtime" + "strings" + "text/template/parse" +) + +// maxExecDepth specifies the maximum stack depth of templates within +// templates. This limit is only practically reached by accidentally +// recursive template invocations. This limit allows us to return +// an error instead of triggering a stack overflow. +var maxExecDepth = initMaxExecDepth() + +func initMaxExecDepth() int { + if runtime.GOARCH == "wasm" { + return 1000 + } + return 100000 +} + +// state represents the state of an execution. It's not part of the +// template so that multiple executions of the same template +// can execute in parallel. +type state struct { + tmpl *Template + wr io.Writer + node parse.Node // current node, for errors + vars []variable // push-down stack of variable values. + depth int // the height of the stack of executing templates. +} + +// variable holds the dynamic value of a variable such as $, $x etc. +type variable struct { + name string + value reflect.Value +} + +// push pushes a new variable on the stack. +func (s *state) push(name string, value reflect.Value) { + s.vars = append(s.vars, variable{name, value}) +} + +// mark returns the length of the variable stack. +func (s *state) mark() int { + return len(s.vars) +} + +// pop pops the variable stack up to the mark. +func (s *state) pop(mark int) { + s.vars = s.vars[0:mark] +} + +// setVar overwrites the last declared variable with the given name. +// Used by variable assignments. +func (s *state) setVar(name string, value reflect.Value) { + for i := s.mark() - 1; i >= 0; i-- { + if s.vars[i].name == name { + s.vars[i].value = value + return + } + } + s.errorf("undefined variable: %s", name) +} + +// setTopVar overwrites the top-nth variable on the stack. Used by range iterations. +func (s *state) setTopVar(n int, value reflect.Value) { + s.vars[len(s.vars)-n].value = value +} + +// varValue returns the value of the named variable. +func (s *state) varValue(name string) reflect.Value { + for i := s.mark() - 1; i >= 0; i-- { + if s.vars[i].name == name { + return s.vars[i].value + } + } + s.errorf("undefined variable: %s", name) + return zero +} + +var zero reflect.Value + +type missingValType struct{} + +var missingVal = reflect.ValueOf(missingValType{}) + +// at marks the state to be on node n, for error reporting. +func (s *state) at(node parse.Node) { + s.node = node +} + +// doublePercent returns the string with %'s replaced by %%, if necessary, +// so it can be used safely inside a Printf format string. +func doublePercent(str string) string { + return strings.ReplaceAll(str, "%", "%%") +} + +// TODO: It would be nice if ExecError was more broken down, but +// the way ErrorContext embeds the template name makes the +// processing too clumsy. + +// ExecError is the custom error type returned when Execute has an +// error evaluating its template. (If a write error occurs, the actual +// error is returned; it will not be of type ExecError.) +type ExecError struct { + Name string // Name of template. + Err error // Pre-formatted error. +} + +func (e ExecError) Error() string { + return e.Err.Error() +} + +func (e ExecError) Unwrap() error { + return e.Err +} + +// errorf records an ExecError and terminates processing. +func (s *state) errorf(format string, args ...interface{}) { + name := doublePercent(s.tmpl.Name()) + if s.node == nil { + format = fmt.Sprintf("template: %s: %s", name, format) + } else { + location, context := s.tmpl.ErrorContext(s.node) + format = fmt.Sprintf("template: %s: executing %q at <%s>: %s", location, name, doublePercent(context), format) + } + panic(ExecError{ + Name: s.tmpl.Name(), + Err: fmt.Errorf(format, args...), + }) +} + +// writeError is the wrapper type used internally when Execute has an +// error writing to its output. We strip the wrapper in errRecover. +// Note that this is not an implementation of error, so it cannot escape +// from the package as an error value. +type writeError struct { + Err error // Original error. +} + +func (s *state) writeError(err error) { + panic(writeError{ + Err: err, + }) +} + +// errRecover is the handler that turns panics into returns from the top +// level of Parse. +func errRecover(errp *error) { + e := recover() + if e != nil { + switch err := e.(type) { + case runtime.Error: + panic(e) + case writeError: + *errp = err.Err // Strip the wrapper. + case ExecError: + *errp = err // Keep the wrapper. + default: + panic(e) + } + } +} + +// ExecuteTemplate applies the template associated with t that has the given name +// to the specified data object and writes the output to wr. +// If an error occurs executing the template or writing its output, +// execution stops, but partial results may already have been written to +// the output writer. +// A template may be executed safely in parallel, although if parallel +// executions share a Writer the output may be interleaved. +func (t *Template) ExecuteTemplate(wr io.Writer, name string, data interface{}) error { + tmpl := t.Lookup(name) + if tmpl == nil { + return fmt.Errorf("template: no template %q associated with template %q", name, t.name) + } + return tmpl.Execute(wr, data) +} + +// Execute applies a parsed template to the specified data object, +// and writes the output to wr. +// If an error occurs executing the template or writing its output, +// execution stops, but partial results may already have been written to +// the output writer. +// A template may be executed safely in parallel, although if parallel +// executions share a Writer the output may be interleaved. +// +// If data is a reflect.Value, the template applies to the concrete +// value that the reflect.Value holds, as in fmt.Print. +func (t *Template) Execute(wr io.Writer, data interface{}) error { + return t.execute(wr, data) +} + +func (t *Template) execute(wr io.Writer, data interface{}) (err error) { + defer errRecover(&err) + value, ok := data.(reflect.Value) + if !ok { + value = reflect.ValueOf(data) + } + state := &state{ + tmpl: t, + wr: wr, + vars: []variable{{"$", value}}, + } + if t.Tree == nil || t.Root == nil { + state.errorf("%q is an incomplete or empty template", t.Name()) + } + state.walk(value, t.Root) + return +} + +// DefinedTemplates returns a string listing the defined templates, +// prefixed by the string "; defined templates are: ". If there are none, +// it returns the empty string. For generating an error message here +// and in html/template. +func (t *Template) DefinedTemplates() string { + if t.common == nil { + return "" + } + var b strings.Builder + t.muTmpl.RLock() + defer t.muTmpl.RUnlock() + for name, tmpl := range t.tmpl { + if tmpl.Tree == nil || tmpl.Root == nil { + continue + } + if b.Len() == 0 { + b.WriteString("; defined templates are: ") + } else { + b.WriteString(", ") + } + fmt.Fprintf(&b, "%q", name) + } + return b.String() +} + +// Walk functions step through the major pieces of the template structure, +// generating output as they go. +func (s *state) walk(dot reflect.Value, node parse.Node) { + s.at(node) + switch node := node.(type) { + case *parse.ActionNode: + // Do not pop variables so they persist until next end. + // Also, if the action declares variables, don't print the result. + val := s.evalPipeline(dot, node.Pipe) + if len(node.Pipe.Decl) == 0 { + s.printValue(node, val) + } + case *parse.CommentNode: + case *parse.IfNode: + s.walkIfOrWith(parse.NodeIf, dot, node.Pipe, node.List, node.ElseList) + case *parse.ListNode: + for _, node := range node.Nodes { + s.walk(dot, node) + } + case *parse.RangeNode: + s.walkRange(dot, node) + case *parse.TemplateNode: + s.walkTemplate(dot, node) + case *parse.TextNode: + if _, err := s.wr.Write(node.Text); err != nil { + s.writeError(err) + } + case *parse.WithNode: + s.walkIfOrWith(parse.NodeWith, dot, node.Pipe, node.List, node.ElseList) + default: + s.errorf("unknown node: %s", node) + } +} + +// walkIfOrWith walks an 'if' or 'with' node. The two control structures +// are identical in behavior except that 'with' sets dot. +func (s *state) walkIfOrWith(typ parse.NodeType, dot reflect.Value, pipe *parse.PipeNode, list, elseList *parse.ListNode) { + defer s.pop(s.mark()) + val := s.evalPipeline(dot, pipe) + truth, ok := isTrue(indirectInterface(val)) + if !ok { + s.errorf("if/with can't use %v", val) + } + if truth { + if typ == parse.NodeWith { + s.walk(val, list) + } else { + s.walk(dot, list) + } + } else if elseList != nil { + s.walk(dot, elseList) + } +} + +// IsTrue reports whether the value is 'true', in the sense of not the zero of its type, +// and whether the value has a meaningful truth value. This is the definition of +// truth used by if and other such actions. +func IsTrue(val interface{}) (truth, ok bool) { + return isTrue(reflect.ValueOf(val)) +} + +func isTrue(val reflect.Value) (truth, ok bool) { + if !val.IsValid() { + // Something like var x interface{}, never set. It's a form of nil. + return false, true + } + switch val.Kind() { + case reflect.Array, reflect.Map, reflect.Slice, reflect.String: + truth = val.Len() > 0 + case reflect.Bool: + truth = val.Bool() + case reflect.Complex64, reflect.Complex128: + truth = val.Complex() != 0 + case reflect.Chan, reflect.Func, reflect.Ptr, reflect.Interface: + truth = !val.IsNil() + case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64: + truth = val.Int() != 0 + case reflect.Float32, reflect.Float64: + truth = val.Float() != 0 + case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr: + truth = val.Uint() != 0 + case reflect.Struct: + truth = true // Struct values are always true. + default: + return + } + return truth, true +} + +func (s *state) walkRange(dot reflect.Value, r *parse.RangeNode) { + s.at(r) + defer s.pop(s.mark()) + val, _ := indirect(s.evalPipeline(dot, r.Pipe)) + // mark top of stack before any variables in the body are pushed. + mark := s.mark() + oneIteration := func(index, elem reflect.Value) { + // Set top var (lexically the second if there are two) to the element. + if len(r.Pipe.Decl) > 0 { + s.setTopVar(1, elem) + } + // Set next var (lexically the first if there are two) to the index. + if len(r.Pipe.Decl) > 1 { + s.setTopVar(2, index) + } + s.walk(elem, r.List) + s.pop(mark) + } + switch val.Kind() { + case reflect.Array, reflect.Slice: + if val.Len() == 0 { + break + } + for i := 0; i < val.Len(); i++ { + oneIteration(reflect.ValueOf(i), val.Index(i)) + } + return + case reflect.Map: + if val.Len() == 0 { + break + } + om := fmtsort.Sort(val) + for i, key := range om.Key { + oneIteration(key, om.Value[i]) + } + return + case reflect.Chan: + if val.IsNil() { + break + } + if val.Type().ChanDir() == reflect.SendDir { + s.errorf("range over send-only channel %v", val) + break + } + i := 0 + for ; ; i++ { + elem, ok := val.Recv() + if !ok { + break + } + oneIteration(reflect.ValueOf(i), elem) + } + if i == 0 { + break + } + return + case reflect.Invalid: + break // An invalid value is likely a nil map, etc. and acts like an empty map. + default: + s.errorf("range can't iterate over %v", val) + } + if r.ElseList != nil { + s.walk(dot, r.ElseList) + } +} + +func (s *state) walkTemplate(dot reflect.Value, t *parse.TemplateNode) { + s.at(t) + tmpl := s.tmpl.Lookup(t.Name) + if tmpl == nil { + s.errorf("template %q not defined", t.Name) + } + if s.depth == maxExecDepth { + s.errorf("exceeded maximum template depth (%v)", maxExecDepth) + } + // Variables declared by the pipeline persist. + dot = s.evalPipeline(dot, t.Pipe) + newState := *s + newState.depth++ + newState.tmpl = tmpl + // No dynamic scoping: template invocations inherit no variables. + newState.vars = []variable{{"$", dot}} + newState.walk(dot, tmpl.Root) +} + +// Eval functions evaluate pipelines, commands, and their elements and extract +// values from the data structure by examining fields, calling methods, and so on. +// The printing of those values happens only through walk functions. + +// evalPipeline returns the value acquired by evaluating a pipeline. If the +// pipeline has a variable declaration, the variable will be pushed on the +// stack. Callers should therefore pop the stack after they are finished +// executing commands depending on the pipeline value. +func (s *state) evalPipeline(dot reflect.Value, pipe *parse.PipeNode) (value reflect.Value) { + if pipe == nil { + return + } + s.at(pipe) + value = missingVal + for _, cmd := range pipe.Cmds { + value = s.evalCommand(dot, cmd, value) // previous value is this one's final arg. + // If the object has type interface{}, dig down one level to the thing inside. + if value.Kind() == reflect.Interface && value.Type().NumMethod() == 0 { + value = reflect.ValueOf(value.Interface()) // lovely! + } + } + for _, variable := range pipe.Decl { + if pipe.IsAssign { + s.setVar(variable.Ident[0], value) + } else { + s.push(variable.Ident[0], value) + } + } + return value +} + +func (s *state) notAFunction(args []parse.Node, final reflect.Value) { + if len(args) > 1 || final != missingVal { + s.errorf("can't give argument to non-function %s", args[0]) + } +} + +func (s *state) evalCommand(dot reflect.Value, cmd *parse.CommandNode, final reflect.Value) reflect.Value { + firstWord := cmd.Args[0] + switch n := firstWord.(type) { + case *parse.FieldNode: + return s.evalFieldNode(dot, n, cmd.Args, final) + case *parse.ChainNode: + return s.evalChainNode(dot, n, cmd.Args, final) + case *parse.IdentifierNode: + // Must be a function. + return s.evalFunction(dot, n, cmd, cmd.Args, final) + case *parse.PipeNode: + // Parenthesized pipeline. The arguments are all inside the pipeline; final must be absent. + s.notAFunction(cmd.Args, final) + return s.evalPipeline(dot, n) + case *parse.VariableNode: + return s.evalVariableNode(dot, n, cmd.Args, final) + } + s.at(firstWord) + s.notAFunction(cmd.Args, final) + switch word := firstWord.(type) { + case *parse.BoolNode: + return reflect.ValueOf(word.True) + case *parse.DotNode: + return dot + case *parse.NilNode: + s.errorf("nil is not a command") + case *parse.NumberNode: + return s.idealConstant(word) + case *parse.StringNode: + return reflect.ValueOf(word.Text) + } + s.errorf("can't evaluate command %q", firstWord) + panic("not reached") +} + +// idealConstant is called to return the value of a number in a context where +// we don't know the type. In that case, the syntax of the number tells us +// its type, and we use Go rules to resolve. Note there is no such thing as +// a uint ideal constant in this situation - the value must be of int type. +func (s *state) idealConstant(constant *parse.NumberNode) reflect.Value { + // These are ideal constants but we don't know the type + // and we have no context. (If it was a method argument, + // we'd know what we need.) The syntax guides us to some extent. + s.at(constant) + switch { + case constant.IsComplex: + return reflect.ValueOf(constant.Complex128) // incontrovertible. + + case constant.IsFloat && + !isHexInt(constant.Text) && !isRuneInt(constant.Text) && + strings.ContainsAny(constant.Text, ".eEpP"): + return reflect.ValueOf(constant.Float64) + + case constant.IsInt: + n := int(constant.Int64) + if int64(n) != constant.Int64 { + s.errorf("%s overflows int", constant.Text) + } + return reflect.ValueOf(n) + + case constant.IsUint: + s.errorf("%s overflows int", constant.Text) + } + return zero +} + +func isRuneInt(s string) bool { + return len(s) > 0 && s[0] == '\'' +} + +func isHexInt(s string) bool { + return len(s) > 2 && s[0] == '0' && (s[1] == 'x' || s[1] == 'X') && !strings.ContainsAny(s, "pP") +} + +func (s *state) evalFieldNode(dot reflect.Value, field *parse.FieldNode, args []parse.Node, final reflect.Value) reflect.Value { + s.at(field) + return s.evalFieldChain(dot, dot, field, field.Ident, args, final) +} + +func (s *state) evalChainNode(dot reflect.Value, chain *parse.ChainNode, args []parse.Node, final reflect.Value) reflect.Value { + s.at(chain) + if len(chain.Field) == 0 { + s.errorf("internal error: no fields in evalChainNode") + } + if chain.Node.Type() == parse.NodeNil { + s.errorf("indirection through explicit nil in %s", chain) + } + // (pipe).Field1.Field2 has pipe as .Node, fields as .Field. Eval the pipeline, then the fields. + pipe := s.evalArg(dot, nil, chain.Node) + return s.evalFieldChain(dot, pipe, chain, chain.Field, args, final) +} + +func (s *state) evalVariableNode(dot reflect.Value, variable *parse.VariableNode, args []parse.Node, final reflect.Value) reflect.Value { + // $x.Field has $x as the first ident, Field as the second. Eval the var, then the fields. + s.at(variable) + value := s.varValue(variable.Ident[0]) + if len(variable.Ident) == 1 { + s.notAFunction(args, final) + return value + } + return s.evalFieldChain(dot, value, variable, variable.Ident[1:], args, final) +} + +// evalFieldChain evaluates .X.Y.Z possibly followed by arguments. +// dot is the environment in which to evaluate arguments, while +// receiver is the value being walked along the chain. +func (s *state) evalFieldChain(dot, receiver reflect.Value, node parse.Node, ident []string, args []parse.Node, final reflect.Value) reflect.Value { + n := len(ident) + for i := 0; i < n-1; i++ { + receiver = s.evalField(dot, ident[i], node, nil, missingVal, receiver) + } + // Now if it's a method, it gets the arguments. + return s.evalField(dot, ident[n-1], node, args, final, receiver) +} + +func (s *state) evalFunction(dot reflect.Value, node *parse.IdentifierNode, cmd parse.Node, args []parse.Node, final reflect.Value) reflect.Value { + s.at(node) + name := node.Ident + function, ok := findFunction(name, s.tmpl) + if !ok { + s.errorf("%q is not a defined function", name) + } + return s.evalCall(dot, function, cmd, name, args, final) +} + +// evalField evaluates an expression like (.Field) or (.Field arg1 arg2). +// The 'final' argument represents the return value from the preceding +// value of the pipeline, if any. +func (s *state) evalField(dot reflect.Value, fieldName string, node parse.Node, args []parse.Node, final, receiver reflect.Value) reflect.Value { + if !receiver.IsValid() { + if s.tmpl.option.missingKey == mapError { // Treat invalid value as missing map key. + s.errorf("nil data; no entry for key %q", fieldName) + } + return zero + } + typ := receiver.Type() + receiver, isNil := indirect(receiver) + if receiver.Kind() == reflect.Interface && isNil { + // Calling a method on a nil interface can't work. The + // MethodByName method call below would panic. + s.errorf("nil pointer evaluating %s.%s", typ, fieldName) + return zero + } + + // Unless it's an interface, need to get to a value of type *T to guarantee + // we see all methods of T and *T. + ptr := receiver + if ptr.Kind() != reflect.Interface && ptr.Kind() != reflect.Ptr && ptr.CanAddr() { + ptr = ptr.Addr() + } + if method := ptr.MethodByName(fieldName); method.IsValid() { + return s.evalCall(dot, method, node, fieldName, args, final) + } + hasArgs := len(args) > 1 || final != missingVal + // It's not a method; must be a field of a struct or an element of a map. + switch receiver.Kind() { + case reflect.Struct: + tField, ok := receiver.Type().FieldByName(fieldName) + if ok { + field := receiver.FieldByIndex(tField.Index) + if !tField.IsExported() { + s.errorf("%s is an unexported field of struct type %s", fieldName, typ) + } + // If it's a function, we must call it. + if hasArgs { + s.errorf("%s has arguments but cannot be invoked as function", fieldName) + } + return field + } + case reflect.Map: + // If it's a map, attempt to use the field name as a key. + nameVal := reflect.ValueOf(fieldName) + if nameVal.Type().AssignableTo(receiver.Type().Key()) { + if hasArgs { + s.errorf("%s is not a method but has arguments", fieldName) + } + result := receiver.MapIndex(nameVal) + if !result.IsValid() { + switch s.tmpl.option.missingKey { + case mapInvalid: + // Just use the invalid value. + case mapZeroValue: + result = reflect.Zero(receiver.Type().Elem()) + case mapError: + s.errorf("map has no entry for key %q", fieldName) + } + } + return result + } + case reflect.Ptr: + etyp := receiver.Type().Elem() + if etyp.Kind() == reflect.Struct { + if _, ok := etyp.FieldByName(fieldName); !ok { + // If there's no such field, say "can't evaluate" + // instead of "nil pointer evaluating". + break + } + } + if isNil { + s.errorf("nil pointer evaluating %s.%s", typ, fieldName) + } + } + s.errorf("can't evaluate field %s in type %s", fieldName, typ) + panic("not reached") +} + +var ( + errorType = reflect.TypeOf((*error)(nil)).Elem() + fmtStringerType = reflect.TypeOf((*fmt.Stringer)(nil)).Elem() + reflectValueType = reflect.TypeOf((*reflect.Value)(nil)).Elem() +) + +// evalCall executes a function or method call. If it's a method, fun already has the receiver bound, so +// it looks just like a function call. The arg list, if non-nil, includes (in the manner of the shell), arg[0] +// as the function itself. +func (s *state) evalCall(dot, fun reflect.Value, node parse.Node, name string, args []parse.Node, final reflect.Value) reflect.Value { + if args != nil { + args = args[1:] // Zeroth arg is function name/node; not passed to function. + } + typ := fun.Type() + numIn := len(args) + if final != missingVal { + numIn++ + } + numFixed := len(args) + if typ.IsVariadic() { + numFixed = typ.NumIn() - 1 // last arg is the variadic one. + if numIn < numFixed { + s.errorf("wrong number of args for %s: want at least %d got %d", name, typ.NumIn()-1, len(args)) + } + } else if numIn != typ.NumIn() { + s.errorf("wrong number of args for %s: want %d got %d", name, typ.NumIn(), numIn) + } + if !goodFunc(typ) { + // TODO: This could still be a confusing error; maybe goodFunc should provide info. + s.errorf("can't call method/function %q with %d results", name, typ.NumOut()) + } + // Build the arg list. + argv := make([]reflect.Value, numIn) + // Args must be evaluated. Fixed args first. + i := 0 + for ; i < numFixed && i < len(args); i++ { + argv[i] = s.evalArg(dot, typ.In(i), args[i]) + } + // Now the ... args. + if typ.IsVariadic() { + argType := typ.In(typ.NumIn() - 1).Elem() // Argument is a slice. + for ; i < len(args); i++ { + argv[i] = s.evalArg(dot, argType, args[i]) + } + } + // Add final value if necessary. + if final != missingVal { + t := typ.In(typ.NumIn() - 1) + if typ.IsVariadic() { + if numIn-1 < numFixed { + // The added final argument corresponds to a fixed parameter of the function. + // Validate against the type of the actual parameter. + t = typ.In(numIn - 1) + } else { + // The added final argument corresponds to the variadic part. + // Validate against the type of the elements of the variadic slice. + t = t.Elem() + } + } + argv[i] = s.validateType(final, t) + } + v, err := safeCall(fun, argv) + // If we have an error that is not nil, stop execution and return that + // error to the caller. + if err != nil { + s.at(node) + s.errorf("error calling %s: %w", name, err) + } + if v.Type() == reflectValueType { + v = v.Interface().(reflect.Value) + } + return v +} + +// canBeNil reports whether an untyped nil can be assigned to the type. See reflect.Zero. +func canBeNil(typ reflect.Type) bool { + switch typ.Kind() { + case reflect.Chan, reflect.Func, reflect.Interface, reflect.Map, reflect.Ptr, reflect.Slice: + return true + case reflect.Struct: + return typ == reflectValueType + } + return false +} + +// validateType guarantees that the value is valid and assignable to the type. +func (s *state) validateType(value reflect.Value, typ reflect.Type) reflect.Value { + if !value.IsValid() { + if typ == nil { + // An untyped nil interface{}. Accept as a proper nil value. + return reflect.ValueOf(nil) + } + if canBeNil(typ) { + // Like above, but use the zero value of the non-nil type. + return reflect.Zero(typ) + } + s.errorf("invalid value; expected %s", typ) + } + if typ == reflectValueType && value.Type() != typ { + return reflect.ValueOf(value) + } + if typ != nil && !value.Type().AssignableTo(typ) { + if value.Kind() == reflect.Interface && !value.IsNil() { + value = value.Elem() + if value.Type().AssignableTo(typ) { + return value + } + // fallthrough + } + // Does one dereference or indirection work? We could do more, as we + // do with method receivers, but that gets messy and method receivers + // are much more constrained, so it makes more sense there than here. + // Besides, one is almost always all you need. + switch { + case value.Kind() == reflect.Ptr && value.Type().Elem().AssignableTo(typ): + value = value.Elem() + if !value.IsValid() { + s.errorf("dereference of nil pointer of type %s", typ) + } + case reflect.PtrTo(value.Type()).AssignableTo(typ) && value.CanAddr(): + value = value.Addr() + default: + s.errorf("wrong type for value; expected %s; got %s", typ, value.Type()) + } + } + return value +} + +func (s *state) evalArg(dot reflect.Value, typ reflect.Type, n parse.Node) reflect.Value { + s.at(n) + switch arg := n.(type) { + case *parse.DotNode: + return s.validateType(dot, typ) + case *parse.NilNode: + if canBeNil(typ) { + return reflect.Zero(typ) + } + s.errorf("cannot assign nil to %s", typ) + case *parse.FieldNode: + return s.validateType(s.evalFieldNode(dot, arg, []parse.Node{n}, missingVal), typ) + case *parse.VariableNode: + return s.validateType(s.evalVariableNode(dot, arg, nil, missingVal), typ) + case *parse.PipeNode: + return s.validateType(s.evalPipeline(dot, arg), typ) + case *parse.IdentifierNode: + return s.validateType(s.evalFunction(dot, arg, arg, nil, missingVal), typ) + case *parse.ChainNode: + return s.validateType(s.evalChainNode(dot, arg, nil, missingVal), typ) + } + switch typ.Kind() { + case reflect.Bool: + return s.evalBool(typ, n) + case reflect.Complex64, reflect.Complex128: + return s.evalComplex(typ, n) + case reflect.Float32, reflect.Float64: + return s.evalFloat(typ, n) + case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64: + return s.evalInteger(typ, n) + case reflect.Interface: + if typ.NumMethod() == 0 { + return s.evalEmptyInterface(dot, n) + } + case reflect.Struct: + if typ == reflectValueType { + return reflect.ValueOf(s.evalEmptyInterface(dot, n)) + } + case reflect.String: + return s.evalString(typ, n) + case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr: + return s.evalUnsignedInteger(typ, n) + } + s.errorf("can't handle %s for arg of type %s", n, typ) + panic("not reached") +} + +func (s *state) evalBool(typ reflect.Type, n parse.Node) reflect.Value { + s.at(n) + if n, ok := n.(*parse.BoolNode); ok { + value := reflect.New(typ).Elem() + value.SetBool(n.True) + return value + } + s.errorf("expected bool; found %s", n) + panic("not reached") +} + +func (s *state) evalString(typ reflect.Type, n parse.Node) reflect.Value { + s.at(n) + if n, ok := n.(*parse.StringNode); ok { + value := reflect.New(typ).Elem() + value.SetString(n.Text) + return value + } + s.errorf("expected string; found %s", n) + panic("not reached") +} + +func (s *state) evalInteger(typ reflect.Type, n parse.Node) reflect.Value { + s.at(n) + if n, ok := n.(*parse.NumberNode); ok && n.IsInt { + value := reflect.New(typ).Elem() + value.SetInt(n.Int64) + return value + } + s.errorf("expected integer; found %s", n) + panic("not reached") +} + +func (s *state) evalUnsignedInteger(typ reflect.Type, n parse.Node) reflect.Value { + s.at(n) + if n, ok := n.(*parse.NumberNode); ok && n.IsUint { + value := reflect.New(typ).Elem() + value.SetUint(n.Uint64) + return value + } + s.errorf("expected unsigned integer; found %s", n) + panic("not reached") +} + +func (s *state) evalFloat(typ reflect.Type, n parse.Node) reflect.Value { + s.at(n) + if n, ok := n.(*parse.NumberNode); ok && n.IsFloat { + value := reflect.New(typ).Elem() + value.SetFloat(n.Float64) + return value + } + s.errorf("expected float; found %s", n) + panic("not reached") +} + +func (s *state) evalComplex(typ reflect.Type, n parse.Node) reflect.Value { + if n, ok := n.(*parse.NumberNode); ok && n.IsComplex { + value := reflect.New(typ).Elem() + value.SetComplex(n.Complex128) + return value + } + s.errorf("expected complex; found %s", n) + panic("not reached") +} + +func (s *state) evalEmptyInterface(dot reflect.Value, n parse.Node) reflect.Value { + s.at(n) + switch n := n.(type) { + case *parse.BoolNode: + return reflect.ValueOf(n.True) + case *parse.DotNode: + return dot + case *parse.FieldNode: + return s.evalFieldNode(dot, n, nil, missingVal) + case *parse.IdentifierNode: + return s.evalFunction(dot, n, n, nil, missingVal) + case *parse.NilNode: + // NilNode is handled in evalArg, the only place that calls here. + s.errorf("evalEmptyInterface: nil (can't happen)") + case *parse.NumberNode: + return s.idealConstant(n) + case *parse.StringNode: + return reflect.ValueOf(n.Text) + case *parse.VariableNode: + return s.evalVariableNode(dot, n, nil, missingVal) + case *parse.PipeNode: + return s.evalPipeline(dot, n) + } + s.errorf("can't handle assignment of %s to empty interface argument", n) + panic("not reached") +} + +// indirect returns the item at the end of indirection, and a bool to indicate +// if it's nil. If the returned bool is true, the returned value's kind will be +// either a pointer or interface. +func indirect(v reflect.Value) (rv reflect.Value, isNil bool) { + for ; v.Kind() == reflect.Ptr || v.Kind() == reflect.Interface; v = v.Elem() { + if v.IsNil() { + return v, true + } + } + return v, false +} + +// indirectInterface returns the concrete value in an interface value, +// or else the zero reflect.Value. +// That is, if v represents the interface value x, the result is the same as reflect.ValueOf(x): +// the fact that x was an interface value is forgotten. +func indirectInterface(v reflect.Value) reflect.Value { + if v.Kind() != reflect.Interface { + return v + } + if v.IsNil() { + return reflect.Value{} + } + return v.Elem() +} + +// printValue writes the textual representation of the value to the output of +// the template. +func (s *state) printValue(n parse.Node, v reflect.Value) { + s.at(n) + iface, ok := printableValue(v) + if !ok { + s.errorf("can't print %s of type %s", n, v.Type()) + } + _, err := fmt.Fprint(s.wr, iface) + if err != nil { + s.writeError(err) + } +} + +// printableValue returns the, possibly indirected, interface value inside v that +// is best for a call to formatted printer. +func printableValue(v reflect.Value) (interface{}, bool) { + if v.Kind() == reflect.Ptr { + v, _ = indirect(v) // fmt.Fprint handles nil. + } + if !v.IsValid() { + return "<no value>", true + } + + if !v.Type().Implements(errorType) && !v.Type().Implements(fmtStringerType) { + if v.CanAddr() && (reflect.PtrTo(v.Type()).Implements(errorType) || reflect.PtrTo(v.Type()).Implements(fmtStringerType)) { + v = v.Addr() + } else { + switch v.Kind() { + case reflect.Chan, reflect.Func: + return nil, false + } + } + } + return v.Interface(), true +} diff --git a/src/text/template/exec_test.go b/src/text/template/exec_test.go new file mode 100644 index 0000000..ef52164 --- /dev/null +++ b/src/text/template/exec_test.go @@ -0,0 +1,1775 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package template + +import ( + "bytes" + "errors" + "flag" + "fmt" + "io" + "reflect" + "strings" + "sync" + "testing" +) + +var debug = flag.Bool("debug", false, "show the errors produced by the tests") + +// T has lots of interesting pieces to use to test execution. +type T struct { + // Basics + True bool + I int + U16 uint16 + X, S string + FloatZero float64 + ComplexZero complex128 + // Nested structs. + U *U + // Struct with String method. + V0 V + V1, V2 *V + // Struct with Error method. + W0 W + W1, W2 *W + // Slices + SI []int + SICap []int + SIEmpty []int + SB []bool + // Arrays + AI [3]int + // Maps + MSI map[string]int + MSIone map[string]int // one element, for deterministic output + MSIEmpty map[string]int + MXI map[interface{}]int + MII map[int]int + MI32S map[int32]string + MI64S map[int64]string + MUI32S map[uint32]string + MUI64S map[uint64]string + MI8S map[int8]string + MUI8S map[uint8]string + SMSI []map[string]int + // Empty interfaces; used to see if we can dig inside one. + Empty0 interface{} // nil + Empty1 interface{} + Empty2 interface{} + Empty3 interface{} + Empty4 interface{} + // Non-empty interfaces. + NonEmptyInterface I + NonEmptyInterfacePtS *I + NonEmptyInterfaceNil I + NonEmptyInterfaceTypedNil I + // Stringer. + Str fmt.Stringer + Err error + // Pointers + PI *int + PS *string + PSI *[]int + NIL *int + // Function (not method) + BinaryFunc func(string, string) string + VariadicFunc func(...string) string + VariadicFuncInt func(int, ...string) string + NilOKFunc func(*int) bool + ErrFunc func() (string, error) + PanicFunc func() string + // Template to test evaluation of templates. + Tmpl *Template + // Unexported field; cannot be accessed by template. + unexported int +} + +type S []string + +func (S) Method0() string { + return "M0" +} + +type U struct { + V string +} + +type V struct { + j int +} + +func (v *V) String() string { + if v == nil { + return "nilV" + } + return fmt.Sprintf("<%d>", v.j) +} + +type W struct { + k int +} + +func (w *W) Error() string { + if w == nil { + return "nilW" + } + return fmt.Sprintf("[%d]", w.k) +} + +var siVal = I(S{"a", "b"}) + +var tVal = &T{ + True: true, + I: 17, + U16: 16, + X: "x", + S: "xyz", + U: &U{"v"}, + V0: V{6666}, + V1: &V{7777}, // leave V2 as nil + W0: W{888}, + W1: &W{999}, // leave W2 as nil + SI: []int{3, 4, 5}, + SICap: make([]int, 5, 10), + AI: [3]int{3, 4, 5}, + SB: []bool{true, false}, + MSI: map[string]int{"one": 1, "two": 2, "three": 3}, + MSIone: map[string]int{"one": 1}, + MXI: map[interface{}]int{"one": 1}, + MII: map[int]int{1: 1}, + MI32S: map[int32]string{1: "one", 2: "two"}, + MI64S: map[int64]string{2: "i642", 3: "i643"}, + MUI32S: map[uint32]string{2: "u322", 3: "u323"}, + MUI64S: map[uint64]string{2: "ui642", 3: "ui643"}, + MI8S: map[int8]string{2: "i82", 3: "i83"}, + MUI8S: map[uint8]string{2: "u82", 3: "u83"}, + SMSI: []map[string]int{ + {"one": 1, "two": 2}, + {"eleven": 11, "twelve": 12}, + }, + Empty1: 3, + Empty2: "empty2", + Empty3: []int{7, 8}, + Empty4: &U{"UinEmpty"}, + NonEmptyInterface: &T{X: "x"}, + NonEmptyInterfacePtS: &siVal, + NonEmptyInterfaceTypedNil: (*T)(nil), + Str: bytes.NewBuffer([]byte("foozle")), + Err: errors.New("erroozle"), + PI: newInt(23), + PS: newString("a string"), + PSI: newIntSlice(21, 22, 23), + BinaryFunc: func(a, b string) string { return fmt.Sprintf("[%s=%s]", a, b) }, + VariadicFunc: func(s ...string) string { return fmt.Sprint("<", strings.Join(s, "+"), ">") }, + VariadicFuncInt: func(a int, s ...string) string { return fmt.Sprint(a, "=<", strings.Join(s, "+"), ">") }, + NilOKFunc: func(s *int) bool { return s == nil }, + ErrFunc: func() (string, error) { return "bla", nil }, + PanicFunc: func() string { panic("test panic") }, + Tmpl: Must(New("x").Parse("test template")), // "x" is the value of .X +} + +var tSliceOfNil = []*T{nil} + +// A non-empty interface. +type I interface { + Method0() string +} + +var iVal I = tVal + +// Helpers for creation. +func newInt(n int) *int { + return &n +} + +func newString(s string) *string { + return &s +} + +func newIntSlice(n ...int) *[]int { + p := new([]int) + *p = make([]int, len(n)) + copy(*p, n) + return p +} + +// Simple methods with and without arguments. +func (t *T) Method0() string { + return "M0" +} + +func (t *T) Method1(a int) int { + return a +} + +func (t *T) Method2(a uint16, b string) string { + return fmt.Sprintf("Method2: %d %s", a, b) +} + +func (t *T) Method3(v interface{}) string { + return fmt.Sprintf("Method3: %v", v) +} + +func (t *T) Copy() *T { + n := new(T) + *n = *t + return n +} + +func (t *T) MAdd(a int, b []int) []int { + v := make([]int, len(b)) + for i, x := range b { + v[i] = x + a + } + return v +} + +var myError = errors.New("my error") + +// MyError returns a value and an error according to its argument. +func (t *T) MyError(error bool) (bool, error) { + if error { + return true, myError + } + return false, nil +} + +// A few methods to test chaining. +func (t *T) GetU() *U { + return t.U +} + +func (u *U) TrueFalse(b bool) string { + if b { + return "true" + } + return "" +} + +func typeOf(arg interface{}) string { + return fmt.Sprintf("%T", arg) +} + +type execTest struct { + name string + input string + output string + data interface{} + ok bool +} + +// bigInt and bigUint are hex string representing numbers either side +// of the max int boundary. +// We do it this way so the test doesn't depend on ints being 32 bits. +var ( + bigInt = fmt.Sprintf("0x%x", int(1<<uint(reflect.TypeOf(0).Bits()-1)-1)) + bigUint = fmt.Sprintf("0x%x", uint(1<<uint(reflect.TypeOf(0).Bits()-1))) +) + +var execTests = []execTest{ + // Trivial cases. + {"empty", "", "", nil, true}, + {"text", "some text", "some text", nil, true}, + {"nil action", "{{nil}}", "", nil, false}, + + // Ideal constants. + {"ideal int", "{{typeOf 3}}", "int", 0, true}, + {"ideal float", "{{typeOf 1.0}}", "float64", 0, true}, + {"ideal exp float", "{{typeOf 1e1}}", "float64", 0, true}, + {"ideal complex", "{{typeOf 1i}}", "complex128", 0, true}, + {"ideal int", "{{typeOf " + bigInt + "}}", "int", 0, true}, + {"ideal too big", "{{typeOf " + bigUint + "}}", "", 0, false}, + {"ideal nil without type", "{{nil}}", "", 0, false}, + + // Fields of structs. + {".X", "-{{.X}}-", "-x-", tVal, true}, + {".U.V", "-{{.U.V}}-", "-v-", tVal, true}, + {".unexported", "{{.unexported}}", "", tVal, false}, + + // Fields on maps. + {"map .one", "{{.MSI.one}}", "1", tVal, true}, + {"map .two", "{{.MSI.two}}", "2", tVal, true}, + {"map .NO", "{{.MSI.NO}}", "<no value>", tVal, true}, + {"map .one interface", "{{.MXI.one}}", "1", tVal, true}, + {"map .WRONG args", "{{.MSI.one 1}}", "", tVal, false}, + {"map .WRONG type", "{{.MII.one}}", "", tVal, false}, + + // Dots of all kinds to test basic evaluation. + {"dot int", "<{{.}}>", "<13>", 13, true}, + {"dot uint", "<{{.}}>", "<14>", uint(14), true}, + {"dot float", "<{{.}}>", "<15.1>", 15.1, true}, + {"dot bool", "<{{.}}>", "<true>", true, true}, + {"dot complex", "<{{.}}>", "<(16.2-17i)>", 16.2 - 17i, true}, + {"dot string", "<{{.}}>", "<hello>", "hello", true}, + {"dot slice", "<{{.}}>", "<[-1 -2 -3]>", []int{-1, -2, -3}, true}, + {"dot map", "<{{.}}>", "<map[two:22]>", map[string]int{"two": 22}, true}, + {"dot struct", "<{{.}}>", "<{7 seven}>", struct { + a int + b string + }{7, "seven"}, true}, + + // Variables. + {"$ int", "{{$}}", "123", 123, true}, + {"$.I", "{{$.I}}", "17", tVal, true}, + {"$.U.V", "{{$.U.V}}", "v", tVal, true}, + {"declare in action", "{{$x := $.U.V}}{{$x}}", "v", tVal, true}, + {"simple assignment", "{{$x := 2}}{{$x = 3}}{{$x}}", "3", tVal, true}, + {"nested assignment", + "{{$x := 2}}{{if true}}{{$x = 3}}{{end}}{{$x}}", + "3", tVal, true}, + {"nested assignment changes the last declaration", + "{{$x := 1}}{{if true}}{{$x := 2}}{{if true}}{{$x = 3}}{{end}}{{end}}{{$x}}", + "1", tVal, true}, + + // Type with String method. + {"V{6666}.String()", "-{{.V0}}-", "-<6666>-", tVal, true}, + {"&V{7777}.String()", "-{{.V1}}-", "-<7777>-", tVal, true}, + {"(*V)(nil).String()", "-{{.V2}}-", "-nilV-", tVal, true}, + + // Type with Error method. + {"W{888}.Error()", "-{{.W0}}-", "-[888]-", tVal, true}, + {"&W{999}.Error()", "-{{.W1}}-", "-[999]-", tVal, true}, + {"(*W)(nil).Error()", "-{{.W2}}-", "-nilW-", tVal, true}, + + // Pointers. + {"*int", "{{.PI}}", "23", tVal, true}, + {"*string", "{{.PS}}", "a string", tVal, true}, + {"*[]int", "{{.PSI}}", "[21 22 23]", tVal, true}, + {"*[]int[1]", "{{index .PSI 1}}", "22", tVal, true}, + {"NIL", "{{.NIL}}", "<nil>", tVal, true}, + + // Empty interfaces holding values. + {"empty nil", "{{.Empty0}}", "<no value>", tVal, true}, + {"empty with int", "{{.Empty1}}", "3", tVal, true}, + {"empty with string", "{{.Empty2}}", "empty2", tVal, true}, + {"empty with slice", "{{.Empty3}}", "[7 8]", tVal, true}, + {"empty with struct", "{{.Empty4}}", "{UinEmpty}", tVal, true}, + {"empty with struct, field", "{{.Empty4.V}}", "UinEmpty", tVal, true}, + + // Edge cases with <no value> with an interface value + {"field on interface", "{{.foo}}", "<no value>", nil, true}, + {"field on parenthesized interface", "{{(.).foo}}", "<no value>", nil, true}, + + // Issue 31810: Parenthesized first element of pipeline with arguments. + // See also TestIssue31810. + {"unparenthesized non-function", "{{1 2}}", "", nil, false}, + {"parenthesized non-function", "{{(1) 2}}", "", nil, false}, + {"parenthesized non-function with no args", "{{(1)}}", "1", nil, true}, // This is fine. + + // Method calls. + {".Method0", "-{{.Method0}}-", "-M0-", tVal, true}, + {".Method1(1234)", "-{{.Method1 1234}}-", "-1234-", tVal, true}, + {".Method1(.I)", "-{{.Method1 .I}}-", "-17-", tVal, true}, + {".Method2(3, .X)", "-{{.Method2 3 .X}}-", "-Method2: 3 x-", tVal, true}, + {".Method2(.U16, `str`)", "-{{.Method2 .U16 `str`}}-", "-Method2: 16 str-", tVal, true}, + {".Method2(.U16, $x)", "{{if $x := .X}}-{{.Method2 .U16 $x}}{{end}}-", "-Method2: 16 x-", tVal, true}, + {".Method3(nil constant)", "-{{.Method3 nil}}-", "-Method3: <nil>-", tVal, true}, + {".Method3(nil value)", "-{{.Method3 .MXI.unset}}-", "-Method3: <nil>-", tVal, true}, + {"method on var", "{{if $x := .}}-{{$x.Method2 .U16 $x.X}}{{end}}-", "-Method2: 16 x-", tVal, true}, + {"method on chained var", + "{{range .MSIone}}{{if $.U.TrueFalse $.True}}{{$.U.TrueFalse $.True}}{{else}}WRONG{{end}}{{end}}", + "true", tVal, true}, + {"chained method", + "{{range .MSIone}}{{if $.GetU.TrueFalse $.True}}{{$.U.TrueFalse $.True}}{{else}}WRONG{{end}}{{end}}", + "true", tVal, true}, + {"chained method on variable", + "{{with $x := .}}{{with .SI}}{{$.GetU.TrueFalse $.True}}{{end}}{{end}}", + "true", tVal, true}, + {".NilOKFunc not nil", "{{call .NilOKFunc .PI}}", "false", tVal, true}, + {".NilOKFunc nil", "{{call .NilOKFunc nil}}", "true", tVal, true}, + {"method on nil value from slice", "-{{range .}}{{.Method1 1234}}{{end}}-", "-1234-", tSliceOfNil, true}, + {"method on typed nil interface value", "{{.NonEmptyInterfaceTypedNil.Method0}}", "M0", tVal, true}, + + // Function call builtin. + {".BinaryFunc", "{{call .BinaryFunc `1` `2`}}", "[1=2]", tVal, true}, + {".VariadicFunc0", "{{call .VariadicFunc}}", "<>", tVal, true}, + {".VariadicFunc2", "{{call .VariadicFunc `he` `llo`}}", "<he+llo>", tVal, true}, + {".VariadicFuncInt", "{{call .VariadicFuncInt 33 `he` `llo`}}", "33=<he+llo>", tVal, true}, + {"if .BinaryFunc call", "{{ if .BinaryFunc}}{{call .BinaryFunc `1` `2`}}{{end}}", "[1=2]", tVal, true}, + {"if not .BinaryFunc call", "{{ if not .BinaryFunc}}{{call .BinaryFunc `1` `2`}}{{else}}No{{end}}", "No", tVal, true}, + {"Interface Call", `{{stringer .S}}`, "foozle", map[string]interface{}{"S": bytes.NewBufferString("foozle")}, true}, + {".ErrFunc", "{{call .ErrFunc}}", "bla", tVal, true}, + {"call nil", "{{call nil}}", "", tVal, false}, + + // Erroneous function calls (check args). + {".BinaryFuncTooFew", "{{call .BinaryFunc `1`}}", "", tVal, false}, + {".BinaryFuncTooMany", "{{call .BinaryFunc `1` `2` `3`}}", "", tVal, false}, + {".BinaryFuncBad0", "{{call .BinaryFunc 1 3}}", "", tVal, false}, + {".BinaryFuncBad1", "{{call .BinaryFunc `1` 3}}", "", tVal, false}, + {".VariadicFuncBad0", "{{call .VariadicFunc 3}}", "", tVal, false}, + {".VariadicFuncIntBad0", "{{call .VariadicFuncInt}}", "", tVal, false}, + {".VariadicFuncIntBad`", "{{call .VariadicFuncInt `x`}}", "", tVal, false}, + {".VariadicFuncNilBad", "{{call .VariadicFunc nil}}", "", tVal, false}, + + // Pipelines. + {"pipeline", "-{{.Method0 | .Method2 .U16}}-", "-Method2: 16 M0-", tVal, true}, + {"pipeline func", "-{{call .VariadicFunc `llo` | call .VariadicFunc `he` }}-", "-<he+<llo>>-", tVal, true}, + + // Nil values aren't missing arguments. + {"nil pipeline", "{{ .Empty0 | call .NilOKFunc }}", "true", tVal, true}, + {"nil call arg", "{{ call .NilOKFunc .Empty0 }}", "true", tVal, true}, + {"bad nil pipeline", "{{ .Empty0 | .VariadicFunc }}", "", tVal, false}, + + // Parenthesized expressions + {"parens in pipeline", "{{printf `%d %d %d` (1) (2 | add 3) (add 4 (add 5 6))}}", "1 5 15", tVal, true}, + + // Parenthesized expressions with field accesses + {"parens: $ in paren", "{{($).X}}", "x", tVal, true}, + {"parens: $.GetU in paren", "{{($.GetU).V}}", "v", tVal, true}, + {"parens: $ in paren in pipe", "{{($ | echo).X}}", "x", tVal, true}, + {"parens: spaces and args", `{{(makemap "up" "down" "left" "right").left}}`, "right", tVal, true}, + + // If. + {"if true", "{{if true}}TRUE{{end}}", "TRUE", tVal, true}, + {"if false", "{{if false}}TRUE{{else}}FALSE{{end}}", "FALSE", tVal, true}, + {"if nil", "{{if nil}}TRUE{{end}}", "", tVal, false}, + {"if on typed nil interface value", "{{if .NonEmptyInterfaceTypedNil}}TRUE{{ end }}", "", tVal, true}, + {"if 1", "{{if 1}}NON-ZERO{{else}}ZERO{{end}}", "NON-ZERO", tVal, true}, + {"if 0", "{{if 0}}NON-ZERO{{else}}ZERO{{end}}", "ZERO", tVal, true}, + {"if 1.5", "{{if 1.5}}NON-ZERO{{else}}ZERO{{end}}", "NON-ZERO", tVal, true}, + {"if 0.0", "{{if .FloatZero}}NON-ZERO{{else}}ZERO{{end}}", "ZERO", tVal, true}, + {"if 1.5i", "{{if 1.5i}}NON-ZERO{{else}}ZERO{{end}}", "NON-ZERO", tVal, true}, + {"if 0.0i", "{{if .ComplexZero}}NON-ZERO{{else}}ZERO{{end}}", "ZERO", tVal, true}, + {"if emptystring", "{{if ``}}NON-EMPTY{{else}}EMPTY{{end}}", "EMPTY", tVal, true}, + {"if string", "{{if `notempty`}}NON-EMPTY{{else}}EMPTY{{end}}", "NON-EMPTY", tVal, true}, + {"if emptyslice", "{{if .SIEmpty}}NON-EMPTY{{else}}EMPTY{{end}}", "EMPTY", tVal, true}, + {"if slice", "{{if .SI}}NON-EMPTY{{else}}EMPTY{{end}}", "NON-EMPTY", tVal, true}, + {"if emptymap", "{{if .MSIEmpty}}NON-EMPTY{{else}}EMPTY{{end}}", "EMPTY", tVal, true}, + {"if map", "{{if .MSI}}NON-EMPTY{{else}}EMPTY{{end}}", "NON-EMPTY", tVal, true}, + {"if map unset", "{{if .MXI.none}}NON-ZERO{{else}}ZERO{{end}}", "ZERO", tVal, true}, + {"if map not unset", "{{if not .MXI.none}}ZERO{{else}}NON-ZERO{{end}}", "ZERO", tVal, true}, + {"if $x with $y int", "{{if $x := true}}{{with $y := .I}}{{$x}},{{$y}}{{end}}{{end}}", "true,17", tVal, true}, + {"if $x with $x int", "{{if $x := true}}{{with $x := .I}}{{$x}},{{end}}{{$x}}{{end}}", "17,true", tVal, true}, + {"if else if", "{{if false}}FALSE{{else if true}}TRUE{{end}}", "TRUE", tVal, true}, + {"if else chain", "{{if eq 1 3}}1{{else if eq 2 3}}2{{else if eq 3 3}}3{{end}}", "3", tVal, true}, + + // Print etc. + {"print", `{{print "hello, print"}}`, "hello, print", tVal, true}, + {"print 123", `{{print 1 2 3}}`, "1 2 3", tVal, true}, + {"print nil", `{{print nil}}`, "<nil>", tVal, true}, + {"println", `{{println 1 2 3}}`, "1 2 3\n", tVal, true}, + {"printf int", `{{printf "%04x" 127}}`, "007f", tVal, true}, + {"printf float", `{{printf "%g" 3.5}}`, "3.5", tVal, true}, + {"printf complex", `{{printf "%g" 1+7i}}`, "(1+7i)", tVal, true}, + {"printf string", `{{printf "%s" "hello"}}`, "hello", tVal, true}, + {"printf function", `{{printf "%#q" zeroArgs}}`, "`zeroArgs`", tVal, true}, + {"printf field", `{{printf "%s" .U.V}}`, "v", tVal, true}, + {"printf method", `{{printf "%s" .Method0}}`, "M0", tVal, true}, + {"printf dot", `{{with .I}}{{printf "%d" .}}{{end}}`, "17", tVal, true}, + {"printf var", `{{with $x := .I}}{{printf "%d" $x}}{{end}}`, "17", tVal, true}, + {"printf lots", `{{printf "%d %s %g %s" 127 "hello" 7-3i .Method0}}`, "127 hello (7-3i) M0", tVal, true}, + + // HTML. + {"html", `{{html "<script>alert(\"XSS\");</script>"}}`, + "<script>alert("XSS");</script>", nil, true}, + {"html pipeline", `{{printf "<script>alert(\"XSS\");</script>" | html}}`, + "<script>alert("XSS");</script>", nil, true}, + {"html", `{{html .PS}}`, "a string", tVal, true}, + {"html typed nil", `{{html .NIL}}`, "<nil>", tVal, true}, + {"html untyped nil", `{{html .Empty0}}`, "<no value>", tVal, true}, + + // JavaScript. + {"js", `{{js .}}`, `It\'d be nice.`, `It'd be nice.`, true}, + + // URL query. + {"urlquery", `{{"http://www.example.org/"|urlquery}}`, "http%3A%2F%2Fwww.example.org%2F", nil, true}, + + // Booleans + {"not", "{{not true}} {{not false}}", "false true", nil, true}, + {"and", "{{and false 0}} {{and 1 0}} {{and 0 true}} {{and 1 1}}", "false 0 0 1", nil, true}, + {"or", "{{or 0 0}} {{or 1 0}} {{or 0 true}} {{or 1 1}}", "0 1 true 1", nil, true}, + {"boolean if", "{{if and true 1 `hi`}}TRUE{{else}}FALSE{{end}}", "TRUE", tVal, true}, + {"boolean if not", "{{if and true 1 `hi` | not}}TRUE{{else}}FALSE{{end}}", "FALSE", nil, true}, + + // Indexing. + {"slice[0]", "{{index .SI 0}}", "3", tVal, true}, + {"slice[1]", "{{index .SI 1}}", "4", tVal, true}, + {"slice[HUGE]", "{{index .SI 10}}", "", tVal, false}, + {"slice[WRONG]", "{{index .SI `hello`}}", "", tVal, false}, + {"slice[nil]", "{{index .SI nil}}", "", tVal, false}, + {"map[one]", "{{index .MSI `one`}}", "1", tVal, true}, + {"map[two]", "{{index .MSI `two`}}", "2", tVal, true}, + {"map[NO]", "{{index .MSI `XXX`}}", "0", tVal, true}, + {"map[nil]", "{{index .MSI nil}}", "", tVal, false}, + {"map[``]", "{{index .MSI ``}}", "0", tVal, true}, + {"map[WRONG]", "{{index .MSI 10}}", "", tVal, false}, + {"double index", "{{index .SMSI 1 `eleven`}}", "11", tVal, true}, + {"nil[1]", "{{index nil 1}}", "", tVal, false}, + {"map MI64S", "{{index .MI64S 2}}", "i642", tVal, true}, + {"map MI32S", "{{index .MI32S 2}}", "two", tVal, true}, + {"map MUI64S", "{{index .MUI64S 3}}", "ui643", tVal, true}, + {"map MI8S", "{{index .MI8S 3}}", "i83", tVal, true}, + {"map MUI8S", "{{index .MUI8S 2}}", "u82", tVal, true}, + {"index of an interface field", "{{index .Empty3 0}}", "7", tVal, true}, + + // Slicing. + {"slice[:]", "{{slice .SI}}", "[3 4 5]", tVal, true}, + {"slice[1:]", "{{slice .SI 1}}", "[4 5]", tVal, true}, + {"slice[1:2]", "{{slice .SI 1 2}}", "[4]", tVal, true}, + {"slice[-1:]", "{{slice .SI -1}}", "", tVal, false}, + {"slice[1:-2]", "{{slice .SI 1 -2}}", "", tVal, false}, + {"slice[1:2:-1]", "{{slice .SI 1 2 -1}}", "", tVal, false}, + {"slice[2:1]", "{{slice .SI 2 1}}", "", tVal, false}, + {"slice[2:2:1]", "{{slice .SI 2 2 1}}", "", tVal, false}, + {"out of range", "{{slice .SI 4 5}}", "", tVal, false}, + {"out of range", "{{slice .SI 2 2 5}}", "", tVal, false}, + {"len(s) < indexes < cap(s)", "{{slice .SICap 6 10}}", "[0 0 0 0]", tVal, true}, + {"len(s) < indexes < cap(s)", "{{slice .SICap 6 10 10}}", "[0 0 0 0]", tVal, true}, + {"indexes > cap(s)", "{{slice .SICap 10 11}}", "", tVal, false}, + {"indexes > cap(s)", "{{slice .SICap 6 10 11}}", "", tVal, false}, + {"array[:]", "{{slice .AI}}", "[3 4 5]", tVal, true}, + {"array[1:]", "{{slice .AI 1}}", "[4 5]", tVal, true}, + {"array[1:2]", "{{slice .AI 1 2}}", "[4]", tVal, true}, + {"string[:]", "{{slice .S}}", "xyz", tVal, true}, + {"string[0:1]", "{{slice .S 0 1}}", "x", tVal, true}, + {"string[1:]", "{{slice .S 1}}", "yz", tVal, true}, + {"string[1:2]", "{{slice .S 1 2}}", "y", tVal, true}, + {"out of range", "{{slice .S 1 5}}", "", tVal, false}, + {"3-index slice of string", "{{slice .S 1 2 2}}", "", tVal, false}, + {"slice of an interface field", "{{slice .Empty3 0 1}}", "[7]", tVal, true}, + + // Len. + {"slice", "{{len .SI}}", "3", tVal, true}, + {"map", "{{len .MSI }}", "3", tVal, true}, + {"len of int", "{{len 3}}", "", tVal, false}, + {"len of nothing", "{{len .Empty0}}", "", tVal, false}, + {"len of an interface field", "{{len .Empty3}}", "2", tVal, true}, + + // With. + {"with true", "{{with true}}{{.}}{{end}}", "true", tVal, true}, + {"with false", "{{with false}}{{.}}{{else}}FALSE{{end}}", "FALSE", tVal, true}, + {"with 1", "{{with 1}}{{.}}{{else}}ZERO{{end}}", "1", tVal, true}, + {"with 0", "{{with 0}}{{.}}{{else}}ZERO{{end}}", "ZERO", tVal, true}, + {"with 1.5", "{{with 1.5}}{{.}}{{else}}ZERO{{end}}", "1.5", tVal, true}, + {"with 0.0", "{{with .FloatZero}}{{.}}{{else}}ZERO{{end}}", "ZERO", tVal, true}, + {"with 1.5i", "{{with 1.5i}}{{.}}{{else}}ZERO{{end}}", "(0+1.5i)", tVal, true}, + {"with 0.0i", "{{with .ComplexZero}}{{.}}{{else}}ZERO{{end}}", "ZERO", tVal, true}, + {"with emptystring", "{{with ``}}{{.}}{{else}}EMPTY{{end}}", "EMPTY", tVal, true}, + {"with string", "{{with `notempty`}}{{.}}{{else}}EMPTY{{end}}", "notempty", tVal, true}, + {"with emptyslice", "{{with .SIEmpty}}{{.}}{{else}}EMPTY{{end}}", "EMPTY", tVal, true}, + {"with slice", "{{with .SI}}{{.}}{{else}}EMPTY{{end}}", "[3 4 5]", tVal, true}, + {"with emptymap", "{{with .MSIEmpty}}{{.}}{{else}}EMPTY{{end}}", "EMPTY", tVal, true}, + {"with map", "{{with .MSIone}}{{.}}{{else}}EMPTY{{end}}", "map[one:1]", tVal, true}, + {"with empty interface, struct field", "{{with .Empty4}}{{.V}}{{end}}", "UinEmpty", tVal, true}, + {"with $x int", "{{with $x := .I}}{{$x}}{{end}}", "17", tVal, true}, + {"with $x struct.U.V", "{{with $x := $}}{{$x.U.V}}{{end}}", "v", tVal, true}, + {"with variable and action", "{{with $x := $}}{{$y := $.U.V}}{{$y}}{{end}}", "v", tVal, true}, + {"with on typed nil interface value", "{{with .NonEmptyInterfaceTypedNil}}TRUE{{ end }}", "", tVal, true}, + + // Range. + {"range []int", "{{range .SI}}-{{.}}-{{end}}", "-3--4--5-", tVal, true}, + {"range empty no else", "{{range .SIEmpty}}-{{.}}-{{end}}", "", tVal, true}, + {"range []int else", "{{range .SI}}-{{.}}-{{else}}EMPTY{{end}}", "-3--4--5-", tVal, true}, + {"range empty else", "{{range .SIEmpty}}-{{.}}-{{else}}EMPTY{{end}}", "EMPTY", tVal, true}, + {"range []bool", "{{range .SB}}-{{.}}-{{end}}", "-true--false-", tVal, true}, + {"range []int method", "{{range .SI | .MAdd .I}}-{{.}}-{{end}}", "-20--21--22-", tVal, true}, + {"range map", "{{range .MSI}}-{{.}}-{{end}}", "-1--3--2-", tVal, true}, + {"range empty map no else", "{{range .MSIEmpty}}-{{.}}-{{end}}", "", tVal, true}, + {"range map else", "{{range .MSI}}-{{.}}-{{else}}EMPTY{{end}}", "-1--3--2-", tVal, true}, + {"range empty map else", "{{range .MSIEmpty}}-{{.}}-{{else}}EMPTY{{end}}", "EMPTY", tVal, true}, + {"range empty interface", "{{range .Empty3}}-{{.}}-{{else}}EMPTY{{end}}", "-7--8-", tVal, true}, + {"range empty nil", "{{range .Empty0}}-{{.}}-{{end}}", "", tVal, true}, + {"range $x SI", "{{range $x := .SI}}<{{$x}}>{{end}}", "<3><4><5>", tVal, true}, + {"range $x $y SI", "{{range $x, $y := .SI}}<{{$x}}={{$y}}>{{end}}", "<0=3><1=4><2=5>", tVal, true}, + {"range $x MSIone", "{{range $x := .MSIone}}<{{$x}}>{{end}}", "<1>", tVal, true}, + {"range $x $y MSIone", "{{range $x, $y := .MSIone}}<{{$x}}={{$y}}>{{end}}", "<one=1>", tVal, true}, + {"range $x PSI", "{{range $x := .PSI}}<{{$x}}>{{end}}", "<21><22><23>", tVal, true}, + {"declare in range", "{{range $x := .PSI}}<{{$foo:=$x}}{{$x}}>{{end}}", "<21><22><23>", tVal, true}, + {"range count", `{{range $i, $x := count 5}}[{{$i}}]{{$x}}{{end}}`, "[0]a[1]b[2]c[3]d[4]e", tVal, true}, + {"range nil count", `{{range $i, $x := count 0}}{{else}}empty{{end}}`, "empty", tVal, true}, + + // Cute examples. + {"or as if true", `{{or .SI "slice is empty"}}`, "[3 4 5]", tVal, true}, + {"or as if false", `{{or .SIEmpty "slice is empty"}}`, "slice is empty", tVal, true}, + + // Error handling. + {"error method, error", "{{.MyError true}}", "", tVal, false}, + {"error method, no error", "{{.MyError false}}", "false", tVal, true}, + + // Numbers + {"decimal", "{{print 1234}}", "1234", tVal, true}, + {"decimal _", "{{print 12_34}}", "1234", tVal, true}, + {"binary", "{{print 0b101}}", "5", tVal, true}, + {"binary _", "{{print 0b_1_0_1}}", "5", tVal, true}, + {"BINARY", "{{print 0B101}}", "5", tVal, true}, + {"octal0", "{{print 0377}}", "255", tVal, true}, + {"octal", "{{print 0o377}}", "255", tVal, true}, + {"octal _", "{{print 0o_3_7_7}}", "255", tVal, true}, + {"OCTAL", "{{print 0O377}}", "255", tVal, true}, + {"hex", "{{print 0x123}}", "291", tVal, true}, + {"hex _", "{{print 0x1_23}}", "291", tVal, true}, + {"HEX", "{{print 0X123ABC}}", "1194684", tVal, true}, + {"float", "{{print 123.4}}", "123.4", tVal, true}, + {"float _", "{{print 0_0_1_2_3.4}}", "123.4", tVal, true}, + {"hex float", "{{print +0x1.ep+2}}", "7.5", tVal, true}, + {"hex float _", "{{print +0x_1.e_0p+0_2}}", "7.5", tVal, true}, + {"HEX float", "{{print +0X1.EP+2}}", "7.5", tVal, true}, + {"print multi", "{{print 1_2_3_4 7.5_00_00_00}}", "1234 7.5", tVal, true}, + {"print multi2", "{{print 1234 0x0_1.e_0p+02}}", "1234 7.5", tVal, true}, + + // Fixed bugs. + // Must separate dot and receiver; otherwise args are evaluated with dot set to variable. + {"bug0", "{{range .MSIone}}{{if $.Method1 .}}X{{end}}{{end}}", "X", tVal, true}, + // Do not loop endlessly in indirect for non-empty interfaces. + // The bug appears with *interface only; looped forever. + {"bug1", "{{.Method0}}", "M0", &iVal, true}, + // Was taking address of interface field, so method set was empty. + {"bug2", "{{$.NonEmptyInterface.Method0}}", "M0", tVal, true}, + // Struct values were not legal in with - mere oversight. + {"bug3", "{{with $}}{{.Method0}}{{end}}", "M0", tVal, true}, + // Nil interface values in if. + {"bug4", "{{if .Empty0}}non-nil{{else}}nil{{end}}", "nil", tVal, true}, + // Stringer. + {"bug5", "{{.Str}}", "foozle", tVal, true}, + {"bug5a", "{{.Err}}", "erroozle", tVal, true}, + // Args need to be indirected and dereferenced sometimes. + {"bug6a", "{{vfunc .V0 .V1}}", "vfunc", tVal, true}, + {"bug6b", "{{vfunc .V0 .V0}}", "vfunc", tVal, true}, + {"bug6c", "{{vfunc .V1 .V0}}", "vfunc", tVal, true}, + {"bug6d", "{{vfunc .V1 .V1}}", "vfunc", tVal, true}, + // Legal parse but illegal execution: non-function should have no arguments. + {"bug7a", "{{3 2}}", "", tVal, false}, + {"bug7b", "{{$x := 1}}{{$x 2}}", "", tVal, false}, + {"bug7c", "{{$x := 1}}{{3 | $x}}", "", tVal, false}, + // Pipelined arg was not being type-checked. + {"bug8a", "{{3|oneArg}}", "", tVal, false}, + {"bug8b", "{{4|dddArg 3}}", "", tVal, false}, + // A bug was introduced that broke map lookups for lower-case names. + {"bug9", "{{.cause}}", "neglect", map[string]string{"cause": "neglect"}, true}, + // Field chain starting with function did not work. + {"bug10", "{{mapOfThree.three}}-{{(mapOfThree).three}}", "3-3", 0, true}, + // Dereferencing nil pointer while evaluating function arguments should not panic. Issue 7333. + {"bug11", "{{valueString .PS}}", "", T{}, false}, + // 0xef gave constant type float64. Issue 8622. + {"bug12xe", "{{printf `%T` 0xef}}", "int", T{}, true}, + {"bug12xE", "{{printf `%T` 0xEE}}", "int", T{}, true}, + {"bug12Xe", "{{printf `%T` 0Xef}}", "int", T{}, true}, + {"bug12XE", "{{printf `%T` 0XEE}}", "int", T{}, true}, + // Chained nodes did not work as arguments. Issue 8473. + {"bug13", "{{print (.Copy).I}}", "17", tVal, true}, + // Didn't protect against nil or literal values in field chains. + {"bug14a", "{{(nil).True}}", "", tVal, false}, + {"bug14b", "{{$x := nil}}{{$x.anything}}", "", tVal, false}, + {"bug14c", `{{$x := (1.0)}}{{$y := ("hello")}}{{$x.anything}}{{$y.true}}`, "", tVal, false}, + // Didn't call validateType on function results. Issue 10800. + {"bug15", "{{valueString returnInt}}", "", tVal, false}, + // Variadic function corner cases. Issue 10946. + {"bug16a", "{{true|printf}}", "", tVal, false}, + {"bug16b", "{{1|printf}}", "", tVal, false}, + {"bug16c", "{{1.1|printf}}", "", tVal, false}, + {"bug16d", "{{'x'|printf}}", "", tVal, false}, + {"bug16e", "{{0i|printf}}", "", tVal, false}, + {"bug16f", "{{true|twoArgs \"xxx\"}}", "", tVal, false}, + {"bug16g", "{{\"aaa\" |twoArgs \"bbb\"}}", "twoArgs=bbbaaa", tVal, true}, + {"bug16h", "{{1|oneArg}}", "", tVal, false}, + {"bug16i", "{{\"aaa\"|oneArg}}", "oneArg=aaa", tVal, true}, + {"bug16j", "{{1+2i|printf \"%v\"}}", "(1+2i)", tVal, true}, + {"bug16k", "{{\"aaa\"|printf }}", "aaa", tVal, true}, + {"bug17a", "{{.NonEmptyInterface.X}}", "x", tVal, true}, + {"bug17b", "-{{.NonEmptyInterface.Method1 1234}}-", "-1234-", tVal, true}, + {"bug17c", "{{len .NonEmptyInterfacePtS}}", "2", tVal, true}, + {"bug17d", "{{index .NonEmptyInterfacePtS 0}}", "a", tVal, true}, + {"bug17e", "{{range .NonEmptyInterfacePtS}}-{{.}}-{{end}}", "-a--b-", tVal, true}, + + // More variadic function corner cases. Some runes would get evaluated + // as constant floats instead of ints. Issue 34483. + {"bug18a", "{{eq . '.'}}", "true", '.', true}, + {"bug18b", "{{eq . 'e'}}", "true", 'e', true}, + {"bug18c", "{{eq . 'P'}}", "true", 'P', true}, +} + +func zeroArgs() string { + return "zeroArgs" +} + +func oneArg(a string) string { + return "oneArg=" + a +} + +func twoArgs(a, b string) string { + return "twoArgs=" + a + b +} + +func dddArg(a int, b ...string) string { + return fmt.Sprintln(a, b) +} + +// count returns a channel that will deliver n sequential 1-letter strings starting at "a" +func count(n int) chan string { + if n == 0 { + return nil + } + c := make(chan string) + go func() { + for i := 0; i < n; i++ { + c <- "abcdefghijklmnop"[i : i+1] + } + close(c) + }() + return c +} + +// vfunc takes a *V and a V +func vfunc(V, *V) string { + return "vfunc" +} + +// valueString takes a string, not a pointer. +func valueString(v string) string { + return "value is ignored" +} + +// returnInt returns an int +func returnInt() int { + return 7 +} + +func add(args ...int) int { + sum := 0 + for _, x := range args { + sum += x + } + return sum +} + +func echo(arg interface{}) interface{} { + return arg +} + +func makemap(arg ...string) map[string]string { + if len(arg)%2 != 0 { + panic("bad makemap") + } + m := make(map[string]string) + for i := 0; i < len(arg); i += 2 { + m[arg[i]] = arg[i+1] + } + return m +} + +func stringer(s fmt.Stringer) string { + return s.String() +} + +func mapOfThree() interface{} { + return map[string]int{"three": 3} +} + +func testExecute(execTests []execTest, template *Template, t *testing.T) { + b := new(bytes.Buffer) + funcs := FuncMap{ + "add": add, + "count": count, + "dddArg": dddArg, + "echo": echo, + "makemap": makemap, + "mapOfThree": mapOfThree, + "oneArg": oneArg, + "returnInt": returnInt, + "stringer": stringer, + "twoArgs": twoArgs, + "typeOf": typeOf, + "valueString": valueString, + "vfunc": vfunc, + "zeroArgs": zeroArgs, + } + for _, test := range execTests { + var tmpl *Template + var err error + if template == nil { + tmpl, err = New(test.name).Funcs(funcs).Parse(test.input) + } else { + tmpl, err = template.New(test.name).Funcs(funcs).Parse(test.input) + } + if err != nil { + t.Errorf("%s: parse error: %s", test.name, err) + continue + } + b.Reset() + err = tmpl.Execute(b, test.data) + switch { + case !test.ok && err == nil: + t.Errorf("%s: expected error; got none", test.name) + continue + case test.ok && err != nil: + t.Errorf("%s: unexpected execute error: %s", test.name, err) + continue + case !test.ok && err != nil: + // expected error, got one + if *debug { + fmt.Printf("%s: %s\n\t%s\n", test.name, test.input, err) + } + } + result := b.String() + if result != test.output { + t.Errorf("%s: expected\n\t%q\ngot\n\t%q", test.name, test.output, result) + } + } +} + +func TestExecute(t *testing.T) { + testExecute(execTests, nil, t) +} + +var delimPairs = []string{ + "", "", // default + "{{", "}}", // same as default + "<<", ">>", // distinct + "|", "|", // same + "(日)", "(本)", // peculiar +} + +func TestDelims(t *testing.T) { + const hello = "Hello, world" + var value = struct{ Str string }{hello} + for i := 0; i < len(delimPairs); i += 2 { + text := ".Str" + left := delimPairs[i+0] + trueLeft := left + right := delimPairs[i+1] + trueRight := right + if left == "" { // default case + trueLeft = "{{" + } + if right == "" { // default case + trueRight = "}}" + } + text = trueLeft + text + trueRight + // Now add a comment + text += trueLeft + "/*comment*/" + trueRight + // Now add an action containing a string. + text += trueLeft + `"` + trueLeft + `"` + trueRight + // At this point text looks like `{{.Str}}{{/*comment*/}}{{"{{"}}`. + tmpl, err := New("delims").Delims(left, right).Parse(text) + if err != nil { + t.Fatalf("delim %q text %q parse err %s", left, text, err) + } + var b = new(bytes.Buffer) + err = tmpl.Execute(b, value) + if err != nil { + t.Fatalf("delim %q exec err %s", left, err) + } + if b.String() != hello+trueLeft { + t.Errorf("expected %q got %q", hello+trueLeft, b.String()) + } + } +} + +// Check that an error from a method flows back to the top. +func TestExecuteError(t *testing.T) { + b := new(bytes.Buffer) + tmpl := New("error") + _, err := tmpl.Parse("{{.MyError true}}") + if err != nil { + t.Fatalf("parse error: %s", err) + } + err = tmpl.Execute(b, tVal) + if err == nil { + t.Errorf("expected error; got none") + } else if !strings.Contains(err.Error(), myError.Error()) { + if *debug { + fmt.Printf("test execute error: %s\n", err) + } + t.Errorf("expected myError; got %s", err) + } +} + +const execErrorText = `line 1 +line 2 +line 3 +{{template "one" .}} +{{define "one"}}{{template "two" .}}{{end}} +{{define "two"}}{{template "three" .}}{{end}} +{{define "three"}}{{index "hi" $}}{{end}}` + +// Check that an error from a nested template contains all the relevant information. +func TestExecError(t *testing.T) { + tmpl, err := New("top").Parse(execErrorText) + if err != nil { + t.Fatal("parse error:", err) + } + var b bytes.Buffer + err = tmpl.Execute(&b, 5) // 5 is out of range indexing "hi" + if err == nil { + t.Fatal("expected error") + } + const want = `template: top:7:20: executing "three" at <index "hi" $>: error calling index: index out of range: 5` + got := err.Error() + if got != want { + t.Errorf("expected\n%q\ngot\n%q", want, got) + } +} + +type CustomError struct{} + +func (*CustomError) Error() string { return "heyo !" } + +// Check that a custom error can be returned. +func TestExecError_CustomError(t *testing.T) { + failingFunc := func() (string, error) { + return "", &CustomError{} + } + tmpl := Must(New("top").Funcs(FuncMap{ + "err": failingFunc, + }).Parse("{{ err }}")) + + var b bytes.Buffer + err := tmpl.Execute(&b, nil) + + var e *CustomError + if !errors.As(err, &e) { + t.Fatalf("expected custom error; got %s", err) + } +} + +func TestJSEscaping(t *testing.T) { + testCases := []struct { + in, exp string + }{ + {`a`, `a`}, + {`'foo`, `\'foo`}, + {`Go "jump" \`, `Go \"jump\" \\`}, + {`Yukihiro says "今日は世界"`, `Yukihiro says \"今日は世界\"`}, + {"unprintable \uFDFF", `unprintable \uFDFF`}, + {`<html>`, `\u003Chtml\u003E`}, + {`no = in attributes`, `no \u003D in attributes`}, + {`' does not become HTML entity`, `\u0026#x27; does not become HTML entity`}, + } + for _, tc := range testCases { + s := JSEscapeString(tc.in) + if s != tc.exp { + t.Errorf("JS escaping [%s] got [%s] want [%s]", tc.in, s, tc.exp) + } + } +} + +// A nice example: walk a binary tree. + +type Tree struct { + Val int + Left, Right *Tree +} + +// Use different delimiters to test Set.Delims. +// Also test the trimming of leading and trailing spaces. +const treeTemplate = ` + (- define "tree" -) + [ + (- .Val -) + (- with .Left -) + (template "tree" . -) + (- end -) + (- with .Right -) + (- template "tree" . -) + (- end -) + ] + (- end -) +` + +func TestTree(t *testing.T) { + var tree = &Tree{ + 1, + &Tree{ + 2, &Tree{ + 3, + &Tree{ + 4, nil, nil, + }, + nil, + }, + &Tree{ + 5, + &Tree{ + 6, nil, nil, + }, + nil, + }, + }, + &Tree{ + 7, + &Tree{ + 8, + &Tree{ + 9, nil, nil, + }, + nil, + }, + &Tree{ + 10, + &Tree{ + 11, nil, nil, + }, + nil, + }, + }, + } + tmpl, err := New("root").Delims("(", ")").Parse(treeTemplate) + if err != nil { + t.Fatal("parse error:", err) + } + var b bytes.Buffer + const expect = "[1[2[3[4]][5[6]]][7[8[9]][10[11]]]]" + // First by looking up the template. + err = tmpl.Lookup("tree").Execute(&b, tree) + if err != nil { + t.Fatal("exec error:", err) + } + result := b.String() + if result != expect { + t.Errorf("expected %q got %q", expect, result) + } + // Then direct to execution. + b.Reset() + err = tmpl.ExecuteTemplate(&b, "tree", tree) + if err != nil { + t.Fatal("exec error:", err) + } + result = b.String() + if result != expect { + t.Errorf("expected %q got %q", expect, result) + } +} + +func TestExecuteOnNewTemplate(t *testing.T) { + // This is issue 3872. + New("Name").Templates() + // This is issue 11379. + new(Template).Templates() + new(Template).Parse("") + new(Template).New("abc").Parse("") + new(Template).Execute(nil, nil) // returns an error (but does not crash) + new(Template).ExecuteTemplate(nil, "XXX", nil) // returns an error (but does not crash) +} + +const testTemplates = `{{define "one"}}one{{end}}{{define "two"}}two{{end}}` + +func TestMessageForExecuteEmpty(t *testing.T) { + // Test a truly empty template. + tmpl := New("empty") + var b bytes.Buffer + err := tmpl.Execute(&b, 0) + if err == nil { + t.Fatal("expected initial error") + } + got := err.Error() + want := `template: empty: "empty" is an incomplete or empty template` + if got != want { + t.Errorf("expected error %s got %s", want, got) + } + // Add a non-empty template to check that the error is helpful. + tests, err := New("").Parse(testTemplates) + if err != nil { + t.Fatal(err) + } + tmpl.AddParseTree("secondary", tests.Tree) + err = tmpl.Execute(&b, 0) + if err == nil { + t.Fatal("expected second error") + } + got = err.Error() + want = `template: empty: "empty" is an incomplete or empty template` + if got != want { + t.Errorf("expected error %s got %s", want, got) + } + // Make sure we can execute the secondary. + err = tmpl.ExecuteTemplate(&b, "secondary", 0) + if err != nil { + t.Fatal(err) + } +} + +func TestFinalForPrintf(t *testing.T) { + tmpl, err := New("").Parse(`{{"x" | printf}}`) + if err != nil { + t.Fatal(err) + } + var b bytes.Buffer + err = tmpl.Execute(&b, 0) + if err != nil { + t.Fatal(err) + } +} + +type cmpTest struct { + expr string + truth string + ok bool +} + +var cmpTests = []cmpTest{ + {"eq true true", "true", true}, + {"eq true false", "false", true}, + {"eq 1+2i 1+2i", "true", true}, + {"eq 1+2i 1+3i", "false", true}, + {"eq 1.5 1.5", "true", true}, + {"eq 1.5 2.5", "false", true}, + {"eq 1 1", "true", true}, + {"eq 1 2", "false", true}, + {"eq `xy` `xy`", "true", true}, + {"eq `xy` `xyz`", "false", true}, + {"eq .Uthree .Uthree", "true", true}, + {"eq .Uthree .Ufour", "false", true}, + {"eq 3 4 5 6 3", "true", true}, + {"eq 3 4 5 6 7", "false", true}, + {"ne true true", "false", true}, + {"ne true false", "true", true}, + {"ne 1+2i 1+2i", "false", true}, + {"ne 1+2i 1+3i", "true", true}, + {"ne 1.5 1.5", "false", true}, + {"ne 1.5 2.5", "true", true}, + {"ne 1 1", "false", true}, + {"ne 1 2", "true", true}, + {"ne `xy` `xy`", "false", true}, + {"ne `xy` `xyz`", "true", true}, + {"ne .Uthree .Uthree", "false", true}, + {"ne .Uthree .Ufour", "true", true}, + {"lt 1.5 1.5", "false", true}, + {"lt 1.5 2.5", "true", true}, + {"lt 1 1", "false", true}, + {"lt 1 2", "true", true}, + {"lt `xy` `xy`", "false", true}, + {"lt `xy` `xyz`", "true", true}, + {"lt .Uthree .Uthree", "false", true}, + {"lt .Uthree .Ufour", "true", true}, + {"le 1.5 1.5", "true", true}, + {"le 1.5 2.5", "true", true}, + {"le 2.5 1.5", "false", true}, + {"le 1 1", "true", true}, + {"le 1 2", "true", true}, + {"le 2 1", "false", true}, + {"le `xy` `xy`", "true", true}, + {"le `xy` `xyz`", "true", true}, + {"le `xyz` `xy`", "false", true}, + {"le .Uthree .Uthree", "true", true}, + {"le .Uthree .Ufour", "true", true}, + {"le .Ufour .Uthree", "false", true}, + {"gt 1.5 1.5", "false", true}, + {"gt 1.5 2.5", "false", true}, + {"gt 1 1", "false", true}, + {"gt 2 1", "true", true}, + {"gt 1 2", "false", true}, + {"gt `xy` `xy`", "false", true}, + {"gt `xy` `xyz`", "false", true}, + {"gt .Uthree .Uthree", "false", true}, + {"gt .Uthree .Ufour", "false", true}, + {"gt .Ufour .Uthree", "true", true}, + {"ge 1.5 1.5", "true", true}, + {"ge 1.5 2.5", "false", true}, + {"ge 2.5 1.5", "true", true}, + {"ge 1 1", "true", true}, + {"ge 1 2", "false", true}, + {"ge 2 1", "true", true}, + {"ge `xy` `xy`", "true", true}, + {"ge `xy` `xyz`", "false", true}, + {"ge `xyz` `xy`", "true", true}, + {"ge .Uthree .Uthree", "true", true}, + {"ge .Uthree .Ufour", "false", true}, + {"ge .Ufour .Uthree", "true", true}, + // Mixing signed and unsigned integers. + {"eq .Uthree .Three", "true", true}, + {"eq .Three .Uthree", "true", true}, + {"le .Uthree .Three", "true", true}, + {"le .Three .Uthree", "true", true}, + {"ge .Uthree .Three", "true", true}, + {"ge .Three .Uthree", "true", true}, + {"lt .Uthree .Three", "false", true}, + {"lt .Three .Uthree", "false", true}, + {"gt .Uthree .Three", "false", true}, + {"gt .Three .Uthree", "false", true}, + {"eq .Ufour .Three", "false", true}, + {"lt .Ufour .Three", "false", true}, + {"gt .Ufour .Three", "true", true}, + {"eq .NegOne .Uthree", "false", true}, + {"eq .Uthree .NegOne", "false", true}, + {"ne .NegOne .Uthree", "true", true}, + {"ne .Uthree .NegOne", "true", true}, + {"lt .NegOne .Uthree", "true", true}, + {"lt .Uthree .NegOne", "false", true}, + {"le .NegOne .Uthree", "true", true}, + {"le .Uthree .NegOne", "false", true}, + {"gt .NegOne .Uthree", "false", true}, + {"gt .Uthree .NegOne", "true", true}, + {"ge .NegOne .Uthree", "false", true}, + {"ge .Uthree .NegOne", "true", true}, + {"eq (index `x` 0) 'x'", "true", true}, // The example that triggered this rule. + {"eq (index `x` 0) 'y'", "false", true}, + {"eq .V1 .V2", "true", true}, + {"eq .Ptr .Ptr", "true", true}, + {"eq .Ptr .NilPtr", "false", true}, + {"eq .NilPtr .NilPtr", "true", true}, + {"eq .Iface1 .Iface1", "true", true}, + {"eq .Iface1 .NilIface", "false", true}, + {"eq .NilIface .NilIface", "true", true}, + {"eq .NilIface .Iface1", "false", true}, + {"eq .NilIface 0", "false", true}, + {"eq 0 .NilIface", "false", true}, + // Errors + {"eq `xy` 1", "", false}, // Different types. + {"eq 2 2.0", "", false}, // Different types. + {"lt true true", "", false}, // Unordered types. + {"lt 1+0i 1+0i", "", false}, // Unordered types. + {"eq .Ptr 1", "", false}, // Incompatible types. + {"eq .Ptr .NegOne", "", false}, // Incompatible types. + {"eq .Map .Map", "", false}, // Uncomparable types. + {"eq .Map .V1", "", false}, // Uncomparable types. +} + +func TestComparison(t *testing.T) { + b := new(bytes.Buffer) + var cmpStruct = struct { + Uthree, Ufour uint + NegOne, Three int + Ptr, NilPtr *int + Map map[int]int + V1, V2 V + Iface1, NilIface fmt.Stringer + }{ + Uthree: 3, + Ufour: 4, + NegOne: -1, + Three: 3, + Ptr: new(int), + Iface1: b, + } + for _, test := range cmpTests { + text := fmt.Sprintf("{{if %s}}true{{else}}false{{end}}", test.expr) + tmpl, err := New("empty").Parse(text) + if err != nil { + t.Fatalf("%q: %s", test.expr, err) + } + b.Reset() + err = tmpl.Execute(b, &cmpStruct) + if test.ok && err != nil { + t.Errorf("%s errored incorrectly: %s", test.expr, err) + continue + } + if !test.ok && err == nil { + t.Errorf("%s did not error", test.expr) + continue + } + if b.String() != test.truth { + t.Errorf("%s: want %s; got %s", test.expr, test.truth, b.String()) + } + } +} + +func TestMissingMapKey(t *testing.T) { + data := map[string]int{ + "x": 99, + } + tmpl, err := New("t1").Parse("{{.x}} {{.y}}") + if err != nil { + t.Fatal(err) + } + var b bytes.Buffer + // By default, just get "<no value>" + err = tmpl.Execute(&b, data) + if err != nil { + t.Fatal(err) + } + want := "99 <no value>" + got := b.String() + if got != want { + t.Errorf("got %q; expected %q", got, want) + } + // Same if we set the option explicitly to the default. + tmpl.Option("missingkey=default") + b.Reset() + err = tmpl.Execute(&b, data) + if err != nil { + t.Fatal("default:", err) + } + want = "99 <no value>" + got = b.String() + if got != want { + t.Errorf("got %q; expected %q", got, want) + } + // Next we ask for a zero value + tmpl.Option("missingkey=zero") + b.Reset() + err = tmpl.Execute(&b, data) + if err != nil { + t.Fatal("zero:", err) + } + want = "99 0" + got = b.String() + if got != want { + t.Errorf("got %q; expected %q", got, want) + } + // Now we ask for an error. + tmpl.Option("missingkey=error") + err = tmpl.Execute(&b, data) + if err == nil { + t.Errorf("expected error; got none") + } + // same Option, but now a nil interface: ask for an error + err = tmpl.Execute(&b, nil) + t.Log(err) + if err == nil { + t.Errorf("expected error for nil-interface; got none") + } +} + +// Test that the error message for multiline unterminated string +// refers to the line number of the opening quote. +func TestUnterminatedStringError(t *testing.T) { + _, err := New("X").Parse("hello\n\n{{`unterminated\n\n\n\n}}\n some more\n\n") + if err == nil { + t.Fatal("expected error") + } + str := err.Error() + if !strings.Contains(str, "X:3: unterminated raw quoted string") { + t.Fatalf("unexpected error: %s", str) + } +} + +const alwaysErrorText = "always be failing" + +var alwaysError = errors.New(alwaysErrorText) + +type ErrorWriter int + +func (e ErrorWriter) Write(p []byte) (int, error) { + return 0, alwaysError +} + +func TestExecuteGivesExecError(t *testing.T) { + // First, a non-execution error shouldn't be an ExecError. + tmpl, err := New("X").Parse("hello") + if err != nil { + t.Fatal(err) + } + err = tmpl.Execute(ErrorWriter(0), 0) + if err == nil { + t.Fatal("expected error; got none") + } + if err.Error() != alwaysErrorText { + t.Errorf("expected %q error; got %q", alwaysErrorText, err) + } + // This one should be an ExecError. + tmpl, err = New("X").Parse("hello, {{.X.Y}}") + if err != nil { + t.Fatal(err) + } + err = tmpl.Execute(io.Discard, 0) + if err == nil { + t.Fatal("expected error; got none") + } + eerr, ok := err.(ExecError) + if !ok { + t.Fatalf("did not expect ExecError %s", eerr) + } + expect := "field X in type int" + if !strings.Contains(err.Error(), expect) { + t.Errorf("expected %q; got %q", expect, err) + } +} + +func funcNameTestFunc() int { + return 0 +} + +func TestGoodFuncNames(t *testing.T) { + names := []string{ + "_", + "a", + "a1", + "a1", + "Ӵ", + } + for _, name := range names { + tmpl := New("X").Funcs( + FuncMap{ + name: funcNameTestFunc, + }, + ) + if tmpl == nil { + t.Fatalf("nil result for %q", name) + } + } +} + +func TestBadFuncNames(t *testing.T) { + names := []string{ + "", + "2", + "a-b", + } + for _, name := range names { + testBadFuncName(name, t) + } +} + +func testBadFuncName(name string, t *testing.T) { + t.Helper() + defer func() { + recover() + }() + New("X").Funcs( + FuncMap{ + name: funcNameTestFunc, + }, + ) + // If we get here, the name did not cause a panic, which is how Funcs + // reports an error. + t.Errorf("%q succeeded incorrectly as function name", name) +} + +func TestBlock(t *testing.T) { + const ( + input = `a({{block "inner" .}}bar({{.}})baz{{end}})b` + want = `a(bar(hello)baz)b` + overlay = `{{define "inner"}}foo({{.}})bar{{end}}` + want2 = `a(foo(goodbye)bar)b` + ) + tmpl, err := New("outer").Parse(input) + if err != nil { + t.Fatal(err) + } + tmpl2, err := Must(tmpl.Clone()).Parse(overlay) + if err != nil { + t.Fatal(err) + } + + var buf bytes.Buffer + if err := tmpl.Execute(&buf, "hello"); err != nil { + t.Fatal(err) + } + if got := buf.String(); got != want { + t.Errorf("got %q, want %q", got, want) + } + + buf.Reset() + if err := tmpl2.Execute(&buf, "goodbye"); err != nil { + t.Fatal(err) + } + if got := buf.String(); got != want2 { + t.Errorf("got %q, want %q", got, want2) + } +} + +func TestEvalFieldErrors(t *testing.T) { + tests := []struct { + name, src string + value interface{} + want string + }{ + { + // Check that calling an invalid field on nil pointer + // prints a field error instead of a distracting nil + // pointer error. https://golang.org/issue/15125 + "MissingFieldOnNil", + "{{.MissingField}}", + (*T)(nil), + "can't evaluate field MissingField in type *template.T", + }, + { + "MissingFieldOnNonNil", + "{{.MissingField}}", + &T{}, + "can't evaluate field MissingField in type *template.T", + }, + { + "ExistingFieldOnNil", + "{{.X}}", + (*T)(nil), + "nil pointer evaluating *template.T.X", + }, + { + "MissingKeyOnNilMap", + "{{.MissingKey}}", + (*map[string]string)(nil), + "nil pointer evaluating *map[string]string.MissingKey", + }, + { + "MissingKeyOnNilMapPtr", + "{{.MissingKey}}", + (*map[string]string)(nil), + "nil pointer evaluating *map[string]string.MissingKey", + }, + { + "MissingKeyOnMapPtrToNil", + "{{.MissingKey}}", + &map[string]string{}, + "<nil>", + }, + } + for _, tc := range tests { + t.Run(tc.name, func(t *testing.T) { + tmpl := Must(New("tmpl").Parse(tc.src)) + err := tmpl.Execute(io.Discard, tc.value) + got := "<nil>" + if err != nil { + got = err.Error() + } + if !strings.HasSuffix(got, tc.want) { + t.Fatalf("got error %q, want %q", got, tc.want) + } + }) + } +} + +func TestMaxExecDepth(t *testing.T) { + if testing.Short() { + t.Skip("skipping in -short mode") + } + tmpl := Must(New("tmpl").Parse(`{{template "tmpl" .}}`)) + err := tmpl.Execute(io.Discard, nil) + got := "<nil>" + if err != nil { + got = err.Error() + } + const want = "exceeded maximum template depth" + if !strings.Contains(got, want) { + t.Errorf("got error %q; want %q", got, want) + } +} + +func TestAddrOfIndex(t *testing.T) { + // golang.org/issue/14916. + // Before index worked on reflect.Values, the .String could not be + // found on the (incorrectly unaddressable) V value, + // in contrast to range, which worked fine. + // Also testing that passing a reflect.Value to tmpl.Execute works. + texts := []string{ + `{{range .}}{{.String}}{{end}}`, + `{{with index . 0}}{{.String}}{{end}}`, + } + for _, text := range texts { + tmpl := Must(New("tmpl").Parse(text)) + var buf bytes.Buffer + err := tmpl.Execute(&buf, reflect.ValueOf([]V{{1}})) + if err != nil { + t.Fatalf("%s: Execute: %v", text, err) + } + if buf.String() != "<1>" { + t.Fatalf("%s: template output = %q, want %q", text, &buf, "<1>") + } + } +} + +func TestInterfaceValues(t *testing.T) { + // golang.org/issue/17714. + // Before index worked on reflect.Values, interface values + // were always implicitly promoted to the underlying value, + // except that nil interfaces were promoted to the zero reflect.Value. + // Eliminating a round trip to interface{} and back to reflect.Value + // eliminated this promotion, breaking these cases. + tests := []struct { + text string + out string + }{ + {`{{index .Nil 1}}`, "ERROR: index of untyped nil"}, + {`{{index .Slice 2}}`, "2"}, + {`{{index .Slice .Two}}`, "2"}, + {`{{call .Nil 1}}`, "ERROR: call of nil"}, + {`{{call .PlusOne 1}}`, "2"}, + {`{{call .PlusOne .One}}`, "2"}, + {`{{and (index .Slice 0) true}}`, "0"}, + {`{{and .Zero true}}`, "0"}, + {`{{and (index .Slice 1) false}}`, "false"}, + {`{{and .One false}}`, "false"}, + {`{{or (index .Slice 0) false}}`, "false"}, + {`{{or .Zero false}}`, "false"}, + {`{{or (index .Slice 1) true}}`, "1"}, + {`{{or .One true}}`, "1"}, + {`{{not (index .Slice 0)}}`, "true"}, + {`{{not .Zero}}`, "true"}, + {`{{not (index .Slice 1)}}`, "false"}, + {`{{not .One}}`, "false"}, + {`{{eq (index .Slice 0) .Zero}}`, "true"}, + {`{{eq (index .Slice 1) .One}}`, "true"}, + {`{{ne (index .Slice 0) .Zero}}`, "false"}, + {`{{ne (index .Slice 1) .One}}`, "false"}, + {`{{ge (index .Slice 0) .One}}`, "false"}, + {`{{ge (index .Slice 1) .Zero}}`, "true"}, + {`{{gt (index .Slice 0) .One}}`, "false"}, + {`{{gt (index .Slice 1) .Zero}}`, "true"}, + {`{{le (index .Slice 0) .One}}`, "true"}, + {`{{le (index .Slice 1) .Zero}}`, "false"}, + {`{{lt (index .Slice 0) .One}}`, "true"}, + {`{{lt (index .Slice 1) .Zero}}`, "false"}, + } + + for _, tt := range tests { + tmpl := Must(New("tmpl").Parse(tt.text)) + var buf bytes.Buffer + err := tmpl.Execute(&buf, map[string]interface{}{ + "PlusOne": func(n int) int { + return n + 1 + }, + "Slice": []int{0, 1, 2, 3}, + "One": 1, + "Two": 2, + "Nil": nil, + "Zero": 0, + }) + if strings.HasPrefix(tt.out, "ERROR:") { + e := strings.TrimSpace(strings.TrimPrefix(tt.out, "ERROR:")) + if err == nil || !strings.Contains(err.Error(), e) { + t.Errorf("%s: Execute: %v, want error %q", tt.text, err, e) + } + continue + } + if err != nil { + t.Errorf("%s: Execute: %v", tt.text, err) + continue + } + if buf.String() != tt.out { + t.Errorf("%s: template output = %q, want %q", tt.text, &buf, tt.out) + } + } +} + +// Check that panics during calls are recovered and returned as errors. +func TestExecutePanicDuringCall(t *testing.T) { + funcs := map[string]interface{}{ + "doPanic": func() string { + panic("custom panic string") + }, + } + tests := []struct { + name string + input string + data interface{} + wantErr string + }{ + { + "direct func call panics", + "{{doPanic}}", (*T)(nil), + `template: t:1:2: executing "t" at <doPanic>: error calling doPanic: custom panic string`, + }, + { + "indirect func call panics", + "{{call doPanic}}", (*T)(nil), + `template: t:1:7: executing "t" at <doPanic>: error calling doPanic: custom panic string`, + }, + { + "direct method call panics", + "{{.GetU}}", (*T)(nil), + `template: t:1:2: executing "t" at <.GetU>: error calling GetU: runtime error: invalid memory address or nil pointer dereference`, + }, + { + "indirect method call panics", + "{{call .GetU}}", (*T)(nil), + `template: t:1:7: executing "t" at <.GetU>: error calling GetU: runtime error: invalid memory address or nil pointer dereference`, + }, + { + "func field call panics", + "{{call .PanicFunc}}", tVal, + `template: t:1:2: executing "t" at <call .PanicFunc>: error calling call: test panic`, + }, + { + "method call on nil interface", + "{{.NonEmptyInterfaceNil.Method0}}", tVal, + `template: t:1:23: executing "t" at <.NonEmptyInterfaceNil.Method0>: nil pointer evaluating template.I.Method0`, + }, + } + for _, tc := range tests { + b := new(bytes.Buffer) + tmpl, err := New("t").Funcs(funcs).Parse(tc.input) + if err != nil { + t.Fatalf("parse error: %s", err) + } + err = tmpl.Execute(b, tc.data) + if err == nil { + t.Errorf("%s: expected error; got none", tc.name) + } else if !strings.Contains(err.Error(), tc.wantErr) { + if *debug { + fmt.Printf("%s: test execute error: %s\n", tc.name, err) + } + t.Errorf("%s: expected error:\n%s\ngot:\n%s", tc.name, tc.wantErr, err) + } + } +} + +// Issue 31810. Check that a parenthesized first argument behaves properly. +func TestIssue31810(t *testing.T) { + // A simple value with no arguments is fine. + var b bytes.Buffer + const text = "{{ (.) }}" + tmpl, err := New("").Parse(text) + if err != nil { + t.Error(err) + } + err = tmpl.Execute(&b, "result") + if err != nil { + t.Error(err) + } + if b.String() != "result" { + t.Errorf("%s got %q, expected %q", text, b.String(), "result") + } + + // Even a plain function fails - need to use call. + f := func() string { return "result" } + b.Reset() + err = tmpl.Execute(&b, f) + if err == nil { + t.Error("expected error with no call, got none") + } + + // Works if the function is explicitly called. + const textCall = "{{ (call .) }}" + tmpl, err = New("").Parse(textCall) + b.Reset() + err = tmpl.Execute(&b, f) + if err != nil { + t.Error(err) + } + if b.String() != "result" { + t.Errorf("%s got %q, expected %q", textCall, b.String(), "result") + } +} + +// Issue 43065, range over send only channel +func TestIssue43065(t *testing.T) { + var b bytes.Buffer + tmp := Must(New("").Parse(`{{range .}}{{end}}`)) + ch := make(chan<- int) + err := tmp.Execute(&b, ch) + if err == nil { + t.Error("expected err got nil") + } else if !strings.Contains(err.Error(), "range over send-only channel") { + t.Errorf("%s", err) + } +} + +// Issue 39807: data race in html/template & text/template +func TestIssue39807(t *testing.T) { + var wg sync.WaitGroup + + tplFoo, err := New("foo").Parse(`{{ template "bar" . }}`) + if err != nil { + t.Error(err) + } + + tplBar, err := New("bar").Parse("bar") + if err != nil { + t.Error(err) + } + + gofuncs := 10 + numTemplates := 10 + + for i := 1; i <= gofuncs; i++ { + wg.Add(1) + go func() { + defer wg.Done() + for j := 0; j < numTemplates; j++ { + _, err := tplFoo.AddParseTree(tplBar.Name(), tplBar.Tree) + if err != nil { + t.Error(err) + } + err = tplFoo.Execute(io.Discard, nil) + if err != nil { + t.Error(err) + } + } + }() + } + + wg.Wait() +} diff --git a/src/text/template/funcs.go b/src/text/template/funcs.go new file mode 100644 index 0000000..fff833e --- /dev/null +++ b/src/text/template/funcs.go @@ -0,0 +1,771 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package template + +import ( + "bytes" + "errors" + "fmt" + "io" + "net/url" + "reflect" + "strings" + "sync" + "unicode" + "unicode/utf8" +) + +// FuncMap is the type of the map defining the mapping from names to functions. +// Each function must have either a single return value, or two return values of +// which the second has type error. In that case, if the second (error) +// return value evaluates to non-nil during execution, execution terminates and +// Execute returns that error. +// +// Errors returned by Execute wrap the underlying error; call errors.As to +// uncover them. +// +// When template execution invokes a function with an argument list, that list +// must be assignable to the function's parameter types. Functions meant to +// apply to arguments of arbitrary type can use parameters of type interface{} or +// of type reflect.Value. Similarly, functions meant to return a result of arbitrary +// type can return interface{} or reflect.Value. +type FuncMap map[string]interface{} + +// builtins returns the FuncMap. +// It is not a global variable so the linker can dead code eliminate +// more when this isn't called. See golang.org/issue/36021. +// TODO: revert this back to a global map once golang.org/issue/2559 is fixed. +func builtins() FuncMap { + return FuncMap{ + "and": and, + "call": call, + "html": HTMLEscaper, + "index": index, + "slice": slice, + "js": JSEscaper, + "len": length, + "not": not, + "or": or, + "print": fmt.Sprint, + "printf": fmt.Sprintf, + "println": fmt.Sprintln, + "urlquery": URLQueryEscaper, + + // Comparisons + "eq": eq, // == + "ge": ge, // >= + "gt": gt, // > + "le": le, // <= + "lt": lt, // < + "ne": ne, // != + } +} + +var builtinFuncsOnce struct { + sync.Once + v map[string]reflect.Value +} + +// builtinFuncsOnce lazily computes & caches the builtinFuncs map. +// TODO: revert this back to a global map once golang.org/issue/2559 is fixed. +func builtinFuncs() map[string]reflect.Value { + builtinFuncsOnce.Do(func() { + builtinFuncsOnce.v = createValueFuncs(builtins()) + }) + return builtinFuncsOnce.v +} + +// createValueFuncs turns a FuncMap into a map[string]reflect.Value +func createValueFuncs(funcMap FuncMap) map[string]reflect.Value { + m := make(map[string]reflect.Value) + addValueFuncs(m, funcMap) + return m +} + +// addValueFuncs adds to values the functions in funcs, converting them to reflect.Values. +func addValueFuncs(out map[string]reflect.Value, in FuncMap) { + for name, fn := range in { + if !goodName(name) { + panic(fmt.Errorf("function name %q is not a valid identifier", name)) + } + v := reflect.ValueOf(fn) + if v.Kind() != reflect.Func { + panic("value for " + name + " not a function") + } + if !goodFunc(v.Type()) { + panic(fmt.Errorf("can't install method/function %q with %d results", name, v.Type().NumOut())) + } + out[name] = v + } +} + +// addFuncs adds to values the functions in funcs. It does no checking of the input - +// call addValueFuncs first. +func addFuncs(out, in FuncMap) { + for name, fn := range in { + out[name] = fn + } +} + +// goodFunc reports whether the function or method has the right result signature. +func goodFunc(typ reflect.Type) bool { + // We allow functions with 1 result or 2 results where the second is an error. + switch { + case typ.NumOut() == 1: + return true + case typ.NumOut() == 2 && typ.Out(1) == errorType: + return true + } + return false +} + +// goodName reports whether the function name is a valid identifier. +func goodName(name string) bool { + if name == "" { + return false + } + for i, r := range name { + switch { + case r == '_': + case i == 0 && !unicode.IsLetter(r): + return false + case !unicode.IsLetter(r) && !unicode.IsDigit(r): + return false + } + } + return true +} + +// findFunction looks for a function in the template, and global map. +func findFunction(name string, tmpl *Template) (reflect.Value, bool) { + if tmpl != nil && tmpl.common != nil { + tmpl.muFuncs.RLock() + defer tmpl.muFuncs.RUnlock() + if fn := tmpl.execFuncs[name]; fn.IsValid() { + return fn, true + } + } + if fn := builtinFuncs()[name]; fn.IsValid() { + return fn, true + } + return reflect.Value{}, false +} + +// prepareArg checks if value can be used as an argument of type argType, and +// converts an invalid value to appropriate zero if possible. +func prepareArg(value reflect.Value, argType reflect.Type) (reflect.Value, error) { + if !value.IsValid() { + if !canBeNil(argType) { + return reflect.Value{}, fmt.Errorf("value is nil; should be of type %s", argType) + } + value = reflect.Zero(argType) + } + if value.Type().AssignableTo(argType) { + return value, nil + } + if intLike(value.Kind()) && intLike(argType.Kind()) && value.Type().ConvertibleTo(argType) { + value = value.Convert(argType) + return value, nil + } + return reflect.Value{}, fmt.Errorf("value has type %s; should be %s", value.Type(), argType) +} + +func intLike(typ reflect.Kind) bool { + switch typ { + case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64: + return true + case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr: + return true + } + return false +} + +// indexArg checks if a reflect.Value can be used as an index, and converts it to int if possible. +func indexArg(index reflect.Value, cap int) (int, error) { + var x int64 + switch index.Kind() { + case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64: + x = index.Int() + case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr: + x = int64(index.Uint()) + case reflect.Invalid: + return 0, fmt.Errorf("cannot index slice/array with nil") + default: + return 0, fmt.Errorf("cannot index slice/array with type %s", index.Type()) + } + if x < 0 || int(x) < 0 || int(x) > cap { + return 0, fmt.Errorf("index out of range: %d", x) + } + return int(x), nil +} + +// Indexing. + +// index returns the result of indexing its first argument by the following +// arguments. Thus "index x 1 2 3" is, in Go syntax, x[1][2][3]. Each +// indexed item must be a map, slice, or array. +func index(item reflect.Value, indexes ...reflect.Value) (reflect.Value, error) { + item = indirectInterface(item) + if !item.IsValid() { + return reflect.Value{}, fmt.Errorf("index of untyped nil") + } + for _, index := range indexes { + index = indirectInterface(index) + var isNil bool + if item, isNil = indirect(item); isNil { + return reflect.Value{}, fmt.Errorf("index of nil pointer") + } + switch item.Kind() { + case reflect.Array, reflect.Slice, reflect.String: + x, err := indexArg(index, item.Len()) + if err != nil { + return reflect.Value{}, err + } + item = item.Index(x) + case reflect.Map: + index, err := prepareArg(index, item.Type().Key()) + if err != nil { + return reflect.Value{}, err + } + if x := item.MapIndex(index); x.IsValid() { + item = x + } else { + item = reflect.Zero(item.Type().Elem()) + } + case reflect.Invalid: + // the loop holds invariant: item.IsValid() + panic("unreachable") + default: + return reflect.Value{}, fmt.Errorf("can't index item of type %s", item.Type()) + } + } + return item, nil +} + +// Slicing. + +// slice returns the result of slicing its first argument by the remaining +// arguments. Thus "slice x 1 2" is, in Go syntax, x[1:2], while "slice x" +// is x[:], "slice x 1" is x[1:], and "slice x 1 2 3" is x[1:2:3]. The first +// argument must be a string, slice, or array. +func slice(item reflect.Value, indexes ...reflect.Value) (reflect.Value, error) { + item = indirectInterface(item) + if !item.IsValid() { + return reflect.Value{}, fmt.Errorf("slice of untyped nil") + } + if len(indexes) > 3 { + return reflect.Value{}, fmt.Errorf("too many slice indexes: %d", len(indexes)) + } + var cap int + switch item.Kind() { + case reflect.String: + if len(indexes) == 3 { + return reflect.Value{}, fmt.Errorf("cannot 3-index slice a string") + } + cap = item.Len() + case reflect.Array, reflect.Slice: + cap = item.Cap() + default: + return reflect.Value{}, fmt.Errorf("can't slice item of type %s", item.Type()) + } + // set default values for cases item[:], item[i:]. + idx := [3]int{0, item.Len()} + for i, index := range indexes { + x, err := indexArg(index, cap) + if err != nil { + return reflect.Value{}, err + } + idx[i] = x + } + // given item[i:j], make sure i <= j. + if idx[0] > idx[1] { + return reflect.Value{}, fmt.Errorf("invalid slice index: %d > %d", idx[0], idx[1]) + } + if len(indexes) < 3 { + return item.Slice(idx[0], idx[1]), nil + } + // given item[i:j:k], make sure i <= j <= k. + if idx[1] > idx[2] { + return reflect.Value{}, fmt.Errorf("invalid slice index: %d > %d", idx[1], idx[2]) + } + return item.Slice3(idx[0], idx[1], idx[2]), nil +} + +// Length + +// length returns the length of the item, with an error if it has no defined length. +func length(item reflect.Value) (int, error) { + item, isNil := indirect(item) + if isNil { + return 0, fmt.Errorf("len of nil pointer") + } + switch item.Kind() { + case reflect.Array, reflect.Chan, reflect.Map, reflect.Slice, reflect.String: + return item.Len(), nil + } + return 0, fmt.Errorf("len of type %s", item.Type()) +} + +// Function invocation + +// call returns the result of evaluating the first argument as a function. +// The function must return 1 result, or 2 results, the second of which is an error. +func call(fn reflect.Value, args ...reflect.Value) (reflect.Value, error) { + fn = indirectInterface(fn) + if !fn.IsValid() { + return reflect.Value{}, fmt.Errorf("call of nil") + } + typ := fn.Type() + if typ.Kind() != reflect.Func { + return reflect.Value{}, fmt.Errorf("non-function of type %s", typ) + } + if !goodFunc(typ) { + return reflect.Value{}, fmt.Errorf("function called with %d args; should be 1 or 2", typ.NumOut()) + } + numIn := typ.NumIn() + var dddType reflect.Type + if typ.IsVariadic() { + if len(args) < numIn-1 { + return reflect.Value{}, fmt.Errorf("wrong number of args: got %d want at least %d", len(args), numIn-1) + } + dddType = typ.In(numIn - 1).Elem() + } else { + if len(args) != numIn { + return reflect.Value{}, fmt.Errorf("wrong number of args: got %d want %d", len(args), numIn) + } + } + argv := make([]reflect.Value, len(args)) + for i, arg := range args { + arg = indirectInterface(arg) + // Compute the expected type. Clumsy because of variadics. + argType := dddType + if !typ.IsVariadic() || i < numIn-1 { + argType = typ.In(i) + } + + var err error + if argv[i], err = prepareArg(arg, argType); err != nil { + return reflect.Value{}, fmt.Errorf("arg %d: %w", i, err) + } + } + return safeCall(fn, argv) +} + +// safeCall runs fun.Call(args), and returns the resulting value and error, if +// any. If the call panics, the panic value is returned as an error. +func safeCall(fun reflect.Value, args []reflect.Value) (val reflect.Value, err error) { + defer func() { + if r := recover(); r != nil { + if e, ok := r.(error); ok { + err = e + } else { + err = fmt.Errorf("%v", r) + } + } + }() + ret := fun.Call(args) + if len(ret) == 2 && !ret[1].IsNil() { + return ret[0], ret[1].Interface().(error) + } + return ret[0], nil +} + +// Boolean logic. + +func truth(arg reflect.Value) bool { + t, _ := isTrue(indirectInterface(arg)) + return t +} + +// and computes the Boolean AND of its arguments, returning +// the first false argument it encounters, or the last argument. +func and(arg0 reflect.Value, args ...reflect.Value) reflect.Value { + if !truth(arg0) { + return arg0 + } + for i := range args { + arg0 = args[i] + if !truth(arg0) { + break + } + } + return arg0 +} + +// or computes the Boolean OR of its arguments, returning +// the first true argument it encounters, or the last argument. +func or(arg0 reflect.Value, args ...reflect.Value) reflect.Value { + if truth(arg0) { + return arg0 + } + for i := range args { + arg0 = args[i] + if truth(arg0) { + break + } + } + return arg0 +} + +// not returns the Boolean negation of its argument. +func not(arg reflect.Value) bool { + return !truth(arg) +} + +// Comparison. + +// TODO: Perhaps allow comparison between signed and unsigned integers. + +var ( + errBadComparisonType = errors.New("invalid type for comparison") + errBadComparison = errors.New("incompatible types for comparison") + errNoComparison = errors.New("missing argument for comparison") +) + +type kind int + +const ( + invalidKind kind = iota + boolKind + complexKind + intKind + floatKind + stringKind + uintKind +) + +func basicKind(v reflect.Value) (kind, error) { + switch v.Kind() { + case reflect.Bool: + return boolKind, nil + case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64: + return intKind, nil + case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr: + return uintKind, nil + case reflect.Float32, reflect.Float64: + return floatKind, nil + case reflect.Complex64, reflect.Complex128: + return complexKind, nil + case reflect.String: + return stringKind, nil + } + return invalidKind, errBadComparisonType +} + +// eq evaluates the comparison a == b || a == c || ... +func eq(arg1 reflect.Value, arg2 ...reflect.Value) (bool, error) { + arg1 = indirectInterface(arg1) + if arg1 != zero { + if t1 := arg1.Type(); !t1.Comparable() { + return false, fmt.Errorf("uncomparable type %s: %v", t1, arg1) + } + } + if len(arg2) == 0 { + return false, errNoComparison + } + k1, _ := basicKind(arg1) + for _, arg := range arg2 { + arg = indirectInterface(arg) + k2, _ := basicKind(arg) + truth := false + if k1 != k2 { + // Special case: Can compare integer values regardless of type's sign. + switch { + case k1 == intKind && k2 == uintKind: + truth = arg1.Int() >= 0 && uint64(arg1.Int()) == arg.Uint() + case k1 == uintKind && k2 == intKind: + truth = arg.Int() >= 0 && arg1.Uint() == uint64(arg.Int()) + default: + if arg1 != zero && arg != zero { + return false, errBadComparison + } + } + } else { + switch k1 { + case boolKind: + truth = arg1.Bool() == arg.Bool() + case complexKind: + truth = arg1.Complex() == arg.Complex() + case floatKind: + truth = arg1.Float() == arg.Float() + case intKind: + truth = arg1.Int() == arg.Int() + case stringKind: + truth = arg1.String() == arg.String() + case uintKind: + truth = arg1.Uint() == arg.Uint() + default: + if arg == zero || arg1 == zero { + truth = arg1 == arg + } else { + if t2 := arg.Type(); !t2.Comparable() { + return false, fmt.Errorf("uncomparable type %s: %v", t2, arg) + } + truth = arg1.Interface() == arg.Interface() + } + } + } + if truth { + return true, nil + } + } + return false, nil +} + +// ne evaluates the comparison a != b. +func ne(arg1, arg2 reflect.Value) (bool, error) { + // != is the inverse of ==. + equal, err := eq(arg1, arg2) + return !equal, err +} + +// lt evaluates the comparison a < b. +func lt(arg1, arg2 reflect.Value) (bool, error) { + arg1 = indirectInterface(arg1) + k1, err := basicKind(arg1) + if err != nil { + return false, err + } + arg2 = indirectInterface(arg2) + k2, err := basicKind(arg2) + if err != nil { + return false, err + } + truth := false + if k1 != k2 { + // Special case: Can compare integer values regardless of type's sign. + switch { + case k1 == intKind && k2 == uintKind: + truth = arg1.Int() < 0 || uint64(arg1.Int()) < arg2.Uint() + case k1 == uintKind && k2 == intKind: + truth = arg2.Int() >= 0 && arg1.Uint() < uint64(arg2.Int()) + default: + return false, errBadComparison + } + } else { + switch k1 { + case boolKind, complexKind: + return false, errBadComparisonType + case floatKind: + truth = arg1.Float() < arg2.Float() + case intKind: + truth = arg1.Int() < arg2.Int() + case stringKind: + truth = arg1.String() < arg2.String() + case uintKind: + truth = arg1.Uint() < arg2.Uint() + default: + panic("invalid kind") + } + } + return truth, nil +} + +// le evaluates the comparison <= b. +func le(arg1, arg2 reflect.Value) (bool, error) { + // <= is < or ==. + lessThan, err := lt(arg1, arg2) + if lessThan || err != nil { + return lessThan, err + } + return eq(arg1, arg2) +} + +// gt evaluates the comparison a > b. +func gt(arg1, arg2 reflect.Value) (bool, error) { + // > is the inverse of <=. + lessOrEqual, err := le(arg1, arg2) + if err != nil { + return false, err + } + return !lessOrEqual, nil +} + +// ge evaluates the comparison a >= b. +func ge(arg1, arg2 reflect.Value) (bool, error) { + // >= is the inverse of <. + lessThan, err := lt(arg1, arg2) + if err != nil { + return false, err + } + return !lessThan, nil +} + +// HTML escaping. + +var ( + htmlQuot = []byte(""") // shorter than """ + htmlApos = []byte("'") // shorter than "'" and apos was not in HTML until HTML5 + htmlAmp = []byte("&") + htmlLt = []byte("<") + htmlGt = []byte(">") + htmlNull = []byte("\uFFFD") +) + +// HTMLEscape writes to w the escaped HTML equivalent of the plain text data b. +func HTMLEscape(w io.Writer, b []byte) { + last := 0 + for i, c := range b { + var html []byte + switch c { + case '\000': + html = htmlNull + case '"': + html = htmlQuot + case '\'': + html = htmlApos + case '&': + html = htmlAmp + case '<': + html = htmlLt + case '>': + html = htmlGt + default: + continue + } + w.Write(b[last:i]) + w.Write(html) + last = i + 1 + } + w.Write(b[last:]) +} + +// HTMLEscapeString returns the escaped HTML equivalent of the plain text data s. +func HTMLEscapeString(s string) string { + // Avoid allocation if we can. + if !strings.ContainsAny(s, "'\"&<>\000") { + return s + } + var b bytes.Buffer + HTMLEscape(&b, []byte(s)) + return b.String() +} + +// HTMLEscaper returns the escaped HTML equivalent of the textual +// representation of its arguments. +func HTMLEscaper(args ...interface{}) string { + return HTMLEscapeString(evalArgs(args)) +} + +// JavaScript escaping. + +var ( + jsLowUni = []byte(`\u00`) + hex = []byte("0123456789ABCDEF") + + jsBackslash = []byte(`\\`) + jsApos = []byte(`\'`) + jsQuot = []byte(`\"`) + jsLt = []byte(`\u003C`) + jsGt = []byte(`\u003E`) + jsAmp = []byte(`\u0026`) + jsEq = []byte(`\u003D`) +) + +// JSEscape writes to w the escaped JavaScript equivalent of the plain text data b. +func JSEscape(w io.Writer, b []byte) { + last := 0 + for i := 0; i < len(b); i++ { + c := b[i] + + if !jsIsSpecial(rune(c)) { + // fast path: nothing to do + continue + } + w.Write(b[last:i]) + + if c < utf8.RuneSelf { + // Quotes, slashes and angle brackets get quoted. + // Control characters get written as \u00XX. + switch c { + case '\\': + w.Write(jsBackslash) + case '\'': + w.Write(jsApos) + case '"': + w.Write(jsQuot) + case '<': + w.Write(jsLt) + case '>': + w.Write(jsGt) + case '&': + w.Write(jsAmp) + case '=': + w.Write(jsEq) + default: + w.Write(jsLowUni) + t, b := c>>4, c&0x0f + w.Write(hex[t : t+1]) + w.Write(hex[b : b+1]) + } + } else { + // Unicode rune. + r, size := utf8.DecodeRune(b[i:]) + if unicode.IsPrint(r) { + w.Write(b[i : i+size]) + } else { + fmt.Fprintf(w, "\\u%04X", r) + } + i += size - 1 + } + last = i + 1 + } + w.Write(b[last:]) +} + +// JSEscapeString returns the escaped JavaScript equivalent of the plain text data s. +func JSEscapeString(s string) string { + // Avoid allocation if we can. + if strings.IndexFunc(s, jsIsSpecial) < 0 { + return s + } + var b bytes.Buffer + JSEscape(&b, []byte(s)) + return b.String() +} + +func jsIsSpecial(r rune) bool { + switch r { + case '\\', '\'', '"', '<', '>', '&', '=': + return true + } + return r < ' ' || utf8.RuneSelf <= r +} + +// JSEscaper returns the escaped JavaScript equivalent of the textual +// representation of its arguments. +func JSEscaper(args ...interface{}) string { + return JSEscapeString(evalArgs(args)) +} + +// URLQueryEscaper returns the escaped value of the textual representation of +// its arguments in a form suitable for embedding in a URL query. +func URLQueryEscaper(args ...interface{}) string { + return url.QueryEscape(evalArgs(args)) +} + +// evalArgs formats the list of arguments into a string. It is therefore equivalent to +// fmt.Sprint(args...) +// except that each argument is indirected (if a pointer), as required, +// using the same rules as the default string evaluation during template +// execution. +func evalArgs(args []interface{}) string { + ok := false + var s string + // Fast path for simple common case. + if len(args) == 1 { + s, ok = args[0].(string) + } + if !ok { + for i, arg := range args { + a, ok := printableValue(reflect.ValueOf(arg)) + if ok { + args[i] = a + } // else let fmt do its thing + } + s = fmt.Sprint(args...) + } + return s +} diff --git a/src/text/template/helper.go b/src/text/template/helper.go new file mode 100644 index 0000000..57905e6 --- /dev/null +++ b/src/text/template/helper.go @@ -0,0 +1,177 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +// Helper functions to make constructing templates easier. + +package template + +import ( + "fmt" + "io/fs" + "os" + "path" + "path/filepath" +) + +// Functions and methods to parse templates. + +// Must is a helper that wraps a call to a function returning (*Template, error) +// and panics if the error is non-nil. It is intended for use in variable +// initializations such as +// var t = template.Must(template.New("name").Parse("text")) +func Must(t *Template, err error) *Template { + if err != nil { + panic(err) + } + return t +} + +// ParseFiles creates a new Template and parses the template definitions from +// the named files. The returned template's name will have the base name and +// parsed contents of the first file. There must be at least one file. +// If an error occurs, parsing stops and the returned *Template is nil. +// +// When parsing multiple files with the same name in different directories, +// the last one mentioned will be the one that results. +// For instance, ParseFiles("a/foo", "b/foo") stores "b/foo" as the template +// named "foo", while "a/foo" is unavailable. +func ParseFiles(filenames ...string) (*Template, error) { + return parseFiles(nil, readFileOS, filenames...) +} + +// ParseFiles parses the named files and associates the resulting templates with +// t. If an error occurs, parsing stops and the returned template is nil; +// otherwise it is t. There must be at least one file. +// Since the templates created by ParseFiles are named by the base +// names of the argument files, t should usually have the name of one +// of the (base) names of the files. If it does not, depending on t's +// contents before calling ParseFiles, t.Execute may fail. In that +// case use t.ExecuteTemplate to execute a valid template. +// +// When parsing multiple files with the same name in different directories, +// the last one mentioned will be the one that results. +func (t *Template) ParseFiles(filenames ...string) (*Template, error) { + t.init() + return parseFiles(t, readFileOS, filenames...) +} + +// parseFiles is the helper for the method and function. If the argument +// template is nil, it is created from the first file. +func parseFiles(t *Template, readFile func(string) (string, []byte, error), filenames ...string) (*Template, error) { + if len(filenames) == 0 { + // Not really a problem, but be consistent. + return nil, fmt.Errorf("template: no files named in call to ParseFiles") + } + for _, filename := range filenames { + name, b, err := readFile(filename) + if err != nil { + return nil, err + } + s := string(b) + // First template becomes return value if not already defined, + // and we use that one for subsequent New calls to associate + // all the templates together. Also, if this file has the same name + // as t, this file becomes the contents of t, so + // t, err := New(name).Funcs(xxx).ParseFiles(name) + // works. Otherwise we create a new template associated with t. + var tmpl *Template + if t == nil { + t = New(name) + } + if name == t.Name() { + tmpl = t + } else { + tmpl = t.New(name) + } + _, err = tmpl.Parse(s) + if err != nil { + return nil, err + } + } + return t, nil +} + +// ParseGlob creates a new Template and parses the template definitions from +// the files identified by the pattern. The files are matched according to the +// semantics of filepath.Match, and the pattern must match at least one file. +// The returned template will have the (base) name and (parsed) contents of the +// first file matched by the pattern. ParseGlob is equivalent to calling +// ParseFiles with the list of files matched by the pattern. +// +// When parsing multiple files with the same name in different directories, +// the last one mentioned will be the one that results. +func ParseGlob(pattern string) (*Template, error) { + return parseGlob(nil, pattern) +} + +// ParseGlob parses the template definitions in the files identified by the +// pattern and associates the resulting templates with t. The files are matched +// according to the semantics of filepath.Match, and the pattern must match at +// least one file. ParseGlob is equivalent to calling t.ParseFiles with the +// list of files matched by the pattern. +// +// When parsing multiple files with the same name in different directories, +// the last one mentioned will be the one that results. +func (t *Template) ParseGlob(pattern string) (*Template, error) { + t.init() + return parseGlob(t, pattern) +} + +// parseGlob is the implementation of the function and method ParseGlob. +func parseGlob(t *Template, pattern string) (*Template, error) { + filenames, err := filepath.Glob(pattern) + if err != nil { + return nil, err + } + if len(filenames) == 0 { + return nil, fmt.Errorf("template: pattern matches no files: %#q", pattern) + } + return parseFiles(t, readFileOS, filenames...) +} + +// ParseFS is like ParseFiles or ParseGlob but reads from the file system fsys +// instead of the host operating system's file system. +// It accepts a list of glob patterns. +// (Note that most file names serve as glob patterns matching only themselves.) +func ParseFS(fsys fs.FS, patterns ...string) (*Template, error) { + return parseFS(nil, fsys, patterns) +} + +// ParseFS is like ParseFiles or ParseGlob but reads from the file system fsys +// instead of the host operating system's file system. +// It accepts a list of glob patterns. +// (Note that most file names serve as glob patterns matching only themselves.) +func (t *Template) ParseFS(fsys fs.FS, patterns ...string) (*Template, error) { + t.init() + return parseFS(t, fsys, patterns) +} + +func parseFS(t *Template, fsys fs.FS, patterns []string) (*Template, error) { + var filenames []string + for _, pattern := range patterns { + list, err := fs.Glob(fsys, pattern) + if err != nil { + return nil, err + } + if len(list) == 0 { + return nil, fmt.Errorf("template: pattern matches no files: %#q", pattern) + } + filenames = append(filenames, list...) + } + return parseFiles(t, readFileFS(fsys), filenames...) +} + +func readFileOS(file string) (name string, b []byte, err error) { + name = filepath.Base(file) + b, err = os.ReadFile(file) + return +} + +func readFileFS(fsys fs.FS) func(string) (string, []byte, error) { + return func(file string) (name string, b []byte, err error) { + name = path.Base(file) + b, err = fs.ReadFile(fsys, file) + return + } +} diff --git a/src/text/template/link_test.go b/src/text/template/link_test.go new file mode 100644 index 0000000..e1d3136 --- /dev/null +++ b/src/text/template/link_test.go @@ -0,0 +1,59 @@ +// Copyright 2019 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package template_test + +import ( + "bytes" + "internal/testenv" + "os" + "os/exec" + "path/filepath" + "testing" +) + +// Issue 36021: verify that text/template doesn't prevent the linker from removing +// unused methods. +func TestLinkerGC(t *testing.T) { + if testing.Short() { + t.Skip("skipping in short mode") + } + testenv.MustHaveGoBuild(t) + const prog = `package main + +import ( + _ "text/template" +) + +type T struct{} + +func (t *T) Unused() { println("THIS SHOULD BE ELIMINATED") } +func (t *T) Used() {} + +var sink *T + +func main() { + var t T + sink = &t + t.Used() +} +` + td := t.TempDir() + + if err := os.WriteFile(filepath.Join(td, "x.go"), []byte(prog), 0644); err != nil { + t.Fatal(err) + } + cmd := exec.Command(testenv.GoToolPath(t), "build", "-o", "x.exe", "x.go") + cmd.Dir = td + if out, err := cmd.CombinedOutput(); err != nil { + t.Fatalf("go build: %v, %s", err, out) + } + slurp, err := os.ReadFile(filepath.Join(td, "x.exe")) + if err != nil { + t.Fatal(err) + } + if bytes.Contains(slurp, []byte("THIS SHOULD BE ELIMINATED")) { + t.Error("binary contains code that should be deadcode eliminated") + } +} diff --git a/src/text/template/multi_test.go b/src/text/template/multi_test.go new file mode 100644 index 0000000..6b81ffe --- /dev/null +++ b/src/text/template/multi_test.go @@ -0,0 +1,464 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package template + +// Tests for multiple-template parsing and execution. + +import ( + "bytes" + "fmt" + "os" + "testing" + "text/template/parse" +) + +const ( + noError = true + hasError = false +) + +type multiParseTest struct { + name string + input string + ok bool + names []string + results []string +} + +var multiParseTests = []multiParseTest{ + {"empty", "", noError, + nil, + nil}, + {"one", `{{define "foo"}} FOO {{end}}`, noError, + []string{"foo"}, + []string{" FOO "}}, + {"two", `{{define "foo"}} FOO {{end}}{{define "bar"}} BAR {{end}}`, noError, + []string{"foo", "bar"}, + []string{" FOO ", " BAR "}}, + // errors + {"missing end", `{{define "foo"}} FOO `, hasError, + nil, + nil}, + {"malformed name", `{{define "foo}} FOO `, hasError, + nil, + nil}, +} + +func TestMultiParse(t *testing.T) { + for _, test := range multiParseTests { + template, err := New("root").Parse(test.input) + switch { + case err == nil && !test.ok: + t.Errorf("%q: expected error; got none", test.name) + continue + case err != nil && test.ok: + t.Errorf("%q: unexpected error: %v", test.name, err) + continue + case err != nil && !test.ok: + // expected error, got one + if *debug { + fmt.Printf("%s: %s\n\t%s\n", test.name, test.input, err) + } + continue + } + if template == nil { + continue + } + if len(template.tmpl) != len(test.names)+1 { // +1 for root + t.Errorf("%s: wrong number of templates; wanted %d got %d", test.name, len(test.names), len(template.tmpl)) + continue + } + for i, name := range test.names { + tmpl, ok := template.tmpl[name] + if !ok { + t.Errorf("%s: can't find template %q", test.name, name) + continue + } + result := tmpl.Root.String() + if result != test.results[i] { + t.Errorf("%s=(%q): got\n\t%v\nexpected\n\t%v", test.name, test.input, result, test.results[i]) + } + } + } +} + +var multiExecTests = []execTest{ + {"empty", "", "", nil, true}, + {"text", "some text", "some text", nil, true}, + {"invoke x", `{{template "x" .SI}}`, "TEXT", tVal, true}, + {"invoke x no args", `{{template "x"}}`, "TEXT", tVal, true}, + {"invoke dot int", `{{template "dot" .I}}`, "17", tVal, true}, + {"invoke dot []int", `{{template "dot" .SI}}`, "[3 4 5]", tVal, true}, + {"invoke dotV", `{{template "dotV" .U}}`, "v", tVal, true}, + {"invoke nested int", `{{template "nested" .I}}`, "17", tVal, true}, + {"variable declared by template", `{{template "nested" $x:=.SI}},{{index $x 1}}`, "[3 4 5],4", tVal, true}, + + // User-defined function: test argument evaluator. + {"testFunc literal", `{{oneArg "joe"}}`, "oneArg=joe", tVal, true}, + {"testFunc .", `{{oneArg .}}`, "oneArg=joe", "joe", true}, +} + +// These strings are also in testdata/*. +const multiText1 = ` + {{define "x"}}TEXT{{end}} + {{define "dotV"}}{{.V}}{{end}} +` + +const multiText2 = ` + {{define "dot"}}{{.}}{{end}} + {{define "nested"}}{{template "dot" .}}{{end}} +` + +func TestMultiExecute(t *testing.T) { + // Declare a couple of templates first. + template, err := New("root").Parse(multiText1) + if err != nil { + t.Fatalf("parse error for 1: %s", err) + } + _, err = template.Parse(multiText2) + if err != nil { + t.Fatalf("parse error for 2: %s", err) + } + testExecute(multiExecTests, template, t) +} + +func TestParseFiles(t *testing.T) { + _, err := ParseFiles("DOES NOT EXIST") + if err == nil { + t.Error("expected error for non-existent file; got none") + } + template := New("root") + _, err = template.ParseFiles("testdata/file1.tmpl", "testdata/file2.tmpl") + if err != nil { + t.Fatalf("error parsing files: %v", err) + } + testExecute(multiExecTests, template, t) +} + +func TestParseGlob(t *testing.T) { + _, err := ParseGlob("DOES NOT EXIST") + if err == nil { + t.Error("expected error for non-existent file; got none") + } + _, err = New("error").ParseGlob("[x") + if err == nil { + t.Error("expected error for bad pattern; got none") + } + template := New("root") + _, err = template.ParseGlob("testdata/file*.tmpl") + if err != nil { + t.Fatalf("error parsing files: %v", err) + } + testExecute(multiExecTests, template, t) +} + +func TestParseFS(t *testing.T) { + fs := os.DirFS("testdata") + + { + _, err := ParseFS(fs, "DOES NOT EXIST") + if err == nil { + t.Error("expected error for non-existent file; got none") + } + } + + { + template := New("root") + _, err := template.ParseFS(fs, "file1.tmpl", "file2.tmpl") + if err != nil { + t.Fatalf("error parsing files: %v", err) + } + testExecute(multiExecTests, template, t) + } + + { + template := New("root") + _, err := template.ParseFS(fs, "file*.tmpl") + if err != nil { + t.Fatalf("error parsing files: %v", err) + } + testExecute(multiExecTests, template, t) + } +} + +// In these tests, actual content (not just template definitions) comes from the parsed files. + +var templateFileExecTests = []execTest{ + {"test", `{{template "tmpl1.tmpl"}}{{template "tmpl2.tmpl"}}`, "template1\n\ny\ntemplate2\n\nx\n", 0, true}, +} + +func TestParseFilesWithData(t *testing.T) { + template, err := New("root").ParseFiles("testdata/tmpl1.tmpl", "testdata/tmpl2.tmpl") + if err != nil { + t.Fatalf("error parsing files: %v", err) + } + testExecute(templateFileExecTests, template, t) +} + +func TestParseGlobWithData(t *testing.T) { + template, err := New("root").ParseGlob("testdata/tmpl*.tmpl") + if err != nil { + t.Fatalf("error parsing files: %v", err) + } + testExecute(templateFileExecTests, template, t) +} + +const ( + cloneText1 = `{{define "a"}}{{template "b"}}{{template "c"}}{{end}}` + cloneText2 = `{{define "b"}}b{{end}}` + cloneText3 = `{{define "c"}}root{{end}}` + cloneText4 = `{{define "c"}}clone{{end}}` +) + +func TestClone(t *testing.T) { + // Create some templates and clone the root. + root, err := New("root").Parse(cloneText1) + if err != nil { + t.Fatal(err) + } + _, err = root.Parse(cloneText2) + if err != nil { + t.Fatal(err) + } + clone := Must(root.Clone()) + // Add variants to both. + _, err = root.Parse(cloneText3) + if err != nil { + t.Fatal(err) + } + _, err = clone.Parse(cloneText4) + if err != nil { + t.Fatal(err) + } + // Verify that the clone is self-consistent. + for k, v := range clone.tmpl { + if k == clone.name && v.tmpl[k] != clone { + t.Error("clone does not contain root") + } + if v != v.tmpl[v.name] { + t.Errorf("clone does not contain self for %q", k) + } + } + // Execute root. + var b bytes.Buffer + err = root.ExecuteTemplate(&b, "a", 0) + if err != nil { + t.Fatal(err) + } + if b.String() != "broot" { + t.Errorf("expected %q got %q", "broot", b.String()) + } + // Execute copy. + b.Reset() + err = clone.ExecuteTemplate(&b, "a", 0) + if err != nil { + t.Fatal(err) + } + if b.String() != "bclone" { + t.Errorf("expected %q got %q", "bclone", b.String()) + } +} + +func TestAddParseTree(t *testing.T) { + // Create some templates. + root, err := New("root").Parse(cloneText1) + if err != nil { + t.Fatal(err) + } + _, err = root.Parse(cloneText2) + if err != nil { + t.Fatal(err) + } + // Add a new parse tree. + tree, err := parse.Parse("cloneText3", cloneText3, "", "", nil, builtins()) + if err != nil { + t.Fatal(err) + } + added, err := root.AddParseTree("c", tree["c"]) + if err != nil { + t.Fatal(err) + } + // Execute. + var b bytes.Buffer + err = added.ExecuteTemplate(&b, "a", 0) + if err != nil { + t.Fatal(err) + } + if b.String() != "broot" { + t.Errorf("expected %q got %q", "broot", b.String()) + } +} + +// Issue 7032 +func TestAddParseTreeToUnparsedTemplate(t *testing.T) { + master := "{{define \"master\"}}{{end}}" + tmpl := New("master") + tree, err := parse.Parse("master", master, "", "", nil) + if err != nil { + t.Fatalf("unexpected parse err: %v", err) + } + masterTree := tree["master"] + tmpl.AddParseTree("master", masterTree) // used to panic +} + +func TestRedefinition(t *testing.T) { + var tmpl *Template + var err error + if tmpl, err = New("tmpl1").Parse(`{{define "test"}}foo{{end}}`); err != nil { + t.Fatalf("parse 1: %v", err) + } + if _, err = tmpl.Parse(`{{define "test"}}bar{{end}}`); err != nil { + t.Fatalf("got error %v, expected nil", err) + } + if _, err = tmpl.New("tmpl2").Parse(`{{define "test"}}bar{{end}}`); err != nil { + t.Fatalf("got error %v, expected nil", err) + } +} + +// Issue 10879 +func TestEmptyTemplateCloneCrash(t *testing.T) { + t1 := New("base") + t1.Clone() // used to panic +} + +// Issue 10910, 10926 +func TestTemplateLookUp(t *testing.T) { + t1 := New("foo") + if t1.Lookup("foo") != nil { + t.Error("Lookup returned non-nil value for undefined template foo") + } + t1.New("bar") + if t1.Lookup("bar") != nil { + t.Error("Lookup returned non-nil value for undefined template bar") + } + t1.Parse(`{{define "foo"}}test{{end}}`) + if t1.Lookup("foo") == nil { + t.Error("Lookup returned nil value for defined template") + } +} + +func TestNew(t *testing.T) { + // template with same name already exists + t1, _ := New("test").Parse(`{{define "test"}}foo{{end}}`) + t2 := t1.New("test") + + if t1.common != t2.common { + t.Errorf("t1 & t2 didn't share common struct; got %v != %v", t1.common, t2.common) + } + if t1.Tree == nil { + t.Error("defined template got nil Tree") + } + if t2.Tree != nil { + t.Error("undefined template got non-nil Tree") + } + + containsT1 := false + for _, tmpl := range t1.Templates() { + if tmpl == t2 { + t.Error("Templates included undefined template") + } + if tmpl == t1 { + containsT1 = true + } + } + if !containsT1 { + t.Error("Templates didn't include defined template") + } +} + +func TestParse(t *testing.T) { + // In multiple calls to Parse with the same receiver template, only one call + // can contain text other than space, comments, and template definitions + t1 := New("test") + if _, err := t1.Parse(`{{define "test"}}{{end}}`); err != nil { + t.Fatalf("parsing test: %s", err) + } + if _, err := t1.Parse(`{{define "test"}}{{/* this is a comment */}}{{end}}`); err != nil { + t.Fatalf("parsing test: %s", err) + } + if _, err := t1.Parse(`{{define "test"}}foo{{end}}`); err != nil { + t.Fatalf("parsing test: %s", err) + } +} + +func TestEmptyTemplate(t *testing.T) { + cases := []struct { + defn []string + in string + want string + }{ + {[]string{"x", "y"}, "", "y"}, + {[]string{""}, "once", ""}, + {[]string{"", ""}, "twice", ""}, + {[]string{"{{.}}", "{{.}}"}, "twice", "twice"}, + {[]string{"{{/* a comment */}}", "{{/* a comment */}}"}, "comment", ""}, + {[]string{"{{.}}", ""}, "twice", ""}, + } + + for i, c := range cases { + root := New("root") + + var ( + m *Template + err error + ) + for _, d := range c.defn { + m, err = root.New(c.in).Parse(d) + if err != nil { + t.Fatal(err) + } + } + buf := &bytes.Buffer{} + if err := m.Execute(buf, c.in); err != nil { + t.Error(i, err) + continue + } + if buf.String() != c.want { + t.Errorf("expected string %q: got %q", c.want, buf.String()) + } + } +} + +// Issue 19249 was a regression in 1.8 caused by the handling of empty +// templates added in that release, which got different answers depending +// on the order templates appeared in the internal map. +func TestIssue19294(t *testing.T) { + // The empty block in "xhtml" should be replaced during execution + // by the contents of "stylesheet", but if the internal map associating + // names with templates is built in the wrong order, the empty block + // looks non-empty and this doesn't happen. + var inlined = map[string]string{ + "stylesheet": `{{define "stylesheet"}}stylesheet{{end}}`, + "xhtml": `{{block "stylesheet" .}}{{end}}`, + } + all := []string{"stylesheet", "xhtml"} + for i := 0; i < 100; i++ { + res, err := New("title.xhtml").Parse(`{{template "xhtml" .}}`) + if err != nil { + t.Fatal(err) + } + for _, name := range all { + _, err := res.New(name).Parse(inlined[name]) + if err != nil { + t.Fatal(err) + } + } + var buf bytes.Buffer + res.Execute(&buf, 0) + if buf.String() != "stylesheet" { + t.Fatalf("iteration %d: got %q; expected %q", i, buf.String(), "stylesheet") + } + } +} + +// Issue 48436 +func TestAddToZeroTemplate(t *testing.T) { + tree, err := parse.Parse("c", cloneText3, "", "", nil, builtins()) + if err != nil { + t.Fatal(err) + } + var tmpl Template + tmpl.AddParseTree("x", tree["c"]) +} diff --git a/src/text/template/option.go b/src/text/template/option.go new file mode 100644 index 0000000..addce2d --- /dev/null +++ b/src/text/template/option.go @@ -0,0 +1,74 @@ +// Copyright 2015 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +// This file contains the code to handle template options. + +package template + +import "strings" + +// missingKeyAction defines how to respond to indexing a map with a key that is not present. +type missingKeyAction int + +const ( + mapInvalid missingKeyAction = iota // Return an invalid reflect.Value. + mapZeroValue // Return the zero value for the map element. + mapError // Error out +) + +type option struct { + missingKey missingKeyAction +} + +// Option sets options for the template. Options are described by +// strings, either a simple string or "key=value". There can be at +// most one equals sign in an option string. If the option string +// is unrecognized or otherwise invalid, Option panics. +// +// Known options: +// +// missingkey: Control the behavior during execution if a map is +// indexed with a key that is not present in the map. +// "missingkey=default" or "missingkey=invalid" +// The default behavior: Do nothing and continue execution. +// If printed, the result of the index operation is the string +// "<no value>". +// "missingkey=zero" +// The operation returns the zero value for the map type's element. +// "missingkey=error" +// Execution stops immediately with an error. +// +func (t *Template) Option(opt ...string) *Template { + t.init() + for _, s := range opt { + t.setOption(s) + } + return t +} + +func (t *Template) setOption(opt string) { + if opt == "" { + panic("empty option string") + } + elems := strings.Split(opt, "=") + switch len(elems) { + case 2: + // key=value + switch elems[0] { + case "missingkey": + switch elems[1] { + case "invalid", "default": + t.option.missingKey = mapInvalid + return + case "zero": + t.option.missingKey = mapZeroValue + return + case "error": + t.option.missingKey = mapError + return + } + } + } + panic("unrecognized option: " + opt) +} diff --git a/src/text/template/parse/lex.go b/src/text/template/parse/lex.go new file mode 100644 index 0000000..6784071 --- /dev/null +++ b/src/text/template/parse/lex.go @@ -0,0 +1,671 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package parse + +import ( + "fmt" + "strings" + "unicode" + "unicode/utf8" +) + +// item represents a token or text string returned from the scanner. +type item struct { + typ itemType // The type of this item. + pos Pos // The starting position, in bytes, of this item in the input string. + val string // The value of this item. + line int // The line number at the start of this item. +} + +func (i item) String() string { + switch { + case i.typ == itemEOF: + return "EOF" + case i.typ == itemError: + return i.val + case i.typ > itemKeyword: + return fmt.Sprintf("<%s>", i.val) + case len(i.val) > 10: + return fmt.Sprintf("%.10q...", i.val) + } + return fmt.Sprintf("%q", i.val) +} + +// itemType identifies the type of lex items. +type itemType int + +const ( + itemError itemType = iota // error occurred; value is text of error + itemBool // boolean constant + itemChar // printable ASCII character; grab bag for comma etc. + itemCharConstant // character constant + itemComment // comment text + itemComplex // complex constant (1+2i); imaginary is just a number + itemAssign // equals ('=') introducing an assignment + itemDeclare // colon-equals (':=') introducing a declaration + itemEOF + itemField // alphanumeric identifier starting with '.' + itemIdentifier // alphanumeric identifier not starting with '.' + itemLeftDelim // left action delimiter + itemLeftParen // '(' inside action + itemNumber // simple number, including imaginary + itemPipe // pipe symbol + itemRawString // raw quoted string (includes quotes) + itemRightDelim // right action delimiter + itemRightParen // ')' inside action + itemSpace // run of spaces separating arguments + itemString // quoted string (includes quotes) + itemText // plain text + itemVariable // variable starting with '$', such as '$' or '$1' or '$hello' + // Keywords appear after all the rest. + itemKeyword // used only to delimit the keywords + itemBlock // block keyword + itemDot // the cursor, spelled '.' + itemDefine // define keyword + itemElse // else keyword + itemEnd // end keyword + itemIf // if keyword + itemNil // the untyped nil constant, easiest to treat as a keyword + itemRange // range keyword + itemTemplate // template keyword + itemWith // with keyword +) + +var key = map[string]itemType{ + ".": itemDot, + "block": itemBlock, + "define": itemDefine, + "else": itemElse, + "end": itemEnd, + "if": itemIf, + "range": itemRange, + "nil": itemNil, + "template": itemTemplate, + "with": itemWith, +} + +const eof = -1 + +// Trimming spaces. +// If the action begins "{{- " rather than "{{", then all space/tab/newlines +// preceding the action are trimmed; conversely if it ends " -}}" the +// leading spaces are trimmed. This is done entirely in the lexer; the +// parser never sees it happen. We require an ASCII space (' ', \t, \r, \n) +// to be present to avoid ambiguity with things like "{{-3}}". It reads +// better with the space present anyway. For simplicity, only ASCII +// does the job. +const ( + spaceChars = " \t\r\n" // These are the space characters defined by Go itself. + trimMarker = '-' // Attached to left/right delimiter, trims trailing spaces from preceding/following text. + trimMarkerLen = Pos(1 + 1) // marker plus space before or after +) + +// stateFn represents the state of the scanner as a function that returns the next state. +type stateFn func(*lexer) stateFn + +// lexer holds the state of the scanner. +type lexer struct { + name string // the name of the input; used only for error reports + input string // the string being scanned + leftDelim string // start of action + rightDelim string // end of action + emitComment bool // emit itemComment tokens. + pos Pos // current position in the input + start Pos // start position of this item + width Pos // width of last rune read from input + items chan item // channel of scanned items + parenDepth int // nesting depth of ( ) exprs + line int // 1+number of newlines seen + startLine int // start line of this item +} + +// next returns the next rune in the input. +func (l *lexer) next() rune { + if int(l.pos) >= len(l.input) { + l.width = 0 + return eof + } + r, w := utf8.DecodeRuneInString(l.input[l.pos:]) + l.width = Pos(w) + l.pos += l.width + if r == '\n' { + l.line++ + } + return r +} + +// peek returns but does not consume the next rune in the input. +func (l *lexer) peek() rune { + r := l.next() + l.backup() + return r +} + +// backup steps back one rune. Can only be called once per call of next. +func (l *lexer) backup() { + l.pos -= l.width + // Correct newline count. + if l.width == 1 && l.input[l.pos] == '\n' { + l.line-- + } +} + +// emit passes an item back to the client. +func (l *lexer) emit(t itemType) { + l.items <- item{t, l.start, l.input[l.start:l.pos], l.startLine} + l.start = l.pos + l.startLine = l.line +} + +// ignore skips over the pending input before this point. +func (l *lexer) ignore() { + l.line += strings.Count(l.input[l.start:l.pos], "\n") + l.start = l.pos + l.startLine = l.line +} + +// accept consumes the next rune if it's from the valid set. +func (l *lexer) accept(valid string) bool { + if strings.ContainsRune(valid, l.next()) { + return true + } + l.backup() + return false +} + +// acceptRun consumes a run of runes from the valid set. +func (l *lexer) acceptRun(valid string) { + for strings.ContainsRune(valid, l.next()) { + } + l.backup() +} + +// errorf returns an error token and terminates the scan by passing +// back a nil pointer that will be the next state, terminating l.nextItem. +func (l *lexer) errorf(format string, args ...interface{}) stateFn { + l.items <- item{itemError, l.start, fmt.Sprintf(format, args...), l.startLine} + return nil +} + +// nextItem returns the next item from the input. +// Called by the parser, not in the lexing goroutine. +func (l *lexer) nextItem() item { + return <-l.items +} + +// drain drains the output so the lexing goroutine will exit. +// Called by the parser, not in the lexing goroutine. +func (l *lexer) drain() { + for range l.items { + } +} + +// lex creates a new scanner for the input string. +func lex(name, input, left, right string, emitComment bool) *lexer { + if left == "" { + left = leftDelim + } + if right == "" { + right = rightDelim + } + l := &lexer{ + name: name, + input: input, + leftDelim: left, + rightDelim: right, + emitComment: emitComment, + items: make(chan item), + line: 1, + startLine: 1, + } + go l.run() + return l +} + +// run runs the state machine for the lexer. +func (l *lexer) run() { + for state := lexText; state != nil; { + state = state(l) + } + close(l.items) +} + +// state functions + +const ( + leftDelim = "{{" + rightDelim = "}}" + leftComment = "/*" + rightComment = "*/" +) + +// lexText scans until an opening action delimiter, "{{". +func lexText(l *lexer) stateFn { + l.width = 0 + if x := strings.Index(l.input[l.pos:], l.leftDelim); x >= 0 { + ldn := Pos(len(l.leftDelim)) + l.pos += Pos(x) + trimLength := Pos(0) + if hasLeftTrimMarker(l.input[l.pos+ldn:]) { + trimLength = rightTrimLength(l.input[l.start:l.pos]) + } + l.pos -= trimLength + if l.pos > l.start { + l.line += strings.Count(l.input[l.start:l.pos], "\n") + l.emit(itemText) + } + l.pos += trimLength + l.ignore() + return lexLeftDelim + } + l.pos = Pos(len(l.input)) + // Correctly reached EOF. + if l.pos > l.start { + l.line += strings.Count(l.input[l.start:l.pos], "\n") + l.emit(itemText) + } + l.emit(itemEOF) + return nil +} + +// rightTrimLength returns the length of the spaces at the end of the string. +func rightTrimLength(s string) Pos { + return Pos(len(s) - len(strings.TrimRight(s, spaceChars))) +} + +// atRightDelim reports whether the lexer is at a right delimiter, possibly preceded by a trim marker. +func (l *lexer) atRightDelim() (delim, trimSpaces bool) { + if hasRightTrimMarker(l.input[l.pos:]) && strings.HasPrefix(l.input[l.pos+trimMarkerLen:], l.rightDelim) { // With trim marker. + return true, true + } + if strings.HasPrefix(l.input[l.pos:], l.rightDelim) { // Without trim marker. + return true, false + } + return false, false +} + +// leftTrimLength returns the length of the spaces at the beginning of the string. +func leftTrimLength(s string) Pos { + return Pos(len(s) - len(strings.TrimLeft(s, spaceChars))) +} + +// lexLeftDelim scans the left delimiter, which is known to be present, possibly with a trim marker. +func lexLeftDelim(l *lexer) stateFn { + l.pos += Pos(len(l.leftDelim)) + trimSpace := hasLeftTrimMarker(l.input[l.pos:]) + afterMarker := Pos(0) + if trimSpace { + afterMarker = trimMarkerLen + } + if strings.HasPrefix(l.input[l.pos+afterMarker:], leftComment) { + l.pos += afterMarker + l.ignore() + return lexComment + } + l.emit(itemLeftDelim) + l.pos += afterMarker + l.ignore() + l.parenDepth = 0 + return lexInsideAction +} + +// lexComment scans a comment. The left comment marker is known to be present. +func lexComment(l *lexer) stateFn { + l.pos += Pos(len(leftComment)) + i := strings.Index(l.input[l.pos:], rightComment) + if i < 0 { + return l.errorf("unclosed comment") + } + l.pos += Pos(i + len(rightComment)) + delim, trimSpace := l.atRightDelim() + if !delim { + return l.errorf("comment ends before closing delimiter") + } + if l.emitComment { + l.emit(itemComment) + } + if trimSpace { + l.pos += trimMarkerLen + } + l.pos += Pos(len(l.rightDelim)) + if trimSpace { + l.pos += leftTrimLength(l.input[l.pos:]) + } + l.ignore() + return lexText +} + +// lexRightDelim scans the right delimiter, which is known to be present, possibly with a trim marker. +func lexRightDelim(l *lexer) stateFn { + trimSpace := hasRightTrimMarker(l.input[l.pos:]) + if trimSpace { + l.pos += trimMarkerLen + l.ignore() + } + l.pos += Pos(len(l.rightDelim)) + l.emit(itemRightDelim) + if trimSpace { + l.pos += leftTrimLength(l.input[l.pos:]) + l.ignore() + } + return lexText +} + +// lexInsideAction scans the elements inside action delimiters. +func lexInsideAction(l *lexer) stateFn { + // Either number, quoted string, or identifier. + // Spaces separate arguments; runs of spaces turn into itemSpace. + // Pipe symbols separate and are emitted. + delim, _ := l.atRightDelim() + if delim { + if l.parenDepth == 0 { + return lexRightDelim + } + return l.errorf("unclosed left paren") + } + switch r := l.next(); { + case r == eof: + return l.errorf("unclosed action") + case isSpace(r): + l.backup() // Put space back in case we have " -}}". + return lexSpace + case r == '=': + l.emit(itemAssign) + case r == ':': + if l.next() != '=' { + return l.errorf("expected :=") + } + l.emit(itemDeclare) + case r == '|': + l.emit(itemPipe) + case r == '"': + return lexQuote + case r == '`': + return lexRawQuote + case r == '$': + return lexVariable + case r == '\'': + return lexChar + case r == '.': + // special look-ahead for ".field" so we don't break l.backup(). + if l.pos < Pos(len(l.input)) { + r := l.input[l.pos] + if r < '0' || '9' < r { + return lexField + } + } + fallthrough // '.' can start a number. + case r == '+' || r == '-' || ('0' <= r && r <= '9'): + l.backup() + return lexNumber + case isAlphaNumeric(r): + l.backup() + return lexIdentifier + case r == '(': + l.emit(itemLeftParen) + l.parenDepth++ + case r == ')': + l.emit(itemRightParen) + l.parenDepth-- + if l.parenDepth < 0 { + return l.errorf("unexpected right paren %#U", r) + } + case r <= unicode.MaxASCII && unicode.IsPrint(r): + l.emit(itemChar) + default: + return l.errorf("unrecognized character in action: %#U", r) + } + return lexInsideAction +} + +// lexSpace scans a run of space characters. +// We have not consumed the first space, which is known to be present. +// Take care if there is a trim-marked right delimiter, which starts with a space. +func lexSpace(l *lexer) stateFn { + var r rune + var numSpaces int + for { + r = l.peek() + if !isSpace(r) { + break + } + l.next() + numSpaces++ + } + // Be careful about a trim-marked closing delimiter, which has a minus + // after a space. We know there is a space, so check for the '-' that might follow. + if hasRightTrimMarker(l.input[l.pos-1:]) && strings.HasPrefix(l.input[l.pos-1+trimMarkerLen:], l.rightDelim) { + l.backup() // Before the space. + if numSpaces == 1 { + return lexRightDelim // On the delim, so go right to that. + } + } + l.emit(itemSpace) + return lexInsideAction +} + +// lexIdentifier scans an alphanumeric. +func lexIdentifier(l *lexer) stateFn { +Loop: + for { + switch r := l.next(); { + case isAlphaNumeric(r): + // absorb. + default: + l.backup() + word := l.input[l.start:l.pos] + if !l.atTerminator() { + return l.errorf("bad character %#U", r) + } + switch { + case key[word] > itemKeyword: + l.emit(key[word]) + case word[0] == '.': + l.emit(itemField) + case word == "true", word == "false": + l.emit(itemBool) + default: + l.emit(itemIdentifier) + } + break Loop + } + } + return lexInsideAction +} + +// lexField scans a field: .Alphanumeric. +// The . has been scanned. +func lexField(l *lexer) stateFn { + return lexFieldOrVariable(l, itemField) +} + +// lexVariable scans a Variable: $Alphanumeric. +// The $ has been scanned. +func lexVariable(l *lexer) stateFn { + if l.atTerminator() { // Nothing interesting follows -> "$". + l.emit(itemVariable) + return lexInsideAction + } + return lexFieldOrVariable(l, itemVariable) +} + +// lexVariable scans a field or variable: [.$]Alphanumeric. +// The . or $ has been scanned. +func lexFieldOrVariable(l *lexer, typ itemType) stateFn { + if l.atTerminator() { // Nothing interesting follows -> "." or "$". + if typ == itemVariable { + l.emit(itemVariable) + } else { + l.emit(itemDot) + } + return lexInsideAction + } + var r rune + for { + r = l.next() + if !isAlphaNumeric(r) { + l.backup() + break + } + } + if !l.atTerminator() { + return l.errorf("bad character %#U", r) + } + l.emit(typ) + return lexInsideAction +} + +// atTerminator reports whether the input is at valid termination character to +// appear after an identifier. Breaks .X.Y into two pieces. Also catches cases +// like "$x+2" not being acceptable without a space, in case we decide one +// day to implement arithmetic. +func (l *lexer) atTerminator() bool { + r := l.peek() + if isSpace(r) { + return true + } + switch r { + case eof, '.', ',', '|', ':', ')', '(': + return true + } + // Does r start the delimiter? This can be ambiguous (with delim=="//", $x/2 will + // succeed but should fail) but only in extremely rare cases caused by willfully + // bad choice of delimiter. + if rd, _ := utf8.DecodeRuneInString(l.rightDelim); rd == r { + return true + } + return false +} + +// lexChar scans a character constant. The initial quote is already +// scanned. Syntax checking is done by the parser. +func lexChar(l *lexer) stateFn { +Loop: + for { + switch l.next() { + case '\\': + if r := l.next(); r != eof && r != '\n' { + break + } + fallthrough + case eof, '\n': + return l.errorf("unterminated character constant") + case '\'': + break Loop + } + } + l.emit(itemCharConstant) + return lexInsideAction +} + +// lexNumber scans a number: decimal, octal, hex, float, or imaginary. This +// isn't a perfect number scanner - for instance it accepts "." and "0x0.2" +// and "089" - but when it's wrong the input is invalid and the parser (via +// strconv) will notice. +func lexNumber(l *lexer) stateFn { + if !l.scanNumber() { + return l.errorf("bad number syntax: %q", l.input[l.start:l.pos]) + } + if sign := l.peek(); sign == '+' || sign == '-' { + // Complex: 1+2i. No spaces, must end in 'i'. + if !l.scanNumber() || l.input[l.pos-1] != 'i' { + return l.errorf("bad number syntax: %q", l.input[l.start:l.pos]) + } + l.emit(itemComplex) + } else { + l.emit(itemNumber) + } + return lexInsideAction +} + +func (l *lexer) scanNumber() bool { + // Optional leading sign. + l.accept("+-") + // Is it hex? + digits := "0123456789_" + if l.accept("0") { + // Note: Leading 0 does not mean octal in floats. + if l.accept("xX") { + digits = "0123456789abcdefABCDEF_" + } else if l.accept("oO") { + digits = "01234567_" + } else if l.accept("bB") { + digits = "01_" + } + } + l.acceptRun(digits) + if l.accept(".") { + l.acceptRun(digits) + } + if len(digits) == 10+1 && l.accept("eE") { + l.accept("+-") + l.acceptRun("0123456789_") + } + if len(digits) == 16+6+1 && l.accept("pP") { + l.accept("+-") + l.acceptRun("0123456789_") + } + // Is it imaginary? + l.accept("i") + // Next thing mustn't be alphanumeric. + if isAlphaNumeric(l.peek()) { + l.next() + return false + } + return true +} + +// lexQuote scans a quoted string. +func lexQuote(l *lexer) stateFn { +Loop: + for { + switch l.next() { + case '\\': + if r := l.next(); r != eof && r != '\n' { + break + } + fallthrough + case eof, '\n': + return l.errorf("unterminated quoted string") + case '"': + break Loop + } + } + l.emit(itemString) + return lexInsideAction +} + +// lexRawQuote scans a raw quoted string. +func lexRawQuote(l *lexer) stateFn { +Loop: + for { + switch l.next() { + case eof: + return l.errorf("unterminated raw quoted string") + case '`': + break Loop + } + } + l.emit(itemRawString) + return lexInsideAction +} + +// isSpace reports whether r is a space character. +func isSpace(r rune) bool { + return r == ' ' || r == '\t' || r == '\r' || r == '\n' +} + +// isAlphaNumeric reports whether r is an alphabetic, digit, or underscore. +func isAlphaNumeric(r rune) bool { + return r == '_' || unicode.IsLetter(r) || unicode.IsDigit(r) +} + +func hasLeftTrimMarker(s string) bool { + return len(s) >= 2 && s[0] == trimMarker && isSpace(rune(s[1])) +} + +func hasRightTrimMarker(s string) bool { + return len(s) >= 2 && isSpace(rune(s[0])) && s[1] == trimMarker +} diff --git a/src/text/template/parse/lex_test.go b/src/text/template/parse/lex_test.go new file mode 100644 index 0000000..6510eed --- /dev/null +++ b/src/text/template/parse/lex_test.go @@ -0,0 +1,557 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package parse + +import ( + "fmt" + "testing" +) + +// Make the types prettyprint. +var itemName = map[itemType]string{ + itemError: "error", + itemBool: "bool", + itemChar: "char", + itemCharConstant: "charconst", + itemComment: "comment", + itemComplex: "complex", + itemDeclare: ":=", + itemEOF: "EOF", + itemField: "field", + itemIdentifier: "identifier", + itemLeftDelim: "left delim", + itemLeftParen: "(", + itemNumber: "number", + itemPipe: "pipe", + itemRawString: "raw string", + itemRightDelim: "right delim", + itemRightParen: ")", + itemSpace: "space", + itemString: "string", + itemVariable: "variable", + + // keywords + itemDot: ".", + itemBlock: "block", + itemDefine: "define", + itemElse: "else", + itemIf: "if", + itemEnd: "end", + itemNil: "nil", + itemRange: "range", + itemTemplate: "template", + itemWith: "with", +} + +func (i itemType) String() string { + s := itemName[i] + if s == "" { + return fmt.Sprintf("item%d", int(i)) + } + return s +} + +type lexTest struct { + name string + input string + items []item +} + +func mkItem(typ itemType, text string) item { + return item{ + typ: typ, + val: text, + } +} + +var ( + tDot = mkItem(itemDot, ".") + tBlock = mkItem(itemBlock, "block") + tEOF = mkItem(itemEOF, "") + tFor = mkItem(itemIdentifier, "for") + tLeft = mkItem(itemLeftDelim, "{{") + tLpar = mkItem(itemLeftParen, "(") + tPipe = mkItem(itemPipe, "|") + tQuote = mkItem(itemString, `"abc \n\t\" "`) + tRange = mkItem(itemRange, "range") + tRight = mkItem(itemRightDelim, "}}") + tRpar = mkItem(itemRightParen, ")") + tSpace = mkItem(itemSpace, " ") + raw = "`" + `abc\n\t\" ` + "`" + rawNL = "`now is{{\n}}the time`" // Contains newline inside raw quote. + tRawQuote = mkItem(itemRawString, raw) + tRawQuoteNL = mkItem(itemRawString, rawNL) +) + +var lexTests = []lexTest{ + {"empty", "", []item{tEOF}}, + {"spaces", " \t\n", []item{mkItem(itemText, " \t\n"), tEOF}}, + {"text", `now is the time`, []item{mkItem(itemText, "now is the time"), tEOF}}, + {"text with comment", "hello-{{/* this is a comment */}}-world", []item{ + mkItem(itemText, "hello-"), + mkItem(itemComment, "/* this is a comment */"), + mkItem(itemText, "-world"), + tEOF, + }}, + {"punctuation", "{{,@% }}", []item{ + tLeft, + mkItem(itemChar, ","), + mkItem(itemChar, "@"), + mkItem(itemChar, "%"), + tSpace, + tRight, + tEOF, + }}, + {"parens", "{{((3))}}", []item{ + tLeft, + tLpar, + tLpar, + mkItem(itemNumber, "3"), + tRpar, + tRpar, + tRight, + tEOF, + }}, + {"empty action", `{{}}`, []item{tLeft, tRight, tEOF}}, + {"for", `{{for}}`, []item{tLeft, tFor, tRight, tEOF}}, + {"block", `{{block "foo" .}}`, []item{ + tLeft, tBlock, tSpace, mkItem(itemString, `"foo"`), tSpace, tDot, tRight, tEOF, + }}, + {"quote", `{{"abc \n\t\" "}}`, []item{tLeft, tQuote, tRight, tEOF}}, + {"raw quote", "{{" + raw + "}}", []item{tLeft, tRawQuote, tRight, tEOF}}, + {"raw quote with newline", "{{" + rawNL + "}}", []item{tLeft, tRawQuoteNL, tRight, tEOF}}, + {"numbers", "{{1 02 0x14 0X14 -7.2i 1e3 1E3 +1.2e-4 4.2i 1+2i 1_2 0x1.e_fp4 0X1.E_FP4}}", []item{ + tLeft, + mkItem(itemNumber, "1"), + tSpace, + mkItem(itemNumber, "02"), + tSpace, + mkItem(itemNumber, "0x14"), + tSpace, + mkItem(itemNumber, "0X14"), + tSpace, + mkItem(itemNumber, "-7.2i"), + tSpace, + mkItem(itemNumber, "1e3"), + tSpace, + mkItem(itemNumber, "1E3"), + tSpace, + mkItem(itemNumber, "+1.2e-4"), + tSpace, + mkItem(itemNumber, "4.2i"), + tSpace, + mkItem(itemComplex, "1+2i"), + tSpace, + mkItem(itemNumber, "1_2"), + tSpace, + mkItem(itemNumber, "0x1.e_fp4"), + tSpace, + mkItem(itemNumber, "0X1.E_FP4"), + tRight, + tEOF, + }}, + {"characters", `{{'a' '\n' '\'' '\\' '\u00FF' '\xFF' '本'}}`, []item{ + tLeft, + mkItem(itemCharConstant, `'a'`), + tSpace, + mkItem(itemCharConstant, `'\n'`), + tSpace, + mkItem(itemCharConstant, `'\''`), + tSpace, + mkItem(itemCharConstant, `'\\'`), + tSpace, + mkItem(itemCharConstant, `'\u00FF'`), + tSpace, + mkItem(itemCharConstant, `'\xFF'`), + tSpace, + mkItem(itemCharConstant, `'本'`), + tRight, + tEOF, + }}, + {"bools", "{{true false}}", []item{ + tLeft, + mkItem(itemBool, "true"), + tSpace, + mkItem(itemBool, "false"), + tRight, + tEOF, + }}, + {"dot", "{{.}}", []item{ + tLeft, + tDot, + tRight, + tEOF, + }}, + {"nil", "{{nil}}", []item{ + tLeft, + mkItem(itemNil, "nil"), + tRight, + tEOF, + }}, + {"dots", "{{.x . .2 .x.y.z}}", []item{ + tLeft, + mkItem(itemField, ".x"), + tSpace, + tDot, + tSpace, + mkItem(itemNumber, ".2"), + tSpace, + mkItem(itemField, ".x"), + mkItem(itemField, ".y"), + mkItem(itemField, ".z"), + tRight, + tEOF, + }}, + {"keywords", "{{range if else end with}}", []item{ + tLeft, + mkItem(itemRange, "range"), + tSpace, + mkItem(itemIf, "if"), + tSpace, + mkItem(itemElse, "else"), + tSpace, + mkItem(itemEnd, "end"), + tSpace, + mkItem(itemWith, "with"), + tRight, + tEOF, + }}, + {"variables", "{{$c := printf $ $hello $23 $ $var.Field .Method}}", []item{ + tLeft, + mkItem(itemVariable, "$c"), + tSpace, + mkItem(itemDeclare, ":="), + tSpace, + mkItem(itemIdentifier, "printf"), + tSpace, + mkItem(itemVariable, "$"), + tSpace, + mkItem(itemVariable, "$hello"), + tSpace, + mkItem(itemVariable, "$23"), + tSpace, + mkItem(itemVariable, "$"), + tSpace, + mkItem(itemVariable, "$var"), + mkItem(itemField, ".Field"), + tSpace, + mkItem(itemField, ".Method"), + tRight, + tEOF, + }}, + {"variable invocation", "{{$x 23}}", []item{ + tLeft, + mkItem(itemVariable, "$x"), + tSpace, + mkItem(itemNumber, "23"), + tRight, + tEOF, + }}, + {"pipeline", `intro {{echo hi 1.2 |noargs|args 1 "hi"}} outro`, []item{ + mkItem(itemText, "intro "), + tLeft, + mkItem(itemIdentifier, "echo"), + tSpace, + mkItem(itemIdentifier, "hi"), + tSpace, + mkItem(itemNumber, "1.2"), + tSpace, + tPipe, + mkItem(itemIdentifier, "noargs"), + tPipe, + mkItem(itemIdentifier, "args"), + tSpace, + mkItem(itemNumber, "1"), + tSpace, + mkItem(itemString, `"hi"`), + tRight, + mkItem(itemText, " outro"), + tEOF, + }}, + {"declaration", "{{$v := 3}}", []item{ + tLeft, + mkItem(itemVariable, "$v"), + tSpace, + mkItem(itemDeclare, ":="), + tSpace, + mkItem(itemNumber, "3"), + tRight, + tEOF, + }}, + {"2 declarations", "{{$v , $w := 3}}", []item{ + tLeft, + mkItem(itemVariable, "$v"), + tSpace, + mkItem(itemChar, ","), + tSpace, + mkItem(itemVariable, "$w"), + tSpace, + mkItem(itemDeclare, ":="), + tSpace, + mkItem(itemNumber, "3"), + tRight, + tEOF, + }}, + {"field of parenthesized expression", "{{(.X).Y}}", []item{ + tLeft, + tLpar, + mkItem(itemField, ".X"), + tRpar, + mkItem(itemField, ".Y"), + tRight, + tEOF, + }}, + {"trimming spaces before and after", "hello- {{- 3 -}} -world", []item{ + mkItem(itemText, "hello-"), + tLeft, + mkItem(itemNumber, "3"), + tRight, + mkItem(itemText, "-world"), + tEOF, + }}, + {"trimming spaces before and after comment", "hello- {{- /* hello */ -}} -world", []item{ + mkItem(itemText, "hello-"), + mkItem(itemComment, "/* hello */"), + mkItem(itemText, "-world"), + tEOF, + }}, + // errors + {"badchar", "#{{\x01}}", []item{ + mkItem(itemText, "#"), + tLeft, + mkItem(itemError, "unrecognized character in action: U+0001"), + }}, + {"unclosed action", "{{", []item{ + tLeft, + mkItem(itemError, "unclosed action"), + }}, + {"EOF in action", "{{range", []item{ + tLeft, + tRange, + mkItem(itemError, "unclosed action"), + }}, + {"unclosed quote", "{{\"\n\"}}", []item{ + tLeft, + mkItem(itemError, "unterminated quoted string"), + }}, + {"unclosed raw quote", "{{`xx}}", []item{ + tLeft, + mkItem(itemError, "unterminated raw quoted string"), + }}, + {"unclosed char constant", "{{'\n}}", []item{ + tLeft, + mkItem(itemError, "unterminated character constant"), + }}, + {"bad number", "{{3k}}", []item{ + tLeft, + mkItem(itemError, `bad number syntax: "3k"`), + }}, + {"unclosed paren", "{{(3}}", []item{ + tLeft, + tLpar, + mkItem(itemNumber, "3"), + mkItem(itemError, `unclosed left paren`), + }}, + {"extra right paren", "{{3)}}", []item{ + tLeft, + mkItem(itemNumber, "3"), + tRpar, + mkItem(itemError, `unexpected right paren U+0029 ')'`), + }}, + + // Fixed bugs + // Many elements in an action blew the lookahead until + // we made lexInsideAction not loop. + {"long pipeline deadlock", "{{|||||}}", []item{ + tLeft, + tPipe, + tPipe, + tPipe, + tPipe, + tPipe, + tRight, + tEOF, + }}, + {"text with bad comment", "hello-{{/*/}}-world", []item{ + mkItem(itemText, "hello-"), + mkItem(itemError, `unclosed comment`), + }}, + {"text with comment close separated from delim", "hello-{{/* */ }}-world", []item{ + mkItem(itemText, "hello-"), + mkItem(itemError, `comment ends before closing delimiter`), + }}, + // This one is an error that we can't catch because it breaks templates with + // minimized JavaScript. Should have fixed it before Go 1.1. + {"unmatched right delimiter", "hello-{.}}-world", []item{ + mkItem(itemText, "hello-{.}}-world"), + tEOF, + }}, +} + +// collect gathers the emitted items into a slice. +func collect(t *lexTest, left, right string) (items []item) { + l := lex(t.name, t.input, left, right, true) + for { + item := l.nextItem() + items = append(items, item) + if item.typ == itemEOF || item.typ == itemError { + break + } + } + return +} + +func equal(i1, i2 []item, checkPos bool) bool { + if len(i1) != len(i2) { + return false + } + for k := range i1 { + if i1[k].typ != i2[k].typ { + return false + } + if i1[k].val != i2[k].val { + return false + } + if checkPos && i1[k].pos != i2[k].pos { + return false + } + if checkPos && i1[k].line != i2[k].line { + return false + } + } + return true +} + +func TestLex(t *testing.T) { + for _, test := range lexTests { + items := collect(&test, "", "") + if !equal(items, test.items, false) { + t.Errorf("%s: got\n\t%+v\nexpected\n\t%v", test.name, items, test.items) + } + } +} + +// Some easy cases from above, but with delimiters $$ and @@ +var lexDelimTests = []lexTest{ + {"punctuation", "$$,@%{{}}@@", []item{ + tLeftDelim, + mkItem(itemChar, ","), + mkItem(itemChar, "@"), + mkItem(itemChar, "%"), + mkItem(itemChar, "{"), + mkItem(itemChar, "{"), + mkItem(itemChar, "}"), + mkItem(itemChar, "}"), + tRightDelim, + tEOF, + }}, + {"empty action", `$$@@`, []item{tLeftDelim, tRightDelim, tEOF}}, + {"for", `$$for@@`, []item{tLeftDelim, tFor, tRightDelim, tEOF}}, + {"quote", `$$"abc \n\t\" "@@`, []item{tLeftDelim, tQuote, tRightDelim, tEOF}}, + {"raw quote", "$$" + raw + "@@", []item{tLeftDelim, tRawQuote, tRightDelim, tEOF}}, +} + +var ( + tLeftDelim = mkItem(itemLeftDelim, "$$") + tRightDelim = mkItem(itemRightDelim, "@@") +) + +func TestDelims(t *testing.T) { + for _, test := range lexDelimTests { + items := collect(&test, "$$", "@@") + if !equal(items, test.items, false) { + t.Errorf("%s: got\n\t%v\nexpected\n\t%v", test.name, items, test.items) + } + } +} + +var lexPosTests = []lexTest{ + {"empty", "", []item{{itemEOF, 0, "", 1}}}, + {"punctuation", "{{,@%#}}", []item{ + {itemLeftDelim, 0, "{{", 1}, + {itemChar, 2, ",", 1}, + {itemChar, 3, "@", 1}, + {itemChar, 4, "%", 1}, + {itemChar, 5, "#", 1}, + {itemRightDelim, 6, "}}", 1}, + {itemEOF, 8, "", 1}, + }}, + {"sample", "0123{{hello}}xyz", []item{ + {itemText, 0, "0123", 1}, + {itemLeftDelim, 4, "{{", 1}, + {itemIdentifier, 6, "hello", 1}, + {itemRightDelim, 11, "}}", 1}, + {itemText, 13, "xyz", 1}, + {itemEOF, 16, "", 1}, + }}, + {"trimafter", "{{x -}}\n{{y}}", []item{ + {itemLeftDelim, 0, "{{", 1}, + {itemIdentifier, 2, "x", 1}, + {itemRightDelim, 5, "}}", 1}, + {itemLeftDelim, 8, "{{", 2}, + {itemIdentifier, 10, "y", 2}, + {itemRightDelim, 11, "}}", 2}, + {itemEOF, 13, "", 2}, + }}, + {"trimbefore", "{{x}}\n{{- y}}", []item{ + {itemLeftDelim, 0, "{{", 1}, + {itemIdentifier, 2, "x", 1}, + {itemRightDelim, 3, "}}", 1}, + {itemLeftDelim, 6, "{{", 2}, + {itemIdentifier, 10, "y", 2}, + {itemRightDelim, 11, "}}", 2}, + {itemEOF, 13, "", 2}, + }}, +} + +// The other tests don't check position, to make the test cases easier to construct. +// This one does. +func TestPos(t *testing.T) { + for _, test := range lexPosTests { + items := collect(&test, "", "") + if !equal(items, test.items, true) { + t.Errorf("%s: got\n\t%v\nexpected\n\t%v", test.name, items, test.items) + if len(items) == len(test.items) { + // Detailed print; avoid item.String() to expose the position value. + for i := range items { + if !equal(items[i:i+1], test.items[i:i+1], true) { + i1 := items[i] + i2 := test.items[i] + t.Errorf("\t#%d: got {%v %d %q %d} expected {%v %d %q %d}", + i, i1.typ, i1.pos, i1.val, i1.line, i2.typ, i2.pos, i2.val, i2.line) + } + } + } + } + } +} + +// Test that an error shuts down the lexing goroutine. +func TestShutdown(t *testing.T) { + // We need to duplicate template.Parse here to hold on to the lexer. + const text = "erroneous{{define}}{{else}}1234" + lexer := lex("foo", text, "{{", "}}", false) + _, err := New("root").parseLexer(lexer) + if err == nil { + t.Fatalf("expected error") + } + // The error should have drained the input. Therefore, the lexer should be shut down. + token, ok := <-lexer.items + if ok { + t.Fatalf("input was not drained; got %v", token) + } +} + +// parseLexer is a local version of parse that lets us pass in the lexer instead of building it. +// We expect an error, so the tree set and funcs list are explicitly nil. +func (t *Tree) parseLexer(lex *lexer) (tree *Tree, err error) { + defer t.recover(&err) + t.ParseName = t.Name + t.startParse(nil, lex, map[string]*Tree{}) + t.parse() + t.add() + t.stopParse() + return t, nil +} diff --git a/src/text/template/parse/node.go b/src/text/template/parse/node.go new file mode 100644 index 0000000..177482f --- /dev/null +++ b/src/text/template/parse/node.go @@ -0,0 +1,972 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +// Parse nodes. + +package parse + +import ( + "fmt" + "strconv" + "strings" +) + +var textFormat = "%s" // Changed to "%q" in tests for better error messages. + +// A Node is an element in the parse tree. The interface is trivial. +// The interface contains an unexported method so that only +// types local to this package can satisfy it. +type Node interface { + Type() NodeType + String() string + // Copy does a deep copy of the Node and all its components. + // To avoid type assertions, some XxxNodes also have specialized + // CopyXxx methods that return *XxxNode. + Copy() Node + Position() Pos // byte position of start of node in full original input string + // tree returns the containing *Tree. + // It is unexported so all implementations of Node are in this package. + tree() *Tree + // writeTo writes the String output to the builder. + writeTo(*strings.Builder) +} + +// NodeType identifies the type of a parse tree node. +type NodeType int + +// Pos represents a byte position in the original input text from which +// this template was parsed. +type Pos int + +func (p Pos) Position() Pos { + return p +} + +// Type returns itself and provides an easy default implementation +// for embedding in a Node. Embedded in all non-trivial Nodes. +func (t NodeType) Type() NodeType { + return t +} + +const ( + NodeText NodeType = iota // Plain text. + NodeAction // A non-control action such as a field evaluation. + NodeBool // A boolean constant. + NodeChain // A sequence of field accesses. + NodeCommand // An element of a pipeline. + NodeDot // The cursor, dot. + nodeElse // An else action. Not added to tree. + nodeEnd // An end action. Not added to tree. + NodeField // A field or method name. + NodeIdentifier // An identifier; always a function name. + NodeIf // An if action. + NodeList // A list of Nodes. + NodeNil // An untyped nil constant. + NodeNumber // A numerical constant. + NodePipe // A pipeline of commands. + NodeRange // A range action. + NodeString // A string constant. + NodeTemplate // A template invocation action. + NodeVariable // A $ variable. + NodeWith // A with action. + NodeComment // A comment. +) + +// Nodes. + +// ListNode holds a sequence of nodes. +type ListNode struct { + NodeType + Pos + tr *Tree + Nodes []Node // The element nodes in lexical order. +} + +func (t *Tree) newList(pos Pos) *ListNode { + return &ListNode{tr: t, NodeType: NodeList, Pos: pos} +} + +func (l *ListNode) append(n Node) { + l.Nodes = append(l.Nodes, n) +} + +func (l *ListNode) tree() *Tree { + return l.tr +} + +func (l *ListNode) String() string { + var sb strings.Builder + l.writeTo(&sb) + return sb.String() +} + +func (l *ListNode) writeTo(sb *strings.Builder) { + for _, n := range l.Nodes { + n.writeTo(sb) + } +} + +func (l *ListNode) CopyList() *ListNode { + if l == nil { + return l + } + n := l.tr.newList(l.Pos) + for _, elem := range l.Nodes { + n.append(elem.Copy()) + } + return n +} + +func (l *ListNode) Copy() Node { + return l.CopyList() +} + +// TextNode holds plain text. +type TextNode struct { + NodeType + Pos + tr *Tree + Text []byte // The text; may span newlines. +} + +func (t *Tree) newText(pos Pos, text string) *TextNode { + return &TextNode{tr: t, NodeType: NodeText, Pos: pos, Text: []byte(text)} +} + +func (t *TextNode) String() string { + return fmt.Sprintf(textFormat, t.Text) +} + +func (t *TextNode) writeTo(sb *strings.Builder) { + sb.WriteString(t.String()) +} + +func (t *TextNode) tree() *Tree { + return t.tr +} + +func (t *TextNode) Copy() Node { + return &TextNode{tr: t.tr, NodeType: NodeText, Pos: t.Pos, Text: append([]byte{}, t.Text...)} +} + +// CommentNode holds a comment. +type CommentNode struct { + NodeType + Pos + tr *Tree + Text string // Comment text. +} + +func (t *Tree) newComment(pos Pos, text string) *CommentNode { + return &CommentNode{tr: t, NodeType: NodeComment, Pos: pos, Text: text} +} + +func (c *CommentNode) String() string { + var sb strings.Builder + c.writeTo(&sb) + return sb.String() +} + +func (c *CommentNode) writeTo(sb *strings.Builder) { + sb.WriteString("{{") + sb.WriteString(c.Text) + sb.WriteString("}}") +} + +func (c *CommentNode) tree() *Tree { + return c.tr +} + +func (c *CommentNode) Copy() Node { + return &CommentNode{tr: c.tr, NodeType: NodeComment, Pos: c.Pos, Text: c.Text} +} + +// PipeNode holds a pipeline with optional declaration +type PipeNode struct { + NodeType + Pos + tr *Tree + Line int // The line number in the input. Deprecated: Kept for compatibility. + IsAssign bool // The variables are being assigned, not declared. + Decl []*VariableNode // Variables in lexical order. + Cmds []*CommandNode // The commands in lexical order. +} + +func (t *Tree) newPipeline(pos Pos, line int, vars []*VariableNode) *PipeNode { + return &PipeNode{tr: t, NodeType: NodePipe, Pos: pos, Line: line, Decl: vars} +} + +func (p *PipeNode) append(command *CommandNode) { + p.Cmds = append(p.Cmds, command) +} + +func (p *PipeNode) String() string { + var sb strings.Builder + p.writeTo(&sb) + return sb.String() +} + +func (p *PipeNode) writeTo(sb *strings.Builder) { + if len(p.Decl) > 0 { + for i, v := range p.Decl { + if i > 0 { + sb.WriteString(", ") + } + v.writeTo(sb) + } + sb.WriteString(" := ") + } + for i, c := range p.Cmds { + if i > 0 { + sb.WriteString(" | ") + } + c.writeTo(sb) + } +} + +func (p *PipeNode) tree() *Tree { + return p.tr +} + +func (p *PipeNode) CopyPipe() *PipeNode { + if p == nil { + return p + } + vars := make([]*VariableNode, len(p.Decl)) + for i, d := range p.Decl { + vars[i] = d.Copy().(*VariableNode) + } + n := p.tr.newPipeline(p.Pos, p.Line, vars) + n.IsAssign = p.IsAssign + for _, c := range p.Cmds { + n.append(c.Copy().(*CommandNode)) + } + return n +} + +func (p *PipeNode) Copy() Node { + return p.CopyPipe() +} + +// ActionNode holds an action (something bounded by delimiters). +// Control actions have their own nodes; ActionNode represents simple +// ones such as field evaluations and parenthesized pipelines. +type ActionNode struct { + NodeType + Pos + tr *Tree + Line int // The line number in the input. Deprecated: Kept for compatibility. + Pipe *PipeNode // The pipeline in the action. +} + +func (t *Tree) newAction(pos Pos, line int, pipe *PipeNode) *ActionNode { + return &ActionNode{tr: t, NodeType: NodeAction, Pos: pos, Line: line, Pipe: pipe} +} + +func (a *ActionNode) String() string { + var sb strings.Builder + a.writeTo(&sb) + return sb.String() +} + +func (a *ActionNode) writeTo(sb *strings.Builder) { + sb.WriteString("{{") + a.Pipe.writeTo(sb) + sb.WriteString("}}") +} + +func (a *ActionNode) tree() *Tree { + return a.tr +} + +func (a *ActionNode) Copy() Node { + return a.tr.newAction(a.Pos, a.Line, a.Pipe.CopyPipe()) + +} + +// CommandNode holds a command (a pipeline inside an evaluating action). +type CommandNode struct { + NodeType + Pos + tr *Tree + Args []Node // Arguments in lexical order: Identifier, field, or constant. +} + +func (t *Tree) newCommand(pos Pos) *CommandNode { + return &CommandNode{tr: t, NodeType: NodeCommand, Pos: pos} +} + +func (c *CommandNode) append(arg Node) { + c.Args = append(c.Args, arg) +} + +func (c *CommandNode) String() string { + var sb strings.Builder + c.writeTo(&sb) + return sb.String() +} + +func (c *CommandNode) writeTo(sb *strings.Builder) { + for i, arg := range c.Args { + if i > 0 { + sb.WriteByte(' ') + } + if arg, ok := arg.(*PipeNode); ok { + sb.WriteByte('(') + arg.writeTo(sb) + sb.WriteByte(')') + continue + } + arg.writeTo(sb) + } +} + +func (c *CommandNode) tree() *Tree { + return c.tr +} + +func (c *CommandNode) Copy() Node { + if c == nil { + return c + } + n := c.tr.newCommand(c.Pos) + for _, c := range c.Args { + n.append(c.Copy()) + } + return n +} + +// IdentifierNode holds an identifier. +type IdentifierNode struct { + NodeType + Pos + tr *Tree + Ident string // The identifier's name. +} + +// NewIdentifier returns a new IdentifierNode with the given identifier name. +func NewIdentifier(ident string) *IdentifierNode { + return &IdentifierNode{NodeType: NodeIdentifier, Ident: ident} +} + +// SetPos sets the position. NewIdentifier is a public method so we can't modify its signature. +// Chained for convenience. +// TODO: fix one day? +func (i *IdentifierNode) SetPos(pos Pos) *IdentifierNode { + i.Pos = pos + return i +} + +// SetTree sets the parent tree for the node. NewIdentifier is a public method so we can't modify its signature. +// Chained for convenience. +// TODO: fix one day? +func (i *IdentifierNode) SetTree(t *Tree) *IdentifierNode { + i.tr = t + return i +} + +func (i *IdentifierNode) String() string { + return i.Ident +} + +func (i *IdentifierNode) writeTo(sb *strings.Builder) { + sb.WriteString(i.String()) +} + +func (i *IdentifierNode) tree() *Tree { + return i.tr +} + +func (i *IdentifierNode) Copy() Node { + return NewIdentifier(i.Ident).SetTree(i.tr).SetPos(i.Pos) +} + +// VariableNode holds a list of variable names, possibly with chained field +// accesses. The dollar sign is part of the (first) name. +type VariableNode struct { + NodeType + Pos + tr *Tree + Ident []string // Variable name and fields in lexical order. +} + +func (t *Tree) newVariable(pos Pos, ident string) *VariableNode { + return &VariableNode{tr: t, NodeType: NodeVariable, Pos: pos, Ident: strings.Split(ident, ".")} +} + +func (v *VariableNode) String() string { + var sb strings.Builder + v.writeTo(&sb) + return sb.String() +} + +func (v *VariableNode) writeTo(sb *strings.Builder) { + for i, id := range v.Ident { + if i > 0 { + sb.WriteByte('.') + } + sb.WriteString(id) + } +} + +func (v *VariableNode) tree() *Tree { + return v.tr +} + +func (v *VariableNode) Copy() Node { + return &VariableNode{tr: v.tr, NodeType: NodeVariable, Pos: v.Pos, Ident: append([]string{}, v.Ident...)} +} + +// DotNode holds the special identifier '.'. +type DotNode struct { + NodeType + Pos + tr *Tree +} + +func (t *Tree) newDot(pos Pos) *DotNode { + return &DotNode{tr: t, NodeType: NodeDot, Pos: pos} +} + +func (d *DotNode) Type() NodeType { + // Override method on embedded NodeType for API compatibility. + // TODO: Not really a problem; could change API without effect but + // api tool complains. + return NodeDot +} + +func (d *DotNode) String() string { + return "." +} + +func (d *DotNode) writeTo(sb *strings.Builder) { + sb.WriteString(d.String()) +} + +func (d *DotNode) tree() *Tree { + return d.tr +} + +func (d *DotNode) Copy() Node { + return d.tr.newDot(d.Pos) +} + +// NilNode holds the special identifier 'nil' representing an untyped nil constant. +type NilNode struct { + NodeType + Pos + tr *Tree +} + +func (t *Tree) newNil(pos Pos) *NilNode { + return &NilNode{tr: t, NodeType: NodeNil, Pos: pos} +} + +func (n *NilNode) Type() NodeType { + // Override method on embedded NodeType for API compatibility. + // TODO: Not really a problem; could change API without effect but + // api tool complains. + return NodeNil +} + +func (n *NilNode) String() string { + return "nil" +} + +func (n *NilNode) writeTo(sb *strings.Builder) { + sb.WriteString(n.String()) +} + +func (n *NilNode) tree() *Tree { + return n.tr +} + +func (n *NilNode) Copy() Node { + return n.tr.newNil(n.Pos) +} + +// FieldNode holds a field (identifier starting with '.'). +// The names may be chained ('.x.y'). +// The period is dropped from each ident. +type FieldNode struct { + NodeType + Pos + tr *Tree + Ident []string // The identifiers in lexical order. +} + +func (t *Tree) newField(pos Pos, ident string) *FieldNode { + return &FieldNode{tr: t, NodeType: NodeField, Pos: pos, Ident: strings.Split(ident[1:], ".")} // [1:] to drop leading period +} + +func (f *FieldNode) String() string { + var sb strings.Builder + f.writeTo(&sb) + return sb.String() +} + +func (f *FieldNode) writeTo(sb *strings.Builder) { + for _, id := range f.Ident { + sb.WriteByte('.') + sb.WriteString(id) + } +} + +func (f *FieldNode) tree() *Tree { + return f.tr +} + +func (f *FieldNode) Copy() Node { + return &FieldNode{tr: f.tr, NodeType: NodeField, Pos: f.Pos, Ident: append([]string{}, f.Ident...)} +} + +// ChainNode holds a term followed by a chain of field accesses (identifier starting with '.'). +// The names may be chained ('.x.y'). +// The periods are dropped from each ident. +type ChainNode struct { + NodeType + Pos + tr *Tree + Node Node + Field []string // The identifiers in lexical order. +} + +func (t *Tree) newChain(pos Pos, node Node) *ChainNode { + return &ChainNode{tr: t, NodeType: NodeChain, Pos: pos, Node: node} +} + +// Add adds the named field (which should start with a period) to the end of the chain. +func (c *ChainNode) Add(field string) { + if len(field) == 0 || field[0] != '.' { + panic("no dot in field") + } + field = field[1:] // Remove leading dot. + if field == "" { + panic("empty field") + } + c.Field = append(c.Field, field) +} + +func (c *ChainNode) String() string { + var sb strings.Builder + c.writeTo(&sb) + return sb.String() +} + +func (c *ChainNode) writeTo(sb *strings.Builder) { + if _, ok := c.Node.(*PipeNode); ok { + sb.WriteByte('(') + c.Node.writeTo(sb) + sb.WriteByte(')') + } else { + c.Node.writeTo(sb) + } + for _, field := range c.Field { + sb.WriteByte('.') + sb.WriteString(field) + } +} + +func (c *ChainNode) tree() *Tree { + return c.tr +} + +func (c *ChainNode) Copy() Node { + return &ChainNode{tr: c.tr, NodeType: NodeChain, Pos: c.Pos, Node: c.Node, Field: append([]string{}, c.Field...)} +} + +// BoolNode holds a boolean constant. +type BoolNode struct { + NodeType + Pos + tr *Tree + True bool // The value of the boolean constant. +} + +func (t *Tree) newBool(pos Pos, true bool) *BoolNode { + return &BoolNode{tr: t, NodeType: NodeBool, Pos: pos, True: true} +} + +func (b *BoolNode) String() string { + if b.True { + return "true" + } + return "false" +} + +func (b *BoolNode) writeTo(sb *strings.Builder) { + sb.WriteString(b.String()) +} + +func (b *BoolNode) tree() *Tree { + return b.tr +} + +func (b *BoolNode) Copy() Node { + return b.tr.newBool(b.Pos, b.True) +} + +// NumberNode holds a number: signed or unsigned integer, float, or complex. +// The value is parsed and stored under all the types that can represent the value. +// This simulates in a small amount of code the behavior of Go's ideal constants. +type NumberNode struct { + NodeType + Pos + tr *Tree + IsInt bool // Number has an integral value. + IsUint bool // Number has an unsigned integral value. + IsFloat bool // Number has a floating-point value. + IsComplex bool // Number is complex. + Int64 int64 // The signed integer value. + Uint64 uint64 // The unsigned integer value. + Float64 float64 // The floating-point value. + Complex128 complex128 // The complex value. + Text string // The original textual representation from the input. +} + +func (t *Tree) newNumber(pos Pos, text string, typ itemType) (*NumberNode, error) { + n := &NumberNode{tr: t, NodeType: NodeNumber, Pos: pos, Text: text} + switch typ { + case itemCharConstant: + rune, _, tail, err := strconv.UnquoteChar(text[1:], text[0]) + if err != nil { + return nil, err + } + if tail != "'" { + return nil, fmt.Errorf("malformed character constant: %s", text) + } + n.Int64 = int64(rune) + n.IsInt = true + n.Uint64 = uint64(rune) + n.IsUint = true + n.Float64 = float64(rune) // odd but those are the rules. + n.IsFloat = true + return n, nil + case itemComplex: + // fmt.Sscan can parse the pair, so let it do the work. + if _, err := fmt.Sscan(text, &n.Complex128); err != nil { + return nil, err + } + n.IsComplex = true + n.simplifyComplex() + return n, nil + } + // Imaginary constants can only be complex unless they are zero. + if len(text) > 0 && text[len(text)-1] == 'i' { + f, err := strconv.ParseFloat(text[:len(text)-1], 64) + if err == nil { + n.IsComplex = true + n.Complex128 = complex(0, f) + n.simplifyComplex() + return n, nil + } + } + // Do integer test first so we get 0x123 etc. + u, err := strconv.ParseUint(text, 0, 64) // will fail for -0; fixed below. + if err == nil { + n.IsUint = true + n.Uint64 = u + } + i, err := strconv.ParseInt(text, 0, 64) + if err == nil { + n.IsInt = true + n.Int64 = i + if i == 0 { + n.IsUint = true // in case of -0. + n.Uint64 = u + } + } + // If an integer extraction succeeded, promote the float. + if n.IsInt { + n.IsFloat = true + n.Float64 = float64(n.Int64) + } else if n.IsUint { + n.IsFloat = true + n.Float64 = float64(n.Uint64) + } else { + f, err := strconv.ParseFloat(text, 64) + if err == nil { + // If we parsed it as a float but it looks like an integer, + // it's a huge number too large to fit in an int. Reject it. + if !strings.ContainsAny(text, ".eEpP") { + return nil, fmt.Errorf("integer overflow: %q", text) + } + n.IsFloat = true + n.Float64 = f + // If a floating-point extraction succeeded, extract the int if needed. + if !n.IsInt && float64(int64(f)) == f { + n.IsInt = true + n.Int64 = int64(f) + } + if !n.IsUint && float64(uint64(f)) == f { + n.IsUint = true + n.Uint64 = uint64(f) + } + } + } + if !n.IsInt && !n.IsUint && !n.IsFloat { + return nil, fmt.Errorf("illegal number syntax: %q", text) + } + return n, nil +} + +// simplifyComplex pulls out any other types that are represented by the complex number. +// These all require that the imaginary part be zero. +func (n *NumberNode) simplifyComplex() { + n.IsFloat = imag(n.Complex128) == 0 + if n.IsFloat { + n.Float64 = real(n.Complex128) + n.IsInt = float64(int64(n.Float64)) == n.Float64 + if n.IsInt { + n.Int64 = int64(n.Float64) + } + n.IsUint = float64(uint64(n.Float64)) == n.Float64 + if n.IsUint { + n.Uint64 = uint64(n.Float64) + } + } +} + +func (n *NumberNode) String() string { + return n.Text +} + +func (n *NumberNode) writeTo(sb *strings.Builder) { + sb.WriteString(n.String()) +} + +func (n *NumberNode) tree() *Tree { + return n.tr +} + +func (n *NumberNode) Copy() Node { + nn := new(NumberNode) + *nn = *n // Easy, fast, correct. + return nn +} + +// StringNode holds a string constant. The value has been "unquoted". +type StringNode struct { + NodeType + Pos + tr *Tree + Quoted string // The original text of the string, with quotes. + Text string // The string, after quote processing. +} + +func (t *Tree) newString(pos Pos, orig, text string) *StringNode { + return &StringNode{tr: t, NodeType: NodeString, Pos: pos, Quoted: orig, Text: text} +} + +func (s *StringNode) String() string { + return s.Quoted +} + +func (s *StringNode) writeTo(sb *strings.Builder) { + sb.WriteString(s.String()) +} + +func (s *StringNode) tree() *Tree { + return s.tr +} + +func (s *StringNode) Copy() Node { + return s.tr.newString(s.Pos, s.Quoted, s.Text) +} + +// endNode represents an {{end}} action. +// It does not appear in the final parse tree. +type endNode struct { + NodeType + Pos + tr *Tree +} + +func (t *Tree) newEnd(pos Pos) *endNode { + return &endNode{tr: t, NodeType: nodeEnd, Pos: pos} +} + +func (e *endNode) String() string { + return "{{end}}" +} + +func (e *endNode) writeTo(sb *strings.Builder) { + sb.WriteString(e.String()) +} + +func (e *endNode) tree() *Tree { + return e.tr +} + +func (e *endNode) Copy() Node { + return e.tr.newEnd(e.Pos) +} + +// elseNode represents an {{else}} action. Does not appear in the final tree. +type elseNode struct { + NodeType + Pos + tr *Tree + Line int // The line number in the input. Deprecated: Kept for compatibility. +} + +func (t *Tree) newElse(pos Pos, line int) *elseNode { + return &elseNode{tr: t, NodeType: nodeElse, Pos: pos, Line: line} +} + +func (e *elseNode) Type() NodeType { + return nodeElse +} + +func (e *elseNode) String() string { + return "{{else}}" +} + +func (e *elseNode) writeTo(sb *strings.Builder) { + sb.WriteString(e.String()) +} + +func (e *elseNode) tree() *Tree { + return e.tr +} + +func (e *elseNode) Copy() Node { + return e.tr.newElse(e.Pos, e.Line) +} + +// BranchNode is the common representation of if, range, and with. +type BranchNode struct { + NodeType + Pos + tr *Tree + Line int // The line number in the input. Deprecated: Kept for compatibility. + Pipe *PipeNode // The pipeline to be evaluated. + List *ListNode // What to execute if the value is non-empty. + ElseList *ListNode // What to execute if the value is empty (nil if absent). +} + +func (b *BranchNode) String() string { + var sb strings.Builder + b.writeTo(&sb) + return sb.String() +} + +func (b *BranchNode) writeTo(sb *strings.Builder) { + name := "" + switch b.NodeType { + case NodeIf: + name = "if" + case NodeRange: + name = "range" + case NodeWith: + name = "with" + default: + panic("unknown branch type") + } + sb.WriteString("{{") + sb.WriteString(name) + sb.WriteByte(' ') + b.Pipe.writeTo(sb) + sb.WriteString("}}") + b.List.writeTo(sb) + if b.ElseList != nil { + sb.WriteString("{{else}}") + b.ElseList.writeTo(sb) + } + sb.WriteString("{{end}}") +} + +func (b *BranchNode) tree() *Tree { + return b.tr +} + +func (b *BranchNode) Copy() Node { + switch b.NodeType { + case NodeIf: + return b.tr.newIf(b.Pos, b.Line, b.Pipe, b.List, b.ElseList) + case NodeRange: + return b.tr.newRange(b.Pos, b.Line, b.Pipe, b.List, b.ElseList) + case NodeWith: + return b.tr.newWith(b.Pos, b.Line, b.Pipe, b.List, b.ElseList) + default: + panic("unknown branch type") + } +} + +// IfNode represents an {{if}} action and its commands. +type IfNode struct { + BranchNode +} + +func (t *Tree) newIf(pos Pos, line int, pipe *PipeNode, list, elseList *ListNode) *IfNode { + return &IfNode{BranchNode{tr: t, NodeType: NodeIf, Pos: pos, Line: line, Pipe: pipe, List: list, ElseList: elseList}} +} + +func (i *IfNode) Copy() Node { + return i.tr.newIf(i.Pos, i.Line, i.Pipe.CopyPipe(), i.List.CopyList(), i.ElseList.CopyList()) +} + +// RangeNode represents a {{range}} action and its commands. +type RangeNode struct { + BranchNode +} + +func (t *Tree) newRange(pos Pos, line int, pipe *PipeNode, list, elseList *ListNode) *RangeNode { + return &RangeNode{BranchNode{tr: t, NodeType: NodeRange, Pos: pos, Line: line, Pipe: pipe, List: list, ElseList: elseList}} +} + +func (r *RangeNode) Copy() Node { + return r.tr.newRange(r.Pos, r.Line, r.Pipe.CopyPipe(), r.List.CopyList(), r.ElseList.CopyList()) +} + +// WithNode represents a {{with}} action and its commands. +type WithNode struct { + BranchNode +} + +func (t *Tree) newWith(pos Pos, line int, pipe *PipeNode, list, elseList *ListNode) *WithNode { + return &WithNode{BranchNode{tr: t, NodeType: NodeWith, Pos: pos, Line: line, Pipe: pipe, List: list, ElseList: elseList}} +} + +func (w *WithNode) Copy() Node { + return w.tr.newWith(w.Pos, w.Line, w.Pipe.CopyPipe(), w.List.CopyList(), w.ElseList.CopyList()) +} + +// TemplateNode represents a {{template}} action. +type TemplateNode struct { + NodeType + Pos + tr *Tree + Line int // The line number in the input. Deprecated: Kept for compatibility. + Name string // The name of the template (unquoted). + Pipe *PipeNode // The command to evaluate as dot for the template. +} + +func (t *Tree) newTemplate(pos Pos, line int, name string, pipe *PipeNode) *TemplateNode { + return &TemplateNode{tr: t, NodeType: NodeTemplate, Pos: pos, Line: line, Name: name, Pipe: pipe} +} + +func (t *TemplateNode) String() string { + var sb strings.Builder + t.writeTo(&sb) + return sb.String() +} + +func (t *TemplateNode) writeTo(sb *strings.Builder) { + sb.WriteString("{{template ") + sb.WriteString(strconv.Quote(t.Name)) + if t.Pipe != nil { + sb.WriteByte(' ') + t.Pipe.writeTo(sb) + } + sb.WriteString("}}") +} + +func (t *TemplateNode) tree() *Tree { + return t.tr +} + +func (t *TemplateNode) Copy() Node { + return t.tr.newTemplate(t.Pos, t.Line, t.Name, t.Pipe.CopyPipe()) +} diff --git a/src/text/template/parse/parse.go b/src/text/template/parse/parse.go new file mode 100644 index 0000000..1a63961 --- /dev/null +++ b/src/text/template/parse/parse.go @@ -0,0 +1,756 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +// Package parse builds parse trees for templates as defined by text/template +// and html/template. Clients should use those packages to construct templates +// rather than this one, which provides shared internal data structures not +// intended for general use. +package parse + +import ( + "bytes" + "fmt" + "runtime" + "strconv" + "strings" +) + +// Tree is the representation of a single parsed template. +type Tree struct { + Name string // name of the template represented by the tree. + ParseName string // name of the top-level template during parsing, for error messages. + Root *ListNode // top-level root of the tree. + Mode Mode // parsing mode. + text string // text parsed to create the template (or its parent) + // Parsing only; cleared after parse. + funcs []map[string]interface{} + lex *lexer + token [3]item // three-token lookahead for parser. + peekCount int + vars []string // variables defined at the moment. + treeSet map[string]*Tree + actionLine int // line of left delim starting action + mode Mode +} + +// A mode value is a set of flags (or 0). Modes control parser behavior. +type Mode uint + +const ( + ParseComments Mode = 1 << iota // parse comments and add them to AST + SkipFuncCheck // do not check that functions are defined +) + +// Copy returns a copy of the Tree. Any parsing state is discarded. +func (t *Tree) Copy() *Tree { + if t == nil { + return nil + } + return &Tree{ + Name: t.Name, + ParseName: t.ParseName, + Root: t.Root.CopyList(), + text: t.text, + } +} + +// Parse returns a map from template name to parse.Tree, created by parsing the +// templates described in the argument string. The top-level template will be +// given the specified name. If an error is encountered, parsing stops and an +// empty map is returned with the error. +func Parse(name, text, leftDelim, rightDelim string, funcs ...map[string]interface{}) (map[string]*Tree, error) { + treeSet := make(map[string]*Tree) + t := New(name) + t.text = text + _, err := t.Parse(text, leftDelim, rightDelim, treeSet, funcs...) + return treeSet, err +} + +// next returns the next token. +func (t *Tree) next() item { + if t.peekCount > 0 { + t.peekCount-- + } else { + t.token[0] = t.lex.nextItem() + } + return t.token[t.peekCount] +} + +// backup backs the input stream up one token. +func (t *Tree) backup() { + t.peekCount++ +} + +// backup2 backs the input stream up two tokens. +// The zeroth token is already there. +func (t *Tree) backup2(t1 item) { + t.token[1] = t1 + t.peekCount = 2 +} + +// backup3 backs the input stream up three tokens +// The zeroth token is already there. +func (t *Tree) backup3(t2, t1 item) { // Reverse order: we're pushing back. + t.token[1] = t1 + t.token[2] = t2 + t.peekCount = 3 +} + +// peek returns but does not consume the next token. +func (t *Tree) peek() item { + if t.peekCount > 0 { + return t.token[t.peekCount-1] + } + t.peekCount = 1 + t.token[0] = t.lex.nextItem() + return t.token[0] +} + +// nextNonSpace returns the next non-space token. +func (t *Tree) nextNonSpace() (token item) { + for { + token = t.next() + if token.typ != itemSpace { + break + } + } + return token +} + +// peekNonSpace returns but does not consume the next non-space token. +func (t *Tree) peekNonSpace() item { + token := t.nextNonSpace() + t.backup() + return token +} + +// Parsing. + +// New allocates a new parse tree with the given name. +func New(name string, funcs ...map[string]interface{}) *Tree { + return &Tree{ + Name: name, + funcs: funcs, + } +} + +// ErrorContext returns a textual representation of the location of the node in the input text. +// The receiver is only used when the node does not have a pointer to the tree inside, +// which can occur in old code. +func (t *Tree) ErrorContext(n Node) (location, context string) { + pos := int(n.Position()) + tree := n.tree() + if tree == nil { + tree = t + } + text := tree.text[:pos] + byteNum := strings.LastIndex(text, "\n") + if byteNum == -1 { + byteNum = pos // On first line. + } else { + byteNum++ // After the newline. + byteNum = pos - byteNum + } + lineNum := 1 + strings.Count(text, "\n") + context = n.String() + return fmt.Sprintf("%s:%d:%d", tree.ParseName, lineNum, byteNum), context +} + +// errorf formats the error and terminates processing. +func (t *Tree) errorf(format string, args ...interface{}) { + t.Root = nil + format = fmt.Sprintf("template: %s:%d: %s", t.ParseName, t.token[0].line, format) + panic(fmt.Errorf(format, args...)) +} + +// error terminates processing. +func (t *Tree) error(err error) { + t.errorf("%s", err) +} + +// expect consumes the next token and guarantees it has the required type. +func (t *Tree) expect(expected itemType, context string) item { + token := t.nextNonSpace() + if token.typ != expected { + t.unexpected(token, context) + } + return token +} + +// expectOneOf consumes the next token and guarantees it has one of the required types. +func (t *Tree) expectOneOf(expected1, expected2 itemType, context string) item { + token := t.nextNonSpace() + if token.typ != expected1 && token.typ != expected2 { + t.unexpected(token, context) + } + return token +} + +// unexpected complains about the token and terminates processing. +func (t *Tree) unexpected(token item, context string) { + if token.typ == itemError { + extra := "" + if t.actionLine != 0 && t.actionLine != token.line { + extra = fmt.Sprintf(" in action started at %s:%d", t.ParseName, t.actionLine) + if strings.HasSuffix(token.val, " action") { + extra = extra[len(" in action"):] // avoid "action in action" + } + } + t.errorf("%s%s", token, extra) + } + t.errorf("unexpected %s in %s", token, context) +} + +// recover is the handler that turns panics into returns from the top level of Parse. +func (t *Tree) recover(errp *error) { + e := recover() + if e != nil { + if _, ok := e.(runtime.Error); ok { + panic(e) + } + if t != nil { + t.lex.drain() + t.stopParse() + } + *errp = e.(error) + } +} + +// startParse initializes the parser, using the lexer. +func (t *Tree) startParse(funcs []map[string]interface{}, lex *lexer, treeSet map[string]*Tree) { + t.Root = nil + t.lex = lex + t.vars = []string{"$"} + t.funcs = funcs + t.treeSet = treeSet +} + +// stopParse terminates parsing. +func (t *Tree) stopParse() { + t.lex = nil + t.vars = nil + t.funcs = nil + t.treeSet = nil +} + +// Parse parses the template definition string to construct a representation of +// the template for execution. If either action delimiter string is empty, the +// default ("{{" or "}}") is used. Embedded template definitions are added to +// the treeSet map. +func (t *Tree) Parse(text, leftDelim, rightDelim string, treeSet map[string]*Tree, funcs ...map[string]interface{}) (tree *Tree, err error) { + defer t.recover(&err) + t.ParseName = t.Name + emitComment := t.Mode&ParseComments != 0 + t.startParse(funcs, lex(t.Name, text, leftDelim, rightDelim, emitComment), treeSet) + t.text = text + t.parse() + t.add() + t.stopParse() + return t, nil +} + +// add adds tree to t.treeSet. +func (t *Tree) add() { + tree := t.treeSet[t.Name] + if tree == nil || IsEmptyTree(tree.Root) { + t.treeSet[t.Name] = t + return + } + if !IsEmptyTree(t.Root) { + t.errorf("template: multiple definition of template %q", t.Name) + } +} + +// IsEmptyTree reports whether this tree (node) is empty of everything but space or comments. +func IsEmptyTree(n Node) bool { + switch n := n.(type) { + case nil: + return true + case *ActionNode: + case *CommentNode: + return true + case *IfNode: + case *ListNode: + for _, node := range n.Nodes { + if !IsEmptyTree(node) { + return false + } + } + return true + case *RangeNode: + case *TemplateNode: + case *TextNode: + return len(bytes.TrimSpace(n.Text)) == 0 + case *WithNode: + default: + panic("unknown node: " + n.String()) + } + return false +} + +// parse is the top-level parser for a template, essentially the same +// as itemList except it also parses {{define}} actions. +// It runs to EOF. +func (t *Tree) parse() { + t.Root = t.newList(t.peek().pos) + for t.peek().typ != itemEOF { + if t.peek().typ == itemLeftDelim { + delim := t.next() + if t.nextNonSpace().typ == itemDefine { + newT := New("definition") // name will be updated once we know it. + newT.text = t.text + newT.Mode = t.Mode + newT.ParseName = t.ParseName + newT.startParse(t.funcs, t.lex, t.treeSet) + newT.parseDefinition() + continue + } + t.backup2(delim) + } + switch n := t.textOrAction(); n.Type() { + case nodeEnd, nodeElse: + t.errorf("unexpected %s", n) + default: + t.Root.append(n) + } + } +} + +// parseDefinition parses a {{define}} ... {{end}} template definition and +// installs the definition in t.treeSet. The "define" keyword has already +// been scanned. +func (t *Tree) parseDefinition() { + const context = "define clause" + name := t.expectOneOf(itemString, itemRawString, context) + var err error + t.Name, err = strconv.Unquote(name.val) + if err != nil { + t.error(err) + } + t.expect(itemRightDelim, context) + var end Node + t.Root, end = t.itemList() + if end.Type() != nodeEnd { + t.errorf("unexpected %s in %s", end, context) + } + t.add() + t.stopParse() +} + +// itemList: +// textOrAction* +// Terminates at {{end}} or {{else}}, returned separately. +func (t *Tree) itemList() (list *ListNode, next Node) { + list = t.newList(t.peekNonSpace().pos) + for t.peekNonSpace().typ != itemEOF { + n := t.textOrAction() + switch n.Type() { + case nodeEnd, nodeElse: + return list, n + } + list.append(n) + } + t.errorf("unexpected EOF") + return +} + +// textOrAction: +// text | comment | action +func (t *Tree) textOrAction() Node { + switch token := t.nextNonSpace(); token.typ { + case itemText: + return t.newText(token.pos, token.val) + case itemLeftDelim: + t.actionLine = token.line + defer t.clearActionLine() + return t.action() + case itemComment: + return t.newComment(token.pos, token.val) + default: + t.unexpected(token, "input") + } + return nil +} + +func (t *Tree) clearActionLine() { + t.actionLine = 0 +} + +// Action: +// control +// command ("|" command)* +// Left delim is past. Now get actions. +// First word could be a keyword such as range. +func (t *Tree) action() (n Node) { + switch token := t.nextNonSpace(); token.typ { + case itemBlock: + return t.blockControl() + case itemElse: + return t.elseControl() + case itemEnd: + return t.endControl() + case itemIf: + return t.ifControl() + case itemRange: + return t.rangeControl() + case itemTemplate: + return t.templateControl() + case itemWith: + return t.withControl() + } + t.backup() + token := t.peek() + // Do not pop variables; they persist until "end". + return t.newAction(token.pos, token.line, t.pipeline("command", itemRightDelim)) +} + +// Pipeline: +// declarations? command ('|' command)* +func (t *Tree) pipeline(context string, end itemType) (pipe *PipeNode) { + token := t.peekNonSpace() + pipe = t.newPipeline(token.pos, token.line, nil) + // Are there declarations or assignments? +decls: + if v := t.peekNonSpace(); v.typ == itemVariable { + t.next() + // Since space is a token, we need 3-token look-ahead here in the worst case: + // in "$x foo" we need to read "foo" (as opposed to ":=") to know that $x is an + // argument variable rather than a declaration. So remember the token + // adjacent to the variable so we can push it back if necessary. + tokenAfterVariable := t.peek() + next := t.peekNonSpace() + switch { + case next.typ == itemAssign, next.typ == itemDeclare: + pipe.IsAssign = next.typ == itemAssign + t.nextNonSpace() + pipe.Decl = append(pipe.Decl, t.newVariable(v.pos, v.val)) + t.vars = append(t.vars, v.val) + case next.typ == itemChar && next.val == ",": + t.nextNonSpace() + pipe.Decl = append(pipe.Decl, t.newVariable(v.pos, v.val)) + t.vars = append(t.vars, v.val) + if context == "range" && len(pipe.Decl) < 2 { + switch t.peekNonSpace().typ { + case itemVariable, itemRightDelim, itemRightParen: + // second initialized variable in a range pipeline + goto decls + default: + t.errorf("range can only initialize variables") + } + } + t.errorf("too many declarations in %s", context) + case tokenAfterVariable.typ == itemSpace: + t.backup3(v, tokenAfterVariable) + default: + t.backup2(v) + } + } + for { + switch token := t.nextNonSpace(); token.typ { + case end: + // At this point, the pipeline is complete + t.checkPipeline(pipe, context) + return + case itemBool, itemCharConstant, itemComplex, itemDot, itemField, itemIdentifier, + itemNumber, itemNil, itemRawString, itemString, itemVariable, itemLeftParen: + t.backup() + pipe.append(t.command()) + default: + t.unexpected(token, context) + } + } +} + +func (t *Tree) checkPipeline(pipe *PipeNode, context string) { + // Reject empty pipelines + if len(pipe.Cmds) == 0 { + t.errorf("missing value for %s", context) + } + // Only the first command of a pipeline can start with a non executable operand + for i, c := range pipe.Cmds[1:] { + switch c.Args[0].Type() { + case NodeBool, NodeDot, NodeNil, NodeNumber, NodeString: + // With A|B|C, pipeline stage 2 is B + t.errorf("non executable command in pipeline stage %d", i+2) + } + } +} + +func (t *Tree) parseControl(allowElseIf bool, context string) (pos Pos, line int, pipe *PipeNode, list, elseList *ListNode) { + defer t.popVars(len(t.vars)) + pipe = t.pipeline(context, itemRightDelim) + var next Node + list, next = t.itemList() + switch next.Type() { + case nodeEnd: //done + case nodeElse: + if allowElseIf { + // Special case for "else if". If the "else" is followed immediately by an "if", + // the elseControl will have left the "if" token pending. Treat + // {{if a}}_{{else if b}}_{{end}} + // as + // {{if a}}_{{else}}{{if b}}_{{end}}{{end}}. + // To do this, parse the if as usual and stop at it {{end}}; the subsequent{{end}} + // is assumed. This technique works even for long if-else-if chains. + // TODO: Should we allow else-if in with and range? + if t.peek().typ == itemIf { + t.next() // Consume the "if" token. + elseList = t.newList(next.Position()) + elseList.append(t.ifControl()) + // Do not consume the next item - only one {{end}} required. + break + } + } + elseList, next = t.itemList() + if next.Type() != nodeEnd { + t.errorf("expected end; found %s", next) + } + } + return pipe.Position(), pipe.Line, pipe, list, elseList +} + +// If: +// {{if pipeline}} itemList {{end}} +// {{if pipeline}} itemList {{else}} itemList {{end}} +// If keyword is past. +func (t *Tree) ifControl() Node { + return t.newIf(t.parseControl(true, "if")) +} + +// Range: +// {{range pipeline}} itemList {{end}} +// {{range pipeline}} itemList {{else}} itemList {{end}} +// Range keyword is past. +func (t *Tree) rangeControl() Node { + return t.newRange(t.parseControl(false, "range")) +} + +// With: +// {{with pipeline}} itemList {{end}} +// {{with pipeline}} itemList {{else}} itemList {{end}} +// If keyword is past. +func (t *Tree) withControl() Node { + return t.newWith(t.parseControl(false, "with")) +} + +// End: +// {{end}} +// End keyword is past. +func (t *Tree) endControl() Node { + return t.newEnd(t.expect(itemRightDelim, "end").pos) +} + +// Else: +// {{else}} +// Else keyword is past. +func (t *Tree) elseControl() Node { + // Special case for "else if". + peek := t.peekNonSpace() + if peek.typ == itemIf { + // We see "{{else if ... " but in effect rewrite it to {{else}}{{if ... ". + return t.newElse(peek.pos, peek.line) + } + token := t.expect(itemRightDelim, "else") + return t.newElse(token.pos, token.line) +} + +// Block: +// {{block stringValue pipeline}} +// Block keyword is past. +// The name must be something that can evaluate to a string. +// The pipeline is mandatory. +func (t *Tree) blockControl() Node { + const context = "block clause" + + token := t.nextNonSpace() + name := t.parseTemplateName(token, context) + pipe := t.pipeline(context, itemRightDelim) + + block := New(name) // name will be updated once we know it. + block.text = t.text + block.Mode = t.Mode + block.ParseName = t.ParseName + block.startParse(t.funcs, t.lex, t.treeSet) + var end Node + block.Root, end = block.itemList() + if end.Type() != nodeEnd { + t.errorf("unexpected %s in %s", end, context) + } + block.add() + block.stopParse() + + return t.newTemplate(token.pos, token.line, name, pipe) +} + +// Template: +// {{template stringValue pipeline}} +// Template keyword is past. The name must be something that can evaluate +// to a string. +func (t *Tree) templateControl() Node { + const context = "template clause" + token := t.nextNonSpace() + name := t.parseTemplateName(token, context) + var pipe *PipeNode + if t.nextNonSpace().typ != itemRightDelim { + t.backup() + // Do not pop variables; they persist until "end". + pipe = t.pipeline(context, itemRightDelim) + } + return t.newTemplate(token.pos, token.line, name, pipe) +} + +func (t *Tree) parseTemplateName(token item, context string) (name string) { + switch token.typ { + case itemString, itemRawString: + s, err := strconv.Unquote(token.val) + if err != nil { + t.error(err) + } + name = s + default: + t.unexpected(token, context) + } + return +} + +// command: +// operand (space operand)* +// space-separated arguments up to a pipeline character or right delimiter. +// we consume the pipe character but leave the right delim to terminate the action. +func (t *Tree) command() *CommandNode { + cmd := t.newCommand(t.peekNonSpace().pos) + for { + t.peekNonSpace() // skip leading spaces. + operand := t.operand() + if operand != nil { + cmd.append(operand) + } + switch token := t.next(); token.typ { + case itemSpace: + continue + case itemRightDelim, itemRightParen: + t.backup() + case itemPipe: + // nothing here; break loop below + default: + t.unexpected(token, "operand") + } + break + } + if len(cmd.Args) == 0 { + t.errorf("empty command") + } + return cmd +} + +// operand: +// term .Field* +// An operand is a space-separated component of a command, +// a term possibly followed by field accesses. +// A nil return means the next item is not an operand. +func (t *Tree) operand() Node { + node := t.term() + if node == nil { + return nil + } + if t.peek().typ == itemField { + chain := t.newChain(t.peek().pos, node) + for t.peek().typ == itemField { + chain.Add(t.next().val) + } + // Compatibility with original API: If the term is of type NodeField + // or NodeVariable, just put more fields on the original. + // Otherwise, keep the Chain node. + // Obvious parsing errors involving literal values are detected here. + // More complex error cases will have to be handled at execution time. + switch node.Type() { + case NodeField: + node = t.newField(chain.Position(), chain.String()) + case NodeVariable: + node = t.newVariable(chain.Position(), chain.String()) + case NodeBool, NodeString, NodeNumber, NodeNil, NodeDot: + t.errorf("unexpected . after term %q", node.String()) + default: + node = chain + } + } + return node +} + +// term: +// literal (number, string, nil, boolean) +// function (identifier) +// . +// .Field +// $ +// '(' pipeline ')' +// A term is a simple "expression". +// A nil return means the next item is not a term. +func (t *Tree) term() Node { + switch token := t.nextNonSpace(); token.typ { + case itemIdentifier: + checkFunc := t.Mode&SkipFuncCheck == 0 + if checkFunc && !t.hasFunction(token.val) { + t.errorf("function %q not defined", token.val) + } + return NewIdentifier(token.val).SetTree(t).SetPos(token.pos) + case itemDot: + return t.newDot(token.pos) + case itemNil: + return t.newNil(token.pos) + case itemVariable: + return t.useVar(token.pos, token.val) + case itemField: + return t.newField(token.pos, token.val) + case itemBool: + return t.newBool(token.pos, token.val == "true") + case itemCharConstant, itemComplex, itemNumber: + number, err := t.newNumber(token.pos, token.val, token.typ) + if err != nil { + t.error(err) + } + return number + case itemLeftParen: + return t.pipeline("parenthesized pipeline", itemRightParen) + case itemString, itemRawString: + s, err := strconv.Unquote(token.val) + if err != nil { + t.error(err) + } + return t.newString(token.pos, token.val, s) + } + t.backup() + return nil +} + +// hasFunction reports if a function name exists in the Tree's maps. +func (t *Tree) hasFunction(name string) bool { + for _, funcMap := range t.funcs { + if funcMap == nil { + continue + } + if funcMap[name] != nil { + return true + } + } + return false +} + +// popVars trims the variable list to the specified length +func (t *Tree) popVars(n int) { + t.vars = t.vars[:n] +} + +// useVar returns a node for a variable reference. It errors if the +// variable is not defined. +func (t *Tree) useVar(pos Pos, name string) Node { + v := t.newVariable(pos, name) + for _, varName := range t.vars { + if varName == v.Ident[0] { + return v + } + } + t.errorf("undefined variable %q", v.Ident[0]) + return nil +} diff --git a/src/text/template/parse/parse_test.go b/src/text/template/parse/parse_test.go new file mode 100644 index 0000000..9b1be27 --- /dev/null +++ b/src/text/template/parse/parse_test.go @@ -0,0 +1,668 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package parse + +import ( + "flag" + "fmt" + "strings" + "testing" +) + +var debug = flag.Bool("debug", false, "show the errors produced by the main tests") + +type numberTest struct { + text string + isInt bool + isUint bool + isFloat bool + isComplex bool + int64 + uint64 + float64 + complex128 +} + +var numberTests = []numberTest{ + // basics + {"0", true, true, true, false, 0, 0, 0, 0}, + {"-0", true, true, true, false, 0, 0, 0, 0}, // check that -0 is a uint. + {"73", true, true, true, false, 73, 73, 73, 0}, + {"7_3", true, true, true, false, 73, 73, 73, 0}, + {"0b10_010_01", true, true, true, false, 73, 73, 73, 0}, + {"0B10_010_01", true, true, true, false, 73, 73, 73, 0}, + {"073", true, true, true, false, 073, 073, 073, 0}, + {"0o73", true, true, true, false, 073, 073, 073, 0}, + {"0O73", true, true, true, false, 073, 073, 073, 0}, + {"0x73", true, true, true, false, 0x73, 0x73, 0x73, 0}, + {"0X73", true, true, true, false, 0x73, 0x73, 0x73, 0}, + {"0x7_3", true, true, true, false, 0x73, 0x73, 0x73, 0}, + {"-73", true, false, true, false, -73, 0, -73, 0}, + {"+73", true, false, true, false, 73, 0, 73, 0}, + {"100", true, true, true, false, 100, 100, 100, 0}, + {"1e9", true, true, true, false, 1e9, 1e9, 1e9, 0}, + {"-1e9", true, false, true, false, -1e9, 0, -1e9, 0}, + {"-1.2", false, false, true, false, 0, 0, -1.2, 0}, + {"1e19", false, true, true, false, 0, 1e19, 1e19, 0}, + {"1e1_9", false, true, true, false, 0, 1e19, 1e19, 0}, + {"1E19", false, true, true, false, 0, 1e19, 1e19, 0}, + {"-1e19", false, false, true, false, 0, 0, -1e19, 0}, + {"0x_1p4", true, true, true, false, 16, 16, 16, 0}, + {"0X_1P4", true, true, true, false, 16, 16, 16, 0}, + {"0x_1p-4", false, false, true, false, 0, 0, 1 / 16., 0}, + {"4i", false, false, false, true, 0, 0, 0, 4i}, + {"-1.2+4.2i", false, false, false, true, 0, 0, 0, -1.2 + 4.2i}, + {"073i", false, false, false, true, 0, 0, 0, 73i}, // not octal! + // complex with 0 imaginary are float (and maybe integer) + {"0i", true, true, true, true, 0, 0, 0, 0}, + {"-1.2+0i", false, false, true, true, 0, 0, -1.2, -1.2}, + {"-12+0i", true, false, true, true, -12, 0, -12, -12}, + {"13+0i", true, true, true, true, 13, 13, 13, 13}, + // funny bases + {"0123", true, true, true, false, 0123, 0123, 0123, 0}, + {"-0x0", true, true, true, false, 0, 0, 0, 0}, + {"0xdeadbeef", true, true, true, false, 0xdeadbeef, 0xdeadbeef, 0xdeadbeef, 0}, + // character constants + {`'a'`, true, true, true, false, 'a', 'a', 'a', 0}, + {`'\n'`, true, true, true, false, '\n', '\n', '\n', 0}, + {`'\\'`, true, true, true, false, '\\', '\\', '\\', 0}, + {`'\''`, true, true, true, false, '\'', '\'', '\'', 0}, + {`'\xFF'`, true, true, true, false, 0xFF, 0xFF, 0xFF, 0}, + {`'パ'`, true, true, true, false, 0x30d1, 0x30d1, 0x30d1, 0}, + {`'\u30d1'`, true, true, true, false, 0x30d1, 0x30d1, 0x30d1, 0}, + {`'\U000030d1'`, true, true, true, false, 0x30d1, 0x30d1, 0x30d1, 0}, + // some broken syntax + {text: "+-2"}, + {text: "0x123."}, + {text: "1e."}, + {text: "0xi."}, + {text: "1+2."}, + {text: "'x"}, + {text: "'xx'"}, + {text: "'433937734937734969526500969526500'"}, // Integer too large - issue 10634. + // Issue 8622 - 0xe parsed as floating point. Very embarrassing. + {"0xef", true, true, true, false, 0xef, 0xef, 0xef, 0}, +} + +func TestNumberParse(t *testing.T) { + for _, test := range numberTests { + // If fmt.Sscan thinks it's complex, it's complex. We can't trust the output + // because imaginary comes out as a number. + var c complex128 + typ := itemNumber + var tree *Tree + if test.text[0] == '\'' { + typ = itemCharConstant + } else { + _, err := fmt.Sscan(test.text, &c) + if err == nil { + typ = itemComplex + } + } + n, err := tree.newNumber(0, test.text, typ) + ok := test.isInt || test.isUint || test.isFloat || test.isComplex + if ok && err != nil { + t.Errorf("unexpected error for %q: %s", test.text, err) + continue + } + if !ok && err == nil { + t.Errorf("expected error for %q", test.text) + continue + } + if !ok { + if *debug { + fmt.Printf("%s\n\t%s\n", test.text, err) + } + continue + } + if n.IsComplex != test.isComplex { + t.Errorf("complex incorrect for %q; should be %t", test.text, test.isComplex) + } + if test.isInt { + if !n.IsInt { + t.Errorf("expected integer for %q", test.text) + } + if n.Int64 != test.int64 { + t.Errorf("int64 for %q should be %d Is %d", test.text, test.int64, n.Int64) + } + } else if n.IsInt { + t.Errorf("did not expect integer for %q", test.text) + } + if test.isUint { + if !n.IsUint { + t.Errorf("expected unsigned integer for %q", test.text) + } + if n.Uint64 != test.uint64 { + t.Errorf("uint64 for %q should be %d Is %d", test.text, test.uint64, n.Uint64) + } + } else if n.IsUint { + t.Errorf("did not expect unsigned integer for %q", test.text) + } + if test.isFloat { + if !n.IsFloat { + t.Errorf("expected float for %q", test.text) + } + if n.Float64 != test.float64 { + t.Errorf("float64 for %q should be %g Is %g", test.text, test.float64, n.Float64) + } + } else if n.IsFloat { + t.Errorf("did not expect float for %q", test.text) + } + if test.isComplex { + if !n.IsComplex { + t.Errorf("expected complex for %q", test.text) + } + if n.Complex128 != test.complex128 { + t.Errorf("complex128 for %q should be %g Is %g", test.text, test.complex128, n.Complex128) + } + } else if n.IsComplex { + t.Errorf("did not expect complex for %q", test.text) + } + } +} + +type parseTest struct { + name string + input string + ok bool + result string // what the user would see in an error message. +} + +const ( + noError = true + hasError = false +) + +var parseTests = []parseTest{ + {"empty", "", noError, + ``}, + {"comment", "{{/*\n\n\n*/}}", noError, + ``}, + {"spaces", " \t\n", noError, + `" \t\n"`}, + {"text", "some text", noError, + `"some text"`}, + {"emptyAction", "{{}}", hasError, + `{{}}`}, + {"field", "{{.X}}", noError, + `{{.X}}`}, + {"simple command", "{{printf}}", noError, + `{{printf}}`}, + {"$ invocation", "{{$}}", noError, + "{{$}}"}, + {"variable invocation", "{{with $x := 3}}{{$x 23}}{{end}}", noError, + "{{with $x := 3}}{{$x 23}}{{end}}"}, + {"variable with fields", "{{$.I}}", noError, + "{{$.I}}"}, + {"multi-word command", "{{printf `%d` 23}}", noError, + "{{printf `%d` 23}}"}, + {"pipeline", "{{.X|.Y}}", noError, + `{{.X | .Y}}`}, + {"pipeline with decl", "{{$x := .X|.Y}}", noError, + `{{$x := .X | .Y}}`}, + {"nested pipeline", "{{.X (.Y .Z) (.A | .B .C) (.E)}}", noError, + `{{.X (.Y .Z) (.A | .B .C) (.E)}}`}, + {"field applied to parentheses", "{{(.Y .Z).Field}}", noError, + `{{(.Y .Z).Field}}`}, + {"simple if", "{{if .X}}hello{{end}}", noError, + `{{if .X}}"hello"{{end}}`}, + {"if with else", "{{if .X}}true{{else}}false{{end}}", noError, + `{{if .X}}"true"{{else}}"false"{{end}}`}, + {"if with else if", "{{if .X}}true{{else if .Y}}false{{end}}", noError, + `{{if .X}}"true"{{else}}{{if .Y}}"false"{{end}}{{end}}`}, + {"if else chain", "+{{if .X}}X{{else if .Y}}Y{{else if .Z}}Z{{end}}+", noError, + `"+"{{if .X}}"X"{{else}}{{if .Y}}"Y"{{else}}{{if .Z}}"Z"{{end}}{{end}}{{end}}"+"`}, + {"simple range", "{{range .X}}hello{{end}}", noError, + `{{range .X}}"hello"{{end}}`}, + {"chained field range", "{{range .X.Y.Z}}hello{{end}}", noError, + `{{range .X.Y.Z}}"hello"{{end}}`}, + {"nested range", "{{range .X}}hello{{range .Y}}goodbye{{end}}{{end}}", noError, + `{{range .X}}"hello"{{range .Y}}"goodbye"{{end}}{{end}}`}, + {"range with else", "{{range .X}}true{{else}}false{{end}}", noError, + `{{range .X}}"true"{{else}}"false"{{end}}`}, + {"range over pipeline", "{{range .X|.M}}true{{else}}false{{end}}", noError, + `{{range .X | .M}}"true"{{else}}"false"{{end}}`}, + {"range []int", "{{range .SI}}{{.}}{{end}}", noError, + `{{range .SI}}{{.}}{{end}}`}, + {"range 1 var", "{{range $x := .SI}}{{.}}{{end}}", noError, + `{{range $x := .SI}}{{.}}{{end}}`}, + {"range 2 vars", "{{range $x, $y := .SI}}{{.}}{{end}}", noError, + `{{range $x, $y := .SI}}{{.}}{{end}}`}, + {"constants", "{{range .SI 1 -3.2i true false 'a' nil}}{{end}}", noError, + `{{range .SI 1 -3.2i true false 'a' nil}}{{end}}`}, + {"template", "{{template `x`}}", noError, + `{{template "x"}}`}, + {"template with arg", "{{template `x` .Y}}", noError, + `{{template "x" .Y}}`}, + {"with", "{{with .X}}hello{{end}}", noError, + `{{with .X}}"hello"{{end}}`}, + {"with with else", "{{with .X}}hello{{else}}goodbye{{end}}", noError, + `{{with .X}}"hello"{{else}}"goodbye"{{end}}`}, + // Trimming spaces. + {"trim left", "x \r\n\t{{- 3}}", noError, `"x"{{3}}`}, + {"trim right", "{{3 -}}\n\n\ty", noError, `{{3}}"y"`}, + {"trim left and right", "x \r\n\t{{- 3 -}}\n\n\ty", noError, `"x"{{3}}"y"`}, + {"trim with extra spaces", "x\n{{- 3 -}}\ny", noError, `"x"{{3}}"y"`}, + {"comment trim left", "x \r\n\t{{- /* hi */}}", noError, `"x"`}, + {"comment trim right", "{{/* hi */ -}}\n\n\ty", noError, `"y"`}, + {"comment trim left and right", "x \r\n\t{{- /* */ -}}\n\n\ty", noError, `"x""y"`}, + {"block definition", `{{block "foo" .}}hello{{end}}`, noError, + `{{template "foo" .}}`}, + + {"newline in assignment", "{{ $x \n := \n 1 \n }}", noError, "{{$x := 1}}"}, + {"newline in empty action", "{{\n}}", hasError, "{{\n}}"}, + {"newline in pipeline", "{{\n\"x\"\n|\nprintf\n}}", noError, `{{"x" | printf}}`}, + {"newline in comment", "{{/*\nhello\n*/}}", noError, ""}, + {"newline in comment", "{{-\n/*\nhello\n*/\n-}}", noError, ""}, + + // Errors. + {"unclosed action", "hello{{range", hasError, ""}, + {"unmatched end", "{{end}}", hasError, ""}, + {"unmatched else", "{{else}}", hasError, ""}, + {"unmatched else after if", "{{if .X}}hello{{end}}{{else}}", hasError, ""}, + {"multiple else", "{{if .X}}1{{else}}2{{else}}3{{end}}", hasError, ""}, + {"missing end", "hello{{range .x}}", hasError, ""}, + {"missing end after else", "hello{{range .x}}{{else}}", hasError, ""}, + {"undefined function", "hello{{undefined}}", hasError, ""}, + {"undefined variable", "{{$x}}", hasError, ""}, + {"variable undefined after end", "{{with $x := 4}}{{end}}{{$x}}", hasError, ""}, + {"variable undefined in template", "{{template $v}}", hasError, ""}, + {"declare with field", "{{with $x.Y := 4}}{{end}}", hasError, ""}, + {"template with field ref", "{{template .X}}", hasError, ""}, + {"template with var", "{{template $v}}", hasError, ""}, + {"invalid punctuation", "{{printf 3, 4}}", hasError, ""}, + {"multidecl outside range", "{{with $v, $u := 3}}{{end}}", hasError, ""}, + {"too many decls in range", "{{range $u, $v, $w := 3}}{{end}}", hasError, ""}, + {"dot applied to parentheses", "{{printf (printf .).}}", hasError, ""}, + {"adjacent args", "{{printf 3`x`}}", hasError, ""}, + {"adjacent args with .", "{{printf `x`.}}", hasError, ""}, + {"extra end after if", "{{if .X}}a{{else if .Y}}b{{end}}{{end}}", hasError, ""}, + // Other kinds of assignments and operators aren't available yet. + {"bug0a", "{{$x := 0}}{{$x}}", noError, "{{$x := 0}}{{$x}}"}, + {"bug0b", "{{$x += 1}}{{$x}}", hasError, ""}, + {"bug0c", "{{$x ! 2}}{{$x}}", hasError, ""}, + {"bug0d", "{{$x % 3}}{{$x}}", hasError, ""}, + // Check the parse fails for := rather than comma. + {"bug0e", "{{range $x := $y := 3}}{{end}}", hasError, ""}, + // Another bug: variable read must ignore following punctuation. + {"bug1a", "{{$x:=.}}{{$x!2}}", hasError, ""}, // ! is just illegal here. + {"bug1b", "{{$x:=.}}{{$x+2}}", hasError, ""}, // $x+2 should not parse as ($x) (+2). + {"bug1c", "{{$x:=.}}{{$x +2}}", noError, "{{$x := .}}{{$x +2}}"}, // It's OK with a space. + // dot following a literal value + {"dot after integer", "{{1.E}}", hasError, ""}, + {"dot after float", "{{0.1.E}}", hasError, ""}, + {"dot after boolean", "{{true.E}}", hasError, ""}, + {"dot after char", "{{'a'.any}}", hasError, ""}, + {"dot after string", `{{"hello".guys}}`, hasError, ""}, + {"dot after dot", "{{..E}}", hasError, ""}, + {"dot after nil", "{{nil.E}}", hasError, ""}, + // Wrong pipeline + {"wrong pipeline dot", "{{12|.}}", hasError, ""}, + {"wrong pipeline number", "{{.|12|printf}}", hasError, ""}, + {"wrong pipeline string", "{{.|printf|\"error\"}}", hasError, ""}, + {"wrong pipeline char", "{{12|printf|'e'}}", hasError, ""}, + {"wrong pipeline boolean", "{{.|true}}", hasError, ""}, + {"wrong pipeline nil", "{{'c'|nil}}", hasError, ""}, + {"empty pipeline", `{{printf "%d" ( ) }}`, hasError, ""}, + // Missing pipeline in block + {"block definition", `{{block "foo"}}hello{{end}}`, hasError, ""}, +} + +var builtins = map[string]interface{}{ + "printf": fmt.Sprintf, + "contains": strings.Contains, +} + +func testParse(doCopy bool, t *testing.T) { + textFormat = "%q" + defer func() { textFormat = "%s" }() + for _, test := range parseTests { + tmpl, err := New(test.name).Parse(test.input, "", "", make(map[string]*Tree), builtins) + switch { + case err == nil && !test.ok: + t.Errorf("%q: expected error; got none", test.name) + continue + case err != nil && test.ok: + t.Errorf("%q: unexpected error: %v", test.name, err) + continue + case err != nil && !test.ok: + // expected error, got one + if *debug { + fmt.Printf("%s: %s\n\t%s\n", test.name, test.input, err) + } + continue + } + var result string + if doCopy { + result = tmpl.Root.Copy().String() + } else { + result = tmpl.Root.String() + } + if result != test.result { + t.Errorf("%s=(%q): got\n\t%v\nexpected\n\t%v", test.name, test.input, result, test.result) + } + } +} + +func TestParse(t *testing.T) { + testParse(false, t) +} + +// Same as TestParse, but we copy the node first +func TestParseCopy(t *testing.T) { + testParse(true, t) +} + +func TestParseWithComments(t *testing.T) { + textFormat = "%q" + defer func() { textFormat = "%s" }() + tests := [...]parseTest{ + {"comment", "{{/*\n\n\n*/}}", noError, "{{/*\n\n\n*/}}"}, + {"comment trim left", "x \r\n\t{{- /* hi */}}", noError, `"x"{{/* hi */}}`}, + {"comment trim right", "{{/* hi */ -}}\n\n\ty", noError, `{{/* hi */}}"y"`}, + {"comment trim left and right", "x \r\n\t{{- /* */ -}}\n\n\ty", noError, `"x"{{/* */}}"y"`}, + } + for _, test := range tests { + t.Run(test.name, func(t *testing.T) { + tr := New(test.name) + tr.Mode = ParseComments + tmpl, err := tr.Parse(test.input, "", "", make(map[string]*Tree)) + if err != nil { + t.Errorf("%q: expected error; got none", test.name) + } + if result := tmpl.Root.String(); result != test.result { + t.Errorf("%s=(%q): got\n\t%v\nexpected\n\t%v", test.name, test.input, result, test.result) + } + }) + } +} + +func TestSkipFuncCheck(t *testing.T) { + oldTextFormat := textFormat + textFormat = "%q" + defer func() { textFormat = oldTextFormat }() + tr := New("skip func check") + tr.Mode = SkipFuncCheck + tmpl, err := tr.Parse("{{fn 1 2}}", "", "", make(map[string]*Tree)) + if err != nil { + t.Fatalf("unexpected error: %v", err) + } + expected := "{{fn 1 2}}" + if result := tmpl.Root.String(); result != expected { + t.Errorf("got\n\t%v\nexpected\n\t%v", result, expected) + } +} + +type isEmptyTest struct { + name string + input string + empty bool +} + +var isEmptyTests = []isEmptyTest{ + {"empty", ``, true}, + {"nonempty", `hello`, false}, + {"spaces only", " \t\n \t\n", true}, + {"comment only", "{{/* comment */}}", true}, + {"definition", `{{define "x"}}something{{end}}`, true}, + {"definitions and space", "{{define `x`}}something{{end}}\n\n{{define `y`}}something{{end}}\n\n", true}, + {"definitions and text", "{{define `x`}}something{{end}}\nx\n{{define `y`}}something{{end}}\ny\n", false}, + {"definition and action", "{{define `x`}}something{{end}}{{if 3}}foo{{end}}", false}, +} + +func TestIsEmpty(t *testing.T) { + if !IsEmptyTree(nil) { + t.Errorf("nil tree is not empty") + } + for _, test := range isEmptyTests { + tree, err := New("root").Parse(test.input, "", "", make(map[string]*Tree), nil) + if err != nil { + t.Errorf("%q: unexpected error: %v", test.name, err) + continue + } + if empty := IsEmptyTree(tree.Root); empty != test.empty { + t.Errorf("%q: expected %t got %t", test.name, test.empty, empty) + } + } +} + +func TestErrorContextWithTreeCopy(t *testing.T) { + tree, err := New("root").Parse("{{if true}}{{end}}", "", "", make(map[string]*Tree), nil) + if err != nil { + t.Fatalf("unexpected tree parse failure: %v", err) + } + treeCopy := tree.Copy() + wantLocation, wantContext := tree.ErrorContext(tree.Root.Nodes[0]) + gotLocation, gotContext := treeCopy.ErrorContext(treeCopy.Root.Nodes[0]) + if wantLocation != gotLocation { + t.Errorf("wrong error location want %q got %q", wantLocation, gotLocation) + } + if wantContext != gotContext { + t.Errorf("wrong error location want %q got %q", wantContext, gotContext) + } +} + +// All failures, and the result is a string that must appear in the error message. +var errorTests = []parseTest{ + // Check line numbers are accurate. + {"unclosed1", + "line1\n{{", + hasError, `unclosed1:2: unclosed action`}, + {"unclosed2", + "line1\n{{define `x`}}line2\n{{", + hasError, `unclosed2:3: unclosed action`}, + {"unclosed3", + "line1\n{{\"x\"\n\"y\"\n", + hasError, `unclosed3:4: unclosed action started at unclosed3:2`}, + {"unclosed4", + "{{\n\n\n\n\n", + hasError, `unclosed4:6: unclosed action started at unclosed4:1`}, + {"var1", + "line1\n{{\nx\n}}", + hasError, `var1:3: function "x" not defined`}, + // Specific errors. + {"function", + "{{foo}}", + hasError, `function "foo" not defined`}, + {"comment1", + "{{/*}}", + hasError, `comment1:1: unclosed comment`}, + {"comment2", + "{{/*\nhello\n}}", + hasError, `comment2:1: unclosed comment`}, + {"lparen", + "{{.X (1 2 3}}", + hasError, `unclosed left paren`}, + {"rparen", + "{{.X 1 2 3 ) }}", + hasError, `unexpected ")" in command`}, + {"rparen2", + "{{(.X 1 2 3", + hasError, `unclosed action`}, + {"space", + "{{`x`3}}", + hasError, `in operand`}, + {"idchar", + "{{a#}}", + hasError, `'#'`}, + {"charconst", + "{{'a}}", + hasError, `unterminated character constant`}, + {"stringconst", + `{{"a}}`, + hasError, `unterminated quoted string`}, + {"rawstringconst", + "{{`a}}", + hasError, `unterminated raw quoted string`}, + {"number", + "{{0xi}}", + hasError, `number syntax`}, + {"multidefine", + "{{define `a`}}a{{end}}{{define `a`}}b{{end}}", + hasError, `multiple definition of template`}, + {"eof", + "{{range .X}}", + hasError, `unexpected EOF`}, + {"variable", + // Declare $x so it's defined, to avoid that error, and then check we don't parse a declaration. + "{{$x := 23}}{{with $x.y := 3}}{{$x 23}}{{end}}", + hasError, `unexpected ":="`}, + {"multidecl", + "{{$a,$b,$c := 23}}", + hasError, `too many declarations`}, + {"undefvar", + "{{$a}}", + hasError, `undefined variable`}, + {"wrongdot", + "{{true.any}}", + hasError, `unexpected . after term`}, + {"wrongpipeline", + "{{12|false}}", + hasError, `non executable command in pipeline`}, + {"emptypipeline", + `{{ ( ) }}`, + hasError, `missing value for parenthesized pipeline`}, + {"multilinerawstring", + "{{ $v := `\n` }} {{", + hasError, `multilinerawstring:2: unclosed action`}, + {"rangeundefvar", + "{{range $k}}{{end}}", + hasError, `undefined variable`}, + {"rangeundefvars", + "{{range $k, $v}}{{end}}", + hasError, `undefined variable`}, + {"rangemissingvalue1", + "{{range $k,}}{{end}}", + hasError, `missing value for range`}, + {"rangemissingvalue2", + "{{range $k, $v := }}{{end}}", + hasError, `missing value for range`}, + {"rangenotvariable1", + "{{range $k, .}}{{end}}", + hasError, `range can only initialize variables`}, + {"rangenotvariable2", + "{{range $k, 123 := .}}{{end}}", + hasError, `range can only initialize variables`}, +} + +func TestErrors(t *testing.T) { + for _, test := range errorTests { + t.Run(test.name, func(t *testing.T) { + _, err := New(test.name).Parse(test.input, "", "", make(map[string]*Tree)) + if err == nil { + t.Fatalf("expected error %q, got nil", test.result) + } + if !strings.Contains(err.Error(), test.result) { + t.Fatalf("error %q does not contain %q", err, test.result) + } + }) + } +} + +func TestBlock(t *testing.T) { + const ( + input = `a{{block "inner" .}}bar{{.}}baz{{end}}b` + outer = `a{{template "inner" .}}b` + inner = `bar{{.}}baz` + ) + treeSet := make(map[string]*Tree) + tmpl, err := New("outer").Parse(input, "", "", treeSet, nil) + if err != nil { + t.Fatal(err) + } + if g, w := tmpl.Root.String(), outer; g != w { + t.Errorf("outer template = %q, want %q", g, w) + } + inTmpl := treeSet["inner"] + if inTmpl == nil { + t.Fatal("block did not define template") + } + if g, w := inTmpl.Root.String(), inner; g != w { + t.Errorf("inner template = %q, want %q", g, w) + } +} + +func TestLineNum(t *testing.T) { + const count = 100 + text := strings.Repeat("{{printf 1234}}\n", count) + tree, err := New("bench").Parse(text, "", "", make(map[string]*Tree), builtins) + if err != nil { + t.Fatal(err) + } + // Check the line numbers. Each line is an action containing a template, followed by text. + // That's two nodes per line. + nodes := tree.Root.Nodes + for i := 0; i < len(nodes); i += 2 { + line := 1 + i/2 + // Action first. + action := nodes[i].(*ActionNode) + if action.Line != line { + t.Fatalf("line %d: action is line %d", line, action.Line) + } + pipe := action.Pipe + if pipe.Line != line { + t.Fatalf("line %d: pipe is line %d", line, pipe.Line) + } + } +} + +func BenchmarkParseLarge(b *testing.B) { + text := strings.Repeat("{{1234}}\n", 10000) + for i := 0; i < b.N; i++ { + _, err := New("bench").Parse(text, "", "", make(map[string]*Tree), builtins) + if err != nil { + b.Fatal(err) + } + } +} + +var sinkv, sinkl string + +func BenchmarkVariableString(b *testing.B) { + v := &VariableNode{ + Ident: []string{"$", "A", "BB", "CCC", "THIS_IS_THE_VARIABLE_BEING_PROCESSED"}, + } + b.ResetTimer() + b.ReportAllocs() + for i := 0; i < b.N; i++ { + sinkv = v.String() + } + if sinkv == "" { + b.Fatal("Benchmark was not run") + } +} + +func BenchmarkListString(b *testing.B) { + text := ` +{{(printf .Field1.Field2.Field3).Value}} +{{$x := (printf .Field1.Field2.Field3).Value}} +{{$y := (printf $x.Field1.Field2.Field3).Value}} +{{$z := $y.Field1.Field2.Field3}} +{{if contains $y $z}} + {{printf "%q" $y}} +{{else}} + {{printf "%q" $x}} +{{end}} +{{with $z.Field1 | contains "boring"}} + {{printf "%q" . | printf "%s"}} +{{else}} + {{printf "%d %d %d" 11 11 11}} + {{printf "%d %d %s" 22 22 $x.Field1.Field2.Field3 | printf "%s"}} + {{printf "%v" (contains $z.Field1.Field2 $y)}} +{{end}} +` + tree, err := New("bench").Parse(text, "", "", make(map[string]*Tree), builtins) + if err != nil { + b.Fatal(err) + } + b.ResetTimer() + b.ReportAllocs() + for i := 0; i < b.N; i++ { + sinkl = tree.Root.String() + } + if sinkl == "" { + b.Fatal("Benchmark was not run") + } +} diff --git a/src/text/template/template.go b/src/text/template/template.go new file mode 100644 index 0000000..776be9c --- /dev/null +++ b/src/text/template/template.go @@ -0,0 +1,238 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package template + +import ( + "reflect" + "sync" + "text/template/parse" +) + +// common holds the information shared by related templates. +type common struct { + tmpl map[string]*Template // Map from name to defined templates. + muTmpl sync.RWMutex // protects tmpl + option option + // We use two maps, one for parsing and one for execution. + // This separation makes the API cleaner since it doesn't + // expose reflection to the client. + muFuncs sync.RWMutex // protects parseFuncs and execFuncs + parseFuncs FuncMap + execFuncs map[string]reflect.Value +} + +// Template is the representation of a parsed template. The *parse.Tree +// field is exported only for use by html/template and should be treated +// as unexported by all other clients. +type Template struct { + name string + *parse.Tree + *common + leftDelim string + rightDelim string +} + +// New allocates a new, undefined template with the given name. +func New(name string) *Template { + t := &Template{ + name: name, + } + t.init() + return t +} + +// Name returns the name of the template. +func (t *Template) Name() string { + return t.name +} + +// New allocates a new, undefined template associated with the given one and with the same +// delimiters. The association, which is transitive, allows one template to +// invoke another with a {{template}} action. +// +// Because associated templates share underlying data, template construction +// cannot be done safely in parallel. Once the templates are constructed, they +// can be executed in parallel. +func (t *Template) New(name string) *Template { + t.init() + nt := &Template{ + name: name, + common: t.common, + leftDelim: t.leftDelim, + rightDelim: t.rightDelim, + } + return nt +} + +// init guarantees that t has a valid common structure. +func (t *Template) init() { + if t.common == nil { + c := new(common) + c.tmpl = make(map[string]*Template) + c.parseFuncs = make(FuncMap) + c.execFuncs = make(map[string]reflect.Value) + t.common = c + } +} + +// Clone returns a duplicate of the template, including all associated +// templates. The actual representation is not copied, but the name space of +// associated templates is, so further calls to Parse in the copy will add +// templates to the copy but not to the original. Clone can be used to prepare +// common templates and use them with variant definitions for other templates +// by adding the variants after the clone is made. +func (t *Template) Clone() (*Template, error) { + nt := t.copy(nil) + nt.init() + if t.common == nil { + return nt, nil + } + t.muTmpl.RLock() + defer t.muTmpl.RUnlock() + for k, v := range t.tmpl { + if k == t.name { + nt.tmpl[t.name] = nt + continue + } + // The associated templates share nt's common structure. + tmpl := v.copy(nt.common) + nt.tmpl[k] = tmpl + } + t.muFuncs.RLock() + defer t.muFuncs.RUnlock() + for k, v := range t.parseFuncs { + nt.parseFuncs[k] = v + } + for k, v := range t.execFuncs { + nt.execFuncs[k] = v + } + return nt, nil +} + +// copy returns a shallow copy of t, with common set to the argument. +func (t *Template) copy(c *common) *Template { + return &Template{ + name: t.name, + Tree: t.Tree, + common: c, + leftDelim: t.leftDelim, + rightDelim: t.rightDelim, + } +} + +// AddParseTree associates the argument parse tree with the template t, giving +// it the specified name. If the template has not been defined, this tree becomes +// its definition. If it has been defined and already has that name, the existing +// definition is replaced; otherwise a new template is created, defined, and returned. +func (t *Template) AddParseTree(name string, tree *parse.Tree) (*Template, error) { + t.init() + t.muTmpl.Lock() + defer t.muTmpl.Unlock() + nt := t + if name != t.name { + nt = t.New(name) + } + // Even if nt == t, we need to install it in the common.tmpl map. + if t.associate(nt, tree) || nt.Tree == nil { + nt.Tree = tree + } + return nt, nil +} + +// Templates returns a slice of defined templates associated with t. +func (t *Template) Templates() []*Template { + if t.common == nil { + return nil + } + // Return a slice so we don't expose the map. + t.muTmpl.RLock() + defer t.muTmpl.RUnlock() + m := make([]*Template, 0, len(t.tmpl)) + for _, v := range t.tmpl { + m = append(m, v) + } + return m +} + +// Delims sets the action delimiters to the specified strings, to be used in +// subsequent calls to Parse, ParseFiles, or ParseGlob. Nested template +// definitions will inherit the settings. An empty delimiter stands for the +// corresponding default: {{ or }}. +// The return value is the template, so calls can be chained. +func (t *Template) Delims(left, right string) *Template { + t.init() + t.leftDelim = left + t.rightDelim = right + return t +} + +// Funcs adds the elements of the argument map to the template's function map. +// It must be called before the template is parsed. +// It panics if a value in the map is not a function with appropriate return +// type or if the name cannot be used syntactically as a function in a template. +// It is legal to overwrite elements of the map. The return value is the template, +// so calls can be chained. +func (t *Template) Funcs(funcMap FuncMap) *Template { + t.init() + t.muFuncs.Lock() + defer t.muFuncs.Unlock() + addValueFuncs(t.execFuncs, funcMap) + addFuncs(t.parseFuncs, funcMap) + return t +} + +// Lookup returns the template with the given name that is associated with t. +// It returns nil if there is no such template or the template has no definition. +func (t *Template) Lookup(name string) *Template { + if t.common == nil { + return nil + } + t.muTmpl.RLock() + defer t.muTmpl.RUnlock() + return t.tmpl[name] +} + +// Parse parses text as a template body for t. +// Named template definitions ({{define ...}} or {{block ...}} statements) in text +// define additional templates associated with t and are removed from the +// definition of t itself. +// +// Templates can be redefined in successive calls to Parse. +// A template definition with a body containing only white space and comments +// is considered empty and will not replace an existing template's body. +// This allows using Parse to add new named template definitions without +// overwriting the main template body. +func (t *Template) Parse(text string) (*Template, error) { + t.init() + t.muFuncs.RLock() + trees, err := parse.Parse(t.name, text, t.leftDelim, t.rightDelim, t.parseFuncs, builtins()) + t.muFuncs.RUnlock() + if err != nil { + return nil, err + } + // Add the newly parsed trees, including the one for t, into our common structure. + for name, tree := range trees { + if _, err := t.AddParseTree(name, tree); err != nil { + return nil, err + } + } + return t, nil +} + +// associate installs the new template into the group of templates associated +// with t. The two are already known to share the common structure. +// The boolean return value reports whether to store this tree as t.Tree. +func (t *Template) associate(new *Template, tree *parse.Tree) bool { + if new.common != t.common { + panic("internal error: associate not common") + } + if old := t.tmpl[new.name]; old != nil && parse.IsEmptyTree(tree.Root) && old.Tree != nil { + // If a template by that name exists, + // don't replace it with an empty template. + return false + } + t.tmpl[new.name] = new + return true +} diff --git a/src/text/template/testdata/file1.tmpl b/src/text/template/testdata/file1.tmpl new file mode 100644 index 0000000..febf9d9 --- /dev/null +++ b/src/text/template/testdata/file1.tmpl @@ -0,0 +1,2 @@ +{{define "x"}}TEXT{{end}} +{{define "dotV"}}{{.V}}{{end}} diff --git a/src/text/template/testdata/file2.tmpl b/src/text/template/testdata/file2.tmpl new file mode 100644 index 0000000..39bf6fb --- /dev/null +++ b/src/text/template/testdata/file2.tmpl @@ -0,0 +1,2 @@ +{{define "dot"}}{{.}}{{end}} +{{define "nested"}}{{template "dot" .}}{{end}} diff --git a/src/text/template/testdata/tmpl1.tmpl b/src/text/template/testdata/tmpl1.tmpl new file mode 100644 index 0000000..b72b3a3 --- /dev/null +++ b/src/text/template/testdata/tmpl1.tmpl @@ -0,0 +1,3 @@ +template1 +{{define "x"}}x{{end}} +{{template "y"}} diff --git a/src/text/template/testdata/tmpl2.tmpl b/src/text/template/testdata/tmpl2.tmpl new file mode 100644 index 0000000..16beba6 --- /dev/null +++ b/src/text/template/testdata/tmpl2.tmpl @@ -0,0 +1,3 @@ +template2 +{{define "y"}}y{{end}} +{{template "x"}} |