From 26a029d407be480d791972afb5975cf62c9360a6 Mon Sep 17 00:00:00 2001 From: Daniel Baumann Date: Fri, 19 Apr 2024 02:47:55 +0200 Subject: Adding upstream version 124.0.1. Signed-off-by: Daniel Baumann --- intl/icu/source/common/uts46.cpp | 1494 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 1494 insertions(+) create mode 100644 intl/icu/source/common/uts46.cpp (limited to 'intl/icu/source/common/uts46.cpp') diff --git a/intl/icu/source/common/uts46.cpp b/intl/icu/source/common/uts46.cpp new file mode 100644 index 0000000000..fce2af4130 --- /dev/null +++ b/intl/icu/source/common/uts46.cpp @@ -0,0 +1,1494 @@ +// © 2016 and later: Unicode, Inc. and others. +// License & terms of use: http://www.unicode.org/copyright.html +/* +******************************************************************************* +* Copyright (C) 2010-2015, International Business Machines +* Corporation and others. All Rights Reserved. +******************************************************************************* +* file name: uts46.cpp +* encoding: UTF-8 +* tab size: 8 (not used) +* indentation:4 +* +* created on: 2010mar09 +* created by: Markus W. Scherer +*/ + +#include "unicode/utypes.h" + +#if !UCONFIG_NO_IDNA + +#include "unicode/idna.h" +#include "unicode/normalizer2.h" +#include "unicode/uscript.h" +#include "unicode/ustring.h" +#include "unicode/utf16.h" +#include "cmemory.h" +#include "cstring.h" +#include "punycode.h" +#include "ubidi_props.h" +#include "ustr_imp.h" + +// Note about tests for UIDNA_ERROR_DOMAIN_NAME_TOO_LONG: +// +// The domain name length limit is 255 octets in an internal DNS representation +// where the last ("root") label is the empty label +// represented by length byte 0 alone. +// In a conventional string, this translates to 253 characters, or 254 +// if there is a trailing dot for the root label. + +U_NAMESPACE_BEGIN + +// Severe errors which usually result in a U+FFFD replacement character in the result string. +const uint32_t severeErrors= + UIDNA_ERROR_LEADING_COMBINING_MARK| + UIDNA_ERROR_DISALLOWED| + UIDNA_ERROR_PUNYCODE| + UIDNA_ERROR_LABEL_HAS_DOT| + UIDNA_ERROR_INVALID_ACE_LABEL; + +static inline UBool +isASCIIString(const UnicodeString &dest) { + const char16_t *s=dest.getBuffer(); + const char16_t *limit=s+dest.length(); + while(s0x7f) { + return false; + } + } + return true; +} + +static UBool +isASCIIOkBiDi(const char16_t *s, int32_t length); + +static UBool +isASCIIOkBiDi(const char *s, int32_t length); + +// IDNA class default implementations -------------------------------------- *** + +IDNA::~IDNA() {} + +void +IDNA::labelToASCII_UTF8(StringPiece label, ByteSink &dest, + IDNAInfo &info, UErrorCode &errorCode) const { + if(U_SUCCESS(errorCode)) { + UnicodeString destString; + labelToASCII(UnicodeString::fromUTF8(label), destString, + info, errorCode).toUTF8(dest); + } +} + +void +IDNA::labelToUnicodeUTF8(StringPiece label, ByteSink &dest, + IDNAInfo &info, UErrorCode &errorCode) const { + if(U_SUCCESS(errorCode)) { + UnicodeString destString; + labelToUnicode(UnicodeString::fromUTF8(label), destString, + info, errorCode).toUTF8(dest); + } +} + +void +IDNA::nameToASCII_UTF8(StringPiece name, ByteSink &dest, + IDNAInfo &info, UErrorCode &errorCode) const { + if(U_SUCCESS(errorCode)) { + UnicodeString destString; + nameToASCII(UnicodeString::fromUTF8(name), destString, + info, errorCode).toUTF8(dest); + } +} + +void +IDNA::nameToUnicodeUTF8(StringPiece name, ByteSink &dest, + IDNAInfo &info, UErrorCode &errorCode) const { + if(U_SUCCESS(errorCode)) { + UnicodeString destString; + nameToUnicode(UnicodeString::fromUTF8(name), destString, + info, errorCode).toUTF8(dest); + } +} + +// UTS46 class declaration ------------------------------------------------- *** + +class UTS46 : public IDNA { +public: + UTS46(uint32_t options, UErrorCode &errorCode); + virtual ~UTS46(); + + virtual UnicodeString & + labelToASCII(const UnicodeString &label, UnicodeString &dest, + IDNAInfo &info, UErrorCode &errorCode) const override; + + virtual UnicodeString & + labelToUnicode(const UnicodeString &label, UnicodeString &dest, + IDNAInfo &info, UErrorCode &errorCode) const override; + + virtual UnicodeString & + nameToASCII(const UnicodeString &name, UnicodeString &dest, + IDNAInfo &info, UErrorCode &errorCode) const override; + + virtual UnicodeString & + nameToUnicode(const UnicodeString &name, UnicodeString &dest, + IDNAInfo &info, UErrorCode &errorCode) const override; + + virtual void + labelToASCII_UTF8(StringPiece label, ByteSink &dest, + IDNAInfo &info, UErrorCode &errorCode) const override; + + virtual void + labelToUnicodeUTF8(StringPiece label, ByteSink &dest, + IDNAInfo &info, UErrorCode &errorCode) const override; + + virtual void + nameToASCII_UTF8(StringPiece name, ByteSink &dest, + IDNAInfo &info, UErrorCode &errorCode) const override; + + virtual void + nameToUnicodeUTF8(StringPiece name, ByteSink &dest, + IDNAInfo &info, UErrorCode &errorCode) const override; + +private: + UnicodeString & + process(const UnicodeString &src, + UBool isLabel, UBool toASCII, + UnicodeString &dest, + IDNAInfo &info, UErrorCode &errorCode) const; + + void + processUTF8(StringPiece src, + UBool isLabel, UBool toASCII, + ByteSink &dest, + IDNAInfo &info, UErrorCode &errorCode) const; + + UnicodeString & + processUnicode(const UnicodeString &src, + int32_t labelStart, int32_t mappingStart, + UBool isLabel, UBool toASCII, + UnicodeString &dest, + IDNAInfo &info, UErrorCode &errorCode) const; + + // returns the new dest.length() + int32_t + mapDevChars(UnicodeString &dest, int32_t labelStart, int32_t mappingStart, + UErrorCode &errorCode) const; + + // returns the new label length + int32_t + processLabel(UnicodeString &dest, + int32_t labelStart, int32_t labelLength, + UBool toASCII, + IDNAInfo &info, UErrorCode &errorCode) const; + int32_t + markBadACELabel(UnicodeString &dest, + int32_t labelStart, int32_t labelLength, + UBool toASCII, IDNAInfo &info, UErrorCode &errorCode) const; + + void + checkLabelBiDi(const char16_t *label, int32_t labelLength, IDNAInfo &info) const; + + UBool + isLabelOkContextJ(const char16_t *label, int32_t labelLength) const; + + void + checkLabelContextO(const char16_t *label, int32_t labelLength, IDNAInfo &info) const; + + const Normalizer2 &uts46Norm2; // uts46.nrm + uint32_t options; +}; + +IDNA * +IDNA::createUTS46Instance(uint32_t options, UErrorCode &errorCode) { + if(U_SUCCESS(errorCode)) { + IDNA *idna=new UTS46(options, errorCode); + if(idna==nullptr) { + errorCode=U_MEMORY_ALLOCATION_ERROR; + } else if(U_FAILURE(errorCode)) { + delete idna; + idna=nullptr; + } + return idna; + } else { + return nullptr; + } +} + +// UTS46 implementation ---------------------------------------------------- *** + +UTS46::UTS46(uint32_t opt, UErrorCode &errorCode) + : uts46Norm2(*Normalizer2::getInstance(nullptr, "uts46", UNORM2_COMPOSE, errorCode)), + options(opt) {} + +UTS46::~UTS46() {} + +UnicodeString & +UTS46::labelToASCII(const UnicodeString &label, UnicodeString &dest, + IDNAInfo &info, UErrorCode &errorCode) const { + return process(label, true, true, dest, info, errorCode); +} + +UnicodeString & +UTS46::labelToUnicode(const UnicodeString &label, UnicodeString &dest, + IDNAInfo &info, UErrorCode &errorCode) const { + return process(label, true, false, dest, info, errorCode); +} + +UnicodeString & +UTS46::nameToASCII(const UnicodeString &name, UnicodeString &dest, + IDNAInfo &info, UErrorCode &errorCode) const { + process(name, false, true, dest, info, errorCode); + if( dest.length()>=254 && (info.errors&UIDNA_ERROR_DOMAIN_NAME_TOO_LONG)==0 && + isASCIIString(dest) && + (dest.length()>254 || dest[253]!=0x2e) + ) { + info.errors|=UIDNA_ERROR_DOMAIN_NAME_TOO_LONG; + } + return dest; +} + +UnicodeString & +UTS46::nameToUnicode(const UnicodeString &name, UnicodeString &dest, + IDNAInfo &info, UErrorCode &errorCode) const { + return process(name, false, false, dest, info, errorCode); +} + +void +UTS46::labelToASCII_UTF8(StringPiece label, ByteSink &dest, + IDNAInfo &info, UErrorCode &errorCode) const { + processUTF8(label, true, true, dest, info, errorCode); +} + +void +UTS46::labelToUnicodeUTF8(StringPiece label, ByteSink &dest, + IDNAInfo &info, UErrorCode &errorCode) const { + processUTF8(label, true, false, dest, info, errorCode); +} + +void +UTS46::nameToASCII_UTF8(StringPiece name, ByteSink &dest, + IDNAInfo &info, UErrorCode &errorCode) const { + processUTF8(name, false, true, dest, info, errorCode); +} + +void +UTS46::nameToUnicodeUTF8(StringPiece name, ByteSink &dest, + IDNAInfo &info, UErrorCode &errorCode) const { + processUTF8(name, false, false, dest, info, errorCode); +} + +// UTS #46 data for ASCII characters. +// The normalizer (using uts46.nrm) maps uppercase ASCII letters to lowercase +// and passes through all other ASCII characters. +// If UIDNA_USE_STD3_RULES is set, then non-LDH characters are disallowed +// using this data. +// The ASCII fastpath also uses this data. +// Values: -1=disallowed 0==valid 1==mapped (lowercase) +static const int8_t asciiData[128]={ + -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, + -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, + // 002D..002E; valid # HYPHEN-MINUS..FULL STOP + -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 0, 0, -1, + // 0030..0039; valid # DIGIT ZERO..DIGIT NINE + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, + // 0041..005A; mapped # LATIN CAPITAL LETTER A..LATIN CAPITAL LETTER Z + -1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, + // 0061..007A; valid # LATIN SMALL LETTER A..LATIN SMALL LETTER Z + -1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1 +}; + +UnicodeString & +UTS46::process(const UnicodeString &src, + UBool isLabel, UBool toASCII, + UnicodeString &dest, + IDNAInfo &info, UErrorCode &errorCode) const { + // uts46Norm2.normalize() would do all of this error checking and setup, + // but with the ASCII fastpath we do not always call it, and do not + // call it first. + if(U_FAILURE(errorCode)) { + dest.setToBogus(); + return dest; + } + const char16_t *srcArray=src.getBuffer(); + if(&dest==&src || srcArray==nullptr) { + errorCode=U_ILLEGAL_ARGUMENT_ERROR; + dest.setToBogus(); + return dest; + } + // Arguments are fine, reset output values. + dest.remove(); + info.reset(); + int32_t srcLength=src.length(); + if(srcLength==0) { + info.errors|=UIDNA_ERROR_EMPTY_LABEL; + return dest; + } + char16_t *destArray=dest.getBuffer(srcLength); + if(destArray==nullptr) { + errorCode=U_MEMORY_ALLOCATION_ERROR; + return dest; + } + // ASCII fastpath + UBool disallowNonLDHDot=(options&UIDNA_USE_STD3_RULES)!=0; + int32_t labelStart=0; + int32_t i; + for(i=0;; ++i) { + if(i==srcLength) { + if(toASCII) { + if((i-labelStart)>63) { + info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; + } + // There is a trailing dot if labelStart==i. + if(!isLabel && i>=254 && (i>254 || labelStart0x7f) { + break; + } + int cData=asciiData[c]; + if(cData>0) { + destArray[i]=c+0x20; // Lowercase an uppercase ASCII letter. + } else if(cData<0 && disallowNonLDHDot) { + break; // Replacing with U+FFFD can be complicated for toASCII. + } else { + destArray[i]=c; + if(c==0x2d) { // hyphen + if(i==(labelStart+3) && srcArray[i-1]==0x2d) { + // "??--..." is Punycode or forbidden. + ++i; // '-' was copied to dest already + break; + } + if(i==labelStart) { + // label starts with "-" + info.labelErrors|=UIDNA_ERROR_LEADING_HYPHEN; + } + if((i+1)==srcLength || srcArray[i+1]==0x2e) { + // label ends with "-" + info.labelErrors|=UIDNA_ERROR_TRAILING_HYPHEN; + } + } else if(c==0x2e) { // dot + if(isLabel) { + // Replacing with U+FFFD can be complicated for toASCII. + ++i; // '.' was copied to dest already + break; + } + if(i==labelStart) { + info.labelErrors|=UIDNA_ERROR_EMPTY_LABEL; + } + if(toASCII && (i-labelStart)>63) { + info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; + } + info.errors|=info.labelErrors; + info.labelErrors=0; + labelStart=i+1; + } + } + } + info.errors|=info.labelErrors; + dest.releaseBuffer(i); + processUnicode(src, labelStart, i, isLabel, toASCII, dest, info, errorCode); + if( info.isBiDi && U_SUCCESS(errorCode) && (info.errors&severeErrors)==0 && + (!info.isOkBiDi || (labelStart>0 && !isASCIIOkBiDi(dest.getBuffer(), labelStart))) + ) { + info.errors|=UIDNA_ERROR_BIDI; + } + return dest; +} + +void +UTS46::processUTF8(StringPiece src, + UBool isLabel, UBool toASCII, + ByteSink &dest, + IDNAInfo &info, UErrorCode &errorCode) const { + if(U_FAILURE(errorCode)) { + return; + } + const char *srcArray=src.data(); + int32_t srcLength=src.length(); + if(srcArray==nullptr && srcLength!=0) { + errorCode=U_ILLEGAL_ARGUMENT_ERROR; + return; + } + // Arguments are fine, reset output values. + info.reset(); + if(srcLength==0) { + info.errors|=UIDNA_ERROR_EMPTY_LABEL; + dest.Flush(); + return; + } + UnicodeString destString; + int32_t labelStart=0; + if(srcLength<=256) { // length of stackArray[] + // ASCII fastpath + char stackArray[256]; + int32_t destCapacity; + char *destArray=dest.GetAppendBuffer(srcLength, srcLength+20, + stackArray, UPRV_LENGTHOF(stackArray), &destCapacity); + UBool disallowNonLDHDot=(options&UIDNA_USE_STD3_RULES)!=0; + int32_t i; + for(i=0;; ++i) { + if(i==srcLength) { + if(toASCII) { + if((i-labelStart)>63) { + info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; + } + // There is a trailing dot if labelStart==i. + if(!isLabel && i>=254 && (i>254 || labelStart0x7f + break; + } + int cData=asciiData[(int)c]; // Cast: gcc warns about indexing with a char. + if(cData>0) { + destArray[i]=c+0x20; // Lowercase an uppercase ASCII letter. + } else if(cData<0 && disallowNonLDHDot) { + break; // Replacing with U+FFFD can be complicated for toASCII. + } else { + destArray[i]=c; + if(c==0x2d) { // hyphen + if(i==(labelStart+3) && srcArray[i-1]==0x2d) { + // "??--..." is Punycode or forbidden. + break; + } + if(i==labelStart) { + // label starts with "-" + info.labelErrors|=UIDNA_ERROR_LEADING_HYPHEN; + } + if((i+1)==srcLength || srcArray[i+1]==0x2e) { + // label ends with "-" + info.labelErrors|=UIDNA_ERROR_TRAILING_HYPHEN; + } + } else if(c==0x2e) { // dot + if(isLabel) { + break; // Replacing with U+FFFD can be complicated for toASCII. + } + if(i==labelStart) { + info.labelErrors|=UIDNA_ERROR_EMPTY_LABEL; + } + if(toASCII && (i-labelStart)>63) { + info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; + } + info.errors|=info.labelErrors; + info.labelErrors=0; + labelStart=i+1; + } + } + } + info.errors|=info.labelErrors; + // Convert the processed ASCII prefix of the current label to UTF-16. + int32_t mappingStart=i-labelStart; + destString=UnicodeString::fromUTF8(StringPiece(destArray+labelStart, mappingStart)); + // Output the previous ASCII labels and process the rest of src in UTF-16. + dest.Append(destArray, labelStart); + processUnicode(UnicodeString::fromUTF8(StringPiece(src, labelStart)), 0, mappingStart, + isLabel, toASCII, + destString, info, errorCode); + } else { + // src is too long for the ASCII fastpath implementation. + processUnicode(UnicodeString::fromUTF8(src), 0, 0, + isLabel, toASCII, + destString, info, errorCode); + } + destString.toUTF8(dest); // calls dest.Flush() + if(toASCII && !isLabel) { + // length==labelStart==254 means that there is a trailing dot (ok) and + // destString is empty (do not index at 253-labelStart). + int32_t length=labelStart+destString.length(); + if( length>=254 && isASCIIString(destString) && + (length>254 || + (labelStart<254 && destString[253-labelStart]!=0x2e)) + ) { + info.errors|=UIDNA_ERROR_DOMAIN_NAME_TOO_LONG; + } + } + if( info.isBiDi && U_SUCCESS(errorCode) && (info.errors&severeErrors)==0 && + (!info.isOkBiDi || (labelStart>0 && !isASCIIOkBiDi(srcArray, labelStart))) + ) { + info.errors|=UIDNA_ERROR_BIDI; + } +} + +UnicodeString & +UTS46::processUnicode(const UnicodeString &src, + int32_t labelStart, int32_t mappingStart, + UBool isLabel, UBool toASCII, + UnicodeString &dest, + IDNAInfo &info, UErrorCode &errorCode) const { + if(mappingStart==0) { + uts46Norm2.normalize(src, dest, errorCode); + } else { + uts46Norm2.normalizeSecondAndAppend(dest, src.tempSubString(mappingStart), errorCode); + } + if(U_FAILURE(errorCode)) { + return dest; + } + UBool doMapDevChars= + toASCII ? (options&UIDNA_NONTRANSITIONAL_TO_ASCII)==0 : + (options&UIDNA_NONTRANSITIONAL_TO_UNICODE)==0; + const char16_t *destArray=dest.getBuffer(); + int32_t destLength=dest.length(); + int32_t labelLimit=labelStart; + while(labelLimit=0x200c)) { + info.isTransDiff=true; + if(doMapDevChars) { + destLength=mapDevChars(dest, labelStart, labelLimit, errorCode); + if(U_FAILURE(errorCode)) { + return dest; + } + destArray=dest.getBuffer(); + // All deviation characters have been mapped, no need to check for them again. + doMapDevChars=false; + // Do not increment labelLimit in case c was removed. + continue; + } + } else if(U16_IS_SURROGATE(c)) { + if(U16_IS_SURROGATE_LEAD(c) ? + (labelLimit+1)==destLength || !U16_IS_TRAIL(destArray[labelLimit+1]) : + labelLimit==labelStart || !U16_IS_LEAD(destArray[labelLimit-1])) { + // Map an unpaired surrogate to U+FFFD before normalization so that when + // that removes characters we do not turn two unpaired ones into a pair. + info.labelErrors|=UIDNA_ERROR_DISALLOWED; + dest.setCharAt(labelLimit, 0xfffd); + destArray=dest.getBuffer(); + } + } + ++labelLimit; + } + // Permit an empty label at the end (0=4 && label[0]==0x78 && label[1]==0x6e && label[2]==0x2d && label[3]==0x2d) { + // Label starts with "xn--", try to un-Punycode it. + // In IDNA2008, labels like "xn--" (decodes to an empty string) and + // "xn--ASCII-" (decodes to just "ASCII") fail the round-trip validation from + // comparing the ToUnicode input with the back-to-ToASCII output. + // They are alternate encodings of the respective ASCII labels. + // Ignore "xn---" here: It will fail Punycode.decode() which logically comes before + // the round-trip verification. + if(labelLength==4 || (labelLength>5 && label[labelLength-1]==u'-')) { + info.labelErrors|=UIDNA_ERROR_INVALID_ACE_LABEL; + return markBadACELabel(dest, labelStart, labelLength, toASCII, info, errorCode); + } + wasPunycode=true; + char16_t *unicodeBuffer=fromPunycode.getBuffer(-1); // capacity==-1: most labels should fit + if(unicodeBuffer==nullptr) { + // Should never occur if we used capacity==-1 which uses the internal buffer. + errorCode=U_MEMORY_ALLOCATION_ERROR; + return labelLength; + } + UErrorCode punycodeErrorCode=U_ZERO_ERROR; + int32_t unicodeLength=u_strFromPunycode(label+4, labelLength-4, + unicodeBuffer, fromPunycode.getCapacity(), + nullptr, &punycodeErrorCode); + if(punycodeErrorCode==U_BUFFER_OVERFLOW_ERROR) { + fromPunycode.releaseBuffer(0); + unicodeBuffer=fromPunycode.getBuffer(unicodeLength); + if(unicodeBuffer==nullptr) { + errorCode=U_MEMORY_ALLOCATION_ERROR; + return labelLength; + } + punycodeErrorCode=U_ZERO_ERROR; + unicodeLength=u_strFromPunycode(label+4, labelLength-4, + unicodeBuffer, fromPunycode.getCapacity(), + nullptr, &punycodeErrorCode); + } + fromPunycode.releaseBuffer(unicodeLength); + if(U_FAILURE(punycodeErrorCode)) { + info.labelErrors|=UIDNA_ERROR_PUNYCODE; + return markBadACELabel(dest, labelStart, labelLength, toASCII, info, errorCode); + } + // Check for NFC, and for characters that are not + // valid or deviation characters according to the normalizer. + // If there is something wrong, then the string will change. + // Note that the normalizer passes through non-LDH ASCII and deviation characters. + // Deviation characters are ok in Punycode even in transitional processing. + // In the code further below, if we find non-LDH ASCII and we have UIDNA_USE_STD3_RULES + // then we will set UIDNA_ERROR_INVALID_ACE_LABEL there too. + UBool isValid=uts46Norm2.isNormalized(fromPunycode, errorCode); + if(U_FAILURE(errorCode)) { + return labelLength; + } + if(!isValid) { + info.labelErrors|=UIDNA_ERROR_INVALID_ACE_LABEL; + return markBadACELabel(dest, labelStart, labelLength, toASCII, info, errorCode); + } + labelString=&fromPunycode; + label=fromPunycode.getBuffer(); + labelStart=0; + labelLength=fromPunycode.length(); + } else { + wasPunycode=false; + labelString=&dest; + } + // Validity check + if(labelLength==0) { + info.labelErrors|=UIDNA_ERROR_EMPTY_LABEL; + return replaceLabel(dest, destLabelStart, destLabelLength, + *labelString, labelLength, errorCode); + } + // labelLength>0 + if(labelLength>=4 && label[2]==0x2d && label[3]==0x2d) { + // label starts with "??--" + info.labelErrors|=UIDNA_ERROR_HYPHEN_3_4; + } + if(label[0]==0x2d) { + // label starts with "-" + info.labelErrors|=UIDNA_ERROR_LEADING_HYPHEN; + } + if(label[labelLength-1]==0x2d) { + // label ends with "-" + info.labelErrors|=UIDNA_ERROR_TRAILING_HYPHEN; + } + // If the label was not a Punycode label, then it was the result of + // mapping, normalization and label segmentation. + // If the label was in Punycode, then we mapped it again above + // and checked its validity. + // Now we handle the STD3 restriction to LDH characters (if set) + // and we look for U+FFFD which indicates disallowed characters + // in a non-Punycode label or U+FFFD itself in a Punycode label. + // We also check for dots which can come from the input to a single-label function. + // Ok to cast away const because we own the UnicodeString. + char16_t *s=(char16_t *)label; + const char16_t *limit=label+labelLength; + char16_t oredChars=0; + // If we enforce STD3 rules, then ASCII characters other than LDH and dot are disallowed. + UBool disallowNonLDHDot=(options&UIDNA_USE_STD3_RULES)!=0; + do { + char16_t c=*s; + if(c<=0x7f) { + if(c==0x2e) { + info.labelErrors|=UIDNA_ERROR_LABEL_HAS_DOT; + *s=0xfffd; + } else if(disallowNonLDHDot && asciiData[c]<0) { + info.labelErrors|=UIDNA_ERROR_DISALLOWED; + *s=0xfffd; + } + } else { + oredChars|=c; + if(disallowNonLDHDot && isNonASCIIDisallowedSTD3Valid(c)) { + info.labelErrors|=UIDNA_ERROR_DISALLOWED; + *s=0xfffd; + } else if(c==0xfffd) { + info.labelErrors|=UIDNA_ERROR_DISALLOWED; + } + } + ++s; + } while(sreplace(labelStart, cpLength, (char16_t)0xfffd); + label=labelString->getBuffer()+labelStart; + labelLength+=1-cpLength; + if(labelString==&dest) { + destLabelLength=labelLength; + } + } + if((info.labelErrors&severeErrors)==0) { + // Do contextual checks only if we do not have U+FFFD from a severe error + // because U+FFFD can make these checks fail. + if((options&UIDNA_CHECK_BIDI)!=0 && (!info.isBiDi || info.isOkBiDi)) { + checkLabelBiDi(label, labelLength, info); + } + if( (options&UIDNA_CHECK_CONTEXTJ)!=0 && (oredChars&0x200c)==0x200c && + !isLabelOkContextJ(label, labelLength) + ) { + info.labelErrors|=UIDNA_ERROR_CONTEXTJ; + } + if((options&UIDNA_CHECK_CONTEXTO)!=0 && oredChars>=0xb7) { + checkLabelContextO(label, labelLength, info); + } + if(toASCII) { + if(wasPunycode) { + // Leave a Punycode label unchanged if it has no severe errors. + if(destLabelLength>63) { + info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; + } + return destLabelLength; + } else if(oredChars>=0x80) { + // Contains non-ASCII characters. + UnicodeString punycode; + char16_t *buffer=punycode.getBuffer(63); // 63==maximum DNS label length + if(buffer==nullptr) { + errorCode=U_MEMORY_ALLOCATION_ERROR; + return destLabelLength; + } + buffer[0]=0x78; // Write "xn--". + buffer[1]=0x6e; + buffer[2]=0x2d; + buffer[3]=0x2d; + int32_t punycodeLength=u_strToPunycode(label, labelLength, + buffer+4, punycode.getCapacity()-4, + nullptr, &errorCode); + if(errorCode==U_BUFFER_OVERFLOW_ERROR) { + errorCode=U_ZERO_ERROR; + punycode.releaseBuffer(4); + buffer=punycode.getBuffer(4+punycodeLength); + if(buffer==nullptr) { + errorCode=U_MEMORY_ALLOCATION_ERROR; + return destLabelLength; + } + punycodeLength=u_strToPunycode(label, labelLength, + buffer+4, punycode.getCapacity()-4, + nullptr, &errorCode); + } + punycodeLength+=4; + punycode.releaseBuffer(punycodeLength); + if(U_FAILURE(errorCode)) { + return destLabelLength; + } + if(punycodeLength>63) { + info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; + } + return replaceLabel(dest, destLabelStart, destLabelLength, + punycode, punycodeLength, errorCode); + } else { + // all-ASCII label + if(labelLength>63) { + info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; + } + } + } + } else { + // If a Punycode label has severe errors, + // then leave it but make sure it does not look valid. + if(wasPunycode) { + info.labelErrors|=UIDNA_ERROR_INVALID_ACE_LABEL; + return markBadACELabel(dest, destLabelStart, destLabelLength, toASCII, info, errorCode); + } + } + return replaceLabel(dest, destLabelStart, destLabelLength, + *labelString, labelLength, errorCode); +} + +// Make sure an ACE label does not look valid. +// Append U+FFFD if the label has only LDH characters. +// If UIDNA_USE_STD3_RULES, also replace disallowed ASCII characters with U+FFFD. +int32_t +UTS46::markBadACELabel(UnicodeString &dest, + int32_t labelStart, int32_t labelLength, + UBool toASCII, IDNAInfo &info, UErrorCode &errorCode) const { + if(U_FAILURE(errorCode)) { + return 0; + } + UBool disallowNonLDHDot=(options&UIDNA_USE_STD3_RULES)!=0; + UBool isASCII=true; + UBool onlyLDH=true; + const char16_t *label=dest.getBuffer()+labelStart; + const char16_t *limit=label+labelLength; + // Start after the initial "xn--". + // Ok to cast away const because we own the UnicodeString. + for(char16_t *s=const_cast(label+4); s63) { + info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; + } + } + return labelLength; +} + +const uint32_t L_MASK=U_MASK(U_LEFT_TO_RIGHT); +const uint32_t R_AL_MASK=U_MASK(U_RIGHT_TO_LEFT)|U_MASK(U_RIGHT_TO_LEFT_ARABIC); +const uint32_t L_R_AL_MASK=L_MASK|R_AL_MASK; + +const uint32_t R_AL_AN_MASK=R_AL_MASK|U_MASK(U_ARABIC_NUMBER); + +const uint32_t EN_AN_MASK=U_MASK(U_EUROPEAN_NUMBER)|U_MASK(U_ARABIC_NUMBER); +const uint32_t R_AL_EN_AN_MASK=R_AL_MASK|EN_AN_MASK; +const uint32_t L_EN_MASK=L_MASK|U_MASK(U_EUROPEAN_NUMBER); + +const uint32_t ES_CS_ET_ON_BN_NSM_MASK= + U_MASK(U_EUROPEAN_NUMBER_SEPARATOR)| + U_MASK(U_COMMON_NUMBER_SEPARATOR)| + U_MASK(U_EUROPEAN_NUMBER_TERMINATOR)| + U_MASK(U_OTHER_NEUTRAL)| + U_MASK(U_BOUNDARY_NEUTRAL)| + U_MASK(U_DIR_NON_SPACING_MARK); +const uint32_t L_EN_ES_CS_ET_ON_BN_NSM_MASK=L_EN_MASK|ES_CS_ET_ON_BN_NSM_MASK; +const uint32_t R_AL_AN_EN_ES_CS_ET_ON_BN_NSM_MASK=R_AL_MASK|EN_AN_MASK|ES_CS_ET_ON_BN_NSM_MASK; + +// We scan the whole label and check both for whether it contains RTL characters +// and whether it passes the BiDi Rule. +// In a BiDi domain name, all labels must pass the BiDi Rule, but we might find +// that a domain name is a BiDi domain name (has an RTL label) only after +// processing several earlier labels. +void +UTS46::checkLabelBiDi(const char16_t *label, int32_t labelLength, IDNAInfo &info) const { + // IDNA2008 BiDi rule + // Get the directionality of the first character. + UChar32 c; + int32_t i=0; + U16_NEXT_UNSAFE(label, i, c); + uint32_t firstMask=U_MASK(u_charDirection(c)); + // 1. The first character must be a character with BIDI property L, R + // or AL. If it has the R or AL property, it is an RTL label; if it + // has the L property, it is an LTR label. + if((firstMask&~L_R_AL_MASK)!=0) { + info.isOkBiDi=false; + } + // Get the directionality of the last non-NSM character. + uint32_t lastMask; + for(;;) { + if(i>=labelLength) { + lastMask=firstMask; + break; + } + U16_PREV_UNSAFE(label, labelLength, c); + UCharDirection dir=u_charDirection(c); + if(dir!=U_DIR_NON_SPACING_MARK) { + lastMask=U_MASK(dir); + break; + } + } + // 3. In an RTL label, the end of the label must be a character with + // BIDI property R, AL, EN or AN, followed by zero or more + // characters with BIDI property NSM. + // 6. In an LTR label, the end of the label must be a character with + // BIDI property L or EN, followed by zero or more characters with + // BIDI property NSM. + if( (firstMask&L_MASK)!=0 ? + (lastMask&~L_EN_MASK)!=0 : + (lastMask&~R_AL_EN_AN_MASK)!=0 + ) { + info.isOkBiDi=false; + } + // Add the directionalities of the intervening characters. + uint32_t mask=firstMask|lastMask; + while(ilabelStart) { + c=s[i-1]; + if(!(0x61<=c && c<=0x7a) && !(0x30<=c && c<=0x39)) { + // Last character in the label is not an L or EN. + return false; + } + } + labelStart=i+1; + } else if(i==labelStart) { + if(!(0x61<=c && c<=0x7a)) { + // First character in the label is not an L. + return false; + } + } else { + if(c<=0x20 && (c>=0x1c || (9<=c && c<=0xd))) { + // Intermediate character in the label is a B, S or WS. + return false; + } + } + } + return true; +} + +// UTF-8 version, called for source ASCII prefix. +// Can contain uppercase A-Z. +// s[length-1] must be the trailing dot. +static UBool +isASCIIOkBiDi(const char *s, int32_t length) { + int32_t labelStart=0; + for(int32_t i=0; ilabelStart) { + c=s[i-1]; + if(!(0x61<=c && c<=0x7a) && !(0x41<=c && c<=0x5a) && !(0x30<=c && c<=0x39)) { + // Last character in the label is not an L or EN. + return false; + } + } + labelStart=i+1; + } else if(i==labelStart) { + if(!(0x61<=c && c<=0x7a) && !(0x41<=c && c<=0x5a)) { + // First character in the label is not an L. + return false; + } + } else { + if(c<=0x20 && (c>=0x1c || (9<=c && c<=0xd))) { + // Intermediate character in the label is a B, S or WS. + return false; + } + } + } + return true; +} + +UBool +UTS46::isLabelOkContextJ(const char16_t *label, int32_t labelLength) const { + // [IDNA2008-Tables] + // 200C..200D ; CONTEXTJ # ZERO WIDTH NON-JOINER..ZERO WIDTH JOINER + for(int32_t i=0; i0) { + info.labelErrors|=UIDNA_ERROR_CONTEXTO_DIGITS; + } + arabicDigits=-1; + } else if(0x6f0<=c) { + if(arabicDigits<0) { + info.labelErrors|=UIDNA_ERROR_CONTEXTO_DIGITS; + } + arabicDigits=1; + } + } + } else if(c==0x30fb) { + // Appendix A.7. KATAKANA MIDDLE DOT (U+30FB) + // Rule Set: + // False; + // For All Characters: + // If Script(cp) .in. {Hiragana, Katakana, Han} Then True; + // End For; + UErrorCode errorCode=U_ZERO_ERROR; + for(int j=0;;) { + if(j>labelEnd) { + info.labelErrors|=UIDNA_ERROR_CONTEXTO_PUNCTUATION; + break; + } + U16_NEXT(label, j, labelLength, c); + UScriptCode script=uscript_getScript(c, &errorCode); + if(script==USCRIPT_HIRAGANA || script==USCRIPT_KATAKANA || script==USCRIPT_HAN) { + break; + } + } + } + } +} + +U_NAMESPACE_END + +// C API ------------------------------------------------------------------- *** + +U_NAMESPACE_USE + +U_CAPI UIDNA * U_EXPORT2 +uidna_openUTS46(uint32_t options, UErrorCode *pErrorCode) { + return reinterpret_cast(IDNA::createUTS46Instance(options, *pErrorCode)); +} + +U_CAPI void U_EXPORT2 +uidna_close(UIDNA *idna) { + delete reinterpret_cast(idna); +} + +static UBool +checkArgs(const void *label, int32_t length, + void *dest, int32_t capacity, + UIDNAInfo *pInfo, UErrorCode *pErrorCode) { + if(U_FAILURE(*pErrorCode)) { + return false; + } + // sizeof(UIDNAInfo)=16 in the first API version. + if(pInfo==nullptr || pInfo->size<16) { + *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; + return false; + } + if( (label==nullptr ? length!=0 : length<-1) || + (dest==nullptr ? capacity!=0 : capacity<0) || + (dest==label && label!=nullptr) + ) { + *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; + return false; + } + // Set all *pInfo bytes to 0 except for the size field itself. + uprv_memset(&pInfo->size+1, 0, pInfo->size-sizeof(pInfo->size)); + return true; +} + +static void +idnaInfoToStruct(IDNAInfo &info, UIDNAInfo *pInfo) { + pInfo->isTransitionalDifferent=info.isTransitionalDifferent(); + pInfo->errors=info.getErrors(); +} + +U_CAPI int32_t U_EXPORT2 +uidna_labelToASCII(const UIDNA *idna, + const char16_t *label, int32_t length, + char16_t *dest, int32_t capacity, + UIDNAInfo *pInfo, UErrorCode *pErrorCode) { + if(!checkArgs(label, length, dest, capacity, pInfo, pErrorCode)) { + return 0; + } + UnicodeString src((UBool)(length<0), label, length); + UnicodeString destString(dest, 0, capacity); + IDNAInfo info; + reinterpret_cast(idna)->labelToASCII(src, destString, info, *pErrorCode); + idnaInfoToStruct(info, pInfo); + return destString.extract(dest, capacity, *pErrorCode); +} + +U_CAPI int32_t U_EXPORT2 +uidna_labelToUnicode(const UIDNA *idna, + const char16_t *label, int32_t length, + char16_t *dest, int32_t capacity, + UIDNAInfo *pInfo, UErrorCode *pErrorCode) { + if(!checkArgs(label, length, dest, capacity, pInfo, pErrorCode)) { + return 0; + } + UnicodeString src((UBool)(length<0), label, length); + UnicodeString destString(dest, 0, capacity); + IDNAInfo info; + reinterpret_cast(idna)->labelToUnicode(src, destString, info, *pErrorCode); + idnaInfoToStruct(info, pInfo); + return destString.extract(dest, capacity, *pErrorCode); +} + +U_CAPI int32_t U_EXPORT2 +uidna_nameToASCII(const UIDNA *idna, + const char16_t *name, int32_t length, + char16_t *dest, int32_t capacity, + UIDNAInfo *pInfo, UErrorCode *pErrorCode) { + if(!checkArgs(name, length, dest, capacity, pInfo, pErrorCode)) { + return 0; + } + UnicodeString src((UBool)(length<0), name, length); + UnicodeString destString(dest, 0, capacity); + IDNAInfo info; + reinterpret_cast(idna)->nameToASCII(src, destString, info, *pErrorCode); + idnaInfoToStruct(info, pInfo); + return destString.extract(dest, capacity, *pErrorCode); +} + +U_CAPI int32_t U_EXPORT2 +uidna_nameToUnicode(const UIDNA *idna, + const char16_t *name, int32_t length, + char16_t *dest, int32_t capacity, + UIDNAInfo *pInfo, UErrorCode *pErrorCode) { + if(!checkArgs(name, length, dest, capacity, pInfo, pErrorCode)) { + return 0; + } + UnicodeString src((UBool)(length<0), name, length); + UnicodeString destString(dest, 0, capacity); + IDNAInfo info; + reinterpret_cast(idna)->nameToUnicode(src, destString, info, *pErrorCode); + idnaInfoToStruct(info, pInfo); + return destString.extract(dest, capacity, *pErrorCode); +} + +U_CAPI int32_t U_EXPORT2 +uidna_labelToASCII_UTF8(const UIDNA *idna, + const char *label, int32_t length, + char *dest, int32_t capacity, + UIDNAInfo *pInfo, UErrorCode *pErrorCode) { + if(!checkArgs(label, length, dest, capacity, pInfo, pErrorCode)) { + return 0; + } + StringPiece src(label, length<0 ? static_cast(uprv_strlen(label)) : length); + CheckedArrayByteSink sink(dest, capacity); + IDNAInfo info; + reinterpret_cast(idna)->labelToASCII_UTF8(src, sink, info, *pErrorCode); + idnaInfoToStruct(info, pInfo); + return u_terminateChars(dest, capacity, sink.NumberOfBytesAppended(), pErrorCode); +} + +U_CAPI int32_t U_EXPORT2 +uidna_labelToUnicodeUTF8(const UIDNA *idna, + const char *label, int32_t length, + char *dest, int32_t capacity, + UIDNAInfo *pInfo, UErrorCode *pErrorCode) { + if(!checkArgs(label, length, dest, capacity, pInfo, pErrorCode)) { + return 0; + } + StringPiece src(label, length<0 ? static_cast(uprv_strlen(label)) : length); + CheckedArrayByteSink sink(dest, capacity); + IDNAInfo info; + reinterpret_cast(idna)->labelToUnicodeUTF8(src, sink, info, *pErrorCode); + idnaInfoToStruct(info, pInfo); + return u_terminateChars(dest, capacity, sink.NumberOfBytesAppended(), pErrorCode); +} + +U_CAPI int32_t U_EXPORT2 +uidna_nameToASCII_UTF8(const UIDNA *idna, + const char *name, int32_t length, + char *dest, int32_t capacity, + UIDNAInfo *pInfo, UErrorCode *pErrorCode) { + if(!checkArgs(name, length, dest, capacity, pInfo, pErrorCode)) { + return 0; + } + StringPiece src(name, length<0 ? static_cast(uprv_strlen(name)) : length); + CheckedArrayByteSink sink(dest, capacity); + IDNAInfo info; + reinterpret_cast(idna)->nameToASCII_UTF8(src, sink, info, *pErrorCode); + idnaInfoToStruct(info, pInfo); + return u_terminateChars(dest, capacity, sink.NumberOfBytesAppended(), pErrorCode); +} + +U_CAPI int32_t U_EXPORT2 +uidna_nameToUnicodeUTF8(const UIDNA *idna, + const char *name, int32_t length, + char *dest, int32_t capacity, + UIDNAInfo *pInfo, UErrorCode *pErrorCode) { + if(!checkArgs(name, length, dest, capacity, pInfo, pErrorCode)) { + return 0; + } + StringPiece src(name, length<0 ? static_cast(uprv_strlen(name)) : length); + CheckedArrayByteSink sink(dest, capacity); + IDNAInfo info; + reinterpret_cast(idna)->nameToUnicodeUTF8(src, sink, info, *pErrorCode); + idnaInfoToStruct(info, pInfo); + return u_terminateChars(dest, capacity, sink.NumberOfBytesAppended(), pErrorCode); +} + +#endif // UCONFIG_NO_IDNA -- cgit v1.2.3