1 files changed, 515 insertions, 0 deletions
diff --git a/src/parser/parser_global.cpp b/src/parser/parser_global.cpp
new file mode 100644
index 0000000..5489e21
--- /dev/null
+++ b/src/parser/parser_global.cpp
@@ -0,0 +1,515 @@
+/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
+/*
+ * This Source Code Form is subject to the terms of the Mozilla Public
+ * License, v. 2.0. If a copy of the MPL was not distributed with this
+ * file, You can obtain one at http://mozilla.org/MPL/2.0/.
+ */
+
+#include <orcus/parser_global.hpp>
+#include <orcus/cell_buffer.hpp>
+#include <orcus/exception.hpp>
+
+#include "numeric_parser.hpp"
+
+#include <cassert>
+#include <cmath>
+#include <iostream>
+#include <limits>
+#include <algorithm>
+#include <cctype>
+
+namespace orcus {
+
+const size_t parse_quoted_string_state::error_no_closing_quote = 1;
+const size_t parse_quoted_string_state::error_illegal_escape_char = 2;
+
+bool is_blank(char c)
+{
+    return is_in(c, " \t\n\r");
+}
+
+bool is_alpha(char c)
+{
+    return std::isalpha(static_cast<unsigned char>(c));
+}
+
+bool is_numeric(char c)
+{
+    return std::isdigit(static_cast<unsigned char>(c));
+}
+
+bool is_in(char c, std::string_view allowed)
+{
+#ifdef __ORCUS_DEBUG_UTILS
+    if (allowed.empty())
+        throw std::invalid_argument("'allowed' string should not be empty.");
+#endif
+    auto f = [c](char c_allowed) { return c == c_allowed; };
+    return std::any_of(allowed.begin(), allowed.end(), f);
+}
+
+const char* parse_numeric(const char* p, const char* p_end, double& value)
+{
+    using numeric_parser_type = detail::numeric_parser<detail::generic_parser_trait>;
+
+    numeric_parser_type parser(p, p_end);
+    double v = parser.parse();
+    if (!std::isnan(v))
+        p = parser.get_char_position();
+
+    value = v;
+    return p;
+}
+
+const char* parse_integer(const char* p, const char* p_end, long& value)
+{
+    if (p >= p_end)
+        return p;
+
+    long result = 0.0;
+    bool negative_sign = false;
+
+    // Check for presence of a sign.
+    if (p != p_end)
+    {
+        switch (*p)
+        {
+            case '+':
+                ++p;
+                break;
+            case '-':
+                negative_sign = true;
+                ++p;
+                break;
+            default:
+                ;
+        }
+    }
+
+    for (; p != p_end; ++p)
+    {
+        if (*p < '0' || '9' < *p)
+        {
+            value = negative_sign ? -result : result;
+            return p;
+        }
+
+        result *= 10;
+        result += *p - '0';
+    }
+
+    value = negative_sign ? -result : result;
+    return p;
+}
+
+string_escape_char_t get_string_escape_char_type(char c)
+{
+    switch (c)
+    {
+        case '"':
+        case '\\':
+        case '/':
+            return string_escape_char_t::valid;
+        case 'b': // backspace
+        case 'f': // formfeed
+        case 'n': // newline
+        case 'r': // carriage return
+        case 't': // horizontal tab
+            return string_escape_char_t::control_char;
+        default:
+            ;
+    }
+
+    return string_escape_char_t::invalid;
+}
+
+namespace {
+
+parse_quoted_string_state parse_string_with_escaped_char(
+    const char*& p, size_t max_length, const char* p_parsed, size_t n_parsed, char c,
+    cell_buffer& buffer)
+{
+    const char* p_end = p + max_length;
+
+    parse_quoted_string_state ret;
+    ret.str = nullptr;
+    ret.length = 0;
+    ret.transient = true;
+    ret.has_control_character = false;
+
+    // Start the buffer with the string we've parsed so far.
+    buffer.reset();
+    if (p_parsed && n_parsed)
+        buffer.append(p_parsed, n_parsed);
+    buffer.append(&c, 1);
+
+    ++p;
+    if (p == p_end)
+    {
+        ret.length = parse_quoted_string_state::error_no_closing_quote;
+        return ret;
+    }
+
+    size_t len = 0;
+    const char* p_head = p;
+    bool escape = false;
+
+    for (; p != p_end; ++p, ++len)
+    {
+        c = *p;
+
+        if (escape)
+        {
+            escape = false;
+
+            switch (get_string_escape_char_type(c))
+            {
+                case string_escape_char_t::valid:
+                    buffer.append(p_head, len-1);
+                    buffer.append(&c, 1);
+                    ++p;
+                    len = 0;
+                    p_head = p;
+                break;
+                case string_escape_char_t::control_char:
+                    // do nothing on control characters.
+                break;
+                case string_escape_char_t::invalid:
+                default:
+                    ret.length = parse_quoted_string_state::error_illegal_escape_char;
+                    return ret;
+            }
+        }
+
+        switch (*p)
+        {
+            case '"':
+            {
+                // closing quote.
+                buffer.append(p_head, len);
+                ++p; // skip the quote.
+                std::string_view s = buffer.str();
+                ret.str = s.data();
+                ret.length = s.size();
+                return ret;
+            }
+            case '\\':
+            {
+                escape = true;
+                continue;
+            }
+            default:
+                ;
+        }
+    }
+
+    ret.length = parse_quoted_string_state::error_no_closing_quote;
+    return ret;
+}
+
+parse_quoted_string_state parse_single_quoted_string_buffered(
+    const char*& p, const char* p_end, cell_buffer& buffer)
+{
+    const char* p0 = p;
+    size_t len = 0;
+    char last = 0;
+
+    parse_quoted_string_state ret;
+    ret.transient = true;
+    ret.has_control_character = false;
+
+    for (; p != p_end; ++p)
+    {
+        if (!p0)
+            p0 = p;
+
+        char c = *p;
+        switch (c)
+        {
+            case '\'':
+            {
+                if (last == c)
+                {
+                    // Second "'" in series.  This is an encoded single quote.
+                    buffer.append(p0, len);
+                    p0 = nullptr;
+                    last = 0;
+                    len = 0;
+                    continue;
+                }
+            }
+            break;
+            default:
+            {
+                if (last == '\'')
+                {
+                    buffer.append(p0, len-1);
+                    auto s = buffer.str();
+                    ret.str = s.data();
+                    ret.length = s.size();
+                    return ret;
+                }
+            }
+        }
+
+        last = c;
+        ++len;
+    }
+
+    if (last == '\'')
+    {
+        buffer.append(p0, len-1);
+        auto s = buffer.str();
+        ret.str = s.data();
+        ret.length = s.size();
+        return ret;
+    }
+
+    ret.str = nullptr;
+    ret.length = parse_quoted_string_state::error_no_closing_quote;
+    return ret;
+}
+
+}
+
+parse_quoted_string_state parse_single_quoted_string(
+    const char*& p, size_t max_length, cell_buffer& buffer)
+{
+    assert(*p == '\'');
+    const char* p_end = p + max_length;
+    ++p;
+
+    parse_quoted_string_state ret;
+    ret.str = p;
+    ret.length = 0;
+    ret.transient = false;
+    ret.has_control_character = false;
+
+    if (p == p_end)
+    {
+        ret.str = nullptr;
+        ret.length = parse_quoted_string_state::error_no_closing_quote;
+        return ret;
+    }
+
+    char last = 0;
+    char c = 0;
+    for (; p != p_end; last = c, ++p, ++ret.length)
+    {
+        c = *p;
+        switch (c)
+        {
+            case '\'':
+            {
+                if (last == c)
+                {
+                    // Encoded single quote.
+                    buffer.reset();
+                    buffer.append(ret.str, ret.length);
+                    ++p;
+                    return parse_single_quoted_string_buffered(p, p_end, buffer);
+                }
+            }
+            break;
+            default:
+            {
+                if (last == '\'')
+                {
+                    --ret.length;
+                    return ret;
+                }
+            }
+
+        }
+    }
+
+    if (last == '\'')
+    {
+        --ret.length;
+        return ret;
+    }
+
+    ret.str = nullptr;
+    ret.length = parse_quoted_string_state::error_no_closing_quote;
+    return ret;
+}
+
+const char* parse_to_closing_single_quote(const char* p, size_t max_length)
+{
+    assert(*p == '\'');
+    const char* p_end = p + max_length;
+    ++p;
+
+    if (p == p_end)
+        return nullptr;
+
+    char last = 0;
+    for (; p != p_end; ++p)
+    {
+        char c = *p;
+        switch (c)
+        {
+            case '\'':
+                if (last == '\'')
+                {
+                    last = 0;
+                    continue;
+                }
+            break;
+            default:
+            {
+                if (last == '\'')
+                    return p;
+            }
+        }
+
+        last = c;
+    }
+
+    if (last == '\'')
+        return p;
+
+    return nullptr;
+}
+
+parse_quoted_string_state parse_double_quoted_string(
+    const char*& p, size_t max_length, cell_buffer& buffer)
+{
+    if (max_length == 0 || !p || *p != '"')
+        throw invalid_arg_error("parse_double_quoted_string: invalid input string");
+
+    parse_quoted_string_state ret;
+    ret.str = nullptr;
+    ret.length = 0;
+    ret.transient = false;
+    ret.has_control_character = false;
+
+    const char* p_end = p + max_length;
+    ++p; // skip the opening quote.
+
+    ret.str = p;
+
+    if (p == p_end)
+    {
+        // The string contains only the opening quote.
+        ret.str = nullptr;
+        ret.length = parse_quoted_string_state::error_no_closing_quote;
+        return ret;
+    }
+
+    bool escape = false;
+
+    for (; p != p_end; ++p, ++ret.length)
+    {
+        char c = *p;
+
+        if (escape)
+        {
+            escape = false;
+
+            switch (get_string_escape_char_type(c))
+            {
+                case string_escape_char_t::valid:
+                    return parse_string_with_escaped_char(p, max_length, ret.str, ret.length-1, c, buffer);
+                case string_escape_char_t::control_char:
+                    // do nothing on control characters.
+                break;
+                case string_escape_char_t::invalid:
+                default:
+                    ret.str = nullptr;
+                    ret.length = parse_quoted_string_state::error_illegal_escape_char;
+                    return ret;
+            }
+        }
+
+        switch (*p)
+        {
+            case '"':
+            {
+                // closing quote.
+                ++p; // skip the quote.
+                return ret;
+            }
+            case '\\':
+            {
+                escape = true;
+                continue;
+            }
+            default:
+                ;
+        }
+
+        if (0x00 <= c && c <= 0x1F)
+        {
+            // This is an unescaped control character.
+            ret.has_control_character = true;
+        }
+    }
+
+    ret.str = nullptr;
+    ret.length = parse_quoted_string_state::error_no_closing_quote;
+    return ret;
+}
+
+const char* parse_to_closing_double_quote(const char* p, size_t max_length)
+{
+    assert(*p == '"');
+    const char* p_end = p + max_length;
+    ++p;
+
+    if (p == p_end)
+        return nullptr;
+
+    bool escape = false;
+
+    for (; p != p_end; ++p)
+    {
+        if (escape)
+        {
+            char c = *p;
+            escape = false;
+
+            if (get_string_escape_char_type(c) == string_escape_char_t::invalid)
+                return nullptr;
+        }
+
+        switch (*p)
+        {
+            case '"':
+                // closing quote.
+                ++p; // skip the quote.
+                return p;
+            case '\\':
+                escape = true;
+            break;
+            default:
+                ;
+        }
+    }
+
+    return nullptr;
+}
+
+std::string_view trim(std::string_view str)
+{
+    const char* p = str.data();
+    const char* p_end = p + str.size();
+
+    // Find the first non-space character.
+    p = std::find_if_not(p, p_end, is_blank);
+
+    if (p == p_end)
+    {
+        // This string is empty.
+        return std::string_view{};
+    }
+
+    // Find the last non-space character.
+    auto last = std::find_if_not(std::reverse_iterator(p_end), std::reverse_iterator(p), is_blank);
+    std::size_t n = std::distance(p, last.base());
+
+    return std::string_view{p, n};
+}
+
+}
+/* vim:set shiftwidth=4 softtabstop=4 expandtab: */