boostorg/regex
1
0
Fork 1
mirror of https://github.com/boostorg/regex.git synced 2025-07-22 16:47:17 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'develop' into git_issue_80

Browse Source 
Fixed Conflicts:
	include/boost/regex/v4/basic_regex_creator.hpp
	include/boost/regex/v4/basic_regex_parser.hpp
This commit is contained in:
jzmaddock
2020-01-21 12:57:09 +00:00
parent f64c22870f 70c3ffa89a
commit bc160a58f0
9 changed files with 102 additions and 104 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
include/boost/regex/v4/basic_regex.hpp
View File

@ -70,13 +70,14 @@ void bubble_down_one(I first, I last)
}
}
static const int hash_value_mask = 1 << (std::numeric_limits<int>::digits - 1);
template <class Iterator>
inline int hash_value_from_capture_name(Iterator i, Iterator j)
{
std::size_t r = boost::hash_range(i, j);
r %= ((std::numeric_limits<int>::max)() - 10001);
r += 10000;
return static_cast<int>(r);
r %= ((std::numeric_limits<int>::max)());
return static_cast<int>(r) | hash_value_mask;
}
class named_subexpressions

10
include/boost/regex/v4/basic_regex_creator.hpp
View File

@ -20,6 +20,8 @@
#ifndef BOOST_REGEX_V4_BASIC_REGEX_CREATOR_HPP
#define BOOST_REGEX_V4_BASIC_REGEX_CREATOR_HPP
#include <boost/regex/v4/indexed_bit_flag.hpp>
#ifdef BOOST_MSVC
#pragma warning(push)
#pragma warning(disable: 4103)
@ -239,7 +241,7 @@ protected:
bool m_icase; // true for case insensitive matches
unsigned m_repeater_id; // the state_id of the next repeater
bool m_has_backrefs; // true if there are actually any backrefs
unsigned m_backrefs; // bitmask of permitted backrefs
indexed_bit_flag m_backrefs; // bitmask of permitted backrefs
boost::uintmax_t m_bad_repeats; // bitmask of repeats we can't deduce a startmap for;
bool m_has_recursions; // set when we have recursive expresisons to fixup
std::vector<unsigned char> m_recursion_checks; // notes which recursions we've followed while analysing this expression
@ -268,7 +270,7 @@ private:
template <class charT, class traits>
basic_regex_creator<charT, traits>::basic_regex_creator(regex_data<charT, traits>* data)
: m_pdata(data), m_traits(*(data->m_ptraits)), m_last_state(0), m_icase(false), m_repeater_id(0),
m_has_backrefs(false), m_backrefs(0), m_bad_repeats(0), m_has_recursions(false), m_word_mask(0), m_mask_space(0), m_lower_mask(0), m_upper_mask(0), m_alpha_mask(0)
m_has_backrefs(false), m_bad_repeats(0), m_has_recursions(false), m_word_mask(0), m_mask_space(0), m_lower_mask(0), m_upper_mask(0), m_alpha_mask(0)
{
m_pdata->m_data.clear();
m_pdata->m_status = ::boost::regex_constants::error_ok;
@ -764,7 +766,7 @@ void basic_regex_creator<charT, traits>::fixup_recursions(re_syntax_base* state)
if(idx < 0)
{
idx = -idx-1;
if(idx >= 10000)
if(idx >= hash_value_mask)
{
idx = m_pdata->get_id(idx);
if(idx <= 0)
@ -796,7 +798,7 @@ void basic_regex_creator<charT, traits>::fixup_recursions(re_syntax_base* state)
bool ok = false;
re_syntax_base* p = base;
std::ptrdiff_t idx = static_cast<re_jump*>(state)->alt.i;
if(idx > 10000)
if(idx >= hash_value_mask)
{
//
// There may be more than one capture group with this hash, just do what Perl

11
include/boost/regex/v4/basic_regex_parser.hpp
View File

@ -558,8 +558,8 @@ bool basic_regex_parser<charT, traits>::parse_open_paren()
//
// allow backrefs to this mark:
//
if((markid > 0) && (markid < sizeof(unsigned) * CHAR_BIT))
this->m_backrefs |= 1u << (markid - 1);
if(markid > 0)
this->m_backrefs.set(markid);
return true;
}
@ -925,7 +925,7 @@ escape_type_class_jump:
}
if(negative)
i = 1 + (static_cast<boost::intmax_t>(m_mark_count) - i);
if(((i > 0) && (i < std::numeric_limits<unsigned>::digits) && (i - 1 < static_cast<boost::intmax_t>(sizeof(unsigned) * CHAR_BIT)) && (this->m_backrefs & (1u << (i-1)))) || ((i > 10000) && (this->m_pdata->get_id(i) > 0) && (static_cast<boost::intmax_t>(this->m_pdata->get_id(i))-1 < static_cast<boost::intmax_t>(sizeof(unsigned) * CHAR_BIT)) && (this->m_backrefs & (1u << (this->m_pdata->get_id(i)-1)))))
if(((i < hash_value_mask) && (i > 0) && (this->m_backrefs.test(i))) || ((i >= hash_value_mask) && (this->m_pdata->get_id(i) > 0) && (this->m_backrefs.test(this->m_pdata->get_id(i)))))
{
m_position = pc;
re_brace* pb = static_cast<re_brace*>(this->append_state(syntax_element_backref, sizeof(re_brace)));
@ -1957,7 +1957,7 @@ bool basic_regex_parser<charT, traits>::parse_backref()
charT c = unescape_character();
this->append_literal(c);
}
else if((i > 0) && (this->m_backrefs & (1u << (i-1))))
else if((i > 0) && (this->m_backrefs.test(i)))
{
m_position = pc;
re_brace* pb = static_cast<re_brace*>(this->append_state(syntax_element_backref, sizeof(re_brace)));
@ -2731,8 +2731,7 @@ option_group_jump:
//
// allow backrefs to this mark:
//
if(markid < (int)(sizeof(unsigned) * CHAR_BIT))
this->m_backrefs |= 1u << (markid - 1);
this->m_backrefs.set(markid);
}
return true;
}

54
include/boost/regex/v4/indexed_bit_flag.hpp Normal file
View File

@ -0,0 +1,54 @@
/*
*
* Copyright (c) 2020
* John Maddock
*
* Use, modification and distribution are subject to the
* Boost Software License, Version 1.0. (See accompanying file
* LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
*
*/
/*
* LOCATION: see http://www.boost.org for most recent version.
* FILE basic_regex_parser.cpp
* VERSION see <boost/version.hpp>
* DESCRIPTION: Declares template class basic_regex_parser.
*/
#include <boost/regex/config.hpp>
#include <set>
#ifndef BOOST_REGEX_V4_INDEXED_BIT_FLAG_HPP
#define BOOST_REGEX_V4_INDEXED_BIT_FLAG_HPP
namespace boost{
namespace BOOST_REGEX_DETAIL_NS{
class indexed_bit_flag
{
boost::uint64_t low_mask;
std::set<std::size_t> mask_set;
public:
indexed_bit_flag() : low_mask(0) {}
void set(std::size_t i)
{
if (i < std::numeric_limits<boost::uint64_t>::digits - 1)
low_mask |= static_cast<boost::uint64_t>(1u) << i;
else
mask_set.insert(i);
}
bool test(std::size_t i)
{
if (i < std::numeric_limits<boost::uint64_t>::digits - 1)
return low_mask & static_cast<boost::uint64_t>(1u) << i ? true : false;
else
return mask_set.find(i) != mask_set.end();
}
};
} // namespace BOOST_REGEX_DETAIL_NS
} // namespace boost
#endif

6
include/boost/regex/v4/perl_matcher_common.hpp
View File

@ -619,7 +619,7 @@ bool perl_matcher<BidiIterator, Allocator, traits>::match_backref()
// or PCRE.
//
int index = static_cast<const re_brace*>(pstate)->index;
if(index >= 10000)
if(index >= hash_value_mask)
{
named_subexpressions::range_type r = re.get_data().equal_range(index);
BOOST_ASSERT(r.first != r.second);
@ -768,7 +768,7 @@ inline bool perl_matcher<BidiIterator, Allocator, traits>::match_assert_backref(
{
// Have we matched subexpression "index"?
// Check if index is a hash value:
if(index >= 10000)
if(index >= hash_value_mask)
{
named_subexpressions::range_type r = re.get_data().equal_range(index);
while(r.first != r.second)
@ -792,7 +792,7 @@ inline bool perl_matcher<BidiIterator, Allocator, traits>::match_assert_backref(
// Have we recursed into subexpression "index"?
// If index == 0 then check for any recursion at all, otherwise for recursion to -index-1.
int idx = -(index+1);
if(idx >= 10000)
if(idx >= hash_value_mask)
{
named_subexpressions::range_type r = re.get_data().equal_range(idx);
int stack_index = recursion_stack.empty() ? -1 : recursion_stack.back().idx;