boostorg/regex
1
0
Fork 1
mirror of https://github.com/boostorg/regex.git synced 2025-07-30 04:27:22 +02:00
Code Issues Packages Projects Releases Wiki Activity

merged from trunk

Browse Source 
[SVN r48605]
This commit is contained in:
Eric Niebler
2008-09-05 04:13:30 +00:00
parent 003f785f4e
commit a388804b3e
11 changed files with 75 additions and 54 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
build/Jamfile.v2
View File

@ -169,7 +169,8 @@ rule check-icu-config ( )
if [ check-icu-config ]
{
BOOST_REGEX_ICU_OPTS = "<define>BOOST_HAS_ICU=1" ;
BOOST_REGEX_ICU_OPTS = "<target-os>freebsd:<include>/usr/local/include" ;
BOOST_REGEX_ICU_OPTS += "<define>BOOST_HAS_ICU=1" ;
if $(ICU_PATH)
{
@ -259,3 +260,4 @@ boost-install boost_regex ;

10
include/boost/regex/concepts.hpp
View File

@ -87,12 +87,12 @@ struct regex_traits_architype
public:
regex_traits_architype();
typedef charT char_type;
typedef std::size_t size_type;
// typedef std::size_t size_type;
typedef std::vector<char_type> string_type;
typedef copy_constructible_archetype<assignable_archetype<> > locale_type;
typedef bitmask_archetype char_class_type;
static size_type length(const char_type* ) { return 0; }
static std::size_t length(const char_type* ) { return 0; }
charT translate(charT ) const { return charT(); }
charT translate_nocase(charT ) const { return static_object<charT>::get(); }
@ -163,21 +163,21 @@ struct RegexTraitsConcept
RegexTraitsConcept();
// required typedefs:
typedef typename traits::char_type char_type;
typedef typename traits::size_type size_type;
// typedef typename traits::size_type size_type;
typedef typename traits::string_type string_type;
typedef typename traits::locale_type locale_type;
typedef typename traits::char_class_type char_class_type;
void constraints()
{
function_requires<UnsignedIntegerConcept<size_type> >();
//function_requires<UnsignedIntegerConcept<size_type> >();
function_requires<RandomAccessContainerConcept<string_type> >();
function_requires<DefaultConstructibleConcept<locale_type> >();
function_requires<CopyConstructibleConcept<locale_type> >();
function_requires<AssignableConcept<locale_type> >();
function_requires<BitmaskConcept<char_class_type> >();
size_type n = traits::length(m_pointer);
std::size_t n = traits::length(m_pointer);
ignore_unused_variable_warning(n);
char_type c = m_ctraits.translate(m_char);

10
include/boost/regex/icu.hpp
View File

@ -355,7 +355,7 @@ inline u32regex make_u32regex(const wchar_t* p, boost::regex_constants::syntax_o
return re_detail::do_make_u32regex(p, p + std::wcslen(p), opt, static_cast<boost::mpl::int_<sizeof(wchar_t)> const*>(0));
}
#endif
#ifndef U_WCHAR_IS_UTF16
#if !defined(U_WCHAR_IS_UTF16) && (U_SIZEOF_WCHAR_T != 2)
inline u32regex make_u32regex(const UChar* p, boost::regex_constants::syntax_option_type opt = boost::regex_constants::perl)
{
return re_detail::do_make_u32regex(p, p + u_strlen(p), opt, static_cast<boost::mpl::int_<2> const*>(0));
@ -455,7 +455,7 @@ inline bool u32regex_match(const UChar* p,
{
return re_detail::do_regex_match(p, p+u_strlen(p), m, e, flags, static_cast<mpl::int_<2> const*>(0));
}
#if !defined(U_WCHAR_IS_UTF16) && !defined(BOOST_NO_WREGEX)
#if !defined(U_WCHAR_IS_UTF16) && (U_SIZEOF_WCHAR_T != 2) && !defined(BOOST_NO_WREGEX)
inline bool u32regex_match(const wchar_t* p,
match_results<const wchar_t*>& m,
const u32regex& e,
@ -519,7 +519,7 @@ inline bool u32regex_match(const UChar* p,
match_results<const UChar*> m;
return re_detail::do_regex_match(p, p+u_strlen(p), m, e, flags, static_cast<mpl::int_<2> const*>(0));
}
#if !defined(U_WCHAR_IS_UTF16) && !defined(BOOST_NO_WREGEX)
#if !defined(U_WCHAR_IS_UTF16) && (U_SIZEOF_WCHAR_T != 2) && !defined(BOOST_NO_WREGEX)
inline bool u32regex_match(const wchar_t* p,
const u32regex& e,
match_flag_type flags = match_default)
@ -640,7 +640,7 @@ inline bool u32regex_search(const UChar* p,
{
return re_detail::do_regex_search(p, p+u_strlen(p), m, e, flags, p, static_cast<mpl::int_<2> const*>(0));
}
#if !defined(U_WCHAR_IS_UTF16) && !defined(BOOST_NO_WREGEX)
#if !defined(U_WCHAR_IS_UTF16) && (U_SIZEOF_WCHAR_T != 2) && !defined(BOOST_NO_WREGEX)
inline bool u32regex_search(const wchar_t* p,
match_results<const wchar_t*>& m,
const u32regex& e,
@ -701,7 +701,7 @@ inline bool u32regex_search(const UChar* p,
match_results<const UChar*> m;
return re_detail::do_regex_search(p, p+u_strlen(p), m, e, flags, p, static_cast<mpl::int_<2> const*>(0));
}
#if !defined(U_WCHAR_IS_UTF16) && !defined(BOOST_NO_WREGEX)
#if !defined(U_WCHAR_IS_UTF16) && (U_SIZEOF_WCHAR_T != 2) && !defined(BOOST_NO_WREGEX)
inline bool u32regex_search(const wchar_t* p,
const u32regex& e,
match_flag_type flags = match_default)

2
include/boost/regex/v4/basic_regex.hpp
View File

@ -191,7 +191,7 @@ class basic_regex : public regbase
{
public:
// typedefs:
typedef typename traits::size_type traits_size_type;
typedef std::size_t traits_size_type;
typedef typename traits::string_type traits_string_type;
typedef charT char_type;
typedef traits traits_type;

2
include/boost/regex/v4/perl_matcher.hpp
View File

@ -331,7 +331,7 @@ public:
typedef typename traits::char_type char_type;
typedef perl_matcher<BidiIterator, Allocator, traits> self_type;
typedef bool (self_type::*matcher_proc_type)(void);
typedef typename traits::size_type traits_size_type;
typedef std::size_t traits_size_type;
typedef typename is_byte<char_type>::width_type width_type;
typedef typename regex_iterator_traits<BidiIterator>::difference_type difference_type;

4
include/boost/regex/v4/perl_matcher_non_recursive.hpp
View File

@ -477,7 +477,9 @@ bool perl_matcher<BidiIterator, Allocator, traits>::match_rep()
take_second = can_start(*position, rep->_map, (unsigned char)mask_skip);
}
if(take_first || (next_count->get_id() != rep->id))
if((m_backup_state->id != saved_state_repeater_count)
|| (static_cast<saved_repeater<BidiIterator>*>(m_backup_state)->count.get_id() != rep->id)
|| (next_count->get_id() != rep->id))
{
// we're moving to a different repeat from the last
// one, so set up a counter object:

2
include/boost/regex/v4/u32regex_iterator.hpp
View File

@ -166,7 +166,7 @@ inline u32regex_iterator<const wchar_t*> make_u32regex_iterator(const wchar_t* p
return u32regex_iterator<const wchar_t*>(p, p+std::wcslen(p), e, m);
}
#endif
#ifndef U_WCHAR_IS_UTF16
#if !defined(U_WCHAR_IS_UTF16) && (U_SIZEOF_WCHAR_T != 2)
inline u32regex_iterator<const UChar*> make_u32regex_iterator(const UChar* p, const u32regex& e, regex_constants::match_flag_type m = regex_constants::match_default)
{
return u32regex_iterator<const UChar*>(p, p+u_strlen(p), e, m);

6
include/boost/regex/v4/u32regex_token_iterator.hpp
View File

@ -282,7 +282,7 @@ inline u32regex_token_iterator<const wchar_t*> make_u32regex_token_iterator(cons
return u32regex_token_iterator<const wchar_t*>(p, p+std::wcslen(p), e, submatch, m);
}
#endif
#ifndef U_WCHAR_IS_UTF16
#if !defined(U_WCHAR_IS_UTF16) && (U_SIZEOF_WCHAR_T != 2)
inline u32regex_token_iterator<const UChar*> make_u32regex_token_iterator(const UChar* p, const u32regex& e, int submatch = 0, regex_constants::match_flag_type m = regex_constants::match_default)
{
return u32regex_token_iterator<const UChar*>(p, p+u_strlen(p), e, m);
@ -313,7 +313,7 @@ inline u32regex_token_iterator<const wchar_t*> make_u32regex_token_iterator(cons
return u32regex_token_iterator<const wchar_t*>(p, p+std::wcslen(p), e, submatch, m);
}
#endif
#ifndef U_WCHAR_IS_UTF16
#if !defined(U_WCHAR_IS_UTF16) && (U_SIZEOF_WCHAR_T != 2)
template <std::size_t N>
inline u32regex_token_iterator<const UChar*> make_u32regex_token_iterator(const UChar* p, const u32regex& e, const int (&submatch)[N], regex_constants::match_flag_type m = regex_constants::match_default)
{
@ -344,7 +344,7 @@ inline u32regex_token_iterator<const wchar_t*> make_u32regex_token_iterator(cons
return u32regex_token_iterator<const wchar_t*>(p, p+std::wcslen(p), e, submatch, m);
}
#endif
#ifndef U_WCHAR_IS_UTF16
#if !defined(U_WCHAR_IS_UTF16) && (U_SIZEOF_WCHAR_T != 2)
inline u32regex_token_iterator<const UChar*> make_u32regex_token_iterator(const UChar* p, const u32regex& e, const std::vector<int>& submatch, regex_constants::match_flag_type m = regex_constants::match_default)
{
return u32regex_token_iterator<const UChar*>(p, p+u_strlen(p), e, m);

83
src/usinstances.cpp
View File

@ -18,46 +18,59 @@
#define BOOST_REGEX_SOURCE
#include <boost/detail/workaround.hpp>
#include <memory>
#include <string>
#if defined(_DLL_CPPLIB) && !defined(_M_CEE_PURE) \
&& !(defined(__SGI_STL_PORT) || defined(_STLPORT_VERSION) || defined(__STD_RWCOMPILER_H__) || defined(_RWSTD_VER))
//
// This is a horrible workaround, but without declaring these symbols extern we get
// duplicate symbol errors when linking if the application is built without
// /Zc:wchar_t
//
#ifdef _CRTIMP2_PURE
# define BOOST_REGEX_STDLIB_DECL _CRTIMP2_PURE
#else
# define BOOST_REGEX_STDLIB_DECL _CRTIMP2
#endif
namespace std{
#if BOOST_WORKAROUND(BOOST_MSVC, >= 1400)
template class BOOST_REGEX_STDLIB_DECL allocator<unsigned short>;
template class BOOST_REGEX_STDLIB_DECL _String_val<unsigned short, allocator<unsigned short> >;
template class BOOST_REGEX_STDLIB_DECL basic_string<unsigned short, char_traits<unsigned short>,
allocator<unsigned short> >;
#endif
#if BOOST_WORKAROUND(BOOST_MSVC, > 1300) && BOOST_WORKAROUND(BOOST_MSVC, BOOST_TESTED_AT(1400))
template<> BOOST_REGEX_STDLIB_DECL std::size_t __cdecl char_traits<unsigned short>::length(unsigned short const*);
#endif
template BOOST_REGEX_STDLIB_DECL bool __cdecl operator==(
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&,
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&);
template BOOST_REGEX_STDLIB_DECL bool __cdecl operator==(
const unsigned short *,
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&);
template BOOST_REGEX_STDLIB_DECL bool __cdecl operator==(
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&,
const unsigned short *);
template BOOST_REGEX_STDLIB_DECL bool __cdecl operator<(
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&,
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&);
template BOOST_REGEX_STDLIB_DECL bool __cdecl operator>(
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&,
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&);
}
#endif
#include <boost/regex/config.hpp>
#if !defined(BOOST_NO_WREGEX) && defined(BOOST_REGEX_HAS_OTHER_WCHAR_T) && !defined(BOOST_REGEX_NO_EXTERNAL_TEMPLATES)
#define BOOST_REGEX_US_INSTANTIATE
#ifdef _DLL_CPPLIB
#include <boost/detail/workaround.hpp>
#include <memory>
#include <string>
//
// This is a horrible workaround, without declaring these symbols extern we get
// duplicate symbol errors when linking if the application is built without
// /Zc:wchar_t
//
namespace std{
template _CRTIMP2 bool __cdecl operator==(
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&,
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&);
template _CRTIMP2 bool __cdecl operator==(
const unsigned short *,
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&);
template _CRTIMP2 bool __cdecl operator==(
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&,
const unsigned short *);
template _CRTIMP2 bool __cdecl operator<(
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&,
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&);
template _CRTIMP2 bool __cdecl operator>(
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&,
const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&);
#if BOOST_WORKAROUND(BOOST_MSVC, > 1300) && BOOST_WORKAROUND(BOOST_MSVC, BOOST_TESTED_AT(1400))
template<> _CRTIMP2 std::size_t __cdecl char_traits<unsigned short>::length(unsigned short const*);
#endif
#if BOOST_WORKAROUND(BOOST_MSVC, >= 1400)
template _CRTIMP2 allocator<unsigned short>::allocator();
#endif
}
#endif
#include <boost/regex.hpp>
#endif

3
test/pathology/bad_expression_test.cpp
View File

@ -46,7 +46,7 @@ int test_main( int , char* [] )
bad_text.assign((std::string::size_type)500000, 'a');
e2.assign("aaa*@");
BOOST_CHECK_THROW(0 == boost::regex_search(bad_text, what, e2), std::runtime_error);
BOOST_CHECK_THROW(boost::regex_search(bad_text, what, e2), std::runtime_error);
good_text.assign((std::string::size_type)5000, 'a');
BOOST_CHECK(0 == boost::regex_search(good_text, what, e2));
@ -54,3 +54,4 @@ int test_main( int , char* [] )
}
#include <boost/test/included/test_exec_monitor.hpp>

3
test/regress/test_tricky_cases.cpp
View File

@ -377,6 +377,9 @@ void test_tricky_cases3()
TEST_REGEX_SEARCH("(?:\\d{9}.*){2}", perl, "123456789dfsdfsdfsfsdfds123456789b", match_default, make_array(0, 34, -2, -2));
TEST_REGEX_SEARCH("(?:\\d{9}.*){2}", perl, "123456789dfsdfsdfsfsdfds12345678", match_default, make_array(-2, -2));
TEST_REGEX_SEARCH("(?:\\d{9}.*){2}", perl, "123456789dfsdfsdfsfsdfds", match_default, make_array(-2, -2));
TEST_REGEX_SEARCH("^(25[0-5]|2[0-4][0-9]|[0-1]?[0-9]?[0-9])(\\.(25[0-5]|2[0-4][0-9]|[0-1]?[0-9]?[0-9])){3}$", perl, "1.2.03", match_default, make_array(-2, -2));
TEST_REGEX_SEARCH("^(25[0-5]|2[0-4][0-9]|[0-1]?[0-9]?[0-9])(\\.(25[0-5]|2[0-4][0-9]|[0-1]?[0-9]?[0-9])){3,4}$", perl, "1.2.03", match_default, make_array(-2, -2));
TEST_REGEX_SEARCH("^(25[0-5]|2[0-4][0-9]|[0-1]?[0-9]?[0-9])(\\.(25[0-5]|2[0-4][0-9]|[0-1]?[0-9]?[0-9])){3,4}?$", perl, "1.2.03", match_default, make_array(-2, -2));
//