Make backstep calculation non-recursive.

Refs https://github.com/boostorg/regex/pull/236.

.github/workflows/ci.yml

@@ -9,6 +9,7 @@ on:
     branches:
       - master
       - develop
+      - feature/**
   pull_request:
   release:
     types: [published, created, edited]

build.jam

@@ -0,0 +1,30 @@
+# Copyright René Ferdinand Rivera Morell 2023-2024
+# Distributed under 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)
+
+require-b2 5.2 ;
+
+constant boost_dependencies :
+    /boost/assert//boost_assert
+    /boost/concept_check//boost_concept_check
+    /boost/config//boost_config
+    /boost/predef//boost_predef
+    /boost/static_assert//boost_static_assert
+    /boost/throw_exception//boost_throw_exception
+    /boost/type_traits//boost_type_traits ;
+
+project /boost/regex
+    : common-requirements
+        <include>include
+    ;
+
+explicit
+    [ alias boost_regex : build//boost_regex ]
+    [ alias all : boost_regex example test ]
+    ;
+
+call-if : boost-library regex
+    : install boost_regex
+    ;
+

build/Jamfile.v2

@@ -1,15 +1,20 @@
 # copyright John Maddock 2003
-# Distributed under the Boost Software License, Version 1.0. 
-# (See accompanying file LICENSE_1_0.txt or copy at 
+# Distributed under 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.
 
 import modules ;
 import testing ;
 import errors ;
 
-project : requirements
-   # default to all warnings on:
-   <warnings>all
+constant boost_dependencies_private :
+    /boost/core//boost_core
+    ;
+
+project
+   : default-build <warnings>all
+   : common-requirements <library>$(boost_dependencies)
+   : requirements <library>$(boost_dependencies_private)
    ;
 
 local disable-icu = [ MATCH (--disable-icu) : [ modules.peek : ARGV ] ] ;
@@ -17,13 +22,13 @@ local disable-icu = [ MATCH (--disable-icu) : [ modules.peek : ARGV ] ] ;
 rule path_options ( properties * )
 {
     local result ;
-    if <address-model>64 in $(properties) && <toolset>msvc in $(properties) 
+    if <address-model>64 in $(properties) && <toolset>msvc in $(properties)
     {
-    	result = <search>$(ICU_PATH)/bin64 <search>$(ICU_PATH)/lib64 ;
+        result = <search>$(ICU_PATH)/bin64 <search>$(ICU_PATH)/lib64 ;
     }
     else
     {
-    	result = <search>$(ICU_PATH)/bin <search>$(ICU_PATH)/lib ;
+        result = <search>$(ICU_PATH)/bin <search>$(ICU_PATH)/lib ;
     }
     return $(result) ;
 }
@@ -34,24 +39,24 @@ rule path_options ( properties * )
 if ! $(disable-icu)
 {
    if [ modules.peek : ICU_LINK ]
-   {    
+   {
        errors.user-error : "The ICU_LINK option is no longer supported by the Boost.Regex build - please refer to the documentation for equivalent options" ;
    }
 
    if [ modules.peek : ICU_PATH ]
-   {    
+   {
        ICU_PATH =  [ modules.peek : ICU_PATH ] ;
    }
    if [ modules.peek : ICU_ICUUC_NAME ]
-   {    
+   {
        ICU_ICUUC_NAME =  [ modules.peek : ICU_ICUUC_NAME ] ;
    }
    if [ modules.peek : ICU_ICUDT_NAME ]
-   {    
+   {
        ICU_ICUDT_NAME =  [ modules.peek : ICU_ICUDT_NAME ] ;
    }
    if [ modules.peek : ICU_ICUIN_NAME ]
-   {    
+   {
        ICU_ICUIN_NAME =  [ modules.peek : ICU_ICUIN_NAME ] ;
    }
 
@@ -102,8 +107,8 @@ if ! $(disable-icu)
       lib icuin : : <name>this_is_an_invalid_library_name ;
    }
 
-   ICU_OPTS = 
-      <include>$(ICU_PATH)/include 
+   ICU_OPTS =
+      <include>$(ICU_PATH)/include
       <runtime-link>shared:<library>icuuc/<link>shared
       <runtime-link>shared:<library>icudt/<link>shared
       <runtime-link>shared:<library>icuin/<link>shared
@@ -111,7 +116,7 @@ if ! $(disable-icu)
       <runtime-link>static:<library>icudt
       <runtime-link>static:<library>icuin
       <target-os>windows,<toolset>clang:<linkflags>"advapi32.lib"
-      <define>BOOST_HAS_ICU=1 
+      <define>BOOST_HAS_ICU=1
       <runtime-link>static:<define>U_STATIC_IMPLEMENTATION=1
       ;
 
@@ -144,16 +149,15 @@ explicit has_icu ;
 
 alias icu_options : : : : [ check-target-builds has_icu : $(ICU_OPTS) : ] ;
 
-SOURCES = 
+SOURCES =
    posix_api.cpp
    wide_posix_api.cpp
 ;
 
 lib boost_regex : ../src/$(SOURCES) icu_options
-    :         
-	     <link>shared:<define>BOOST_REGEX_DYN_LINK=1
-	     <toolset>gcc-cygwin:<link>static
+    :
+        <link>shared:<define>BOOST_REGEX_DYN_LINK=1
+        <toolset>gcc-cygwin:<link>static
+    : :
+        <define>BOOST_REGEX_NO_LIB=1
     ;
-
-boost-install boost_regex ;
-

example/Jamfile.v2

@@ -1,32 +1,34 @@
 # copyright John Maddock 2003
-# Distributed under the Boost Software License, Version 1.0. 
-# (See accompanying file LICENSE_1_0.txt or copy at 
+# Distributed under 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.
 
 project
-    : requirements 
-      <threading>multi 
+    : requirements
+      <library>/boost/regex//boost_regex
+      <library>/boost/timer//boost_timer
+      <threading>multi
       <link>shared:<define>BOOST_REGEX_DYN_LINK=1
       # There are unidentified linker problems on these platforms:
       <toolset>mipspro-7.4:<link>static
       <toolset>sun-5.9:<link>static
       <toolset>msvc:<warnings>all
-      <toolset>gcc:<warnings>all
-      <toolset>gcc:<cxxflags>-Wextra
+      <toolset>gcc:<warnings>extra
       <define>U_USING_ICU_NAMESPACE=0
       #<toolset>gcc-mw:<link>static
 	  #<toolset>gcc-mingw:<link>static
 	  <toolset>gcc-cygwin:<link>static
+      <define>BOOST_TIMER_ENABLE_DEPRECATED
     ;
 
 rule regex-test-run ( sources + : input * : name * )
 {
-        return [ 
+        return [
            run
              # sources
              $(sources)
              # dependencies
-             ../build//boost_regex
+             /boost/regex//boost_regex
          :  # additional args
             $(input)
          :  # test-files
@@ -37,32 +39,33 @@ rule regex-test-run ( sources + : input * : name * )
         ] ;
 }
 
+path-constant HERE : . ;
+
 test-suite regex-examples :
 
-[ regex-test-run timer/regex_timer.cpp : $(BOOST_ROOT)/libs/regex/example/timer/input_script.txt ]
-[ regex-test-run grep/grep.cpp ../../program_options/build//boost_program_options/<link>static : -n -b $(BOOST_ROOT)/boost/regex.hpp $(BOOST_ROOT)/boost/type_traits.hpp : test_grep ]
+[ regex-test-run timer/regex_timer.cpp /boost/smart_ptr//boost_smart_ptr : $(HERE)/timer/input_script.txt ]
+[ regex-test-run grep/grep.cpp /boost/program_options//boost_program_options/<link>static : -n -b -E include $(HERE)/../include/boost/regex.hpp : test_grep ]
 [ regex-test-run snippets/credit_card_example.cpp ]
 [ regex-test-run snippets/mfc_example.cpp ]
 [ regex-test-run snippets/icu_example.cpp ]
-[ regex-test-run snippets/partial_regex_grep.cpp : $(BOOST_ROOT)/libs/regex/index.htm ]
-[ regex-test-run snippets/partial_regex_iterate.cpp : $(BOOST_ROOT)/libs/regex/index.htm ]
+[ regex-test-run snippets/partial_regex_grep.cpp : $(HERE)/../index.htm ]
+[ regex-test-run snippets/partial_regex_iterate.cpp : $(HERE)/../index.htm ]
 [ regex-test-run snippets/partial_regex_match.cpp : 1234-5678-8765-4 ]
-[ regex-test-run snippets/regex_grep_example_1.cpp : $(BOOST_ROOT)/boost/regex/v5/regex_token_iterator.hpp ]
-[ regex-test-run snippets/regex_grep_example_2.cpp : $(BOOST_ROOT)/boost/regex/v5/regex_token_iterator.hpp ]
-[ regex-test-run snippets/regex_grep_example_3.cpp : $(BOOST_ROOT)/boost/regex/v5/regex_token_iterator.hpp ]
-[ regex-test-run snippets/regex_grep_example_4.cpp : $(BOOST_ROOT)/boost/regex/v5/regex_token_iterator.hpp ]
+[ regex-test-run snippets/regex_grep_example_1.cpp : $(HERE)/../include/boost/regex/v5/regex_iterator.hpp ]
+[ regex-test-run snippets/regex_grep_example_2.cpp : $(HERE)/../include/boost/regex/v5/regex_iterator.hpp ]
+[ regex-test-run snippets/regex_grep_example_3.cpp : $(HERE)/../include/boost/regex/v5/regex_iterator.hpp ]
+[ regex-test-run snippets/regex_grep_example_4.cpp : $(HERE)/../include/boost/regex/v5/regex_iterator.hpp ]
 [ regex-test-run snippets/regex_match_example.cpp : -auto ]
-[ regex-test-run snippets/regex_merge_example.cpp : $(BOOST_ROOT)/boost/regex/v5/regex_token_iterator.hpp ]
-[ regex-test-run snippets/regex_replace_example.cpp : $(BOOST_ROOT)/boost/regex/v5/regex_token_iterator.hpp ]
-[ regex-test-run snippets/regex_search_example.cpp : $(BOOST_ROOT)/boost/regex/v5/regex_token_iterator.hpp ]
+[ regex-test-run snippets/regex_merge_example.cpp : $(HERE)/../include/boost/regex/v5/regex_iterator.hpp ]
+[ regex-test-run snippets/regex_replace_example.cpp : $(HERE)/../include/boost/regex/v5/regex_iterator.hpp ]
+[ regex-test-run snippets/regex_search_example.cpp : $(HERE)/../include/boost/regex/v5/regex_iterator.hpp ]
 [ regex-test-run snippets/regex_split_example_1.cpp : -auto ]
-[ regex-test-run snippets/regex_split_example_2.cpp : $(BOOST_ROOT)/libs/regex/doc/html/index.html ]
+[ regex-test-run snippets/regex_split_example_2.cpp : $(HERE)/../doc/html/index.html ]
 [ regex-test-run snippets/regex_token_iterator_eg_1.cpp : -auto ]
-[ regex-test-run snippets/regex_token_iterator_eg_2.cpp : $(BOOST_ROOT)/libs/regex/doc/html/index.html ]
-[ regex-test-run snippets/regex_iterator_example.cpp : $(BOOST_ROOT)/boost/regex/v5/regex_token_iterator.hpp ]
+[ regex-test-run snippets/regex_token_iterator_eg_2.cpp : $(HERE)/../doc/html/index.html ]
+[ regex-test-run snippets/regex_iterator_example.cpp : $(HERE)/../include/boost/regex/v5/regex_iterator.hpp ]
 [ run snippets/captures_example.cpp
   ../test/captures//boost_regex_extra
   : : :  <threading>multi <define>BOOST_REGEX_MATCH_EXTRA=1 ]
 
 ;
-

include/boost/regex/concepts.hpp

@@ -76,13 +76,10 @@ inline long hash_value(char_architype val)
 //
 } // namespace boost
 namespace std{
-   template<> struct char_traits<boost::char_architype>
-   {
-      // The intent is that this template is not instantiated,
-      // but this typedef gives us a chance of compilation in
-      // case it is:
-      typedef boost::char_architype char_type;
-   };
+   //
+   // We should never use this, if we do it should be an error:
+   //
+   template<> struct char_traits<boost::char_architype>;
 }
 //
 // Allocator architype:
@@ -412,6 +409,10 @@ struct BaseRegexConcept
       Regex e5(in1, in2, m_flags);
       ignore_unused_variable_warning(e5);
 
+      // equals:
+      e1 == e2;
+      e1 != e2;
+
       // assign etc:
       Regex e;
       e = m_pointer;

include/boost/regex/v5/basic_regex.hpp

@@ -19,6 +19,15 @@
 #ifndef BOOST_REGEX_V5_BASIC_REGEX_HPP
 #define BOOST_REGEX_V5_BASIC_REGEX_HPP
 
+#include <boost/regex/v5/regbase.hpp>
+#include <boost/regex/v5/syntax_type.hpp>
+#include <boost/regex/v5/regex_traits.hpp>
+#include <boost/regex/v5/states.hpp>
+#include <boost/regex/v5/regex_raw_buffer.hpp>
+
+#include <algorithm>
+#include <limits>
+#include <memory>
 #include <vector>
 
 namespace boost{
@@ -488,7 +497,7 @@ public:
    }
    //
    // swap:
-   void  swap(basic_regex& that)throw()
+   void  swap(basic_regex& that)noexcept
    {
       m_pimpl.swap(that.m_pimpl);
    }
@@ -533,7 +542,21 @@ public:
          return status() - that.status();
       if(flags() != that.flags())
          return flags() - that.flags();
-      return str().compare(that.str());
+      
+      const char_type* i = m_pimpl->begin();
+      const char_type* j = that.m_pimpl->begin();
+      while ((i != m_pimpl->end()) && (j != that.m_pimpl->end()))
+      {
+         if (*i != *j)
+            return *i < *j ? -1 : 1;
+         ++i;
+         ++j;
+      }
+      if (i != m_pimpl->end())
+         return *i > static_cast<char_type>(0) ? 1 : -1;
+      if (j != that.m_pimpl->end())
+         return *j > static_cast<char_type>(0) ? -1 : 1;
+      return 0;
    }
    bool  operator==(const basic_regex& e)const
    { 

include/boost/regex/v5/basic_regex_creator.hpp

@@ -28,6 +28,9 @@
 #endif
 #endif
 
+#include <boost/regex/v5/basic_regex.hpp>
+
+#include <vector>
 #include <set>
 
 namespace boost{
@@ -971,7 +974,12 @@ template <class charT, class traits>
 int basic_regex_creator<charT, traits>::calculate_backstep(re_syntax_base* state)
 {
    typedef typename traits::char_class_type m_type;
+
    int result = 0;
+   int last_alternative_result = -1;
+
+   std::vector<std::tuple<int, re_syntax_base*>> stack;
+
    while(state)
    {
       switch(state->type)
@@ -990,9 +998,28 @@ int basic_regex_creator<charT, traits>::calculate_backstep(re_syntax_base* state
          }
          break;
       case syntax_element_endmark:
-         if((static_cast<re_brace*>(state)->index == -1)
+         if ((static_cast<re_brace*>(state)->index == -1)
             || (static_cast<re_brace*>(state)->index == -2))
-            return result;
+         {
+            // We've finished the calculation, check against any previous alternatives:
+            if (last_alternative_result >= 0)
+            {
+               if (last_alternative_result != result)
+                  return -1;
+            }
+            else
+               last_alternative_result = result;
+
+            if (stack.size())
+            {
+               // Skip to next alternative and calculate that as well:
+               std::tie(result, state) = stack.back();
+               stack.pop_back();
+               continue;
+            }
+            else
+               return result;
+         }
          break;
       case syntax_element_literal:
          result += static_cast<re_literal*>(state)->length;
@@ -1048,11 +1075,13 @@ int basic_regex_creator<charT, traits>::calculate_backstep(re_syntax_base* state
          continue;
       case syntax_element_alt:
          {
-            int r1 = calculate_backstep(state->next.p);
-            int r2 = calculate_backstep(static_cast<re_alt*>(state)->alt.p);
-            if((r1 < 0) || (r1 != r2))
+            // Push the alternative if we haven't pushed too many already:
+            if(stack.size() > BOOST_REGEX_MAX_BLOCKS)
                return -1;
-            return result + r1;
+            stack.push_back(std::make_tuple(result, static_cast<re_alt*>(state)->alt.p));
+            // and take the first one:
+            state = state->next.p;
+            continue;
          }
       default:
          break;

include/boost/regex/v5/basic_regex_parser.hpp

@@ -19,6 +19,13 @@
 #ifndef BOOST_REGEX_V5_BASIC_REGEX_PARSER_HPP
 #define BOOST_REGEX_V5_BASIC_REGEX_PARSER_HPP
 
+#include <boost/regex/v5/basic_regex_creator.hpp>
+
+#include <climits>
+#include <cstdint>
+#include <limits>
+#include <type_traits>
+
 namespace boost{
 namespace BOOST_REGEX_DETAIL_NS{
 

include/boost/regex/v5/c_regex_traits.hpp

@@ -21,7 +21,12 @@
 
 #include <boost/regex/config.hpp>
 #include <boost/regex/v5/regex_workaround.hpp>
+#include <boost/regex/v5/primary_transform.hpp>
+#include <boost/regex/v5/regex_traits_defaults.hpp>
+
 #include <cctype>
+#include <cstdint>
+#include <cwctype>
 
 namespace boost{
 

include/boost/regex/v5/char_regex_traits.hpp

@@ -20,6 +20,8 @@
 #ifndef BOOST_REGEX_V5_CHAR_REGEX_TRAITS_HPP
 #define BOOST_REGEX_V5_CHAR_REGEX_TRAITS_HPP
 
+#include <boost/regex/v5/regex_traits.hpp>
+
 namespace boost{
 
 namespace deprecated{

include/boost/regex/v5/iterator_traits.hpp

@@ -19,6 +19,8 @@
 #ifndef BOOST_REGEX_V5_ITERATOR_TRAITS_HPP
 #define BOOST_REGEX_V5_ITERATOR_TRAITS_HPP
 
+#include <iterator>
+
 namespace boost{
 namespace BOOST_REGEX_DETAIL_NS{
 

include/boost/regex/v5/match_results.hpp

@@ -19,6 +19,14 @@
 #ifndef BOOST_REGEX_V5_MATCH_RESULTS_HPP
 #define BOOST_REGEX_V5_MATCH_RESULTS_HPP
 
+#include <boost/regex/v5/match_flags.hpp>
+#include <boost/regex/v5/sub_match.hpp>
+#include <boost/regex/v5/basic_regex.hpp>
+#include <boost/regex/v5/regex_format.hpp>
+
+#include <string>
+#include <vector>
+
 namespace boost{
 #ifdef BOOST_REGEX_MSVC
 #pragma warning(push)
@@ -97,15 +105,23 @@ public:
    bool empty() const
    { return m_subs.size() < 2; }
    // element access:
-   difference_type length(int sub = 0) const
+private:
+   difference_type do_get_length(int sub = 0) const
    {
-      if(m_is_singular)
+      if (m_is_singular)
          raise_logic_error();
       sub += 2;
-      if((sub < (int)m_subs.size()) && (sub > 0))
+      if ((sub < (int)m_subs.size()) && (sub > 0))
          return m_subs[sub].length();
       return 0;
    }
+public:
+   template <class Integer>
+   typename std::enable_if<std::is_integral<Integer>::value, difference_type>::type length(Integer sub) const
+   {
+      return do_get_length(static_cast<int>(sub));
+   }
+   difference_type length() const { return do_get_length(0); }
    difference_type length(const char_type* sub) const
    {
       if(m_is_singular)
@@ -161,7 +177,8 @@ public:
    {
       return position(sub.c_str());
    }
-   string_type str(int sub = 0) const
+private:
+   string_type do_get_string(int sub = 0) const
    {
       if(m_is_singular)
          raise_logic_error();
@@ -177,6 +194,13 @@ public:
       }
       return result;
    }
+public:
+   template <class Integer>
+   typename std::enable_if<std::is_integral<Integer>::value, string_type>::type str(Integer sub) const
+   {
+      return do_get_string(static_cast<int>(sub));
+   }
+   string_type str() const { return do_get_string(0); }
    string_type str(const char_type* sub) const
    {
       return (*this)[sub].str();
@@ -196,7 +220,8 @@ public:
    {
       return (*this)[sub].str();
    }
-   const_reference operator[](int sub) const
+   private:
+   const_reference get_at(int sub) const
    {
       if(m_is_singular && m_subs.empty())
          raise_logic_error();
@@ -207,6 +232,12 @@ public:
       }
       return m_null;
    }
+public:
+   template <class Integer>
+   typename std::enable_if<std::is_integral<Integer>::value, const_reference>::type operator[](Integer sub) const
+   {
+      return get_at(static_cast<int>(sub));
+   }
    //
    // Named sub-expressions:
    //
@@ -530,7 +561,7 @@ public:
 
    void  set_named_subs(std::shared_ptr<named_sub_type> subs)
    {
-      m_named_subs = subs;
+      m_named_subs = std::move(subs);
    }
 
 private:

include/boost/regex/v5/mem_block_cache.hpp

@@ -18,6 +18,8 @@
 #ifndef BOOST_REGEX_V5_MEM_BLOCK_CACHE_HPP
 #define BOOST_REGEX_V5_MEM_BLOCK_CACHE_HPP
 
+#include <boost/regex/config.hpp>
+
 #include <new>
 #ifdef BOOST_HAS_THREADS
 #include <mutex>
@@ -31,6 +33,7 @@
   #endif
 #endif
 
+
 namespace boost{
 namespace BOOST_REGEX_DETAIL_NS{
 

include/boost/regex/v5/pattern_except.hpp

@@ -52,7 +52,7 @@ public:
       , m_position(0)
    {
    }
-   ~regex_error() noexcept override {}
+   ~regex_error() noexcept override = default;
    regex_constants::error_type code()const
    { return m_error_code; }
    std::ptrdiff_t position()const

include/boost/regex/v5/perl_matcher.hpp

@@ -12,7 +12,18 @@
 #ifndef BOOST_REGEX_MATCHER_HPP
 #define BOOST_REGEX_MATCHER_HPP
 
+#include <boost/regex/v5/match_flags.hpp>
+#include <boost/regex/v5/match_results.hpp>
+#include <boost/regex/v5/regbase.hpp>
 #include <boost/regex/v5/iterator_category.hpp>
+#include <boost/regex/v5/states.hpp>
+#include <boost/regex/v5/regex_traits.hpp>
+
+#ifndef BOOST_REGEX_STANDALONE
+#include <boost/throw_exception.hpp>
+#endif
+
+#include <climits>
 
 #ifdef BOOST_REGEX_MSVC
 #  pragma warning(push)
@@ -28,25 +39,34 @@
 #endif
 #endif
 
+#ifndef BOOST_REGEX_STANDALONE
+#  define BOOST_REGEX_DETAIL_THROW(ex) boost::throw_exception(ex)
+#else
+#  define BOOST_REGEX_DETAIL_THROW(ex) throw ex
+#endif
+
 namespace boost{
 namespace BOOST_REGEX_DETAIL_NS{
 
 //
 // error checking API:
 //
-inline void  verify_options(boost::regex_constants::syntax_option_type, match_flag_type mf)
+inline void verify_options(boost::regex_constants::syntax_option_type, match_flag_type mf)
 {
+   auto is_perl = (mf & match_perl);
+   auto is_posix = (mf & match_posix);
+
+   if (is_perl && is_posix)
+   {
+      BOOST_REGEX_DETAIL_THROW(std::logic_error("Usage Error: Can't mix Perl and POSIX matching rules"));
+   }
+
    //
    // can't mix match_extra with POSIX matching rules:
    //
-   if ((mf & match_extra) && (mf & match_posix))
+   if ((mf & match_extra) && is_posix)
    {
-      std::logic_error msg("Usage Error: Can't mix regular expression captures with POSIX matching rules");
-#ifndef BOOST_REGEX_STANDALONE
-      throw_exception(msg);
-#else
-      throw msg;
-#endif
+      BOOST_REGEX_DETAIL_THROW(std::logic_error("Usage Error: Can't mix regular expression captures with POSIX matching rules"));
    }
 }
 //

include/boost/regex/v5/perl_matcher_common.hpp

@@ -20,6 +20,22 @@
 #ifndef BOOST_REGEX_V5_PERL_MATCHER_COMMON_HPP
 #define BOOST_REGEX_V5_PERL_MATCHER_COMMON_HPP
 
+#include <boost/regex/config.hpp>
+
+#ifndef BOOST_REGEX_STANDALONE
+
+#include <boost/config.hpp>
+#if defined(BOOST_HAS_PRAGMA_ONCE)
+#pragma once
+#include <boost/regex/v5/perl_matcher.hpp>
+#endif
+
+#endif
+
+#include <boost/regex/v5/basic_regex.hpp>
+#include <boost/regex/v5/match_flags.hpp>
+#include <boost/regex/v5/match_results.hpp>
+
 #ifdef BOOST_REGEX_MSVC
 #  pragma warning(push)
 #pragma warning(disable:4459)
@@ -44,12 +60,7 @@ void perl_matcher<BidiIterator, Allocator, traits>::construct_init(const basic_r
    if(e.empty())
    {
       // precondition failure: e is not a valid regex.
-      std::invalid_argument ex("Invalid regular expression object");
-#ifndef BOOST_REGEX_STANDALONE
-      boost::throw_exception(ex);
-#else
-      throw e;
-#endif
+      BOOST_REGEX_DETAIL_THROW(std::invalid_argument("Invalid regular expression object"));
    }
    pstate = 0;
    m_match_flags = f;
@@ -82,7 +93,11 @@ void perl_matcher<BidiIterator, Allocator, traits>::construct_init(const basic_r
    match_any_mask = static_cast<unsigned char>((f & match_not_dot_newline) ? BOOST_REGEX_DETAIL_NS::test_not_newline : BOOST_REGEX_DETAIL_NS::test_newline);
    // Disable match_any if requested in the state machine:
    if(e.get_data().m_disable_match_any)
+   {
+      if (m_match_flags & match_posix)
+         BOOST_REGEX_DETAIL_THROW(std::logic_error("Invalid regex for POSIX-style matching"));
       m_match_flags &= regex_constants::match_not_any;
+   }
 }
 #ifdef BOOST_REGEX_MSVC
 #  pragma warning(pop)

include/boost/regex/v5/perl_matcher_non_recursive.hpp

@@ -20,6 +20,18 @@
 #ifndef BOOST_REGEX_V5_PERL_MATCHER_NON_RECURSIVE_HPP
 #define BOOST_REGEX_V5_PERL_MATCHER_NON_RECURSIVE_HPP
 
+#include <boost/regex/config.hpp>
+
+#ifndef BOOST_REGEX_STANDALONE
+
+#include <boost/config.hpp>
+#if defined(BOOST_HAS_PRAGMA_ONCE)
+#pragma once
+#include <boost/regex/v5/perl_matcher.hpp>
+#endif
+
+#endif
+
 #include <boost/regex/v5/mem_block_cache.hpp>
 
 #ifdef BOOST_REGEX_MSVC
@@ -1204,7 +1216,10 @@ bool perl_matcher<BidiIterator, Allocator, traits>::skip_until_paren(int index,
       else if(pstate->type == syntax_element_startmark)
       {
          int idx = static_cast<const re_brace*>(pstate)->index;
-         pstate = pstate->next.p;
+         if(idx > 0)
+            match_startmark();
+         else
+            pstate = pstate->next.p;
          skip_until_paren(idx, false);
          continue;
       }

include/boost/regex/v5/regex_format.hpp

@@ -21,6 +21,10 @@
 #ifndef BOOST_REGEX_FORMAT_HPP
 #define BOOST_REGEX_FORMAT_HPP
 
+#include <boost/regex/v5/match_flags.hpp>
+#include <boost/regex/v5/sub_match.hpp>
+#include <boost/regex/v5/regex_traits_defaults.hpp>
+
 #include <type_traits>
 #include <functional>
 

include/boost/regex/v5/regex_grep.hpp

@@ -19,6 +19,10 @@
 #ifndef BOOST_REGEX_V5_REGEX_GREP_HPP
 #define BOOST_REGEX_V5_REGEX_GREP_HPP
 
+#include <boost/regex/v5/basic_regex.hpp>
+#include <boost/regex/v5/match_flags.hpp>
+#include <boost/regex/v5/match_results.hpp>
+#include <boost/regex/v5/perl_matcher.hpp>
 
 namespace boost{
 

include/boost/regex/v5/regex_iterator.hpp

@@ -19,6 +19,9 @@
 #ifndef BOOST_REGEX_V5_REGEX_ITERATOR_HPP
 #define BOOST_REGEX_V5_REGEX_ITERATOR_HPP
 
+#include <boost/regex/v5/basic_regex.hpp>
+#include <boost/regex/v5/match_results.hpp>
+
 #include <memory>
 
 namespace boost{

include/boost/regex/v5/regex_match.hpp

@@ -22,6 +22,9 @@
 #ifndef BOOST_REGEX_MATCH_HPP
 #define BOOST_REGEX_MATCH_HPP
 
+#include <boost/regex/v5/match_results.hpp>
+#include <boost/regex/v5/perl_matcher.hpp>
+
 namespace boost{
 
 //

include/boost/regex/v5/regex_raw_buffer.hpp

@@ -27,6 +27,7 @@
 
 #include <algorithm>
 #include <cstddef>
+#include <cstring>
 
 namespace boost{
    namespace BOOST_REGEX_DETAIL_NS{

include/boost/regex/v5/regex_replace.hpp

@@ -22,6 +22,10 @@
 #define BOOST_REGEX_V5_REGEX_REPLACE_HPP
 
 
+#include <boost/regex/v5/basic_regex.hpp>
+#include <boost/regex/v5/match_flags.hpp>
+#include <boost/regex/v5/regex_iterator.hpp>
+
 namespace boost{
 
 template <class OutputIterator, class BidirectionalIterator, class traits, class charT, class Formatter>

include/boost/regex/v5/regex_search.hpp

@@ -20,6 +20,9 @@
 #define BOOST_REGEX_V5_REGEX_SEARCH_HPP
 
 
+#include <boost/regex/v5/match_results.hpp>
+#include <boost/regex/v5/perl_matcher.hpp>
+
 namespace boost{
 
 template <class BidiIterator, class Allocator, class charT, class traits>

include/boost/regex/v5/regex_split.hpp

@@ -21,6 +21,9 @@
 #ifndef BOOST_REGEX_SPLIT_HPP
 #define BOOST_REGEX_SPLIT_HPP
 
+#include <boost/regex/v5/basic_regex.hpp>
+#include <boost/regex/v5/match_results.hpp>
+
 namespace boost{
 
 #ifdef BOOST_REGEX_MSVC

include/boost/regex/v5/regex_token_iterator.hpp

@@ -19,6 +19,10 @@
 #ifndef BOOST_REGEX_V5_REGEX_TOKEN_ITERATOR_HPP
 #define BOOST_REGEX_V5_REGEX_TOKEN_ITERATOR_HPP
 
+#include <boost/regex/v5/basic_regex.hpp>
+#include <boost/regex/v5/match_results.hpp>
+#include <boost/regex/v5/sub_match.hpp>
+
 #include <memory>
 
 namespace boost{

include/boost/regex/v5/states.hpp

@@ -19,6 +19,11 @@
 #ifndef BOOST_REGEX_V5_STATES_HPP
 #define BOOST_REGEX_V5_STATES_HPP
 
+#include <boost/regex/v5/regex_raw_buffer.hpp>
+
+#include <climits>
+#include <cstddef>
+
 namespace boost{
 namespace BOOST_REGEX_DETAIL_NS{
 

include/boost/regex/v5/sub_match.hpp

@@ -19,6 +19,9 @@
 #ifndef BOOST_REGEX_V5_SUB_MATCH_HPP
 #define BOOST_REGEX_V5_SUB_MATCH_HPP
 
+#include <iterator>
+#include <utility>
+
 namespace boost{
 
 template <class BidiIterator>

include/boost/regex/v5/u32regex_iterator.hpp

@@ -19,6 +19,21 @@
 #ifndef BOOST_REGEX_V5_U32REGEX_ITERATOR_HPP
 #define BOOST_REGEX_V5_U32REGEX_ITERATOR_HPP
 
+#include <boost/regex/config.hpp>
+
+#ifndef BOOST_REGEX_STANDALONE
+
+#include <boost/config.hpp>
+#if defined(BOOST_HAS_PRAGMA_ONCE)
+#pragma once
+#include <boost/regex/v5/icu.hpp>
+#endif
+
+#endif
+
+#include <boost/regex/v5/match_flags.hpp>
+#include <boost/regex/v5/match_results.hpp>
+
 namespace boost{
 
 template <class BidirectionalIterator>

include/boost/regex/v5/u32regex_token_iterator.hpp

@@ -19,6 +19,18 @@
 #ifndef BOOST_REGEX_V5_U32REGEX_TOKEN_ITERATOR_HPP
 #define BOOST_REGEX_V5_U32REGEX_TOKEN_ITERATOR_HPP
 
+#include <boost/regex/config.hpp>
+
+#ifndef BOOST_REGEX_STANDALONE
+
+#include <boost/config.hpp>
+#if defined(BOOST_HAS_PRAGMA_ONCE)
+#pragma once
+#include <boost/regex/v5/icu.hpp>
+#endif
+
+#endif
+
 namespace boost{
 
 #ifdef BOOST_REGEX_MSVC

src/wide_posix_api.cpp

@@ -16,6 +16,7 @@
   *   DESCRIPTION: Implements the wide character POSIX API wrappers.
   */
 
+#define _CRT_SECURE_NO_WARNINGS // for std::wcscpy
 #define BOOST_REGEX_SOURCE
 
 #include <boost/regex/config.hpp>
@@ -169,11 +170,7 @@ BOOST_REGEX_DECL regsize_t BOOST_REGEX_CCALL regerrorW(int code, const regex_tW*
       {
          result = std::wcslen(wnames[code]) + 1;
          if(buf_size >= result)
-#if BOOST_WORKAROUND(BOOST_MSVC, >= 1400) && !defined(_WIN32_WCE) && !defined(UNDER_CE)
-            ::wcscpy_s(buf, buf_size, wnames[code]);
-#else
             std::wcscpy(buf, wnames[code]);
-#endif
          return result;
       }
       return result;
@@ -193,11 +190,7 @@ BOOST_REGEX_DECL regsize_t BOOST_REGEX_CCALL regerrorW(int code, const regex_tW*
             (boost::core::swprintf)(localbuf, 5, L"%d", i);
 #endif
             if(std::wcslen(localbuf) < buf_size)
-#if BOOST_WORKAROUND(BOOST_MSVC, >= 1400) && !defined(_WIN32_WCE) && !defined(UNDER_CE)
-               ::wcscpy_s(buf, buf_size, localbuf);
-#else
                std::wcscpy(buf, localbuf);
-#endif
             return std::wcslen(localbuf) + 1;
          }
       }
@@ -207,11 +200,7 @@ BOOST_REGEX_DECL regsize_t BOOST_REGEX_CCALL regerrorW(int code, const regex_tW*
       (boost::core::swprintf)(localbuf, 5, L"%d", 0);
 #endif
       if(std::wcslen(localbuf) < buf_size)
-#if BOOST_WORKAROUND(BOOST_MSVC, >= 1400) && !defined(_WIN32_WCE) && !defined(UNDER_CE)
-         ::wcscpy_s(buf, buf_size, localbuf);
-#else
          std::wcscpy(buf, localbuf);
-#endif
       return std::wcslen(localbuf) + 1;
    }
    if(code <= (int)REG_E_UNKNOWN)

test/Jamfile.v2

@@ -1,15 +1,16 @@
 # copyright John Maddock 2003
-# Distributed under the Boost Software License, Version 1.0. 
-# (See accompanying file LICENSE_1_0.txt or copy at 
+# Distributed under 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.
 
 project
-    : requirements 
-      <threading>multi 
-      <link>shared:<define>BOOST_REGEX_DYN_LINK=1 
-      <toolset>msvc-7.1:<define>TEST_MFC=1 
+    : requirements
+      <library>/boost/regex//boost_regex
+      <library>/boost/detail//boost_detail
+      <threading>multi
+      <link>shared:<define>BOOST_REGEX_DYN_LINK=1
+      <toolset>msvc-7.1:<define>TEST_MFC=1
       <toolset>msvc-7.0:<define>TEST_MFC=1
-      <toolset>msvc:<asynch-exceptions>on
       # There are unidentified linker problems on these platforms:
       <toolset>mipspro-7.4:<link>static
       <toolset>sun-5.9:<link>static
@@ -21,6 +22,8 @@ project
       #<toolset>gcc-mingw:<link>static
       #<toolset>gcc-cygwin:<link>static
       <toolset>sun:<link>static
+      <include>../src
+      <include>.
     ;
 
 #
@@ -28,14 +31,14 @@ project
 #
 rule regex-test ( name : sources + : requirements * : input-files * )
 {
-    return [ run $(sources) ../build//boost_regex
+    return [ run $(sources) /boost/regex//boost_regex
                :
                : $(input-files)
                : $(requirements)
                : $(name) ] ;
 }
 
-R_SOURCE = 
+R_SOURCE =
 basic_tests.cpp
 main.cpp
 wmain.cpp
@@ -70,7 +73,7 @@ run regress/$(R_SOURCE) ../build//boost_regex ../build//icu_options
    : regex_regress ;
 
 run regress/$(R_SOURCE) ../build//boost_regex
-                           ../../thread/build//boost_thread ../build//icu_options
+                           /boost/thread//boost_thread ../build//icu_options
    : # command line
    : # input files
    : # requirements
@@ -78,7 +81,7 @@ run regress/$(R_SOURCE) ../build//boost_regex
    : regex_regress_threaded ;
 
 run regress/$(R_SOURCE) ../build//boost_regex
-                           ../../thread/build//boost_thread ../build//icu_options
+                           /boost/thread//boost_thread ../build//icu_options
    : # command line
    : # input files
    : # requirements
@@ -98,15 +101,15 @@ run unicode/unicode_iterator_test.cpp : : : release <define>TEST_UTF16 : unicode
 run unicode/unicode_casefold_test.cpp  ../build//boost_regex ../build//icu_options ;
 run object_cache/object_cache_test.cpp ;
 run config_info/regex_config_info.cpp  ../build//boost_regex/<link>static : : : <test-info>always_show_run_output ;
-run config_info/regex_config_info.cpp ../build//boost_regex : : : <test-info>always_show_run_output : regex_dll_config_info ;    
+run config_info/regex_config_info.cpp ../build//boost_regex : : : <test-info>always_show_run_output : regex_dll_config_info ;
 run collate_info/collate_info.cpp ../build//boost_regex : : : <test-info>always_show_run_output : test_collate_info ;
-link concepts/concept_check.cpp : <toolset>gcc:<cxxflags>-Wno-deprecated-copy ;
+link concepts/concept_check.cpp /boost/range//boost_range : <toolset>gcc:<cxxflags>-Wno-deprecated-copy ;
 link concepts/concept_check.cpp : <define>BOOST_REGEX_STANDALONE <toolset>gcc:<cxxflags>-Wno-deprecated-copy : standalone_concept_check ;
 link concepts/icu_concept_check.cpp : <define>BOOST_REGEX_STANDALONE <toolset>gcc:<cxxflags>-Wno-deprecated-copy ;
 link concepts/icu_concept_check.cpp : <toolset>gcc:<cxxflags>-Wno-deprecated-copy : standalone_icu_concept_check ;
-link concepts/range_concept_check.cpp : <toolset>gcc:<cxxflags>-Wno-deprecated-copy ;
-run concepts/test_bug_11988.cpp ;      
-run captures/captures_test.cpp ../build//icu_options : : : <threading>multi <define>BOOST_REGEX_MATCH_EXTRA=1 <define>BOOST_REGEX_NO_LIB=1 : captures_test ;
+link concepts/range_concept_check.cpp : <toolset>gcc:<cxxflags>-Wno-deprecated-copy <library>/boost/range//boost_range ;
+run concepts/test_bug_11988.cpp ;
+run captures/captures_test.cpp ../build//icu_options /boost/array//boost_array : : : <threading>multi <define>BOOST_REGEX_MATCH_EXTRA=1 <define>BOOST_REGEX_NO_LIB=1 : captures_test ;
 run regress/$(R_SOURCE) ./noeh_test//boost_regex_noeh ../build//icu_options : : : <define>BOOST_NO_EXCEPTIONS=1 <exception-handling>off <link>static <runtime-link>shared : regex_regress_noeh ;
 compile test_consolidated.cpp ;
 
@@ -114,12 +117,12 @@ build-project ../example ;
 
 # `quick` target (for CI)
 run quick.cpp ../build//boost_regex ;
-compile test_warnings.cpp 
+compile test_warnings.cpp
    :     <toolset>msvc:<warnings>all <toolset>msvc:<warnings-as-errors>on
          <toolset>gcc:<warnings>all <toolset>gcc:<warnings-as-errors>on
          <toolset>clang:<warnings>all <toolset>clang:<warnings-as-errors>on ;
 
-compile test_warnings.cpp 
+compile test_warnings.cpp
    :     <toolset>msvc:<warnings>all <toolset>msvc:<warnings-as-errors>on
          <toolset>gcc:<warnings>all <toolset>gcc:<warnings-as-errors>on
          <toolset>clang:<warnings>all <toolset>clang:<warnings-as-errors>on
@@ -131,4 +134,8 @@ compile test_windows_defs_2.cpp ;
 compile test_windows_defs_3.cpp ;
 compile test_windows_defs_4.cpp ;
 
-run issue153.cpp : : : <toolset>msvc:<linkflags>-STACK:2097152 ;
+run issue153.cpp : : : "<toolset>msvc:<linkflags>-STACK:2097152" ;
+run issue227.cpp ;
+run issue232.cpp ;
+run lookbehind_recursion_stress_test.cpp ;
+

test/cmake_install_test/CMakeLists.txt

@@ -15,4 +15,4 @@ target_link_libraries(quick Boost::regex Boost::core)
 enable_testing()
 add_test(quick quick)
 
-add_custom_target(check COMMAND ${CMAKE_CTEST_COMMAND} --output-on-failure -C $<CONFIG>)
+add_custom_target(check VERBATIM COMMAND ${CMAKE_CTEST_COMMAND} --output-on-failure -C $<CONFIG>)

test/cmake_subdir_test/CMakeLists.txt

@@ -20,4 +20,4 @@ target_link_libraries(quick Boost::regex Boost::core)
 enable_testing()
 add_test(quick quick)
 
-add_custom_target(check COMMAND ${CMAKE_CTEST_COMMAND} --output-on-failure -C $<CONFIG>)
+add_custom_target(check VERBATIM COMMAND ${CMAKE_CTEST_COMMAND} --output-on-failure -C $<CONFIG>)

test/cmake_subdir_test_icu/CMakeLists.txt

@@ -19,4 +19,4 @@ target_link_libraries(quick_icu Boost::regex_icu)
 enable_testing()
 add_test(quick_icu quick_icu)
 
-add_custom_target(check COMMAND ${CMAKE_CTEST_COMMAND} --output-on-failure -C $<CONFIG>)
+add_custom_target(check VERBATIM COMMAND ${CMAKE_CTEST_COMMAND} --output-on-failure -C $<CONFIG>)

test/config_info/regex_config_info.cpp

@@ -19,7 +19,7 @@
 #endif
 
 #define main OLD_MAIN
-#include <libs/config/test/config_info.cpp>
+#include <config_info.cpp>
 #undef main
 #ifndef NEW_MAIN
 #  define NEW_MAIN main

test/issue227.cpp

@@ -0,0 +1,19 @@
+/*
+* Copyright (c) 2024
+* Christian Mazakas
+*
+* 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)
+*
+*/
+
+#include <boost/regex.hpp>
+#include <string>
+
+int main() {
+   boost::regex rx("(*ACCEPT)*+\\1((*ACCEPT)*+\\K)");
+   std::string str = "Z";
+   boost::smatch what;
+   boost::regex_search(str, what, rx, boost::match_default | boost::match_partial);
+}

test/issue232.cpp

@@ -0,0 +1,71 @@
+#include <boost/core/lightweight_test.hpp>
+
+#include <boost/regex.hpp>
+#include <cstddef>
+#include <vector>
+
+template<std::size_t N0, std::size_t N = N0 - 1>
+void tester( char const (&str)[ N0 ] )
+{
+   std::vector<char> s(N, '\0');
+   std::memcpy(s.data(), str, N);
+   boost::regex rx(s.begin(), s.end());
+
+   std::vector<std::string> wheres;
+   wheres.push_back(std::string(15, 'H'));
+   wheres.push_back("");
+   wheres.push_back("                         ");
+
+   // Perl-style matching
+   for (auto const& where : wheres) {
+      boost::match_results<std::string::const_iterator> what;
+      bool match = boost::regex_match(where, what, rx, boost::match_default | boost::match_partial | boost::match_any | boost::match_perl);
+      (void) match;
+   }
+
+   // POSIX-style matching
+   for (auto const& where : wheres) {
+      try {
+         boost::match_results<std::string::const_iterator> what;
+         bool match = boost::regex_match(where, what, rx, boost::match_default | boost::match_partial | boost::match_any | boost::match_posix);
+         (void) match;
+      } catch(...) {}
+   }
+
+}
+
+int main()
+{
+   // test strings derived from fuzzing
+   // we keep a simple human-readable version
+   char const str1[] = "(Y(*COMMIT)|\\K\\D|.)+";
+   char const str2[] = "(Y(*COMMIT){||\\K\\D|||||||||\\K|||ss|||||.|\232*(?(50)\027\0204657|H)\020}\031\000.* 6.'?-i)+[L??.\000\000\000\004\000\000\000\000?..<[\000\024R]*+";
+   char const str3[] = "(Y(*COMMIT)\xFF\x80|\\L\\K||||||||||.|||||\x84|||||\x00\x00\x10||||||.*  .'?-i)[L??...-i)[L??...[\x00\x14R]*+";
+   char const str4[] = "(Y(*COMMIT)\x96||.*  .*  .\\K|||\x9F||||\x9C|.|||||\x84\x99|||\x01\x00\x00\x00|||'?-i#PL\x00\x01.\x86??OMMIT)?...[\x00\x14R]*+";
+
+   tester(str1);
+   tester(str2);
+   tester(str3);
+   tester(str4);
+
+   // prove that we catch certain impossible scenarios
+
+   {
+      char const* str = "abcd";
+      boost::regex rx(str);
+      boost::match_results<std::string::const_iterator> what;
+      std::string where(15, 'H');
+      BOOST_TEST_THROWS(boost::regex_match(where, rx, boost::match_posix | boost::match_perl), std::logic_error);
+   }
+
+   {
+      char const* str = "ab(*COMMIT)cd";
+      boost::regex rx(str);
+      boost::match_results<std::string::const_iterator> what;
+      std::string where(15, 'H');
+      BOOST_TEST_THROWS(boost::regex_match(where, rx, boost::match_posix), std::logic_error);
+   }
+
+   return boost::report_errors();
+}
+

test/lookbehind_recursion_stress_test.cpp

@@ -0,0 +1,28 @@
+/*
+ *
+ * Copyright (c) 2024
+ * 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)
+ *
+ */
+
+#include <boost/regex.hpp>
+
+int main()
+{
+   std::string s("(?<=(");
+   s.append(1000, '|');
+   s += "))";
+   boost::regex rx(s);
+
+   s = "(?<=(a";
+   for (unsigned i = 0; i < 1000; ++i)
+   {
+      s += "|a";
+   }
+   s += "))";
+   boost::regex rx2(s);
+}

test/test_consolidated.cpp

@@ -3,12 +3,12 @@
  * Copyright (c) 2011
  * John Maddock
  *
- * Use, modification and distribution are subject to the 
- * Boost Software License, Version 1.0. (See accompanying file 
+ * 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)
  *
  */
 
 
-#include <libs/regex/src/posix_api.cpp>
-#include <libs/regex/src/wide_posix_api.cpp>
+#include <posix_api.cpp>
+#include <wide_posix_api.cpp>