boostorg/regex
1
0
Fork 1
mirror of https://github.com/boostorg/regex.git synced 2025-07-23 17:17:22 +02:00
Code Issues Packages Projects Releases Wiki Activity

de-fuzz: improve set creation so as not to allow duplicate characters.

Browse Source
This commit is contained in:
jzmaddock
2017-02-14 19:41:35 +00:00
parent febd44ddd6
commit b05fafe1c5
2 changed files with 41 additions and 35 deletions
Download Patch File Download Diff File Expand all files Collapse all files

73
include/boost/regex/v4/basic_regex_creator.hpp
View File

@ -77,7 +77,7 @@ public:
void add_single(const digraph_type& s) void add_single(const digraph_type& s)
{ {
m_singles.insert(m_singles.end(), s); m_singles.insert(s);
if(s.second) if(s.second)
m_has_digraphs = true; m_has_digraphs = true;
m_empty = false; m_empty = false;
@ -136,11 +136,12 @@ public:
return m_negate; return m_negate;
} }
typedef typename std::vector<digraph_type>::const_iterator list_iterator; typedef typename std::vector<digraph_type>::const_iterator list_iterator;
list_iterator singles_begin()const typedef typename std::set<digraph_type>::const_iterator set_iterator;
set_iterator singles_begin()const
{ {
return m_singles.begin(); return m_singles.begin();
} }
list_iterator singles_end()const set_iterator singles_end()const
{ {
return m_singles.end(); return m_singles.end();
} }
@ -152,11 +153,11 @@ public:
{ {
return m_ranges.end(); return m_ranges.end();
} }
list_iterator equivalents_begin()const set_iterator equivalents_begin()const
{ {
return m_equivalents.begin(); return m_equivalents.begin();
} }
list_iterator equivalents_end()const set_iterator equivalents_end()const
{ {
return m_equivalents.end(); return m_equivalents.end();
} }
@ -173,14 +174,14 @@ public:
return m_empty; return m_empty;
} }
private: private:
std::vector<digraph_type> m_singles; // a list of single characters to match std::set<digraph_type> m_singles; // a list of single characters to match
std::vector<digraph_type> m_ranges; // a list of end points of our ranges std::vector<digraph_type> m_ranges; // a list of end points of our ranges
bool m_negate; // true if the set is to be negated bool m_negate; // true if the set is to be negated
bool m_has_digraphs; // true if we have digraphs present bool m_has_digraphs; // true if we have digraphs present
m_type m_classes; // character classes to match m_type m_classes; // character classes to match
m_type m_negated_classes; // negated character classes to match m_type m_negated_classes; // negated character classes to match
bool m_empty; // whether we've added anything yet bool m_empty; // whether we've added anything yet
std::vector<digraph_type> m_equivalents; // a list of equivalence classes std::set<digraph_type> m_equivalents; // a list of equivalence classes
}; };
template <class charT, class traits> template <class charT, class traits>
@ -365,6 +366,7 @@ re_syntax_base* basic_regex_creator<charT, traits>::append_set(
{ {
typedef typename traits::string_type string_type; typedef typename traits::string_type string_type;
typedef typename basic_char_set<charT, traits>::list_iterator item_iterator; typedef typename basic_char_set<charT, traits>::list_iterator item_iterator;
typedef typename basic_char_set<charT, traits>::set_iterator set_iterator;
typedef typename traits::char_class_type m_type; typedef typename traits::char_class_type m_type;
re_set_long<m_type>* result = static_cast<re_set_long<m_type>*>(append_state(syntax_element_long_set, sizeof(re_set_long<m_type>))); re_set_long<m_type>* result = static_cast<re_set_long<m_type>*>(append_state(syntax_element_long_set, sizeof(re_set_long<m_type>)));
@ -395,24 +397,25 @@ re_syntax_base* basic_regex_creator<charT, traits>::append_set(
// now extend with all the singles: // now extend with all the singles:
// //
item_iterator first, last; item_iterator first, last;
first = char_set.singles_begin(); set_iterator sfirst, slast;
last = char_set.singles_end(); sfirst = char_set.singles_begin();
while(first != last) slast = char_set.singles_end();
while(sfirst != slast)
{ {
charT* p = static_cast<charT*>(this->m_pdata->m_data.extend(sizeof(charT) * (first->first ? 1 : first->second ? 3 : 2))); charT* p = static_cast<charT*>(this->m_pdata->m_data.extend(sizeof(charT) * (sfirst->first == static_cast<charT>(0) ? 1 : sfirst->second ? 3 : 2)));
p[0] = m_traits.translate(first->first, m_icase); p[0] = m_traits.translate(sfirst->first, m_icase);
if(first->first) if(sfirst->first == static_cast<charT>(0))
{ {
p[0] = 0; p[0] = 0;
} }
else if(first->second) else if(sfirst->second)
{ {
p[1] = m_traits.translate(first->second, m_icase); p[1] = m_traits.translate(sfirst->second, m_icase);
p[2] = 0; p[2] = 0;
} }
else else
p[1] = 0; p[1] = 0;
++first; ++sfirst;
} }
// //
// now extend with all the ranges: // now extend with all the ranges:
@ -476,24 +479,24 @@ re_syntax_base* basic_regex_creator<charT, traits>::append_set(
// //
// now process the equivalence classes: // now process the equivalence classes:
// //
first = char_set.equivalents_begin(); sfirst = char_set.equivalents_begin();
last = char_set.equivalents_end(); slast = char_set.equivalents_end();
while(first != last) while(sfirst != slast)
{ {
string_type s; string_type s;
if(first->second) if(sfirst->second)
{ {
charT cs[3] = { first->first, first->second, charT(0), }; charT cs[3] = { sfirst->first, sfirst->second, charT(0), };
s = m_traits.transform_primary(cs, cs+2); s = m_traits.transform_primary(cs, cs+2);
} }
else else
s = m_traits.transform_primary(&first->first, &first->first+1); s = m_traits.transform_primary(&sfirst->first, &sfirst->first+1);
if(s.empty()) if(s.empty())
return 0; // invalid or unsupported equivalence class return 0; // invalid or unsupported equivalence class
charT* p = static_cast<charT*>(this->m_pdata->m_data.extend(sizeof(charT) * (s.size()+1) ) ); charT* p = static_cast<charT*>(this->m_pdata->m_data.extend(sizeof(charT) * (s.size()+1) ) );
BOOST_REGEX_DETAIL_NS::copy(s.begin(), s.end(), p); BOOST_REGEX_DETAIL_NS::copy(s.begin(), s.end(), p);
p[s.size()] = charT(0); p[s.size()] = charT(0);
++first; ++sfirst;
} }
// //
// finally reset the address of our last state: // finally reset the address of our last state:
@ -522,6 +525,7 @@ re_syntax_base* basic_regex_creator<charT, traits>::append_set(
{ {
typedef typename traits::string_type string_type; typedef typename traits::string_type string_type;
typedef typename basic_char_set<charT, traits>::list_iterator item_iterator; typedef typename basic_char_set<charT, traits>::list_iterator item_iterator;
typedef typename basic_char_set<charT, traits>::set_iterator set_iterator;
re_set* result = static_cast<re_set*>(append_state(syntax_element_set, sizeof(re_set))); re_set* result = static_cast<re_set*>(append_state(syntax_element_set, sizeof(re_set)));
bool negate = char_set.is_negated(); bool negate = char_set.is_negated();
@ -530,17 +534,18 @@ re_syntax_base* basic_regex_creator<charT, traits>::append_set(
// handle singles first: // handle singles first:
// //
item_iterator first, last; item_iterator first, last;
first = char_set.singles_begin(); set_iterator sfirst, slast;
last = char_set.singles_end(); sfirst = char_set.singles_begin();
while(first != last) slast = char_set.singles_end();
while(sfirst != slast)
{ {
for(unsigned int i = 0; i < (1 << CHAR_BIT); ++i) for(unsigned int i = 0; i < (1 << CHAR_BIT); ++i)
{ {
if(this->m_traits.translate(static_cast<charT>(i), this->m_icase) if(this->m_traits.translate(static_cast<charT>(i), this->m_icase)
== this->m_traits.translate(first->first, this->m_icase)) == this->m_traits.translate(sfirst->first, this->m_icase))
result->_map[i] = true; result->_map[i] = true;
} }
++first; ++sfirst;
} }
// //
// OK now handle ranges: // OK now handle ranges:
@ -627,13 +632,13 @@ re_syntax_base* basic_regex_creator<charT, traits>::append_set(
// //
// now process the equivalence classes: // now process the equivalence classes:
// //
first = char_set.equivalents_begin(); sfirst = char_set.equivalents_begin();
last = char_set.equivalents_end(); slast = char_set.equivalents_end();
while(first != last) while(sfirst != slast)
{ {
string_type s; string_type s;
BOOST_ASSERT(static_cast<charT>(0) == first->second); BOOST_ASSERT(static_cast<charT>(0) == sfirst->second);
s = m_traits.transform_primary(&first->first, &first->first+1); s = m_traits.transform_primary(&sfirst->first, &sfirst->first+1);
if(s.empty()) if(s.empty())
return 0; // invalid or unsupported equivalence class return 0; // invalid or unsupported equivalence class
for(unsigned i = 0; i < (1u << CHAR_BIT); ++i) for(unsigned i = 0; i < (1u << CHAR_BIT); ++i)
@ -643,7 +648,7 @@ re_syntax_base* basic_regex_creator<charT, traits>::append_set(
if(s == s2) if(s == s2)
result->_map[i] = true; result->_map[i] = true;
} }
++first; ++sfirst;
} }
if(negate) if(negate)
{ {

1
include/boost/regex/v4/regex_workaround.hpp
View File

@ -33,6 +33,7 @@
#include <algorithm> #include <algorithm>
#include <iosfwd> #include <iosfwd>
#include <vector> #include <vector>
#include <set>
#include <map> #include <map>
#include <boost/limits.hpp> #include <boost/limits.hpp>
#include <boost/assert.hpp> #include <boost/assert.hpp>