xenia/home
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Optimize Regex CharacterClass matching

Browse Source 
Take advantage of ranges sorting to early out, make the logic
inline.
This commit is contained in:
Maxime Coste 2023-02-19 11:15:31 +11:00
parent afaa47e93f
commit f115af7a57
2 changed files with 22 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
src/regex_impl.cc
View File

@ -959,7 +959,7 @@ private:
{ {
for (Codepoint cp = 0; cp < CompiledRegex::StartDesc::count; ++cp) for (Codepoint cp = 0; cp < CompiledRegex::StartDesc::count; ++cp)
{ {
if (start_desc.map[cp] or is_character_class(character_class, cp)) if (start_desc.map[cp] or character_class.matches(cp))
start_desc.map[cp] = true; start_desc.map[cp] = true;
} }
} }
@ -1165,20 +1165,6 @@ CompiledRegex compile_regex(StringView re, RegexCompileFlags flags)
return RegexCompiler{RegexParser::parse(re), flags}.get_compiled_regex(); return RegexCompiler{RegexParser::parse(re), flags}.get_compiled_regex();
} }
bool is_character_class(const CharacterClass& character_class, Codepoint cp)
{
if (character_class.ignore_case)
cp = to_lower(cp);
auto it = std::find_if(character_class.ranges.begin(),
character_class.ranges.end(),
[cp](auto& range) { return range.min <= cp and cp <= range.max; });
bool found = it != character_class.ranges.end() or (character_class.ctypes != CharacterType::None and
is_ctype(character_class.ctypes, cp));
return found != character_class.negative;
}
bool is_ctype(CharacterType ctype, Codepoint cp) bool is_ctype(CharacterType ctype, Codepoint cp)
{ {
auto check = [&](CharacterType bit, CharacterType not_bit, auto&& func) { auto check = [&](CharacterType bit, CharacterType not_bit, auto&& func) {

26
src/regex_impl.hh
View File

@ -31,6 +31,8 @@ enum class CharacterType : unsigned char
}; };
constexpr bool with_bit_ops(Meta::Type<CharacterType>) { return true; } constexpr bool with_bit_ops(Meta::Type<CharacterType>) { return true; }
bool is_ctype(CharacterType ctype, Codepoint cp);
struct CharacterClass struct CharacterClass
{ {
struct Range struct Range
@ -45,10 +47,24 @@ struct CharacterClass
bool ignore_case = false; bool ignore_case = false;
friend bool operator==(const CharacterClass&, const CharacterClass&) = default; friend bool operator==(const CharacterClass&, const CharacterClass&) = default;
};
bool is_character_class(const CharacterClass& character_class, Codepoint cp); bool matches(Codepoint cp) const
bool is_ctype(CharacterType ctype, Codepoint cp); {
if (ignore_case)
cp = to_lower(cp);
for (auto& range : ranges)
{
if (cp < range.min)
break;
else if (cp <= range.max)
return not negative;
}
return (ctypes != CharacterType::None and is_ctype(ctypes, cp)) != negative;
}
};
struct CompiledRegex : RefCountable, UseMemoryDomain<MemoryDomain::Regex> struct CompiledRegex : RefCountable, UseMemoryDomain<MemoryDomain::Regex>
{ {
@ -418,7 +434,7 @@ private:
case CompiledRegex::CharClass: case CompiledRegex::CharClass:
if (pos == config.end) if (pos == config.end)
return failed(); return failed();
return is_character_class(m_program.character_classes[inst.param.character_class_index], codepoint(pos, config)) ? return m_program.character_classes[inst.param.character_class_index].matches(codepoint(pos, config)) ?
consumed() : failed(); consumed() : failed();
case CompiledRegex::CharType: case CompiledRegex::CharType:
if (pos == config.end) if (pos == config.end)
@ -552,7 +568,7 @@ private:
else if (op >= Lookaround::CharacterClass and op < Lookaround::CharacterType) else if (op >= Lookaround::CharacterClass and op < Lookaround::CharacterType)
{ {
auto index = to_underlying(op) - to_underlying(Lookaround::CharacterClass); auto index = to_underlying(op) - to_underlying(Lookaround::CharacterClass);
if (not is_character_class(m_program.character_classes[index], cp)) if (not m_program.character_classes[index].matches(cp))
return false; return false;
} }
else if (op >= Lookaround::CharacterType and op < Lookaround::OpEnd) else if (op >= Lookaround::CharacterType and op < Lookaround::OpEnd)