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

Regex: Support any char and character classes in lookarounds

Browse Source 
Lookarounds still need to be fixed size, but accept character classes
as well as plain literals.
This commit is contained in:
Maxime Coste 2017-10-09 11:12:42 +08:00
parent b8cb65160a
commit cca730193c
2 changed files with 46 additions and 23 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

33
src/regex_impl.cc
View File

@ -474,8 +474,13 @@ private:
void validate_lookaround(const AstNodePtr& node)
{
for (auto& child : node->children)
if (child->op != ParsedRegex::Literal)
parse_error("Lookaround can only contain literals");
{
if (child->op != ParsedRegex::Literal and child->op != ParsedRegex::Matcher and
child->op != ParsedRegex::AnyChar)
parse_error("Lookaround can only contain literals, any chars or character classes");
if (child->quantifier.type != ParsedRegex::Quantifier::One)
parse_error("Quantifiers cannot be used in lookarounds");
}
}
ParsedRegex m_parsed_regex;
@ -679,15 +684,27 @@ private:
return res;
}
uint32_t push_lookaround(const Vector<ParsedRegex::AstNodePtr>& literals, bool reversed = false)
uint32_t push_lookaround(const Vector<ParsedRegex::AstNodePtr>& characters, bool reversed = false)
{
uint32_t res = m_program.lookarounds.size();
if (reversed)
for (auto& literal : literals | reverse())
m_program.lookarounds.push_back(literal->value);
auto write_lookaround = [this](auto&& characters) {
for (auto& character : characters)
{
if (character->op == ParsedRegex::Literal)
m_program.lookarounds.push_back(character->value);
else if (character->op == ParsedRegex::AnyChar)
m_program.lookarounds.push_back(0xF000);
else if (character->op == ParsedRegex::Matcher)
m_program.lookarounds.push_back(0xF0001 + character->value);
else
for (auto& literal : literals)
m_program.lookarounds.push_back(literal->value);
kak_assert(false);
}
};
if (reversed)
write_lookaround(characters | reverse());
else
write_lookaround(characters);
m_program.lookarounds.push_back((Codepoint)-1);
return res;

38
src/regex_impl.hh
View File

@ -294,28 +294,14 @@ private:
break;
case CompiledRegex::LookAhead:
case CompiledRegex::NegativeLookAhead:
{
auto ref = m_program.lookarounds.begin() + inst.param;
for (auto it = pos; *ref != -1 and it != m_end; ++it, ++ref)
if (*it != *ref)
break;
if ((inst.op == CompiledRegex::LookAhead and *ref != -1) or
(inst.op == CompiledRegex::NegativeLookAhead and *ref == -1))
if (lookaround<MatchDirection::Forward>(inst.param, pos) != (inst.op == CompiledRegex::LookAhead))
return StepResult::Failed;
break;
}
case CompiledRegex::LookBehind:
case CompiledRegex::NegativeLookBehind:
{
auto ref = m_program.lookarounds.begin() + inst.param;
for (auto it = pos; *ref != -1 and it > m_begin; --it, ++ref)
if (*(it-1) != *ref)
break;
if ((inst.op == CompiledRegex::LookBehind and *ref != -1) or
(inst.op == CompiledRegex::NegativeLookBehind and *ref == -1))
if (lookaround<MatchDirection::Backward>(inst.param, pos) != (inst.op == CompiledRegex::LookBehind))
return StepResult::Failed;
break;
}
case CompiledRegex::Match:
return StepResult::Matched;
}
@ -392,6 +378,26 @@ private:
++start;
}
template<MatchDirection look_direction>
bool lookaround(uint32_t index, Utf8It pos) const
{
for (auto it = m_program.lookarounds.begin() + index; *it != -1; ++it)
{
if (pos == (look_direction == MatchDirection::Forward ? m_end : m_begin))
return false;
auto cp = (look_direction == MatchDirection::Forward ? *pos : *(pos-1)), ref = *it;
if (ref == 0xF000)
{} // any character matches
else if (ref > 0xF0000 and ref <= 0xFFFFD and not m_program.matchers[ref - 0xF0001](cp))
return false;
else if (ref != cp)
return false;
(look_direction == MatchDirection::Forward) ? ++pos : --pos;
}
return true;
}
bool is_line_start(const Utf8It& pos) const
{
if (not (m_flags & RegexExecFlags::PrevAvailable) and pos == m_begin)