beenull/ladybird
1
0
Fork 0
mirror of https://github.com/LadybirdBrowser/ladybird.git synced 2024-11-21 23:20:20 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 6

Merge e383f52d04 into 7f989765f5

Browse source
This commit is contained in:
Ali Mohammad Pur 2024-11-21 08:00:13 -05:00 committed by GitHub
commit c98baaa1d3
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 25 additions and 12 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

8
Libraries/LibRegex/RegexByteCode.h
View file

@ -791,7 +791,13 @@ public:
ByteString arguments_string() const override ByteString arguments_string() const override
{ {
auto reps = id() < state().repetition_marks.size() ? state().repetition_marks.at(id()) : 0; auto reps = id() < state().repetition_marks.size() ? state().repetition_marks.at(id()) : 0;
return ByteString::formatted("offset={} count={} id={} rep={}, sp: {}", offset(), count() + 1, id(), reps + 1, state().string_position); return ByteString::formatted("offset={} [&{}] count={} id={} rep={}, sp: {}",
static_cast<ssize_t>(offset()),
state().instruction_position - offset(),
count() + 1,
id(),
reps + 1,
state().string_position);
} }
}; };

1
Libraries/LibRegex/RegexMatcher.h
View file

@ -28,6 +28,7 @@ namespace Detail {
struct Block { struct Block {
size_t start; size_t start;
size_t end; size_t end;
StringView comment { "N/A"sv };
}; };
} }

28
Libraries/LibRegex/RegexOptimizer.cpp
View file

@ -59,18 +59,18 @@ typename Regex<Parser>::BasicBlockList Regex<Parser>::split_basic_blocks(ByteCod
auto& op = static_cast<T const&>(opcode); auto& op = static_cast<T const&>(opcode);
ssize_t jump_offset = op.size() + op.offset(); ssize_t jump_offset = op.size() + op.offset();
if (jump_offset >= 0) { if (jump_offset >= 0) {
block_boundaries.append({ end_of_last_block, state.instruction_position }); block_boundaries.append({ end_of_last_block, state.instruction_position, "Jump ahead"sv });
end_of_last_block = state.instruction_position + opcode.size(); end_of_last_block = state.instruction_position + opcode.size();
} else { } else {
// This op jumps back, see if that's within this "block". // This op jumps back, see if that's within this "block".
if (jump_offset + state.instruction_position > end_of_last_block) { if (jump_offset + state.instruction_position > end_of_last_block) {
// Split the block! // Split the block!
block_boundaries.append({ end_of_last_block, jump_offset + state.instruction_position }); block_boundaries.append({ end_of_last_block, jump_offset + state.instruction_position, "Jump back 1"sv });
block_boundaries.append({ jump_offset + state.instruction_position, state.instruction_position }); block_boundaries.append({ jump_offset + state.instruction_position, state.instruction_position, "Jump back 2"sv });
end_of_last_block = state.instruction_position + opcode.size(); end_of_last_block = state.instruction_position + opcode.size();
} else { } else {
// Nope, it's just a jump to another block // Nope, it's just a jump to another block
block_boundaries.append({ end_of_last_block, state.instruction_position }); block_boundaries.append({ end_of_last_block, state.instruction_position, "Jump"sv });
end_of_last_block = state.instruction_position + opcode.size(); end_of_last_block = state.instruction_position + opcode.size();
} }
} }
@ -92,15 +92,16 @@ typename Regex<Parser>::BasicBlockList Regex<Parser>::split_basic_blocks(ByteCod
check_jump.template operator()<OpCode_ForkStay>(opcode); check_jump.template operator()<OpCode_ForkStay>(opcode);
break; break;
case OpCodeId::FailForks: case OpCodeId::FailForks:
block_boundaries.append({ end_of_last_block, state.instruction_position }); block_boundaries.append({ end_of_last_block, state.instruction_position, "FailForks"sv });
end_of_last_block = state.instruction_position + opcode.size(); end_of_last_block = state.instruction_position + opcode.size();
break; break;
case OpCodeId::Repeat: { case OpCodeId::Repeat: {
// Repeat produces two blocks, one containing its repeated expr, and one after that. // Repeat produces two blocks, one containing its repeated expr, and one after that.
auto repeat_start = state.instruction_position - static_cast<OpCode_Repeat const&>(opcode).offset(); auto& repeat = static_cast<OpCode_Repeat const&>(opcode);
auto repeat_start = state.instruction_position - repeat.offset() - repeat.size();
if (repeat_start > end_of_last_block) if (repeat_start > end_of_last_block)
block_boundaries.append({ end_of_last_block, repeat_start }); block_boundaries.append({ end_of_last_block, repeat_start, "Repeat"sv });
block_boundaries.append({ repeat_start, state.instruction_position }); block_boundaries.append({ repeat_start, state.instruction_position, "Repeat after"sv });
end_of_last_block = state.instruction_position + opcode.size(); end_of_last_block = state.instruction_position + opcode.size();
break; break;
} }
@ -116,7 +117,7 @@ typename Regex<Parser>::BasicBlockList Regex<Parser>::split_basic_blocks(ByteCod
} }
if (end_of_last_block < bytecode_size) if (end_of_last_block < bytecode_size)
block_boundaries.append({ end_of_last_block, bytecode_size }); block_boundaries.append({ end_of_last_block, bytecode_size, "End"sv });
quick_sort(block_boundaries, [](auto& a, auto& b) { return a.start < b.start; }); quick_sort(block_boundaries, [](auto& a, auto& b) { return a.start < b.start; });
@ -664,7 +665,7 @@ void Regex<Parser>::attempt_rewrite_loops_as_atomic_groups(BasicBlockList const&
RegexDebug dbg; RegexDebug dbg;
dbg.print_bytecode(*this); dbg.print_bytecode(*this);
for (auto const& block : basic_blocks) for (auto const& block : basic_blocks)
dbgln("block from {} to {}", block.start, block.end); dbgln("block from {} to {} (comment: {})", block.start, block.end, block.comment);
} }
// A pattern such as: // A pattern such as:
@ -1221,6 +1222,7 @@ void Optimizer::append_alternation(ByteCode& target, Span<ByteCode> alternatives
ssize_t jump_offset; ssize_t jump_offset;
auto is_jump = true; auto is_jump = true;
auto patch_location = state.instruction_position + 1; auto patch_location = state.instruction_position + 1;
bool should_negate = false;
switch (opcode.opcode_id()) { switch (opcode.opcode_id()) {
case OpCodeId::Jump: case OpCodeId::Jump:
@ -1243,6 +1245,7 @@ void Optimizer::append_alternation(ByteCode& target, Span<ByteCode> alternatives
break; break;
case OpCodeId::Repeat: case OpCodeId::Repeat:
jump_offset = static_cast<ssize_t>(0) - static_cast<ssize_t>(static_cast<OpCode_Repeat const&>(opcode).offset()) - static_cast<ssize_t>(opcode.size()); jump_offset = static_cast<ssize_t>(0) - static_cast<ssize_t>(static_cast<OpCode_Repeat const&>(opcode).offset()) - static_cast<ssize_t>(opcode.size());
should_negate = true;
break; break;
default: default:
is_jump = false; is_jump = false;
@ -1272,7 +1275,10 @@ void Optimizer::append_alternation(ByteCode& target, Span<ByteCode> alternatives
intended_jump_ip); intended_jump_ip);
VERIFY_NOT_REACHED(); VERIFY_NOT_REACHED();
} }
target[patch_location] = static_cast<ByteCodeValueType>(*target_ip - patch_location - 1); ssize_t target_value = *target_ip - patch_location - 1;
if (should_negate)
target_value = -target_value + 2; // from -1 to +1.
target[patch_location] = static_cast<ByteCodeValueType>(target_value);
} else { } else {
patch_locations.append({ QualifiedIP { ip.alternative_index, intended_jump_ip }, patch_location }); patch_locations.append({ QualifiedIP { ip.alternative_index, intended_jump_ip }, patch_location });
} }