1
0
Fork 0
mirror of https://github.com/LadybirdBrowser/ladybird.git synced 2025-06-08 05:27:14 +09:00

LibRegex: Pick the right target for OpCode_Repeat

Repeat's 'offset' field is a bit odd in that it is treated as a negative
offset, causing a backwards jump when positive; the optimizer didn't
correctly model this behaviour, which caused crashes and misopts when
dealing with Repeats.
This commit fixes that behaviour.
This commit is contained in:
Ali Mohammad Pur 2024-12-11 10:56:32 +01:00 committed by Ali Mohammad Pur
parent 4a8d3e35a3
commit 358378c1c0
Notes: github-actions[bot] 2024-12-13 09:01:22 +00:00

View file

@ -97,7 +97,8 @@ typename Regex<Parser>::BasicBlockList Regex<Parser>::split_basic_blocks(ByteCod
break;
case OpCodeId::Repeat: {
// 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)
block_boundaries.append({ end_of_last_block, repeat_start, "Repeat"sv });
block_boundaries.append({ repeat_start, state.instruction_position, "Repeat after"sv });
@ -1221,6 +1222,7 @@ void Optimizer::append_alternation(ByteCode& target, Span<ByteCode> alternatives
ssize_t jump_offset;
auto is_jump = true;
auto patch_location = state.instruction_position + 1;
bool should_negate = false;
switch (opcode.opcode_id()) {
case OpCodeId::Jump:
@ -1243,6 +1245,7 @@ void Optimizer::append_alternation(ByteCode& target, Span<ByteCode> alternatives
break;
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());
should_negate = true;
break;
default:
is_jump = false;
@ -1272,7 +1275,10 @@ void Optimizer::append_alternation(ByteCode& target, Span<ByteCode> alternatives
intended_jump_ip);
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 {
patch_locations.append({ QualifiedIP { ip.alternative_index, intended_jump_ip }, patch_location });
}