RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-25 14:27:35 +00:00
Code Issues Activity Actions

LibRegex: Flatten bytecode before performing optimizations

Browse source 
This avoids doing DisjointChunks traversal for every bytecode access,
significantly reducing startup time for large regular expressions.
This commit is contained in:
Andreas Kling 2021-09-29 18:43:58 +02:00
parent 7bdf0be667
commit 2758d99bbc
1 changed files with 2 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

2
Userland/Libraries/LibRegex/RegexOptimizer.cpp
View file

@ -17,6 +17,8 @@ using Detail::Block;
template<typename Parser> template<typename Parser>
void Regex<Parser>::run_optimization_passes() void Regex<Parser>::run_optimization_passes()
{ {
parser_result.bytecode.flatten();
// Rewrite fork loops as atomic groups // Rewrite fork loops as atomic groups
// e.g. a*b -> (ATOMIC a*)b // e.g. a*b -> (ATOMIC a*)b
attempt_rewrite_loops_as_atomic_groups(split_basic_blocks()); attempt_rewrite_loops_as_atomic_groups(split_basic_blocks());