fergalmoran/ladybird
1
0
Fork 0
mirror of https://github.com/fergalmoran/ladybird.git synced 2025-12-22 09:19:03 +00:00
Code Issues Packages Projects Releases Wiki Activity

LibRegex: Use depth-first search in regex optimizer

Browse Source 
use depth-first search in optimizer code bacause using breadth-first
search generate a bug. Add test example in test lib.
This commit is contained in:
mikiubo
2025-02-24 08:41:04 +01:00
committed by Ali Mohammad Pur
parent 2797f9f73e
commit 8a6f7b787e
2 changed files with 6 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
Libraries/LibRegex/RegexOptimizer.cpp
View File

@@ -6,11 +6,11 @@
#include <AK/Debug.h> #include <AK/Debug.h>
#include <AK/Function.h> #include <AK/Function.h>
#include <AK/Queue.h>
#include <AK/QuickSort.h> #include <AK/QuickSort.h>
#include <AK/RedBlackTree.h> #include <AK/RedBlackTree.h>
#include <AK/Stack.h> #include <AK/Stack.h>
#include <AK/Trie.h> #include <AK/Trie.h>
#include <AK/Vector.h>
#include <LibRegex/Regex.h> #include <LibRegex/Regex.h>
#include <LibRegex/RegexBytecodeStreamOptimizer.h> #include <LibRegex/RegexBytecodeStreamOptimizer.h>
#include <LibUnicode/CharacterTypes.h> #include <LibUnicode/CharacterTypes.h>
@@ -1176,8 +1176,8 @@ void Optimizer::append_alternation(ByteCode& target, Span<ByteCode> alternatives
patch_locations.append({ node_ip, target_ip }); patch_locations.append({ node_ip, target_ip });
}; };
Queue<Tree*> nodes_to_visit; Vector<Tree*> nodes_to_visit;
nodes_to_visit.enqueue(&trie); nodes_to_visit.append(&trie);
HashMap<size_t, NonnullOwnPtr<RedBlackTree<u64, u64>>> instruction_positions; HashMap<size_t, NonnullOwnPtr<RedBlackTree<u64, u64>>> instruction_positions;
if (has_any_backwards_jump) if (has_any_backwards_jump)
@@ -1195,7 +1195,7 @@ void Optimizer::append_alternation(ByteCode& target, Span<ByteCode> alternatives
// forkjump child2 // forkjump child2
// ... // ...
while (!nodes_to_visit.is_empty()) { while (!nodes_to_visit.is_empty()) {
auto const* node = nodes_to_visit.dequeue(); auto const* node = nodes_to_visit.take_last();
for (auto& patch : patch_locations) { for (auto& patch : patch_locations) {
if (!patch.done && node_is(node, patch.source_ip)) { if (!patch.done && node_is(node, patch.source_ip)) {
auto value = static_cast<ByteCodeValueType>(target.size() - patch.target_ip - 1); auto value = static_cast<ByteCodeValueType>(target.size() - patch.target_ip - 1);
@@ -1291,7 +1291,7 @@ void Optimizer::append_alternation(ByteCode& target, Span<ByteCode> alternatives
target.append(static_cast<ByteCodeValueType>(OpCodeId::ForkJump)); target.append(static_cast<ByteCodeValueType>(OpCodeId::ForkJump));
add_patch_point(child_node, target.size()); add_patch_point(child_node, target.size());
target.append(static_cast<ByteCodeValueType>(0)); target.append(static_cast<ByteCodeValueType>(0));
nodes_to_visit.enqueue(child_node); nodes_to_visit.append(child_node);
} }
} }

1
Tests/LibRegex/Regex.cpp
View File

@@ -716,6 +716,7 @@ TEST_CASE(ECMA262_match)
""sv, ""sv,
false, }, // See above, also ladybird#2931. false, }, // See above, also ladybird#2931.
{ "[^]*[^]"sv, "i"sv, true }, // Optimizer bug, ignoring an enabled trailing 'invert' when comparing blocks, ladybird#3421. { "[^]*[^]"sv, "i"sv, true }, // Optimizer bug, ignoring an enabled trailing 'invert' when comparing blocks, ladybird#3421.
{ "xx|...|...."sv, "cd"sv, false },
}; };
// clang-format on // clang-format on