kyle/kte
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Highlighter engine returns cached spans by reference; drop unneeded mutexes

Browse Source 
HighlighterEngine::GetLine copied the span vector on every call (twice
on a miss) and did a dynamic_cast per call; it now returns the cached
entry (valid until the next call) and dispatches on Stateful().

PieceTable and HighlighterEngine took a mutex in their read paths while
their writes took none, and PieceTable.h claimed concurrent reads were
safe; they are not (reads update caches). Every Buffer is used from the
main thread only (the swap writer thread receives copies; the regex
large-stack thread runs while its caller waits), so the mutexes are
removed and the headers say so. ASan and the threaded tests under TSan
are clean.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015rcVQiiWyTzoNrsvLSaAfB
This commit is contained in:
Claude
2026-09-24 12:19:59 -07:00
committed by Kyle Isom
parent 7039c14507
commit d490245f30
8 changed files with 43 additions and 73 deletions
Download Patch File Download Diff File Expand all files Collapse all files
PieceTable.cc
-10
View File
@@ -321,7 +321,6 @@ PieceTable::addPieceFront(Source src, std::size_t start, std::size_t len)
void
PieceTable::materialize() const
{
std::lock_guard<std::mutex> lock(mutex_);
if (!dirty_) {
return;
}
@@ -401,7 +400,6 @@ PieceTable::coalesceNeighbors(std::size_t index)
void
PieceTable::InvalidateLineIndex() const
{
std::lock_guard<std::mutex> lock(mutex_);
line_index_dirty_ = true;
}
@@ -435,7 +433,6 @@ PieceTable::applyLineShift() const
void
PieceTable::lineIndexOnInsert(std::size_t offset, const char *text, std::size_t len) const
{
std::lock_guard<std::mutex> lock(mutex_);
if (line_index_dirty_)
return;
// Line starts after the insertion point move right; a start exactly at
@@ -461,7 +458,6 @@ PieceTable::lineIndexOnInsert(std::size_t offset, const char *text, std::size_t
void
PieceTable::lineIndexOnDelete(std::size_t offset, std::size_t len) const
{
std::lock_guard<std::mutex> lock(mutex_);
if (line_index_dirty_)
return;
// Starts in (offset, offset+len] followed a deleted newline: drop them.
@@ -482,7 +478,6 @@ PieceTable::lineIndexOnDelete(std::size_t offset, std::size_t len) const
void
PieceTable::RebuildLineIndex() const
{
std::lock_guard<std::mutex> lock(mutex_);
if (!line_index_dirty_) {
return;
@@ -930,7 +925,6 @@ PieceTable::GetRange(std::size_t byte_offset, std::size_t len) const
// Fast path: return cached value if version/offset/len match
{
std::lock_guard<std::mutex> lock(mutex_);
if (range_cache_.valid && range_cache_.version == version_ &&
range_cache_.off == byte_offset && range_cache_.len == len) {
return range_cache_.data;
@@ -940,7 +934,6 @@ PieceTable::GetRange(std::size_t byte_offset, std::size_t len) const
std::string out;
out.reserve(len);
if (!dirty_) {
std::lock_guard<std::mutex> lock(mutex_);
// Already materialized; slice directly
out.assign(materialized_.data() + static_cast<std::ptrdiff_t>(byte_offset), len);
} else {
@@ -965,7 +958,6 @@ PieceTable::GetRange(std::size_t byte_offset, std::size_t len) const
// Update cache
{
std::lock_guard<std::mutex> lock(mutex_);
range_cache_.valid = true;
range_cache_.version = version_;
range_cache_.off = byte_offset;
@@ -984,7 +976,6 @@ PieceTable::Find(const std::string &needle, std::size_t start) const
if (start > total_size_)
return std::numeric_limits<std::size_t>::max();
{
std::lock_guard<std::mutex> lock(mutex_);
if (find_cache_.valid &&
find_cache_.version == version_ &&
find_cache_.needle == needle &&
@@ -996,7 +987,6 @@ PieceTable::Find(const std::string &needle, std::size_t start) const
materialize();
std::size_t pos;
{
std::lock_guard<std::mutex> lock(mutex_);
pos = materialized_.find(needle, start);
if (pos == std::string::npos)
pos = std::numeric_limits<std::size_t>::max();