Parser + lexer performance: consolidated 2–3× end-to-end speedup - #378

Closed

JanJakes wants to merge 30 commits into

trunkfrom

Commits on Jun 5, 2026

Inline terminal matching and defer parse node allocation
JanJakes
committed
5a3f211
Use per-branch FIRST sets to skip unreachable branches
JanJakes
committed
cdc9713
Short-circuit nullable-fallback and inline single-branch fragments
JanJakes
committed
332c10a
Strip epsilon markers and cache grammar refs on the parser
JanJakes
committed
402554c
Return fragment results as children arrays, skip the intermediate node
JanJakes
committed
c43425f
Append end-of-input sentinel token to drop range checks
JanJakes
committed
e35e14f
Embed branch symbol sequences directly in the per-token selector
JanJakes
committed
a3b8a08
Compare selectStatement by rule id instead of by name
JanJakes
committed
25dad62
Re-align grammar and parser whitespace after recent changes
JanJakes
committed
e4a0951
Deduplicate selector entries while embedding branch sequences
JanJakes
committed
0e23320
Add direct-return fast path for single-candidate rules
JanJakes
committed
3164da9
Speed up the lexer with cheaper byte checks
JanJakes
committed
45c344a
Skip parent constructor in WP_MySQL_Token
JanJakes
committed
a763087
Use ! empty() in WP_Parser_Node::has_child()
JanJakes
committed
60981f6
Inline leading-whitespace skip in lexer's token loops
JanJakes
committed
1aa326a
Catch identifier and keyword tokens at the top of the chain
JanJakes
committed
7256e62
Add a single-byte operator dispatch table
JanJakes
committed
ee75d4b
Document non-obvious lexer dispatch conditions
JanJakes
committed
e36e2d2
Unroll whitespace check in '--' line-comment dispatch
JanJakes
committed
3867de9
Extract WP_Parser::set_tokens() helper for shared sentinel handling
JanJakes
committed
0e91429
Update native parser to consume branches_for_token + nullable_branches
JanJakes
committed
58d84c6
Materialize native AST eagerly into plain WP_Parser_Node instances
JanJakes
committed
b330676
Mark WP_Parser_Node as final
JanJakes
committed
ec5a74e
Document parent-ctor bypass and remaining_tokens duplication
JanJakes
committed
2e20dc5
Add unit tests for grammar build-time transforms
JanJakes
committed
849c596
Refresh native-extension benchmark numbers
JanJakes
committed
00a502a
Fall back to pure PHP when the native extension grammar ABI mismatches
JanJakes
committed
096a44e
Speed up grammar construction with a worklist fixpoint and lazy selectors
JanJakes
committed
0679908
Unify unit-test CI into one matrix and disable Xdebug
JanJakes
committed
7368e34
Cache the Rust extension build in CI
JanJakes
committed
d3716d3