Repository navigation

Parser performance: 2-3x speedup via grammar preprocessing and interpreter optimisations - #373

Closed

JanJakes wants to merge 20 commits into

trunkfrom

parser-performance

Commits on Apr 24, 2026
- Inline terminal matching and defer parse node allocation
  JanJakes
  committed
  f935d03
- Use per-branch FIRST sets to skip unreachable branches
  JanJakes
  committed
  0e34f25
- Short-circuit nullable-fallback and inline single-branch fragments
  JanJakes
  committed
  6df347a
- Strip epsilon markers and cache grammar refs on the parser
  JanJakes
  committed
  9eea849
- Return fragment results as children arrays, skip the intermediate node
  JanJakes
  committed
  ddfe4b6
- Append end-of-input sentinel token to drop range checks
  JanJakes
  committed
  77756bf
- Embed branch symbol sequences directly in the per-token selector
  JanJakes
  committed
  daec1cb
- Compare selectStatement by rule id instead of by name
  JanJakes
  committed
  2609898
- Add parse-only benchmark tool and fix coding-standard alignment
  JanJakes
  committed
  c4e7f8f
- Deduplicate identical selector entries to shrink grammar memory
  JanJakes
  committed
  cd0b609
- Cache branch symbol sequences in the deduplicated selector
  JanJakes
  committed
  e8b006f
- Add grammar compilation experiment and supporting benchmarks
  JanJakes
  committed
  b5959e8
- Pack switch case labels to halve compiled parser line count
  JanJakes
  committed
  bf1b1fe
- Add fast path for rules with a single branch per token
  JanJakes
  committed
  fa2fd65
- Direct-return fast path for single-candidate rules
  JanJakes
  committed
  1e7e3cf
- Mark WP_Parser_Node as final
  JanJakes
  committed
  9fcfb27
- Add bench-final.php helper for multi-config parser benchmarking
  JanJakes
  committed
  b9b64a6
Commits on Apr 25, 2026
- Add regex-based grammar matcher experiment
  JanJakes
  committed
  e0c09f8
- Add hybrid regex-pre-validate + parser experiment
  JanJakes
  committed
  9d36df4
- Document why PHP cannot expose PCRE2 callouts (the only way to get an AST)
  JanJakes
  committed
  dea9df7

Navigation Menu

Repository navigation

Parser performance: 2-3x speedup via grammar preprocessing and interpreter optimisations - #373

Parser performance: 2-3x speedup via grammar preprocessing and interpreter optimisations#373
JanJakes wants to merge 20 commits into
trunkfrom
parser-performance

Commits on Apr 24, 2026

Inline terminal matching and defer parse node allocation

Use per-branch FIRST sets to skip unreachable branches

Short-circuit nullable-fallback and inline single-branch fragments

Strip epsilon markers and cache grammar refs on the parser

Return fragment results as children arrays, skip the intermediate node

Append end-of-input sentinel token to drop range checks

Embed branch symbol sequences directly in the per-token selector

Compare selectStatement by rule id instead of by name

Add parse-only benchmark tool and fix coding-standard alignment

Deduplicate identical selector entries to shrink grammar memory

Cache branch symbol sequences in the deduplicated selector

Add grammar compilation experiment and supporting benchmarks

Pack switch case labels to halve compiled parser line count

Add fast path for rules with a single branch per token

Direct-return fast path for single-candidate rules

Mark WP_Parser_Node as final

Add bench-final.php helper for multi-config parser benchmarking

Commits on Apr 25, 2026

Add regex-based grammar matcher experiment

Add hybrid regex-pre-validate + parser experiment

Document why PHP cannot expose PCRE2 callouts (the only way to get an AST)