feat(lang): GDScript — vendored grammar, tags.scm, gdscriptcheck.sh

CMakeLists.txt

@@ -250,6 +250,7 @@ set(_ripwire_query_names
   java
   javascript
   json
+  gdscript
   kotlin
   lua
   objc
@@ -474,6 +475,17 @@ add_ts_grammar(lua    https://github.com/tree-sitter-grammars/tree-sitter-lua.gi
 # deliberately not a kLangTable row yet — see the note there.
 add_ts_grammar(kotlin https://github.com/fwcd/tree-sitter-kotlin.git 1852ea17b7f60fb3f9d84e0b1555d56b46b39fb1)  # main, post-0.3.8
 
+# GDScript (.gd): PrestonKnopp/tree-sitter-gdscript — the only maintained GDScript grammar; there is no
+# first-party one under tree-sitter/. LANGUAGE_VERSION 14, inside [MIN_COMPATIBLE, LANGUAGE_VERSION]=[13,15]
+# for the vendored core. Ships a scanner.c (significant indentation: INDENT/DEDENT/NEWLINE, like Python's)
+# which the macro picks up. Untagged upstream, so the SHA below is a plain master commit, pinned as always.
+# PARSE RATE MEASURED, not assumed, before vendoring (STEP 0 of prompts/add-a-language.md): 98.88% of 2938
+# real .gd files parse with zero ERROR/MISSING nodes — 13 open-source Godot projects (2611 files, 98.81%)
+# plus a private 327-file game (99.39%); 8159 error bytes out of 25.3 MB (0.032%). The three known blind
+# spots are disclosed in queries/gdscript/tags.scm and test/gdscriptcheck.sh; all three are upstream grammar
+# gaps, NOT patched here (guardrail G3 — this repo carries no grammar fork).
+add_ts_grammar(gdscript https://github.com/PrestonKnopp/tree-sitter-gdscript.git c5c8fa4861b5a4f04a7e60d97587fc3b6cc5639e)  # master, ABI 14
+
 # PHP (.php): the official tree-sitter org grammar. Like TypeScript this repo hosts TWO sub-grammars
 # in subdirectories — `php/` (the full one: HTML text interleaved with `<?php … ?>` blocks) and
 # `php_only/` (no surrounding text). ONLY `php/` is vendored and built: a real .php file in a real
@@ -535,6 +547,7 @@ set(RIPWIRE_TS_OBJECTS
   $<TARGET_OBJECTS:ts_lua>
   $<TARGET_OBJECTS:ts_dart>
   $<TARGET_OBJECTS:ts_kotlin>
+  $<TARGET_OBJECTS:ts_gdscript>
   $<TARGET_OBJECTS:ts_elixir>
   $<TARGET_OBJECTS:ts_php>
   $<TARGET_OBJECTS:ts_toml>
@@ -745,7 +758,7 @@ endif()
 set(RIPWIRE_GRAMMAR_TARGETS
   ts_cpp ts_python ts_go ts_rust ts_tsts ts_tstsx ts_swift ts_objc
   ts_javascript ts_bash ts_java ts_ruby ts_json ts_toml ts_yaml ts_csharp ts_c ts_cuda ts_markdown
-  ts_php ts_lua ts_elixir ts_dart ts_kotlin)
+  ts_php ts_lua ts_elixir ts_dart ts_kotlin ts_gdscript)
 set(RIPWIRE_RUNTIME_COMPILE_TARGETS ripwire_probe ripwire tree-sitter ${RIPWIRE_GRAMMAR_TARGETS})
 set(RIPWIRE_RUNTIME_LINK_TARGETS ripwire_probe ripwire)
 # The C++ subset of the compile list — the only targets that include a C++ standard library at all, and so
@@ -1072,5 +1085,6 @@ if(RIPWIRE_FUZZ)
   add_ripwire_fuzzer(elixir     tree_sitter_elixir     elixir)
   add_ripwire_fuzzer(lua        tree_sitter_lua        lua)
   add_ripwire_fuzzer(kotlin     tree_sitter_kotlin     kotlin)
+  add_ripwire_fuzzer(gdscript   tree_sitter_gdscript   gdscript)
   add_custom_target(ripwire_fuzzers DEPENDS ${RIPWIRE_FUZZ_TARGETS})
 endif()

README.md

@@ -1816,9 +1816,9 @@ wrong, and it has. These are the results that say so, all in-tree, all published
 ### In the tests
 
 <details>
-<summary><b>618 gate scripts</b>, five contracts no unit test can hold, and the house rule: write the gate before the code it measures</summary> <!-- gatecount -->
+<summary><b>619 gate scripts</b>, five contracts no unit test can hold, and the house rule: write the gate before the code it measures</summary> <!-- gatecount -->
 
-`test/regression.sh` names **618 gate scripts** and is the authoritative list; <!-- gatecount -->
+`test/regression.sh` names **619 gate scripts** and is the authoritative list; <!-- gatecount -->
 `python3 test/pargates.py . ./build/ripwire -j 6` runs the same set in parallel. On top of them sit the
 contracts that do not fit a unit test: two runs byte-identical, warm output identical to cold, output
 that pipes clean through `xmllint --noout`, a sanitizer build with `-fno-sanitize-recover=all`, and a
@@ -2649,7 +2649,7 @@ file, and one row in the extension table.
 | Metal (MSL) | `.metal` | Indexed with the C++ grammar. |
 | CUDA | `.cu`, `.cuh` | `<<<>>>` launch sites are call edges. |
 | Python | `.py` | |
-| TypeScript / JavaScript | `.ts`, `.tsx`, `.js`, `.jsx` | Named imports and default imports resolve. One vendored dependency supplies two of the 24 grammars, `typescript` and `tsx`. |
+| TypeScript / JavaScript | `.ts`, `.tsx`, `.js`, `.jsx` | Named imports and default imports resolve. One vendored dependency supplies two of the 25 grammars, `typescript` and `tsx`. |
 | Java | `.java` | Qualified `new` calls resolve in a precise tier. |
 | Kotlin | `.kt` | Shares one call graph with Java. A file with string templates past 128 levels is refused and listed by `--skipped`. |
 | Ruby | `.rb` | Superclasses, mixins, `autoload`, and constant receivers are read. |
@@ -2662,6 +2662,7 @@ file, and one row in the extension table.
 | Go | `.go` | Qualified calls are rejected and fenced, not guessed. |
 | Rust | `.rs` | Scoped, turbofish, and `Self::` calls resolve in a precise tier. |
 | Bash | `.sh`, `.bash` | |
+| GDScript | `.gd` | A file is a class body: `class_name` names it and file-scope `func`/`var` are its members. A signal is indexed as a member. `preload`/`load` produce no dependency edge. `.tscn`, `.tres`, and `.gdshader` are not indexed. |
 | JSON | `.json` | Config keys become symbols. The lane emits no call edges. |
 | TOML | `.toml` | A table header is one symbol. Keys below it are one level down. |
 | YAML | `.yml`, `.yaml` | Mapping depth 2 is the cut. Sequence levels are transparent. |

THIRD_PARTY.md

@@ -69,13 +69,16 @@ is why the sizes are what they are — `parser.c` is one big static table, not h
 | `deps/lua` | tree-sitter-lua (v0.5.0) | Munif Tanjim | MIT | `10fe0054734eec83049514ea2e718b2a56acd0c9` | https://github.com/tree-sitter-grammars/tree-sitter-lua | 392 KB |
 | `deps/markdown` | tree-sitter-markdown (v0.5.3; the block grammar `tree-sitter-markdown/` only) | Matthias Deiml | MIT | `f969cd3ae3f9fbd4e43205431d0ae286014c05b5` | https://github.com/tree-sitter-grammars/tree-sitter-markdown | 2.1 MB |
 | `deps/kotlin` | tree-sitter-kotlin (ABI 14) | fwcd | MIT | `1852ea17b7f60fb3f9d84e0b1555d56b46b39fb1` | https://github.com/fwcd/tree-sitter-kotlin | 32 MB |
+| `deps/gdscript` | tree-sitter-gdscript (master) | Preston Knopp | MIT | `c5c8fa4861b5a4f04a7e60d97587fc3b6cc5639e` | https://github.com/PrestonKnopp/tree-sitter-gdscript | 2624 KB |
 | `deps/doctest` | doctest (v2.4.12) | Viktor Kirilov | MIT | `1da23a3e8119ec5cce4f9388e91b065e20bf06f5` | https://github.com/doctest/doctest | 0.7 MB |
 
 Notes:
 
 - `deps/tree_sitter` keeps upstream's own `CMakeLists.txt` (the build `add_subdirectory`s it),
   `lib/src`, `lib/include` and `lib/tree-sitter.pc.in`. Its `lib/src/unicode/` is a subset of ICU
   carrying its own `LICENSE` (Unicode-DFS-2016) and `ICU_SHA` provenance file, left untouched.
+- `deps/gdscript` is pinned to a bare commit for the same reason as `deps/swift`: upstream
+  publishes no release tags at all, so a 40-hex commit is the only immutable pin available.
 - `deps/swift` is pinned to a bare commit rather than a tag because upstream's default branch does
   not carry a generated `parser.c`; that commit's generated output is what is vendored here.
 - `deps/kotlin` is pinned to a bare commit rather than the last tag (`v0.3.8`, Aug 2024) because

docs/ARCHITECTURE.md

@@ -252,6 +252,31 @@ Elixir extraction landed at revision 78 (rich 79) — `kParserVer` in `src/inges
 `kIngestParserVerMirror` in `src/quality.h`. The required `qschemetrip` source-change pin is refreshed
 for this extraction change; snapshot scheme 8 is unchanged.
 
+<a id="gdscript-extraction"></a>
+
+GDScript needs no capture-filter module: unlike Elixir, its grammar carries real definition nodes, so
+`queries/gdscript/tags.scm` alone is the extraction. A `.gd` FILE IS A CLASS BODY — `class_name` names
+the class and file-scope `func`/`var` are its members — so a file-scope `func` is `fn` (as every other
+language treats a file-scope definition) while a `func` inside an explicit `class Inner:` is `method`.
+`enum` rides `@definition.type` as Java/C#/TypeScript do. Two capture choices are forced by gates in
+`ingest_names.h`, not taste: enum MEMBERS ride `@definition.constant` because `@definition.enummember`
+is gated by `isPyEnumMemberTarget` and would silently drop every GDScript enumerator, and member
+variables ride `@definition.var` because `fieldCaptureKept()` returns false for every language but
+Python and C/C++. A `signal` has no SymKind of its own and is DISCLOSED as `t="var"`. The Godot 4
+spellings were read off real parses, not node types; `queries/gdscript/tags.scm` records which shapes
+are not what the node-type list implies.
+
+THE FLOOR, measured before vendoring: 98.88% of 2938 real `.gd` files parse clean. Three upstream
+grammar gaps survive — the `%` unique-name inside a path, a column-0 comment in an indented block, and
+the Godot 3 RPC keywords still reserved — and none is patched (guardrail G3). Recovery is LOCAL, so a
+file holding them still yields every definition and call edge; `test/gdscriptcheck.sh` asserts that
+survival rather than the failure. `preload`/`load("res://…")` resolution is NOT implemented, so a `.gd`
+file is never a node in the `--deps`/`--arch` graph and `dependencyCapable()` is not claimed for it.
+`.tscn`, `.tres` and `.gdshader` are not indexed.
+
+GDScript extraction landed at revision 98 — `kParserVer` in `src/ingest_cache.h`, mirrored by
+`kIngestParserVerMirror` in `src/quality.h`; snapshot scheme is unchanged.
+
 The three config lanes are *data*, not code: they emit `t="sec"` symbols and **zero call edges**, and
 `langCompatible` keeps a config key from ever resolving a same-spelled code symbol. They differ in
 where the navigable unit sits. JSON cuts at document depth — top-level and second-level object

docs/COMMANDS.md

@@ -16,9 +16,9 @@ current shape.
 > ripwire — the "ripgrep of AI context": parse a codebase, rank symbols by Personalized PageRank,
 > stream a deterministic minified XML map to stdout. Zero runtime deps. Languages: C++, C, ObjC/ObjC++,
 > Metal (MSL, .metal — C++ grammar), CUDA (.cu/.cuh — tree-sitter-cuda, <<<>>> launches are call edges),
-> Python, TypeScript, JavaScript, Java, Ruby, PHP (.php/.phtml), Lua, Elixir (.ex/.exs), Dart (.dart), Kotlin (.kt), Bash, Go, Rust, Swift, C#;
+> Python, TypeScript, JavaScript, Java, Ruby, PHP (.php/.phtml), Lua, Elixir (.ex/.exs), Dart (.dart), Kotlin (.kt), Bash, Go, Rust, Swift, C#,
+> GDScript (.gd — Godot; .tscn/.tres/.gdshader are NOT indexed);
 > JSON, TOML, YAML (config keys); Markdown (.md/.markdown — headings are section symbols with spans).
-> usage: ripwire <dir> [flags]            # default = the ranked map of <dir> on stdout
 
 ## How to read a section
 

docs/EVALS.md

@@ -21,7 +21,7 @@ section, and it is not an afterthought.
 | **Co-change / known-item evals** | `--eval`, `--eval-retrieval` (see `bench/ANSWERQUALITY.md`) | Whether the tool surfaces the other files a real historical commit touched; and known-item retrieval across four rankers. |
 | **Ensemble calibration harness** | `bench/ensemblecal/` | Whether `--ensemble`'s four evidence families are actually orthogonal, how often each fires, how stable each is across commits — and the preset ladder derived from that (§9). |
 | **Differential argv harness** | `test/argvdiffcheck.sh` | That a refactor changed *nothing observable*: two binaries, every argv vector, stdout + stderr + exit code byte-identical. |
-| **The gate suite** | `test/regression.sh`, `test/pargates.py` | 618 gate scripts plus the determinism, cache-transparency and golden contracts. <!-- gatecount --> |
+| **The gate suite** | `test/regression.sh`, `test/pargates.py` | 619 gate scripts plus the determinism, cache-transparency and golden contracts. <!-- gatecount --> |
 | **`--quality-delta`** | `src/quality.h` | Ten measured code-quality failure modes, reported only where a change made them worse. |
 
 ### The labeling protocol (why the held-out eval is allowed to disagree with the ranker)
@@ -5834,7 +5834,7 @@ copy here would be exactly the dialect divergence that gate exists to catch. Com
 tags, wrap, stable-order defaults), seven individually invoked standalone gates (`g1freshcheck`,
 `skillscan`, `htmlexport`, `compresscheck`, `handoffcheck`, `releaseinstallcheck`,
 `taskroutecheck`), and a single loop
-naming **618 gate scripts**, all of which exist on disk. <!-- gatecount -->
+naming **619 gate scripts**, all of which exist on disk. <!-- gatecount -->
 
 `python3 test/pargates.py . ./build/ripwire -j 6` runs the same scripts in parallel so a full
 verification fits in one sitting. It does not modify `regression.sh`.
@@ -6846,7 +6846,7 @@ Listed because the reason is more useful than the silence.
   shipped**. See `bench/locbench/anchorhop_calib.json`. The mention anchor's reproducible numbers are
   the ablations in §4.
 - **A single round gate-count.** Two in-tree numbers disagree (`test/pargates.py`'s docstring says
-  ~210; `test/argvdiffcheck.sh` says 200+), while the loop in `test/regression.sh` names 618. The <!-- gatecount -->
+  ~210; `test/argvdiffcheck.sh` says 200+), while the loop in `test/regression.sh` names 619. The <!-- gatecount -->
   loop is the authority; the stale docstrings are a known drift. Since 2026-09-10 the number is not
   written by hand anywhere: `docs/gatecount_build.py` derives it from the loop and rewrites every
   published site, `test/gatecountcheck.sh` fails if any of them drifts, and `test/manifestcheck.sh`

present/deck5_ripwire_build.js

@@ -1123,7 +1123,7 @@ function storyCards(s, { kick, head, stories, footText }){
   kicker(s, "// how it stays true", AMBER);
   title(s, "Proven, not promised");
   const cards = [
-    ["618 gate scripts", "the suite runs on every push — plus determinism, cache-transparency and golden contracts; the gate count itself is gated against the runner's own loop"], // gatecount
+    ["619 gate scripts", "the suite runs on every push — plus determinism, cache-transparency and golden contracts; the gate count itself is gated against the runner's own loop"], // gatecount
     ["byte-identical, always", "two runs over the same tree produce the same bytes; warm equals cold. Enforced in CI, twice — Release AND a plain flavour, because NDEBUG once blinded a whole class of checks"],
     ["differential refactoring", "a refactor must prove it changed nothing observable: two binaries, hundreds of argv vectors, stdout + stderr + exit codes byte-identical"],
     ["held-out labels, authored blind", "eval labels were written by reading source before the ranker ever ran on them — so the eval is allowed to say the ranker is wrong. It has."],
@@ -1147,7 +1147,7 @@ function storyCards(s, { kick, head, stories, footText }){
   title(s, "Claims you can trust, because we publish what failed", { size: 32 });
 
   card(s, MX, 1.72, 3.86, 1.72);
-  stat(s, "618", "gate scripts named by test/regression.sh — and the COUNT itself is gated against the runner's own loop, so it cannot go stale quietly", // gatecount
+  stat(s, "619", "gate scripts named by test/regression.sh — and the COUNT itself is gated against the runner's own loop, so it cannot go stale quietly", // gatecount
     MX+0.15, 1.86, 3.56, CYAN, { bsize: 42, bh: 0.66, lsize: 9.5 });
   card(s, 4.68, 1.72, 3.86, 1.72, CARD2);
   stat(s, "8", "registered NEGATIVES — changes built, gated green, measured against a band written before the code, and reverted rather than tuned",
@@ -1397,7 +1397,7 @@ function storyCards(s, { kick, head, stories, footText }){
     ["180 long flags · 34 slides",        "bash test/deckclaimcheck.sh"],
     ["every --flag named here exists",    "bash test/deckcheck.sh"],
     ["74.7% fewer element bytes",         "bash test/showcasecapturecheck.sh"],
-    ["618 gate scripts",                  "bash test/manifestcheck.sh"], // gatecount
+    ["619 gate scripts",                  "bash test/manifestcheck.sh"], // gatecount
     ["49 repos · 71 papers · 237 surveyed","bash test/readmedriftcheck.sh"],
     ["the ten moments, any row",          "ripwire . --callers=SYM | wc -c"],
     ["the head-to-head table",            "bench/headtohead/r4-2026-08-06/"],