Feat/fortran language support#1085
Open
IppearPeng wants to merge 2 commits into
Open
Conversation
Wire the tree-sitter-fortran grammar (vendored ABI-15 wasm) and a visitNode-driven extractor into the pipeline: modules/programs, subroutines & functions (incl. CONTAINS-block procedures), derived types with components and type-bound procedures, generic interfaces, PARAMETER constants, and `use` imports — across modern free-form (.f90/.f95/.f03/.f08/.f18) and legacy fixed-form (.f/.for/.f77/.ftn). Fortran's AST is irregular relative to the declarative LanguageExtractor model (definition nodes carry no name/body field), so every program-unit and type construct is handled in the visitNode hook; the declarative type lists stay empty so the core never re-dispatches a Fortran definition. Wiring follows the project's add-lang convention: LANGUAGES (types.ts), WASM_GRAMMAR_FILES / EXTENSION_MAP / display-name / vendored-wasm branch (grammars.ts), and the EXTRACTORS barrel. Docs (README, site languages table, CHANGELOG) updated. Ported from the team's proven implementation, re-applied onto 1.1.6. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Make Fortran's call graph resolve correctly across files: - extractCall normalizes `CALL obj%method()` to obj.method (chained receivers unwrap to the immediate component; this/self kept for typed resolution). - name-matcher: case-insensitive lookup (Fortran declares `subroutine foo`, calls `CALL FOO()`) via a lowercase-index fallback gated to fortran; receiver-type inference rides the shared localReceiverTypePatterns table (colbymchenry#1108) with a CLASS(t)/TYPE(t) declaration pattern, so `eng%Integrate()` lands on that type's binding; non-callable kinds are filtered from `calls` refs so array indexing never resolves to a variable; TBP binding refs can't self-loop. - resolution pre-filter consults the lowercase name set for case-insensitive languages so a case-mismatched ref isn't dropped before any matcher runs. - callback-synthesizer: fortranOverrideEdges bridges base-type bindings to extending types' overrides (CHA), mirroring cpp-override — so polymorphic dispatch through an abstract base follows into every override. Tests: Fortran extraction block in extraction.test.ts plus an end-to-end fortran-pipeline integration test. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
IppearPeng
force-pushed
the
feat/fortran-language-support
branch
from
4643a53 to
ea9ecca
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds Fortran language support to CodeGraph — ported from a proven in-house implementation (developed on 0.9.x) and re-applied onto the current 1.1.6 resolution architecture, following the repo's own add-lang convention.
CodeGraph now indexes modern free-form (.f90 .f95 .f03 .f08 .f18) and legacy fixed-form (.f .for .f77 .ftn .fpp) Fortran:
imports. Emits contains, imports, extends, and calls edges.
How it was wired
Standard add-lang touch points: LANGUAGES (types.ts), grammar/extension/display-name maps + vendored-wasm branch (grammars.ts), the EXTRACTORS barrel, and a vendored tree-sitter-fortran.wasm (ABI 15). The extractor is visitNode-driven
because Fortran's AST is irregular relative to the declarative LanguageExtractor model. Resolution changes live in tree-sitter.ts, name-matcher.ts, resolution/index.ts, and callback-synthesizer.ts, all Fortran-gated.
Tests / validation
Notes