feat: tree-sitter integration — fn, class, smells commands + enhanced file/symbol/status#17
Draft
feat: tree-sitter integration — fn, class, smells commands + enhanced file/symbol/status#17
Conversation
- src/tree-sitter/types.ts: ParameterInfo, FunctionInfo, MethodInfo, ClassInfo, FileMetrics - src/schemas/treesitter.ts: Zod schemas + FnResult, SmellsResult, ClassResult - src/schemas/index.ts: export treesitter schemas Refs: dora-zp1e
- typescript.ts: function + class queries with full capture extraction - javascript.ts: same for JS/JSX (no type annotations) - registry.ts: maps extensions to grammar WASM names and query modules Refs: dora-xe7i
Refs: dora-9mh7
Refs: dora-rj0c
Refs: dora-f9wi
Refs: dora-q94q
Refs: dora-ypef
Refs: dora-y30c
Refs: dora-us7d
- Fix fn/class/smells calls to use single object param in index.ts and handlers.ts - Fix import order in index.ts - Remove duplicate src/tree-sitter/types.ts, use schemas/treesitter.ts as source of truth - Deduplicate parseFunctions/parseClasses into shared parseFile helper - Make getDbConnection synchronous (getDb is sync) - Rename inBlockComment to isInBlockComment per boolean naming convention
…andling, cookbook registration
…rics - parseFunctionCaptures: all 5 capture types, dedup, async, params, return type, jsdoc, line numbers - parseClassCaptures: extends, implements, abstract, decorators, methods, properties - calculateFileMetrics: blank/comment/sloc counting, block comment state machine, complexity aggregation - export calculateFileMetrics from parser.ts - extend biome:format script to include ./test
Deploying with
|
| Status | Name | Latest Commit | Updated (UTC) |
|---|---|---|---|
| ✅ Deployment successful! View logs |
dora-docs | 8bfce36 | Mar 04 2026, 03:27 AM |
The nullish coalescing operator is a binary_expression node in tree-sitter with operator.text === '??', not a standalone '??' node type. The previous else-if branch was dead code and never fired. Move ?? into the binary_expression operator check alongside && and ||. Refs: dora-gn1s
Parser.init() was called on every getLanguage() cache miss, meaning loading a second grammar (e.g. tsx after typescript) would re-initialize the WASM runtime. Extract initialization into getInitializedModule() which caches the init promise and resolves it only once regardless of concurrent callers. Refs: dora-kudd
Bun 1.1+ uses bun.lock (text format) instead of bun.lockb (binary). Repos using newer Bun only have bun.lock and were not detected as Bun workspaces, resulting in wrong scip-typescript indexer flags. Refs: dora-otpb
Replace template literal path construction (`${root}/${path}`)
with resolveAbsolute() from utils/paths.ts in fn, class, smells,
file, and symbol commands. String interpolation is fragile on
Windows and can produce double slashes if root has a trailing
slash.
Refs: dora-n3ob
Previously tree_sitter status only probed the project's primary
language (e.g. 'typescript'), giving a misleading picture when
multiple grammars are installed. Now probes every language in
languageRegistry in parallel and reports each one individually.
Schema change: tree_sitter.grammars replaces
tree_sitter.{available, language, grammar_path}.
Refs: dora-2stw
The previous regex-based scanner only matched single-line // and inline /* */ comments on a single line. TODO/FIXME/HACK markers inside multi-line block comments (/* \n TODO ... \n */) were silently skipped. Rewrite scanTodoComments to use the same line-by-line state machine used in calculateFileMetrics, tracking isInBlockComment across lines. Refs: dora-t75k
MCP protocol delivers tool args as untyped JSON - no schema is available at the boundary, justifying the use of any. Refs: dora-zt88
Detect god_class (too many methods) and large_class (too many properties) using the existing parseClasses path. Both checks run in parallel with parseFunctions via Promise.all. New options: --methods-threshold (default 20), --properties-threshold (default 15). Exposed in CLI and MCP. Refs: dora-vhbg
Previously an unrecognized sort value silently fell through to the default. Now throws with a clear message listing valid values. Refs: dora-d6s0
findGlobalNodeModulesPath() spawned 'bun pm ls -g' on every grammar cache miss. Multiple tree-sitter commands in one session would each pay that subprocess cost. Cache the promise at module level so the subprocess runs at most once per process. Refs: dora-ommx
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
1 participant
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.
Summary
Adds tree-sitter powered static analysis to dora via
web-tree-sitter. Introduces three new commands (dora fn,dora class,dora smells) and enhances existing commands with richer output from AST parsing.Changes
Core infrastructure
node_modules, then globalbunpackagesNew commands
Enhanced commands
Docs
dora cookbook show tree-sitter)Tests
loc = sloc + comment_lines + blank_linesinvariant across all cases, complexity aggregation and rounding