code-review
Review changes to a multilingual source code parser, focusing on whether relationships between code elements are correct. Check consistency across languages and verify that test examples reproduce the issue and actually validate the fix.
中文简介正在准备中。当前显示英语简介。
原文查看原文说明
Review pull requests to code-graph-rag, a multi-language code graph parser. Focuses on graph-edge correctness in language parsers, cross-language consistency, and test fixtures that actually exercise the fix.
适用场景
- Checking constructor call and instantiation relationships
- Reviewing name resolution and scope handling
- Checking whether a parser fix applies to other languages
- Verifying that test fixtures reproduce the exact issue
包含的文件(1)
- SKILL.md2.7 KB
SKILL.md(原文)
安装前可以查看交给 AI 代理的指令内容。
Reviewing code-graph-rag
This repository parses source in many languages with tree-sitter and ingests the
result into a code graph: nodes for modules, classes and functions, edges such as
CALLS, INSTANTIATES, DEFINES and IMPORTS. Most defects that reach main
are wrong or missing edges for a specific language construct, not crashes.
Weight the review accordingly.
Highest-value checks
Graph-edge correctness
- Does a new or changed construct produce every edge it should, and no edge it
should not? A constructor call is the recurring example: it should record
INSTANTIATESon the class andCALLSon the constructor. - Check the qualified name the edge is attached to. Name-mangling bugs recur
here: duplicate-suffixed classes, verbatim identifiers (
@classin C#), named constructors, generics (Box<int>()) and marker-stripping passes have each produced an edge pointing at a name that no node has. - Resolution must respect scope. A local variable shadowing a type means the receiver is not a construction; flag resolution that matches on bare name without checking what is actually in scope.
Language coverage and consistency
- A fix in one language's parser usually applies to its siblings. If the change
touches
dart/, ask whethercsharp/,cpp/, and the rest share the shape and need the same fix — or say explicitly why they do not. - Grammars are optional at install time. Code and tests must not assume a grammar is present; tests for one language should skip cleanly without it.
Tests
- Every behavioral fix needs a fixture reproducing the exact construct, and the fixture must be valid source in that language — an invalid fixture can pass for the wrong reason. Verify the assertion would fail without the fix.
- Prefer asserting on specific edges between specific qualified names over asserting on counts, which pass accidentally.
Repository conventions
- Python 3.12,
ruff(line length 88), anduvfor dependencies. Any dependency change must updateuv.lock; CI fails on drift viauv lock --check. - Comments explain why, not what. Match the density of surrounding code.
Calibration
Report correctness problems: a wrong edge, a missed construct, a fixture that
does not test the fix, a resolution that ignores scope. Do not report style that
ruff already enforces, and do not restate what the diff does. If the change is
correct, say so briefly rather than inventing findings.
评价
还没有评价。欢迎分享你的使用感受。
评论
无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。