Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 590d12dd36
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| if (std.mem.indexOf(u8, stripped, "(")) |paren| { | ||
| if (paren > 0) { | ||
| // Skip control flow keywords | ||
| const control = [_][]const u8{ "if ", "if(", "else ", "for ", "for(", "foreach ", "foreach(", "while ", "while(", "switch ", "switch(", "catch ", "catch(", "return ", "return(", "new ", "throw ", "lock ", "lock(" }; |
There was a problem hiding this comment.
This branch marks almost any parenthesized C# line as a .method symbol, so ordinary calls like Console.WriteLine(...), LINQ calls, or constructor invocations inside method bodies are indexed as declarations. That will significantly inflate symbol lists and degrade outline/search accuracy on real C# files, even though no new declarations exist on those lines. Method extraction needs a declaration-specific guard instead of relying on ( plus a small control-keyword denylist.
Useful? React with 👍 / 👎.
| if (startsWith(line, "using ")) { | ||
| const symbol_copy = try a.dupe(u8, line); | ||
| errdefer a.free(symbol_copy); | ||
| try outline.symbols.append(a, .{ |
There was a problem hiding this comment.
The parser treats any line beginning with using as an import and appends it to outline.imports, but C# also uses using for local resource-management statements (using var ... / using (...)). Those common in-method statements will be recorded as dependencies, which corrupts import/dependency data and downstream getImportedBy results for C# files.
Useful? React with 👍 / 👎.
Cherry-picked from justrach/codedb PR justrach#106 (unliftedq). Adds parseCSharpLine with support for using directives, namespaces, classes, interfaces, structs, enums, delegates, constants, and methods. Includes stripCSharpModifiers and lastIdent helpers. Co-Authored-By: unliftedq <unliftedq@users.noreply.github.com> Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
Hey @unliftedq, thanks for adding C# support — the parser itself looks solid (using directives, namespaces, classes, interfaces, structs, enums, delegates, methods, constants with modifier stripping). Appreciate the contribution! One issue though — the PR includes ~262 lines of whitespace reformatting on existing methods (
This reformatting creates massive merge conflicts with every other open PR (we just merged PHP support in #87 and are about to merge HCL in #109), and makes it very hard to review the actual C# changes. Could you resubmit with only the C# parser additions? Specifically just:
No reformatting of existing code — keep the original indentation as-is. That way it'll rebase cleanly and we can get it merged quickly. Thanks again for the work! |
unliftedq
force-pushed
the
add-csharp-support
branch
from
590d12d to
4aaa6f0
Compare
It should be fixed now, previous unrelated change is unexpectedly introduced by the zig language server of my local zig extension. |
2 participants
As title, Added basic C# support to the explorer.