Repository navigation
fix(cli): name both recorded-core architectures in the path contract - #3234
fengqiaoqu wants to merge 1 commit into
Conversation
|
Codex usage limits have been reached for code reviews. Please check with the admins of this repo to increase the limits by adding credits. |
|
| const serve = async ( | ||
| cli: CliInstance, | ||
| argv: readonly string[], | ||
| ): Promise<{ readonly stdout: string; readonly exitCode: number }> => { | ||
| let stdout = ""; | ||
| let exitCode = 0; | ||
| await cli.serve([...argv], { | ||
| env: {}, | ||
| exit: (code) => { | ||
| exitCode = code; | ||
| }, | ||
| stdout: (text) => { | ||
| stdout += text; | ||
| }, | ||
| }); | ||
| return { stdout, exitCode }; | ||
| }; |
There was a problem hiding this comment.
The new serve helper repeats the output and exit-code capture already provided by serveCli in cliServeHarness.ts. Future changes to the shared setup would need to be repeated here. Import serveCli and pass the command arguments directly instead.
Prompt To Fix With AI
This is a comment left during a code review.
Path: tests/boundary/cli/recordedCrashHelpArchitectures.test.ts
Line: 7-23
Comment:
**CLI test setup is duplicated**
The new `serve` helper repeats the output and exit-code capture already provided by `serveCli` in `cliServeHarness.ts`. Future changes to the shared setup would need to be repeated here. Import `serveCli` and pass the command arguments directly instead.
---
For each issue above, determine whether it is valid and should be fixed. If so, fix it directly.Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!
|
This PR's behavior is now published on main (8dde637) in Covered scope: Recorded-crash help names Linux ELF64 amd64 and AArch64 little-endian core paths already admitted by the provider. Main evidence: tests/boundary/binaryDiagnostics/recordedCrashProvider.test.ts; tests/boundary/mcp/recordedCrashMcp.test.ts. Validation limits: Documentation/help change only; no claim that every core subtype or architecture is accepted. I checked the current published source and the substantive review threads; no unique behavior or dependency change remains on this PR. Closing as superseded without resolving individual review threads. |
|
@fengqiaoqu, thank you for catching the recorded-crash help omitting AArch64 and for supplying the CLI schema regression. Main names both supported core architectures. |
Problem
inspect-recorded-crashdecodes both supported Linux core profiles —x86_64-little-endian(machine 62) andaarch64-little-endian(machine 183) — and the MCP tool contract already says "Linux ELF64 amd64 or AArch64 little-endian core". The CLI path argument still described onlyLocal Linux amd64 ELF core(#3167), so the CLI help and exported schema omitted the documented AArch64 support that PR #2505 added.Fix
Align the CLI
pathdescription with the MCP contract wording:Local Linux ELF64 amd64 or AArch64 little-endian core. No decoder, acceptance, or option behavior changes.Validation
tests/boundary/cli/recordedCrashHelpArchitectures.test.tsexports the command schema via--schema --jsonand asserts thepathdescription names bothamd64andAArch64. It fails before the fix and passes after.npm run check:fastpasses.Fixes #3167