xiaoxiang781216 commented on PR #17514: URL: https://github.com/apache/nuttx/pull/17514#issuecomment-3680861976
> > @xiaoxiang781216: it is an unrelated topic, please provide a patch to fix instead blocking this patch, thanks. @cederom . > > Respectfully, there are constantly PRs getting submitted with very poor descriptions and no testing, and I often see some frustration when other reviewers (often me, I will admit) request testing information or a better description. This PR has a very poorly formatted summary and almost 0 testing information. > > I know that this PR is a very small and straightforward change, but we enforce the contributing guidelines for everyone. It's important to see testing information that actually confirms the issue is fixed as the contributor claims. I know it is sometimes frustrating to have back-and-forth comments on sometimes very small patches, but it's also important to have quality PRs. I've noticed a lot better testing information from frequent contributors since they have started receiving this feedback and that's a good thing for NuttX's stability. Xiaomi is an important NuttX contributor and for a long time there have been PRs getting merged with almost zero description, since other Xiaomi-employed reviewers have "insider information" about how the feature was tested that us external reviewers do not. NuttX is not a Xiaomi product and it is important that everyone who is reviewing the PR understand its implications, not just Xiaomi employees, even if it is frustrating to have to describe a PR in detail or repeat internal testing to obtain logs. > > So please, try to avoid the hostility even when a PR is being held up by review. That is the point of review after all: to get everyone's feedback. But the message in pr already desribe the situation clearly, what' additional mesage do you still want? > ## Summary > bugfix:When the stat() returns ret < 0 and the error is not ENOENT, mkstemp will hang here and keep retrying. > > for example: When: path_template:/data/quickapp/cache/ncm_XXXXXX xlen is equal to BIG_XS retries will be set to UINT32_MAX then: The stat operation fails due to issues with drivers, memory problems, or file system operations and continuously returns an error other than ENOENT, it will keep retrying, which may cause the watchdog timer to time out. > > ## Impact > mkstem will not hang and keep retry when stat() failed but not a ENOENT > > ## Testing > Test passed on qemu-armeabi-v7a-ap From the above descrption, do you think there is any mainline test case could expose the type of error? please point out, so @guohao15 can try it. If not, how can you -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
