On Apr 4, 2014, at 1:33 PM, Zoltan Horvath <[email protected]> wrote:
> On Fri, Apr 4, 2014 at 12:43 PM, James Craig <[email protected]> wrote: > >> On Apr 4, 2014, at 9:08 AM, Zoltan Horvath <[email protected]> wrote: >> >>> Hi there, >>> >>> prepare-Changelog: >>> - The style checker script runs on the ChangeLogs as well, and it gives >>> warning if the bug number or the tests' line are not filled. >> >> I consider this an expected behavior. If you don't include --bug=12345 >> parameter when running prepare-Changelog, your diff is currently in a state >> that would fail style checks. Having the Terminal output immediately tell >> you that detail is useful. >> >>> - I think the best behavior for this case would be to abort the change log >>> creation, if there are errors on the actual code check. (Then you can fix >>> those, and when you rerun prepare-Changelog, it puts the recent things into >>> the change log. [e.g. you've got a lower/uppercase function naming warning, >>> and the function name appears in the change log]) >> >> I don't think this is right. If we aborted the change log creation, you'd >> never be able to run prepare-Changelog without the --bug parameter, unless >> you specifically disabled the style check with --no-style. > > I'd avoid running the style-checker on the change logs themselves <since > these're being created> when running prepare-Changelog, and abort changelog > creation, if the _rest of the code_ have coding style violation <since fact > of violation is our interest at this point>. What if we file another bug for that part of the work? It seems like a nice-to-have, but not necessary for this patch. _______________________________________________ webkit-dev mailing list [email protected] https://lists.webkit.org/mailman/listinfo/webkit-dev
