[Webkit-unassigned] [Bug 183888] Add a bisect button to automatically schedule bisecting A/B tasks.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=183888

--- Comment #6 from Ryosuke Niwa <rniwa at webkit.org> ---
(In reply to dewei_zhu from comment #5)
> Comment on attachment 336889 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=336889&action=review
> 
> >> Websites/perf.webkit.org/public/v3/models/analysis-task.js:175
> >> +        const allTestGroupsInTask = TestGroup.findAllByTask(this.id());
> > 
> > What guarantees that we've already fetched all the test groups?
> 
> It only works for the UI part. I think I need to use fetchForTask instead.

Okay. Do we cache the results there? We should make sure we don't end up fetching it over & over.

> >> Websites/perf.webkit.org/public/v3/models/analysis-task.js:191
> >> +        try {
> > 
> > I don't really use try-catch elsewhere.
> > I'm not certain if it's a good idea to use it here either because it can swallow unexpected exceptions.
> 
> When will those error eventually get caught?

The point is that it won't be caught and shows up on WebInspector.

> >> Websites/perf.webkit.org/public/v3/models/commit-set.js:186
> >> +                nameParts.push(`${repository.name()}: Patch-${firstPatch.id()} - Patch-${secondPatch.id()}`);
> > 
> > I don't think it makes sense to expose the attachment ID like this. We don't do elsewhere in the UI.
> > If we're concerned about the length of the name, then we can just diff the name with the maximum length of name to show.
> > e.g. if we had WebKit-WebComponents-A.patch and WebKit-WebComponents-B.patch, we can just show:
> > WebKit: "...A.patch" - "...B.patch"
> 
> What if two different patch have the same name?

We should probably add the file size, uploaded date, etc... This seems like an unlikely scenario though. It would be confusing for humans anyway.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20180401/467ebba4/attachment-0002.html>