[Webkit-unassigned] [Bug 146056] [GTK] [Wayland] Allow building and testing the Wayland target with the default JHBuild moduleset.

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Wed Jun 17 09:37:00 PDT 2015


https://bugs.webkit.org/show_bug.cgi?id=146056

--- Comment #6 from Martin Robinson <mrobinson at webkit.org> ---
(In reply to comment #5)
> (In reply to comment #4)
> > (In reply to comment #3)
> > > I think we should go straight to the latest stable release, 3.16.4, since
> > > it's closer to what our users will actually have. There's less value in
> > > testing with 3.12.x since it's already quite old.
> > 
> > I also agree with Michael here.
> 
> Everytime I have proposed that, for exactly that same reason, the response
> has been that the rule is to not bump anything unless required by tests. I
> guess it applies here. Anyway, I never liked that rule, so I, of course,
> agree on *always* testing with the version as close as possible to what most
> users have.

I think we shouldn't bump unless we absolutely need to, as it introduces a lot of churn in test results and forces every developer to recompile the *entire* JHBuild moduleset. It's a lot of work to update many results and the JHBuild compilation time is nothing to sneeze at. On the other hand, if we are forced to update, I don't think there is any issue updating to a recent version.

Still perhaps it is time to reexamine the rule. One idea is that it is fine to update JHBuild moduleset if the updater personally takes responsibility for carefully updating all pixel and layout test results that change as a result.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20150617/6a0980cc/attachment.html>


More information about the webkit-unassigned mailing list