[webkit-reviews] review requested: [Bug 36128] [Gtk] nameFromChildren is obsolete : [Attachment 58885] combined patch (code+test) to fix this bug

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Wed Jun 16 04:58:17 PDT 2010


Mario Sanchez Prada <msanchez at igalia.com> has asked  for review:
Bug 36128: [Gtk] nameFromChildren is obsolete
https://bugs.webkit.org/show_bug.cgi?id=36128

Attachment 58885: combined patch (code+test) to fix this bug
https://bugs.webkit.org/attachment.cgi?id=58885&action=review

------- Additional Comments from Mario Sanchez Prada <msanchez at igalia.com>
Attached a new version of the combined patch with a better test case, where I
got rid of the second (and really weird) timeout, to ensure that notifyDone()
was called after the post JS file being successfully loaded.

I found that just setting window.jsTestIsAsync to true is far more convenient
for this, and also avoids all the mess about dynamically loading the post JS
file to avoid calling to finishJSTest() to early, while keeping the same
functionality, so I did it that way (and also reduced a bit the code in the
test).

What I did not find yet is a way to get rid of the other timeout, which must be
there (even when is set to 0ms) to kind of simulate a g_idle_add() in JS, and
therefore ensure the layout has been properly done before checking the results.
I know this is weird but (1) after deeply checking DRT I still haven't found a
way to fix it there (if possible), and on top of that (2) I've seen more tests
needing to do the same thing so I guess the current patch proposal could be
good enough for now.

Hopefully in the future we will be able to find a more general solution in DRT
code to avoid this problems (if possible), but so far whatever I've tried
didn't work in a consistent way, and current proposal is the only one I got
that didn't break any other tests.

Eager to read some opinions/review on this last patch :-)


More information about the webkit-reviews mailing list