[Webkit-unassigned] [Bug 11645] Table with percentage column widths doesn't scale to fill the entire width of a table containing it

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=11645





--- Comment #14 from Arpita Bahuguna <arpitabahuguna at gmail.com>  2012-10-05 04:14:57 PST ---
(In reply to comment #12)
> (From update of attachment 162957 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=162957&action=review
> 
Thanks for the review Julien.

> The code change is fine, AFAICT the test is progressing but would like to see another round with the proper justifications.
> 
> > LayoutTests/ChangeLog:14
> > +        * platform/chromium-linux/fast/table/025-expected.png:
> > +        * platform/chromium-win/fast/table/025-expected.png:
> 
> You should only need the chromium-linux baseline here. The fact that it fails leads me to think that your machine's configuration doesn't match the EWS. Unfortunately, they don't upload their baselines to the bugs anymore and I don't know if those are available somewhere.
> 
I faced the same problem with 13598 as well and tried several combinations but all attempts failed.
Meanwhile, I added the expected files for the qt port (they show similar diff as cr-linux) and have added this test in the chromium testexpectations file to be rebaselined later.

> > LayoutTests/ChangeLog:16
> > +        Existing test case's results have been modified as per this change.
> 
> As pointed out by Dan, you should explain what the difference is and why this is a progression.
> 
> As written, this line answers none of those questions and thus adds no value to the reader.
> 
Have added explanation for the above in the changelog.

> > LayoutTests/fast/table/scale-nested-percent-width-cols-expected.html:13
> > +                        <table style="width: 100%; background: green;">
> 
> The difference between the expected and the normal result is really not clear cut which makes it hard to say if it's correct.
> 
Normally, i.e. without the fix, I would not expect my innermost table (with the green bg color to expand up to its containing table. This would make the containing table's bg color (which is red) to cover most of the table dimensions.
After the fix, since our innermost table would now scale, the green bg color of this table would now mask that of the containing tables.

> Also if you used to not scale when you encountered any percent width, the test as written would pass while actually failing. The expected result should probably have fixed lengths to avoid any such ambiguity.
I understand how it can perhaps be confusing, have thus changed the widths to fixed for my expected test.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.