[Webkit-unassigned] [Bug 48110] getBoundingClientRect: Do not truncate the coordinates to integers

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=48110





--- Comment #22 from Nikolas Zimmermann <zimmermann at kde.org>  2010-11-19 05:37:01 PST ---
(In reply to comment #21)
> Created an attachment (id=74353)
 --> (https://bugs.webkit.org/attachment.cgi?id=74353&action=review) [details]
> Fix and layout tests
> 
> (In reply to comment #19)
> > Ok, new suggestion: Use the "Ahem" font, that should give you exactly the same results on all platforms (hopefully). Text will appear as black blocks (rects), all of the same width/height and advance.
> > Can you give that another try?
> 
> I did, it worked, thanks for the suggestion. One thing that I found convenient is that I no longer need two -expected files for each of the two test cases.
> 
> But I still see a virtue in not having the font-dependent coordinates hardcoded in the HTML source: in the moment when I loaded the HTML in the browser, my first instinct was, oh! now there are FAILures!
> In fact, everything was fine. It was the browser just showing the regular font (thus producing the errors), while the tests are now run using the Ahem font.
Sure, these tests are really intended to be run in DRT only, it doesn't make much sense to test these zooming things in the browser itself.
If you want to be sure you can add a:
if (!window.layoutTestController) { alert('please run in drt'); return; } on top of the test.

> 
> > I'd love to avoid SKIP ... messages, or not checking something.
> 
> But they are still being checked, using the -expected file.
Yeah, but errors are not immediately obvious, unlike where you expect to see 'PASS' messages and now get 'FAIL' messages.

> 
> One of the following two things should be done: either leave the SKIPs in there and use the expectation file to check the actual coordinates; or insert a code snippet that only allows the two tests to run in testshell, so that they don't look alarming and misleadingly incorrect in the browser. In my latest patch, I did the former, because I think that loading the HTML tests in the browser is useful. However, if you still want me to do the latter, please tell me, and I'll resubmit.
It would be great, if you would warn the user if anyone tries to execute the tests in a non-DRT environment (eg. in Safari) and replace all SKIP usages by shouldBe(...). Then you also don't need to dump the actual results in the expecaction files.

Thanks for you hard work!

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