Bug 17696 – form tag causing misalignment in footer

• Component: Layout and Rendering
• Status: VERIFIED FIXED

[jasneet]

I Steps:
Go to
http://www.vente-privee.com

II Issue:
The footer link is misaligned.

III Conclusion: Removing the form tag resolves the issue.

IV Other browsers:
IE7: ok
FF: ok
Opera: ok

V Nightly tested: 30628

[jasneet]

Created attachment 19572 [details]
screenshot

[jasneet]

Created attachment 19573 [details]
reduction

[Robert Blaut]

*** Bug 17784 has been marked as a duplicate of this bug. ***

[Robert Blaut]

Created attachment 19682 [details]
even reduced test case

The test case shows problem with default margin-bottom other than 0 for form tag. The same behavior is also visible in Opera 9.5 and Minefield but only if the form is loaded inside an iframe! The issue is not visible in Gecko if the test case is loaded in browser window.

Check this test case in Minefield: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%3Cp%3EThere%20should%20be%20no%20red%20on%20the%20page%3C%2Fp%3E%3Cdiv%20%20style%3D%22background-color%3Ared%3B%20border%3A1px%20solid%22%3E%3Cform%3E%3C%2Fform%3E%3C%2Fdiv%3E%0A%0A

Click rendered view to compare both behaviors.

[Robert Blaut]

html4.css contains such definition:

form {
    display: block;
    margin: 0__qem 0 1em 0
}

margin-bottom is set to 1em.

Is it fixable? If not, the bug should be classified as an evangelism bug. 

[Robert Blaut]

Gecko allows margin-bottom: 1em in quirks mode only: https://bugzilla.mozilla.org/show_bug.cgi?id=41806 

I think we should consider the same behavior for Webkit.

BTW. It sad to say but Live DOM Viewer - the absolutely wonderful tool for debugging html/css code - has a sort of bug that makes Gecko to render the page in iframe on that page in quirks mode instead of standard mode, but it's of cause unrelated to the bug report.

[Robert Blaut]

Created attachment 19697 [details]
patch for setting margin-bottom in quirk mode only

The patch is based on Gecko solution.

[Darin Adler]

Comment on attachment 19697 [details]
patch for setting margin-bottom in quirk mode only

I think it's misleading that the quirks.css claims that the margin-bottom if a fix for bug 17696. The fix is to do the margin only in quirks mode, but the comma doesn't make that clear.

But the fix is otherwise great.

r=me

[Robert Blaut]

(In reply to comment #8)
> (From update of attachment 19697 [details] [edit])
> I think it's misleading that the quirks.css claims that the margin-bottom if a
> fix for bug 17696. The fix is to do the margin only in quirks mode, but the
> comma doesn't make that clear.

Darin, would you be so kind to correct the comment in quirks.css file before landing?

[Darin Adler]

Comment on attachment 19697 [details]
patch for setting margin-bottom in quirk mode only

When I applied this patch and ran regression tests to prepare to check in, 8 regression tests failed:

fast/block/margin-collapse/103.html
fast/frames/viewsource-empty-attribute-value.html
tables/mozilla/bugs/bug44505.html
tables/mozilla/bugs/bug51727.html
tables/mozilla/bugs/bug52505.html
tables/mozilla/bugs/bug52506.html
tables/mozilla_expected_failures/bugs/bug2479-2.html
tables/mozilla_expected_failures/bugs/bug56024.html

We'll need a new version of this patch that includes changes to the expected results for tests that we expect to change. Marking this review- until then.

[Darin Adler]

Created attachment 19810 [details]
my version of the patch (with slightly improved comment)

This patch is not ready to land because of the 8 other regression tests affected by the change.

[Robert Blaut]

(In reply to comment #11)
> Created an attachment (id=19810) [edit]
> my version of the patch (with slightly improved comment)
> 
> This patch is not ready to land because of the 8 other regression tests
> affected by the change.
> 

Text only results are no problem for me I can generated the new ones. But there are some tests with pixel results and also platform depended. So I ask for help for generating the new results for these tests.

[Robert Blaut]

Created attachment 19833 [details]
patch with darin comment and new text layout test results (without pixel and qt platform results)

Darin, this patch contains new text layout results and deletes old pixel and qt results. I tried to regenerate new pixel results bug these tests are highly system settings depended. Almost all existing pixel layout tests fails on my machine. I also changed comment in quirks.css file.

[Darin Adler]

Committed revision 31100.