Bug 5771 – WebKit does not implicitly close anchor tags when its containing block ends

• Status: RESOLVED
• Resolution: CONFIGURATION CHANGED
• Component: Layout and Rendering

[Daniel Udey]

When an anchor tag is not closed but its containing element closes, WebKit continues to treat all text 
afterwards as a link regardless. From my understanding of the HTML spec, the anchor tag should be 
implicitly closed when its containing block ends, and this behaviour is thus incorrect.

[Daniel Udey]

Created attachment 4725 [details]
Testcase for bug 5771, the unclosing link bug

[Eric Seidel (no email)]

Related to 3905  They don't really depend, but there is no other good way to show relations in bugzilla.

[Darin Adler]

We actually go out of our way to re-open the anchor tag, using the "residual style" mechanism. It should be possible to change the rules around.

[Ryan Schmidt]

Created attachment 10141 [details]
test case 2

This bug has been responsible for some weirdness over at http://pecl.php.net/ :

http://pear.php.net/bugs/bug.php?id=7183

The original test case attached to this bug looks the same to me in Safari and Firefox, so I don't know how wrong the behavior is. But I have a test case which shows how Safari is behaving differently from Firefox, attached as "test case 2." In Safari, the entire document is red, but in Firefox, the text "not a link 1" is shown in black, because Firefox has auto-closed the open anchor before going into the table. And presumably Safari should do the same.

[Robert Hogan]

The bug is now that 'not a link 1' in test case 2 displays red but with no underline. So close, but no cigar.

[Brent Fulgham]

Testcase 2 still shows "not a link 1" as not underlined, while Firefox and Chrome show with an underline.

[Radar WebKit Bug Importer]

<rdar://problem/96543957>

[Ahmad Saleem]

Created attachment 468048 [details]
STP179, WebKit ToT, Chrome Canary 119 and FX120

All matching each other.