Bug 107710 – Empty P in table cells elements are not ignored

• Component: Tables
• Status: NEW

[Marcelo]

Created attachment 184271 [details]
Example of empty tag P that's not ignored

Sometimes WebKit is not ignoring empty P elements.
According W3C (http://www.w3.org/TR/html401/struct/text.html#h-9.3.1) "User agents should ignore empty P elements."
I've attached a test case reduction.

[Alexey Proskuryakov]

HTML 4.0.1 is obsolete, we do not consider it authoritative.

However, this doesn't happen in Firefox, so seems worth a look as a compatibility issue.

Created attachment 213245 [details]
Patch

[Build Bot]

Comment on attachment 213245 [details]
Patch

Attachment 213245 [details] did not pass mac-ews (mac):
Output: http://webkit-queues.appspot.com/results/3127093

New failing tests:
fast/table/cell-absolute-child.html
fast/table/table-insert-before-non-anonymous-block.html
http/tests/misc/acid3.html
tables/mozilla/bugs/bug3037-1.html
tables/mozilla_expected_failures/bugs/bug101759.html
css2.1/20110323/abspos-containing-block-initial-004e.htm
fast/table/table-section-split-with-after-content.html
css2.1/20110323/abspos-containing-block-initial-004f.htm

[Build Bot]

Created attachment 213262 [details]
Archive of layout-test-results from webkit-ews-07 for mac-mountainlion

The attached test failures were seen while running run-webkit-tests on the mac-ews.
Bot: webkit-ews-07  Port: mac-mountainlion  Platform: Mac OS X 10.8.5

[Build Bot]

Comment on attachment 213245 [details]
Patch

Attachment 213245 [details] did not pass mac-ews (mac):
Output: http://webkit-queues.appspot.com/results/3127106

New failing tests:
fast/table/cell-absolute-child.html
fast/table/table-insert-before-non-anonymous-block.html
http/tests/misc/acid3.html
tables/mozilla/bugs/bug3037-1.html
tables/mozilla_expected_failures/bugs/bug101759.html
css2.1/20110323/abspos-containing-block-initial-004e.htm
fast/table/table-section-split-with-after-content.html
css2.1/20110323/abspos-containing-block-initial-004f.htm

[Build Bot]

Created attachment 213269 [details]
Archive of layout-test-results from webkit-ews-04 for mac-mountainlion

The attached test failures were seen while running run-webkit-tests on the mac-ews.
Bot: webkit-ews-04  Port: mac-mountainlion  Platform: Mac OS X 10.8.5

[Build Bot]

Comment on attachment 213245 [details]
Patch

Attachment 213245 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.appspot.com/results/3126031

New failing tests:
fast/table/cell-absolute-child.html
fast/table/table-insert-before-non-anonymous-block.html
http/tests/misc/acid3.html
tables/mozilla/bugs/bug3037-1.html
tables/mozilla_expected_failures/bugs/bug101759.html
css2.1/20110323/abspos-containing-block-initial-004e.htm
fast/table/table-section-split-with-after-content.html
css2.1/20110323/abspos-containing-block-initial-004f.htm

[Build Bot]

Created attachment 213281 [details]
Archive of layout-test-results from webkit-ews-15 for mac-mountainlion-wk2

The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: webkit-ews-15  Port: mac-mountainlion-wk2  Platform: Mac OS X 10.8.5

[Darin Adler]

Comment on attachment 213245 [details]
Patch

Change is possibly correct, but affects the following regression tests:

  css2.1/20110323/abspos-containing-block-initial-004e.htm [ Failure ]
  css2.1/20110323/abspos-containing-block-initial-004f.htm [ Failure ]
  fast/table/cell-absolute-child.html [ Failure ]
  fast/table/table-insert-before-non-anonymous-block.html [ Failure ]
  fast/table/table-section-split-with-after-content.html [ Failure ]
  http/tests/misc/acid3.html [ Failure ]
  tables/mozilla/bugs/bug3037-1.html [ Failure ]
  tables/mozilla_expected_failures/bugs/bug101759.html [ Failure ]

Need a new patch that changes expected results for those tests.

[Gurpreet]

>   tables/mozilla_expected_failures/bugs/bug101759.html [ Failure ]
Hi Darin. Thanks a lot for reviewing the patch. But for the above test case the actual result after my patch shows the wrong output as compared to mozilla. The existing output is correct and same as mozilla's behaviour.
So I thought maybe this patch is not correct :(

[Gurpreet]

For example for the below content the first td gets a width of 1% and remaining is set for the second td
<!DOCTYPE html>
<html>
<body>
<table cellspacing="0" cellpadding="0">
 <tr>
  <td width="1%" bgcolor="blue">XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX</td>
  <td bgcolor="green"><p></p></td>
 </tr>
</table>
</body>
</html>

And for the below content the first td takes up the full width and second td background color is not shown
<!DOCTYPE html>
<html>
<body>
<table cellspacing="0" cellpadding="0">
 <tr>
  <td bgcolor="blue">XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX</td>
  <td bgcolor="green"><p></p></td>
 </tr>
</table>
</body>
So which means if width is set the td then the empty td gets the remaining width.
Mozilla's behaviour is confusing :(
</html>