421460 - table cells' height expanded incorrectly when crossing rowspanned cell

Status: UNCONFIRMED

(Core :: Layout: Tables, defect)

Description

[Roy Tam]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1.13pre) Gecko/20080302 Firefox/2.0.0.13pre (Firefox musume)
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1.13pre) Gecko/20080302 Firefox/2.0.0.13pre (Firefox musume)

I found that Gecko engine expands the next row's table cells' incorrectly when corssing a rowspaned cell. IE6/7, Safari, Opera are tested without this issue.

Reproducible: Always

Steps to Reproduce:
1.go that website.
2.view the table by eye or through DOM inspector.

Actual Results:  
the row expanded incorrectly.

Expected Results:  
the row will not expand, should look like normal row in table.

Comment 1

[Roy Tam]

Attachment: testcase

testcase atteched.
if the height of #placeholder increases, the red row's height will increase too.

Comment 2

[Mardeg]

Attachment: Standards compliance mode version

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9b3) Gecko/2008020514 Firefox/3.0b3
Confirmed, also happens in the Standards Compliance version.

Comment 3

[Roy Tam]

Attachment: testcase 2

for making workaround, I tried setting height on red row.
Yes, red row looks like other rows but the lime cells expands.

Comment 4

[Amin]

Attachment: Testcase3

Simplest testcase for this bug.

Comment 5

[Daniel Holbert [:dholbert]]

Attachment: testcase 4 (reduced, with both row orderings & col orderings)

"Ordering 1" in the attached testcase 4 is just a simplified version of testcase 3, excluding the final row & final column, which are unnecessary.

I've then repeated the testcase using different row & column orderings, to see how consistent the various browsers' behaviors are.

IIUC, the reporter's expected behavior here is for the orange cell 'a' to remain short (fixed at 20px, its specified height).  However, no browser consistently shows this behavior, when rows/columns are reversed.  Here are the behaviors I observe on testcase 4:

 Firefox 2 & 3:  orange cell is tall
 Opera & WebKit: orange cell is short on orderings 1-2, tall on orderings 3-4
 IE 8:           orange cell is medium-height. (some height goes to gray cell)

So, it appears that Opera & WebKit are the only browsers that do what the reporter wants, and they only do it when the rows are in a particular order.

Comment 6

[Daniel Holbert [:dholbert]]

I'd recommend resolving this as INVALID, since this appears to just be a case where we're consistent across row-ordering, but Opera/WebKit aren't. (as demonstrated in comment 5)

(This is based on the assumption that testcase 3 was an accurate reduction of the original testcase.)