From Bugzilla Helper: User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; m16) Gecko/20000525 BuildID: 2000052520 the CSS Style Visibility:collapse causes a row from a table to collapse. That works fine. But if there is a table containing this table that collapses, the outside table is no longer resized to fit accordingly. (It does not shrink when the inside table is collapsed) Reproducible: Always Steps to Reproduce: 1.Go to http://www.abika.com 2.Click one of the items in the table on the left side of the screen 3.Click the same item again (to contract the sub-list) Actual Results: The outside table is expanded fine when the visibility is set to "visible". But when the visibility is set to "collapse" the outside table does not shrink back to original size. Expected Results: The outside table shrinks back to fit collapsed inside table. This is a regression which occurred between builds 200002508 and 200002520. The current build (200002608) also has the same problem.

Changing summary from "[Regression] visibility:collapse no longer causes containing table to be resized" to "[Regression] height change in nested table no longer causes containing table to be resized" since there are a lot of other conditions affected.

Nominating for nsbeta2 since it is a recent regression, it is fixed, and has little risk.

[nsbeta2+] will take fix until [6/22]

Adding testcase keyword so this doesn't show up on the bugathon search page.

Fixed.

Verified 2000061508

This is now broken again in 2000061608

I backed out the change yesterday, thinking it might have contributed to the problems. I'll probably add it back tomorrow.

Cleaning up whiteboard and changing to beta2 minus (6/22 has passed). I saw in the white-board "fix in hand," but the comment above suggests that this was fixed, then backed out, then was "going" to be fixed again. I can't quite figure out what the status is... so it would be good to summarize where we are, and possibly nominate for beta3 if needed. Thanks, Jim

Some more strangeness: Open the Simplified Testcase for this bug. Click the button 3 times. Poof! All gone! If the outer, containing table is removed, this behavior does not occur. For a more involved testcase, visit the above URL: http://www.abika.com and click any of the links in the "Information Center". All of the sub- categories are expanded when only the one clicked should be.

I've put back the patch. It was backed out at a time when the tree was in trouble and had nothing to do with it.

Verified

I haven't been following closely, so I don't know exactly when it happened, but as of 2000081020, this bug is now back.

Karnaze, any word on this? Did you back out your changes again, or is something else now causing this?

I didn't back out the fix a 2nd time.

Marking mozilla0.8.

I don't the table code well enough to say it will fix the bug, but it does look right. r=rods

sr=buster

The patch was checked in.

QA contact update

The problem no longer exist.. Marking verified on 2001-07-27-00-0.9.2 build

Comment on attachment 21019 [details] [diff] [review] revised patch (previous one is based on old code) - TallestCellGotShorter(oldCellDesAscent, initCellDesDescent, - cellMet.descent, mMaxCellDescent); + TallestCellGotShorter(oldCellDesAscent, cellMet.descent, mMaxCellDescent); Looks like there was an inadvertent typo: s/oldCellDesAscent/oldCellDesDescent/

karnaze/kin, r/sr?

Comment on attachment 101183 [details] [diff] [review] little patch to fix the typo sr=kin@netscape.com

Comment on attachment 101183 [details] [diff] [review] little patch to fix the typo r=karnaze

Comment on attachment 101183 [details] [diff] [review] little patch to fix the typo Checked in.

I'm not sure when it happened, but this is back to not working again. (See testcase)

WD, is this still an issue in current builds? The testcase worksforme in Linux trunk build 2004-08-06-05....

Ok, this is WFM with a nightly. (Mozilla 1.7.1 doesn't work)

WD, this was 'FIXED' (i.e., it took a patch to fix the problem). FYI, WFM means that no patch (or sometimes, an unknown patch elsewhere) fixed the problem.

rbs, that's exactly what happened between 1.7 and now... (this was fixed by a patch, then regressed, and then unregressed again for unknown reasons).