Closed Bug 128930 Opened 22 years ago Closed 22 years ago

border-collapse: border paint -invalidate problem

Categories

(Core :: Layout: Tables, defect, P3)

x86
Linux
defect

Tracking

()

RESOLVED FIXED
mozilla1.3beta

People

(Reporter: ewv, Assigned: karnaze)

References

()

Details

Attachments

(2 files, 1 obsolete file)

 
Using a linux build from 03-04-2002, viewing attachment 41276 [details] from bug 24113, a
number of the horizontal rules stretch the entire window width. Forcing
repaints by dragging another window over or scrolling off and back on causes
these lines to randomly disappear and reappear on rules that were fine before.
I don't see this with 0.9.9. Does anyone else?
I still see it with 2002041203 win98
i am seeing this also, 2002041321/Linux ...

but it seems that this bug is a bit inconsistent ... when i scroll the page back 
and forth, sometimes there is a problem but sometimes there is not ...
I believe this is distinctly related to the border-collapse series of bugs.  An
example page is at http://blue-labs.org/mozbugs/table-lines.php, in order to
reproduce the page width lines, a minimum of three things are required to
reproduce this fairly reliably for me, and those are:

TABLE { border-collapse: collapse; }
TD { border: 12px solid black; }  
TD.yes { border: 12px solid blue;}

A table with border-collapse, and two different border colors.  With the above,
(read text on page), I can fairly reliably repeat the occurance.  If I remove
any particular of the three, I can't repeat it.

Visit the blue-labs.org main page, view either of the sendmail pages the
bluelist page, or the financial shames page.  All have tables that fairly
reliably repeat the occurance.
Summary: Horizontal rules stretch entire window width. → Horizontal rules stretch entire window width. [border-collapse and multi color borders related]
Note in http://bugzilla.mozilla.org/showattachment.cgi?attach_id=41726, there
are three example tables that produce the horizontal lines, Rules="cols"
Frame="hsides" or "border". And Frame="box", Rules="all".
Summary: Horizontal rules stretch entire window width. [border-collapse and multi color borders related] → border-collapse: border paint -invalidate problem
 Confirmed! I see the problem on Linux 7.1 only but not with WIN2K. Adding
testcase keyword.
Keywords: testcase
Priority: -- → P3
Target Milestone: --- → Future
I'm seeing what appears to be the same bug in Mozilla 2002110208 under Windows 98.
  http://www.moo.ca/789.%3A!
But I don't see the bug in the testcases presented so far for this bug.
Removing the testcase keyword. Hoping to attach one soon.
Keywords: testcase
*** Bug 163883 has been marked as a duplicate of this bug. ***
Any interest in fixing this bug yet?  It's a longstanding and rather glaring bug
for anyone who collapses borders on their tables.

The blue-labs.org URL in comment #6 shows it perfectly 100% of the time now.
>Any interest in fixing this bug yet?  

Do you volunteer???? 

Can I reasign the bug to you? 

If not, can you please take the time to query for the number of bugs in layout
tables before posting such a question. Or even better buy a C++ book and help.

What I observe is that this only occurs when you scroll slowly (i.e., using the
arrows in the scroll bar, or by dragging the elevator in the scroll bar). If you
scroll by jumping half a page (pressing PgUp/PgDn, or the grey areas in the
scroll bars), the bug does not occur.
*** Bug 182868 has been marked as a duplicate of this bug. ***
I placed at
http://lxr.mozilla.org/seamonkey/source/layout/html/style/src/nsCSSRendering.cpp#4003
to assertion verifying that rect.x and rect.width are below 4000. If you scroll
down the url then you will hit the assertion the values for rect.x look like
unitialized variables
Status: NEW → ASSIGNED
Target Milestone: Future → mozilla1.3beta
The patch fixes Bernd's assertion which is likely causing the problem, but I
can't see the problem without the patch, so I'm not sure if it fixes the bug.
Attachment #108295 - Flags: superreview?(bzbarsky)
Attachment #108295 - Flags: review?(bernd.mielke)
the patch fixes the assertions and those stray lines are gone. I think the null
for row is contradicts the assignment to row above and cell is nulled twice. I
will r' it if you could explain me the row thing. 
Attachment #108295 - Flags: superreview?(bzbarsky) → superreview+
Attachment #108295 - Flags: review?(bernd.mielke)
Attachment #108295 - Attachment is obsolete: true
Attachment #108380 - Flags: review?(bernd.mielke)
Attachment #108380 - Flags: review?(bernd.mielke) → review+
Hopefully people who expressed so much love in this bug ( see comment 12) will
finally help and test it once it is in.
Another site that shows the problem rather consistently is
http://plone.org/collector (which also affects my intranet, based on plone).
I'm using 1.2.1.
Is the patch in comment 17 in the nightlies/1.3 alpha?
Ok, I just downloaded the latest nighly for windows (2002122308) and I still see
the problem.
Yes, because the patch hasn't been checked in yet.
*** Bug 149303 has been marked as a duplicate of this bug. ***
Polite reminder, I'll be very happy and quick to test this once the patch is
checked in ;)
*** Bug 187888 has been marked as a duplicate of this bug. ***
fix checked in, thank you Chris for the permission to do that. I know thats in
principal bad style and dont want to see that to happen with my patches that are
under danger to bitrot.
Just checked with the latest nighly for windows (2003010708) and the problem
seems to be solved.
I checked with the url in bug 149303 (since it always displays the problem while
other examples do it only from time to time) as well as with my intranet and the
plone.org site.
Thanks.
seems to be fixed
Status: ASSIGNED → RESOLVED
Closed: 22 years ago
Resolution: --- → FIXED
*** Bug 197222 has been marked as a duplicate of this bug. ***
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: