Closed Bug 84879 Opened 24 years ago Closed 22 years ago

Crash when using color picker in preference menu

Categories

(Core :: XUL, defect)

x86
Windows 98
defect
Not set
critical

Tracking

()

RESOLVED DUPLICATE of bug 160656
mozilla1.0

People

(Reporter: doctor__j, Assigned: saari)

References

Details

(Keywords: crash)

Attachments

(3 files)

From Bugzilla Helper: User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:0.9.1+) Gecko/20010607 BuildID: 2001060809 Reproducible: Always Steps to Reproduce: 1. Open Preference menu 2. Go to "Appearance -> Colors" 3. Open the color picker to choose a color Actual Results: Pref menu title bar flashes rapidly and then crash occured. Crash message from Windows: MOZILLA caused a stack fault in module KERNEL32.DLL at 017f:bff7429f. Talkback failed to start. I can't submit a talkback crash info. Any relations to bug 64160?
wfm with win2k build 20010609.. (CVS opt)
Keywords: crash
Im seeing this on win98 build:2001060820 Looks like an infinite loop error of some sort, if I was to guess
--> cmanske Not seeing in win2k, must be 9x-specific. I'll bet the cause here is similar to bug 63452.
Assignee: mcafee → cmanske
Component: Preferences → XP Apps: GUI Features
Keywords: mozilla0.9.2
Toolkit, actually.
Component: XP Apps: GUI Features → XP Toolkit/Widgets
QA Contact: sairuh → aegis
wfm on NT4.0 with both Classic and Modern Skin. My changes to colorpicker for bug 64160 were carefully crafted to not affect the existing colorpicker. The only change that would reawaken a problem (that does very much seem like bug 63452) would be the addition of: colorpicker { -moz-user-focus: normal; } to all colorpicker.css files. But why does this only occur on Win9x !!!! What about other platforms? Please be sure to test both Classic and Modern themes, as (unfortunately), there are subtle differences in focus handling in each. Suggested interim fix would be to remove "-moz-user-focus: normal;" from colorpicker.css and add it only to Composer's CSS, because Composer is the only user of "colorpicker" without the "colorpicker-button". Thus only Composer needs the "normal" focus (without it, user can't tab to/from the colorpicker swatches in Composer's colorpicker dialog.)
Status: NEW → ASSIGNED
Target Milestone: --- → mozilla0.9.2
Attached patch Possible fix?Splinter Review
I can reproduce the crash using both Classic and Modern themes. BTW, talkback can't catch the crash... I can't even start the Talkback agent. <sigh>
Since I'm only running WinNT 4.0, I can't test this bug at all, so can someone please try the attachment "Better suggested fix?". Please test with both Classic and Modern themes, and test if Prefs -> Appearance -> color buttons work in both Browser and Composer. Also: Does anyone see any problems using Composer's colorpicker (with or without the new patch) under Win9x?
Hrm...I'm not sure I understand why this would fix it (I can't test either, though). Wasn't bug 63452 fixed by *allowing* the colorpicker to take focus, so the keystrokes went to it and not the pref tree, so the user had to dismiss the colorpicker before switching panes?
I have no idea if my suggested patch will fix this, but it gets us to very close to the state before my changes for bug 64160, so I'm trying to eliminate my changes as the cause.
I can't reproduce this on win98 using the opt build from 6/11, asking QA to verify
no can repro in Win95 (p133, 48 MB of RAM), build 20010608. tried both classic and modern
yes, i can repro this using 2001.06.11.09 comm trunk bits on a win32 machine in the testlab. i see the same flashing effect as dr_j originally reported, then crash. the test machine was running Win Me v4.90.3000; it's agonizingly slow [only 32M RAM]. unfortunately, the stack trace is paltry --no symbols :( Incident ID 31595372 Stack Signature KERNEL32.DLL + 0x4277 (0xbff64277) c243d2bd Bug ID Trigger Time 2001-06-11 15:48:16 User Comments win Me, 4.90.3000: choose color from picker Build ID 2001061109 Product ID MozillaTrunk Platform ID Win32 Stack Trace KERNEL32.DLL + 0x4277 (0xbff64277) 0x07200720
But can anyone reproduce this after applying my 6/10/01 09:20 patch?
blake, i cannot remember if you have access to WinMe, much less a build system on it? beppe, out of curiosity --was your 6/11 opt build from the branch or trunk?
We can just hand-roll cmanske's patch into a release build and then try to reproduce this on the machine where you could crash, sairuh. So, to quote an obscure song "I'll do the rolling, you do the details."
John's suggestion is good. Basically, my patch is totally safe and can't hurt! And if we're lucky, it just might help! Anyone for r=, sr=, please?
jrgm added your changes to the 2001.06.11.09 trunk build on the WinMe machine --and the color picker worked just dandy [both modern and classic, as well as both the prefs and editor color picker].
Super! Then let's get this approved and checked in.
Whiteboard: fix in hand, need r= and sr=
Why are we patching up a crash with CSS changes? It sounds like this is some kind of focus-related infinite loop (cc'ing saari). Can anyone get a stack?
I think Charley's fixes definitely should be checked in since they are supposed to be for Composer-only color pickers (not global). However, I am afraid that we are simply masking the problem. I'd like someone to generate a stack crawl or point the finger to some specific code that is causing this problem before this fix goes in (or put the fix only on a branch). conditional r=brade
I suggest we checkin the additions of #ColorPicker { -moz-user-focus: normal; } to the Editor CSS files as in the patch. This will assure that the colorpicker focus is correct for editor, but will not actually fix this bug. It will simply be a redundant rule. We then reassign (to Saari?) to investigate the real problem. If real solution can't be found before RTM, then it is easier to just remove the "-moz-user-focus: normal;" from the global CSS files as an interim fix. It is very important that the color buttons not crash in the Browser!
purify on my NT build doesn't show anything unusual... :-(
Note that this crash has only been seen on 9x systems (98/ME). I just tried to get a stack trace on a 98 build, but vc++ would not come up (at least on that machine).
The editor CSS now has the required "-moz-user-focus: normal" for its own colorpicker. This bug is being reassigned to the menu/event expert for further investigation. But it no fix is found, the patch on 06/13/01 14:47 MUST BE CHECKED IN to prevent the crash on Win9x platforms. If we do have to use that patch, this bug should be kept open for future investigation. To restore back to the "bug" condition, simply add back: colorpicker { -moz-user-focus: normal; } to colorpicker.css in themes/modern/global, and themes/classic/global/win. brade has supplied the "conditional" r= for this patch.
Assignee: cmanske → saari
Status: ASSIGNED → NEW
Whiteboard: fix in hand, need r= and sr= → fix in hand, need sr=
a sad sad sr=hewitt
Whiteboard: fix in hand, need sr= → fix in hand
Blocks: 83989
A sad, sad a=blizzard on behalf of drivers for the trunk. :(
work around checked in for RTM; moving to mozilla 1.0 reminder: to reproduce this bug you'll need to back out the attachment from 6/13/01 14:47.
Keywords: mozilla0.9.2
Whiteboard: fix in hand
Target Milestone: mozilla0.9.2 → mozilla1.0
*** Bug 86756 has been marked as a duplicate of this bug. ***
wfm win98 build 2001063008
This seems to be fixed now (build:2001070708, win98)
May God have mercy on us all. The 212 bug spam-o-rama is Now!
QA Contact: aegis → jrgm
*** Bug 98415 has been marked as a duplicate of this bug. ***
is this still a problem on win98? this is not a problem for me on winNT using 2001.11.20.09-comm bits, using the modern theme. marking wfm, but do reopen if still a problem.
Keywords: mozilla1.0
Worksforme (Build ID: 2002 01 11 03/Win98). Could we close that one now ? /jc
Marking so...
Status: NEW → RESOLVED
Closed: 23 years ago
Resolution: --- → WORKSFORME
dupe of bug 160656
Status: RESOLVED → REOPENED
Resolution: WORKSFORME → ---
*** This bug has been marked as a duplicate of 160656 ***
Status: REOPENED → RESOLVED
Closed: 23 years ago22 years ago
Resolution: --- → DUPLICATE
I disagree with this bug being a duplicate for several reasons. If it was going to be a duplicate, it should've been duplicated to bug 156486 (not a bug that itself is a duplicate). Both of those bugs were regressions that were filed well after this bug was filed and resolved as worksforme.
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: