Closed Bug 152557 Opened 23 years ago Closed 22 years ago

crash while accepting cookies

Categories

(Core :: Networking: Cookies, defect)

Product:
Core
Component:
Networking: Cookies
Platform:
x86
Linux
Type:
defect
Priority:
Not set
Severity:
critical

Tracking

()

Status:
VERIFIED WORKSFORME

People

(Reporter: ewilson, Assigned: morse)

References

(
URL
)

Details

(Keywords: crash)

From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0rc2) Gecko/20020510 BuildID: 2002051013 I'm afraid this is not easily repeatable. In the Preference->Privacy->Cookies config pane I have: -enable cookies for the originating web site -Disable cookies in mail and News. -Ask before storing cookies -Limit max life of cookie - 14 days On some sites, the popup dialog box for aprroving the cookie pops up as expected. In most cases I can approve/disapprove the cookie and continue. In other instances, including, but not limited to an IBM.com URL, the cookie approval popup display as expected. Once I click on yes(accept the cookie) the entire browser, all windows, crashes. It apears to have stored the cookie, because once the browser is restarted, I can go back the site and I am not asked for the cookie approval, and can browse on. Reproducible: Sometimes Steps to Reproduce: 1.see above 2. 3. Actual Results: Required to restart the browser. Expected Results: store the cookie and continue.
WFM 1.1 alpha 2002061712 Windows 98. Reporter, please install Mozilla 1.1 alpha and create a new profile. Can you then reproduce the problem?
URL: http:// multiple → http://IBM.com
Keywords: crash
I've been searching for days on here trying to figure out if I'm the only one who's seeing this. I'm on a Compaq Armada 7400 laptop running RH 7.1 (up2date'd) and I'm finding it a lot. It is actually to the point where it's almost unusable. I can agree with the information Mr Wilson has provided so far with one addition, when I was running 0.99, I was getting this too (with an error regarding a stack boundry of some form). I upgraded to 1.0 and am getting repeats of it. I have also noticed that when the crash is coming (it is not immediate) the drive can be heard thrashing then "blink" all windows (for Mozilla) are gone. I also agree with him that it seems that the subsequent encounter with that cookie goes fine. Lastly, if I had a 'debug' version of the browser, I'd be glad to capture what I can when the crash occurrs. I have wondered if the problem is associated with how long it takes to respond to the cookie storage request but I have no proof of that.
WFM, BuildID 2002061717 (trunk) on RedHat Linux 7.3+
If it is helpfull, my config is: Redhat 7.3 IBM ThinkPad A30p Intel P III,1.4 Ghz 512 Meg RAM The Mozilla is an RPM'ed version: $ rpm -qa | grep mozilla mozilla-nspr-1.0rc2-0 mozilla-nss-1.0rc2-0 mozilla-devel-1.0rc2-0 mozilla-chat-1.0rc2-0 mozilla-mail-1.0rc2-0 mozilla-psm-1.0rc2-0 mozilla-1.0rc2-0 mozilla-js-debugger-1.0rc2-0 mozilla-nss-devel-1.0rc2-0 mozilla-dom-inspector-1.0rc2-0 mozilla-nspr-devel-1.0rc2-0 The first statement from Bugzilla is to upgrade to Mozilla 1.1 Alpha. Two reasons make me hesitant to do so: 1.) The RPM databases keeps the software auditable. From an app dev standpoint that may be a lame point, but from a commerical / security / software audit standpoint it becomes a requirement. 2.) There is nothing in the response to suggest the effort would solve the problem. I will study the conditions under which Mozilla crashes, to produce a test scenario.
I agree about using RPM versions. I've not changed from the 1.0 because of the same reasons.
more notes: can crash even when accepting images or cookies. Once restarted, browser can access the site at which it crashed, as if nothing heppened.
URL: http://IBM.com → http://IBM.com, http://pricewatch.com...
Upgraded to latest known available RPMS: # rpm -qa | egrep moz mozilla-nspr-devel-1.0.0-9 mozilla-chat-1.0.0-9 mozilla-psm-1.0.0-9 mozilla-1.0.0-9 mozilla-js-debugger-1.0.0-9 mozilla-nss-devel-1.0.0-9 mozilla-nspr-1.0.0-9 mozilla-dom-inspector-1.0.0-9 mozilla-nss-1.0.0-9 mozilla-devel-1.0.0-9 mozilla-mail-1.0.0-9 Still fails in the same manor.
Wilson, Tiernan, Can the two of you try using a clean profile. Is the problem still occuring in that case?
Stephen; Can you be more specific with regard to a clean profile? I have tried with cache off, disk and mem cache cleared, and no proxy. I do not see any options for the Netscape Profile manager type of profile, if that is what you are referring. Cheers; E!
Enter profile manager and create a new profile. That's what I meant by a clean profile.
I've done as you asked, I (after backing up) nuked my .mozilla subdir and then started from scratch with a new user ID. It took a bit of time but it finally did it. What I'll try to do is capture a core file for you (if you want).
Clarification, I've gotten it to crash.
> What I'll try to do is capture a core file for you (if you want). Did you submit a talkback crash report?
Two pieces of follow up, 1) I dont' know how to do a 'talk-back' and can't for the life of me remember how to force a core-dump when it **** out. 2) The error message I was trying to find FINALLY occurred, I was always seeing "Detected an attempt to write across stack boundary." when it crashes.
RPM builds are not talkback enabled. Mozilla will not dump core because of bug 148453. the RPM also disables core-dumping. to get a stack with an RPM build, you have to use gdb. it would be easier to just grab a nightly (non-RPM) talkback-enabled build, dump it in a directory, run/crash/etc, and then nuke the directory if you don't like non-rpm apps. Mozilla does not stick its fingers outside of its own directory.
Worksforme, Mozilla 1.3 alpha 2002121204, Windows 98. Reporter, please download and install Mozilla 1.2.1 or later and try this problem again. Does it still crash? Please also test it with a fresh profile. You can create a new profile by running Mozilla with this command: mozilla -profilemanager
URL: http://IBM.com, http://pricewatch.com... → http://IBM.com
God help me, I thought that this bug entry had been closed. The problem seemed to have been resolved with V1.0 and I've not seen it since. What actions must I take to close out my part of this bug?
then close it.
Status: UNCONFIRMED → RESOLVED
Closed: 22 years ago
Resolution: --- → WORKSFORME
worksforme too, verified - linux rh6 12/13 builds
Status: RESOLVED → VERIFIED
You need to log in before you can comment on or make changes to this bug.