Closed Bug 19187 Opened 21 years ago Closed 20 years ago

[dogfood]Memory leaks leading to crash with dsMemFullErr.

Categories

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

PowerPC
Mac System 8.5
defect

Tracking

()

VERIFIED FIXED

People

(Reporter: sdagley, Assigned: sfraser_bugs)

References

()

Details

(Whiteboard: [PDT+][12/10])

Attachments

(1 obsolete file)

Loading http://www.stomped.com/articles/q3acr1.html seems to consume lots of
memory and crash consistently for me.  If you don't crash on the first load try
loading one of the character renderings from the links in the article then going
back.
Status: NEW → ASSIGNED
Summary: Memory leaks leading to crash with dsMemFullErr. → Memory leaks leading to crash with dsMemFullErr.
This is GWorld badness. Image leaks (bugs 15818, 15820, 15821) are part of the
raeson, but we should still not crash.
Depends on: 15818
Depends on: 15820
Depends on: 15821
Target Milestone: M12
Severity: normal → critical
Summary: Memory leaks leading to crash with dsMemFullErr. → [dogfood]Memory leaks leading to crash with dsMemFullErr.
Making dogfood. This bug causes the browser to be unusable for more than a few
minutes on Mac.
Whiteboard: [PDT+]
Putting on PDT+ radar.
Status: ASSIGNED → RESOLVED
Closed: 20 years ago
Resolution: --- → FIXED
This will have been fixed by the fixes for 19114. I verified using ZoneRanger,
and the debugger, that we are actually freeing image data now.
Whiteboard: [PDT+] → [PDT+] 12/6: Requested verification by reporter
sdagley@netscape.com:

I need to get this verified today if possible. I do not have sufficient info to
tract leaks. Can you take a look at this and, if it is fixed, verify it so?
Thank.
If sfraser sez he verified it that's good enough for me.  Of course QA is the
ultimate authority on marking things VERIFIED so I suspect they'll test it
themselves and they can ask me if they have any questions on the repro steps.
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
pink says he's still getting crashes, so I'm reopening.
Status: REOPENED → ASSIGNED
clearing reopen and assigning
Depends on: 20743
Whiteboard: [PDT+] 12/6: Requested verification by reporter
Target Milestone: M12 → M13
I'm gonna take this off the PDT list, since a real solution requires some low-
memory handling work (bug 20743).
How frequent are the crashes?  Less frequent PDT-, more frequent...impacting
daily work PDT+...fill us in please.  Is there a workaround by increasing the
partition size?
Yes, increasing the partition size, so that the amount of free memory just after
startup is > 2Mb, should work. Only some people have reported crashes
(pinkerton).
Whiteboard: [PDT+]
Target Milestone: M13 → M12
Cool - Putting on PDT+ radar.  Let's crank up the partition for dogfood. Moving
to M12 per chofmann.
Whiteboard: [PDT+] → [PDT+][12/10]
Status: ASSIGNED → RESOLVED
Closed: 20 years ago20 years ago
Resolution: --- → FIXED
This bug has been fixed by kmcclusk's change to fix bug 19114 -- reducing the
imglib cache from 2 to 1Mb. I don't need to do anything.
Status: RESOLVED → VERIFIED
Marking verified per last comments.
Attachment #8738796 - Flags: review?(philip.chee)
Attachment #8738796 - Flags: review?(mkmelin+mozilla)
Attachment #8738796 - Flags: review?(clokep)
Oh dear, bzexport messed up and picked the wrong bug number.
Attachment #8738796 - Attachment is obsolete: true
Attachment #8738796 - Flags: review?(philip.chee)
Attachment #8738796 - Flags: review?(mkmelin+mozilla)
Attachment #8738796 - Flags: review?(clokep)
You need to log in before you can comment on or make changes to this bug.