Closed
Bug 383692
Opened 17 years ago
Closed 17 years ago
hixie's twitter status script causes firefox to leak memory
Categories
(Core :: DOM: Core & HTML, defect)
Tracking
()
RESOLVED
DUPLICATE
of bug 373462
People
(Reporter: moco, Assigned: sicking)
References
()
Details
Attachments
(1 file, 2 obsolete files)
701 bytes,
text/html
|
Details |
hixie's twitter status script causes firefox to leak memory See http://www.whatwg.org/ I've got a reduced test case, which I'll attach.
Reporter | ||
Comment 1•17 years ago
|
||
Reporter | ||
Comment 2•17 years ago
|
||
I may have overly reduced that test case, working on it...
Reporter | ||
Comment 3•17 years ago
|
||
Attachment #267680 -
Attachment is obsolete: true
Assignee | ||
Updated•17 years ago
|
Flags: blocking-firefox3?
Assignee | ||
Comment 4•17 years ago
|
||
You probably want to add an end </script> to that testcase. This looks like a DOM bug, so moving.
Component: General → DOM
Flags: blocking-firefox3?
Product: Firefox → Core
QA Contact: general → general
Assignee | ||
Updated•17 years ago
|
Flags: blocking1.9?
Reporter | ||
Comment 5•17 years ago
|
||
Attachment #267684 -
Attachment is obsolete: true
Reporter | ||
Comment 6•17 years ago
|
||
> You probably want to add an end </script> to that testcase Thanks Jonas. Note, that Hixie wasn't so sloppy in his original HTML, that was all me. Also note with the WHATWG site, I am not logged into twitter. And on my console when visit http://www.whatwg.org, I get: JavaScript error: http://www.twitter.com/t/status/user_timeline/3963051?callback =twitterCallback&count=1, line 1: syntax error
Assignee | ||
Updated•17 years ago
|
Assignee: nobody → jonas
Flags: blocking1.9? → blocking1.9+
Comment 7•17 years ago
|
||
I can't reproduce this. Running the testcase increases the mem usage a bit for a while, but after that mem usage stabilizes. And when closing the browser XPCOM_MEM_LEAK_LOG doesn't show any leaks related to the testcase. First I thought this might depend on bug 385322, but apparently not.
Depends on: 385322
Comment 8•17 years ago
|
||
Seth, can you still reproduce the leak?
Reporter | ||
Comment 9•17 years ago
|
||
Olli, sorry for the delay. I wasn't seeing it either, until I opened up the error console. Could this leak really be a leak in the error console? This test cause (and hixie's original page) did flood the error console with messages like: Error: syntax error Source File: about:blank Line: 1 Source Code: <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"><html><head><title></title></head><body></body></html> Security Error: Content at about:blank may not load data from https://bugzilla.mozilla.org/attachment.cgi?id=267720.
Comment 10•17 years ago
|
||
I see the mem usage going up when error console is opened. This is a dup of bug 385322 (and the patch for bug 373462 seems to fix this one too).
Status: NEW → RESOLVED
Closed: 17 years ago
Resolution: --- → DUPLICATE
Comment 11•17 years ago
|
||
I don't think this is an exact duplicate of bug 385322, which seems to be about events. This bug on the other hand seems to be about XMLHttpRequest, which currently does GC from OnStopRequest. We should consider changing that to do cycle collection instead.
Comment 12•17 years ago
|
||
Why you think this is about xhr? The testcase at least isn't using any xhr objects, and I don't think the error console is using those either.
Comment 13•17 years ago
|
||
Ah, hmm, I missed the reduced testcase. You might want to change the summary somewhat to focus the bug better. It'd be nice to know what exactly about this testcase and the error console cause the leak (and I gather there's still no shutdown leak?). I suppose there's a cycle not being collected?
Updated•17 years ago
|
Status: RESOLVED → REOPENED
Resolution: DUPLICATE → ---
Assignee | ||
Updated•17 years ago
|
Comment 14•17 years ago
|
||
This is actually a dup of bug 373462, or at least sort of. Fixed anyway.
Status: REOPENED → RESOLVED
Closed: 17 years ago → 17 years ago
Resolution: --- → DUPLICATE
Updated•5 years ago
|
Component: DOM → DOM: Core & HTML
You need to log in
before you can comment on or make changes to this bug.
Description
•