Closed
Bug 1269364
Opened 8 years ago
Closed 3 years ago
m.trulia.com reliably leaks windows on firefox for android
Categories
(Firefox for Android Graveyard :: General, defect)
Firefox for Android Graveyard
General
Tracking
(fennec+)
RESOLVED
INCOMPLETE
Tracking | Status | |
---|---|---|
fennec | + | --- |
People
(Reporter: bkelly, Unassigned)
References
Details
(Whiteboard: [MemShrink:P2])
STR: 1) Launch a fresh firefox for android process. 2) Navigate to m.trulia.com 3) Click on a property 4) Scroll down the property page 5) Close the trulia window 6) Open about:memory 7) Click minimize memory usage 8) Click measure At this point you should see an entry like: top(none)/detached window(http://m.trulia.com/property/...)
Reporter | ||
Comment 1•8 years ago
|
||
I tried with Settings > Advanced > Plugins set to disabled, but it still leaked. So its probably not flash.
Updated•8 years ago
|
tracking-fennec: --- → ?
Updated•8 years ago
|
Whiteboard: [MemShrink] → [MemShrink:P2]
Updated•8 years ago
|
tracking-fennec: ? → +
Reporter | ||
Comment 2•8 years ago
|
||
Andrew, any hints on how to hunt down the "1 unknown edge" here? $ /c/devel/heapgraph/find_roots.py cc-edges.26140.1464481627.log 0xb7766000 Parsing cc-edges.26140.1464481627.log. Done loading graph. 0xc2fe8560 [FragmentOrElement (xhtml) div class='overlayContainer' (orphan) http://m.trulia.com/prop erty/3222512124-400-E-54th-St-16D-New-York-NY-10022] --[Preserved wrapper]--> 0xc42211c0 [JS Object (HTMLDivElement)] --[global]--> 0xca0de160 [JS Object (Window)] --[UnwrapDOMObject(obj)]--> 0xb7766000 [nsGlobalWindow #204 inner http://m.trulia.com/property/3 222512124-400-E-54th-St-16D-New-York-NY-10022] Root 0xc2fe8560 is a ref counted object with 1 unknown edge(s). known edges: 0xc59acda0 [Event] --[mEvent->target, mEvent->originalTarget]--> 0xc2fe8560 0xc59406a0 [Touch] --[mTarget]--> 0xc2fe8560 0xc2cfc300 [nsGenericDOMDataNode (len=5)] --[GetParent()]--> 0xc2fe8560 0xc59ad3c0 [Event] --[mEvent->target, mEvent->originalTarget]--> 0xc2fe8560 0xc3085510 [FragmentOrElement (xhtml) div class='line mbxs' (orphan) http://m.trulia.com/prop erty/3222512124-400-E-54th-St-16D-New-York-NY-10022] --[GetParent()]--> 0xc2fe8560 0xc4dd8320 [Event] --[mEvent->target, mEvent->originalTarget]--> 0xc2fe8560 0xc42211c0 [JS Object (HTMLDivElement)] --[UnwrapDOMObject(obj)]--> 0xc2fe8560 0xc2fe8150 [FragmentOrElement (xhtml) div class='propertyGroupCardView' (orphan) http://m.tru lia.com/property/3222512124-400-E-54th-St-16D-New-York-NY-10022] --[mAttrsAndChildren[i]]--> 0xc2fe8 560 0xc2e736c0 [nsGenericDOMDataNode (len=5)] --[GetParent()]--> 0xc2fe8560 0xc59404e0 [Event] --[mEvent->target, mEvent->originalTarget]--> 0xc2fe8560 0xc2f30240 [nsGenericDOMDataNode (len=2)] --[GetParent()]--> 0xc2fe8560 0xc3085060 [FragmentOrElement (xhtml) div class='line mbxs' (orphan) http://m.trulia.com/prop erty/3222512124-400-E-54th-St-16D-New-York-NY-10022] --[GetParent()]--> 0xc2fe8560 0xc41b7820 [Touch] --[mTarget]--> 0xc2fe8560 0xc4dd9f20 [Touch] --[mTarget]--> 0xc2fe8560 0xc5940be0 [Touch] --[mTarget]--> 0xc2fe8560 0xc5940710 [Event] --[mEvent->target, mEvent->originalTarget]--> 0xc2fe8560 0xc2cdf6c0 [nsGenericDOMDataNode (len=3)] --[GetParent()]--> 0xc2fe8560 0xc2fe85b0 [FragmentOrElement (xhtml) div class='line mbxs' (orphan) http://m.trulia.com/prop erty/3222512124-400-E-54th-St-16D-New-York-NY-10022] --[GetParent()]--> 0xc2fe8560 0xc4dd9c10 [Event] --[mEvent->target, mEvent->originalTarget]--> 0xc2fe8560 0xbf96d400 [Touch] --[mTarget]--> 0xc2fe8560
Flags: needinfo?(continuation)
Reporter | ||
Comment 3•8 years ago
|
||
Kats, do we have any android builds handy with c++ apz? Since I had a slow fling still active I wonder if this leak is the java apzc in fennec.
Flags: needinfo?(bugmail.mozilla)
Comment 4•8 years ago
|
||
The C++ APZ is the default in Fennec in 48+ (so currently Aurora and Nightly builds).
Flags: needinfo?(bugmail.mozilla)
Reporter | ||
Comment 5•8 years ago
|
||
Ok, I tried in latest nightly and it still reproduces. So probably unrelated to apzc.
Comment 6•8 years ago
|
||
(In reply to Ben Kelly [:bkelly] from comment #2) > Andrew, any hints on how to hunt down the "1 unknown edge" here? Like Olli said, the "easiest" thing to do is to use a debugger to watch the release of this class, then exit and see what the stacks are. You can also do refcount logging, though that's a pain. DMD also can give you a snapshot of the heap and then you can dig around and try to figure out what points to the leaking object. I've never debugged a leak on Android so I don't know how it would work there.
Flags: needinfo?(continuation)
Comment 7•3 years ago
|
||
We have completed our launch of our new Firefox on Android. The development of the new versions use GitHub for issue tracking. If the bug report still reproduces in a current version of [Firefox on Android nightly](https://play.google.com/store/apps/details?id=org.mozilla.fenix) an issue can be reported at the [Fenix GitHub project](https://github.com/mozilla-mobile/fenix/). If you want to discuss your report please use [Mozilla's chat](https://wiki.mozilla.org/Matrix#Connect_to_Matrix) server https://chat.mozilla.org and join the [#fenix](https://chat.mozilla.org/#/room/#fenix:mozilla.org) channel.
Status: NEW → RESOLVED
Closed: 3 years ago
Resolution: --- → INCOMPLETE
Assignee | ||
Updated•3 years ago
|
Product: Firefox for Android → Firefox for Android Graveyard
You need to log in
before you can comment on or make changes to this bug.
Description
•