New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 648317 link

Starred by 1 user

Issue metadata

Status: WontFix
Owner:
Closed: Nov 2016
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 1
Type: Bug



Sign in to add a comment

Chrome_Blimp_Engine: Crash Report - AddRenderWidget

Project Member Reported by scf@chromium.org, Sep 19 2016

Issue description

Product name: Chrome_Blimp_Engine
Magic Signature: AddRenderWidget

Current link:
https://crash.corp.google.com/browse?q=product.name%3D'Chrome_Blimp_Engine'%20AND%20custom_data.ChromeCrashProto.ptype%3D'browser'%20AND%20(product.version%3D'55.0.2858.0'%20OR%20product.version%3D'55.0.2860.0')%20AND%20custom_data.ChromeCrashProto.magic_signature_1.name%3D'AddRenderWidget'%20AND%20ReportID%3D'c503897500000000'&ignore_case=false&enable_rewrite=true&omit_field_name=&omit_field_value=&omit_field_opt=%3D#3


Search properties:
product.name: Chrome_Blimp_Engine
custom_data.chromecrashproto.ptype: browser
product.version: 55.0.2858.0' OR product.version='55.0.2860.0
custom_data.chromecrashproto.magic_signature_1.name: AddRenderWidget
reportid: c503897500000000

Metadata :
Product Name: Chrome_Blimp_Engine
Product Version: 55.0.2858.0
Report ID: c503897500000000
Report Time: Mon, 12 Sep 2016 23:08:22 GMT
Uptime: 230060 ms
Cumulative Uptime: 0 ms
User Email: 
OS Name: Linux
OS Version: 0.0.0 Linux 3.16.0-4-amd64 #1 SMP Debian 3.16.7-ckt20-1+deb8u4google (2016-01-26) x86_64
CPU Architecture: amd64
CPU Info: family 6 model 63 stepping 0


 

Comment 1 by scf@chromium.org, Nov 7 2016

Owner: ----
Putting back into the up for grabs. After talking with @kushal this might need some rewiring of some client events.
Cc: scf@chromium.org khushals...@chromium.org
Owner: steimel@chromium.org
Status: Assigned (was: Untriaged)
+cc khushalsagar
scf@, can you add more clarifying detail and/or repro steps to this bug?

Comment 3 by w...@chromium.org, Nov 15 2016

Labels: M-57
khushalsagar@, scf@: I looked at this a bit and IIUC AddRenderWidget DCHECKs that the same widget/id is never added twice. However, when we navigate back, it reinitializes the old render widget, which refires the "Created" events, which calls AddRenderWidget to these widgets that were already added, violating the DCHECKs. 

Assuming all of that is correct, I'm not sure which piece is incorrect here. Are the DCHECKs invalid? Should the "Created" events not be firing on reinitialization? Should we not be reinitializing?

Any insights you guys have would be greatly appreciated

Thanks! 
The same widget can be added twice. It just DCHECKs that the Create and Destroy calls come in pairs, so you don't try to re-create a compositor for the same RenderWidget on the client. Does navigating back to the old widget fire a Create again? I thought it would only give a OnRenderWidgetInitialized event?

If there are cases where we can get a Create call again, then how about just early outing when that happens then?
Status: WontFix (was: Assigned)
Obsolete, WontFix.
Labels: Archive-Blimp

Sign in to add a comment