Bug 117328 - Completely initialise ChildProcess state on launch
Summary: Completely initialise ChildProcess state on launch
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: 528+ (Nightly build)
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Oliver Hunt
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2013-06-06 17:53 PDT by Oliver Hunt
Modified: 2013-06-06 18:12 PDT (History)
0 users

See Also:


Attachments
Patch (1.30 KB, patch)
2013-06-06 17:53 PDT, Oliver Hunt
no flags Details | Formatted Diff | Diff
Patch (1.50 KB, patch)
2013-06-06 18:03 PDT, Oliver Hunt
sam: review+
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Oliver Hunt 2013-06-06 17:53:04 PDT
Completely initialise ChildProcess state on launch
Comment 1 Oliver Hunt 2013-06-06 17:53:41 PDT
Created attachment 203984 [details]
Patch
Comment 2 Sam Weinig 2013-06-06 17:59:59 PDT
Comment on attachment 203984 [details]
Patch

Process initialization should happen in ChildProcess::initialize() or one of its subclasses.  It seems like setProcessSuppressionEnabled(false); is already called from ChildProcess::platformInitialize(), why is that not enough?
Comment 3 Oliver Hunt 2013-06-06 18:03:07 PDT
Created attachment 203986 [details]
Patch
Comment 4 Sam Weinig 2013-06-06 18:07:36 PDT
Comment on attachment 203986 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=203986&action=review

> Source/WebKit2/ChangeLog:8
> +        Don't rely on receiving a message from the UI process as part of our initialisation

I don't think this make sense.

> Source/WebKit2/Shared/mac/ChildProcessMac.mm:151
> -    setProcessSuppressionEnabled(false);
> +    setProcessSuppressionEnabledInternal(false);

You should note why you are using the weird internal variant here.
Comment 5 Oliver Hunt 2013-06-06 18:12:26 PDT
Committed r151305: <http://trac.webkit.org/changeset/151305>