Bug 193243 – Stop using NetworkProcess::singleton in NetworkCache code

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Alex Christensen]

Stop using NetworkProcess::singleton in NetworkCache code

[Alex Christensen]

Created attachment 358604 [details]
Patch

[Alex Christensen]

Created attachment 358609 [details]
Patch

[Brent Fulgham]

Comment on attachment 358609 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=358609&action=review

> Source/WebKit/ChangeLog:4
> +        https://bugs.webkit.org/show_bug.cgi?id=193243

Missing radar...

> Source/WebKit/NetworkProcess/cache/NetworkCache.h:166
> +    Ref<NetworkProcess> m_networkProcess;

This means that the NetworkCache will keep the NetworkProcess alive. Is that the right ownership model? (I'm not disagreeing, just want to be sure that's correct).

Otherwise we could make this a WeakPtr.

However, since it was a singleton before, I assume the expectation is we never have a cache without a process (and vice versa).

[Chris Dumez]

(In reply to Brent Fulgham from comment #3)
> Comment on attachment 358609 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=358609&action=review
> 
> > Source/WebKit/ChangeLog:4
> > +        https://bugs.webkit.org/show_bug.cgi?id=193243
> 
> Missing radar...

We do not have a requirement have having a radar, do we? radars get auto-filed when the patch gets committed.

[Alex Christensen]

The NetworkCache will soon be a member of the NetworkSession, which should keep the NetworkProcess alive.
Thanks for the review.

[Alex Christensen]

http://trac.webkit.org/r239740

[Radar WebKit Bug Importer]

<rdar://problem/47125976>