Bug 238800 – In case of COOP-based process swap, we need to make sure the document gets controlled by its matching service worker registration

• Status: RESOLVED
• Resolution: FIXED
• Component: Service Workers

[youenn fablet]

In case of COOP-based process swap, we need to make sure the document gets controlled by its matching service worker registration

[Radar WebKit Bug Importer]

<rdar://problem/91288849>

[youenn fablet]

Created attachment 456705 [details]
Patch

[Chris Dumez]

Comment on attachment 456705 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=456705&action=review

r=me

> Source/WebKit/NetworkProcess/NetworkLoadParameters.h:47
> +    mutable WebPageProxyIdentifier webPageProxyID;

These mutable seem a little odd. Maybe we should imply not mark the NetworkResourceLoader data member as const?

[youenn fablet]

(In reply to Chris Dumez from comment #3)
> Comment on attachment 456705 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=456705&action=review
> 
> r=me
> 
> > Source/WebKit/NetworkProcess/NetworkLoadParameters.h:47
> > +    mutable WebPageProxyIdentifier webPageProxyID;
> 
> These mutable seem a little odd. Maybe we should imply not mark the
> NetworkResourceLoader data member as const?

There used to be no mutable, I guess we should remove them, and const_cast m_parameters when needed instead. Or make the member no longer const.

[youenn fablet]

> There used to be no mutable, I guess we should remove them, and const_cast
> m_parameters when needed instead. Or make the member no longer const.

I'll look into cleaning all these mutables in a follow-up.

[Chris Dumez]

(In reply to youenn fablet from comment #4)
> (In reply to Chris Dumez from comment #3)
> > Comment on attachment 456705 [details]
> > Patch
> > 
> > View in context:
> > https://bugs.webkit.org/attachment.cgi?id=456705&action=review
> > 
> > r=me
> > 
> > > Source/WebKit/NetworkProcess/NetworkLoadParameters.h:47
> > > +    mutable WebPageProxyIdentifier webPageProxyID;
> > 
> > These mutable seem a little odd. Maybe we should imply not mark the
> > NetworkResourceLoader data member as const?
> 
> There used to be no mutable, I guess we should remove them, and const_cast
> m_parameters when needed instead. Or make the member no longer const.

Right, I think keeping the m_parameters const at this point no longer makes sense since we're updating its data member. Mutable has some use cases (e.g. lazy init) but IMO this isn't a good use case for it.

[EWS]

Committed r292468 (249319@main): <https://commits.webkit.org/249319@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 456705 [details].

[Brent Fulgham]

This fix shipped with Safari 15.5 (all platforms).