Bug 201830 – NetworkProcessConnection does not need HashMaps of sessionIDs for Service Workers and IDB

• Status: RESOLVED
• Resolution: FIXED
• Component: WebKit2

[Chris Dumez]

NetworkProcessConnection does not need HashMap of sessionIDs for Service Workers and IDB, since we now have one sessionID per WebContent process.

[Chris Dumez]

Created attachment 378878 [details]
Patch

[Chris Dumez]

Created attachment 378879 [details]
Patch

[Chris Dumez]

Created attachment 378880 [details]
Patch

[Chris Dumez]

Created attachment 378885 [details]
Patch

[Geoffrey Garen]

Comment on attachment 378885 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=378885&action=review

r=me

> Source/WebKit/NetworkProcess/IndexedDB/WebIDBConnectionToClient.h:76
> +    uint64_t messageSenderDestinationID() const final { return 0; }

Are these zero accessors going to be removed in a future patch?

[Chris Dumez]

Comment on attachment 378885 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=378885&action=review

>> Source/WebKit/NetworkProcess/IndexedDB/WebIDBConnectionToClient.h:76
>> +    uint64_t messageSenderDestinationID() const final { return 0; }
> 
> Are these zero accessors going to be removed in a future patch?

I don't think they can. messageSenderDestinationID() is pure virtual on IPC::MessageSender so we have to provide an implementation. It is equivalent to the many send() calls we have that pass 0 as destinationID as last parameter (because there is only one destination on this IPC connection).

[Chris Dumez]

Comment on attachment 378885 [details]
Patch

Clearing flags on attachment: 378885

Committed r249923: <https://trac.webkit.org/changeset/249923>

[Chris Dumez]

All reviewed patches have been landed.  Closing bug.

[Radar WebKit Bug Importer]

<rdar://problem/55415603>