[Webkit-unassigned] [Bug 61287] [UNIX] Use SOCK_SEQPACKET when available

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=61287





--- Comment #32 from Carlos Garcia Campos <cgarcia at igalia.com>  2011-05-26 09:29:13 PST ---
(In reply to comment #31)
> (From update of attachment 94982 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=94982&action=review
> 
> This approach seems optimal for systems with SOCKET_SEQPACKET.
> 
> > Source/WebKit2/Platform/CoreIPC/unix/ConnectionUnix.cpp:196
> > +        // Connection has been closed.
> > +        connectionDidClose();
> > +        return;
> > +    }
> 
> I think the method call here makes the comment redundant.

Right, I added the comment because it's not obvious that a 0-byte read means the other end closed the connection, but the method name makes it clear, so I'll remove the comment

> > Source/WebKit2/UIProcess/Launcher/gtk/ProcessLauncherGtk.cpp:47
> > +#ifdef SOCK_SEQPACKET
> > +#define SOCKET_TYPE SOCK_SEQPACKET
> > +#else
> > +#define SOCKET_TYPE SOCK_DGRAM
> > +#endif
> > +
> 
> It's probably better to use SOCK_SEQPACKET directly until we have the fallback. People packaging for platforms that don't support SOCK_SEQPACKET will have some advance notice that we currently fail for them.

Well, it will work with SOCK_DGRAM, it's just in a few corner cases that the web process might get blocked for a while.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.