[Bug 17195] NamedPipe datagrams need to be _really_ datagrams

wine-bugs at winehq.org wine-bugs at winehq.org
Mon Dec 19 14:05:56 CST 2011


http://bugs.winehq.org/show_bug.cgi?id=17195

--- Comment #52 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> 2011-12-19 14:05:56 CST ---
> There's no reason to make the server multithreaded. All blocking calls can be
> implemented with a non-blocking state machine, including calls to userspace,
> that BTW the server already supports.

 i'm aware that it's a non-blocking state machine, alexandre.  that's
absolutely
 required, and it is good that it is a non-blocking state machine, alexandre.


 however, unfortunately, in order to get a proper implementation of NamedPipes,
 it is necessary to use blocking calls on sockets.

 that is simply and unavoidable a fact.  there's no point in trying to deny
this
 fact.  it is simply a fact.

 the only POSIX system call that can be used - period - is a blocking call
 on a socket.

 fact.

 now.

 how, exactly, when there is only one thread / process in wineserver, which,
 if it is blocked because it is waiting for a response from a socket,
 will it be possible to service any other operations?

 answer: more threads are needed... *in wineserver*.

 so, please can i ask you the following:

 1) you stop implying that i am a fucking moron.  it'll get us nowhere, fast.

 2) that you evaluate whether there are *any* other options - i don't believe
    that there are - any other POSIX functions and/or tricks available.

 please feel free to entirely ignore both these requests.  i don't have to
 put up with being told that i don't know what i'm talking about, and you
 should know by now not to do that.

-- 
Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email
Do not reply to this email, post in Bugzilla using the
above URL to reply.
------- You are receiving this mail because: -------
You are watching all bug changes.



More information about the wine-bugs mailing list