This is the mail archive of the
cygwin@sourceware.cygnus.com
mailing list for the Cygwin project.
RE: New patch
- To: Sergey Okhapkin <sos at buggy dot prospect dot com dot ru>, "'Gnu-Win32'" <gnu-win32 at cygnus dot com>
- Subject: RE: New patch
- From: "Parker, Ron" <rdparker at butlermfg dot org>
- Date: Wed, 22 Apr 1998 10:45:37 -0500
- Cc: "'cygwin32-developers at cygnus dot com'" <cygwin32-developers at cygnus dot com>
No we don't really need it. I was wondering why you implemented this
in the first place. Using select() explains it all. I agree that
PeekMessage should be used.
I thought you were trying to replace the entire message loop
functionality with a UNIX like interface. (^8
-----Original Message-----
From: Sergey Okhapkin [mailto:sos@buggy.prospect.com.ru]
Sent: Wednesday, April 22, 1998 9:47 AM
To: 'Gnu-Win32'; 'Parker, Ron'
Cc: 'cygwin32-developers@cygnus.com'
Subject: RE: New patch
Parker, Ron wrote:
> It there a method to check for content on the queue, like
PeekMessage?
> Would it make sense to have one?
>
Do we really need it? Just call PeekMessage() :-) The main goal of my
patch
is to provide working select() on a mix of handles and message queue,
it's
really neccessary to compile GDI version of Xemacs with cygwin.
Anyway, it's easy to implement a read() call without removing a
message
from the queue, have you any suggestions how to control the behavior
of
read() call?
--
Sergey Okhapkin, http://www.lexa.ru/sos
Moscow, Russia
application/ms-tnef