Re: PATCH: Support Windows in event-loop.c

[Christopher Faylor <me@cgf.cx>]

On Mon, Apr 25, 2005 at 09:16:12AM -0400, Daniel Jacobowitz wrote:
>On Mon, Apr 25, 2005 at 12:24:14AM -0400, Christopher Faylor wrote:
>> On Sun, Apr 24, 2005 at 04:57:36PM -0700, Mark Mitchell wrote:
>> >Daniel Jacobowitz wrote:
>> >>I guess I don't see this as a problem, while you do.  In any case,
>> >>since Chris has raised technical objections, I'm going to sit back and
>> >>see what the next revision looks like.  Hopefully it will make us both
>> >>happier.
>> >
>> >I, too, am waiting on Chris' comments re.  my justification for using
>> >WaitForMultipleObjects.  If it turns out that this is not the right
>> >primitive to use, then we'll have to revisit that side of things, but I
>> >suspect that the choice between directly modifying the file and
>> >providing a (almost-certainly incomplete) implementation of "select"
>> >will probably remain.
>> 
>> Sorry.  I didn't know you were waiting for any further feedback from me.
>> 
>> Your last message implied that the only handle which ever makes it into
>> a select call is a console handle.  Is that right?  I thought that it
>> was also used for serial I/O and GDB/MI.
>
>It is only the console handle - in the configuration that Mark's been
>testing.

Ok...  So, is it acceptable to include a console-only implementation in
event-loop.c?  I would think that it wasn't.

That seems to suggest that some kind of generic select or poll
implementation needs to be developed, probably using threads.

cgf