[Fwd: Re: Some info on Squid and MinGW]

-----Forwarded Message-----

> From: Robert Collins <robertc@squid-cache.org>
> To: Guido Serassio <serassio@libero.it>
> Subject: Re: Some info on Squid and MinGW
> Date: 16 Sep 2002 22:12:05 +1000
>
> On Mon, 2002-09-16 at 22:03, Guido Serassio wrote:
>
>
> > I think that separate disk, socket and pipe I/O in HEAD is the better help
> > for the Windows port, and this should be made using a modularized
> > approach.
>
> Thats exactly what I'm talking about. There are a number of steps to get
> there though. One such way that will keep HEAD working at every stage is
> to:
>
> identify what operations we do on all types of fd, socket fds, disk fds
> etc.
>
> mark the fds as being a disk/pipe/socket.
>
> Alter the fd accessing routines to call the correct operations based on
> the type of the fd. Place all those operations in a file somewhere.
>
> Override the routines that are called on win32 with the correct native
> functions.
>
> Done.
>
> Each step can be done safely and tested carefully. No #ifdefs are needed
> except for the very last step, and that can be modularised if desired.
> I'd get it working first though, and as the #ifdefs would be in one and
> only one file, I don't see that being such an issue.
>
> Rob