Path: utzoo!attcan!uunet!ficc!peter
From: peter@ficc.uu.net (Peter da Silva)
Newsgroups: comp.lang.c
Subject: Re: Coroutines in C
Message-ID: <5797@ficc.uu.net>
Date: 21 Aug 89 02:18:33 GMT
References: <5663@ficc.uu.net> <2151@netcom.UUCP> <5773@ficc.uu.net> <1989Aug19.230124.860@light.uucp>
Organization: Xenix Support, FICC
Lines: 39

In article <1989Aug19.230124.860@light.uucp>, bvs@light.uucp (Bakul Shah) writes:
> In article <5773@ficc.uu.net> peter@ficc.uu.net (Peter da Silva) writes:
> >                            In many C compilers I wouldn't even need
> >to write any assembly code, since setjmp/longjmp can be used to implement
> >the context switch.

> You are better off not depending on setjmp/longjmp for this because

I'm not depending on setjmp/longjmp. I'm just pointing out that, *in many
C compilers* it is possible to use setjmp/longjmp to implement the context
switch. In my original posting on this subject I went into more detail on
this subject. I am well aware of the Berkeley VAX compiler's stack unwinding
on a longjmp.

> Even in cases where you can get away with using them for coroutine
> `context switch' now, the compiler vendor may rightfully decide to
> check for 4. above in a later version.

True. The compiler vendor can also change their calling sequence and
invalidate your low level context-switch routines. A new version of
the compiler is effectively a new compiler and a new porting effort for
routines like this. Hopefully if they become standardised the compiler
vendor will include them in the new release.

> It would be nice to have a standardized (but optional) coroutine
> facility (if an implementation provides it, one can depend on a standard
> inteface and behavior) but there is not reason to mix them up with
> setjmp/longjmp.

You should pop back a few stack levels yourself :->, and have a look at my
original proposal. I do keep the two concepts seperate, but designing things
so a naive implementation of setjmp/longjmp (store all regs in jmp_buf in
setjmp, restore them and do a return in longjmp) can be used to bootstrap
the system is IMO a good idea.
-- 
Peter da Silva, *NIX support guy @ Ferranti International Controls Corporation.
Biz: peter@ficc.uu.net, +1 713 274 5180. Fun: peter@sugar.hackercorp.com. `-_-'
"Optimization is not some mystical state of grace, it is an intricate act   U
   of human labor which carries real costs and real risks." -- Tom Neff