Re: Stack unwinding for C code

[Dirk Herrmann]

Marius Vollmer wrote:

> Dirk Herrmann <address@hidden> writes:
>  
>
> > I wouldn't prefer any of the two solutions, but I am currently not
> > sure what you actually suggest - especially since in the example
> > given below you don't pass any argument to scm_begin_frame.
> >    
>
> The first variant (with scm_prevent_rewind) would be more elegant from
> an implementational point of view.  The latter (with
> SCM_F_REWINDABLE_FRAME) leads to a more desirable default behavior.  I
> think people should explicitely allow rewinding when they have unwind
> handlers.
Hmm, probably I am missing something, but wouldn't it just be possible 
to make the non-rewindable frame the default, and offer scm_allow_rewind 
as a function then?

> > It is a nice coincidence that 'free' matches the void (*func) (void
> > *) signature, especially since free will probably be one of the most
> > frequently used functions with scm_on_unwind. fclose, however, does
> > not match and is another candidate that may be commonly
> > used. Unfortunately it wouldn't be standard conforming to just cast
> > fclose to match the signature.
> >    
>
> Is that a theoretical problem or do indeed platforms exist where you
> can't cast fclose to (void (*)(void *))?
>
> If it is only theoretical, I'm inclined not to worry about it...
>  
Huh, such casts would need to be added explicitly, both in guile and in 
user code. I wouldn't recommend users to use such a programming style. 
What we do within guile is a different thing, but even here I am much in 
favor of having our code as standard conforming as possible - or at 
least encapsulate places where it isn't. Maybe we should first wait what 
kinds of cleanup functions we detect to be useful within guile and then 
decide about this issue.

Best regards
Dirk