On Sat, 15 Dec 2007 21:31:25 +0300
Bulat Ziganshin <[EMAIL PROTECTED]> wrote:

> Hello Juanma,
> 
> Saturday, December 15, 2007, 4:24:43 AM, you wrote:
> 
> > Because what Yitzchak Gale proposed and I seconded does not mean that
> > getHomeDirectory will not "follow the Windows API", unless very
> > specifically requested by setting HOME.
> 
> i'm against this idea. one can setup HOME for some specific program
> and then find that all his ghc-compiled programs are changed their
> behavior. ghc don't have a goal of emulating Unix standards in windows
> environment so such behavior will look unexpectedly

I too agree, and would add that (as someone pointed out earlier) it's trivial 
to wrap the function in question.  Further, not only is it trivial but it's 
"more correct" in the sense that O/S specific behavior should be isolated 
whenever possible, and such isolation is certainly possible here.  Create a 
class that defines, but does not implement, the required methods, and create an 
instance for the O/S in use.  That's clean, simple, and is guaranteed to not 
break existing working programs.

-- 
Seth Kurtzberg <[EMAIL PROTECTED]>

> 
> -- 
> Best regards,
>  Bulat                            mailto:[EMAIL PROTECTED]
> 
> _______________________________________________
> Glasgow-haskell-users mailing list
> Glasgow-haskell-users@haskell.org
> http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
> 

_______________________________________________
Glasgow-haskell-users mailing list
Glasgow-haskell-users@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-users

Reply via email to