Re: [PATCH] Delete default __getreent() implementation

[Corinna Vinschen <vinschen@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1646 bytes --]

On Nov  7 20:18, Sebastian Huber wrote:
> 
> 
> ----- Am 7. Nov 2017 um 17:12 schrieb Craig Howland howland@LGSInnovations.com:
> 
> > On 11/07/2017 10:03 AM, Corinna Vinschen wrote:
> >> On Nov  7 12:53, Sebastian Huber wrote:
> [...] 
> > I suggest that at a minimum the contents of getreent.c should be kept.  This
> > keeps a starting point for people who are needing to supply such a function.
> 
> The default implementation is a bad example.  It makes no sense to use __DYNAMIC_REENT__ with this implementation, e.g. you get the same thing without a function call overhead:
> 
> #if defined(__DYNAMIC_REENT__) && !defined(__SINGLE_THREAD__)
> #ifndef __getreent
>   struct _reent * _EXFUN(__getreent, (void));
> #endif
> # define _REENT (__getreent())
> #else /* __SINGLE_THREAD__ || !__DYNAMIC_REENT__ */
> # define _REENT _impure_ptr
> #endif /* __SINGLE_THREAD__ || !__DYNAMIC_REENT__ *
> 
> > (Leaving it out of the library stops the link problem which is the stated goal
> > of the patch.)  But I think it needs to go further, and this patch should not be
> > done in this form because of the cited backwards-compatibility reasons.  If
> > something were to be done it should be a configure flag to leave it out of the
> > library.
> 
> Then we are back to Joel's patch:
> 
> https://sourceware.org/ml/newlib/2017/msg01026.html
> 
> I don't know why Joel added this _dummy_getreent, but something similar is used elsewhere:

The "I don't know" is the key here.  And as long as Joel doesn't
answer this question, this is on hold.


Corinna

-- 
Corinna Vinschen
Cygwin Maintainer
Red Hat

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 819 bytes --]