Re: darwin, gcc, and static inline

[Max Horn]

On 12.12.2012, at 20:49, Paul Eggert wrote:

> On 12/12/12 11:21, Jeremy Huddleston Sequoia wrote:
>> Can you please instead just '#define _DONT_USE_CTYPE_INLINE_ 1'
> 
> I had considered that, but unfortunately as I understand it
> we'd still have problems when compiling C code
> with GCC in the now-typical case where _FORTIFY_SOURCE > 0,
> because in that case <secure/_string.h> misuses static inline
> in the same way when implementing memcpy etc.
> 
> I suppose one option would be to put this into config.h:
> 
>  #ifdef __APPLE__
>  # define _DONT_USE_CTYPE_INLINE_ 1
>  # define _FORTIFY_SOURCE 0
>  #endif
> 
> but it's pretty drastic to disable Fortify, and I'd rather
> not do that.
> 
> I'll CC: this to bug-gnulib to give Gnulibers a heads-up
> on these suggestions -- maybe someone else who's
> an OS X expert can think of a better workaround.

As a long time OS X developer (been working on porting unix stuff to OS X for 
over a decade), I'd love to chime in -- but from the above, I cannot determine 
what the issue at hand actually is. Not even where this discussion was taking 
place before. Care to explain and/or point to an archive with the preceding 
discussion?

Thanks a lot,
Max

> 
> In the long run, I hope Apple fixes the bugs (as they're clearly
> violations of the C99 standard) and I hope that there
> will be a way for gnulib to detect that the bugs have been fixed
> so that it can stop using the current workaround.