Re: GNUstep Base 1.11.0

[Richard Frith-Macdonald]

On 2005-07-23 14:25:34 +0100 Benoit Astruc <benoit.astruc@free.fr> wrote:

> > What weere the other two comple errors?  Perhaps I can put fixes for them 
> > into CVS?
>
>         Bug #13893 in NSRunLoop.h
>
> You have this code :
>
> typedef enum {
> #ifdef __MINGW__
> ET_HANDLE,
> #else
> ET_RDESC, /* Watch for descriptor becoming readable. */
> ET_WDESC, /* Watch for descriptor becoming writeable. */
> #endif
> ET_RPORT, /* Watch for message arriving on port. */
> /* For binary compatibility we have an extra ifdef... */
> #ifndef __MINGW__
> ET_EDESC /* Watch for descriptor with out-of-band data. */
> #endif
> } RunLoopEventType;
>
>         which don't compil when MINGW isn't defined.

I don't understand that ... (since it certainly compiles for me) ... if 
__MINGW__ is not defined, the above reduces to -

 typedef enum {
 ET_RDESC, /* Watch for descriptor becoming readable. */
 ET_WDESC, /* Watch for descriptor becoming writeable. */
 ET_RPORT, /* Watch for message arriving on port. */
 ET_EDESC /* Watch for descriptor with out-of-band data. */
 } RunLoopEventType;

Which is fine.

If __MINGW__ *is* defined, we get code  -

 typedef enum {
 ET_HANDLE,
 ET_RPORT, /* Watch for message arriving on port. */
 } RunLoopEventType;

which has a trailing comma, which it strictly shouldn't ... though this 
seems to be acceptable to the windows/mingw compiler (the versions people 
mhave used so far anyway).

I think your suggested improvement is good ... but I don't see how you got a 
problem unless you defined __MINGW__ for some reason.