[beports] Re: curl.diff

  • From: Andreas Färber <andreas.faerber@xxxxxx>
  • To: beports@xxxxxxxxxxxxx
  • Date: Sat, 24 May 2008 09:02:46 +0200


Am 24.05.2008 um 08:53 schrieb Schrijvers Luc:

On za, 2008-05-24 at 07:22 +0000, Grzegorz Dąbrowski wrote:
On Fri, 23 May 2008 15:41:21 -0700
"scott mc" <scottmc2@xxxxxxxxx> wrote:

Andreas,
On the Curl.diff you have,

#ifdef __BEOS__
#ifdef __HAIKU__
/* Haiku has FD_SET defined in sys/socket.h and no socket.h */
#include <sys/socket.h>
#else
/* BeOS has FD_SET defined in socket.h */
#include <socket.h>
#endif
#endif

Shouldn't this be rewritten like this, for the day when __BEOS__ is
removed from Haiku?  Also what should happen for Bone/Zeta on this?


As a porter I am already shivering when you say drop __BEOS__, BeOS and
ZETA can be dead OS's but they are still around, Haiku, being
selfhosting only for a short while (wich we can only applaud) is not
directly fit for every day use. BeOS has been dead for 8 yrs now, but
still people are using it.
So for all times sake don't send any patches upstream that brake things
in BeOS/ZETA, instead add fixes related to Haiku without breaking down
(if possible!) other bulds.

Seems you're misunderstand something here: We're not talking about removing __BEOS__ from external software such as cURL. It's about removing __BEOS__ from Haiku's GCC, so that on Haiku we would no longer walk into #ifdef __BEOS__ sections. Doing so would have no effect on BeOS/Zeta.

Andreas
--
BePorts homepage - http://tools.assembla.com/BePorts
List archives: //www.freelists.org/archives/beports
Administrative contact: brecht@xxxxxxxxxxx

Other related posts: