[Kamailio-Devel] Build failures with seas and mi_xmlrpc modules on OpenBSD 4.3

[Henning Westerholt]

On Saturday 18 October 2008, Jason Creighton wrote:
> Henning Westerholt wrote:
> > On Thursday 16 October 2008, Jason Creighton wrote:
> >> [Sorry if the devel list isn't the right place for this; Please steer me
> >> in the right direction if it is not.]
> >
> > Hi Jason,
> >
> > you could also open a bug report on our tracker for this issues, but now
> > lets discuss this on the list.
>
> For future reference, which one is the preferred method?

Hi Jason,

ok. :-) If there are compile failure for a release or a stable branch, its 
better to submit a bug report, or report this to the user lists. If the code 
in the trunk fails to compile, this is dicussed at the devel list. 

> >> But apparently sys/types.h isn't being included like they say it will, I
> >> can get it to compile if I force inclusion of <sys/types.h> (eg, with
> >> "DEFS+=-include sys/types.h" in that module's Makefule), but I have no
> >> idea what the right solution would be.
> >
> > Strange, perhaps the comment in the file is obselet. I don't think that
> > adding a '#include <sys/types.h>' to this file will break something. But
> > perhaps we can add this with an #ifdef? Do you know what type of #define
> > is usually checked against an openbsd compiler/ system?
>y
> There is an __OpenBSD__ define, eg,
>
> #ifdef __OpenBSD__
> #include <sys/types.h>
> #endif
>
> If the above style is acceptable, let me know and I can submit a patch
> to fix the affected files. (encode_allow.c is just the first to fail; I
> believe the all the encode_*.c include <netinet/in.h> without
> <sys/types.h>)

Sure, this is acceptable. You can upload your patch to our tracker as well, in 
the patches section: 
http://sourceforge.net/tracker/?group_id=139143&atid=743022

Perhaps this error also exists for other *BSDs as well, then we can revise the 
#ifdefs.

Henning