On Mon, 23 Feb 2004, Tor Arntsen wrote:

> >Tor (or anyone else with AIX), can you see if including <sys/select.h> in
> >ares.h helps fixing it?
>
> Some like this? Yes, that would fix it. I have been holding back on a patch
> for this because I was thinking of doing it the ./configure way like it's
> done in lib and src, i.e. with HAVE_SYS_SELECT_H wrappers.

I added that for the sys/select.h one and I added checks for a few other
header files as well, while hacking at it.

> On the other hand, the latest POSIX demands that one should always include
> <sys/select.h> for anything select() (thus also for fd_set), so should it
> really be necessary with wrappers for e.g. sys/types.h and sys/select.h? Or
> are there some curl-supported platforms without these files?

I'm not sure, but I prefer to use the #ifdefs just in case.

> (completely unrelated: The CVS as of 15:15 GMT today hangs in test 127, on
> AIX at least)

That sounds strange. I can't recall anything in particular I've done that
should break 127... The 127 one is very similar to a few other tests, it just
doesn't use EPRT so it sounds very odd that this particular test goes on
strike right now!

(I did break one of the base64 functions during a couple of hours this
morning, and that's why several builds reported 7 test case failures...)

-- 
    Daniel Stenberg -- http://curl.haxx.se/ -- http://daniel.haxx.se/
   [[ Do not send mails to this email address. They won't reach me. ]]