[Date Prev][Date Next]
[Chronological]
[Thread]
[Top]
Re: autoconfig on Solaris (ITS#1337)
On Monday, 2001-09-17 at 12:54:03 -0700, Kurt D. Zeilenga wrote:
> At 11:53 PM 2001-09-16, lupe@lupe-christoph.de wrote:
> >On Sunday, 2001-09-16 at 20:21:56 +0000, Kurt Zeilenga wrote:
> >> Please provide the config.log fragment of the failed test
> >I couldn't find a failed test. That's the problem - it does not test if it needs -lsocket.
> In looking at your config.log, it did in fact test for -lsocket
> but failed to detect that it was in working order due to the
> inclusion of -lbind as a library.
Ah, you're right. I was distracted by "checking for main". And the
program does not use any symbol except main.
configure:4765: checking for main in -lsocket
configure:4780: gcc -o conftest -g -O2 -I/opt/local/include/bind -I/opt/local/include -L/opt/local/lib conftest.c -lsocket -ldl -lbind 1>&5
Undefined first referenced
symbol in file
yp_first /opt/local/i386/lib/libbind.a(nis_ho.o) (symbol belongs to implicit dependency /usr/lib/libnsl.so.1)
yp_next /opt/local/i386/lib/libbind.a(nis_ho.o) (symbol belongs to implicit dependency /usr/lib/libnsl.so.1)
yp_match /opt/local/i386/lib/libbind.a(nis_ho.o) (symbol belongs to implicit dependency /usr/lib/libnsl.so.1)
yp_get_default_domain /opt/local/i386/lib/libbind.a(nis.o) (symbol belongs to implicit dependency /usr/lib/libnsl.so.1)
Fails, because the missing functions are in -lnsl. Never use -lsocket
without -lnsl on Solaris. No program I know uses one without
the other.
> >Here is the part of config.log that is somehow related. You can
> >also see that even though I have -lbind (not part of Solaris),
> >it uses -lresolv, too. But I consider this my problem because
> >I created an unusual situation by installing a -lbind on
> >Solaris.
> Well, yes. If you add -lbind to LIBS you need to be sure
> you add whatever -lbind depends upon to the environment
> as well.
I've never needed to include -lsocket and -lnsl. They are normally
added automatically. Why would you want to use a socket without
a nameservice.
OK, please close the bug. But maybe leaving off other libraries for this
test would be a good idea.
Thanks,
Lupe Christoph
--
| lupe@lupe-christoph.de | http://free.prohosting.com/~lupe |
| I have challenged the entire ISO-9000 quality assurance team to a |
| Bat-Leth contest on the holodeck. They will not concern us again. |
| http://public.logica.com/~stepneys/joke/klingon.htm |