This is the mail archive of the
libc-hacker@sourceware.cygnus.com
mailing list for the glibc project.
Re: glibc-2.0.93 homebrew distribution tests
- To: drepper@cygnus.com (Ulrich Drepper), John Kennedy <jk@csuchico.edu>
- Subject: Re: glibc-2.0.93 homebrew distribution tests
- From: Franz Sirl <Franz.Sirl-kernel@lauterbach.com>
- Date: Fri, 22 May 1998 06:45:48 -0700 (PDT)
- Cc: libc-alpha@cygnus.com
- References: John Kennedy's message of "Wed, 20 May 1998 08:03:35 -0700"<199805201503.IAA18885@hircine.jk.ml.org>
At 16:05 Uhr -0000 20.05.1998, Ulrich Drepper wrote:
>> Many configure programs laying around check for HAVE_UTMP_H and
>> HAVE_UTMPX_H and end up in a strange state when you have both. So far
>> I've been able to fix all of them by ignoring HAVE_UTMPX_H if I have
>> HAVE_UTMP_H defined.
>
>utmpx.h is the standard interface. utmp.h is there only for
>compatibility reasons.
Yes, and eg. ssh-1.2.23 tries to use it, but also makes some (invalid?)
assumptions as soon as HAVE_UTMPX_H is defined:
1. it tries to include <sys/mkdev.h> for minor()
2. it expects a structure member ut_syslen (the strlen() of ut_host)
3. it expects WTMPX_FILE to be defined, but this is only defined if
_GNU_SOURCE is defined
Probably there are other packages out there which make similar assumptions
and I think a better solution than to undefine HAVE_UTMPX_H is to correct
the sources and send a diff to the package maintainer.
Franz.