samba2.0.7 crashes on install

[David Lee]

On Thu, 27 Apr 2000, Lanny Baron wrote:

> Hello,
> I tried to ./configure --with-utmp --with-ssl and the following happens. Any
> clue as to why?
> 
> TIA
> 
> -Lanny Baron
> 
> 
> " -DHAVE_INCLUDES_H -DPASSWD_PROGRAM="/bin/passwd"
> -DSMB_PASSWD_FILE="/usr/local/samba/private/smbpasswd"
> Using LIBS = -lssl -lcrypto -lreadline  -lcrypt
> Compiling smbd/server.c
> Compiling smbd/files.c
> Compiling smbd/chgpasswd.c
> Compiling smbd/connection.c
> smbd/connection.c: In function `utmp_fill':
> smbd/connection.c:355: structure has no member named `ut_pid'
> smbd/connection.c: In function `utmp_yield':
> smbd/connection.c:545: structure has no member named `ut_type'
> smbd/connection.c:545: `DEAD_PROCESS' undeclared (first use in this function)
> smbd/connection.c:545: (Each undeclared identifier is reported only once
> smbd/connection.c:545: for each function it appears in.)
> smbd/connection.c:546: structure has no member named `ut_exit'
> smbd/connection.c:547: structure has no member named `ut_exit'
> smbd/connection.c: In function `utmp_claim':
> smbd/connection.c:592: structure has no member named `ut_type'
> smbd/connection.c:592: `USER_PROCESS' undeclared (first use in this function)
> *** Error code 1

The utmp stuff is very new, and still experimental.  We have discovered
that almost every operating system has its own little utmp quirks.
At the "pre3" version of Samba 2.0.7, the rapidly evolving utmp code was
deliberately frozen for the purpose of overall Samba stability.

Since then, I have developed the utmp code, which I am about to retrofit
to the official 2.0.7 release, and then make available as an evolving
patch onto that release.

I hope this evolving patch is much more portable;  it addresses many such
errors, by detecting structure members and existence of system-provided
functions (at least one OS has ".h" files which are not exactly
truthful...).  And a couple of the errors above worry me.

Although the aim is to increase portability, there are probably some
systems to which it will never easily be ported, particularly those which
lack subroutine interfaces onto the utmp/wtmp files, such as the old SunOS
4.x .  So I hope that clean compilation on every OS will be achievable,
even though full functionality may not be. 

Could you tell me what your OS is (and its version).  Also email to me
your system's "utmp.h" and "utmpx.h" files (probably in a directory such
as "/usr/include").  If there are subsidiary "...utmp..." (e.g. Redhat
seems to have "bits/utmp.h"), could you send those, too, please.

I hope to be able to circulate the patch in a few days for you to try.

Thanks again.

-- 

:  David Lee                                I.T. Service          :
:  Systems Programmer                       Computer Centre       :
:                                           University of Durham  :
:  http://www.dur.ac.uk/~dcl0tdl            South Road            :
:                                           Durham                :
:  Phone: +44 191 374 2882                  U.K.                  :