Very Computer

Hi folks,

I've been struggling for the past few months with Wietse Venema's
logdaemon program.  First, I couldn't compile the version 4 code without
posix -- which wasn't available until NextStep 3.2.  So, after a while I
managed to upgrade my system to 3.2.  Then, I found out that posix is
really, seriously broken in NextStep 3.2, and no matter what Wietse and I
tried, there was no way to make it work.

Then, I thought I'd just go back to logdaemon version 3.x, which did
compile and was running just fine under 3.1.  Wrong!  All logdaemon
version 3 code now bus errors when I compile it under 3.2 -- even though
the binary which was compiled under 3.1 worked famously.

So, in the end it seems, I need to find a kind soul who is still running
NextStep 3.1, so I can build the logdaemon version 3 code and use it on
my system.

If you or someone you know can help me out, please let me know!  :)

Thanks,
David.

Dimitri
--
+------------------------------------------------------------------------+

+------------------------------------+-----------------------------------+
| Delft University of Technology     | NeGeN                             |
| Fac Applied Math & Computer Science| NEXTSTEP Gebruikers Nederland     |
| Dep of Statistics, Probabilitistics| NiNe                              |
|  and Operations Research           | NEXTSTEP In the Netherlands       |
+------------------------------------+-----------------------------------+

Sorry for the long delay...

: You probably have tried everything, but just to be sure: how exactly
: did you try to compile it? I've had problems with posix myself, but
: eventually, every problem was solvable through some flags and defines.

:)  Yes, we definitely tried everything.

To start with, we tinkered with the compiler flag -posix and the
library -lposix.

After Wietse showed me the broken instance of strftime() that you can see
by doing

        nm -op /usr/lib/libposix.a | grep strftime

I tried using a third-party strftime.  Which seemed to take care of those
problems.  Add in Wietse's environ.c to take care of the missing putenv()
and setenv(), and finally it compiled.

Then the error on execution was --

        Warning: no access to tty; thus no job control in this shell

We tried some open() calls, as well as setsgid and setpgrp(0,0), all to
no avail.  So it compiles, but it is broken in a severe way.

Attempting to compile the older code (version 3.9 or earlier) results in
bus errors.

Do you have, or know of, a version 3.1 system which could help me out?

Thanks,
David.

I have found someone who got this to work for me.  Many thanks to Jack

The code will be publically available via anonymous FTP on
vector.casti.com in /ftp/pub/next soon.

David.