Christian "Xtra" Schiffler wrote:
> Did anyone of you guys ever had a similar problem?
> I'm really lost in how to tackle this problem.
> System is Slackware 10.2 which I want to upgrade on that testing machine
> before upgrading the life system.

Well, no one has had any real advice yet. The only thing I can say is
I have a vague memory of someone having a line in a startup script that
began with a './' instead of just '/' when executing a command that for
some reason caused the script to just exit. I wish I could remember the
subject - I'd probably be able to google it up easy enough.

- Kurt

It had nothing to do with the script itself.
After some more testing (booting into the system via ini=/bin/bash and
executing the startup scripts by hand), I found out the old udev was
unable to detect any tty/pty on the testing machine.
This resulted in loosing all terminals (along with several other devices).
I have no clue why it was like that but after upgrading udev to the
slack12.1 contained version I was able to boot in just normal and do the
rest of the upgrade.

Very strange issue though.
- Chris

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 2008-06-04, ~kurt wrote:
> I have a vague memory of someone having a line in a startup script that
> began with a './' instead of just '/' when executing a command that for
> some reason caused the script to just exit.

Close, but no cigar.

Many of the init scripts in Slackware source other init scripts and
they do so thusly:

# Start Apache web server:
if [ -x /etc/rc.d/rc.httpd ]; then
. /etc/rc.d/rc.httpd start
fi

In this case, rc.httpd is included as part of the calling script (in
this case, rc.M). The problem occurs when sysadmins do this with their
own custom scripts that include an exit command. Example:

#rc.foo
if [ -x /usr/sbin/foo ]; then
/usr/sbin/foo
elif
exit
fi

In this case, if some one were to say, edit rc.M and source rc.foo, all
of rc.M would quite running if /usr/sbin/foo was not executable for
some reason.

The proper way when doing something along those lines, is not to source
the script but simply to call it directly.

.. /etc/rc.d/rc.foo # bad if rc.foo contains an explicit termination
/etc/rc.d/rc.foo # good if rc.foo contains an explicit termination

Savy?

- --
It is better to hear the rebuke of the wise,
Than for a man to hear the song of fools.
Ecclesiastes 7:5
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)

iEYEARECAAYFAkhGupUACgkQrZS6hX/gvjoPsQCglvSvC8e7IRi2zpvDbfzf9Smw
5nEAmgKXM0f9fyacXz/HZBn+LTXbhepX
=BDJe
-----END PGP SIGNATURE-----

On 2008-06-04, ~kurt wrote:
> Christian "Xtra" Schiffler wrote:
>> Did anyone of you guys ever had a similar problem?
>> I'm really lost in how to tackle this problem.
>> System is Slackware 10.2 which I want to upgrade on that testing machine
>> before upgrading the life system.
>
> Well, no one has had any real advice yet. The only thing I can say is
> I have a vague memory of someone having a line in a startup script that
> began with a './' instead of just '/' when executing a command that for
> some reason caused the script to just exit. I wish I could remember the
> subject - I'd probably be able to google it up easy enough.


I believe you're thinking sourcing a file from another; IOW, you
could do this:
source /some/file
or:
. /some/file

That's roughly equivalent to an include statement in C, i.e.
#include "/some/file"

For clarity's sake, the 'source' command inserts the contents of the
sourced file into the calling file at whatever location the sourcing
was done. For example:

_script.sh:_ _/some/file_ _RESULT:_
#!/bin/sh echo middle #!/bin/sh
echo begin exit 0 echo begin
.. /some/file echo middle
echo end exit 0
echo end

Note that because of the "exit 0" in the resulting commands to be
executed, the "end" will never been seen. I used the "exit 0"
example here because quite a few pure SysV init scripts have that
in them.

In case it's not obvious to a reader, the way around this problem
is to *run* or *execute* the /some/file instead of sourcing it.
You can accomplish that with either:
./some/file (if it has executable flag set)
or
sh /some/file (if it does not have executable flag set)


-RW

Robby Workman wrote:
>
> I believe you're thinking sourcing a file from another; IOW, you
> could do this:

Yep, this is what I was thinking of - I think you were the guy who had
found the problem....

- Kurt

+Alan Hicks+ wrote:
>
> The proper way when doing something along those lines, is not to source
> the script but simply to call it directly.
>
> . /etc/rc.d/rc.foo # bad if rc.foo contains an explicit termination
> /etc/rc.d/rc.foo # good if rc.foo contains an explicit termination
>
> Savy?

Yep, makes sense - thanks,

- Kurt