[Nagiosplug-devel] subst.in patch

[Phil Dibowitz]

Ton Voon wrote:
> Phil,
> 
> I've been taking a look at this. I could not recreate your problem. I 
> ran a ./configure with /usr/local/nagios/libexec in existence and  then
> ran a ./configure after rm -r /usr/local/nagios and all the 
> substitutions for the perl scripts in plugins-scripts were the same. 
> There is a specific part of subst.in which says:
> 
> # used to replace "use lib utils.pm" with "use lib @libexecdir"
> #
> function led() {
>         led1 = "@libexecdir@";
> 
> So I think your diagnosis is wrong: the libexecdir is being taken  from
> the configure script and passing through to subst.in correctly.
> 
> I put a debugging line into the main awk loop here:
> 
> # Specific programs
> 
> #
> /^[^#]/ && /(\/.*)?\/(bin|sbin|lib|libexec)\// {
>         print STDERR "Got here!!!: "$0;
>         #match($0,/(\/.*)?\/(bin|sbin|lib|libexec)\/[-_a-zA-Z0-9]+/);
>         #start=RSTART+RLENGTH;
>         #c=substr($0,RSTART,RLENGTH);
>         #sub(c,which(c,path));
> }

I'm not sure what this proves - my debugging showed I always got there
as well.

> Eg, check_log.in has GREP="/bin/grep", but after running make 
> check_log, it replaces that line with GREP="/usr/bin/grep", which is 
> the location of grep on my MacOSX system.

Ah. In that case the script is still in error, but not completely in the
way I thought it was. It's checking for things that may nto be in
libexec - great for grep. But if it doesn't find it, it gives it a
relative path, and that's an *incorrect* action. In this case which()
should return whatever it was sent.

In theory, I don't think the "use" lines should be passed through the
"which" functions at all.... since "which utils.pm" or "which
/usr/lib/nagios/plugins" never makes sense anyway...

One of those two needs to be remedied.

> You mentioned something about building the plugins in a chroot 
> environment. My guess is that your chroot is missing a lot of these 
> commands, and thus the substitutions are failing.

The chroot has all the basics... it's just using mach(1) to build and
tear it down each time to ensure it's always identical.

> So I am going to reject your patch because it looks like there is 
> specific functionality the which routine is providing, even though it 
> is failing in your specialised case. However, I will try and make the 
> comments more informative.

I think the patch is over-zealous, but a fix is still needed as
described above.

> Long term, the substitutions shouldn't be done at compile time, but 
> should be done at configure time instead. However, this will be a 
> tougher job to sort out. I welcome any patches in this area.

Agreed, but in the short term, I think which() should be fixed to never
replace "/usr/lib/nagios/plugins" whith "nagios/plugins" because that's
never a good thing.

-- 
Phil Dibowitz                             phil at ipom.com
Freeware and Technical Pages              Insanity Palace of Metallica
http://www.phildev.net/                   http://www.ipom.com/

"They that can give up essential liberty to obtain a little temporary
safety deserve neither liberty nor safety."
 - Benjamin Franklin, 1759

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 256 bytes
Desc: OpenPGP digital signature
URL: <http://nagios-plugins.org/archive/devel/attachments/20050830/6d2b3412/attachment.sig>