[rrd-users] Re: Can not parse any arguments with Perl RRDs

[Shawn Asmussen]

On 4/18/06, mugi at pinesalad.net <mugi at pinesalad.net> wrote:
>
>
> Possible to post a link to your compiled rrdtool (or at least the perl
> modules)?

 I could probably send you a compiled rrdtool if you can't get anything else
working. Let me know if you can't get it resolved. The only boxes I have the
IBM compiler on are 5.2, but if I had to guess, I'd say that the binaries
would probably still work on a 5.1 or 5.3 machine.

Some other thoughts:
>
> 1.What I'm not sure about is why the Xlc compiler works fine with no
> code modification with the getopt issue.
> Should be the same issues that exists for getopt and threading for any
> compiler?


 That one baffled me a little too, but I was just happy to finally get it
working after a lot of struggling with trying to make it work with gcc.

2. It looks like when the perl modules are built they use compiler
> options that appear to come from the perl version installed. When it
> builds with gcc you get
>
> cc_r: unrecognized option '-qmaxmem=-1'
> cc_r: unrecognized option '-qnoansialias'
> cc_r: unrecognized option '-q32'
> cc_r: unrecognized option '-qlonglong'
>
> Which are Xlc based options.
>
> Not sure where those are stored (or if they can be modified) but would
> be nice to be able to change these for gcc options.


 These options are stored somewhere in Config.pm <http://config.pm/>, or in
a file included from there. I tried using gcc options when compiling the
shared perl module, but it didn't help. I can't recall if the rrdtool build
process supports the environment variables that override the
Config.pm<http://config.pm/>values (PERLCC, PERLCCFLAGS, etc...), or
if I had to edit the Makefile by
hand. Either way though, compiling with the options from a gcc compiled
version of perl didn't change the outcome any.


Right now I'm compiling perl 5.8.8 with no multi-threading with gcc to
> see if it will work that way.
> I can use an alternate path to this version of perl when using RRDs.
>
> Horribly unclean, but I need to get this working.
>
> Any further thoughts on this?


 Don't know if that'll work for you or not, although it might. I'd be
interested in hearing, either way. Ultimately, I'd like to see the code
changed so that the perl shared module doesn't have to go through the
thread-unsafe versions of the functions. Either getopt could be written out
of the code altogether, or possibly the RRDs module could be redone so
that the portion of the argument parsing that is done with getopt happens in
pure perl, and then RRDs could use the thread-safe versions of the functions
in librrd, like rrd_create_r. At any rate, something needs to be done with
it, because a lot of people using AIX don't have the IBM compiler.

Shawn Asmussen

--
Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:rrd-users-request at list.ee.ethz.ch?subject=help
Archive     http://lists.ee.ethz.ch/rrd-users
WebAdmin    http://lists.ee.ethz.ch/lsg2.cgi