The big problem I see is going to be with DBD::Oracle. I'm not sure how it works with non-Unix installations, but when I've had to install it, it uses the native, local copies of the Oracle client libraries. If you can't build DBD::Oracle on your development machine in such a way that the Oracle client libraries are the same on the production machine, I don't think this is going to work. The other modules don't (to my knowledge) require much in the way of extra-Perl code/libraries.
You might have the most luck using CPAN and setting yourself up with a bundle file, and just have the IS people use CPAN to install that bundle. Else, just archive up the whole Perl lib directory that you're using for this project and distribute it en masse (assuming nothing in there is proprietary or is licensed in such a way as to disallow it).
Without access to a Novell system, though, I'm wondering if DBD::Oracle might be the difficult piece to get installed right. | [reply] |
Its never recommended, but its always possible. If you've got
the same hardware and OS then its probably OK. We had HP's but
the production box did not have an HP compilier license, so
we built modules (the ones needing compiling) on the development box, then copied stuff to production.
Sure, we maybe we could have installed gcc on both boxes, but let's not
go there... | [reply] |
Is there any way for you to get remote shell access to
their server? Since you seem to have ftp access to their
production directories, they might give it to you.
Then you could work remotely.
I have to say that I doubt that you can just copy over
something like DBD::Oracle. And the CPAN module won't work
for that, too. Maybe you can follow up on a CPAN module
build attempt - patch the files in the build directory with
the right values for environment variables etc. with
a simple perl script, and rerun the CPAN module command.
Christian Lemburg
Brainbench MVP for Perl
http://www.brainbench.com
| [reply] |