[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Koha-devel] CPAN and re-inventing the wheel
From: |
Alan Millar |
Subject: |
Re: [Koha-devel] CPAN and re-inventing the wheel |
Date: |
Sat Jul 6 23:03:02 2002 |
User-agent: |
Mutt/1.2.5i |
On Sat, Jul 06, 2002 at 01:51:42PM -0700, Tonnesen Steve wrote:
> I agree that there are some things that we should be relying on modules
> for (MARC support, Z39.50), but I don't think that we should be adding a
> module dependency for the 4 lines of code required to check the validity
> of an ISBN.
I don't want to be argumentative, but those four lines of code have
been in Koha for 8 months (CVS start of marcimport.pl), and the bug
was just now caught. How many other latent bugs are throughout Koha
because we are re-inventing the wheel, that wouldn't be there by
using a pre-debugged CPAN module? It's impossible to know.
However, I expect that solving the fundamental CPAN dependency
problem for one module will solve it for all modules. If we can
properly detect and install one CPAN module to solve a dependency,
we can do that with a list of modules too.
> It would also be nice to have a short tutorial on diagnosing module
> problems (If you see something like _this_ in your apache error logs, that
> means you are missing _this_. Explain what the @INC paths are that should
> also be down in the apache error logs, and how to use CPAN to install the
> missing modules). If this could be automated, that'd be great.
Yes, that's a great idea.
> At this point in Koha's development, installation problems are losing us a
> lot of potential users, in my opinion. It's getting better, of course,
> but we still need to pay attention to it.
Rhetorical question: Which is worse, a tough installation that
provides a working system afterwards, or an easy installation that
provides a buggy system afterwards? No quick simple answer, but
you can guess which way I'm leaning.
Just my food for thought. Thanks for the dialog.
- Alan
----
Alan Millar --==> address@hidden <==--