Andreas Koenig:
Thanks for your bug report of last year. I just a moment ago
became aware of it. I can only imagine how I might have
missed that when it first came through. But it certainly
explains all those test failures I see from the smoke testers
on this module.
My Makefile.PL at:
http://cpansearch.perl.org/src/HESCO/LedgerSMB-API-
0.04a/Makefile.PL
spends a lot of energy asking a human operator to manually
configure their environment for the tests.
This thing does depend on an application written in perl, but
which is not in CPAN. Any advice on how to test for it and
bomb out or configure to it, would be appreciated.
Looking back over that dialogue, I can see that with
appropriate care and sanity checks I could likely automate
that entire process. Will need to research with the
LedgerSMB team how to fire off that database population code
without a browser, or perhaps write some mechanize code to
handle that, though that would not be firewall proof, coming
from localhost and not the public interface, it might be
blocked.
I have a Thursday launch I'm now preparing for. But your bug
report is now on my radar, and has just been added to the
queue I actually work from on a daily basis.
Thanks for the feedback.
-- Hugh Esco
