On Sun, Oct 21, 2007 at 04:48:23PM +0200, demerphq wrote:

> As an aside, it seems to me that both Devel::CheckLib and
> configure_requires suffer from a fatal flaw in that they do not solve
> the problem for existing modules.

Agreed.

There are actually two problems to solve:

* not sending bogus FAIL notices;
* catching real failures

an example of the first one would be sending a FAIL notice when a
library is missing.  And it is in fact easy to fix - just don't send any
failure notices at all for the 'Makefile.PL' and 'make' steps.

Unfortunately this means that an important class of error will be
completely ignored.  Number 2 above is all about catching things like
"the author assumed everyone uses gcc" or "SVs don't look like that in
5.9.5".

Devel::CheckLib etc will at least help us to get closer to these two
ideals.  I hope, however, that the CP6AN will make it obsolete by
having:

* tighter integration and better co-operation between the CP6AN, build
  tools, and testers;
* clear documentation about what diagnostics to spit out and when

-- 
David Cantrell | http://www.cantrell.org.uk/david

  On the bright side, if sendmail is tied up routing spam and pointless
  uknot posts, it's not waving its arse around saying "root me!"
      -- Peter Corlett, in uknot

Reply via email to