perl-test-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Geoffrey Young <>
Subject Re: [rfc] a function to adjust @INC for live modules
Date Mon, 19 Sep 2005 19:08:47 GMT

Stas Bekman wrote:
> When you develop several independent modules which are using each other,
> it's usually a pain to always run 'make install' in all of them to get
> the latest changes in, while working on the current module. I usually
> try to add the lib subdirs to @INC so that I can test against liv,
> un-installed dirs.

yes, I recall :)

> Things get more complicated when you have .xs
> modules, since now you have to include blib as well, but lib must come
> before blib.

what's the point of that?  if you need blib for xs modules that implies that
you're using make to update the .so files, which automatically moves changed
perl libraries to blib anyway.  to me, using lib here starts getting really
confusing, with the possibility that you would have made changes to .xs
files that would not be picked up...

but to developer his own methods :)

> Moreover you need to deal with -T, so I've come up with the
> following sub:
> sub adjust_INC {
> I was thinking to add this to Apache::TestUtil. If you like it, we
> should come up with an intuitive name, as adjust_INC is just a temp name
> I use in my project.

I'd much rather see this kind of thing in one of two places

  - a separate CPAN module
  - incorporated into standard perl tools, like MakeMaker or Test::Builder

the reason being that this looks to be really useful to people that want to
develop they way you like to, so it would be more universally useful than
being isolated in Apache-Test.

in truth, I think Apache-Test contains lots of stuff that really ought to
have been places somewhere else for the betterment of all.
t_write_perl_script() and its associated mkpath and cleanup functionality,
for example, is such a useful testing feature that it's a shame nobody else
can use it without installing Apache-Test or duplicating the code therein...


View raw message