morning all,
Apologies in advance for an annoying and quite possible off-topic question,
but I can't seem to get my head around this bug, and I'm hoping perhaps
someone on the list can enlighten me.
in the wake of the release of PDL 2.014 and some associated buglets
manifesting in a few of my CPAN modules, I've noticed some very strange
behavior coming from cpantesters, to wit: systems reporting near-identical
configurations aren't testing identically; compare e.g.
http://www.cpantesters.org/cpan/report/c9c3665a-87b1-11e5-99df-94052cde3c15
(PASS) with
http://www.cpantesters.org/cpan/report/e77c5196-897a-11e5-b552-e159351a082c
(FAIL). Both systems report themselves as "perl v5.22.0, GNU/Linux,
3.16.0-4-amd64, x86_64-linux-thread-multi", both running PDL-2.014, yet the
one fails with a "PDL::VectorValued::Utils needs to be recompiled against
the newly installed PDL" message during testing, while the other succeeds.
I understand the why such a message can be useful; in this case, I'm not
sure that situation (re-linking and/or re-compilation required) applies.
The module in question, PDL::Ngrams
<https://metacpan.org/release/PDL-Ngrams>, depends on my CPAN module
PDL::VectorValued <https://metacpan.org/release/PDL-VectorValued> both for
runtime utilities (rlevec() & co) and compile-time macros (vv_indx_sig(),
vv_indx_typedef()). For consistency reasons, I've put the version
information for PDL::VectorValued in a specialized file
PDL/VectorValued/Version.pm
<https://metacpan.org/source/MOOCOW/PDL-VectorValued-1.0.3/VectorValued/Version.pm>
(which itself does not require PDL). Then, in PDL-Ngrams/Makefile.PL
<https://metacpan.org/source/MOOCOW/PDL-Ngrams-0.06/Makefile.PL>, I had to
specify
PREREQ_PM => { ..., 'PDL::VectorValued::Version' => ... }
because ExtUtils::MakeMaker doesn't seem to resolve it correctly if I
specify it as PREREQ_PM=> { ..., 'PDL::VectorValued' => ... } when the
actual package $VERSION declaration isn't really in PDL/VectorValued.pm (as
it isn't). On the other hand, I want PDL::VectorValued::Dev
<https://metacpan.org/pod/PDL::VectorValued::Dev> to carry the same
$VERSION as PDL::VectorValued itself, and I even need the Dev module in
order to build PDL::VectorValued::Utils
<https://metacpan.org/pod/PDL::VectorValued::Utils>, which provides the
PDL::PP guts for PDL::VectorValued.
I'm guessing that the "needs re-compilation" messages from the FAIL report
are occurring on a smoker with an older version of PDL::VectorValued
installed, which passes the PREREQ_PM test because the referenced
PDL::VectorValued::Version package isn't itself built with PDL::PP, but
then fails during 'make test' because the actual PDL::VectorValued::Utils
guts aren't current. Can anyone suggest a configuration that would allow
me to (a) specify a $VERSION for PDL::VectorValued and its sub-packages
once and only once, (b) specify PREREQ_PM for dependent modules like
PDL::Ngrams in a manner which is more palatable to CPAN smokers than the
current hacks, and (c) continue to use PDL::VectorValued::Dev when building
PDL::VectorValued::Utils itself? Or is there something else going on here?
I realize it's rather a complex situation but I'm about at my wits' end
here, and would be grateful for any suggestions.
marmosets,
Bryan
--
Bryan Jurish "There is *always* one more bug."
moocow.bov...@gmail.com -Lubarsky's Law of Cybernetic Entomology
------------------------------------------------------------------------------
Presto, an open source distributed SQL query engine for big data, initially
developed by Facebook, enables you to easily query your data on Hadoop in a
more interactive manner. Teradata is also now providing full enterprise
support for Presto. Download a free open source copy now.
http://pubads.g.doubleclick.net/gampad/clk?id=250295911&iu=/4140
_______________________________________________
pdl-devel mailing list
pdl-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/pdl-devel
