Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 7 Sep 2012 03:00:08 GMT
From:      "John W. O'Brien" <john@saltant.com>
To:        gnome@FreeBSD.org
Subject:   Re: ports/171381: [patch] finance/gnucash fails to generate and install .pyc/.pyo files
Message-ID:  <201209070300.q87308xn021325@freefall.freebsd.org>

next in thread | raw e-mail | index | archive | help
The following reply was made to PR ports/171381; it has been noted by GNATS.

From: "John W. O'Brien" <john@saltant.com>
To: bug-followup@FreeBSD.org
Cc:  
Subject: Re: ports/171381: [patch] finance/gnucash fails to generate and install
 .pyc/.pyo files
Date: Thu, 06 Sep 2012 22:59:28 -0400

 -----BEGIN PGP SIGNED MESSAGE-----
 Hash: SHA1
 
 OK, so I did some more sleuthing, and while I am still nowhere near
 being the autotools ninja I think I would need to be to propose a
 proper fix, I will share my findings in the hope that such a ninja is
 listening.
 
 My first stop of note was the automake manual where I learned that the
 py-compile script, which is actually part of AM, is applied to
 properly-identified Python files.
 
 https://www.gnu.org/software/automake/manual/automake.html#Python
 
 So how are those files marked? See lines 5-10:
 
 http://svn.gnucash.org/trac/browser/gnucash/tags/2.4.11/src/optional/python-bindings/Makefile.am
 
 That doesn't look like what the AM manual says. How did it get to be
 that way? Here's the offending changeset in trunk...
 
 http://svn.gnucash.org/trac/changeset/21905
 
 ... and the backport to the 2.4 series.
 
 http://svn.gnucash.org/trac/changeset/22020
 
 Following the breadcrumbs back to the source, we find the report of a
 vague, unconfirmed problem with the patch that breaks the AM use of
 py-compile.
 
 https://bugzilla.gnome.org/show_bug.cgi?id=668196
 
 The upshot is that any upstream-facing way to resolve this needs to
 accommodate this upstream bug. However, "accommodate" may mean "show
 why installing in different places is correct and not the cause of any
 issues on those platforms," but that sounds excessively difficult.
 
 In the mean time, I would offer an alternative patch, which merely
 performs a local revert of r22020, but I can't figure out how to tell
 ports that I want to regenerate the Makefile.in from the reverted
 Makefile.am. I've tried several basic combinations of the
 USE_AUTOTOOLS variable to no avail. If somebody gives me a gently push
 in the right direction, I'll gladly prepare that new patch.
 
 Regards,
 John
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v1.4.11 (GNU/Linux)
 Comment: Using GnuPG with Mozilla - http://www.enigmail.net/
 
 iQEcBAEBAgAGBQJQSWMNAAoJEEdKvTwaez9wnGQH/2+LAPtnAYcSKBQlcM2BYRPu
 IC2D2pboL9rFA7QF5z3vEhijohnVSgpgHZPEw8pHX8uDcbKQd3HmPk1T1Onk8OCz
 uqUqQtGpv5NH6Ch8hOafR2BZXDc+zkBonAR/Xq5K8yuYfZvcIwE+azPhEiKyZJV4
 eJCrUYmztW2A4BVbuxBKpMY+hDCK6OBkunKu9p+jTv5ptn6UDW8Y5T5MhED4Ig6i
 0B4HBZKz7wMomyl1Xi6gXnLGnKImPdaLKWXeB6qSi8M2XYa6o1iGvppqx4388JRS
 n0H4uFWh9oBhxhLRnAHYvSPB3oAFYNKBis+PlVX6144HCwpZxAQgwhJOcc7t08Y=
 =YYU4
 -----END PGP SIGNATURE-----



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201209070300.q87308xn021325>