bzr: warning: some compiled extensions could not be loaded;

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Barry Warsaw wrote:
> Public bug reported:
>
> On Karmic:
>
> % bzr --version
> Bazaar (bzr) 2.1.0dev
> Python interpreter: /usr/bin/python 2.6.2
> Python standard library: /usr/lib/python2.6
> Platform: Linux-2.6.31-11-generic-x86_64-with-Ubuntu-9.10-karmic
> bzrlib: /usr/lib/python2.6/dist-packages/bzrlib
> Bazaar configuration: /home/barry/.bazaar
> Bazaar log file: /home/barry/.bzr.log
>
> Copyright 2005, 2006, 2007, 2008, 2009 Canonical Ltd.
> http://bazaar-vcs.org/
>
> bzr comes with ABSOLUTELY NO WARRANTY. bzr is free software, and
> you may use, modify and redistribute it under the terms of the GNU
> General Public License version 2 or later.
>
> bzr: warning: some compiled extensions could not be loaded; see
> <https://answers.launchpad.net/bzr/+faq/703>
>

^- did you read the +faq that we asked you to read?

Basically, you didn't run 'make' in your copy of bzr, which means that
lots of operations are going to run slower than they usually do.

Though it looks like you are running from an installed copy, which
doesn't quite make sense.

Are you running a nightly? Did the extensions start failing to build there?

John
=:->

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAkrCxz4ACgkQJdeBCYSNAANr1gCfd8cMY7IqQln1g0CzTdwZ9RD2
xQYAoIUf3XyoYhlg1ppKgyFDezg8Rq6V
=CJ0J
-----END PGP SIGNATURE-----

John: It's only complaining about the *_pyx extensions, not the *_c ones. So it appears that "make" probably was run, but that Pyrex isn't installed.

Seeing as this appears to be a packaged copy of bzr, this is presumably a bug in the packaging rather than a mistake by Barry.

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Andrew Bennetts wrote:
> John: It's only complaining about the *_pyx extensions, not the *_c
> ones. So it appears that "make" probably was run, but that Pyrex isn't
> installed.
>
> Seeing as this appears to be a packaged copy of bzr, this is presumably
> a bug in the packaging rather than a mistake by Barry.
>

So in both cases unless '--allow-python-fallbacks' was set, 'make'
should fail if it cannot build the extensions (whether pyrex is
available or not).

I've certainly made use of that locally to do:

make && bzr selftest ...

So it may be that we need to look at the build chain, and figure out why
the builds are succeeding when they aren't getting the extensions built.

John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAkrDYxgACgkQJdeBCYSNAAM/zgCgh1O93MTYbX02x/lr+QYi8UCw
xJcAoIJLbGYlLRD4QatIGWnn60Os9yd8
=SVeG
-----END PGP SIGNATURE-----

Barry is using nightlies; apparently they use the PPA packaging so that is a) at risk and b) needs to be fixed.

We should have a rule that there must be 6+.so files in <pkg>/usr/shared/pyshared/bzrlib/ (or whereever the temp path is) so that any error (the bzrlib/bzrlib thing; other things) causes a build failure.

As Martin pointed out, if this is going to be Critical, then someone should be assigned to work on it, and complete it. I personally don't feel that I have the setup/knowledge to do so. I believe Robert is away sprinting so he is out. Does anyone else feel like they could step up?