Requirements for Pyramid 1.0b1

36 views
Skip to first unread message

Chris McDonough

unread,
Jan 21, 2011, 4:49:00 AM1/21/11
to pylons-devel
Hi folks,

In my estimation, we have exactly three issues left before we can
release a Pyramid 1.0b1. "Beta" status means feature freeze, and
bugfixes only.

Here are those issues:

- Add MANIFEST.in to each Pyramid paster template, and verify its
interaction with setuptools version controlled resource
inclusion. See
https://github.com/Pylons/pyramid/issues#issue/95

- Explain how to use i18n localization in Mako better. Currently we
just link to
http://blog.abourget.net/2011/1/13/pyramid-and-mako:-how-to-do-i18n-the-pylons-way/,
but I've asked Alexandre if we can include his text in the
documentation.

- Explain ``{foo:<regex>}`` pattern formatting and purpose better in
within URL dispatch chapter. Currently it's just kind of mentioned
as an aside.

If you know something about any of these topics, I'd appreciate any help
with these issues you can provide. If we can get a beta out, the final
1.0 should follow quickly.

- C


Marius Gedminas

unread,
Jan 21, 2011, 7:32:39 AM1/21/11
to pylons-devel
On Fri, Jan 21, 2011 at 04:49:00AM -0500, Chris McDonough wrote:
> Here are those issues:
>
> - Add MANIFEST.in to each Pyramid paster template, and verify its
> interaction with setuptools version controlled resource
> inclusion. See
> https://github.com/Pylons/pyramid/issues#issue/95
...

> If you know something about any of these topics, I'd appreciate any help
> with these issues you can provide.

As you can probably tell from my comments on that github issue, I've
been burned a few times by setuptools not including files in packages
I've released on PyPI. My solution, which I offer to your
consideration, is to make sure every package I maintain has a makefile
rule 'distcheck' that ensures my sdists are okay.

The test is basically this:

1. $vcs export the source tree into a temporary directory #1
2. setup.py sdist and extract the .tar.gz into a temporary directory #2
3. diff the two trees (ignoring some files that sdist generates, such as
setup.cfg or $package.egg-info), abort if there are any differences

You can see an example in objgraph's Makefile, e.g.:
http://bazaar.launchpad.net/~mgedmin/objgraph/trunk/view/head:/Makefile#L60

It does a few things more: run the testsuite in temporary directory #2,
builds a second sdist from there, extracts it into a temporary directory
#3 and compares the two again. That's probably overkill and, I think, a
vestige from earlier times when I just compared two sdists, not
realizing that this way I may miss source files not mentioned in the
manifest.

Marius Gedminas
--
Writing setattr hooks properly is a black art. Writing persistent
setattr hooks is more like hearding bees blindfolded...
-- Casey Duncan

signature.asc

Rocky Burt

unread,
Jan 23, 2011, 5:45:04 PM1/23/11
to pylons...@googlegroups.com
FYI I created DistChecker for almost this exact reason...

Typically what I end up doing is going into the directory containing my egg and simply going:
  distchecker compare_sdist --with-python=/path/to/virtualenv/bin/python

- Rocky


Reply all
Reply to author
Forward
0 new messages