Packaging is generally assigned to the programmers that are too incompetent to work on the main application or the programmers that have severely pissed off management :-).
For that tarball, it works for everything except for the csv file and the requirements.txt file. It'll break the moment you add something new to mercurial with a new extension. I know your build/develop/deploy style is to abuse symlinks. But I come from the viewpoint that symlinks make for a nightmare of repeatability and maintainability. I should be able to setup and activate a virtualenv without site packages, and then use pip -r requirements.txt -f http://whever/downloads/v0.9 and I should be able to get the same install every time. I have no expectation of that happening as 4 of the requirements are pulled from mercurial tip and the rest have no version information. pip freeze is showing some promise on generating a stable requirements file, but presently it's spitting out garbage on the two '-e http://' items. This is probably working as designed since '-f http://someurl' is the preferred way for pip to find tarballs at odd places. Thanks, Chris Dukes On Oct 26, 9:00 pm, Chris Moffitt <[email protected]> wrote: > Chris, > > Despite my many good qualities, apparently I suck at packaging ;) > > Anyway, I have just checked in a bunch of changes to the MANIFEST and > setup.py that should resolve these issues. I tested in my virtual > environment and everything seems to come through fine. > > When you get a second, would someone mind doing a sanity check of > downloading this -http://bitbucket.org/chris1610/satchmo/get/a0e07cff2526.gz > and doing an install. > > As far as I can tell it fixes all these issues. Assuming I can get some > positive confirmation, I will re-tag 0.9 so that people won't have to deal > with these issues going forward! > > Thanks, > Chris > > On Mon, Oct 26, 2009 at 5:36 PM, C <[email protected]> wrote: > > > Chris, > > > I don't know what you are doing to manage the source distributions, > > but it appears to be the hard way. > > > I am not suggesting to modify setup.py again. The more folks poke at > > it, the more it gets screwed up. > > > I am suggesting that you make use of easy_setup's ability to manage > > source distributions combined with the setup_requires of setuptools_hg > > to handle management of the SOURCES.txt based on mercurial metadata. > > It had to be done anyways for 'setup.py install' to continue to work > > after the move to mercurial. > > > I am generating tarballs for my own use with > > rm -rf satchmo > > hg clonehttp://bitbucket.org/chris1610/satchmo > > cd satchmo > > hg co v0.9 > > python setup.py sdist --formats=bztar,zip > > > And I get a Satchmo-0.9-0.tar.bz2 that I can feed to easy_setup and it > > successfully installs with all templates. > > Or I can untar it, run python setup.py install, and it installs with > > all templates. > > > Here is a diff of the contents of the easy_setup sdist tarball vs your > > tarball > > > --- 0 2009-10-26 18:09:07.000000000 -0400 > > +++ 1 2009-10-26 18:09:11.000000000 -0400 > > @@ -1,4 +1,3 @@ > > - > > AUTHORS > > CHANGELOG > > CONTRIBUTORS > > @@ -71,12 +70,12 @@ > > docs/upsell.txt > > ez_setup.py > > .gitignore > > +.hg_archival.txt > > .hgignore > > .hgtags > > __init__.py > > LICENSE > > MANIFEST.in > > -PKG-INFO > > satchmo/apps/keyedcache/__init__.py > > satchmo/apps/keyedcache/locale/de/LC_MESSAGES/django.mo > > satchmo/apps/keyedcache/locale/de/LC_MESSAGES/django.po > > @@ -453,11 +452,6 @@ > > satchmo/apps/product/views/adminviews.py > > satchmo/apps/product/views/filters.py > > satchmo/apps/product/views/__init__.py > > -satchmo/apps/Satchmo.egg-info/dependency_links.txt > > -satchmo/apps/Satchmo.egg-info/not-zip-safe > > -satchmo/apps/Satchmo.egg-info/PKG-INFO > > -satchmo/apps/Satchmo.egg-info/SOURCES.txt > > ---------------------------------------------------------------------- > > -satchmo/apps/Satchmo.egg-info/top_level.txt > > satchmo/apps/satchmo_ext/brand/admin.py > > satchmo/apps/satchmo_ext/brand/config.py > > satchmo/apps/satchmo_ext/brand/__init__.py > > > satchmo/apps/Satchmo.egg-info/SOURCES.txt is the magic > > > If I do a python setup.py sdist > > in an untarred v0.9.tar.bz2 I have a SOURCES.txt that is missing > > All of the images from docs > > All templates HTML > > satchmo/apps/l10n/bin > > All fixtures and tax tables > > All of satchmo/static > > > More importantly, templates and static are properly installed on the > > system, or within the virutalenv. > > > Thanks, > > Chris Dukes > > On Oct 26, 4:47 pm, Chris Moffitt <[email protected]> wrote: > > > So, are you saying that if I add the following line to Manifest.in > > > recursive-include satchmo/apps *.html *.rml > > > > It will all get included? > > > > Can someone test and confirm that fixes the issue? If it does, I can > > retag. > > > > -Chris > > > > On Mon, Oct 26, 2009 at 3:04 PM, C <[email protected]> wrote: > > > > > I have no idea what a mercurial tarball is. > > > > If you were going from a mercurial clone, you would have the mercurial > > > > metadata required for setuptools_hg to generate a SOURCES.txt > > > > containing the templates. > > > > The v0.9.{tar.bz,tar.gz,zip} files lack the SOURCES.txt file and the > > > > mercurial metadata so setuptools blindly assumes only python files and > > > > things explicitly mentioned in setup.py should be installed. > > > > This is broken as designed. > > > > > Until Mr. Moffitt does a python setup.py sdist --formats=... and makes > > > > the resulting tar/zip balls available, just grab v0.9 from mercurial. > > > > See my earlier posting for instructions on getting the working copy to > > > > the v0.9 tag. > > > > > Thanks, > > > > Chris Dukes > > > > > On Oct 26, 3:48 pm, Gloria <[email protected]> wrote: > > > > > So, from where do I download these missing files in the interim, for > > a > > > > > manual install, while I await this fix? > > > > > > This worked before from a Mercurial tar ball, so I am confused as to > > > > > why it would fail now. > > > > > > Gloria --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Satchmo users" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/satchmo-users?hl=en -~----------~----~----~----~------~----~------~--~---
