Well, unfortunately you lost me at Windows XP. I have no business need for Windows XP at work and I'm too cheap to have it at home.
Here are the three things I think might be happening. 1) python is not finding the mercurial modules and windows equivalent of system(3) is not finding 'hg' 2) mercurial modules are missing, hg is found, the output is parsed incorrectly by setuptools_hg 3) mercurial modules are present and are spewing garbage. I can give you a test for presence of mercurial modules python from mercurial import hg, ui, cmdutil If that works, then it's a bug in the mercurial modules for python, or setuptools_hg. Chat with whoever is responsible for your mercurial for windows and the setuptools_hg maintainer. If it doesn't work... well, have fun reading setuptools_hg.py to figure out if it's not running 'hg' . If it doesn't bomb out, try changing the setup_requires in setup.py from "setuptools_hg" to "hg-setuptools". It's an older plugin to extract package data from mercurial, but it might work better on Windows. I would suggest as a kludge around to do the hg clone on a linux, *BSD, or OSX, and do a "./setup.py sdist --formats=zip" and copy of the dist/*.zip to your Windows box. At that point setup.py should work as it should have a complete SOURCES.txt file. On Aug 12, 10:47 pm, Pranav <[email protected]> wrote: > I appreciate your enthusiasm to get to the bottom of this. > > OS - Win XP SP3 > version of mercurial - using TortoiseHG (v 0.8.1) with mercurial 1.3.1 > python setuptools has been installed > other python packages (before starting satchmo install) - mysql, > python imaging, S3 (storage backend), django-thumbs. > > Let me explain how i found out the problem (that fixtures and other > folders werent getting copied over) -- because I think that can help > other folks who are installing satchmo as well. > > I was running the satchmo_load_l10n command - the command used to > complete and return successfully to the command prompt (no ouput > printed though). At that point, i wasnt sure if the task / command did > what it is supposed to do or not. Giving it the benefit of doubt, i > went ahead and ran the satchmo_load_store command. And it failed due > to the Country Does not exist error. > Tried searching through the forums -- found a copule of similar issues > but didnt seem to relate to why i was getting the error. > > Digged into the satchmo_load_l10n command and found that it just loads > the fixture with the xml file l10n_data.xml -- so went to the site- > packages and tried to locate this file and try and load this fixture > myself. Thats when i found that this folder / file didnt exist under > the site-packages. > > So then deleted the satchmo egg from the site-packages directory -- > tried running the setup.py again. same problem. > > The reason why i listed the above -- (1) to help anyone else who might > run into the same (2) probably you could improve the install ation doc > and add stuff related to what the user should expect when the > satchmo_load_l10n is run -- data in the db they could check for -- > ditto for other util commands that you provide. > > let me know if that makes sense and if you need any additional doc > from me. > > -p > > On Aug 12, 9:44 pm, C <[email protected]> wrote: > > > Thank you, I now know that the problem is happening before the > > setuptools that ez_setup.py downloads installs the egg it creates. > > I would like to have enough information that I have a chance to > > recreate the problem. > > > 1) What OS and distribution > > 2) What version of mercurial is installed? > > 3) Is python-setuptools installed? > > 4) What other python packages are installed? > > > While I can understand why you went ahead with the kludge around to > > "Just make it work," I'd rather find out if it's a bug in setuptools, > > a bug in setuptools-hg, or something completely unexpected. Plus I'm > > somewhat obsessed because to get these kinds on inconsistencies 6 > > years ago I had to use software billed as Enterprise :-). > > > Unfortunately, the only breakage I can reproduce every time right now > > is if I grab the last version on subversion with subversion 1.6, and > > run 'setup.py install' > > > On Aug 12, 8:04 pm, Pranav <[email protected]> wrote: > > > > thanks for the fast response. > > > > I dont have an older install - my first time installing satchmo. > > > > So the first theory you provided doesnt apply. > > > > the 2nd theory, I went into the clone created and unzipped the dist/ > > > *.egg file -- there was a tax directory in there -- but only > > > subfolders of tax were modules and templatetags - no data directory > > > was present. > > > > thanks, > > > pranav > > > > On Aug 12, 7:49 pm, C <[email protected]> wrote: > > > > > Pranav, > > > > If you'd looked through you'd know that it has been discussed in some > > > > detail. > > > > And yes, I reproduced something similar a week or so ago when I had > > > > garbage left from an earlier install of Satchmo 0.9pre scattered all > > > > over site-packages. > > > > > Due to the switch from subversion to mercurial there seems to be some > > > > confusion between the subversion trunk (Old, deprecated, and probably > > > > won't install data files correctly if retrieved with subversion 1.6.x) > > > > and the mercurial tip. Between your choice or words and the > > > > documentation's choice of words, I have no idea which you actually > > > > did. I also have no idea if you had an earlier version installed > > > > before. I'm guessing that you do. > > > > > In theory setup.py should have installed to site-packages/ > > > > Satchmo-0.9_pre-py2.?.egg/ > > > > Please check this theory. > > > > If you have directories that belong in site-packages/Satchmo-0.9...egg > > > > that are in site-packages from an earlier installation of Satchmo... > > > > they seem to be searched first and lack the files you need. > > > > In theory running 'unzip -v dist/*.egg | grep tax/data' in that > > > > directory you just got by cloning tip from mercurial should result in > > > > something like > > > > 424 Defl:N 205 52% 2009-08-12 19:15 dfcaef56 tax/data/tax- > > > > table.csv > > > > Please check this theory. > > > > In theory mercurial managed data files should be present in the egg > > > > file. > > > > > If either theory I requested that you check is confirmed... You have > > > > cruft from an earilier install. Remove it. > > > > If either theory is false, especially the unzip one, I'd like to know > > > > a bit more about your environment. > > > > > On Aug 12, 6:56 pm, Pranav <[email protected]> wrote: > > > > > > I was following the install readme at this url: > > > > > >http://www.satchmoproject.com/docs/svn/new_installation.html > > > > > > after cloning the latest satchmo release, i ran the python setup.py > > > > > install. > > > > > > Then, when i go into the python site-packages folder, i find that some > > > > > of the directories dont get copied into the site-packages satchmo > > > > > directory. > > > > > > For instance, the latest satchmo release has the following folders in > > > > > the apps/tax directory: data. locale, modles, templatetags. however, > > > > > the site-packages only has the tax/modules and tax/templatetags > > > > > folders. > > > > > > fixtures folders for the apps dont get copied at all to the site- > > > > > packages directory. > > > > > > I'm on windows, python 2.6, django 1.1 and the latest trunk release of > > > > > satchmo. > > > > > > i'm guessing instead of running the load_l10n i will just run / load > > > > > the fixture data using django-admin.py ? > > > > > > but just wanted to run this behavior by others in case you're run into > > > > > it. > > > > > > -pranav --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---
