Mailman 3 September 2008 - Distutils-SIG

[issue45] distutils.log not imported in sdist.py
by johannes raggam 14 Sep '08

14 Sep '08

New submission from johannes raggam <raggam-nl(a)adm.at>: please add from distutils import log in setuptools.command.sdist.py setuptools version: setuptools-0.6c8-py2.4.egg i got an error with installing a python module which i checked out from svn. from traceback: """log.warn("unrecognized .svn/entries format in %s", dirname) NameError: global name 'log' is not defined """ ---------- messages: 170 nosy: thet priority: bug status: unread title: distutils.log not imported in sdist.py _______________________________________________ Setuptools tracker <setuptools(a)bugs.python.org> <http://bugs.python.org/setuptools/issue45> _______________________________________________

1 0

[issue44] distutils.log not imported in sdist.py
by johannes raggam 14 Sep '08

14 Sep '08

New submission from johannes raggam <raggam-nl(a)adm.at>: please add from distutils import log in setuptools.command.sdist.py setuptools version: setuptools-0.6c8-py2.4.egg i got an error with installing a python module which i checked out from svn. from traceback: """log.warn("unrecognized .svn/entries format in %s", dirname) NameError: global name 'log' is not defined """ ---------- messages: 169 nosy: thet priority: bug status: unread title: distutils.log not imported in sdist.py _______________________________________________ Setuptools tracker <setuptools(a)bugs.python.org> <http://bugs.python.org/setuptools/issue44> _______________________________________________

1 0

Re: [Distutils] compiling an extension on solaris: genshi, mercurial, ... python-dev installed?
by Phillip J. Eby 13 Sep '08

13 Sep '08

At 10:39 AM 9/13/2008 -0700, rupert.thurner wrote: >phillip eby suggested that "python-dev" is not installed and this >should be the cause. blastwave does not really offer such a thing, so >i tried to check the files, and took debian solaris as example: >http://packages.debian.org/sid/sparc/python2.5-dev/filelist > >the header files are there, as well as the .so file /opt/csw/lib/ >libpython2.5.so. but > >/opt/csw/lib/python2.5/config/Setup >/opt/csw/lib/python2.5/config/Setup.config >/opt/csw/lib/python2.5/config/Setup.local >/opt/csw/lib/python2.5/config/config.c >/opt/csw/lib/python2.5/config/config.c.in >/opt/csw/lib/python2.5/config/install-sh >/opt/csw/lib/python2.5/config/libpython2.5-pic.a >/opt/csw/lib/python2.5/config/libpython2.5.a >/opt/csw/lib/python2.5/config/libpython2.5.so >/opt/csw/lib/python2.5/config/makesetup >/opt/csw/lib/python2.5/config/python.o > >is missing. could it be that setuptools would miss this? It's not setuptools that cares; it's the distutils. Or more likely, the linker. IIUC, these files are not optional. This appears to suggest that there's a bug in the Python configuration on the platform: http://blogs.sun.com/dp/entry/a_screencasting_toolchain You might want to follow up with the person who reported(?) the bug back to Sun or whoever it is. This is not a setuptools problem nor a distutils problem: the missing files are clearly symptomatic of breakage in your platform's Python installation. You might want to consider building your own Python from source, as often these kinds of breakage are caused by platform distributors "cleaning things up" that shouldn't be moved or deleted. (For example, separating the headers and libraries into a separate "lib", "dev", or "devel" package.) If that's what happened here, then building a fresh Python from source (without the platform vendor's "improvements") would fix it.

1 0

compiling an extension on solaris: genshi, mercurial, ...
by rupert.thurner 13 Sep '08

13 Sep '08

hi, we run http://www.blastwave.org and try to install http://genshi.edgewall.org which bails out with (btw, the same happens with mercurial version control): # ls -l /opt/csw/lib/libpython2.5.so lrwxrwxrwx 1 root root 19 Jul 18 23:14 /opt/csw/lib/ libpython2.5.so -> libpython2.5.so.1.0 # easy_install --prefix /opt/csw/testing genshi Searching for genshi Best match: Genshi 0.6dev-r914 Processing Genshi-0.6dev_r914-py2.5-solaris-2.10-sun4v.egg Genshi 0.6dev-r914 is already the active version in easy-install.pth Using /opt/csw/testing/lib/python2.5/site-packages/Genshi-0.6dev_r914- py2.5-solaris-2.10-sun4v.egg Processing dependencies for genshi Finished processing dependencies for genshi root@chvp011vs035 / # rm /opt/csw/testing/lib/python2.5/site-packages/Genshi-0.6dev_r914- py2.5-solaris-2.10-sun4v.egg root@chvp011vs035 / # easy_install --prefix /opt/csw/testing genshi Searching for genshi Reading http://pypi.python.org/simple/genshi/ Couldn't find index page for 'genshi' (maybe misspelled?) Scanning index of all packages (this may take a while) Reading http://pypi.python.org/simple/ Reading http://pypi.python.org/simple/Genshi/ Reading http://genshi.edgewall.org/ Reading http://genshi.edgewall.org/wiki/Download Best match: Genshi 0.5.1 Downloading http://ftp.edgewall.com/pub/genshi/Genshi-0.5.1.zip Processing Genshi-0.5.1.zip Running Genshi-0.5.1/setup.py -q bdist_egg --dist-dir /tmp/ easy_install-kz7muh/Genshi-0.5.1/egg-dist-tmp--y0Vju warning: no previously-included files found matching 'doc/ 2000ft.graffle' warning: no previously-included files matching '*' found under directory 'doc/logo.lineform' ld: fatal: library -lpython2.5: not found ld: fatal: File processing errors. No output written to build/ lib.solaris-2.10-sun4v-2.5/genshi/_speedups.so ********************************************************************** WARNING: An optional C extension could not be compiled, speedups will not be available. ********************************************************************** error: Setup script exited with error: can't copy 'Genshi.egg-info/ native_libs.txt': doesn't exist or not a regular file anybody has an idea how to convince setuptools to use the python library? python works and it should be in the library path though.

1 1

Re: [Distutils] bdist_egg doesn't call build?
by Phillip J. Eby 12 Sep '08

12 Sep '08

At 12:27 PM 9/12/2008 -0500, chris wrote: >I created a custom build command, and was hoping it would be called >when I ran bdist_egg. However, what appears to happen is that >bdist_egg runs all of build's subcommands directly, and my custom >build command never runs. Is this what's supposed to happen? What happens is bdist_egg runs install_lib, install_scripts, and install_data. It's these commands that run any build subcommands that apply. bdist_egg doesn't actually invoke *any* build commands directly. > If so, >how do I make it call my custom build command short of subclassing >bdist_egg and using self.run_command('build')? Short of that? You can't.

1 0

bdist_egg doesn't call build?
by chris 12 Sep '08

12 Sep '08

I created a custom build command, and was hoping it would be called when I ran bdist_egg. However, what appears to happen is that bdist_egg runs all of build's subcommands directly, and my custom build command never runs. Is this what's supposed to happen? If so, how do I make it call my custom build command short of subclassing bdist_egg and using self.run_command('build')? On a side note, would it be better for bdist_egg (and other commands that do builds) to call build directly, or are there other considerations? It seems like it would make extending easier... Thanks, -Chris

1 0

Buildout to run setup.py
by Yves Moisan 12 Sep '08

12 Sep '08

Hi All, I'm new at buildout and I was wondering if it was possible to emulate a system command like "python setup.py install" in buildout. Use Case : [buildout] parts = MapFishTrunk [MapFishTrunk] recipe = infrae.subversion urls = http://www.mapfish.org/svn/mapfish/trunk/MapFish MapFish interpreter = Scripts\python There is a setup.py file a couple of levels down the directory that I would like to execute to build the eggs (and dependencies) and to store the resulting eggs in a location accessible to Scripts\python (which was generated using virtualenv). I can run Scripts/python setup.py install but I wonder if that can be done within buildout. TIA, Yves Moisan

3 3

Getting source from __file__ in egg?
by skip＠pobox.com 11 Sep '08

11 Sep '08

If I have a package installed as a zip file egg and ask for its __file__ I might get a reference which looks like a normal file system path but isn't: % python Python 2.4.5 (#4, Apr 12 2008, 09:09:16) [GCC 3.4.1] on sunos5 Type "help", "copyright", "credits" or "license" for more information. >>> import magnitude >>> magnitude.__file__ '/opt/app/g++lib6/python-2.4/lib/python2.4/site-packages/magnitude-0.9.3-py2.4.egg/magnitude.pyc' >>> magnitude.__file__[:-1] '/opt/app/g++lib6/python-2.4/lib/python2.4/site-packages/magnitude-0.9.3-py2.4.egg/magnitude.py' >>> import os >>> os.path.exists(magnitude.__file__[:-1]) False How do I actually lay my hands on the .py file source for the magnitude module in this case? Do I have to recognize somehow that .../magnitude-0.9.3-py2.4.egg is a zip archive and resort to extracting the Python source with the zipfile module? Thx, Skip Montanaro

5 5

[issue43] add force_shared Library option to create shared lib even with use_stubs=False
by vajda 11 Sep '08

11 Sep '08

New submission from vajda <vajda(a)osafoundation.org>: setuptools is growing the capability to build regular shared libraries (as opposed to python extensions). JCC (http://svn.osafoundation.org/pylucene/trunk/jcc/jcc) uses this capability to build the JCC runtime into a regular shared library shared by all python extensions it builds and by programs embedding python (such as a Java VM when running JCC-built eggs from Apache Tomcat). This bug is about adding another option to the setuptools Library class called force_shared which forces setuptools to create a shared library from a Library instance even though the dl module may not be present to generate stubs. This is important on Linux. Note that using this flag then implies that the library itself is responsible for calling dlopen(buf, RTLD_NOW | RTLD_GLOBAL) on the relevant libpython.so before initializing the python runtime is initialized. A patch against the setuptools 0.6 branch svn is attached. The idea for this patch came from a conversation on IRC: http://chandlerproject.org/script/getIrcTranscript.cgi?channel=chandler&dat… ---------- assignedto: pje files: patch.st messages: 162 nosy: pje, vajda priority: feature status: unread title: add force_shared Library option to create shared lib even with use_stubs=False Added file: http://bugs.python.org/setuptools/file22/patch.st _______________________________________________ Setuptools tracker <setuptools(a)bugs.python.org> <http://bugs.python.org/setuptools/issue43> _______________________________________________

1 0

passing arguments to self.run_command
by chris 10 Sep '08

10 Sep '08

Is there a way to pass arguments to a command called by self.run_command? Anytime I try to add arguments by putting them inside the quotes I get an invalid command error, and self.run_command itself doesn't take any other arguments. Am I missing something? -Chris

1 0