[AstroPy] ANN: vo.table 0.6 released

Erik Tollerud erik.tollerud@gmail....
Fri Jul 23 19:48:52 CDT 2010


PyPI is perfectly happy with just getting the source code and no
binary packages, although as Russell noted, it's also possible to
upload compiled binaries (the distribute package has the tools to help
the process). I've never really had any version trouble with PyPI as
long as the package is actually using the proper distutils way of
specifying version numbers - the rules are spelled out in
http://docs.python.org/distutils/setupscript.html#relationships-between-distributions-and-packages
, but if you use  an incompatible numbering scheme, PyPI thinks you're
uploading an older version, and that can indeed cause trouble...

I agree PyPI can be trouble for binary eggs, though... some people
seem to be able to make it work, but I always just compile it from
source anyway, like pip does.  I'd say the best approach is to just
post the source code and if someone really cares about it, leave it to
them to build binaries for their architecture as they want them.

The main thing I think is important is to have some sort of download
link on PyPI so the automated tools like pip or setuptools can resolve
dependencies automatically.


On Wed, Jul 21, 2010 at 3:06 PM, Russell Owen <rowen@uw.edu> wrote:
> I like PyPI for serving source code.
>
> I am not personally very impressed with binary eggs -- in my limited
> experience it's easy to install in an incompatible system or have
> easy_install get confused and download stuff you've already got. In
> fact I use pip instead of easy_install partly for this reason (also to
> avoid easy_install's mega-pth file).
>
> For Mac OS X I prefer to ordinary installers using bdist_mpkg. If you
> go that route you could serve the source code from PyPI and have a
> link to a separate site for downloading binary installers.
>
> -- Russell
>
> On Jul 21, 2010, at 2:44 PM, Michael Droettboom wrote:
>
>> I have no experience submitting to PyPI (I generally get annoyed at
>> packages that go out to PyPI because there's a better than even chance
>> it loads the wrong version of something).  Since vo.table includes
>> compiled code, does that mean I have to provide packages for all
>> platforms?
>>
>> Mike
>>
>> On 07/21/2010 01:53 PM, Erik Tollerud wrote:
>>> Do you have plans to submit vo.table as a package to pypi? It would
>>> be
>>> useful to be able to specify it as a requirement for packages that
>>> might want to access VOTables, but currently the python install tools
>>> need a package to have a pypi entry to resolve the dependencies (at
>>> least, automatically)...
>>>
>>> On Wed, Jul 21, 2010 at 7:28 AM, Michael
>>> Droettboom<mdroe@stsci.edu>  wrote:
>>>
>>>> I have made a new release of vo.table, a library for parsing,
>>>> validating
>>>> and generating VOTable XML format files.
>>>>
>>>> This release adds the following features since 0.5:
>>>>
>>>> Changes since 0.5
>>>> =================
>>>>
>>>> - Initial support for conesearch, simple image access and simple
>>>>   spectrum access protocols.  This support should be considered
>>>>   experimental and subject to drastic changes in the future.
>>>>
>>>> - Support for VOTable 1.2
>>>>
>>>> - Improved compatibility with VOTable 1.0
>>>>
>>>> - Support for Python 3.x
>>>>
>>>> - All vo-specific warnings and exceptions have a number.  These
>>>>   warnings and exceptions are defined in the documentation.
>>>>
>>>> - Warnings and error messages are more detailed.
>>>>
>>>> - Only warnings indicating a violation of the specification are
>>>> given
>>>>   type VOTableSpecWarning.
>>>>
>>>> - Improved compatibility with Python 2.6
>>>>
>>>> - C extension for writing XML, resulting in 2x speedup
>>>>
>>>> - Generally faster
>>>>
>>>> - More workarounds for invalid VOTABLE files in the wild
>>>>
>>>> BUGS:
>>>>
>>>> - Loading binary tables without an explicit 'nrows' is now fixed.
>>>>
>>>> - pyfits is not required -- will only be used when necessary.
>>>>
>>>> Mike
>>>>
>>>> --
>>>> Michael Droettboom
>>>> Science Software Branch
>>>> Space Telescope Science Institute
>>>> Baltimore, Maryland, USA\
>
> _______________________________________________
> AstroPy mailing list
> AstroPy@scipy.org
> http://mail.scipy.org/mailman/listinfo/astropy
>



-- 
Erik Tollerud


More information about the AstroPy mailing list