Merge proposal: zope.interface/branches/tseaver-no_2to3
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Having merged the 'tseaver-better_unittests' branch to the zope.interface trunk last Friday, I now have a new branch ready for merging: http://svn.zope.org/zope.interface/branches/tseaver-no_2to3 The branch excises the use of the 'lib2to3' module and associated fixers in favor of a "compatible subset" which supports Python 2.6, 2.7, 3.2, and PyPy 1.8. At this point all tests pass under all four environments using 'python setup.py test' as well as using 'nosetests.' Code coverage is at 99.9% (two edge-case lines are untested). The biggest advantage of the subset approach is that installing the package now takes on the order of seconds (2 seconds on my machine) rather than minutes (2 minutes my machine). Running the tests is also much faster. The secondary win is that running the same code under Py3k makes for better support / bug reports (line numbers match, etc.) Unless the community's consensus is against the branch, I plan to merge it to the zope.interface trunk by early next week. Tres. - -- =================================================================== Tres Seaver +1 540-429-0999 tseaver@palladion.com Palladion Software "Excellence by Design" http://palladion.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk9+AkcACgkQ+gerLs4ltQ5LSQCgnnf/dW4ENZXaFkG/42e469LZ CzEAmwQOGqVcauz7bpugJ+824vBzt6iw =fl2m -----END PGP SIGNATURE-----
On Thu, Apr 5, 2012 at 4:36 PM, Tres Seaver <tseaver@palladion.com> wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Having merged the 'tseaver-better_unittests' branch to the zope.interface trunk last Friday, I now have a new branch ready for merging:
http://svn.zope.org/zope.interface/branches/tseaver-no_2to3
The branch excises the use of the 'lib2to3' module and associated fixers in favor of a "compatible subset" which supports Python 2.6, 2.7, 3.2, and PyPy 1.8. At this point all tests pass under all four environments using 'python setup.py test' as well as using 'nosetests.' Code coverage is at 99.9% (two edge-case lines are untested).
The biggest advantage of the subset approach is that installing the package now takes on the order of seconds (2 seconds on my machine) rather than minutes (2 minutes my machine). Running the tests is also much faster. The secondary win is that running the same code under Py3k makes for better support / bug reports (line numbers match, etc.)
Unless the community's consensus is against the branch, I plan to merge it to the zope.interface trunk by early next week.
Woo hoo! Jim -- Jim Fulton http://www.linkedin.com/in/jimfulton
Am 05.04.2012, 22:36 Uhr, schrieb Tres Seaver <tseaver@palladion.com>:
Unless the community's consensus is against the branch, I plan to merge it to the zope.interface trunk by early next week.
+ many/lots! Charlie -- Charlie Clark Managing Director Clark Consulting & Research German Office Kronenstr. 27a Düsseldorf D- 40217 Tel: +49-211-600-3657 Mobile: +49-178-782-6226
On Thu, Apr 05, 2012 at 04:36:23PM -0400, Tres Seaver wrote:
Having merged the 'tseaver-better_unittests' branch to the zope.interface trunk last Friday, I now have a new branch ready for merging:
http://svn.zope.org/zope.interface/branches/tseaver-no_2to3
The branch excises the use of the 'lib2to3' module and associated fixers in favor of a "compatible subset" which supports Python 2.6, 2.7, 3.2, and PyPy 1.8.
Yay! Curious: why getattr(meth, _IM_FUNC)? AFAICS starting with Python 2.6 you can use the Python 3 spelling of meth.__func__. The same applies to _FUNC_CODE, _FUNC_DEFAULTS etc. Is it a relic of an older effort to also support Python 2.5?
Unless the community's consensus is against the branch, I plan to merge it to the zope.interface trunk by early next week.
+1 Marius Gedminas -- http://pov.lt/ -- Zope 3/BlueBream consulting and development
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 04/06/2012 10:39 AM, Marius Gedminas wrote:
Curious: why getattr(meth, _IM_FUNC)? AFAICS starting with Python 2.6 you can use the Python 3 spelling of meth.__func__. The same applies to _FUNC_CODE, _FUNC_DEFAULTS etc. Is it a relic of an older effort to also support Python 2.5?
Nope, just my ignorance: the code I extracted that from still used 'im_self', 'im_func', 'func_code', and 'func_defaults', and I didn't check that the newer spellings would work on 2.6. Thanks for the catch! Tres. - -- =================================================================== Tres Seaver +1 540-429-0999 tseaver@palladion.com Palladion Software "Excellence by Design" http://palladion.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk9/BSsACgkQ+gerLs4ltQ4X7gCgjAALlpdmdbRT//1uxbBNRlXN Kh0Anidld1IPE1RKApot5DgThkvcXsau =wJ3l -----END PGP SIGNATURE-----
* Tres Seaver <tseaver@palladion.com> [2012-04-05 22:36]:
Having merged the 'tseaver-better_unittests' branch to the zope.interface trunk last Friday, I now have a new branch ready for merging:
http://svn.zope.org/zope.interface/branches/tseaver-no_2to3
The branch excises the use of the 'lib2to3' module and associated fixers in favor of a "compatible subset" which supports Python 2.6, 2.7, 3.2, and PyPy 1.8. At this point all tests pass under all four environments using 'python setup.py test' as well as using 'nosetests.' Code coverage is at 99.9% (two edge-case lines are untested).
Woot! However alluring the 2to3 idea might have been, it seems to me that the single-source approach (with the six package if needed) is the only sane way to go.
Unless the community's consensus is against the branch, I plan to merge it to the zope.interface trunk by early next week.
+1, please do! Wolfgang
participants (5)
-
Charlie Clark -
Jim Fulton -
Marius Gedminas -
Tres Seaver -
Wolfgang Schnerring