Message82297
| Author |
gsakkis |
| Recipients |
dripton, gsakkis, tarek |
| Date |
2009年02月17日.03:31:38 |
| SpamBayes Score |
4.3317017e-12 |
| Marked as misclassified |
No |
| Message-id |
<1234841503.0.0.592566070878.issue2279@psf.upfronthosting.co.za> |
| In-reply-to |
| Content |
> > FWIW I wrote a module that overrides the default build_py and sdist
> > commands with versions that allow specifying package_data recursively
>
> Maybe that could be a new feature ?
That would be nice, especially if we want to reimplement MANIFEST.in as
setup() option at some point. My current implementation doesn't extend
the API, so there's no way to specify a subset of files under a
directory like recursive-include; every directory matched by a glob is
copied in whole (recursively):
import os
from distutils.command.build_py import build_py as _build_py
class build_py(_build_py):
def find_data_files(self, package, src_dir):
files = []
for p in _build_py.find_data_files(self, package, src_dir):
if os.path.isdir(p):
files.extend(os.path.join(par,f)
for par,dirs,files in os.walk(p)
for f in files)
else:
files.append(p)
return files
> > (while preserving file permissions, unlike the - buggy IMO - behavior of
> > distutils) so that I can get rid of the MANIFEST.in.
>
> Sounds like a bug to me, could you fill an issue on that ?
If it's a bug, it's certainly not accidental; there's a big XXX comment
justifying this choice but I'm not convinced. I posted about it at
http://mail.python.org/pipermail/python-list/2009-January/524263.html;
if you think it's a bug I'll fill an issue. |
|
History
|
|---|
| Date |
User |
Action |
Args |
| 2009年02月17日 03:31:43 | gsakkis | set | recipients:
+ gsakkis, dripton, tarek |
| 2009年02月17日 03:31:43 | gsakkis | set | messageid: <1234841503.0.0.592566070878.issue2279@psf.upfronthosting.co.za> |
| 2009年02月17日 03:31:40 | gsakkis | link | issue2279 messages |
| 2009年02月17日 03:31:38 | gsakkis | create |
|