On Tue, Jul 18, 2006 at 02:44:47PM -0500, Dirk Eddelbuettel wrote:
> On 18 July 2006 at 12:35, Keith Hellman wrote:
> | Package: python-rpy
> | Version: 0.99.2-4
> | Severity: important
> | 
> | The rpy python modules appear to be installed into the pycentral
> | tree, but they are not linked to the pythonX.Y/site-packages
> | directories.
> 
> Do they have to be? Is that specified somewhere?

I'm not familiar enough with either the new python policy to know if and
where it is specified.  I simply noticed that the latest Numeric
packages (using dh_pycentral have the links).  

What I do know is that my python is not finding the rpy modules, because
it doesn't look in /usr/share/pycentral/blah/blah/blah for modules.  It
sounds (from what limited I've read about pycentral) that it (pycentral) 
is supposed to make these links *for* the python module packager.  It
doesn't appear to be happening in the case of rpy.
 
> As far as I can see, this just works:
> 
> [EMAIL PROTECTED]:~> python
> Python 2.3.5 (#2, Jun 13 2006, 23:12:55)
> [GCC 4.1.2 20060613 (prerelease) (Debian 4.1.1-4)] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
> >>> from rpy import *
> RHOME= /usr/lib/R
> RVERSION= 2.3.1
> RVER= 2031
> RUSER= /home/edd
> Loading Rpy version 2031 .. Done.
> Creating the R object 'r' ..  Done
> >>>
> 
> The supplied examples chisqure.py and animation.py [ in python-rpy-doc ]
> still work...   This is on my current testing box:
> 
> [EMAIL PROTECTED]:~> dpkg -l python python-central python-rpy python-numeric| 
> grep ^ii | cut -c-78
> ii  python         2.3.5-11       An interactive high-level object-oriented la
> ii  python-central 0.5.1          register and build utility for Python packag
> ii  python-numeric 24.2-5         Numerical (matrix-oriented) Mathematics for
> ii  python-rpy     0.99.2-4       Python interface to the GNU R language and e
> 
> which appear to be the same packages as you have.
> 
> So I think I would need you to ask you to show me where this is a bug, esp.
> one of Severity: important. Right now I don't see it. 

Do you have RPY in your PYTHONPATH?  Perhaps for development reasons?
If you uninstall python-rpy* do you have rpy files still in
/usr/lib/pythonX.Y/site-packages/ ?  Perhaps they are remnants of a
previous install (I think this is in fact how stumbled across this,
there were residual rpy files in site-packages that didn't like the new
R version string).
 
> | I assume that they need to be symlinked by pycentral, that is what
> | seems to happen with Numeric (which works).  Not being a debian
> | developer, I'm not really able to make heads or tails of the Debian
> | Python Policy.  I downloaded the source to the python-rpy package,
> | but found all the ``How to Use pycentral'' steps seemingly done.
> | So I'm sorta at the stopping point of my knowledge and abilities.
> 
> In a case like this, wouldn't it be better to first inquire about the
> situation at hand, either on the debian-python list, or with the respective
> maintainer. [ I can't help much here, I am much more familiar with R than
> with Python ... ]
> 

Well, now I feel like a putz for submitting a bug-report.  I honestly
didn't know what the best course was (I didn't know there was a
debian-python list), and I guess I thought Emailing the maintainer
directly would be a little presumptuous.  If you'd like me to delete
(nullify? close? whack-with-a-big-stick?) the bug report I'm happy to do
so, didn't know it was the wrong course to take.

Sorry for the trouble.

-- 
Keith Hellman                             #include <disclaimer.h>
[EMAIL PROTECTED]                from disclaimer import standard
public key @ pgp.mit.edu B5354B76


"While it might seem that a simple increment operator is an atomic
operation, there's no guarantee that it is.  It's actually possible for
Thread 1 to update half of a 32-bit x while Thread 2 reads the full 32
bits, getting a mishmash instead of a valid integer.  (And that's only
one thing that can go wrong.  Another is that compiler optimization might
leave the integer in a register.  You really can't ever let threads
simultaneously access data without protection.)"

-- Marc J. Rochkind, *Advanced Unix Programming*, Second Ed.

Attachment: signature.asc
Description: Digital signature

Reply via email to