Re: Experimental Python packages

[dman <dsh8290@rit.edu>]

On Fri, Sep 07, 2001 at 08:33:28AM -0500, Ben Burton wrote:
| 
| > In any case, Jython and CPython really do need to be able to co-exist
| > peacfully.
| 
| They certainly coexist peacefully.  No problem there.  All I'm saying is that 
| it doesn't support *.so CPython modules.  And this is somehow unavoidable 
| since jython is pure java.
| 
| So my concern is that if an administrator has /usr/bin/python pointing to 
| jython and random package foo provides its own foo.so cpython module and 
| proceeds to run /usr/bin/python to open it, there will be problems.

Here's the problem -- the random package foo depends on a certain
version of CPython for it's C extension.  Rather than explicitly using
that version it is assuming that /usr/bin/python is that version of
CPython.  This foo package is likely to break even in the face of
multiple co-existing cpython installations.

| Hence one of the reasons I wouldn't want to offer jython as a /usr/bin/python 
| alternative.

The real problem is making assumptions about what /usr/bin/python is
beyond what the RefMan says.  The same sort of problem occurs if a
script writer assumes /bin/sh is bash and uses bash-isms rather than
sticking to the POSIX specification because /bin/sh could be any POSIX
compliant shell (ie ash).

-D