[Python-Dev] Re: [Import-sig] Re: Proposal for a modified import mechanism.

Prabhu Ramachandran Prabhu Ramachandran <prabhu@cyberwaveindia.com>
Tue, 13 Nov 2001 13:09:31 +0530

>>>>> "FL" == Fredrik Lundh <fredrik@pythonware.com> writes:

    >> Its not the application that I'm concerned about - an
    >> application is typically a single/few file(s) and editing them
    >> to suit things is certainly not an issue. But editing 100 files
    >> inside a package each time the parent changes is nuts.

    FL> I'd recommend that you don't put the above in your PEP.

    FL> changing the language to cater to package owners who have more
    FL> files than users (and don't appear to give a damn about the
    FL> few users they have) isn't just nuts, it's total madness.

Ouch!  Yes, I might be mad but I guess you are taking my statement
totally out of context.

  (1) I was talking of _my_ particular application where I have a
  script that imports the package depending on whether the application
  is installed or frozen.  I wasn't talking of every Python

  (2) How does modifying the behaviour of import or better still
  adding a couple of new keywords (aimport, rimport etc.) that do
  things a little differently affect the all the users of Python?
  What we proposed will only affect the package in question and not
  every user in the system.  AFAIK, what was proposed would also not
  break backward compatibility in any way.

  (3) As to me not giving a 'damn' about the 'few users' that I may
  have, I think that is fairly untrue and all I have to say is that I
  wouldn't have considered posting to c.l.p* with this proposal if it
  wouldn't be of use to someone else.  It would have been very easy
  for me to have solved my problem and left things at that.

    FL> if you do this a lot, write a 20-line script to do it for you.

I'm afraid I already did better than that. <wink> I modified knee.py
to do exactly what I wanted and it does work.  Also, even without such
a hack I can get my package running, thanks.  However, the point is
that Gordon, Eric, some others and myself think that it would be nice
if the import mechanism were a little nicer to package developers.
The problem is not something that I am personally desperate for.  I'm
just trying to see if Python's import mechanism can be improved (if
necessary).  Flaming me is pointless. :)