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

[Steve Holden]

"Toby Dickenson" <tdickenson at devmail.geminidataloggers.co.uk> wrote in
message news:5q52vtgjkn38etn7tpu8eji4r9k0oaguk9 at 4ax.com...
> barry at zope.com (Barry A. Warsaw) wrote:
>
> >Why does B have to add packages to A's namespace?  Why can't the B
> >author simply use distutils to ensure that vanilla A is installed,
> >import the bits and pieces of A that you want to expose, overriding
> >what you want to change, and export an interface through B that
> >clients can use instead of A?
>
> Two reasons have ocurred in my projects over the last year:
>
> 1. A and B have the same author, and he want to
>    *permanently* repackage B inside A. Maybe he
>    developed A independantly without considering
>    where it would be finally packaged
>
In which case the re-packaging work is one-off (US: one-time-only) and
should not be seen as an unbearable load.

> 2. Maybe A needs to use an old version of B, or
>    a version of B that has been hacked for an
>    A-specific reason.
>
In which case it would seem more sensible to rename it as hacked_B, since
surely madness otherwise lies down the road.

regards
 Steve
--
http://www.holdenweb.com/