It turns out that functools.reduce() is simply __builtins__.reduce().
That does not stop the DeprecationWarning from using reduce() from being
raised even though the message says to use functools.reduce()!

Easiest solution is to create a wrapper for __builtin__.reduce() that
silences the warning when called.

Can't this be handled by 2-to-3 instead of a -3 warning?

On Wed, Jul 30, 2008 at 11:32 PM, Raymond Hettinger
<report@bugs.python.org> wrote:
>
> Raymond Hettinger <rhettinger@users.sourceforge.net> added the comment:
>
> Can't this be handled by 2-to-3 instead of a -3 warning?
>

Well, there is the problem of someone having a function already named
reduce() and the import. Since 2to3 can't go back and add an import,
every found use of reduce() would need to have an import inserted
somewhere. That won't work if reduce() is being used in a expression
context instead of a statement context, etc.

Alternatively, you could move reduce to functools and have the builtin
module issue a warning and call it.

On Thu, Jul 31, 2008 at 11:56 AM, Benjamin Peterson
<report@bugs.python.org> wrote:
>
> Benjamin Peterson <musiccomposition@gmail.com> added the comment:
>
> Alternatively, you could move reduce to functools and have the builtin
> module issue a warning and call it.
>

That's what 3.0 did; created _functools and that holds the original
implementation from __builtins__. But that is a lot more work. =)

Here's a patch that actually imports reduce from functools to stop
repetition.

Reviewing the patch at http://codereview.appspot.com/2968 .

Reviewed in Rietveld. Only a minor change is suggested. Otherwise I
think the patch is ready to be committed, Benjamin.

Thanks for the review! Committed in r65802.