You can never go down the drain...

[Tim Peters]

[/F posts a one-line patch to sre_parse.py]

[Philip]
> Thanks, but... EEEK!!
>
> If we could get all our heads out from under the hood for a moment,

There was nothing wrong with original regexp.  That's why our heads are
under the hood.

> I feel compelled to restate my question. Is there any regular expression
> I can code besides this one that does what I was trying? The point of
> the exercize is to _not_ change the source of the package.
>
>     assert ('Gold Man' == re.sub('([a-z])([A-Z])', r'\1 \2', 'GoldMan'))

Sure:  change

    r'\1 \2'

to

    r'\1\040\2'

That worms around the bug under the hood without changing the source of the
package.

someday-you'll-think-regexps-are-really-boring<wink>-ly y'rs  - tim