[Python-ideas] Combining test and assignment

[Eike Hein]

On 1/22/2012 4:48 AM, Nick Coghlan wrote:
> Simple embedded assignment, though, only works when the predicate is
> just "bool" - as soon as the condition differs from the value you want
> to access, you need to revert to the existing idiom *anyway*. Given
> the non-trivial costs of adding new syntax, that gets the idea put
> into the "not worth the hassle" bucket.

Actually, just to be clear, in the use case I had in mind
spam() is returning a numeric value. That satisfies a truth
test and is still useful in other ways (but I assume you
are actually on the same page, hence the quotation marks
around bool).

Another case I come across frequently is with the re module,
where re.match() returns either None or a MatchObject.

if re.match(pattern, string) as m:
     ... make use of the match object ...

Anyway, thank you for a comprehensive reply and recap of
the history.


> To date, nobody has been interested enough in the latter idea to put
> together a formal PEP and reference implementation for python-dev's
> consideration, and the former idea has been informally rejected
> several times due to the lack of generality.

I thought a bit about what Python offers today in terms
of built-in tools to approximate what I want. My first
thought was to try and abuse context managers, but aside
from being horribly ugly, a with statement also can't
prevent its body from being run, it seems.

Another, possibly better way is to modify locals(), i.e.:

if test(expr, 'var'):
    ... do stuff ...

Where the test function returns the expression but also
stores its results in locals()['var'].

I wonder if a function like this (with a better name
obviously) would have a place in the stdlib somewhere,
or is it too much of a limited-use hack?


> Cheers,
> Nick.

-- 
Best regards,
Eike Hein