guthrie wrote:
> I'm pretty new to Python, and trying to parse the grammar.
>
> Q: What is the scope of the testlist in a list_for?
>
> For example;
> Instead of;
> for x in [ x in dict if dict[x]=="thing" ]:
> in this:
> for x in dict and dict[x]=="thing":
> x is undefined.
In the above two examples, x is a regular variable (global or local
depending on whether it's in a function). It will remain undefined
until the first time something gets assigned to it.
In your first example (where you appear to have left out a for), it get
bound inside the list comprehension, after evaluating 'dict' but before
evaluating the test 'dict[x]=="thing"'.
I doubt the second example does what you want it to, but here's what
happens. Python tries to evaluate the expression 'dict and
dict[x]=="thing"' first, before assigning anything to x. Because that
expression uses x, you get an undefined error. If it hadn't had the
undefined error, it would have tried to iterate over the result (which
would have been True or False), which would have raised a TypeError
exception.
> And why doesn't this work:
> for x in dict if dict[x]=="thing":
Because it would would just be another way to spell:
for x in dict:
if dict[x] == "thing:
Yes, we've heard all the arguments before. The Python language
designers like to have only one way to do things when they can, and in
this case they felt that none of the objections (an extra level of
indentation or unsightly continue) warranted a new spelling.
I recall there was to be a (prerejected) PEP submitted about this
recently, but it doesn't appear to be up.
> Any insights/hints on why it is broken?
It's not broken :)
Carl Banks
--
http://mail.python.org/mailman/listinfo/python-list
