[Python-ideas] Fwd: unpacking generalisations for list comprehension

[Steven D'Aprano]

On Thu, Oct 13, 2016 at 03:28:27PM +0000, אלעזר wrote:

> It may also suggest that there are currently two ways to understand the
> *[...] construct, 

This thread is about allowing sequence unpacking as the internal 
expression of list comprehensions:

    [ *(expr) for x in iterable ]

It isn't about unpacking lists:

    *[...]

so I don't see what relevance your comment has.

There may be two or three or ten or 100 ways to (mis)understand list 
comprehensions in Python, but only one of them is the correct way. List 
comprehensions are (roughly) syntactic sugar for:

    result = []
    for x in iterable:
        result.append(expression)

Any other understanding of them is incorrect.

Now if people wish to make an argument for changing the meaning of 
comprehensions so that the suggested internal unpacking makes sense, 
then by all means try making that argument! That's absolutely fine.

In the past, I've tried a similar thing: I argued for a variant list 
comprehension that halts early:

    [expr for x in iterable while condition]

(as found in at least one other language), but had that knocked back 
because it doesn't fit the existing list comprehension semantics. I 
wasn't able to convince people that the value of this new comprehension 
was worth breaking the existing semantics of comprehensions.

Maybe you will be able to do better than me.

But understand that:

    [*(expr) for x in iterable] 

also fails to fit the existing list comprehension semantics. To make it 
work requires changing the meaning of Python list comps. It isn't enough 
to just deny the existing meaning and insist that your own personal 
meaning is correct.



-- 
Steve