bpo-46072: Include length in stats for UNPACK_SEQUENCE.

[markshannon]

Even more stats!

https://bugs.python.org/issue46072

[brandtbucher]

Nice, I'm curious to see the results.

(I'm gonna be really surprised if anybody is out here unpacking zero-length iterables, though.)

[sweeneyde]

(I'm gonna be really surprised if anybody is out here unpacking zero-length iterables, though.)

To my surprise, this works:

() = [] 

[brandtbucher]

To my surprise, this works:

() = [] 

...for some definition of "works". The whole thing gets peepholed out. ;)

I think maybe once long ago I used something like this to "assert" that an iterator I used was completely exhausted (the idea being that it would raise a TypeError if not):

it = iter(something) # ...do some stuff... [] = it # Fails if there are any remaining items.

I thought it was pretty clever at the time. :)

[markshannon]

UNPACK_SEQUENCE 0 does occur. In this:

def f(x): match x: case T(): pass

 1 0 RESUME 0 2 2 LOAD_FAST 0 (x) 3 4 LOAD_GLOBAL 0 (T) 6 LOAD_CONST 1 (()) 8 MATCH_CLASS 0 10 COPY 1 12 POP_JUMP_IF_NONE 10 (to 20) 14 UNPACK_SEQUENCE 0 4 16 LOAD_CONST 0 (None) 18 RETURN_VALUE 3 >> 20 POP_TOP 22 LOAD_CONST 0 (None) 24 RETURN_VALUE 

[brandtbucher]

UNPACK_SEQUENCE 0 does occur.

Interesting. That's a common case that's worth improving.