[Python] What does <type>.<attr> do during a match/case?

In general, the code <type>(arg1=value1) is changed to the code <type>(attr1=value1) during a case in a match statement. This allows for some very interesting and complex capturing.

For example, we can ask "does my instance have a attr called 'sort', if so capture it":

match [1,2,3]:
case list(sort=f): # this isn't a call, it is an attr-capture statement
print(f)


Since the first (and only) positional argument of the list instance is the list values, you can capture those items as well:

match [1,2,3]:
case list((x,*y),sort=f): # here we are capturing the 'attr' called 'sort' as 'f'
print(x,y,f)


Likewise, you can specify the matching values instead of capturing them into variables:

g = list.sort
match [1,2,3]:
case list((1,*y),sort=g): # here we ask if the attr 'sort' is the same as 'g'
print(y)


However, I am now confused by the following:

match [1,2,3]:
case list(x,sort=list.sort): # fails to match, but why???
print(x)


I would have thought that this fails because list.sort means something different than to specify "the function 'sort' of 'list'". Maybe sort=list.sort means "do I have an attr called 'sort' which itself is of 'type' 'list' that also has an attr called 'sort'". So I tried:

match [1,2,3]:
case list.sort: # also fails, why?
print('ya')


So, list.sort does not mean that I want type list with an attr of sort. So, then I tried:

match list.sort:
case list.sort: # does match, wait, what?
print('ya')


Which works. So, list.sort during a case does mean to use the function of sort from the type of list, but then why did list(sort=list.sort) fail?


Example of capturing an attr 'y' and matching an attr 'x':

class P:
def __init__(self,a,b):
self.x,self.y = a+b,a-b
match P(5,6):
case P(x=11,y=a): # a is a new variable, 11 is a match
print(a)

Continue reading...