Simplification about tortilla HOT 2 CLOSED

Do you have a specific need to use your actual implementation,
instead of simply concatenating path segments ?

Yes.

The wraps in my implementation are programmed to overload settings
from their parents. So, when a request is executed, the information
will flow from top to bottom. Every wrap can add their default config
if it's not there.

Now, this could also be implemented by going from bottom to top.
However, wraps are made to be saved for later use. This means that a
bottom to top solution could behave weirdly. For example:

example = tortilla.wrap('https://api.example.org')

# worst case scenario of endpoint
users = example('get').users('for', 'each', params={'this': 'that'})

# we've logged in and set the authorization header
example.headers.authorization = 'secret key'

# this won't send the authorization header because the `example` parent
# can't modify the request information
users.get('someone')

If we use a top to bottom method, the request will fall through our
example wrap which can add the headers and other config.

URLs are created seperately from the request in the url() method.
This method concatenates URL parts from top to bottom and could be
done in the request method. But, because the __repr__ method also
uses the URL, I chose to abstract the URL creation in a different
function to keep things DRY.

I'm completely open to modifications to improve performance, but I
don't want it to impact the flexibility of using the library. :)

from tortilla.

OK I understand better your point of view, my basic implementation clearly can't handle this, but I never though about storing intermediates values, nor modifing them and expecting it to still work ;-)

from tortilla.