[ new ] Adding Basic and Bearer Authorization

[Matthew-Mosior]

This PR focuses on adding Authorization header functionality:

• Both Basic authentication and Bearer authentication parsing functionality.
• Functionality to base64 encode a username and password to create a proper Basic authentication Header.
• Functionality to create a proper Bearer authentication Header.

[jarlah]

Lgtm

[jarlah]

But no tests though for it

[jarlah]

And cant start workflow since im not maintainer role

[Matthew-Mosior]

But no tests though for it

Let me look into adding a simple test or two.

[Matthew-Mosior]

But no tests though for it

@jarlah This should be ready to look at again.

[jarlah]

First, good work 👊

But, I guess this is the one that fails ? Can you add different error messages for the two tests? So its easier to see what test fails? And i dont think using idris crash is how the other tests are designed to fail. I could be wrong of course, but i dont remember seing a test crash like this before.

[Matthew-Mosior]

@jarlah

There was a small oversight on my part here (easy fix):

basicauth here is Basic [base64encodedstring], but I was accidentally treating it as [base64encodedstring].

Thus I needed to break basicauth apart, and call base64DecodeString only on the [base64encodedstring] part.

This should be fixed via 24e30c8

[Matthew-Mosior]

Hmm, it failed again, let me take a look.

[jarlah]

In a test, code like this is fine. In normal code i would have asked to create a method to keep my brain from trying to decode it. 🤣

[jarlah]

Are you running tests locally btw. I can try running the tests locally to see if its mismatch between workflow and local, when i get time. Not at computer right now

[jarlah]

no hacks in user config. You could try pack update-db? i destroyed my setup now and its bootstrapping pack .. taking ages

[jarlah]

if rebuilding you should get these two when running tests

[ warning ] Package base64 uses custom build hooks. Continue (yes/*no)?
yes
[ warning ] Package tls uses custom build hooks. Continue (yes/*no)?
yes

if not prompt is disabled which it is in github workflow

[jarlah]

@Matthew-Mosior after full rebuild of pack and with default contents in .pack/user/pack.toml, i can still run the tests. Maybe remove stuff from .pack/.cache/tls? Or maybe i should just reinstall pack completely and see if its an issue with clean start.

[jarlah]

I will test the stuff in repl later today ;) you could also use pack repl and test the individual operations

[Matthew-Mosior]

@jarlah

I've fixed this function, turned out to be a really silly oversight:

It had to do with the break function.

break wasn't stripping the leading white space in Basic YWxhZGRpbjpvcGVuc2VzYW1l, causing base64DecodeString to be called on completely erroneous input (should have been called on YWxhZGRpbjpvcGVuc2VzYW1l, not YWxhZGRpbjpvcGVuc2VzYW1l).

Using split accomplished what needed to be done here.

[jarlah]

Some whitespace but otherwise green as a bean 🫘

[jarlah]

"Fix whitespace or not fix whitespace. Its your choice." 😂

[jarlah]

@Matthew-Mosior agree? ☝️

[Matthew-Mosior]

@Matthew-Mosior agree? ☝️

Thank you for fixing the white space, looks great! 👍