Switch to email header parser from deprecated cgi library

[pib]

Fixes #160

Based on the answer here https://stackoverflow.com/a/77225775

Also removed the use of cgi.parse_header for the location header since as far as I can tell, that was never actually necessary.

[peterbe]

Hmm. Why aren't tests running??

[hartwork]

@peterbe it says "2 workflows awaiting approval". You need to press the "Approve" button.

[peterbe]

I think it's a bug. That option isn't appearing.

[hartwork]

I think it's a bug. That option isn't appearing.

@peterbe could be!

What I see:

[peterbe]

When I switch to the old merge box experience, that button appears.

[hartwork]

When I switch to the old merge box experience

@peterbe could you elaborate what you mean by that and what you did to change now?

[peterbe]

When I switch to the old merge box experience

@peterbe could you elaborate what you mean by that and what you did to change now?

Perhaps because I'm GitHub staff, I have access to a preview version of the merge-box. So mine looks different.
I disabled the preview version and went back to the regular one, and then that option appears.

[peterbe]

if tests are passing, I'm cool with it.

[peterbe]

if tests are passing, I'm cool with it.

You, @hartwork ?

[hartwork]

if tests are passing, I'm cool with it.

You, @hartwork ?

@peterbe I'm looking at it now from a few different angles, I'm hoping to come back with a result within the next hour.

[hartwork]

@peterbe I have done two things now for review:

1. Check what the HTTP 1.1 spec had to say on the value of header Location back then at https://datatracker.ietf.org/doc/html/rfc2616#section-14.30. Good news is it's just a URI so the change here is fine. Bad news is: It no longer has to be absolute, I'll open a dedicated issue in a minute about how that breaks hashin code (but it is unrelated to the issue at hand here).

2. Play with the new code with a non-trivial example of a Content-type value, looks good, see below:

In [1]: content_type = 'application/json; charset="utf8"'

In [2]: import cgi

In [3]: cgi.parse_header(content_type)
Out[3]: ('application/json', {'charset': 'utf8'})

In [4]: from email.headerregistry import HeaderRegistry

In [5]: parsed = HeaderRegistry()("content-type", content_type)

In [6]: parsed.content_type
Out[6]: 'application/json'

In [7]: parsed.params['charset']
Out[7]: 'utf8'

So approving… 👍

Thanks @pib for taking us into the Python 3.13 era with hashin! 🙏