pathname2url: normalize forward slashes to backslashes before processing

[Kreijstal]

fixes #182
Fix pathname2url to handle paths with forward slashes correctly by normalizing all forward slashes to backslashes at the start of processing. This prevents issues where paths containing forward slashes would result in malformed URLs with double slashes after the drive letter (e.g. "C://Users" instead of "C:/Users").

Before:

>>> pathname2url("C:/Users/example")
'//C://Users/example'

After:

>>> pathname2url("C:/Users/example")
'//C:/Users/example'

The function now handles both forward and backward slashes consistently.

[Kreijstal]

is test_urllib buggy in purpose? why 3 forward slashes, that doesn't work

[lazka]

How is this related to the mingw port? Is it a regression we introduced?

[Kreijstal]

How is this related to the mingw port? Is it a regression we introduced?

jupyterlab on normal windows seem to open the browser but on mingw it doesn't
would you recommend to patch jupyterlab on mingw instead?

[lazka]

Using the official CPython build:

Python 3.12.7 (tags/v3.12.7:0b05ead, Oct  1 2024, 03:06:41) [MSC v.1941 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import urllib.request
>>> urllib.request.pathname2url("C:/Users/example")
'///C://Users/example'
>>>

Not saying that there isn't a problem, but I doubt this is the root cause.

[naveen521kk]

I think it has to do something with our conversion system from \\ to /. One thing I could observe is:

Python 3.12.7 (main, Nov  3 2024, 19:54:34)  [GCC 14.2.0 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> from urllib.request import pathname2url
>>> pathname2url(r"C:/hello")
'///C://hello'
>>> pathname2url(r"C:\\hello")
'///C:/hello'
>>>

that's why I think things work with the official Python but not with ours. Trying out MSYSTEM= jupyter lab seems to open my browser without any issues.

I think this should be fixed upstream.