Url path seems to be os-dependant

[tweber36]

The construc of the url seems to be very os-dependant because it uses os.path
On Windows the generator didn't work because path.join uses os.sep (which is \ on Windows) :

>>>from os import path
>>>path.join("https://projecteuler.net", "problem=1")
'https://projecteuler.net\\problem=1'
>>>path.join("https://projecteuler.net/", "problem=1")
'https://projecteuler.net/problem=1'

I had to add a forward slash in the base_url to make it work.

I'd suggest maybe using urllib :

>>>import urllib
>>>urllib.parse.urljoin("https://projecteuler.net", "problem=1")
'https://projecteuler.net/problem=1'

[ewjoachim]

I highly appreciate the fact you took the time to try out this project, and submit an issue here ! I think there probably are not many people who've used it so far, and I guess you're the first one to try it on a Windows machine.

It's quite clear you have cornered the bug. For my part, I haven't touched this project since Nov 2015, and I realize it's lacking quite a few basic features (continuous integration, real documentation beyond Readme, ...)

I suggest 3 pathes, and you can let me know which one you'd like best :

• You make a PR with urllib, in exchange, I plug Tox & Travis in with the existing tests, linting and such. And because the project needs to be tested on Windows, we'll plug AppVeyor too, so no more surprises. Of course, anything we fix gets released right away.
• If, for any reason, you're not submitting a PR, I will fix the bug myself, but I'm not sure I have the motivation to setup a CI for myself on a project I don't use anymore 😅 . Of course, I'll release it too.
• Or the other way around : if you're using the project, and I'm not, I can give you maintenance rights. This way, you can make the project evolve regarding any limitation you find while using it, and I'll be glad to review PRs. If you want, you can have PyPI access and push your own releases too.

So what's your choice ?