Allow joining URL and pathlib.Path

[ukch]

Given the following URL and Path objects:

>>> url = URL("http://whatever/foo")
>>> path = pathlib.Path("bar/baz")

I would expect them to be joinable using code such as:

>>> url / path
URL('http://whatever/foo/bar/baz')

Instead, I receive this traceback:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
    url / path
  File "/path/to/lib/python3.6/site-packages/yarl/__init__.py", line 281, in __truediv__
    name = self._PATH_QUOTER(name)
  File "yarl/_quoting.pyx", line 226, in yarl._quoting._Quoter.__call__
TypeError: Argument should be str

[asvetlov]

How do you want to join absolute path?

[ukch]

Currently, attempting to join a (string) path starting with a slash onto an URL raises ValueError: Appending path starting from slash is forbidden. I would say the same behaviour should occur with absolute Path objects.

[asvetlov]

Another question is: should symlinks be resolved?

Now I have a feeling that url / str(path) is simple and good enough. Why do we need to blow up the API?

[asvetlov]

I have a feeling that the request is out of the library scope.

[Bobronium]

I stumbled upon this issue while trying to answer this question: is pathlib.Path suitable for dealing with url path?

The immediate concern that I have is that on windows str(pathlib.Path('foo') / 'bar') would result in a 'foo\bar'.

So, answering my own question: you probably shouldn't use pathlib.Path as url path constructor.

But you desperate to do otherwise, using pathlib.PurePosixPath should at least avoid problem with separator described above.