Documentation misleading on query string encoding

[bsolomon1124]

Long story short

The docs in Passing Parameters in URLs is misleading and the behavior is essentially the opposite of what is stated:

You can also pass str content as param, but beware – content is not encoded by library. Note that + is not encoded.

Expected behaviour

Passing the fully encoded querystring params=url=https%3A%2F%2Fwww.katherinetimes.com.au%2Fstory%2F6005621%2Fwife-of-ex-nissan-boss-ghosn-leaves-japan%2F%3Fsrc%3Drss&api_key=XXXX should not requote the querystring. I'm assuming that's what "content is not encoded by library" is supposed to mean.

What is more accurate is that passing a URL with encoded=True to the url of request() is the only way to get this behavior.

Actual behaviour

From client_reqrep.py:

            q = MultiDict(url.query)
            url2 = url.with_query(params)
            q.extend(url2.query)
            url = url.with_query(q)

Okay, let's try to interpret the above:

>>> import urllib.parse
>>> from yarl import URL
>>> 
>>> base = "https://api.example.com" 
>>> params = {
...     "url": "https://www.katherinetimes.com.au/story/6005621/wife-of-ex-nissan-boss-ghosn-leaves-japan/?src=rss",
...     "api_key": "XXXX",
... }
>>> u = URL(base).with_query(urllib.parse.urlencode(params))
>>> u
URL('https://api.example.com/?url=https://www.katherinetimes.com.au/story/6005621/wife-of-ex-nissan-boss-ghosn-leaves-japan/?src%3Drss&api_key=XXXX')
>>> urllib.parse.urlencode(params)
'url=https%3A%2F%2Fwww.katherinetimes.com.au%2Fstory%2F6005621%2Fwife-of-ex-nissan-boss-ghosn-leaves-japan%2F%3Fsrc%3Drss&api_key=XXXX'

So the query string in u is requoted! (encoded=True always in .with_query().)

What are the docs trying to say here? Am I misinterpreting the phrase "content is not encoded by library"? How so?

Steps to reproduce

See above.

Your environment

aiohttp client 3.5.4, Python 3.7.2, yarl 1.3.0.