Properly preserve encoded slashes when using base_url

[florimondmanca]

Update: Prompted by #1405

An alternative to #1406 that tweaks URL merging to operate on .raw_path, rather than .path. Rationale: .path is URL-decoded and might swallow pre-existing %2F characters into /. On the other hand .raw_path doesn't do URL-decoding, so it is suitable for handling encoded slashes as well.

[lovelydinosaur]

Great, yup this is a reasonable change.

I'm going to scrub the "Closes 1405" notation in the description, since the OP there doesn't note any usage of base_url, so although it's been a useful prompt onto this, which is a nice little tweak, I don't think it's relevant to what they're seeing.

[florimondmanca]

@tomchristie Makes sense! Are we ✅ on this one?

[lovelydinosaur]

Lovely & neat, yup!

[jbaayen]

I can confirm that this PR does fix the concrete test of https://github.com/encode/httpx/issues/1405#issuecomment-737163659.

[ Apologies for previous confusion in this comment, I had ended up with parallel installations of 0.15.0 and this branch. ]

12:33 $ pip show httpx
Name: httpx
Version: 0.16.1
Summary: The next generation HTTP client.
Home-page: https://github.com/encode/httpx
Author: Tom Christie
Author-email: tom@tomchristie.com
License: BSD
Location: /home/jbaayen/httpx
Requires: certifi, sniffio, rfc3986, httpcore
Required-by: kisters.water.time-series, kisters.water.time-series.tsa, kisters.water.operational.access-control
✔ ~

[lovelydinosaur]

👍 Thanks all!

[jbaayen]

Thank you!

Do you know yet when a new release might be made?