bpo-30380: Pin the version of Sphinx used to build the documentation #1612
Conversation
|
@Mariatta, thanks for your PR! By analyzing the history of the files in this pull request, we identified @birkenfeld, @benjaminp and @tiran to be potential reviewers. |
Doc/Makefile
Outdated
| @@ -107,7 +107,7 @@ clean: | |||
|
|
|||
| venv: | |||
| $(PYTHON) -m venv venv | |||
| ./venv/bin/python3 -m pip install -U Sphinx | |||
| ./venv/bin/python3 -m pip install sphinx~=1.5.6 | |||
There was a problem hiding this comment.
I think, we depend upon the latest version intentionally.
So, after this lands, #1613 can be merged and we can revert back to using the latest version?
There was a problem hiding this comment.
I think it might be better to specify the version in .travis.yml and leave the Makefile alone. My thinking is that locking Travis for testing makes sure it stays green and it's obvious what we are requiring for CI to pass, but leaving the Makefile alone makes it easier to update the docs to the latest version of Sphinx offline.
Python 2.7 also doesn't have the venv command in the Makefile but it does have the explicit pip call in .travis.yml.
Or maybe I'm just nuts and that logic is flawed. 😄
There was a problem hiding this comment.
Keeping version details out of the Makefile sounds like a better option to me. Makefiles can increase the complexity of its own :)
|
I have edited Mariatta's branch to show what I'm suggesting for editing |
|
Travis passed for both approaches so they both solve the problem, so either solution will work. |
|
No that's ok 😄 Whatever's needed to get this fixed soon. I'm at the airport so if it looks good then please merge 😃 |
brettcannon
changed the title
|
It looks like @serhiy-storchaka has fixed the 1.6.1-specific issues for at least |
|
Thanks @brettcannon and @serhiy-storchaka for actually taking care of this :) |
No description provided.