-
-
Notifications
You must be signed in to change notification settings - Fork 7.9k
DOC: Fix missing-reference generation on Windows #26850
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
d435f00
to
a75e062
Compare
Ah, I think because on load and comparison it runs through: matplotlib/doc/sphinxext/missing_references.py Lines 135 to 142 in 2cbdff8
which with the Windows-style paths gives us always doc/C so everything matches, perhaps?
|
The missing references file confuses me/I also I think more or less had the linter ignore it - I'm just not sure which of those I'm supposed to care about. |
e1651de
to
f762b4c
Compare
So it looks like CI is getting additional unused references:
but I don't see them locally. I guess that's probably because of some differing versions of things. I'll leave that option off for now. |
A possibly good first issue is running through those and commenting out the corresponding ignored_warnings on the linters |
This file causes the |
On Windows, the path is absolute and contains a colon after the drive letter, so splitting on colon results in trying to relativize 'c' (or whatever the drive is) to the base directory, which just makes the final path into base directory + file path.
f762b4c
to
5e5378e
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The changes look fine, but I am slightly confused as to why we weren't getting warnings for all of the things in that file that had the windows paths.
Are we ignoring the path portion entirely somehow? if so should it just be removed in total from the json file?
I get that we have config on whether or not to warn about unused entries in this file (which is currently set to not warn for those), but presumably many of these are valid entries that should be included in the ignores.
Where were you expecting those to show up? |
I guess I was expecting the warning that should have been silenced to still show up (and flag as a warning/get extracted by the circle build step etc. |
I looked again; it appears that the answer is yes. Only the keys are used for ignoring warnings: The file paths appear to be only to help people to investigate ignored things, and to make the warnings for unused ignores. |
…850-on-v3.8.x Backport PR #26850 on branch v3.8.x (DOC: Fix missing-reference generation on Windows)
PR summary
On Windows, the path is absolute and contains a colon after the drive letter, so splitting on colon results in trying to relativize 'c' (or whatever the drive is) to the base directory, which just makes the final path into base directory + file path.
These missing references all broke in #26628, though I'm not sure why builds never failed since they aren't on Windows, or using the same base paths.
PR checklist