Strip RIGHT-TO-LEFT OVERRIDE and LEFT-TO-RIGHT OVERRIDE characters from AUTHORS.txt #12517
Conversation
|
This is marked draft because it has no changelog entry -- I suppose it does not need one. Happy to receive feedback. |
…om AUTHORS.txt Since we don't know anything about the names, we assume anything in between RIGHT-TO-LEFT OVERRIDE and LEFT-TO-RIGHT OVERRIDE (or the end of the name) should be spelled backwards. This resulted in a duplicate author name, because it uses different Unicode form. So I also added a call to unicodedata.normalize. Fixes pypa#12467
hroncok
force-pushed
the
stripoverrides
branch
from
cd91152 to
239c1d4
Compare
DiddiLeija
added
the
skip news
|
Feel welcome to mark this ready for review! |
|
Possible improvement: It might be useful to put the conversion into the pre-commit config. For example, we have author sorting on that level and it's nice that it can be autofixed: https://github.com/aio-libs/aiohttp/blob/0e91eb0/.pre-commit-config.yaml#L79. |
|
+1 to having a pre-commit hook for this. |
|
I am a bit lost. A pre-commit hook to do what exactly? Instead of generating the AUTHORS file as we desire, we would generate it with the RTL OVERRIDE characters, and only then would we fix it by a pre-commit hook? How is that better? |
|
I thought this file is also modified externally. I didn't realize you're modifying the script that produces it in the first place. I suppose your solution is correct, then. |
|
Gonna merge and see what happens. |
Since we don't know anything about the names, we assume anything in between RIGHT-TO-LEFT OVERRIDE and LEFT-TO-RIGHT OVERRIDE (or the end of the name) should be spelled backwards.
This resulted in a duplicate author name,
because it uses different Unicode form.
So I also added a call to unicodedata.normalize.
Fixes #12467