Two issues with python-mode snippets

[joostkremers]

When I open a Python file, I get two yasnippet warnings. First:

Ignoring unknown directive "NOTE" in file: /home/joost/.emacs.d/elpa/yasnippet-snippets-20240115.1138/snippets/python-mode/function_docstring

function_docstring has the following line in its header:

# NOTE: Use minimum indentation, because Emacs 25+ doesn't dedent docstrings.

Removing the colon (or the entire "NOTE: " part) solves the issue.

Second:

Multiple snippets with same identity: "__iter__"

There are two snippets for __iter__, one in a file called iter and one in a file called __iter__. Unfortunately, both use __iter__ for the name: directive. Since both snippets essentially do the same thing, perhaps one can be removed?

I know these are minor issues, but the warnings are displayed in the echo area and may push out other, more important, warnings. (Such as an issue with Eglot, in my case...)

[swinkels]

I just had a look at the warning about the unknown directive "NOTE". That exact line has been in the snippet for 7 years so it must be something in yasnippet itself. Indeed, yasnippet commit joaotavora/yasnippet@25f5d88 (19 January) added the warning message for unknown directives.

Removing the colon seems like a valid workaround, but of course a pull request to modify the snippet is better.

[swinkels]

About the message for the multiple snippets with the same identity, I see that message is added in yasnippet commit joaotavora/yasnippet@8e9ce76 (12 February). Looking at the code comment, it is allowed for one snippet to overwrite another one, but apparently this is not one of these cases.

[swinkels]

@JulienPalard Your commit 9f67370 adds Python snippet file __iter__. This file adds a snippet with the name __iter__, just as the much older snippet file iter does. This leads to a "Multiple snippets with same identity" message.

The snippets are slightly different. Is there a need to support both but if not, which should be the one to keep?

[JulienPalard]

Thanks for reporting.

I opened pr #500 to remove one of the two iter. I kept the one consistent with the other snippets.