Raymond, you merged this just before I tried to submit a review suggesting changes and expansion to include all the fixes needed in the file. I cannot submit the review and AFAIK, cannot turn pending review comments into regular comments, so I am copying most of them here before before cancelling the review.

Tests: In the 3 tests before this, the suffix 0, 2, or 3 is the number of arguments given. This would suggest 0a, 0b, 2a, and 3a for the suffixes here. I like 1min, 2min, 1max, and 2max even better. You can wait a bit for other reviews before changing.

Blurb:

Let us be more specific. For the patch as is,
Fix pluralization of 'argument' in TypeError message with 'expected at least/most 1 argument', instead of '1 arguments'.

But if we add all the remaining fixes needed in this file (see below), we could write.
Fix pluralization in TypeError messages in getargs.c: '1 argument' instead of '1 arguments' and '1 element' instead of '1 elements'.

Fixes: This patch follows existing expressions formatted into argument%s:
line 380: (nargs < min ? min : max) == 1 ? "" : "s",
lines 1667 and 2093: (len == 1) ? "" : "s", (parens not needed)

Scope: I think we should finish the job in the file. The next line in both else clauses,
"unpacked tuple should have %s%zd elements,"
should get the same treatment. This may break existing tests. (These are the only two formatted 'elements'.

Fixes for the following will break some test, which will then need fixing also.
542: "expected %d arguments, not %" PY_FORMAT_SIZE_T "d",
1720, "%.200s%s takes %s %d positional arguments"
1799, "%.200s%s takes %s %d positional arguments"
2106, "%200s%s takes %s %d positional arguments (%d given)",
2154: "%.200s%s takes %s %d positional arguments"

I checked all PyErr_Format entries and did not see anything other than %d arguments and %d elements.