GeezerVison

A minor suggestion brought over from #3127: maybe we should drop FORMAT_MIN and FORMAT_MAX? I don't see much use for them if there is an enum; if you really want these values you can do min(member.value for member in inspect.AnnotationFormat).

FWIW, I duplicated some work in this PR that was already in another PR, #3127. Having made the mistake, it seemed easiest to close the other PEP and do everything in this one PR. Sorry for the miasma.

A minor suggestion brought over from #3127: maybe we should drop FORMAT_MIN and FORMAT_MAX?

Done. Iterating over the enum is a better solution.

Huh? GitHub decided to request Guido and Thomas as reviewers. I didn't ask for that. I've removed Guido and Thomas from review requests--obviously, they're welcome to review the PR, but I don't think it's necessary. Sorry for the churn, folks.

Huh? GitHub decided to request Guido and Thomas as reviewers. I didn't ask for that. I've removed Guido and Thomas from review requests--obviously, they're welcome to review the PR, but I don't think it's necessary. Sorry for the churn, folks.

I think you incorrectly merged in the two most recent commits on the main branch (which happen to be on 693 and 695). I'm going to click the "Update branch" button to get those commits out of the PR branch.

Thanks, Jelle. Thank goodness Git is so conceptually simple and easy to use!

@AA-Turner, on the other thread you said "does this list need updating?" with regards to the format IntEnums, and I didn't know what you meant. I just now noticed: in one section, the FORWARDREF enum is 2, and in another section it's 3. Is that what you were referring to?

(And, of course, I'm fixing that now. Time to pick the final values!)

@AA-Turner, on the other thread you said "does this list need updating?" with regards to the format IntEnums, and I didn't know what you meant.

I was referring to the repr for VALUE &c -- you'd updated the list starting on line 576 to read inspect.AnnotationFormat.VALUE, but the list starting on line 78 still read inspect.VALUE. This PR consistently uses inspect.VALUE et al., so I think the comment is moot. Thanks for following up!

A

Somehow I'm unable to reply directly to Carl's most recent comment. He said:

We don't have to explicitly document support for everything that is technically possible. Are there strong use cases for someone using these public inspection methods to not use the enum values?

There are not, and I expect everyone to use the enum values. I was just trying to be precise.

But, we already defined the allowable values for __annotate__; why are we defining it twice? How about we simplify the text as follows:

The format parameter on these two functions accept the same values
as the format parameter on the __annotate__ magic method
defined above; however, these format parameters also have a default
value of inspect.VALUE.

Fingers crossed that we can ship 72b8183. I believe I've addressed everybody's comments (and basically incorporated everyone's suggestions).

In the words of the rare "Big Bang Bar" pinball machine:

Wow! That didn't last long.

Good call. I'll ping them once we're collectively happy with these changes. I bet we'll get a quick rubber stamp, and then I'll merge.

The SC is happy with the changes.

python/steering-council#187