gh-130928: Fix error message during bytes formatting for the 'i' flag

[ApostolFet]

Error output before correction:

>>> b"%i" % "str"
Traceback (most recent call last):
  File "<python-input-0>", line 1, in <module>
    b"%i" % "str"
    ~~~~~~^~~~~~~
TypeError: %d format: a real number is required, not str

Error output after correction:

>>> b"%i" % "str"
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
    b"%i" % "str"
    ~~~~~~^~~~~~~
TypeError: %i format: a real number is required, not str

• Issue: Not accurate when outputting an error during byte formatting for the 'i' flag, the 'd' flag is output, which can be misleading #130928

[jstasiak]

I think I'd add a unit test for this.

[sobolevn]

Yes, please, add a unit test :)
And also a NEWS entry. This is a user-visible change.

[sobolevn]

Thanks, I think that the fix is correct.
But, let's wait for @serhiy-storchaka review, because his is the code owner.

[sobolevn]

You can also add yourself to https://github.com/python/cpython/blob/main/Misc/ACKS

[ApostolFet]

You can also add yourself to https://github.com/python/cpython/blob/main/Misc/ACKS

Thanks, but I've already taken advantage of that opportunity earlier

[sobolevn]

I would also suggest to create a new issue to add tests for bytes formatting options. Please, tag me - I can assign you this issue if you want.

[sobolevn]

@serhiy-storchaka friendly ping, what do you think? :)

[serhiy-storchaka]

I think that test_common_format in test_format.py may be a better place for the tests. There are already tests for %d and %x. You can just add

        test_exc_common('%i', '1', TypeError,
                        "%i format: a real number is required, not str")

Or you can leave tests in both places.

[serhiy-storchaka]

LGTM.

[sobolevn]

Thanks again!

[sobolevn]

We usually do not backport error message changes.