f-string internal formatting to be consistent with non-f-string formatting

[ColemanTom]

Is your feature request related to a problem? Please describe.

Consider a simple python script:

import numpy as np
a=np.arange(9).reshape((3,3))
print(a[1,2])
print(f"a[1,2] = {a[1,2]}")

Running black on this it would by default provide:

$ black --diff a.py
--- a.py        2024-04-19 01:29:28.562943+00:00
+++ a.py        2024-04-19 01:29:52.064944+00:00
@@ -1,4 +1,5 @@
 import numpy as np
-a=np.arange(9).reshape((3,3))
-print(a[1,2])
+
+a = np.arange(9).reshape((3, 3))
+print(a[1, 2])
 print(f"a[1,2] = {a[1,2]}")
would reformat a.py

All done! ✨ 🍰 ✨
1 file would be reformatted.

You can see the final two print statements are inconsistent formatting now. The first has inserted a space after the comma in a[1,2], but, the f-string has not adjusted the formatting of {a[1,2]} to be {a[1, 2]}.

If I accept the formatting black suggests, and then run flake8 on the same piece of code, flake8 will complain about the lack of a space after the comma in the f-string.

$ black a.py
reformatted a.py

All done! ✨ 🍰 ✨
1 file reformatted.
$ flake8 a.py
a.py:5:22: E231 missing whitespace after ','

Describe the solution you'd like

Formatting applied to code outside of f-strings should also apply to the code within the {..} in the f-strings.

Describe alternatives you've considered

Disable that warning in flake8.

Additional context

N/A.

[JelleZijlstra]

Duplicate of #567.