Avoid parsing ForwardRef contents as type references

[charliermarsh]

Summary

Today, we parse the inner contents of ForwardRef (e.g., the "List[int]" in ForwardRef("List[int]")) as a forward annotation. However, no other tools treat them like that. E.g., if you run this code through Flake8, or Mypy, or Pyright:

from typing import TypeVar

X = TypeVar("X", "List[int]", int)
X = ForwardRef("List[int]")

...none of them flag the second List[int] as an undefined reference to List.

Closes #3678 although there's another fix coming to avoid a panic there.

[github-actions]

PR Check Results

Ecosystem

✅ ecosystem check detected no changes.

Benchmark

Linux

group                                      main                                   pr
-----                                      ----                                   --
linter/all-rules/large/dataset.py          1.00     13.5±0.04ms     3.0 MB/sec    1.00     13.5±0.05ms     3.0 MB/sec
linter/all-rules/numpy/ctypeslib.py        1.00      3.5±0.01ms     4.7 MB/sec    1.01      3.6±0.01ms     4.7 MB/sec
linter/all-rules/numpy/globals.py          1.00    485.9±0.75µs     6.1 MB/sec    1.03    499.1±5.00µs     5.9 MB/sec
linter/all-rules/pydantic/types.py         1.00      5.9±0.01ms     4.3 MB/sec    1.01      6.0±0.02ms     4.3 MB/sec
linter/default-rules/large/dataset.py      1.00      7.1±0.01ms     5.7 MB/sec    1.01      7.2±0.01ms     5.7 MB/sec
linter/default-rules/numpy/ctypeslib.py    1.00   1608.3±2.31µs    10.4 MB/sec    1.00   1602.9±2.38µs    10.4 MB/sec
linter/default-rules/numpy/globals.py      1.00    179.0±0.61µs    16.5 MB/sec    1.02    181.8±9.41µs    16.2 MB/sec
linter/default-rules/pydantic/types.py     1.00      3.4±0.02ms     7.6 MB/sec    1.00      3.4±0.03ms     7.5 MB/sec

Windows

group                                      main                                   pr
-----                                      ----                                   --
linter/all-rules/large/dataset.py          1.02     15.3±0.14ms     2.7 MB/sec    1.00     15.1±0.07ms     2.7 MB/sec
linter/all-rules/numpy/ctypeslib.py        1.01      4.2±0.03ms     4.0 MB/sec    1.00      4.1±0.02ms     4.0 MB/sec
linter/all-rules/numpy/globals.py          1.00    456.1±4.20µs     6.5 MB/sec    1.00    455.5±4.47µs     6.5 MB/sec
linter/all-rules/pydantic/types.py         1.00      6.7±0.05ms     3.8 MB/sec    1.00      6.8±0.05ms     3.8 MB/sec
linter/default-rules/large/dataset.py      1.00      8.2±0.06ms     5.0 MB/sec    1.00      8.2±0.06ms     5.0 MB/sec
linter/default-rules/numpy/ctypeslib.py    1.01  1780.1±11.06µs     9.4 MB/sec    1.00  1763.9±10.79µs     9.4 MB/sec
linter/default-rules/numpy/globals.py      1.00    181.9±1.76µs    16.2 MB/sec    1.01   184.4±29.31µs    16.0 MB/sec
linter/default-rules/pydantic/types.py     1.01      3.8±0.03ms     6.6 MB/sec    1.00      3.8±0.02ms     6.7 MB/sec