Improve for loop index variable type narrowing

[cdce8p]

Preserve the literal type of index expressions a bit longer (until the next assignment) to support TypedDict lookups.

fromtypingimportTypedDictclassX(TypedDict):
hourly: intdaily: intdeffunc(x: X) ->None:
forvarin ("hourly", "daily"):
print(x[var])

Closes#9230

[cdce8p]

Mypy primer results

scrapy

+ scrapy/downloadermiddlewares/cookies.py:145: error: "str" has no attribute "decode"; maybe "encode"? [attr-defined]+ scrapy/downloadermiddlewares/cookies.py:152: error: "str" has no attribute "decode"; maybe "encode"? [attr-defined]

These are true positives. All TypedDict values str or bool. The intention is probably to apply decode to extra bytes items. Annotating it would require PEP 728.

https://github.com/scrapy/scrapy/blob/b4bad97eae6bcce790a626d244c63589f4134408/scrapy/downloadermiddlewares/cookies.py#L141-L145

[JukkaL]

Nice to see several now-unnecessary type ignores in mypy_primer output! Looks good, just one minor test-related suggestion.

[JukkaL]

Test b = a followed by reveal_type(b).

[github-actions]

Diff from mypy_primer, showing the effect of this PR on open source code:

pydantic (https://github.com/pydantic/pydantic)
+ pydantic/v1/networks.py:345: error: Unused "type: ignore" comment [unused-ignore]
scrapy (https://github.com/scrapy/scrapy)
+ scrapy/downloadermiddlewares/cookies.py:141: error: Unused "type: ignore" comment [unused-ignore]+ scrapy/downloadermiddlewares/cookies.py:142: error: Unused "type: ignore" comment [unused-ignore]+ scrapy/downloadermiddlewares/cookies.py:145: error: Unused "type: ignore" comment [unused-ignore]+ scrapy/downloadermiddlewares/cookies.py:145: error: "str" has no attribute "decode"; maybe "encode"? [attr-defined]+ scrapy/downloadermiddlewares/cookies.py:145: note: Error code "attr-defined" not covered by "type: ignore" comment+ scrapy/downloadermiddlewares/cookies.py:152: error: Unused "type: ignore" comment [unused-ignore]+ scrapy/downloadermiddlewares/cookies.py:152: error: "str" has no attribute "decode"; maybe "encode"? [attr-defined]+ scrapy/downloadermiddlewares/cookies.py:152: note: Error code "attr-defined" not covered by "type: ignore" comment
discord.py (https://github.com/Rapptz/discord.py)
- discord/message.py:2265: error: TypedDict key must be a string literal; expected one of ("channel_id", "guild_id", "id", "author", "content", ...) [literal-required]
core (https://github.com/home-assistant/core)
+ homeassistant/components/image_processing/__init__.py:226: error: Unused "type: ignore" comment [unused-ignore]+ homeassistant/components/energy/data.py:334: error: Unused "type: ignore" comment [unused-ignore]
schemathesis (https://github.com/schemathesis/schemathesis)
+ src/schemathesis/cli/__init__.py:260: error: Unused "type: ignore" comment [unused-ignore]