Kinda curious about this feedback.

Isn't @phpstan-ignore-line function.alreadyNarrowedType clearer ?

I understand it as Ignore function.alreadyNarrowedType in the current line
As opposite to phpstan-ignore-next-line

Are phpstan-ignore-line and phpstan-ignore the same ?

my understanding is

• @phpstan-ignore-line always ignores all errors of a single line (does not support a by-identifier filter)
• @phpstan-ignore-next-line always ignores all errors of a single line (does not support a by-identifier filter)
• @phpstan-ignore requires at least one identifier and ignores only errors of such identifier (on the same line)

Exactly. I suspect people do this mistake often but I don't know how to prevent it.

I didn't know that, and made the mistake a lot then
https://github.com/search?q=repo%3AVincentLanglet%2FTwig-CS-Fixer+phpstan-ignore-next-line&type=code

Especially because I don't like adding @phpstan-ignore on the same line and prefer to ignore before the line (so I use ignore-next-line a lot).

as far as I remember there is a open feature request (cannot find it right now), which goal is to disallow @phpstan-ignore-line and @phpstan-ignore-next-line and allow only ignoring errors by identifier (via opt-in config)

alternatively we could look at @phpstan-ignore-line and @phpstan-ignore-next-line like we do for @phpstan-ignore - but report a error if we find a valid issue-identifier after it