SA Checks: fromIterable, empty #96
Conversation
| /** @var array<NewTKey, NewT> $emptyArray */ | ||
| $emptyArray = []; | ||
|
|
||
| return self::fromIterable($emptyArray); |
There was a problem hiding this comment.
this is the only way to get around having template parameters for the empty array, otherwise PHPStan complains because it cannot figure out the template types:
301 Method loophp\collection\Collection::empty() should
return loophp\collection\Collection<NewTKey of
(int|string), NewT> but returns
loophp\collection\Collection<*NEVER*, *NEVER*>.
There was a problem hiding this comment.
Should we report that as a bug ?
There was a problem hiding this comment.
yeah leave it with me. Looks like a valid issue though maybe I thought it makes sense because it has no way to know where the template parameters are coming from
There was a problem hiding this comment.
done! phpstan/phpstan#5065
Let's see what they say
| * @template TKey of int | ||
| * @template T of array{0: NewTKey, 1: NewT} | ||
| * | ||
| * @template NewTKey of array-key | ||
| * @template NewT |
There was a problem hiding this comment.
I literally must've spent over an hour getting these type hints right 😄
There was a problem hiding this comment.
but it's beautiful when the static analysis finally turns green ✅ 😆
There was a problem hiding this comment.
damn looks like this is not actually supported yet in PHPStan 😄. Not sure why it was showing all good in another file
There was a problem hiding this comment.
trying to work around it makes this worse and incorrect, so I think I'll keep this. Psalm supports it and it passes, and PHPStan should support it soon: phpstan/phpstan#3931
There was a problem hiding this comment.
Can't wait for this to be supported !
There was a problem hiding this comment.
It's now supported! phpstan/phpstan-src#673
| - phpstan-docs-baseline.neon | ||
| - phpstan-unsupported-baseline.neon |
There was a problem hiding this comment.
-
the docs baseline is just issues in
docs/so nothing important but fixing them requires diving into other methods, which I will do anyway in future PRs. -
the unsupported baseline refers to the PHPStan unsupported feature I mentioned here. It's limited to the
unpackoperation and I expect we'll be able to get rid of it once@template T of arrayis supported
There was a problem hiding this comment.
Nice and clean!
This is why I was using @psalm-template
By the way, what do you think about removing that of array-key ?
As this is a lazy library, the TKey can be of any type, so, I don't think I'm wrong by saying we can get rid of it. WDYT ?
There was a problem hiding this comment.
I think you're right we could get rid of array-key but I don't think it can only be done in one place because it will probably result in static analysis errors.
I'll look to prepare a PR after this to remove it everywhere and see how many errors we get 😄
There was a problem hiding this comment.
Yeah I did it in #95 maybe it could be a starting point.
|
@drupol btw what's the process for releases and how often do we do them in this project? |
There is actually no strict release process and I'm absolutely not against to setup one. If you want, I can make a point release this morning. Do you prefer to wait until this PR is merged or I do it before? |
This PR is ready from my point of view so let me know if there's anything left to address 😁 . |
|
4.0.7 released! |
|
Thanks again for your amazing contributions! |
This PR adds to the static analysis checks:
fromIterable,empty.tests/in GrumPHPIterableIteratorUnpack-> surfaced by thesquashoperation now thatfromIterablehas the right type hints