[parallel] bump default threads from 16 to 32 to keep up with modern technology #6877
Conversation
| @@ -190,7 +190,7 @@ final public function enterNode(Node $node): int|Node|null | |||
| * Replacing nodes in leaveNode() method avoids infinite recursion | |||
| * see"infinite recursion" in https://github.com/nikic/PHP-Parser/blob/master/doc/component/Walking_the_AST.markdown | |||
| */ | |||
| public function leaveNode(Node $node): array|int|Node|null | |||
| final public function leaveNode(Node $node): array|int|Node|null | |||
There was a problem hiding this comment.
On purpose, this should have never been allowed.
|
I will check the rector-downgrade-php unit test error, it possibly due to latest phpstan |
Thanks 👍 PHPStan x.14 now returns |
|
It seems due to this change phpstan/phpstan-src@3854cbc It seems the only way we can do is just follow it since it not error when set as magic https://3v4l.org/IrE2q#v7.4.33 I will update the fixture. |
|
I restarted the CI, seems green now 👍 |
|
@TomasVotruba rebase is needed 👍 |
…technology, alligned with phpstan setup
…g mechanism in child classes
|
Thanks 👍 |
| ->withImportNames(removeUnusedImports: true) | ||
| ->withImportNames() |
There was a problem hiding this comment.
It seems you were testing speed? I think this can be re-enabled again, it already fast on rector-src, and this also enabled in other rector-* packages rector.php.
There was a problem hiding this comment.
I thought the removeUnusedImports: true is enabled by default.
I think we should enable it, so default behavior is the one with least work. What do you think?
There was a problem hiding this comment.
imo, current optional implementation already correct, it should be optional, as it "remove things" that can possibly there on purpose.
There was a problem hiding this comment.
I see, I'll restore it here then.
No description provided.