return types were incorrectly hinted as null, should be void

[mindplay-dk]

Corrected all return type-hints from null to void to avoid failed inspections in e.g. Php Storm.

(strict inspection would state that a function must return if annotated with return null - as opposed to return void, which means the function does not use a return-statement and has no defined return-value.)

[GrahamCampbell]

This is technically a breaking change to the interfaces.

[Seldaek]

It's not really though, functionally speaking, and as we have no type hints, it doesn't matter at all. I'd be +1 for merging here, seeing as PHP7.1 made void official, but to merge the change on the PSR itself I guess you'll have to go through the ML.

[michaelcullum]

Please send a message to the mailing list if anyone objects, as this might be considered a breaking change; if nobody does object then we can get this merged here and on the spec. If people do object then it can either go to a vote (if people care enough either way) or it needs to remain as-is here and on the spec.

[mindplay-dk]

This is technically a breaking change to the interfaces.

Technically, but in practice, no one has been writing return null statements at the end of every method implementation.

So in practice, this is not a breaking change, it's a bugfix.

For that matter, it's not even a bugfix, it's a documentation fix :-)

[michaelcullum]

Sure, it's just it's better to make sure on the ML nobody objects first. :)

[GrahamCampbell]

It's not really though, functionally speaking, and as we have no type hints, it doesn't matter at all.

That thing is really a typehint documentation wise. It says all implementations must return null, but changing it to void means all implementations must not return. See why I'd say this is breaking?

[Seldaek]

That's quite a pedantic argument though, given that return; and return null; are 99% equivalent in PHP, and that it would not break anything at all in userland.

[mindplay-dk]

See why I'd say this is breaking?

@GrahamCampbell yes, but show me one implementation that returns? It's technically a break change, and I already said as much, but since there's not a line of code actually implementing it "correctly" according to the incorrect doc-block, changing this effectively breaks nothing.

[GrahamCampbell]

Yeh. I'm just playing devil's advocate. I do think this change is fine. I just want to make sure all points are properly considered. :)

[mindplay-dk]

By the way, it just occurred to me - the included AbstractLogger, LoggerTrait and NullLogger implementations also do not implement the interfaces according to their own annotations.

So this really is a bugfix.

[Seldaek]

OK I think we all agree here this is insignificant so merging, we wasted enough time on this IMO.