It was before

I want to tell most participants on the True asnyc thread to just fuck off. It annoys me. The RFC is great but in a few details. And people discuss such basic stuff. You have no fucking clue about how amphp works. And nor understand how true async RFC actually works.

I don't agree it's great, but the thread is noisy for no purpose mostly ...

I don't agree with the approach technically or in a procedural sense ... any RFC this large and complex, written by one person is going to devolve into these kinds of conversations, and I tried to impress on the author that before we merge something this complicated we need to have discussions about the most basic of decisions taken when the code was written, so that we can agree on the fundamentals of the proposal .. things are being said and done that I don't think make much sense

when I asked the author privately about them, they couldn't really explain those decisions, they said they were going to change course and move very slowly and build consensus and they haven't done that ... so I don't much see the point in talking publicly ....

we can disagree, and we can ask each other why we are thinking the way we are thinking ... but someone who wrote this as their first ever C project failed that test (and it is their first C project, by the way) ... this isn't the way it is because of deep analytical thought about why it should be this way, it's because it's the only way they could see to do it ...

Also, he did split up his initial RFC. So that's great. The initial RFC was definitely far too large.

oh I must have misunderstood, you said you were ready to merge it ...

Sorry, I meant accept the RFC.

But this is a thing where the impl exists and shows that the concept works. Then we have the RFC and vote on this.

Refining the impl can always be done later, in doubt.

But their impl is not that bad. Just a bit chaotic.

the text of the RFC is simpler, but the implementation is more or less unchanged, and decisions like deprecating fibers are still built in, there is still the assertion that "we can do this magical thing and nobody needs to rewrite their code or think about it differently", and that is incorrect ... we pass this, with the declaration that "everything will be magically okay" and we open a can of worms trying to make that statement true ... it will not be okay ...

krakjoe.github.io/em partials v2 rfc available for playing ...

what was actually said wasn't "I lack the knowledge to know how to implement this on the zend side", it was "these things are fundamentally incompatible" ... they are a world apart, and communicate very different things ...

sorrry, I think it's fixed now.... in case not just for the record 3v4l.org/kuLmD#v8.4.15

I need some education about building PHP extensions, more specifically installing them.

So I filed opcache fails on install modules/* to extension dir #20557 and a patch Fix GH-20557: Build make -C ext/* install-modules w/o modules/* #20558

So it's about the Opcache extension that since PHP-8.5 does not build an opcache.so file any longer, right?

So why does the build still have one header file in it's install image?