Michał "phoe" Herda

Please attach your build logs in text files instead of storing them in the message body. The messages that you post are too large for sensible conversation.

> What is weird is the make-semaphore function has no count argument, instead it just supplies 0. Must be old news. See the current source of BT: https://github.com/sionescu/bordeaux-threads/blob/6a63fed849e34b3c36f6025d301e673d4b49b164/src/impl-clozure.lisp#L97 ```lisp (defun...

Sorry, I misread that. Yes, `ccl:make-semaphore` has no `:count` argument, as we can see here: https://github.com/Clozure/ccl/blob/275105afd94706d95ac955178316074931822c42/level-0/l0-aprims.lisp#L194-L208 We can also see that the value is initialized to `0`. This means that...

Correct - I was also thinking of passing a count argument to `semaphore-signal`. I've taken that discussion to #309.

I've investigated a little bit further and it seems that accessing the actual semaphore value will be platform-dependent and should therefore be done in C, next to `new_semaphore` in https://github.com/Clozure/ccl/blob/master/lisp-kernel/thread_manager.c#L675.

Extending on my previous comment a bit: #310 added an ability to set the semaphore count from the Lisp side. The real code that creates a semaphore is on the...

So, in other words, @sionescu - it seems to me that adding semaphore counts on macOS requires a rework of the whole semaphore CCL mechanism, and that makes it non-trivial...

Oh right - that could be done on the CL side instead of the foreign one. Welp, that simplifies the design. > If for example you passed the underlying true,...

SICL isn't yet usable as a standalone implementation.

Core compression is an important feature and I wouldn't build without it. As for detecting zlib headers, try compilling any trivial C file that includes the `zlib.h` header. If it...