Sigwait support

[Dan Pat]

Hi. A third-party library I am trying to link in is making use of the "sigwait" function, so I am getting an unresolved reference. By grepping through the emscripten sources I was able to find an implementation of the function in the musl library. While I could just get the file compiled as part of my project, I suspect there is a mechanism for passing this requirement to emscripten so that it compiles what's necessary and puts it in the cache for later reuse. Is there such a mechanism?

[Dan Pat]

And, it turns out compiling those system files it is not as straightforward as I thought) So I definitely need a way to tell emscripten what I need!

среда, 9 декабря 2020 г. в 12:46:46 UTC+5, Dan Pat:

[Floh]

It looks like the emscripten SDK support for the signal-related syscalls is only very rudimentary (see https://github.com/emscripten-core/emscripten/blob/master/src/library_signals.js). In case of sigwait() that's understandable because it wouldn't fit well into the browser-loop model, since it's not simply possible to yield execution back to the browser loop (which is a problem for any blocking POSIX function), so a sigwait() implementation would probably need to be built around ASYNCIFY (https://emscripten.org/docs/porting/asyncify.html).

What's a bit strange is that the musl source in emscripten implements sigwait (which indirectly calls the 'syscall' SYS_rt_sigtimedwait), but library_signals.js doesn't at least provide an empty syscall stub for sigtimedwait. This looks like an oversight or TODO to me.

As a workaround Maybe you can provide en empty sigwait() C function (inside an 'extern "C"' { }' if you're in C++) in your own code, which then would be choosen by the linker instead of musl's implementation. This would at least fix the linking problems (but if the library actually calls sigwait this would break of course, a "proper" sigwait() implementation would need to be more advanced and involve ASYNCIFY).

[Dan Pat]

I got my hands on the library's source code and disabled use of the posix signals. Thanks for the advice, though - this made be update my knowledge on the subject) 

ср, 9 дек. 2020 г. в 16:42, Floh <flo...@gmail.com>:

--
You received this message because you are subscribed to the Google Groups "emscripten-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to emscripten-disc...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/emscripten-discuss/d52776e9-3eda-402b-bc23-08306c006e35n%40googlegroups.com.