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).
On Wednesday, 9 December 2020 at 08:54:56 UTC+1 [email protected] wrote:
> 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:
>
>> 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?
>>
>
--
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 [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/emscripten-discuss/d52776e9-3eda-402b-bc23-08306c006e35n%40googlegroups.com.
