Hi folks,
writing an email [1] to answer a few RPC-related questions, I started to
wonder:
Am I missing something, or does PMT *not* export a C ABI, though it's
API is pretty much C++-agnostic: Everything has signatures of the pattern
pmt_t pmt_operation(pmt_t input, void* parameters...)
Trying to test my hypothesis, I ran a
nm --defined-only libgnuradio-pmt.so|grep -v _Z
which in fact showed not a single un-C++-mangled pmt_* symbol.
Is there a reason we're doing this? As far as I can tell, the only
obstacle when declaring the whole of pmt.h 'extern "C"' would be that
pmt_t is actually a boost::intrusive_pointer... I'm not 100% sure, but
wouldn't having an external C API just handle the typedef as "um, yes,
that is a specific type", and work as is? Or do we need to add C
wrappers, after having named every PMT function very C naming-ish?
Best regards,
Marcus
[1] "Remote access to running GNU Radio application's USRP objects (was
Re: [USRP-users] Multiple parallel connections to x310)"
_______________________________________________
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
No comments:
Post a Comment