I’m continuing to make progress on libnng. Some folks have started testing
it, reported problems, and that has helped me find and fix bugs early. I
hope to have a compatibility layer available in the next day or two. It
won’t be complete, but it should be enough for most apps.
Meanwhile, observers may notice that I’ve gone ahead and closed a bunch of
issues against nanomsg itself, without actually integrating code. That’s
because a number of issues are already fixed in libnng.
What this means is that I believe the future of nanomsg is nng. We’re not
ready to retire libnanomsg yet, but I strongly feel that the day that we
make that decision is actually not that far off. So in the meantime, I’m
declining to make non-critical changes to libnanomsg, so that I can stay
focused on getting nng as stable and solid as possible, and to make sure
that we don’t have any regressions in it relative to nanomsg.
Folks that have reservations about this, feel free to give me a shout out.
I’m especially interested to hear from people using nanomsg in platforms or
environments where nng might be viewed as difficult — for example older
versions of nanomsg running on Windows XP or older, or embedded systems
that don’t have a reasonable threading model or condition variables. Also
platforms where people are stuck using pre-C99 compilers. (I am using a
very limited subset of C99 extensions to C89 — mostly // comments and
structure initializers at this point.)
nng also insists on Windows on having Vista or better, and at least Visual
Studio 2013 (mostly because of some modest but important improvements in
the C runtime — like having a functional snprintf().) Do you have to build
your product on Windows systems with older compilers than this? (Why?!?)
Let me know!
- Garrett
