Dale Glass

I think setting it later might work?

Built, doesn't start. Gets stuck on: ``` [06/01 23:34:30] [CRITICAL] [hifi.networking] PAGE: Failed to lookup address of STUN server stun1.l.google.com [06/01 23:34:30] [WARNING] [hifi.networking] PAGE: Failed to lookup public address...

Okay, interesting. We already have a suitable test. Build with `cmake -DBUILD_TESTS=ON`, `make all-tests` Run the test in `tests/networking/networking-PacketTests` Here's from master: ``` ********* Start testing of PacketTests ********* Config:...

False alarm, it's an assertion triggered by a test. Fixed in #993

After #994, test `networking-ResourceTests` works, and resolves the lookup correctly. Hm. ``` QDEBUG : ResourceTests::downloadFirst() hifi.networking: QHostInfo lookup result for "stun1.l.google.com" with lookup ID 1 is "74.125.250.129" QDEBUG : ResourceTests::downloadFirst()...

Okay, I finally appear to be getting somewhere. Lookups actually work. The problem is that the code never receives the answer, because the interface thread gets stuck. It gets stuck...

It gets stuck on this: ``` #0 syscall () at ../sysdeps/unix/sysv/linux/x86_64/syscall.S:38 #1 0x00007fffacab8010 in __cxxabiv1::__cxa_guard_acquire (g=0x7ffff43ff660 ) at ../../../../libstdc++-v3/libsupc++/guard.cc:348 #2 0x00007ffff437834f in (anonymous namespace)::Q_QGS_globalData::innerFunction () at ssl/qsslsocket.cpp:387 #3 QGlobalStatic::operator() (this=)...

Right. PR #996 when applied to the conan branch reproduces the problem. Second test just hangs. There's something very wrong here given that this runs just Qt's own functionality.

> Which version of OpenSSL do you have installed on your system? It appears to be some sort of Qt internal initialization bug and you can actually fix it by...

Can't this break some existing voxels?