mSource assertion fails in OpenALStream::init() about opentesarena

[Ragora]

It looks like it could potentially just be the wrong variable being used there. Let me explain. According to the OpenALStream constructor, the mSource value will be set to 0: https://github.com/afritz1/OpenTESArena/blob/master/OpenTESArena/src/Media/AudioManager.cpp#L406

Cool, but the reason I say there might be wrong values in use here is the init itself using mSource:
https://github.com/afritz1/OpenTESArena/blob/master/OpenTESArena/src/Media/AudioManager.cpp#L477

Why? It appears maybe it wants to assert(source != 0) because we're initializing mSource with the source value input to ::init: https://github.com/afritz1/OpenTESArena/blob/master/OpenTESArena/src/Media/AudioManager.cpp#L510

Which makes sense as to why mSource might not be initialized (well, why it might be 0).

[afritz1]

I have a feeling it should say assert(source != 0). @kcat?

@Ragora, I think you should use the link of the actual commit instead of using master. It would go out of date if master was updated, right?

[Ragora]

Also ::reset will pretty much just reinitialize the unique_ptr with a new pointer value, destroying the old one if there is a non-NULL value stored: http://en.cppreference.com/w/cpp/memory/unique_ptr/reset

[kcat]

I have a feeling it should say assert(source != 0).

Actually that should've been assert(mSource == 0); to ensure the init method isn't being called again. The source parameter is valid, there's no way it can be called with 0 (if it is, there's more serious issues to worry about). I'll get a fix for that shortly.

Really annoying that cmake's RelWithDebInfo build type disables assertions, even though debug info is supposed to remain. I would've caught it earlier.