Mainloop lock needs to be held to access PulseAudio object from outside the
event loop thread, in particular in GetPosition() calls. Otherwise we might hit
a race condition and have the call fail. There's also a possibility that the
call might randomly fail if it's called before first timing data is received
for the stream, so handle that case as well.
Also fix the stream write callback to call Mix() with correct parameters, as
well as do the work in a loop in case pa_stream_begin_write() returns smaller
memory buffer than was requested.
PulseAudio driver used to report position of last frame written to audio stream
instead of actual playback position. This kind of worked, since on average the
write position doesn't diverge too much from actual elaped time, but it can
result in note stuttering and in general isn't very accurate. Change the
implementation to use actual playback position from PulseAudio stream.
Remove some compatibity code with old PulseAudio, since version 0.9.10 was
released in 2008.
Also, as an optimization to avoid unnecessary memory copying, use
pa_stream_begin_write() to mix audio directly into PulseAudio memory.
- Remove checking for standard functions from the build system
- Prefix all invocations with std::
- Replace suffixed functions with unprefixed versions
- Include <cmath> in all files that use it and remove the global include
e.g. floorf(x) -> std::floor(x)
On linux / is mounted to /rootfs in RageFile, which allows access to the
_whole_ filesystem from lua. This means themes and mod files can
re-write user data and extract files via USB profiles.
/rootfs was only added for internal use by the alsa, oss and memory card
drivers, so it can be easily replaced with direct fstream file access.
This attempts to organize all needed items and places appropriate defines
in a single location. Redundant comparisons/defines were removed when noticed.
A few caveats, however:
* This may be better targeted towards the 5_1_0 branch instead of master right now.
* Cmake will run a little slower on configuration/generation now. That's due to the sanity checks it runs.
* There are some more checks to be added later, but this should be a solid start.
Thanks to the [libical](https://github.com/libical/libical) project for inspiration.
The Yosemite crash happens when we call CFRunLoopAddObserver. It doesn't seem to like runLoopRef. Replacing the first argument to this call with CFRunLoopGetCurrent() avoids the crash, but the app then hangs because CFRunLoopWakeUp doesn't cause observerRef (NameHALThread) to fire, so the semaphore is not posted. Commenting the semaphore wait line allows the game to boot in Yosemite, and seems to play totally normally. This doesn't really fix our problem (we are trying to name the HAL thread, and a catastrophic error occurs), but it shows us what the problem is.
We are obtaining runLoopRef using an API call that was deprecated years ago. However, the Apple-approved replacement code doesn't produce a runLoopRef that avoids the crash, either. So I'm not sure how we get the correct runLoopRef so we can name the thread.
This is my first time looking at the code base, but I don't see the value of naming the thread. In the interest of keeping Stepmania compatible with the upcoming operating system, I propose we just remove the thread-naming logic.
This is meant to be a safer alternative since
NULL can often be 0. Let's not rely on that.
And yes, I know this is a lot of files. This is
a safer thing to do in big commits vs for loops.
5f7001e: "Added a new branch"
01456ed: "Fixed a lot of memory leaks"
dac4493: "Fixed all remaining memory leaks that I could figure out"
0792db7: "Removed the smnew macro and the call to _CrtSetDbgFlag()"
Some of these caused destructor-time problems due to static initialization
order fiasco and related issues. Notably, the program would no longer exit on
OSX and had to be killed.
There were probably legitimate fixes in here, but since these are monolithic
commits it's too much work to extract them now. Let's reapply them
individually and in the forward direction.
sukibaby