mirror of
https://github.com/nzp-team/fteqw.git
synced 2024-11-29 23:22:01 +00:00
0c8ad17f7c
Added sv_guidkey cvar, allowing cross-server guid key generation (although it lacks auth). Support .ico, because we can. preliminary support for sdl 2.0.6's vulkan stuff. will wait till its actually released before its properly used. Fix capturedemo. videomap should typically use premultiplied alpha, apparently. Updated sound drivers. No more old drivers. Better cvar registration. More drivers optionally support float output. Added certificate log for dtls connections. Rewrote font char cache, now supports full unicode char range, not just ucs-2. Attempt to support FreeType 2.5+ rgba fonts. XMPP now supports carbons, and shows avatars in conversations. Updated xmpp's scram auth to be more strict, including the plus variation (hopefully), to block evil tls proxies. ffmpeg plugin now uses the decoupled api for decoding too. Cef plugin updated to support fte-scheme post data properly, as well as request/response headers (like cross-origin). git-svn-id: https://svn.code.sf.net/p/fteqw/code/trunk@5148 fc73d0e0-1445-4013-8a0c-d673dee63da5
154 lines
4.1 KiB
C
154 lines
4.1 KiB
C
/*
|
|
this file is basically a copy of the SDL one
|
|
java code has a function or two which just periodically calls us to ask us to dump out audio for it
|
|
*/
|
|
#include "quakedef.h"
|
|
#include <jni.h>
|
|
#include <pthread.h>
|
|
|
|
static soundcardinfo_t *sys_sc = NULL;
|
|
extern int sys_soundflags;
|
|
|
|
//called by the java code when it wants to know what sort of AudioTrack format to use.
|
|
JNIEXPORT jint JNICALL Java_com_fteqw_FTEDroidEngine_audioinfo(JNIEnv *env, jclass this, jint arg)
|
|
{
|
|
soundcardinfo_t *sc = sys_sc;
|
|
if (!sc)
|
|
return 0;
|
|
|
|
switch(arg)
|
|
{
|
|
case 1:
|
|
return sc->sn.numchannels;
|
|
case 2:
|
|
return sc->sn.samplebytes*8;
|
|
default:
|
|
return sc->sn.speed;
|
|
}
|
|
}
|
|
|
|
extern int S_GetMixerTime(soundcardinfo_t *sc);
|
|
//transfer the 'dma' buffer into the buffer it requests, called from a dedicated sound thread created by the java code.
|
|
JNIEXPORT jint JNICALL Java_com_fteqw_FTEDroidEngine_paintaudio(JNIEnv *env, jclass this, jbyteArray stream, jint len)
|
|
{
|
|
int offset = 0;
|
|
soundcardinfo_t *sc = sys_sc;
|
|
int framesz;
|
|
|
|
if (sc)
|
|
{
|
|
int buffersize = sc->sn.samples*sc->sn.samplebytes;
|
|
|
|
int curtime = S_GetMixerTime(sc);
|
|
framesz = sc->sn.numchannels * sc->sn.samplebytes;
|
|
|
|
S_PaintChannels (sc, curtime + (len / framesz));
|
|
|
|
if (len > buffersize)
|
|
{
|
|
len = buffersize; //whoa nellie!
|
|
}
|
|
|
|
if (len + sc->snd_sent%buffersize > buffersize)
|
|
{ //buffer will wrap, fill in the rest
|
|
(*env)->SetByteArrayRegion(env, stream, offset, buffersize - (sc->snd_sent%buffersize), (char*)sc->sn.buffer + (sc->snd_sent%buffersize));
|
|
offset += buffersize - (sc->snd_sent%buffersize);
|
|
sc->snd_sent += buffersize - (sc->snd_sent%buffersize);
|
|
len -= buffersize - (sc->snd_sent%buffersize);
|
|
if (len < 0) /*this must be impossible, surely?*/
|
|
len = 0;
|
|
}
|
|
//and finish from the start
|
|
(*env)->SetByteArrayRegion(env, stream, offset, len, (char*)sc->sn.buffer + (sc->snd_sent%buffersize));
|
|
offset += len;
|
|
sc->snd_sent += len;
|
|
}
|
|
else
|
|
offset = len; /*so the playback thread ends up blocked properly*/
|
|
return offset;
|
|
}
|
|
|
|
|
|
static void Droid_Shutdown(soundcardinfo_t *sc)
|
|
{
|
|
//fixme: what if we're currently inside Java_com_fteqw_FTEDroidEngine_paintaudio?
|
|
sys_sc = NULL;
|
|
free(sc->sn.buffer);
|
|
sys_soundflags = 0;
|
|
}
|
|
|
|
//return the number of samples that have already been submitted to the device.
|
|
static unsigned int Droid_GetDMAPos(soundcardinfo_t *sc)
|
|
{
|
|
sc->sn.samplepos = sc->snd_sent / sc->sn.samplebytes;
|
|
return sc->sn.samplepos;
|
|
}
|
|
|
|
static void Droid_UnlockBuffer(soundcardinfo_t *sc, void *buffer)
|
|
{
|
|
}
|
|
|
|
static void *Droid_LockBuffer(soundcardinfo_t *sc, unsigned int *sampidx)
|
|
{
|
|
return sc->sn.buffer;
|
|
}
|
|
|
|
/*
|
|
static void Droid_SetEnvironmentReverb(soundcardinfo_t *sc, qboolean uw)
|
|
{
|
|
}
|
|
*/
|
|
|
|
static void Droid_Submit(soundcardinfo_t *sc, int start, int end)
|
|
{
|
|
}
|
|
|
|
//on android, 16bit audio is 'guarenteed'.
|
|
//8bit is not guarenteed.
|
|
//there's no reference to sample rates. I assume 44.1khz will always work, though we want to avoid that cpu+mem load if we can
|
|
//nor any guarentee about channels supported. I assume mono will always work.
|
|
static qboolean Droid_InitCard (soundcardinfo_t *sc, const char *cardname)
|
|
{
|
|
if (sys_sc)
|
|
return false; //can only cope with one device.
|
|
if (cardname && *cardname)
|
|
return false; //only the default device
|
|
|
|
sc->selfpainting = true;
|
|
// sc->sn.speed = 11025;
|
|
// sc->sn.samplebytes = 2;
|
|
// sc->sn.numchannels = 1;
|
|
|
|
if (sc->sn.samplebytes == 1)
|
|
sc->sn.sampleformat = QSF_U8;
|
|
else /*if (sc->sn.samplebytes == 2)*/
|
|
{
|
|
sc->sn.samplebytes = 2;
|
|
sc->sn.sampleformat = QSF_S16;
|
|
}
|
|
|
|
/*internal buffer should have 1 sec audio*/
|
|
sc->sn.samples = sc->sn.speed*sc->sn.numchannels;
|
|
|
|
sc->Lock = Droid_LockBuffer;
|
|
sc->Unlock = Droid_UnlockBuffer;
|
|
// sc->SetEnvironmentReverb = Droid_SetEnvironmentReverb;
|
|
sc->Submit = Droid_Submit;
|
|
sc->Shutdown = Droid_Shutdown;
|
|
sc->GetDMAPos = Droid_GetDMAPos;
|
|
|
|
sc->sn.buffer = malloc(sc->sn.samples*sc->sn.samplebytes);
|
|
|
|
sys_sc = sc;
|
|
|
|
sys_soundflags = 3;
|
|
|
|
return 1;
|
|
}
|
|
|
|
sounddriver_t Droid_AudioOutput =
|
|
{
|
|
"Android",
|
|
Droid_InitCard,
|
|
NULL
|
|
};
|