Fixed bug 3662 - Error message when using the audio conversion setup without an initialized audio subsystem is a bit vague

Simon Hug

This issue actually raises the question if this API change (requirement of initialized audio subsystem) is breaking backwards compatibility. I don't see the documentation saying it is needed in 2.0.5.

src/audio/SDL_audio.c

@@ -874,8 +874,6 @@ SDL_GetAudioDriver(int index)
     return NULL;
 }
 
-extern void SDL_ChooseAudioConverters(void);
-
 int
 SDL_AudioInit(const char *driver_name)
 {
@@ -890,8 +888,6 @@ SDL_AudioInit(const char *driver_name)
     SDL_zero(current_audio);
     SDL_zero(open_devices);
 
-    SDL_ChooseAudioConverters();
-
     /* Select the proper audio driver */
     if (driver_name == NULL) {
         driver_name = SDL_getenv("SDL_AUDIODRIVER");

src/audio/SDL_audio_c.h

@@ -54,7 +54,10 @@ extern SDL_AudioFormat SDL_NextAudioFormat(void);
 /* Function to calculate the size and silence for a SDL_AudioSpec */
 extern void SDL_CalculateAudioSpec(SDL_AudioSpec * spec);
 
-/* These pointers get set during init to various SIMD implementations. */
+/* Choose the audio filter functions below */
+extern void SDL_ChooseAudioConverters(void);
+
+/* These pointers get set during SDL_ChooseAudioConverters() to various SIMD implementations. */
 extern SDL_AudioFilter SDL_Convert_S8_to_F32;
 extern SDL_AudioFilter SDL_Convert_U8_to_F32;
 extern SDL_AudioFilter SDL_Convert_S16_to_F32;

src/audio/SDL_audiocvt.c

@@ -895,11 +895,6 @@ SDL_BuildAudioCVT(SDL_AudioCVT * cvt,
         return SDL_InvalidParamError("cvt");
     }
 
-    /* Conversions from and to float require the audio subsystem to be initialized */
-    if (!SDL_WasInit(SDL_INIT_AUDIO)) {
-        return SDL_SetError("Audio subsystem has not been initialized");
-    }
-
     /* Make sure we zero out the audio conversion before error checking */
     SDL_zerop(cvt);
 
@@ -932,6 +927,9 @@ SDL_BuildAudioCVT(SDL_AudioCVT * cvt,
     cvt->len_ratio = 1.0;
     cvt->rate_incr = ((double) dst_rate) / ((double) src_rate);
 
+    /* Make sure we've chosen audio conversion functions (MMX, scalar, etc.) */
+    SDL_ChooseAudioConverters();
+
     /* SDL now favors float32 as its preferred internal format, and considers
        everything else to be a degenerate case that we might have to make
        multiple passes over the data to convert to and from float32 as