- 26 Oct, 2016 8 commits
-
-
David Reid authored
Rationale: 1) It allows the callbacks to be set at initialization time which feels a bit more intuitive to me. 2) It avoids the need to call mal_device_set_send_callback(), etc. 3) It's a bit more consistent with the onLog callback. Previously, onLog would be passed to mal_device_init(), whereas onSend, etc were set with mal_device_set_send_callback(), etc. which feels needlessly inconsistent.
-
David Reid authored
-
David Reid authored
Rationale: 1) These formats just aren't common enough to justify the added maintenance cost. 2) It makes software format conversion a lot simpler since it reduces the number of permutations.
-
David Reid authored
-
David Reid authored
-
David Reid authored
Previously, there was only a single default buffer size and period count. This forces the library to use a lowest common denominator for every backend which forces higher quality backends to suffer because of the lesser quality ones. This change adds basic infrastructure for future work to tweak these properties for each individual backend to try and get the best out them.
-
David Reid authored
Rationale: 1) The number of parameters is just getting too much. 2) It makes it a bit easier to add new configuration properties in the future. In particular, there's a chance there will be support added for backend-specific properties.
-
David Reid authored
The rationale for this change is to ensure the user data is available for the logging callback during initialization.
-
- 25 Oct, 2016 6 commits
-
-
David Reid authored
-
David Reid authored
-
David Reid authored
-
David Reid authored
-
David Reid authored
-
David Reid authored
-
- 23 Oct, 2016 1 commit
-
-
David Reid authored
-
- 22 Oct, 2016 2 commits
-
-
David Reid authored
-
David Reid authored
-
- 21 Oct, 2016 13 commits
-
-
David Reid authored
-
David Reid authored
-
David Reid authored
-
David Reid authored
-
David Reid authored
-
David Reid authored
-
David Reid authored
Add onStop callback which is called when a device is stopped either explicitly or internally as a result of an error.
-
David Reid authored
-
David Reid authored
-
David Reid authored
-
David Reid authored
The rationale for this change is to keep it consistent with the rest of the library.
-
David Reid authored
-
David Reid authored
-
- 20 Oct, 2016 6 commits
-
-
David Reid authored
-
David Reid authored
-
David Reid authored
-
David Reid authored
-
David Reid authored
Rename mal_device__read_samples_from_client() to mal_device__read_frames_from_client() and mal_device__send_samples_to_client() to mal_device__send_frames_to_client(). Rationale for this is to keep it consistent with the rest of the library.
-
David Reid authored
-
- 19 Oct, 2016 4 commits
-
-
David Reid authored
-
David Reid authored
The rationale for this change is that during testing I noticed that snd_pcm_open() would sometimes fail with "Invalid value for card" when using the "NAME" hint returned by snd_device_name_get_hint() verbatim. Using "hw:%d,%d" seems to fix this problem.
-
David Reid authored
-
David Reid authored
-