symbolic voice-types versus synthesis voices

[Andrei Kholodnyi]

> The only remaining question is compatibility. ?Can we safely remove the 
> current symbolic
> voices and support native voices only? ?I guess not. ?So we need some ideas 
> how to
> maintain compatibility with current clients if we want to make such a change.

i'd rather claim following API as obsolete:

int spd_set_voice_type(SPDConnection*, SPDVoiceType type);
int spd_set_voice_type_all(SPDConnection*, SPDVoiceType type);
int spd_set_voice_type_uid(SPDConnection*, SPDVoiceType type, unsigned int uid);
char** spd_list_voices(SPDConnection *connection);

and extend SPDVoice as

typedef struct {
    char *name;      /* Name of the voice (id) */
    char *language;  /* 2-letter ISO language code */
    char *variant;   /* bcp47 language tags */
    SPDVoiceGender gender;
    unsigned int age;
} SPDVoice;