if (i == frames_read) {
- to_read = min (chunksize, (_length - current_frame));
+ if (current_frame >= _length) {
- if (to_read == 0) {
- /* XXX ARGH .. out by one error ... need to figure out why this happens
- and fix it rather than do this band-aid move.
+ /* hmm, error condition - we've reached the end of the file
+ without generating all the peak data. cook up a zero-filled
+ data buffer and then use it. this is simpler than
+ adjusting zero_fill and npeaks and then breaking out of
+ this loop early
*/
- zero_fill = npeaks - nvisual_peaks;
- break;
- }
- if ((frames_read = read_unlocked (raw_staging, current_frame, to_read)) == 0) {
- error << string_compose(_("AudioSource[%1]: peak read - cannot read %2 samples at offset %3 of %4 (%5)"),
- _name, to_read, current_frame, _length, strerror (errno))
- << endmsg;
- goto out;
+ memset (raw_staging, 0, sizeof (Sample) * chunksize);
+
+ } else {
+
+ to_read = min (chunksize, (_length - current_frame));
+
+
+ if ((frames_read = read_unlocked (raw_staging, current_frame, to_read)) == 0) {
+ error << string_compose(_("AudioSource[%1]: peak read - cannot read %2 samples at offset %3 of %4 (%5)"),
+ _name, to_read, current_frame, _length, strerror (errno))
+ << endmsg;
+ goto out;
+ }
}
i = 0;