CA5: Interrupting ASWA correction results in errors

[rualark]

Also, algorithm does not obey max_correct_ms

23:23:26 Algorithm is running in low performance mode. To increase performance, reset calculate_correlation, calculate_blocking, calculate_stat, calculate_ssf, best_rejected (check config)
23:23:28 Attemp to use persistent data element '13' while it is not ready yet
23:23:30 - Flag does not appear on second full evaluation: [225] + Slurs (suspension) at 1:1 (beat 3:5) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [220] - Suspension resolution (to discord) at 1:1 (beat 3:5) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [268] + First harmony (CC) at 1:1 (beat 3:5) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [1] + Uncompensated leap (start 3rd) at 1:1 (beat 3:5) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [241] + Suspension (1/2 + 1/4) at 1:5 (beat 4:1) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [187] - Dissonance upbeat (leap) at 1:7 (beat 4:3) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [112] + Uncompensated: precompensated (3rd) at 1:9 (beat 4:5) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [108] + Close + no MDC (3rd) at 1:9 (beat 4:5) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [248] + Perfect consonances < 4/4 apart (upbeat contrary) at 1:11 (beat 4:7) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [225] + Slurs (suspension) at 1:17 (beat 5:5) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [245] + Many notes in measure (5) at 1:21 (beat 6:1) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [241] + Suspension (1/2 + 1/4) at 1:21 (beat 6:1) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [227] + 1/8 note (beat 2) at 1:23 (beat 6:3) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [249] + Perfect consonances < 4/4 apart (non-harmonic) at 1:24 (beat 6:4) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [42] + Compensation deviation to 2nd (3rd) at 1:25 (beat 6:5) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [59] + Far + close MDC (3rd) at 1:25 (beat 6:5) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [169] + Dissonance upbeat (stepwize) at 1:33 (beat 7:5) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [250] - Perfect consonances < 4/4 apart (other) at 1:35 (beat 7:7) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [32] + Tritone (on climax) at 1:37 (beat 8:1) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [31] - Tritone (unresolved) at 1:37 (beat 8:1) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [133] - No + close MDC (4th) at 1:37 (beat 8:1) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [225] + Slurs (suspension) at 1:41 (beat 8:5) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [242] + Suspension (1/2 + 1/2) at 1:45 (beat 9:1) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [225] + Slurs (suspension) at 1:49 (beat 9:5) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [241] + Suspension (1/2 + 1/4) at 1:53 (beat 10:1) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [227] + 1/8 note (beat 2) at 1:55 (beat 10:3) midi\good-cp5.mid
23:23:30 - Flag does not appear on second full evaluation: [238] + 1/2 after 1/4 or 1/8 in measure (penultimate) at 1:57 (beat 10:5) midi\good-cp5.mid
23:23:31 SendMIDI got buf underrun in 704 ms (steps 16 - 96)

[rualark]

This bug is only relevant for calculate_blocking=1.
Found out that calculate_blocking decreases SAS emulator speed greatly (probably 20 times). Cause of this is probably creating fblock for each SAS emulator run. This should be fixed in #313
Bug happens only when Generation is stopped while SAS emulator is run.

[rualark]

Currently fixed by two commits (last attached)