$ diskstack tests/fixtures/capture_a.scp tests/fixtures/capture_b.scp -o merged.img
Reading capture_a.scp
Reading capture_b.scp
Format: ibm.360 (capture_a.scp (IBM 40 cyls, 2 heads, 9 sectors of 512 bytes))

720 sectors: 714 clean (from 2 files), 6 unresolved

Sources
  file           type  revs  attempts  good  supplied  sole
  capture_a.scp  scp      3      2058  2013       694    20
  capture_b.scp  scp      3      2061  2017       694    20

Tracks needing attention
  cyl  head  clean  voted  unresolved  missing  sector ids still bad
   17     0      7      0           2        0  4,7
   18     0      7      0           2        0  4,7
   19     0      7      0           2        0  4,7

Re-read just these tracks, then run diskstack again with the new capture added:
  gw read retry.scp --tracks=c=17-19:h=0

Wrote merged.img (368640 bytes) and diskstack-report.json
$ echo $?
2

$ gw read capture_c.scp --tracks=c=17-19:h=0        # on the real drive

$ diskstack tests/fixtures/capture_a.scp tests/fixtures/capture_b.scp tests/fixtures/capture_c.scp -o merged.img
Reading capture_a.scp
Reading capture_b.scp
Reading capture_c.scp
Format: ibm.360 (capture_a.scp (IBM 40 cyls, 2 heads, 9 sectors of 512 bytes))

720 sectors: 714 clean (from 3 files), 6 recovered by vote
Since the last report: 6 recovered.

Sources
  file           type  revs  attempts  good  supplied  sole
  capture_a.scp  scp      3      2058  2013       700     0
  capture_b.scp  scp      3      2061  2017       700     0
  capture_c.scp  scp      3      2060  2017       700     0

2 of 3 inputs came out of the cache instead of being decoded again. --no-cache turns that off.

Every sector confirmed by CRC. Nothing left to re-read.

Wrote merged.img (368640 bytes) and diskstack-report.json
$ echo $?
0
