It doesn’t seem right to me that 5000 files should cause a problem. 50,000 maybe, but Picard ought not to lock up at just 5000. A ticket that says
Fairly often we get messages from users like “I loaded 50k files on Picard and it’s going very slowly”. We should have some sort of alert for these when someone tries to load more than 5k or 10k files in one go
seems pretty vague to me and not addressing the underlying problem. The cause of the lock-ups/hangs has not been explained. Frequently, it is not a crash, but just a (longish) temporary non-responsiveness, which may be reduced by increasing the priority of the process.
Picard is a great program (and free) but the random nature of this problem is its most irritating feature and puts off a good many new users who might otherwise become useful contributors to the MusicBrainz community.
That said, one should ask why load even 5,000 tracks in one go? If they are only scantily tagged and the hope is that Picard will fix them, then that is a bit optimistic. Even assuming they are all in the MusicBrainz database, some human review will be required afterwards, so some smaller chunking would make the task easier. If they already have MBIDs and it is just a retagging exercise, however, then Picard ought not to complain.
I suspect that this is not a simple fix and am reluctant to raise a new ticket given the current backlog. Also, I am aware that v2.0 is still in development - might it be possible that this will operate a little more robustly?