The patch 1) works great for me.

Can this be merged? This is needed for v0.10 to continue the user acceptance test.
Let's merge another PR to devel if this is not needed (case 2).

If it's needed for user acceptance testing, I think we should merge it as you suggest.

I think @jsorg71 was investigating an anomaly that this shouldn't be necessary; but if it fixes a user facing issue then we should still merge it. No objections here.

@matt335672 I think we still need this fix. I just want to figure out why it crashed instead of just returning a few less tiles.
I did try to reproduce this crash and I can not. I don't have a 4k screen but I set the max buffer to 1000 and I got one tile back which is expected. Anyway lets not hold this up.

Thanks, merging and will cherry-pick to v0.10.

Agreed that this needs more investigation. I've re-opened the original fault report which was closed automatically.

@jsorg71 - I don't have a 4k monitor either, but I managed to frig something with the console on a windows VM which was bigger than my virtual monitor. Also, I think that you can probably just edit a mstsc ini file which isn't full sceen, and just put the larger dimensions in. I don't have access to an environment to check this today though.

Thanks Matt, I can repro and I found the issue. I'll test a couple of days, then do a PR.

That's great news @jsorg71.

Can you undo 90e4aca as part of the PR? I was never really happy with it as anything other than a short-term fix.