the road to 2.3.0

[dankamongmen]

Remaining issues for 2.3.0, priorities, flow:

• ✅ visual render always fails after a visual resize #1559 handled except for bitmaps less than 6 rows tall
• ✅ draw sprixels bottom-to-top #1575 has to get fixed, and we need ensure it introduces no breakage -- i saw some
• ✅ don't allow arbitrary sprixel placement within a plane #1572 let's first decide policy here, and put it into place
  • ✅ sixel deepclean isn't entirely accurate #1580 fix up deepclean it's possibly causing yield demo's text pulses in sixel terminals #1557 (narrator: it was not causing yield demo's text pulses in sixel terminals #1557)
  • ✅ need restore bitmaps when no longer occluded #1440 is the last big bit of work, depends on don't allow arbitrary sprixel placement within a plane #1572
  • segfault when resizing with bitmaps present #1531 will be a full day's hard work, but is tractable
• (*bumped to 3.0) nctree needs to support collapse/expand internally #1366 is unrelated to bitmap stuff, and just needed for raptorial
• ✅ "view" demo's picture-in-a-picture is printed on top line rather than second #1577 has to get fixed, or else we need tear out the view PIP chrome, independent of others
• bitmap problems with large margins #1588 has to get fixed, jfc

✅ #1557 worries me. if it's not due to sixel editing slipups, i really have no idea why it's happening. we don't see this flicker anywhere but XTerm, yet i can't help but think it's our bug. i don't think it needs to be a blocker for 2.3.0, but goddamn, i hate it.

✅ we should look into #1538 and figure out what the story is there.

so i'm thinking we're at least a week out. i'd love to have #1559, #1575, and #1572 wrapped up by tomorrow (friday) evening. i don't want to burn my entire weekend on notcurses, but figure #1580 gets done sunday.

#1440 and #1531 will both be pains in the ass, figure two to six hours each.

[dankamongmen]

I think I knocked out #1559, which turned out to be a real issue, or at least related to a real issue. I'm pretty surprised that bug has been there for so long, but I guess you did need a resize + scaling, and why would you resize if you were then just going to scale anyway? Either way thanks @joseluis for flushing this out.

One bug has been found and fixed regarding #1580, and I'm hoping there's at least one more. I'm not calling this done until I have a rigorous unit test on it. How I hope it's somehow responsible for #1557, but that's looking less likely.

[dankamongmen]

working on #1440 now, expecting to have it done tomorrow. i think these remaining issues will get handled over the course of next week, and we're looking for a May Day 2.3.0 release.

[dankamongmen]

and...she's done (#1440 that is)

[dankamongmen]

2021-05-09: 2.3.0 has been released. Locking.