This one is definitely fun to reproduce, it's sporadic as hell, and it would often work on 2nd try for me anyway. However, I've put in two fixes now that hopefully address this. One, the font resource was purgable, so it could get purged from memory, and on real hardware with more memory pressure, this was more likely to happen than on QEMU. Second, I picked a randomized higher font ID, to ensure there's less likelihood of a collision
You aren't kidding, that type of information is like a lost art anymore. I was originally thinking maybe the character set commands were getting corrupted, like, split up or cut-off/offset midstream type thing, or even some kind of network corruption issue.
However, while testing on the quadra 700, the issue was happening way more consistently, but almost always on the first connection attempt. The fact it wasn't just consistently broke (at least for me), and seemed to "fix" itself, lead me down a slightly different avenue, possibly to do with the glyph font not being consistently loaded in memory, and/or getting purged from memory, and on an older machine like the quadra @ 25mhz, it can't keep up on that first load. I have put some fixes in place, but still testing. My last fix was definitely not 100%, but seemed to reduce it somewhat. I suspect there's still something else going on.
Also been doing a whole bunch of screen-refresh optimizations, using gworld to reduce flicker, added a proper scrollbar as well. I'm definitely down another rabbithole, and there's so many pending features now, I might bump the next version to 1.2.0. We'll see!
