Strobing on CLEAR using Tek4010 emulator / How to do Tektronix animation properly?

[hornc]

Using https://github.com/rricharz/Tek4010 static graphics work perfectly:

tek4010 -noexit ./ct.py <(./CTBASIC.py examples/teapot.bas)

but the animations cause the screen to strobe white on CLEAR, which is very jarring.

tek4010 ./ct.py --hold 130 <(./CTBASIC.py examples/CUBE.BAS)

and

tek4010 ./ct.py --hold 130 <(./CTBASIC.py examples/ANIMTEST003.bas)

It's not because of the extra ANSI / RIS code I added (surprisingly), but it appears that the Tektronix ESC + FF causes the screen flash on reset, and that is the authentic behaviour -- which does not occur under xterm.

It looks like using ESC+FF between frames in not the standard way to run animations on a Tektronix display.

TODO:

• Investigate https://github.com/rricharz/Tek4010/blob/master/animation.plt (an original clean running (no flicker / strobing) animation and see how that works.

Is is using "Write-Thru" mode, or specific erasing? I couldn't find clear descriptions of these features in the manual.

[hornc]

I last tested using Tek4010 SHA: commit 9c3cd7be921e3c27de8420dc807f7097b2561078

[hornc]

PR #35 demonstrates write-thru mode animation and display storage which works on the Tek4010 emulator: https://github.com/rricharz/Tek4010

Write-thru (normal lines) is set by sending ESC + p. Other (dashed/dotted) line modes can be set with other charcter combinations, see Table F-3, Appendix F of https://vt100.net/tektronix/4014-um.pdf

Need to think on how best to manage timing...
ZX BASIC has PAUSE n, where n is in TV frames (50 or 60 Hz). This can't be translated reliably to BCT, other than padding with junk data to take time, but would depend on processing frequency.

Perhaps add an argument to set the baud rate / frequency / frame rate of ct.py output?

• Frame rate might be the most natural setting: time taken for each CT program loop, we can force it to be at least t duration.
• baud rate would set the speed of transmitted output characters... from a buffer

Current --hold seems to hang a program without any CLS. Seems like a bug.