Don't know if this is still under development but there is a bug regarding being crouched while pressing the reset button (i.e. "tas frame goto"). If you are crouched while being in the brief replay state afterwards , the player position is lower than it should be. So every time you press the reset button, you have to make sure you are uncrouched until you can move again.
Reproducable with this recording:
https://drive.google.com/file/d/1eS1er9QA5mVxDAfg4p1mqnJfFVS8oRmm/view?usp=sharing
Hmm I haven't actually worked on the plugin in a very long time, but based on some of my old code that is now commented out I feel like this should be handled:
/* No real need for this any more with the ghost tick thing, it accounts for it
if (ucmd->buttons & IN_DUCK && !(recording.getCommand(teleportToFrame - ghostTicks).buttons & IN_DUCK)) {
Msg("[TAS] Crouch Mismatch\n");
}
if ( !(ucmd->buttons & IN_DUCK ) && recording.getCommand(teleportToFrame - ghostTicks).buttons & IN_DUCK) {
Msg("[TAS] Crouch Mismatch\n");
}
*/
Basically, for the period of the "ghost ticks" (the time before it starts recording again while it has control) it should be completely overriding player input and teleporting you every tick, so even if there was initially a mismatch I feel like it should smooth out (and apparently that's what I thought in the past too).
Is this something you can consistently reproduce? As in it's not some funky edge case right at the bounds of the ghost tick window? I haven't gotten a chance to look at the recording yet - sorry if it's clear from looking at it. Any extra info you have would be helpful until I have a chance to investigate more.
Also, I was actually working on getting a working dev environment on my current computer a few weeks ago, but worst case I still have my other computer lying around somewhere that has it working.