tragic. pygame.display.flip seems to always imply a vsync stall if you aren't using it to create an opengl context, and so solving the input events merging problem is going to be a pain in the ass. it is, however, an ass pain for another night: it is now time to "donkey kong"

EDIT: pygame.display.flip does not imply vsync! I just messed up my throttling logic 😁 huzzah!

> (which linux sees as a mouse)

Awful. What kind of laptop is that? Only seen that with shitty USB touchpads in lapdocks (...and Pinebooks), presumably done for compatibility with Android(?)

typedef struct SDL_MouseButtonEvent
{
...
Uint8 clicks; /**< 1 for single-click, 2 for double-click, etc. */
}

Did you correctly handle this?

someone brought up a really good point, which is that depending on how the touch screen works, it may be fundamentally ambiguous whether or not rapidly alternating adjacent taps can be interpreted as one touch event wiggling vs two separate touch events

easy to verify if that is the case, but damn, that might put the kibosh on one of my diabolical plans

Have you seen OSU?

osu.ppy.sh/

Just a demo of what some expect some touch screens to do.

with a lot of objects in memory, occasional hitches sounds suspiciously like gc generation 2. If you haven't already...

import gc
gc.set_debug(gc.DEBUG_STATS)

Usually I escalate to gc.callbacks, then discover the cycles are created by a (networking-related, in my cases) library.

I once got it so I could've turned off gc but the pace of allocations was then low enough that we chose to tolerate hitches in case of future cycles.

gc.freeze() seemed a possible next step but I didn't try.

the "proper" way of doing this would be to maintain a ringbuffer that's big enough to fetch as much history as you wish to support, then get the desired range from the rb, collect min/max amplitude for all samples per vertical line and render the bounds (effectively a very simple candle graph).

but i have done this before and it's not fun to write.

I've started having ngscopeclient filters output a quiet NaN when they have a scalar output and are requested to do something that doesn't make sense (missing inputs, contradictory configuration, divide by zero, etc).

(For waveform outputs we already had the ability to simply not output anything, but there's no concept of a null scalar in our object model)

the EMS VCS3 and Synthi called them three different things according to what they were: Sources, Treatments, or Outputs – and Sources only had output arrows, Treatments had inputs and outputs arrows, outputs only had input arrows

anaphonic.com/wp-content/uploads/ems_synthi-vcs3-a-dk1_om.pdf