Daniel Kochmanski

For the reference, this came up when I've tried to improve the invalid interpretation - I'm attaching patches for curious readers. ```diff From c1ed00cf22746811eb2b5b0427caba61fa27a653 Mon Sep 17 00:00:00 2001 From:...

regarding width, when we disable the cache the width is still 640 even when the viewport is much bigger and the cache is disabled - this should be the case...

Incidentally this demo reveals another bug, where line-height is not reset before display (so the cursor position after the initial click is always on the baseline of the first call...

Issue is impossible to solve completely because of asynchronous nature of events. window-repaint-event may come at any time. solution I've taken: displaying the function first records the output, then dispatches...

Thanks for the report. Input editing needs some work (both the buffer abstraction and the command reader).

n.b we should probably introduce an operator `(setf %port)` instead of `(setf port)` unless we trust our users :_)

accept-from-string calls underneath read-token (maybe it shouldn't?) -- for example, when the type is real, then: presentation method accept calls: ``` accept-using-read read-token read-from-string * ``` read-from-string returns the position....

This pr will also fix "flashing" in menu bar, as reported in #1171

I've noticed a regression: when I have a slider and drag it, and there is another mirror along the way, then dragging stops. I think that's one of a few...

also notice, that a few operators allow specifying multiple-window as an argument; I think that we should get rid of the "pressed-sheet" copy distribution but that needs to be replaced...