Another run at the problem 57516164de was meant to solve. The
new solution just gives up when a `drawRect:' method is called
in the wrong thread, which might create refresh glitches but
avoids a deadlock in the case that the Racket thread is blocked
on the update happening in the foreign thread.
On Cocoa, a view's `drawRect:' method can be called from a
heartbeat thread that animates controls. Such a call happens
rarely for a `canvas%' or other class where `drawRect:'
is overridden, but since it can happen, ensure that the
callback runs on the Racket thread.
The relevant `on-paint' call is one that is forced for a GL
canvas to try to draw a frame's content before the frame is
shown. Sometimes, thread scheduling would let the frame get
sufficiently initialized at the Cocoa level before the `on-paint'
call happened, but sometimes not.
"My" -> "Racket". The "My" prefix was from initial experiments,
of course, and I just never got around to changing it before.
I think these names go into a global namespace, though, at the
ObjC level, so they need to have a distinct and Racket-specific
prefix.
Also, use keywords for `make-pen' and `make-brush'.
Adding `make-pen' and `make-color' creates many conflicts among
teaching libraries, such as `2htdp/image'. These are easy to fix
up in the tree, but adding such obvious names to `racket/draw'
may create other compatibility problems, so we might have to reconsider
the names.
In consultation with Asumu.
- call `on-size' on setting the canvas size, like Win32 and Cocoa,
which mainly shows/hides auto-shown scrollbars more on time
(e.g., DrRacket doesn't start with horizontal scrollbars that
immediately disappear)
- switch mouse tracking to "hint" mode, instead of catching every
single mouse position
- when queueing `on-size' methos in response to client-size changes,
cancel and queued but undelivered calls; in the case of editor
canvases, this doesn't really have an effect, due to the other
`on-size' call
Closes PR 11369, hopefully
