Robby Findler

Yes that could be expensive. In practice for systems that don't have a lot of different ways to build the intermediate proofs it will be as efficient as redex can...

I looked into this one specifically, under monterey. It looks like [these calls to the os](https://github.com/racket/gui/blob/master/gui-lib/mred/private/wx/cocoa/procs.rkt#L182-L195) always return the same color (until the app is restarted). If the call changed...

Oh-- an it is specifically the results of `get-label-background-color` and `get-label-foreground-color` that do not change but I would have expected them to change.

Here's the docs on the [`backgroundColor`](https://developer.apple.com/documentation/appkit/nswindow/1419751-backgroundcolor) method that's being called and it makes me wonder if `NSColor` somehow has to be more dynamic?

Oops actually [this is the method being called](https://developer.apple.com/documentation/appkit/nscolor/1528630-windowbackgroundcolor).

Thanks, @david-christiansen . (The errortrace thing is probably because redex's implementation isn't being instrumented by errortrace.) @paulstansifer do you have time to look into this one?

This looks like one of those situations where it may turn out to be many separate things that need improving. It looks like one of them is saving the file....

This commit is one piece of the puzzle (meant to note the connection in the commit message but forgot): https://github.com/racket/gui/commit/e947d596bb1eefcc660740a1ba0146188e0e1fc4

With the latest set of commits, the time it takes to call `save-file` in the code above goes from 12 seconds to 20ish msec. This matches the time to save...

This seems like a good idea. Three concerns: 1) what is the new minimum width of the preferences dialog? (It has to be kept narrow) 2) user's old preferences will...