I just noticed the Random stdlib has an UnsafeView type for filling arrays. With the new layout optimizations, this should no longer be necessary. We should try using the normal SubArray type for this. If it's still too slow/complex, then at least the type can be made safe by holding the underlying array directly instead of a Ptr.

It would be nice to add graphviz-like functionality natively to Penrose. (Before starting this, talk with @hypotext and @strout18 as there is some existing work in the system)

➜ corral --help=false
ERRR: Error: missing subcommand at: 'corral'

and yet...

Options:
   -g, --debug=0             Configure debug output: 0=off, 1=err, 2=warn, 3=info, 4=fine.
   -n, --nothing=false       Don't actually apply changes.
   -h, --help=false

according to the options, that should work.

Options shouldn't have "=false" as the default displayed really.

Current behavior:

.> view ##Nat

  ⚠️
  
  The following names were not found in the codebase. Check your spelling.
    ##Nat

.> names ##Nat

  😶
  
  I couldn't find anything by that name.

.> alias.type ##Nat Nat

  Done.

.> names ##Nat

  Type
  Hash:  ##Nat
  Names: Nat

.> view ##Nat

  -- Nat is built-in.

Desired behavior would be something like:

Motivation

In supercollider/supercollider#4572 we discussed the need for better documentation on when collection functions test on equality vs identity. For instance, SequenceableCollection:indexOf tests on identity, but you wouldn't know it f

Really nice project.

Writing in coconut files, you lose all the IDE integrations for Python. For example, I use VSCode and all the intellicode and hints of course doesn't work for cocount.

I am not too familiar but perhaps theres needs to be a separate language server or maybe you could leverage the existing Python integrations. Either way, better IDE integration wo