Sam Tobin-Hochstadt
Sam Tobin-Hochstadt
Not much. On Tue, Jun 6, 2017, 5:32 PM Magnus Morton wrote: > @samth how much do you care about clean > history? > > — > You are receiving...
This isn't going to happen for pldi. It needs concurrency etc that will be hard to do.
I think there's significant racket code that uses iteration and other primitives on weak maps, probably including the expander.
One thing I immediately notice: we spend a bunch of time in assignment conversion and normalization. We should do that ahead of time instead of every time.
This now gets to a genuine pycket error in structs.
@vkirilichev can you open issues for the remaining things, and link them here?
I can think of a couple places this could go wrong. The `net/git-checkout` code, or the package archiving code.
Could we fix this by using the same syntax property approach that's used for []?
This is too hard to do for the deadline, postponing.
Should this be closed now?