Open dannypsnl opened 2 months ago
Now I'm thinking, if interval-map
cannot cross the boundary, it seems forward every message to place is more reasonable now……Though that is a big refactoring.
Although interval-map
cannot cross the boundary, the content of it can, so if each updation sent back to record-maintainer, will be a relatively simple solution.
Is your feature request related to a problem? Please describe. The thread by default created with current-thread-custodian, which will slow down the whole DrRacket, particular if I'm working on some large project (not really large in normal scale, but racket level large).
The fundamental problem is https://github.com/dannypsnl/sauron/blob/fb2eaa94ca305371088f93cb8b9828b690de5f86/collect/record-maintainer.rkt#L79, this
collect-from
is computation intensive operation, move it into a place worker seems reasonable.Describe the solution you'd like
Describe alternatives you've considered No idea
Additional context
I found
(place-message-allowed? (make-record))
is#f
, the reason are fields that involve(make-interval-map)
.