No clear plan, just a sense of what I'd like the experience to be. It feels so close, but maybe there's deeper limitations I'm unaware of. Do you tend to hack on shadow standalone, or in active use for developing something else? By "library code", I mean having shadow's build/repl services running and being able to hack on them without restarting the whole command line process to integrate changes.
@robert.mather.rmm I work on shadow-cljs standalone. most of the time using the :browser
, :script
or :library
builds. sometimes just a browser-repl or node-repl. started like https://github.com/thheller/shadow-cljs/blob/master/CONTRIBUTING.md#starting-a-repl
there is also a bunch of code in src/repl
that may look like tests but arent. each deftest
is just an isolated snippet of code I can run via keypress so experiment with specific things
if you have suggestions for the REPL stuff I'm happy to work through them with you
it is all a tad more complicated than I prefer right now since I rewrote the REPL from the ground up for shadow.remote
https://github.com/thheller/shadow-cljs/blob/master/doc/remote.md
combining shadow.remote with nrepl has been challenging since it is meant to replace it entirely
but I wrote that stuff mostly for the UI. maybe I'll work on the REPL interface there for a bit today
just to show how its meant to work