Refs & state #24

New Issue

scott · 2023-12-27T00:40:03Z

scott commented

2023-12-27 00:40:03 +00:00

The object that run returns should not be a bespoke object that's managed and exported by the postlude, but actually a dump of all the refs in the executed Ludus code. Easy peasy: the result of a script is not only the result, but also the state of its stateful parts upon exiting.

The object that `run` returns should not be a bespoke object that's managed and exported by the postlude, but actually a dump of all the refs in the executed Ludus code. Easy peasy: the result of a script is not only the result, but also the state of its stateful parts upon exiting.

scott added the

labels 2023-12-27 00:40:33 +00:00

scott self-assigned this 2023-12-27 04:18:16 +00:00

scott added this to the Computer Class milestone 2024-04-26 21:14:52 +00:00

scott commented

2024-05-18 22:29:18 +00:00

One design issue that's not so easy-peasy: refs would now need globally unique names.

That's probably good enough for Computer Class, which probably won't use refs except under the hood.

Possible solutions include:

don't export refs in this way, but only specific, hard-coded refs,
manually export refs with some magic function,
give each ref a unique ID that isn't a lexical name

(In the latter case, naming conflicts would of course be checked.)

I'm actually leaning towards one of the first two.

This is mostly an issue for #31 and #23. Maybe work locally with those two, and then make a decision here.

One design issue that's not so easy-peasy: refs would now need globally unique names. That's probably good enough for Computer Class, which probably won't use refs except under the hood. Possible solutions include: 1. don't export refs in this way, but only specific, hard-coded refs, 1. manually export refs with some magic function, 1. give each ref a unique ID that isn't a lexical name (In the latter case, naming conflicts would of course be checked.) I'm actually leaning towards one of the first two. This is mostly an issue for #31 and #23. Maybe work locally with those two, and then make a decision here.

scott added

research

and removed

accepted

labels 2024-05-19 19:25:49 +00:00

scott modified the milestone from Computer Class to 0.2.0

2024-05-29 19:09:52 +00:00

scott commented

2024-05-29 19:10:08 +00:00

Everything stays hardcoded for Computer Class; fix this for 0.1.0.