mirror of
https://github.com/ruuda/rcl.git
synced 2025-12-23 04:47:19 +00:00
Add some more ideas
I was thinking today about how the | operator is way briefer than writing a comprehension in a collection, but writing the comprehension is nice to format and | is awkward. Also when writing down the type of the | operator ... it's a challenge (also to typecheck it). Perhaps, adding a spread/unpack operator will resolve both issues. It would be slightly more verbose, but not terribly so. And I believe it would be more readable too.
This commit is contained in:
parent
5f29bc19b6
commit
6f746f3f8c
1 changed files with 8 additions and 1 deletions
|
|
@ -1,5 +1,9 @@
|
|||
# Roadmap
|
||||
|
||||
## In progress
|
||||
|
||||
* Type system.
|
||||
|
||||
## Near term
|
||||
|
||||
* Accept an expression through `--expr` for `rcl evaluate`.
|
||||
|
|
@ -14,10 +18,13 @@
|
|||
* Absorb the `highlight` command into `fmt` and make it use the same coloring?
|
||||
* Evaluate whether to add a tuple type to describe heterogeneous lists of
|
||||
statically known length.
|
||||
* Add a spread operator for inside collections, and evaluate whether it should
|
||||
supersede the dict/set union operator. It would solve some typing issues, and
|
||||
it's easier to format in a pleasant way.
|
||||
* Add nested fields to record syntax, i.e. `{x.y = "z"} == {x = {y = "z" }}`.
|
||||
|
||||
## Long-term
|
||||
|
||||
* Type system.
|
||||
* Add a more efficient runtime system, possibly with a garbage collector.
|
||||
* Multithreaded evaluation. Due to strict evaluation and being purely
|
||||
functional without side effects, this is actually feasible: when we evaluate
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue