Const traits

[nikomatsakis]

Metadata
Owner(s)	@fee1-dead
Team(s)	lang, types
Goal document	2024h2/const-traits

Summary

Experiment with effects-based desugaring for "maybe-const" functionality

Tasks and status

• Implementation (@fee1-dead)
• Discussion and moral support (types, lang )

[nikomatsakis]

This issue is intended for status updates only.

For general questions or comments, please contact the owner(s) directly.

[fee1-dead]

Update 2024-07-27:

• Key developments: We recently landed my new proposed desugaring in #120639, which is a big step towards having a robust model of const traits in the type system.

[fee1-dead]

Update 2024-08-25:

• Key developments: A handful of cleanups (#126552 #126737 #127452 #128424 #129500), opened a project group repo for tracking design considerations. A proposed solution to making implied bounds with super traits work (project-const-traits#2) is open: #129499

[fee1-dead]

Update 2024-10-02:

• Key developments: With the PR for supporting implied super trait bounds landed (#129499), the current implementation is mostly complete in that it allows most code that should compile, and should reject all code that shouldn't. Further testing is required, with our next steps being improving diagnostics (#131152), and fixing more holes before we'd be comfortable with adding const traits back to core.

[fee1-dead]

Update 2024-11-20:

#131985 by @compiler-errors completely restructured the way const traits are desugared, which made the design easier to understand and more robust against our current unit tests. Since then, there had been significant development and cleanup for the feature, with #131968 #132118 #132168 #132169 #132227 #132275 #132344 #132366 #132368 #132479 #132823 among the ones that got merged, and #133216 #133218 that are still open.

With the recent work, we are very close to being able to dogfooding the feature on the standard library again, and the current implementation is the closest we have towards something that can be stabilized.