Relative scoping for hierarchical ADT arguments

I think the best way is to comment on the PR with your doubts or questions. The more we have comments (with structured arguments) the better informed committee decision will be

6 Likes

I’m surprised I didn’t suggest backticks for the distinguishing syntax, as I’ve said that we don’t have enough uses for backticks.

It would add a minor incompatibility for identifiers in scope; but I’m assuming the new scoping would be highest precedence, not lowest as suggested for a prefixless proposal.

I plan to change my CAPS LOCK from CTRL to backquote as soon as I remember how.

You know what has made me ā€œbacktick-fluentā€ is not infix notation but markdown on tickets and PRs.

(Speaking of backticks, be aware it’s a bad year for tick-borne disease in the U.S. Don’t neglect to check your back.)

1 Like

I have, kind of, in the Github thread. I would say backticks count as a form of stropping :grin:

Indeed, that is what reminded me!

Worth replying here to recognize your groundbreaking nudging.

Everyone, I had a thought about a possible compromise.
Please read my latest comment on the SIP thread https://github.com/scala/improvement-proposals/pull/134#issuecomment-4392978166 and share your thoughts.

Eventually the proposal was updated back to #(hash) based leading character.
The SIP PR is the same one

The implementation has a new PR

1 Like

Third time’s a charm :crossed_fingers:

After further thought and experimentation with implementation and diagnostics, I decided to adopt’s @odersky sigil-free proposal (not limited to enums). It’s aptly named ā€œCompanion Scope Inferenceā€.

Implementation:

5 Likes

Sigil-free seems most natural to me, resulting in nice and natural source. I understand concerns about contextuality of this feature, which is a bit unusual, but still I think the benefits overweight downsides.

From the variants offered, I like this most, and I hope this will be accepted.

3 Likes
1 Like