Approach to case tableswitch may consume significant bytecode space

Question

The issue:
https://ask.clojure.org/index.php/13407/method-large-compiling-expression-unfortunate-clause-hashes

Logged as:
https://clojure.atlassian.net/browse/CLJ-2808

Possible approaches:
> One approach is to only allow smaller lookup tables.
Another would be to lift the lookup code for large tables into a function or a separate compiled method...

I suggest an approach that just measures a gap between tests.
No need to big span and shift/mask model.
If the gap is less than 128, then we use tableswitch; otherwise, we use lookupswitch.
https://github.com/clojure/clojure/compare/master...magnalium:clojure:dev-2808

Comments

This will be closed as a duplicate, I expect. Why not just post it as a comment on the original issue (and vote for it)?

---

Unfortunately, I don't this due to access level.

---

You mean you can't comment on https://ask.clojure.org/index.php/13407/method-large-compiling-expression-unfortunate-clause-hashes ? Are you sure?

---

In this topic I can to comment, but I want to discuss in JIRA.
If I need first to discuss in ask site, ok close this topic.

---

Per https://www.clojure.org/community/contributing#_reporting_problems_and_requesting_enhancements ask.clojure.org is the place to discuss issues. Jira is specifically for contributors. That page overall discusses the whole process.