make super_relate_consts use trait objects

[lcnr]

super_relate_consts generates quite a lot of llvm-ir considering that it is still fairly rarely used, so let's try using trait objects for that.

[rust-highfive]

r? @oli-obk

(rust_highfive has picked a reviewer for you, use r? to override)

[lcnr]

@bors try @rust-timer queue

[rust-timer]

Awaiting bors try build completion

[lcnr]

@bors try

[bors]

⌛ Trying commit 1a28bb0 with merge 49d7eb9f08603d5d5d3a7f1e34de7ab706096d3c...

[bors]

☀️ Try build successful - checks-actions
Build commit: 49d7eb9f08603d5d5d3a7f1e34de7ab706096d3c (49d7eb9f08603d5d5d3a7f1e34de7ab706096d3c)

[rust-timer]

Queued 49d7eb9f08603d5d5d3a7f1e34de7ab706096d3c with parent 7bade6e, future comparison URL.

[rust-timer]

Finished benchmarking try commit (49d7eb9f08603d5d5d3a7f1e34de7ab706096d3c): comparison url.

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. Please note that if the perf results are neutral, you should likely undo the rollup=never given below by specifying rollup- to bors.

Importantly, though, if the results of this run are non-neutral do not roll this PR up -- it will mask other regressions or improvements in the roll up.

@bors rollup=never

[jackh726]

Perf is basically neutral and this just increases complexity. I'm not sure the tradeoff is worth it.

[lcnr]

It looks like a small improvement to me, but as @jackh726 mentioned it does worsen the experience of working on it in the future so I am going ahead and close this for now