David Renshaw

Elaborating `List Float` is slow too. Does that also go through `OfNat`? ```lean set_option trace.profiler true in def a : List Float := [ 0.5, 0.5, 0.5, 0.5, 0.5, 0.5,...

The test suite can sometimes trigger this exception. ``` $ ./capnp-evolution-test --seed=1467142715 Randomly testing backwards-compatibility scenarios with seed: 1467142715 *** Uncaught exception *** ../src/capnp/compiler/evolution-test.c++:879: context: seed = 1467142715; PLEASE REPORT...

The usual thing is to use the `message::Reader::canonicalize()` method: https://github.com/capnproto/capnproto-rust/blob/c702d16dd9c8b5acdf2ae8937c954171fb164d49/capnp/src/message.rs#L286-L297 But that only works if your data-to-canonicalize is the root struct of a `message::Reader`, which is not the case for...

> In my case, message.set_root_canonical(root)?; would be message.set_root_canonical(txn)?;? yes. > I get a panic with that at this assert: What if you do ```rust let size = txn.total_size()?.word_count + 1;...

> Happy to try a PR for a built-in method if you want. Up to you. I don't intend to implement it myself in the near future.

Are you aware of [IMOLean](https://github.com/jsm28/IMOLean)? It has formalizations of problem statements for recent years. [Here](https://github.com/jsm28/IMOLean/blob/main/IMO/IMO2025P4.lean) is its formalization of this problem.

I agree that the capnpc could use a lot of refactoring. I don't think we need to bring in new dependencies, though. `quote` seems like overkill. We don't need to...

As far as I know, `bindgen` does not work in that direction. You give it a C library, and it gives you a Rust interface to it. Here, what you...

> So the suggestion is to add a anyptr-debug-print hook to register a method to interpret the AnyPointer as a pointer to something else and/or to print it in the...

> I see that to implement this, one has to implement the entire orphan concept. Am I seeing this correctly? Yeah, I think so. > Alternatively, is there a way...