name mangling cannot be reversed

[thestinger]

We're using hacks like identity numbers instead of paths to traits and types.

[thestinger]

Nominating for the production ready milestone. Debugging and profiling really need to be sane, and these are going to be visible in any traceback.

[brson]

Please do not assign issues to milestones. The nomination process is documented here: https://github.com/mozilla/rust/wiki/Note-milestone-nomination

[thestinger]

@brson: Ah, I had the wrong impression about how the process worked.

[pnkfelix]

Visiting for triage, email from 2013-09-02.

I might argue that this should be part of Milestone 4, Well-Covered, before production-ready; Tools (both internal and third-party) to evaluate code quality and code correctness are going to need this.

[alexcrichton]

Improvement will be made in #8875, although it's mostly fixing bugs there in name collisions, and I think that there's still room for improvement.

[catamorphism]

Accepted for production-ready

[pnkfelix]

Assiging to 1.0 milestone, since it would be good to have this and it sounds like we might almost already have it. But also giving it priority P-low.

[alexcrichton]

Nominating for removal from a milestone. This is pretty backwards compatible, I think it's safe to have the exact definition of our symbols in flux at 1.0.

[brson]

P-low, not 1.0.

[alexcrichton]

http://mentorembedded.github.io/cxx-abi/abi.html#mangle.template-param appears to be a good summary of C++ name mangling and how we may be able to leverage the support from C++ for things like generics.

[steveklabnik]

Triage: no changes here.

[steveklabnik]

Triage: no change

[bstrie]

I'm not sure what the complete extent of the scope of this issue was intended to be, but our mangled symbols are indeed reversible (see @alexcrichton 's repo at https://github.com/alexcrichton/rustc-demangle/blob/master/src/lib.rs ). I highly suspect that the exact symbol mangling scheme isn't considered stable though, but I'd consider that to be a distinct issue that we'd need to tackle if we ever wanted to stabilize our ABI.