Conversation
I'd be happy to help flesh it out, though I don't have strong opinions on whether it is necessary. It would probably be fine to push the rest of the changes, point out the gap, and ask the libs team what they think rather than adding it proactively, so I guess my preference would be to wait and see, but if you'd prefer to expand on it I'm happy to support that work. |
That sounds good to me, save additional work for future me only if it's necessary :). So this is a PR to update your PR -- is that the best way to proceed or should I just re-open RFC-2895 so I can work on and get feedback directly on the PR that will be merged into master? Would the RFC name change in that case? |
That sounds like a smart plan. And if by name change you mean would the RFC number change? yeah |
|
@yaahc here's the new PR targeted at master branch with my updates added: rust-lang#3461 |
This PR does three things:
backtraceunstable feature (which I believe has been stabilized) and the defunctbacktracemethod onErrorprovide_anyanderror_generic_member_accessbe mergedOne thing I think is still missing from my changes to this RFC is a bit more detail about how
<dyn Error>.request_refand<dyn Error>.request_valueare implemented in terms of the new private typecore::any::TaggedOptionand modulecore::any::tags. I think that's appropriate because of the documentation on the reference-level explanation section: https://github.com/rust-lang/rfcs/blob/master/0000-template.md#reference-level-explanation@yaahc I think your offer to sync up and discuss this further would help me flesh this section out, unless you don't think it's necessary to go into any further detail.