Stream: project-error-handling

Topic: Meeting 2021-10-25


view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 18:58):

Meeting time soon

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 18:59):

agenda: https://hackmd.io/@rust-libs/BywtRd4It

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 18:59):

some exciting updates today!

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:00):

https://github.com/rust-lang/project-error-handling/issues/3

view this post on Zulip Charles Ellis O'Riley Jr. (Oct 25 2021 at 19:00):

cool

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:00):

the PR adding support for negative traits to coherence has merged

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:01):

so now we should just need to sort out the issues around fn backtrace to get Error moved into core

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:01):

so the next priority will be landing generic member access

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:01):

@Nick Cameron lmk if you're not going to be able to work on that soon, if so I'd like to take over the RFC / PR adding it

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:01):

since I see generic member access as the best path forward

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:02):

we should try to land that, then create a migration plan for removing fn backtrace from nightly and moving Error into core

view this post on Zulip Nick Cameron (Oct 25 2021 at 19:04):

Hey, sorry, my contract/paperwork nonsense has been dragging on, but I'll be back at work from next Monday and will get on the RFC ASAP after that

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:04):

perfect, thank you Nick!

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:06):

in related news, we realized that we don't need to start with an RFC for std::error::Report since libs changed their process for adding new unstable APIs, so @Sean Chen (he/him) went ahead and opened a PR adding std::error::Report

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:06):

https://github.com/rust-lang/rust/pull/90174

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:06):

working out the bugs right now but we should have that on nightly soon enough

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:06):

so the plans to create a trial-and-error crate are likely going to just be canceled

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:07):

since the only experimental APIs it had were for Report and for a Box<dyn Error> alternative

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:07):

and along those lines, I have a new goal for how to resolve the issues around Box<dyn Error>

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:07):

https://rust-lang.zulipchat.com/#narrow/stream/219381-t-libs/topic/ThinBox

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:07):

separately the need for a ThinBox type came up which I've started looking into adding as an unstable API

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:08):

and I realized in that conversation that such a type would be a perfect opportunity to provide a Box type for handling dyn Errors that does actually impl Error

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:08):

while also acting as a thin pointer to the error trait object

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:09):

which is separately a desirable feature as shown by eyre and anyhow.

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:09):

so if we do still do trial-and-error it will likely only end up containing the Result type that has the proper FromResidual impls to ? arbitrary error types into a ThinBox<dyn Error>

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:10):

since with the proper error impl ThinBox<dyn Error> would not work with ? when used with std::result::Result

view this post on Zulip Jane Lusby [she/her] (Oct 25 2021 at 19:11):

So that's all I've been focused on recently

view this post on Zulip Sean Chen (he/him) (Oct 25 2021 at 19:19):

Sweet. I'm continuing to work on Mara's feedback on the std::error::Report PR. Afterwards, I'm keen to help out with the other tasks you have on the docket :smile:

view this post on Zulip Charles Ellis O'Riley Jr. (Oct 25 2021 at 19:26):

I'm open to getting my feet wet in any way I can


Last updated: Jan 26 2022 at 14:02 UTC