Net WG Lead Meeting #1
Maximum duration: 60 minutes Date picker: https://doodle.com/poll/xttrxdp3f8xukfad
Objective
This will be the first meeting Net WG meeting since we kicked off the different sub-working groups. Let’s figure out ways we can keep momentum going, and effectively communicate with the wider Rust community.
Agenda
Feel free to add more items that you think are important.
- General Introduction - 5m (Yosh)
- Agree someone to take notes
- Check if any agenda items are missing
- WG Status - 25m
- Web - 5m 1. Progress 2. Goals 3. Issues
- Async - 5m 1. Progress 2. Goals 3. Issues
- Embedded - 5m 1. Progress 2. Goals 3. Issues
- Blog - 10m
- Content planning?
- Newsletter (link) - 10m
- Agree on date
- Content planning?
- Schedule next meeting - 5m
- Have a doodle poll ready
Minutes
Attendees: yoshuawuyts, lkurusa, MajorBreakfast, Nemo157, cramertj
- Brief overview from each WG (goals, progress and blockers; to have a shared sense of what’s going on)
- WG-Web
- Two goals since kickoff:
- to build out a learning web framework (tide)
- to investigate the web ecosystem and help out where pieces are missing
- aturon primarily responsible for the former, yoshuawuyts for the latter
- Blog posts for tide in progress, placeholder repo at https://github.com/rust-net-web/tide
- Two issues for ecosystem investigation:
- https://github.com/rust-lang-nursery/wg-net/issues/44, a call for example web projects to help inform what folks are running into
- https://github.com/rust-lang-nursery/wg-net/issues/54, to help document missing parts of tokio
- Inaugural WG-Web meeting next Thursday
- Two goals since kickoff:
- WG-Async
- New methodology for testing, the
futures-testcrate. Contains various helpers to make writing tests for 0.3 futures easier. Will be shipped as part of the upcoming alpha 4 of thefuturescrate. - Disruption with the sudden changes to the
stdPinAPIs. Came as a surprise, will also be fixed in the alpha 4 release. - Work on the compatibility layer, blog post is ready.
tokio-async-awaitwas released by the tokio team, good opportunity for some collaboration; maybe move all the compatibility stuff to this crate. - cramertj has started discussion about a
FusedFuturetrait that should make it possible to check whether a future has already completed and produced its value (https://github.com/rust-lang-nursery/futures-rs/issues/1219). PR ready.- Easy to support with
asyncproduced futures as they already track this state for memory safety, - Difficult to decide on the best implementation strategy, discuss on the issue.
- Easy to support with
- boats has been reworking the pin types, but has hit some stumbling blocks in method resolution, so that’s stuck at the moment (https://github.com/rust-lang/rust/issues/53843).
- cramertj has been working on a PR for a new “scoped tasks” feature, similar to crossbeam’s, that would let you spawn futures that could reference your local stack.
- cramertj has also been giving some thought on how to solve the issue of wanting to return early from a function that returns an async block; finding it quite common in fuchsia so been exploring options there
- discussion on the wg-net repo about the async programming book WG-embedded
- nemo mostly been working on https://github.com/nemo157/embrio-rs - a
no_stdcompatible cross-target executor + futures targeted Hardware Abstraction Layer (HAL) for interacting with micro-controller peripherals. - nemo major blocker is lack of async/await in no-std environments.
- nemo async/await is also needed for traits in HAL because it’s basically a big collection of traits.
- nemo blog post + docs are in the works.
- nemo been meaning to have more discussion about standardizing async IO traits (https://github.com/rust-lang-nursery/wg-net/issues/53) + traits for datagram protocol like UDP and bring up ideas around removing the
poll_*layer and having traits that return futures directly. - lkurusa got hold of a bunch 16 ARM controllers - hoping to use them to pinpoint problems with rust.
- lkurusa released cgroups crate — also been attending the embedded WG meetings.
- New methodology for testing, the
- WG-Web
Blog
aturon emphasised that we do have a blog, and we should probably make more use of it discussion on possible blog posts/plan for content MajorBreakfast: what about encouraging framework authors to present their framework in a blog post, we could come up with some kind of limited length format that could be followed MB: it would also mean that the authors could influence directly what is written, instead of someone else describing their work it’d be the authors themselves lkurusa: could we agree on a set of questions/bulletpoints beforehand MB: create an issue for it and agree on a semi-flexible style for the posts MB: Limit length so that we don’t get tutorials, but instead a glance at what the project’s are about yoshuawuyts will open issue, MB is lacking in spare time to followup ATM yoshuawuyts: aturon mentioned wanting to blog about tide a bunch, but not sure where the posts are intended to go
Newsletter
https://github.com/rust-lang-nursery/wg-net/issues/51 similar to the embedded WG, regular newsletter with what’s been happening suggestions to post to internals and blog (for an easily accessible archive) target date for first issue: 2018-09-12