T-spec meeting 2025-04-17

--- title: T-spec meeting 2025-04-17 tags: ["T-spec", "meeting", "minutes"] date: 2025-04-17 discussion: https://rust-lang.zulipchat.com/#narrow/stream/399173-t-spec/topic/Meeting.202025-04-17 url: https://hackmd.io/pr4G1IDFQT271RFvijlAEA --- Attendees: - Joel Marcey - Pierre-Emmanuel Patry - Eric Huss - Pete LeVasseur - Josh Triplett - TC - Jack Huey Regrets: - Niko Matsakis Agenda: - Any updates to the agenda? - New Team Member - FLS Transition Update - The Reference Rocks - Reference and FLS chapter comparison - Reference Railroad Diagram update ## Agenda Updates - Uncertain team assignments for pull requests ## New Team Member Josh Triplett is officially joining the spec team 👏 Welcome! Excited to have Josh on the team! ## FLS Transition Update Joel: We now officially have signed CLA/DCO documents from Ferrous Systems and AdaCore to the Foundation. Joel: We're now ready to go with moving the repo over to `rust-lang/`. When Felix returns next week, we can get this started. Joel: We'll move it over as we agreed to a repo named `fls` on our side. We'll leave a note in the README about the history of the document. We still need to work out what we want to do about the test annotations and test suite. If we upstream the test annotations, they'll offer up the tooling for the test matrix. TC: Had a thought the other day about what we could do on the test identifiers. The FLS has hash based identifiers, and people have had some reservations about these. Instead of committing these hash based identifiers to the test suite, we could create a mapping from these to the corresponding chapter or section of the Reference. We've been similarly annotating tests with Reference identifiers. So we'd use the FLS identifiers to add Reference identifiers to the test suite, then we'd have a mapping table from the Reference identifiers to the FLS hash based identifiers for building the FLS test matrix. This way, we get immediate test coverage on the Reference, and we only have one set of identifiers merged in the test suite. The FLS identifiers are coarse grained, but this would give us a place to start, and we would later work to make these more fine grained. This sort of mapping between Reference chapters and FLS chapters is something that we had talked about as being useful in other contexts, such as helping people move between these documents, so creating this mapping could be useful here also. ## Reference Rocks https://github.com/rust-lang/reference/pull/1792 TC: People were sometimes taking, I had noticed, a dim view of the Reference based upon front matter wording touting it as something that is not necessarily seen as official. So let's drop that. Josh: 100% believes that we should remove such wording. Pete: How does the FCP process work? TC: (Explains.) ## Reference Railroad Diagram Update https://github.com/rust-lang/rust/pull/139884 https://doc.rust-lang.org/nightly/reference/ The full grammar: https://doc.rust-lang.org/nightly/reference/grammar.html Here's the commit to look at for the change to the source: https://github.com/rust-lang/reference/commit/3a2ea7ef41484992803e5c8ea58736074e5c4d61 Basically everyone thinks this is awesome stuff! ## Uncertain team assignments - https://github.com/rust-lang/reference/pull/1752 - Guarantee soundness of pointer-to-int transmutes - https://github.com/rust-lang/reference/pull/1731 - Refactor unsizing coercion documentation - https://github.com/rust-lang/reference/pull/1661 - Define raw pointer transmute behavior Eric and TC work on some sort of PR owner template to show what they need to do to push the PR forward. Also think about how ping groups could be used in this situation. ## Reference and FLS chapter comparison ## Chat https://hackmd.io/pr4G1IDFQT271RFvijlAEA 8:01 PL Pete LeVasseur Pete LeVasseur says:👏 8:05 Pete LeVasseur says:Putting an example diff showing how the change improves readability for myself mostly: https://github.com/rust-lang/reference/commit/3340922df189bddcbaad17dc3927d51a76bcd5ed#diff-4481eac96a83ce35802b28fdcb117bbcc6a1caa78912efbc17fdd9113ed124d8R357 8:19 Pete LeVasseur says:👏 8:19 avatar Josh Triplett Josh Triplett says:👏 8:19 Jack Jack says:Sorry y'all, i gotta drop. I have an important meeting tomorrow i have to prep for. Grammar rendering looks great!