minutes.md

Coding Guidelines Subcommittee Meeting on 2026-03-27 @ 0800 JST / 1600 PDT

Link to meeting time in common time zones.

Search Key	Description
[todo]	Action Item
[decision]	Something decided on
[important]	Key information

Agenda

1. Solicitation of notetaker
2. Acceptance of Previous Meeting Minutes
3. Introduction of new members
4. Review batches for CERT C => Rust Mapping (Pete)
   • #427 [CERT C Review Batch 1/5] Review proposed Rust categorization
   • #428 [CERT C Review Batch 2/5] Review proposed Rust categorization
   • #429 [CERT C Review Batch 3/5] Review proposed Rust categorization
   • #430 [CERT C Review Batch 4/5] Review proposed Rust categorization
   • #431 [CERT C Review Batch 5/5] Review proposed Rust categorization
5. Discuss running Asia-Pacific meeting organization
6. Proposals and ideas for new rules (all)
7. Progress on ongoing tasks (all)
8. Round table

Check-in area

• Mikhail Antoshkin ☕
• Pete LeVasseur 🖖
• Yuchen Shen ☕

Please add your name, and an emoji that describes your day.

Notetaker:

• Yuchen

For tips on how we take notes in the Safety-Critical Rust Consortium, please see the Meeting Notetaker Role doc.

Housekeeping section

• Document space: coding-guidelines
• Zulip: safety-critical-consortium: Coding Guidelines
• Kanban board
  • contributor experience view
  • coding guideline view

Tasks

• xx

Meeting Minutes

• Have mapping between Cert C to Rust and Misra C to Rust in FY26
• MEM34 partially applicable to Rust, only in unsafe
  • We don’t do a lot manual allocations in Rust
  • There is case that user creates bindings to libc memory API
  • Do we need a specific rule for FFI?
  • List all the possibilities for Rust programmer to use free() and dealloc() in the mapped rule
• CON35 applicable but probably change the name to not mention “by the predefined order” but just “avoid deadlock”. Rust has better API to avoid deadlock
  • Parking_lot has the experimental mechanism to prevent deadlock
  • https://fuchsia-docs.firebaseapp.com/rust/lock_order/index.html
  • https://crates.io/crates/lock_ordering
• EXP36 cast itself should be safe? Only the access can cause undefined behavior.
  • But it doesn’t make sense to just cast the pointer without using the casted pointer. So better to just say do not cast the pointers.
  • What “strictly aligned” mean here?
    • Like cast a pointer to byte to pointer to a data structure, then the pointer can point into unaligned location in the data structure
• MEM35-C
  • We probably can reduce the work we do to have a single rule to say “do not break safety contracts” to cover all possible cases and refine later to specific cases.
  • It is good to explicitly document “do not break safety contracts” even it is obvious.
• Copy the discussion above into issue #431

Material

Any material to read before the meeting should be included here.