08:13 <AshconMohseninia> does anyone with probe-rs know why vscode might be saying "Unknown source" but can list the function stack?

08:27 <Noah[m]> probably it's just not implemented properly. But it could also be bad debug symbols :) having the symbol and having the file are two different things :)

11:50 <Ralph[m]> maybe interested for the RE / REC owned crates?

11:50 <Ralph[m]> https://blog.rust-lang.org/2025/07/11/crates-io-development-update-2025-07/#trusted-publishing

11:50 <Ralph[m]> it's now possible to set up a release workflow on GH which does not need to have an explicit token stored in a secret to publish to crates.io. for a team-based setup this should simplify things (just make sure that only the right people have the possibility to trigger the release workflow and you should be set)

11:51 <Ralph[m]> s/interested/interesting/

14:30 <thejpster[m]> yeah I saw that. Sounds like a great idea. But I don't have the time to set it up.

14:39 <thejpster[m]> Tier 2 targets - get good or get gone: https://github.com/rust-lang/compiler-team/issues/896

14:41 <thejpster[m]> I propose wg-embedded/t-arm adopts armv7r-none-eabi, armebv7r-none-eabi, armebv7r-none-eabihf, armv7a-none-eabi and armv7r-none-eabi.

18:25 <RobinMueller[m]> is there a specific reason that the nop function was not completely feature gated like the other asm method inside the cortex-m crate? (see here: https://github.com/rust-embedded/cortex-m/blob/6d566b220b9fe1c8e67f7a6808bf105e3f01dc03/cortex-m/src/asm.rs#L56)

18:26 <dirbaio[m]> probably so it compiles for all targets

18:26 <dirbaio[m]> since you can "implement" nop in all targets by just ... doing nothing

18:31 <thejpster[m]> in an ideal world we'd have had an aarch32-cpu crate, and then cortex-m and cortex-ar wouldn't each have copies of the same intrinsics

18:36 <RobinMueller[m]> thejpster: for https://github.com/rust-embedded/cortex-ar/pull/46, go for consistency and feature gate them completely? or should be allow using nop on non cortex-ar platforms?

18:36 <RobinMueller[m]> (don't really see where/how this would be useful, but maybe you do?)

18:36 <thejpster[m]> I'd got for consistency. If you're using one asm intrinsic, you're likely using several, and having only one available is weird.

18:36 <thejpster[m]> Which means you could just gate the whole module?

18:36 <thejpster[m]> s/got/go/

18:37 <RobinMueller[m]> yes

18:40 <dirbaio[m]> thejpster[m]: it's a breaking change tho.

18:41 <thejpster[m]> the users of cortex-ar are basically me and Robin at the moment - I can live with breaking it

18:41 <dirbaio[m]> ah okay you aren't talking about changing this in cortex-m. sorry I misunderstood.

18:41 <thejpster[m]> RobinMueller[m]: what host machine were you on?

18:42 <thejpster[m]> dirbaio[m]: having spent hours on the cortex-m delete-and-start-again PR, my appetite for breaking changes in the cortex-m crate is as low as can be

18:42 <dirbaio[m]> yeah 🤣 that's why I was mildly surprised you'd suggest doing a breaking change

18:42 <RobinMueller[m]> oh, I'm on regular x86_64.. I'll investigate this..

18:46 <RobinMueller[m]> okay, I guess the PR is not required. just need to do some specific feature gating in my host tool. maybe the feature gate can be made cortex-ar specific though?

19:52 <thejpster[m]> I guess the functions will work fine if you're on a 32-bit Raspberry Pi.