Sean Cross

Specialization would make so many of these things easier, I agree. I considered the `Box` acceptable for now since it only happens once per session, seeing as how it's a...

As an example: ``` $ RUSTUP_LOG=trace RUSTUP_DIST_SERVER=https://xous.xobs.io/dist rustup +stable target install riscv32imac-unknown-xous-elf 2025-12-09T08:58:42.927239Z DEBUG run_rustup:run_rustup_inner:main:set_globals: rustup::cli::common: read metadata version: '12' current_dir="/Users/seancross/Code/39c3-talk" args="[\"rustup\", \"+nightly\", \"target\", \"install\", \"riscv32imac-unknown-xous-elf\"]" current_dir="/Users/seancross/Code/39c3-talk" quiet=false 2025-12-09T08:58:42.927348Z TRACE...

I'm sorry, I must have mis-copied the output. I've been switching back and forth between +nightly and +stable. I would have expected the +nightly version of rustup to have the...

It's not hitting my manifest file at all. I've made this patch: ```patch diff --git a/src/download/mod.rs b/src/download/mod.rs index 37ae9b32..f1053c2c 100644 --- a/src/download/mod.rs +++ b/src/download/mod.rs @@ -254,6 +254,7 @@ impl Backend...

I think part of the problem may be that it likely uses a channel toml file from the sysroot and not the server? I could see there being security considerations...

That wouldn't be a problem, and in fact would be preferred. I'd rather have the minimum amount of stuff to specialize a tier 3 target and keep everything else upstream....

I'm not sure that works, but I'm not entirely understanding the proposal. How would that allow for installing a `+stable` toolchain for tier-3 targets? The proposal looks like a way...

We [already manually install it](https://github.com/betrusted-io/xous-core/blob/4b8a202985e90781a5adaa1efeeb06f2666ab1dd/xtask/src/utils.rs#L166-L357) by unzipping a file to `$(rustc +stable --print sysroot)/lib/rustlib` which gets us the ability to use normal rustc to compile. I opened this issue because...

I should be able to come up with a Zephyr build for Raspberry Pi 4 that exhibits the problem.

I've seen this as well. Can you check if https://github.com/probe-rs/probe-rs/pull/3699 fixes it for you? I've been running with this patch for about three weeks and it's fixed the reset issues...