Daniel Silverstone

On I've just had an interesting thought - I wonder if this is a question of partial vs. full target specifications... Could you try the following: `rustup update stable-x86_4-pc-windows-msvc` If...

@retep998 Yes, we're trying to resolve the issue, we're aware of what it ought to be and we can fix it fairly easily. A fix is being worked on here...

In #2854 the fix is in for partial toolchain names, and has a test, so the next release of rustup should fix the major pain of Windows at least. If...

Or you can acknowledge that the installer needs an exec-compatible temp dir and just set TMPDIR before running the installer.

OpenBSD is all tier 3 at this time, which means the required dockers for rustup to build will not be available since they are not produced for tier 3 platforms...

We have removed snap support so you will need to rebase and remove that bit of your change, also since this adds a proxy, we must be sure to put...

I am loathe to merge this until/unless we are certain it won't confuse others. Is the right answer to get it merged/released just before the stable release which contains r-a...

@ehuss I thought r-a had switched to some kind of thing in the compiler component for proc-macros -- do we still want this?

I think that ensuring `rust-toolchain{,.toml}` is properly honoured is an all-or-nothing thing - it couldn't "just" be the toolchain itself, no matter what. You're right that #1397 is about ensuring...

Given all the other arguments `rustup toolchain install` can be given, I'd favour a separate verb.