Should Thumb bugs get increased visibility, e.g. Tier 1 status?

[pnkfelix]

Spawned off of discussion here: #58605 (comment)

Do we need to explicitly revise stuff in our tagging/tests/release processes to try to decrease risk of Thumb bugs?

Namely these two points:

we may want to expand the number of items tested in the thumb-* tests contained in https://github.com/rust-lang/rust/tree/master/src/test/run-make

and:

There was a discussion at the 2018 all hands about making the thumb targets tier 1, however this was not done as @alexchrichton mentioned that the concept of tiers was going away, or was being retooled.

If not, I'd like to reintroduce the discussion of making the thumb targets an official tier one target, as it already has some of the ground work laid to claim this

[pnkfelix]

cc @alexcrichton

[pietroalbini]

One of the questions we should try to solve is the requirements we want for tier 1 target and the process to promote them.

[Centril]

Another question is to clarify what Tier 1 means. For example:

• Can it ever be demoted?
  • Under what circumstances? (e.g. people stopped using the target because it got really old)

[alexcrichton]

I don't personally think this is really a question we can directly answer at this time in the sense that it feels tied up in a larger discussion about tiers, what they mean, etc. Until that higher level question is answered we don't have a direct answer to the question here other than "well we'll try to fix them in a timely fashion like all other bugs if they crop up"

[jonas-schievink]

Another question is to clarify what Tier 1 means. For example:

• Can it ever be demoted?

  • Under what circumstances? (e.g. people stopped using the target because it got really old)

This now has an answer in the form of #67724

Also: cc rust-lang/rfcs#2803

I hope to reopen the discussion on this as I am also interested in getting tier-1 embedded platform support.