Require approval from t-infra instead of t-release on tier bumps

[Kobzol]

Discussed at https://rust-lang.zulipchat.com/#narrow/channel/242791-t-infra/topic/Tier.201.20target.20promotion.20RFC.20FCP.20sign-offs/with/532735844.

I also changed "viability and value" to just "viability". I think that t-infra should decide whether it's viable to support a given target on our CI. The value should be determined by t-compiler.

r? @jieyouxu

[jieyouxu]

Thanks, this looks good to me. I would like a compiler lead to sign-off, since it's technically a change in how we do Tier 1 target promotions.

[jieyouxu]

r? compiler_leads

[wesleywiser]

Has the release team weighed in on this? I agree this conceptually makes more sense but since the original policy was the result of an RFC, I think we should do a T-release/T-infra FCP to confirm this is what both teams want.

[Kobzol]

I asked Mark, who was fine with it (along with most members of t-infra), but yeah, we should probably involve the whole release team.

[Noratrieb]

@rust-lang/release hi everyone, this change takes you off the compiler target tier changes FCPs, are y'all happy with that or do you see a reason to involve the release team there?

[Mark-Simulacrum]

@rfcbot fcp merge

Historically the release team was on the FCP because early discussions imagined we'd be checking the "value" of a target, but I think T-compiler and T-infra are better positioned for this anyway (e.g., pulling statistics on usage, deciding how different individual targets are - and so how much usage of one indicates reason to promote others) and T-release doesn't have a large body of stakeholders in various platforms. T-compiler also has established processes for exploring these questions as part of tier 2 and 3 target proposals.

So proposing we merge this, happy to see any concerns registered.

[rfcbot]

Team member @Mark-Simulacrum has proposed to merge this. The next step is review by the rest of the tagged team members:

• @BoxyUwU
• @Dylan-DPC
• @Kobzol
• @Mark-Simulacrum
• @cuviper
• @jdno
• @jieyouxu
• @marcoieni
• @pietroalbini
• @shepmaster

No concerns currently listed.

Once a majority of reviewers approve (and at most 2 approvals are outstanding), this will enter its final comment period. If you spot a major issue that hasn't been raised at any point in this process, please speak up!

See this document for info about what commands tagged team members can give me.

[cuviper]

@rfcbot reviewed

I don't have a problem with switching release to infra here. However, on a tangent, I wonder why the "value" question has been delegated to compiler and not also libs, especially when non-host targets are considered.
(although I do realized that adding another team to fcps would be hard)

[jieyouxu]

I don't have a problem with switching release to infra here. However, on a tangent, I wonder why the "value" question has been delegated to compiler and not also libs, especially when non-host targets are considered.

In practice, I think libs do get asked about if the libs team would have concerns with promoting a target beyond Tier 3, but mostly from an impl POV (and not necessarily "value"). For example: rust-lang/compiler-team#864 (comment). (But this is also not necessarily "written down".)