remove untrue/contentious statement

[tshepang]

One can use unstable features on stable too, like via an env var

r? ghost

[BoxyUwU]

That env var is not supposed to be used by anyone. It is very much true that for pretty much all purposes nightly is the only channel where you can use unstable features

[Noratrieb]

yeah it's probably more useful to people for understanding the toolchains to leave it as is

[tshepang]

the statement is misleading though, and it's not true... I will add intended

[jieyouxu]

The current statement I find is accurate, and less confusing than trying to caveat it with "intended".

[tshepang]

it would be accurate if a RUSTC_BOOTSTRAP did not exist, and that is something widely used

[BoxyUwU]

RUSTC_BOOTSTRAP is very much not widely used

[tshepang]

I should of chosen a different word

[jieyouxu]

I actually find removing the intended usage to be more confusing¹, so I think we should retain current wording.

Footnotes

1. Or less helpful. ↩

[rustbot]

⚠️ Warning ⚠️

• This PR is based on an upstream commit that is older than 7 days.

  It's recommended to update your branch according to the rustc-dev-guide.

[tshepang]

I see now how I made it less helpful. How about now...

• I restored the explanation
• I removed the false claim
• I added the intent (to imply that people should avoid using feature gates in beta and stable)

[jieyouxu]

Suggestion: can you say something more like "are intended/designed to be usable through opt-in via feature gates"? (Well, ideally with a less confusing wording than what I wrote.) "Should" sounds a bit to me like a "weak recommendation/vibe", but this really is the intended stability versus experimentation design setup. RUSTC_BOOTSTRAP has nothing to do with the design intention, it is merely an internal implementation detail to make bootstrapping possible. I would also just say "unstable", since that's why we have different release channels. Whether said unstable feature is incomplete/experimental doesn't really matter.