Reserve prefixed identifiers and literals (RFC 3101)

[lrh2000]

This PR denies any identifiers immediately followed by one of three tokens ", ' or #, which is stricter than the requirements of RFC 3101 but may be necessary according to the discussion at Zulip.

The tracking issue #84599 says we'll add a feature gate named reserved_prefixes, but I don't think I can do this because it is impossible for the lexer to know whether a feature is enabled or not. I guess determining the behavior by the edition information should be enough.

Fixes #84599

[rust-highfive]

r? @jackh726

(rust-highfive has picked a reviewer for you, use r? to override)

[petrochenkov]

cc #84037 (one more in-progress implementation) @Julian-Wollersberger

[petrochenkov]

Could you assign this to me for the final implementation review, once it's decided who will implement this and what are the rules regarding raw strings etc.

[steffahn]

Added some more comments on zulip (archive).

[lrh2000]

cc #84037 (one more in-progress implementation) @Julian-Wollersberger

I guess it mainly focuses on implements the k# proposal now?

Could you assign this to me for the final implementation review, once it's decided who will implement this and what are the rules regarding raw strings etc.

OK, I'll do this.

Added some more comments on zulip (archive).

Thanks for your detailed feedback!

[jackh726]

r? @petrochenkov

[Julian-Wollersberger]

I guess it mainly focuses on implements the k# proposal now?

Yes, and #84037 touches almost the same code. But since the k#foo RFC isn't accepted yet, it'll take some time. So I'd suggest that you (@lrh2000) go forward with this PR, and I'll later rebase or close the other one. (I don't have a strong opinion there.) Sounds good?

[lrh2000]

Yes, and #84037 touches almost the same code. But since the k#foo RFC isn't accepted yet, it'll take some time. So I'd suggest that you (@lrh2000) go forward with this PR, and I'll later rebase or close the other one. (I don't have a strong opinion there.) Sounds good?

@Julian-Wollersberger Yes, it sounds good. Sorry for the reply delay since I'm really busy in the past several days.

I don't think we have reached an agreement on the implementation details now (see discussions at #84978), so this PR may also take some time and the code may be completely rewritten.

[jack-champagne]

I think you @'ed the wrong jack @lrh2000

[lrh2000]

I think you @'ed the wrong jack @lrh2000

Oops, what a big mistake! I've fixed this and really sorry to bother you.

[m-ou-se]

Hope you don't mind, but I pushed the changes i suggested above, and added a migration lint for cargo fix --edition together with a test for that.

[nikomatsakis]

This looks good to me. I left some nits.

[petrochenkov]

r? @nikomatsakis @bors r=nikomatsakis

[bors]

📌 Commit f6dd137 has been approved by nikomatsakis

[bors]

⌛ Testing commit f6dd137 with merge e8cb1a4...

[bors]

☀️ Test successful - checks-actions
Approved by: nikomatsakis
Pushing e8cb1a4 to master...

[asdf8dfafjk]

This may be known to most people here but Scala allows people to make their own prefixes, say sql"" without the use of any compiler plugins. Here are some more details https://www.scala-lang.org/api/2.12.2/scala/StringContext.html

I'd love to see that being supported by compiler and a very small number of useful individual prefixes in the std-lib.