Revert "Rollup merge of #82296 - spastorino:pubrules, r=nikomatsakis"

[spastorino]

This reverts commit e2561c5, reversing
changes made to 2982ba5.

As discussed in #83641 this feature is not complete and in particular doesn't work cross macros and given that this is not going to be included in edition 2021 nobody seems to be trying to fix the underlying problem. When can add this again I guess, whenever somebody has the time to make it work cross crates.

r? @nikomatsakis

[nikomatsakis]

@bors r+

[bors]

📌 Commit bfb5c48aba30a976b7a8e73bb67db9ac657f2bc8 has been approved by nikomatsakis

[nikomatsakis]

Actually, @bors r-

Is there any special procedure for removing a feature gate? I don't think so, right? (cc @rust-lang/compiler)

[nikomatsakis]

Also we should probably discuss this in @rust-lang/lang meeting

[nikomatsakis]

Nominating for the @rust-lang/lang meeting:

Given the current status of the effort around pub macro_rules! and the buggy state of the implementation, I was saying I'd probably rather just back out the PR if it is not under active development.

But I'd like @rylev to weigh in here too.

[varkor]

Is there any special procedure for removing a feature gate?

The removed feature needs to be added to compiler/rustc_feature/src/removed.rs.

[crlf0710]

Personally I do feel there's some value in using pub(crate) macro_rules! even in its current form (I was writing a few macro_rules!-heavy crates and it's nice to use this because i don't have to rearrange the modules according the order the macros are used).

It would be nice if instead removing this, just add it to the list of incomplete features...

[danielhenrymantilla]

@crlf0710 yeah, the current approach to get that is a bit uglier: https://stackoverflow.com/a/67140319/10776437

• and can cause a false positive lint: Incorrect unused_import lint firing for a function-local macro_rules! #78894

[nikomatsakis]

We discussed this in the @rust-lang/lang meeting yesterday. There is no doubt that the feature has value, but right now there is nobody actively working to resolve the various problems that were encountered with its design, so we are inclined to remove the code (it can be added again later if the situation changes).

[nikomatsakis]

Can you add the feature to the list of deprecated features as @varkor suggested, @spastorino ?

[spastorino]

@nikomatsakis I guess you meant as a removed feature, right?. I got confused also because someone else was suggesting to add it to incomplete features and to not remove it.

[nikomatsakis]

@spastorino that is what i meant, yes.

[spastorino]

I guess since is the first version in which the feature would be removed (which would be 1.53) but I wasn't 100% sure, so please check this in particular when reviewing.

[spastorino]

Also, always check my broken english :)

[spastorino]

@nikomatsakis done, have made a couple of comments here and here

[spastorino]

Forgot to comment something I was chatting with Niko. The feature was introduced in 1.52 and the current stable version is 1.51, so we could also just apply the revert and backport the revert to beta so the feature would never reach stable.

[nikomatsakis]

@bors r+

[bors]

📌 Commit 83767d9 has been approved by nikomatsakis

[nikomatsakis]

Note: I would rather just revert the normal way and and let it ride the trains than bother with the increasing hassle/risk of backporting.

[bors]

⌛ Testing commit 83767d9 with merge 855c2d1...

[bors]

☀️ Test successful - checks-actions
Approved by: nikomatsakis
Pushing 855c2d1 to master...