Revert "Add rustc_test_entrypoint_marker"

[JonathanBrouwer]

Reverts #158854 because it caused two beta regressions
Revert required some minor solving of conflicts
We can think about re-landing it after

Fixes #161920
Fixes #161917

cc: @Nadrieril @N1ark @aDotInTheVoid @jdonszelmann

[rustbot]

Some changes occurred in compiler/rustc_passes/src/check_attr.rs

cc @jdonszelmann

Some changes occurred in compiler/rustc_attr_ir

cc @jdonszelmann

Some changes occurred in compiler/rustc_attr_parsing

cc @jdonszelmann

[rustbot]

r? @dingxiangfei2009

rustbot has assigned @dingxiangfei2009.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

• Owners of files modified in this PR: compiler
• compiler expanded to 75 candidates
• Random selection from 19 candidates

[N1ark]

this is really unfortunate :( i wonder if we can fix this later; #161920 seems to just be an issue in the test_retry macro that needs fixing anyways, but idk what to do about #161917

[JonathanBrouwer]

Yeah #161917 is the annoying one of the two.
For beta the safe thing to do is just to revert.
For on nightly we might be able to re-land this as a synthetic attribute? I'd love to find a way to still have this without breakage

[mejrs]

Consider also cases like

#![feature(macro_metavar_expr_concat)]
#![feature(macro_attr)]

macro_rules! ohno {
    attr() { $(#[$a:meta])* fn $name:ident () $body: block } => {
        $(#[$a])*
        fn ${ concat(not_, $name) }() $body
    }
}

#[ohno]
#[test]
fn works() {}

#[test]
#[ohno]
fn no_works() {}

error[E0658]: use of an internal attribute
  --> src/lib.rs:15:1
   |
15 | #[test]
   | ^^^^^^^
   |
   = help: add `#![feature(rustc_attrs)]` to the crate attributes to enable
   = note: the `rustc_test_entrypoint_marker` attribute is an internal implementation detail that will never be stable

error[E0425]: cannot find function `no_works` in this scope
  --> src/lib.rs:17:1
   |
17 | fn no_works() {}
   | ^^^^^^^^^^^^^^^^
   | |
   | you might have meant to refer to this constant
   | not found in this scope

So it was already possible to combine misbehaving attribute macros and #[test] to create malformed output.

Another way to fix this and these issues is to make the test macro reorder itself, i.e. for

#[test] // not last
#[ohno]
fn blah() {}

we expand into

#[ohno]
#[test] // last!
fn blah() {}

[mejrs]

The following also does not work (despite not being weird) , can you add it as a test case?

#![feature(macro_attr)]

macro_rules! ohno {
    attr() { $(#[$a:meta])* fn $name:ident () $body: block } => {
        $(#[$a])*
        fn $name () $body
    }
}

#[test]
#[ohno]
fn my_test() {}

produces:

error[E0658]: use of an internal attribute
  --> src/lib.rs:10:1
   |
10 | #[test]
   | ^^^^^^^
   |
   = help: add `#![feature(rustc_attrs)]` to the crate attributes to enable
   = note: the `rustc_test_entrypoint_marker` attribute is an internal implementation detail that will never be stable

For more information about this error, try `rustc --explain E0658`.

View changes since this review

[JonathanBrouwer]

Good idea! Added
@rustbot ready

[mejrs]

r? me

r=me with test fixed

View changes since this review

[mejrs]

Suggested change

	//@ check-pass
	//@ check-pass
	//@ compile-flags: --test

[JonathanBrouwer]

Oh whooops that's terrible, thanks for catching that!

[JonathanBrouwer]

Confirmed that the test now fails before the revert.
@bors r=mejrs rollup

[rust-bors]

📌 Commit 9b70ab0 has been approved by mejrs

It is now in the queue for this repository.