feat: add expansion_growth_limit attr as another expansion limit

[vincenzopalazzo]

This PR required a little bit of content, as discussed on Zulip

The following code triggers a recursion that will keep adding tokens inside the list of tokens, as described here

/// Code that cause the rust issue #95698
macro_rules! from_cow_impls {
    ($( $from: ty ),+ $(,)? ) => {
        // recursion call
        from_cow_impls!(
            $( $from, Cow::from ),+
        );
    };

    ($( $from: ty, $normalizer: expr ),+ $(,)? ) => {
        $( impl<'a> From<$from> for LhsValue<'a> {
            fn from(val: $from) -> Self {
                LhsValue::Bytes($normalizer(val))
            }
        } )+
    };
}

from_cow_impls!(
    &'a [u8], /*callback*/
    Vec<u8>, /*callback*/
);

This PR adds a new attribute expansion_growth_limit which allows us to fix this corner case where this happens and also can catch some performance problems in case of macros resolution.

Closes #95698

@rustbot r? @Aaron1011

[vincenzopalazzo]

@rustbot r? rust-lang/compiler

[compiler-errors]

Not going to be able to review this for a while.

r? parser

[rustbot]

rustc_error_messages was changed

cc @davidtwco, @compiler-errors, @JohnTitor, @estebank, @TaKO8Ki

[wesleywiser]

r? rust-lang/compiler

[oli-obk]

I'm on vacation and will review in 2-3 weeks

[petrochenkov]

I need to be aware of what happens here, but right now I'm taking some break from reviewing, I'll hope to return to this some time later in January.

Right now this PR seems to set a limit on number of tokens produced by a macro (*), which is not right because macros can simply produce a lot of code without any recursion.

(*) Only a fn-like macro, but not an attribute or derive macro, which also doesn't look correct.
@rustbot author

[WaffleLapkin]

One day I'll learn how to use crater properly 😓 (it needs a try build...)

@bors try

[bors]

⌛ Trying commit 7ddb63a with merge d42a007e04d1d7df0de258d39ac2402f35aba932...

[bors]

☀️ Try build successful - checks-actions
Build commit: d42a007e04d1d7df0de258d39ac2402f35aba932 (d42a007e04d1d7df0de258d39ac2402f35aba932)

[WaffleLapkin]

@craterbot check p=1 crates=https://crater-reports.s3.amazonaws.com/pr-103029/retry-regressed-list.txt

[craterbot]

👌 Experiment pr-103029-3 created and queued.
🤖 Automatically detected try build d42a007e04d1d7df0de258d39ac2402f35aba932
🔍 You can check out the queue and this experiment's details.

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[craterbot]

🚧 Experiment pr-103029-3 is now running

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[craterbot]

🎉 Experiment pr-103029-3 is completed!
📊 1 regressed and 0 fixed (125 total)
📰 Open the full report.

⚠️ If you notice any spurious failure please add them to the blacklist!
ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[Dylan-DPC]

@vincenzopalazzo any updates on this?

[vincenzopalazzo]

@Dylan-DPC thinking how to move on on it, I guess it need a MCP and a compile team discussion

[Dylan-DPC]

@vincenzopalazzo any updates on this?

[vincenzopalazzo]

any updates on this?

Now really @Dylan-DPC I was unable to work on anything in the past few months, so I think I will start the discussion again.

Now that we have the @rust-lang/wg-macros I think the discussion should start from there before creating the MCP

[apiraino]

I'll set this to S-blocked to signal that it needs design work

@rustbot S-blocked