Regression: `align` is ambiguous

[Mark-Simulacrum]

• https://crater-reports.s3.amazonaws.com/beta-1.89-4/beta-2025-07-03/reg/bitbuffer_derive-0.11.2/log.txt
• https://crater-reports.s3.amazonaws.com/beta-1.89-4/beta-2025-07-03/reg/java_asm-0.1.1/log.txt
• https://crater-reports.s3.amazonaws.com/beta-1.89-4/beta-2025-07-03/reg/soa-rs-0.8.1/log.txt
• https://crater-reports.s3.amazonaws.com/beta-1.89-4/beta-2025-07-03/reg/soapy-0.2.9/log.txt

[INFO] [stdout] error[E0659]: `align` is ambiguous
[INFO] [stdout]    --> tests/write.rs:329:3
[INFO] [stdout]     |
[INFO] [stdout] 329 | #[align]
[INFO] [stdout]     |   ^^^^^ ambiguous name
[INFO] [stdout]     |
[INFO] [stdout]     = note: ambiguous because of a name conflict with a builtin attribute
[INFO] [stdout]     = note: `align` could refer to a built-in attribute
[INFO] [stdout] note: `align` could also refer to the derive helper attribute defined here
[INFO] [stdout]    --> tests/write.rs:327:10
[INFO] [stdout]     |
[INFO] [stdout] 327 | #[derive(BitWrite, PartialEq, Debug)]
[INFO] [stdout]     |          ^^^^^^^^
[INFO] [stdout] 

[workingjubilee]

cc @folkertdev was this one of yours?

[folkertdev]

I implemented the attribute in #142507, but adding it was decided on in rust-lang/rfcs#3806.

[workingjubilee]

I don't think we should revert the change necessarily, I'm just contemplating what we should do (including gathering information).

[Jules-Bertholet]

Shouldn’t the derive attribute shadow the built-in one? Especially as the built-in attribute is unstable, and currently not permitted on ADT fields.

[Jules-Bertholet]

@workingjubilee Why was this closed?

[Mark-Simulacrum]

This will at least need relnotes, but I'm not convinced we should break users here without some kind of migration path.

[Jules-Bertholet]

The #[align(…)] language attribute is still feature-gated, and currently isn’t even allowed in that position (on ADT fields). We should not be breaking users even with a migration path. This is a bug.