Liam Miller-Cushon

Thanks for the report. Do you have a minimal, self-contained example that reproduces the crash?

There was some related discussion to `-XepAllWarningsAsErrors` in #1589 and #648. I wondered if setting javac's `--should-stop=` policy to a later phase would help but it doesn't. I think the...

We are planning to update the style guide to allow this, the intent is not to discourage unnamed variables or patterns.

https://github.com/google/error-prone/issues/4918#issuecomment-2952215463 is a good workaround for projects using versions of Error Prone affected by the crash. I looked at the repro in https://github.com/google/error-prone/issues/4918#issuecomment-2825502199: ``` javac \ -J--add-exports=jdk.compiler/com.sun.tools.javac.api=ALL-UNNAMED \ -J--add-exports=jdk.compiler/com.sun.tools.javac.file=ALL-UNNAMED \...

> But a small change in Error Prone like I made in #5003 will solve this problem. This change ignores the situation rather than simply throwing an NPE. This makes...

> Lombok authors are open to communicate and find solutions 🙏 [projectlombok/lombok#3917](https://github.com/projectlombok/lombok/issues/3917) I left a note in https://github.com/projectlombok/lombok/issues/3917#issuecomment-3580606317

I think this might be a dupe of #4819, which was just fixed. @graememorgan is also working on some additional improvements that should make Inliner more robust.

Sorry, I was confused earlier, as Graeme said in https://github.com/google/error-prone/issues/4866#issuecomment-2701584213 the fix I was thinking of was for InlineMe, not for refaster

Thanks for the report. I'd be happy to take a PR to add `@hidden` or any other standard block tags this check is missing

There isn't a line break after `return` because that token is a fixed length. Theoretically a line break there could avoid going over the style guide's 100 character limit, but...