Sometimes compiler panics!

[omid]

Sometimes I face this long error message:

thread 'rustc' panicked at 'failed to lookup `SourceFile` in new context', compiler/rustc_query_impl/src/on_disk_cache.rs:500:22
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

error: internal compiler error: unexpected panic

note: the compiler unexpectedly panicked. this is a bug.

note: we would appreciate a bug report: https://github.com/rust-lang/rust/issues/new?labels=C-bug%2C+I-ICE%2C+T-compiler&template=ice.md

note: rustc 1.59.0-nightly (83b15bfe1 2021-12-28) running on x86_64-unknown-linux-gnu

note: compiler flags: -Z binary-dep-depinfo -Z save-analysis -C embed-bitcode=no -C debug-assertions=off -C overflow-checks=on -C incremental

note: some of the compiler flags provided by cargo are hidden

query stack during panic:
#0 [typeck] type-checking `some::where::something`
#1 [analysis] running analysis passes on this crate
end of query stack

[est31]

Thanks! Can you provide a setting where, from a non-existant target directory, one can reproduce the issue?

[omid]

Hell @est31, I don't know how to debug it. I don't know how I can have a non-existent target directory. Should I run it with RUST_BACKTRACE=1?

[est31]

Non existent target directory means that you delete it before you run it, so that others can reproduce the issue (incremental bugs depend on the target directory for example).

[omid]

yep, no errors after removing the target.

[omid]

Sorry, I don't know how udeps works. Does this test mean it's not a udeps issue?

[est31]

It is likely a bug in the rust compiler when invoking the save-analysis backend. Maybe it's even a dupe of https://github.com/rust-lang/rust/issues/92355 . As it doesn't seem like a bug in udeps, I'm closing this issue.