`./x.py test` panics

[RalfJung]

$ ./x.py test
Building bootstrap
    Finished dev [unoptimized] target(s) in 0.18s
thread 'main' panicked at 'fs::read_dir(builder.src.join(&relative_path).join("redirects")) failed with No such file or directory (os error 2)', doc.rs:228:21
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
Build completed unsuccessfully in 0:00:01

This might not be a valid request (I have to tell it what to test), but surely it shouldn't just result in such an obscure panic.

Cc @rust-lang/bootstrap

[chenyukang]

you don't have dir src/doc/book/redirects?

[onur-ozkan]

it doesn't exists until fetching the submodule

[RalfJung]

Seems like I don't have that submodule checked out; I don't work on the book so it's not needed (and I keep the checked-out submodules to a minimum to avoid having to update them all the time).

[chenyukang]

I reproduced it in local, without

submodules = false

in config.toml and don't update submodule manually, x.py test will fail like this.

[onur-ozkan]

I can fix it tomorrow, feel free to work on it if you want to fix it earlier.

[chenyukang]

@ozkanonur by the way, do you know why we use those panic! for compiletest here?

rust/src/tools/compiletest/src/lib.rs

Line 165 in 548978f

panic!()

[onur-ozkan]

@ozkanonur by the way, do you know why we use those panic! for compiletest here?

rust/src/tools/compiletest/src/lib.rs

Line 165 in 548978f

panic!()

Not particularly, I am not so familiar with the compiletest code. Probably we panic there so the compiletest runner process can handle(being able to understand compiletest was failed) that.

[onur-ozkan]

The fix shouldn't make any changes on the compiletest side. We just need to avoid using submodules(if doesn't exists) instead of assuming that they exist.

[chenyukang]

Yes, that's another issue, I just met it today.