cargo valgrind halts and catches fire even though I have valgrind installed?

[workingjubilee]

$ RUST_BACKTRACE=full cargo valgrind
Oooops. cargo valgrind unexpectedly crashed. This is a bug!

This is an error in this program, which should be fixed. If you can, please submit a bug report at

    https://github.com/jfrimmel/cargo-valgrind/issues/new/choose

To make fixing the error more easy, please provide the information below as well as additional information on which project the error occurred or how to reproduce it.

cargo-valgrind: version 2.2.1
thread 'main' panicked at /home/jubilee/.cargo/registry/src/index.crates.io-6f17d22bba15001f/cargo-valgrind-2.2.1/src/main.rs:127:30:
Could not execute subcommand: Os { code: 2, kind: NotFound, message: "No such file or directory" }
stack backtrace:
   0:     0x58db00f99cd6 - <std::sys::backtrace::BacktraceLock::print::DisplayBacktrace as core::fmt::Display>::fmt::h9dff740d6699a4a3
   1:     0x58db00fbbd13 - core::fmt::write::h405f94e1289aa138
   2:     0x58db00f7c49f - std::io::Write::write_fmt::h927ba1da1a503d38
   3:     0x58db00f99b73 - std::sys::backtrace::BacktraceLock::print::hb6c2f528a894346c
   4:     0x58db00f8b38e - std::panicking::default_hook::{{closure}}::h55c73f2fbe5ce7fc
   5:     0x58db00f8b1a3 - std::panicking::default_hook::had1197a6b487ba38
   6:     0x58db00f4f8c7 - cargo_valgrind::panic::replace_hook::{{closure}}::h8ef61ae13f909380
   7:     0x58db00f8b90f - std::panicking::rust_panic_with_hook::h1b43f6801257ba51
   8:     0x58db00f9a09a - std::panicking::begin_panic_handler::{{closure}}::h3e6710bf0274d6f5
   9:     0x58db00f99ed9 - std::sys::backtrace::__rust_end_short_backtrace::h54384f38f0e8ee43
  10:     0x58db00f8b46c - rust_begin_unwind
  11:     0x58db00fbdc30 - core::panicking::panic_fmt::hd7c9cbd134eec45f
  12:     0x58db00fc75d6 - core::result::unwrap_failed::h752d82d4204beb8e
  13:     0x58db00f3fdeb - cargo_valgrind::main::he56d4d7302f54c63
  14:     0x58db00f45e73 - std::sys::backtrace::__rust_begin_short_backtrace::hca6489dccd0feffb
  15:     0x58db00f4a879 - std::rt::lang_start::{{closure}}::h7a7a74f1363663d6
  16:     0x58db00f895c0 - std::rt::lang_start_internal::h9c49934ec68a8718
  17:     0x58db00f400f5 - main
  18:     0x7ccd2a5ece08 - <unknown>
  19:     0x7ccd2a5ececc - __libc_start_main
  20:     0x58db00f32755 - _start
  21:                0x0 - <unknown>
$ valgrind --version
valgrind-3.24.0

[workingjubilee]

And no, cargo valgrind run does not seem meaningfully different.

[jfrimmel]

Thanks for your report. Looks kind of funny, since this implies, that the tool cannot invoke itself as the driver?

cargo-valgrind/src/main.rs

Line 127 in de70356

if !driver::driver().expect("Could not execute subcommand") {

This suggests, that the issue is entirely unrelated to the valgrind-binary itself (the program crashes before even running valgrind).

Unfortunately, I cannot reproduce locally, so I have some questions:

1. On which distribution are you running this?
2. The stack trace suggest, that you run the tool installed via crates.io, is this correct?
3. Which Rust version is used?

[workingjubilee]

The stack trace suggest, that you run the tool installed via crates.io, is this correct?

Yes.

$ rustc --version --verbose
rustc 1.83.0-nightly (f7c53a2df 2024-11-25)
binary: rustc
commit-hash: f7c53a2df0aea1187879a37653480b31afb11417
commit-date: 2024-11-25
host: x86_64-unknown-linux-gnu
release: 1.83.0-nightly
LLVM version: 19.1.1

[workingjubilee]

I use Arch Linux.

$ which cargo-valgrind
/home/jubilee/.cargo/bin/cargo-valgrind

[jfrimmel]

Hm, I've tried for a bit, but I cannot reproduce. I also use Arch Linux, hence our setup should be similar enough. Nevertheless, it works for me (a really bad sentence, I know):

$ cargo +nightly-2024-11-25 install cargo-valgrind
$ which cargo-valgrind
/home/jfrimmel/.cargo/bin/cargo-valgrind
$ cd tests/ffi-bug/
$ cargo +nightly-2024-11-25 valgrind run --example no-leak -q
Hello, world!
$ cargo +stable valgrind run --example no-leak -q
Hello, world!
       Error leaked 56 B in 1 block
        Info stack trace (user code at the bottom)
             <snip>
     Summary Leaked 56 B total (0 other errors)
$ valgrind --version
valgrind-3.24.0

I try a different setup, but would you mind running strace on it to see, where the the problem is located? Alternatively I can provide a branch with additional logging to narrow down the problem.

[jfrimmel]

Hey @workingjubilee, do you mind to try the release v2.3.0 if the behavior is still present?