blaind / hstrace Goto Github PK

Waiting for nix-rust/nix#1422

After that,

• update nix crate to latest
• fix the ptrace::trace call

This is an issue for tracking implementation of syscalls

This is a thread for discussing the API & usage of hstrace -library for producing MVP of it.
See https://github.com/blaind/hstrace/blob/master/docs/01_hstrace_plan.md for the plan.
Any features missing, any suggestions for the API? What specific use cases would you have for such a library?

This is a build based off of #13 but I didn't think it would break things this badly. Any idea how to solve this one?

thread 'build_zsh' panicked at 'primitive conversion', /home/palfrey/.cargo/git/checkouts/hstrace-de21f6094a22c832/6f7a453/src/call/mod.rs:29:1
stack backtrace:
   0:     0x55a9180c334c - <unknown>
   1:     0x55a9180ede5c - <unknown>
   2:     0x55a9180bb1a8 - <unknown>
   3:     0x55a9180c5927 - <unknown>
   4:     0x55a9180c55df - <unknown>
   5:     0x55a9180c618b - <unknown>
   6:     0x55a9180c5e77 - <unknown>
   7:     0x55a9180c3814 - <unknown>
   8:     0x55a9180c5b89 - <unknown>
   9:     0x55a9180ec203 - <unknown>
  10:     0x55a9180ec0c1 - <unknown>
  11:     0x55a9180ec06b - <unknown>
  12:     0x55a9180ebdb6 - <unknown>

(lots more frames with <unknown>)

Hey, I'm looking into using hstrace for a utility im working on but strugling tracing both parent and child processes. It seems like coupling pid with syscall in the crossbeam_chanel as a touple would keep things clearer when collecting calls from parent and child processes. You could simply forward each child into a central channel or chain iterators into a single match.

Is there an easier way to do this?

Please report unknown error codes to this issue