ThreadLock::new — use RAII for ptrace detach cleanup (#117)

[charles-dyfis-net]

Construct ThreadLock { tid } immediately after a successful ptrace::seize so that Drop becomes responsible for detach on every exit path, including ? propagation. Closes all three error paths listed in #117 with a single change.

[benfred]

Thanks for the change! I like the core fix here - but running the new test locally and I'm seeing some intermittent failures on my local ubuntu machine (fails maybe 1/8th of the time for me). The failures look like

running 5 tests
test tests::test_copy_pointer ... ok
test tests::test_copy_struct ... ok
test linux::test_parse_active_stat ... ok
test linux::test_parse_ppid_stat ... ok
test linux::test_threadlock_leak_on_waitpid_eintr ... FAILED

failures:

---- linux::test_threadlock_leak_on_waitpid_eintr stdout ----

thread 'linux::test_threadlock_leak_on_waitpid_eintr' (111640) panicked at src/linux/mod.rs:459:9:
seize leaked: first=NixError(EINTR), second=NixError(EPERM)
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace


failures:
    linux::test_threadlock_leak_on_waitpid_eintr

test result: FAILED. 4 passed; 1 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.04s

error: test failed, to rerun pass `--lib`

[benfred]

btw, I think the GHA build failure is unrelated to this PR and I will fix this up in a separate PR