w: Fix formatting of command line value

[alxndrv]

The command line value exposed by the kernel in /proc/<pid>/cmdline represents whitespace as NULL bytes.
Using read_to_string() on this directly will cause the output to join all parts of the original command into one long string.
This PR adds a very simple fix to avoid this.

Before:

$ cargo run -q w
 22:16:22 up  2:00,  1 user,  load average: 0.32, 0.37, 0.38
USER     TTY       LOGIN@   IDLE   JCPU   PCPU  WHAT
alx      tty1      20:16    1:59m  0.00   0     /bin/sh/usr/bin/niri-session-l

After:

$ cargo run -q w
 22:17:02 up  2:00,  1 user,  load average: 0.67, 0.46, 0.41
USER     TTY       LOGIN@   IDLE   JCPU   PCPU  WHAT
alx      tty1      20:16    2:00m  0.00   0     /bin/sh /usr/bin/niri-session -l 

[codecov]

Codecov Report

❌ Patch coverage is 0% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 19.19%. Comparing base (ee98c0a) to head (45fbba4).
⚠️ Report is 6 commits behind head on main.

Files with missing lines	Patch %	Lines
src/uu/w/src/w.rs	0.00%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #666   +/-   ##
=======================================
  Coverage   19.19%   19.19%           
=======================================
  Files          68       68           
  Lines       10135    10135           
  Branches      545      545           
=======================================
  Hits         1945     1945           
  Misses       8190     8190           

Flag	Coverage Δ
macos_latest	6.74% <ø> (ø)
ubuntu_latest	19.54% <0.00%> (ø)
windows_latest	0.02% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Bluemangoo]

Can you fix the tests?

[alxndrv]

Tests are fixed now, but I'm a bit confused by the coverage check. Is there anything I can do to make it come up as successful?

[Krysztal112233]

I feel the current test coverage is not very important, and it's already sufficient :)

[Krysztal112233]

Could you tell me why it needs to be changed to this?

[alxndrv]

The original implementation was basically a 1:1 copy of the code that's inside the fetch_cmdline function (both just build a Path from a PID and read_to_string it).
So previously, we had a test where both sides of the assert_eq are results of basically identical code. This kind of approach guarantees that things "work", but does not test for correctness. Both sides could be outputting the same incorrect result, but the test would pass. You can probably see how that's potentially problematic 😅

Since we're already using std::process::id() to figure out the PID of the process, I thought it would make more sense to compare the result of our fetch_cmdline with a "known correct" value which in this case would be std::env::args(). Hope my explanation makes sense 😅

[Krysztal112233]

Can you add a comment for the fetch_cmdline function? I noticed that the man page mentions this point after checking it, so you could reference the man page there:

https://man7.org/linux/man-pages/man5/proc_pid_cmdline.5.html

[Krysztal112233]

Great work, I think this work can inspire us to consider similar issues elsewhere :)

[alxndrv]

@Krysztal112233 Thanks for the thorough review. My Rust is... rusty, so these pointers are very much needed.
@Bluemangoo Feel free to hold off on the review until I address all of these comments 🙂