fix: handle errors when decoding bytes from subprocesses

[tigarmo]

Since these bytes come from the output of unknown processes, there is no guarantee that they are valid utf-8 text. We also have no way to know what encoding they have (if any), so the best we can do is replace the invalid bytes with the usual unicode marker for that.

Fixes #221

• [ ] Have you followed the guidelines for contributing?
• [ ] Have you signed the CLA?
• [ ] Have you successfully run tox?

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Comparison is base (8b61db8) 94.81% compared to head (a5eacf6) 94.81%.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #222 +/- ## ======================================= Coverage 94.81% 94.81% ======================================= Files 7 7 Lines 1080 1080 Branches 200 200 ======================================= Hits 1024 1024 Misses 53 53 Partials 3 3 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[tigarmo]

Confirmed manually that this works in Snapcraft:

Building my-part                                                                                                                                                                            
:: + echo -e 'hi \x96 bye'                                                                                                                                                                  
:: hi � bye                                                                                                                                                                                 
Staging my-part