Challenge 3: Spot the Swallowed Exit Code — Possible Solution
====================================================================

THE PROBLEM
-------------
run: npm test | tee test-output.log

WHY THIS CAN REPORT SUCCESS EVEN WHEN TESTS FAILED
--------------------------------------------------------
In a standard shell pipeline (command1 | command2), the EXIT CODE of
the whole pipeline is, by default, the exit code of the LAST command
in the pipe — not the first one. Here, the last command is tee, and
tee's job is simply to copy its input to both stdout and a file
(test-output.log) — tee itself almost always succeeds and exits with
code 0, REGARDLESS of whether npm test (the first command, piped INTO
tee) actually failed with a non-zero exit code of its own.

This means: even if npm test genuinely fails (returns exit code 1 or
higher because real tests failed), the overall run: step's reported
exit code is tee's exit code (0, success) — not npm test's real exit
code. GitHub Actions (or any CI system) sees a 0 exit code and marks
the step, and therefore the whole job, as PASSED — silently defeating
the entire "fail the build on test failure" mechanism this chapter
describes as the actual enforcement behind CI, even though the tests
genuinely did fail.

WHY THIS IS THE EXACT KIND OF ISSUE THE CHAPTER WARNS ABOUT
------------------------------------------------------------------
This is precisely the chapter's own warning about "swallowing/hiding a
test command's real exit code" via "piping test output through another
command that always exits 0" — tee is a completely ordinary, everyday
command developers add for a reasonable-seeming purpose (saving test
output to a log file for later inspection), with no obvious intent to
break anything — which is exactly why this mistake is so easy to make
and so easy to miss in a code review.

THE FIX
--------
run: |
  set -o pipefail
  npm test | tee test-output.log

Adding "set -o pipefail" (a bash shell option) changes how the
pipeline's overall exit code is determined: instead of always using
the LAST command's exit code, it uses the exit code of the FIRST
command in the pipe that actually fails (or 0 if none fail). With this
option set, if npm test fails, the pipeline's reported exit code
correctly reflects that failure, regardless of tee's own (always
successful) exit code — restoring the "fail the build on test failure"
guarantee while still keeping the convenience of writing test output to
a log file.