e362d69972
- `test_section "..."` replaces `echo "Now we're testing ..."`
- `test_run ...` replaces `... || { ...; eval "testsfailed"; }`
- `test_run not ...` replaces `... && { ...; eval "testsfailed"; }`
`test_section` saves the output of the program and shows it only in the
case of failures.
`test_run` arranges to exit with non-zero status if a test fails.
Use `set -e` to force early exit. Conversely use `set +e` to continue
running the remaining tests when one fails -- this will be very useful
in reducing the number of CI test runs (e.g., GitHub Actions), thus
saving time and money.
This is Claude-generated code, guided by me, with minor corrections.
43 lines
974 B
Bash
Executable File
43 lines
974 B
Bash
Executable File
#!/bin/sh
|
|
#
|
|
# Inner test script for testing test-lib.sh
|
|
# This script has two test sections: one that fails and one that succeeds.
|
|
# It's run by check-test-lib which validates the output.
|
|
#
|
|
|
|
top_srcdir="@top_srcdir@"
|
|
objdir="@objdir@"
|
|
|
|
# Source test-lib.sh
|
|
. "${top_srcdir}/tests/bin/test-lib.sh"
|
|
|
|
# Create a fake messages.log in a temp directory
|
|
tmpdir="${TMPDIR:-/tmp}/test-lib-test.$$"
|
|
mkdir -p "$tmpdir"
|
|
cd "$tmpdir"
|
|
|
|
test_init messages.log
|
|
|
|
# Section 1: This one will fail
|
|
test_section "Failing test section"
|
|
|
|
# Simulate some library/KDC output in messages.log
|
|
echo "KDC: some trace output" >> messages.log
|
|
echo "KDC: more trace output" >> messages.log
|
|
|
|
# Run a command that will fail
|
|
test_run_x sh -c 'echo "command stdout"; echo "command stderr" >&2; exit 1'
|
|
|
|
# Section 2: This one will succeed
|
|
test_section "Succeeding test section"
|
|
|
|
# Run a command that succeeds
|
|
test_run sh -c 'echo "success output"; exit 0'
|
|
|
|
# Clean up
|
|
cd /
|
|
rm -rf "$tmpdir"
|
|
|
|
test_finish
|
|
exit $?
|