-rwxr-xr-x test/test.sh

---

 #!/usr/bin/env bash
 : '
 
 Testing gitja
 =============
 
 Running `make test` or `./test/test.sh` from the root project folder will run
 this script, which tests gitja.
 
 Beside this file is the config file that gitja uses. It runs using the
 templates in this folder and outputs into this folder. Then the output is
 compared with expected results.
 
 Hopefully this will do until a "proper" test setup is needed.
 
 '
 
 set -u
 
 if [[ $(basename "$PWD") != gitja ]]
 then
     echo "The test script should be run from the gitja root project folder."
     exit 1
 fi
 
 let errors="0"
 EXPECTED="test/expected"
 RESULT="test/result"
 rm -fr "$RESULT"
 
 declare -a TESTS
 TESTS=(
     "index.html"  # Index scope
     "link.html"  # Symbolic link at top level
     "style.css"  # Static file at top level
     "static/a_nice_file"  # Static folder at top level
     "gitja/index.html"  # Repo scope
     "gitja/commit/0292014748caae952bbc8dd6225680d83c0a5135.html"  # A commit
     "gitja/blob/test.templates.style.css.html"  # A plain text file
     "so_called_binary_file"  # Static file at top level that git considers binary
     "gitja/blob/test.templates.so_called_binary_file.html"  # The binary file
     "gitja/blob/github.FUNDING.yml.html"  # HREF drops leading period
     "gitja/log.html"  # Symbolic link inside repo/
     "gitja/static/another_file"  # Static folder inside repo/
 )
 
 stack run -- -c test/config.dhall -q || exit 1
 
 for test in "${TESTS[@]}"
 do
     diff --text --strip-trailing-cr \
 	"$EXPECTED/$test" "$RESULT/$test" || {
 	    let errors+=1
 	    echo "^ ERRORED ON: $test"
 	}
 done
 
 test -f "$RESULT/title.html.include" && {
     let errors+=1
     echo "title.html.include found in results"
 }
 
 case "$errors" in
     0)
 	echo "All good!"
 	rm -r "$RESULT"
 	;;
     *)
 	echo "Had this many errors: $errors"
 	;;
 esac
 
 exit $errors