Shell: Make tests use PASS/FAIL instead of exit codes

There's no guarantee that the last executed command will have a zero
exit code, and so the shell exit code may or may not be zero, even if
all the tests pass.
Also changes the `test || echo fail && exit` to
`if not test { echo fail && exit }`, since that's nicer-looking.

Userland/Shell/Tests/subshell.sh

@@ -1,5 +1,7 @@
 #/bin/sh
 
+source test-commons.inc
+
 setopt --verbose
 
 rm -rf shell-test
@@ -8,21 +10,30 @@ cd shell-test
 
     # Simple sequence (grouping)
     { echo test > testfile }
-    test "$(cat testfile)" = "test" || echo cannot write to file in subshell && exit 1
+    if not test "$(cat testfile)" = "test" {
+        fail cannot write to file in subshell
+    }
 
     # Simple sequence - many commands
     { echo test1 > testfile; echo test2 > testfile }
-    test "$(cat testfile)" = "test2" || echo cannot write to file in subshell 2 && exit 1
-
+    if not test "$(cat testfile)" = "test2" {
+        fail cannot write to file in subshell 2
+    }
 
     # Does it exit with the last exit code?
     { test -z "a" }
     exitcode=$?
-    test "$exitcode" -eq 1 || echo exits with $exitcode when it should exit with 1 && exit 1
+    if not test "$exitcode" -eq 1 {
+        fail exits with $exitcode when it should exit with 1
+    }
 
     { test -z "a" || echo test }
     exitcode=$?
-    test "$exitcode" -eq 0 || echo exits with $exitcode when it should exit with 0 && exit 1
+    if not test "$exitcode" -eq 0 {
+        fail exits with $exitcode when it should exit with 0
+    }
 
 cd ..
 rm -rf shell-test
+
+echo PASS