Allow multiple expectations per test.

test/README.markdown

@@ -26,6 +26,5 @@ call Setup()
 ### Limitations
 
 - Currently tests are done by writing out a file and comparing it to a known good one.
-- Only 1 file per test is supported.
 - There's no support for assertions within the testcase code.
 

test/test.sh

@@ -2,24 +2,20 @@
 
 # TODO: exit with non-zero status code when tests fail.
 
-canonicalise_test_name() {
-  local testname=$1                                              # testFoo.vim
-  name=${testname%.*}                                            # testFoo
-  name=${name:4}                                                 # Foo
-  name="$(tr '[:upper:]' '[:lower:]' <<< ${name:0:1})${name:1}"  # foo
-}
-
 
 rm -f *.out
 
+# Execute the tests.
+for testcase in test*.vim; do
+  vim -N -u NONE -S $testcase -c 'quit!'
+done
+
+# Verify the results.
 count_ok=0
 count_fail=0
 
-for testcase in test*.vim; do
+for expected in *.ok; do
-  vim -N -u NONE -S $testcase -c 'quit!'
+  name=${expected%.*}
-
-  canonicalise_test_name $testcase
-  expected=$name.ok
   actual=$name.out
 
   if diff $expected $actual; then
@@ -31,6 +27,7 @@ for testcase in test*.vim; do
   fi
 done
 
+git reset HEAD fixture.txt > /dev/null
 git checkout fixture.txt
 
 echo