Add notes on adding failure test cases to CONTRIBUTING

CONTRIBUTING.md

@@ -4,3 +4,11 @@ Please raise issues for discussion before opening PRs (or ideally before startin
 just to save time if its not the direction we want to take the project.
 
 For small changes (anywhere from typos to non-breaking bug fixes) going straight to a PR is fine, but not required.
+
+## Testing
+
+This library is organized so `go test -v` will test tokenization, parsing, code generation, and the validity of the output generated code. Output code is validated via some integrated tests, and by relying on `go test` itself checking that output can be compiled.
+
+If an example file fails tokenization, it can be added to `testTokenizeFiles` to track that failure and prevent regressions. Likewise for parsing, `TestReadFile` notes expected parsing output. Adding a variant that just ensures a file -can- be parsed, not ensuring a particular AST, is a welcome augmentation.
+
+For generated code, add the example file to `generated/base` and add it's name, minus `.bop`, to `genTestFiles`.

tokenize_test.go

@@ -8,7 +8,7 @@ import (
 	"testing"
 )
 
-var testFiles = []string{
+var testTokenizeFiles = []string{
 	"array_of_strings",
 	"all_consts",
 	"basic_arrays",
@@ -31,7 +31,7 @@ var testFiles = []string{
 
 func TestTokenize(t *testing.T) {
 	t.Parallel()
-	for _, filename := range testFiles {
+	for _, filename := range testTokenizeFiles {
 		filename := filename
 		t.Run(filename, func(t *testing.T) {
 			t.Parallel()
@@ -111,7 +111,7 @@ func TestTokenizeNoSemis(t *testing.T) {
 		"import":             "import semis cannot be added",
 		"msgpack_comparison": "we are naively removing semis from within comments",
 	}
-	for _, filename := range testFiles {
+	for _, filename := range testTokenizeFiles {
 		filename := filename
 		t.Run(filename, func(t *testing.T) {
 			t.Parallel()