Mutation tests only

[JekaMas]

Description

As far as consensus is extremely sensitive functionality, I believe we need a way to check our tests for usefulness and correctness. Code coverage, unfortunately, doesn't tell us about the quality of tests, either about usefulness. Even with high code coverage it's possible to write not very useful tests that always 'green', or tests that are 'green' if being executed in particular order.
For that reason the PR introduces changes:

1. As far as we run tests with t.Parallel() it makes sense to use -race flag as well. At the moment tests do have data races that'd be great to fix in separate issues.
2. To prevent incorrect global state the PR introduces -shuffle flag
3. To check that tests are useful the PR adds mutation testing https://github.com/avito-tech/go-mutesting

Changes include

• Bugfix (non-breaking change that solves an issue)
• Hotfix (change that solves an urgent issue, and requires immediate attention)
• New feature (non-breaking change that adds functionality)
• Breaking change (change that is not backwards-compatible and/or changes current functionality)

Checklist

• I have assigned this PR to myself
• I have added at least 1 reviewer
• I have added the relevant labels
• I have added sufficient documentation in code

Comments

This PR is for discussion about new tooling. At the moment data race tests and mutation tests are red. If we'd like to merge the PR, we should comment out this CI steps and make tasks to fix tests with race and a bunch of tasks to add/check tests with respect of mutation testing results.

A few examples of mutation testing:

 func (i *IBFT) handlePrepare(view *proto.View, quorum uint64) bool {
 	isValidPrepare := func(message *proto.Message) bool {
 		// Verify that the proposal hash is valid
@@ -799,7 +799,7 @@
 
 	if len(prepareMessages) < int(quorum)-1 {
 		//	quorum not reached, keep polling
-		return false
+
 	}
 
func (i *IBFT) handlePrepare(view *proto.View, quorum uint64) bool {
 	isValidPrepare := func(message *proto.Message) bool {
 		// Verify that the proposal hash is valid
@@ -852,7 +852,7 @@
 		case <-sub.SubCh:
 			if !i.handleCommit(view, quorum) {
 				//	quorum not reached, retry
-				continue
+
 			}
 

 func (i *IBFT) handleCommit(view *proto.View, quorum uint64) bool {
 	isValidCommit := func(message *proto.Message) bool {
 		var (
@@ -880,7 +880,7 @@
 	commitMessage
[result.txt](https://github.com/0xPolygon/go-ibft/files/9723038/result.txt)
s := i.messages.GetValidMessages(view, proto.MessageType_COMMIT, isValidCommit)
 	if len(commitMessages) < int(quorum) {
 		//	quorum not reached, keep polling
-		return false
+
 	}
 


 func (i *IBFT) runStates(ctx context.Context) {
 	var timeout error
 
@@ -594,7 +594,7 @@
 
 	//	is proposer
 	if !i.backend.IsProposer(msg.From, height, round) {
-		return false
+
 	}

 func (i *IBFT) runStates(ctx context.Context) {
 	var timeout error
 
@@ -615,7 +615,7 @@
 
 	//	proposal must be for round 0
 	if msg.View.Round != 0 {
-		return false
+
 	}
 
 func (i *IBFT) handlePrePrepare(view *proto.View) *proto.Message {
 	isValidPrePrepare := func(message *proto.Message) bool {
 		if view.Round == 0 {
@@ -772,7 +772,7 @@
 		case <-sub.SubCh:
 			if !i.handlePrepare(view, quorum) {
 				//	quorum of valid prepare messages not received, retry
-				continue
+
 			}
 

More results are in the file result.txt

[codecov]

Codecov Report

Merging #38 (da4fb4f) into main (12435ef) will not change coverage.
The diff coverage is n/a.

❗ Current head da4fb4f differs from pull request most recent head 428332e. Consider uploading reports for the commit 428332e to get more accurate results

@@           Coverage Diff           @@
##             main      #38   +/-   ##
=======================================
  Coverage   95.51%   95.51%           
=======================================
  Files           6        6           
  Lines        1292     1292           
=======================================
  Hits         1234     1234           
  Misses         42       42           
  Partials       16       16           

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more