make blip tests fail with stack traces using testify #6748
Conversation
- Create TB() helper functions to make it easier to pick up the right TB function since it can get reset or hard to pick up - use require.EventuallyT instead of testing.T.Fatalf to get tracebacks when they fail. Because these functions fail anyway, removed the return of boolean found. - switch panic to require statements while the code was here
| @@ -2257,8 +2257,7 @@ func TestRevocationMessage(t *testing.T) { | |||
| assert.NoError(t, err) | |||
|
|
|||
| // Wait for doc revision to come over | |||
| _, ok := btcRunner.WaitForBlipRevMessage(btc.id, "doc", version) | |||
| require.True(t, ok) | |||
| _ = btcRunner.WaitForBlipRevMessage(btc.id, "doc", version) | |||
There was a problem hiding this comment.
Nitpick: Do we need to have the empty return values on these in this file? Could keep it in line with what you have done in most other places and not have the empty return values.
There was a problem hiding this comment.
WaitForBlipRevMessage does use return value in blip_api_delta_sync_test.go
There was a problem hiding this comment.
What I meant was you could have btcRunner.WaitForBlipRevMessage(btc.id, "doc", version) as opposed to _ = btcRunner.WaitForBlipRevMessage(btc.id, "doc", version)
It doesn't have any impact on tests so i'm good merging, was just a nitpick.
| @@ -2040,8 +2035,7 @@ func TestRemovedMessageWithAlternateAccess(t *testing.T) { | |||
|
|
|||
| err = btcRunner.StartOneshotPull(btc.id) | |||
| assert.NoError(t, err) | |||
| _, ok := btcRunner.WaitForVersion(btc.id, docID, version) | |||
| assert.True(t, ok) | |||
| _ = btcRunner.WaitForVersion(btc.id, docID, version) | |||
There was a problem hiding this comment.
Nitpick: Same as above in this file too?
There was a problem hiding this comment.
WaitForMessage return values are used in blip_api_delta_sync_test.go
* Cleanup tests - Create TB() helper functions to make it easier to pick up the right TB function since it can get reset or hard to pick up - use require.EventuallyT instead of testing.T.Fatalf to get tracebacks when they fail. Because these functions fail anyway, removed the return of boolean found. - switch panic to require statements
This is really only to help debug the CBG-3818 ticket part 1, I have a second round of refactoring just for this test to see what documents aren't coming through.
To review:
Pre-review checklist
fmt.Print,log.Print, ...)base.UD(docID),base.MD(dbName))docs/apiIntegration Tests
GSI=true,xattrs=truehttps://jenkins.sgwdev.com/job/SyncGateway-Integration/2369/