Test main loop #6
Conversation
| ingester.Run(context.Background(), &wg) | ||
| go func() { | ||
| defer wg.Done() | ||
| err := ingester.Run(context.Background(), cfg.BlockHeight, 0 /* maxCount */) |
There was a problem hiding this comment.
I didn't realize /* */ was valid in Go!
| @@ -12,7 +11,8 @@ import ( | |||
| ) | |||
|
|
|||
| type Ingester interface { | |||
| Run(ctx context.Context, wg *sync.WaitGroup) error | |||
| // Run starts the ingester and blocks until the context is cancelled or maxCount blocks are ingested | |||
| Run(ctx context.Context, startBlockNumber, maxCount int64) error | |||
There was a problem hiding this comment.
maxCount 🤔 EDIT: Nevermind, tests made it clear why we want this
| inFlightChan := make(chan models.RPCBlock, i.cfg.MaxBatchSize) | ||
| defer close(inFlightChan) | ||
|
|
||
| var err error | ||
|
|
||
| startBlockNumber := i.cfg.StartBlockHeight | ||
| if startBlockNumber <= 0 { | ||
| if startBlockNumber < 0 { |
There was a problem hiding this comment.
Good! We do indeed want block 0
| "github.com/stretchr/testify/require" | ||
| ) | ||
|
|
||
| func TestBlockConsumptionLoopErrors(t *testing.T) { |
There was a problem hiding this comment.
It's not immediately clear to me what this test is testing. Is it testing that we exit the function cleanly under these scenarios + what blocks have been ingested? Or am I misunderstanding?
There was a problem hiding this comment.
this test is bogus right now, because I realized (while trying to get this test any decent) that the logic I created is just bad...
The problem is:
- we should be able to cancel execution (using the context) and have a good understanding if the execution was smooth, flaky, or we were under a busted scenario.
- for example, with everything broken, I think there should be an early return and just abort execution ? or should we just stay alive and keep trying?
- if execution is flaky, we should still consume blocks, but skip some, right ?
in short, this test just shows we're a bit confused on what to expect from the test.
| } | ||
| for _, testcase := range testcases { | ||
| t.Run(testcase, func(t *testing.T) { | ||
| t.Skip("not implemented") |
| } | ||
|
|
||
| func TestRunLoop(t *testing.T) { | ||
| t.Skip("not implemented") | ||
| func TestRunLoopBaseCase(t *testing.T) { |
Obvious, creating tests found bugs and forced some changes :-)
but at least we validate some basic cases