So specifically the issue is the way the test checks (ie chunk2!=chunk1) - but since streaming is very much timing/performance related it's possible two chunks could be the same - for example two spaces. I saw this erratically when testing locally. It may depend on system performance (ie tps) etc.
That makes a lot of sense why the issue would be so rare. I would vote to loosen the assertion to validate the end result (reassembly only) as well as add the stubbed tests, assuming that the stubbed tests keep the chunk1 != chunk2 assertions.
Originally posted by @jakelorocco in #567 (comment)
That makes a lot of sense why the issue would be so rare. I would vote to
loosen the assertion to validate the end result (reassembly only) as well as add the stubbed tests, assuming that the stubbed tests keep the chunk1 != chunk2 assertions.Originally posted by @jakelorocco in #567 (comment)