27 Oct
2020
27 Oct
'20
5:42 p.m.
On 10/27/20 8:20 PM, Derek Lesho wrote:
On 10/27/20 11:53 AM, Nikolay Sivov wrote:
Is there any evidence that it works this way?
No, but I doubt that any application relies on ReadSample blocking forever when RequestSample fails. The only reason I put it in was so the mfreadwrite tests don't fail on patch 4 before path 5. Yes, I could merge the patches but they are already fairly large and implement different functionality.
If you want to change this you'll have to show with a dummy source that it does happen. If you want to skip tests differently, simply return failure from Start() in earlier patch, then add RequestSample(), then fix return value for Start(), if there is no other way.