Closed jkarder closed 3 years ago
Merging #966 (0e0c470) into master (10a263b) will decrease coverage by
0.18%
. The diff coverage is40.74%
.
@@ Coverage Diff @@
## master #966 +/- ##
==========================================
- Coverage 66.10% 65.92% -0.19%
==========================================
Files 148 148
Lines 9037 9059 +22
Branches 1486 1495 +9
==========================================
- Hits 5974 5972 -2
- Misses 2461 2472 +11
- Partials 602 615 +13
Impacted Files | Coverage Δ | |
---|---|---|
src/NetMQ/AsyncReceiveExtensions.cs | 39.60% <40.74%> (+0.36%) |
:arrow_up: |
src/NetMQ/NetMQConfig.cs | 45.09% <0.00%> (-9.81%) |
:arrow_down: |
src/NetMQ/Core/SessionBase.cs | 72.38% <0.00%> (-7.47%) |
:arrow_down: |
src/NetMQ/Core/Patterns/Utils/FairQueueing.cs | 93.33% <0.00%> (-2.23%) |
:arrow_down: |
src/NetMQ/Core/Patterns/Utils/LoadBalancer.cs | 68.42% <0.00%> (-1.76%) |
:arrow_down: |
src/NetMQ/Core/Transports/StreamEngine.cs | 63.51% <0.00%> (ø) |
|
src/NetMQ/Core/SocketBase.cs | 71.19% <0.00%> (+0.35%) |
:arrow_up: |
src/NetMQ/Core/Patterns/Dish.cs | 53.84% <0.00%> (+3.41%) |
:arrow_up: |
The same treatment is needed in ReceiveFrameStringAsync
. I don't know why SkipFrameAsync
doesn't support cancellation, but you could add it there too as the pattern is similar.
Also, it's worth checking token.CanBeCanceled
, as default(CancellationToken)
will never be cancelled and so there's no point registering. For many users, no cancellation will be required, so we can avoid the callback registration overhead completely.
LGTM
The tests succeed locally on my Windows 10 installation for netcoreapp2.1, netcoreapp3.1, and net47. Maybe the build server had some hickups, because the failing test is a timeout.
Thanks for your approvals. The tests succeed here as well (Windows 10.0.19041).
netcoreapp2.1:
Test Run Successful.
Total tests: 256
Passed: 240
Skipped: 16
Total time: 37.3524 Seconds
netcoreapp3.1:
Test Run Successful.
Total tests: 258
Passed: 242
Skipped: 16
Total time: 38.2665 Seconds
net47:
Test Run Successful.
Total tests: 259
Passed: 243
Skipped: 16
Total time: 39.0455 Seconds
BTW, would you like me to revert the whitespace changes I accidentally committed in d297a27?
Re-running jobs to see if that helps.
BTW, would you like me to revert the whitespace changes I accidentally committed in d297a27?
Sure, but only if it's convenient.
Yep, just needed a kick. I'll merge as-is. Thanks very much!
This PR updates ReceiveFrameBytesAsync to detach competing listeners before finishing the TaskCompletionSource and call
TrySetResult
andTrySetCanceled
instead ofSetResult
andSetCanceled
.Fixes #965.