fix(remoteingress-core): wait for hub-to-client draining before cleanup and reliably send close frames
This commit is contained in:
@@ -1,5 +1,12 @@
|
||||
# Changelog
|
||||
|
||||
## 2026-03-16 - 4.5.5 - fix(remoteingress-core)
|
||||
wait for hub-to-client draining before cleanup and reliably send close frames
|
||||
|
||||
- switch CLOSE frame delivery on the data channel from try_send to send().await to avoid dropping it when the channel is full
|
||||
- delay stream cleanup until the hub-to-client task finishes or times out so large downstream responses continue after upload EOF
|
||||
- add a bounded 5-minute wait for download draining to prevent premature termination of asymmetric transfers such as git fetch
|
||||
|
||||
## 2026-03-15 - 4.5.4 - fix(remoteingress-core)
|
||||
preserve stream close ordering and add flow-control stall timeouts
|
||||
|
||||
|
||||
Reference in New Issue
Block a user