Yicong-Huang opened a new issue, #4799:
URL: https://github.com/apache/texera/issues/4799
### What happened?
`amber/src/main/scala/org/apache/texera/amber/engine/architecture/scheduling/config/ChannelConfig.scala::ChannelConfig.generateChannelConfigs`
uses `fromWorkerIds.zip(toWorkerIds)` for the `OneToOnePartition` arm. Scala's
`List.zip` truncates to the shorter side, so calling this method with
mismatched-length worker lists silently drops the surplus pairings instead of
raising. The other partitioning arms either compute a full cross product or
assert exact-length expectations (`SinglePartition` requires `toWorkerIds.size
== 1`); only `OneToOne` has this silent-truncation hazard.
### How to reproduce?
```scala
import org.apache.texera.amber.core.virtualidentity.ActorVirtualIdentity
import org.apache.texera.amber.core.workflow.{OneToOnePartition,
PortIdentity}
import
org.apache.texera.amber.engine.architecture.scheduling.config.ChannelConfig
val w = (1 to 3).map(i => ActorVirtualIdentity(s"w$i")).toList
val u = (1 to 2).map(i => ActorVirtualIdentity(s"u$i")).toList
val out = ChannelConfig.generateChannelConfigs(w, u, PortIdentity(0, false),
OneToOnePartition())
// out.size == 2 — w3 silently has no channel even though OneToOne implies a
1:1 contract
```
### Version
1.1.0-incubating (Pre-release/Master)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
