[
https://issues.apache.org/jira/browse/IGNITE-5137?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Pavel Tupitsyn updated IGNITE-5137:
-----------------------------------
Description:
When exception occurs in user callback on .NET side we should serialize it
entirely, send to caller node and re-throw there (wrapping in IgniteException)
to preserve original stack trace.
We already do this for Compute (see {{UnmanagedCallbacks.ComputeJobExecute}}),
but not for ScanQuery, where only exception message gets propagated (which
complicates debugging).
was:
When exception occurs in user callback on .NET side we should serialize it
entirely, send to caller node and re-throw there (wrapping in IgniteException)
to preserve original stack trace.
We already do this for Compute, but not for ScanQuery, where only exception
message gets propagated (which complicates debugging).
> .NET: Propagate native ScanQuery exceptions
> -------------------------------------------
>
> Key: IGNITE-5137
> URL: https://issues.apache.org/jira/browse/IGNITE-5137
> Project: Ignite
> Issue Type: Improvement
> Components: platforms
> Reporter: Pavel Tupitsyn
> Labels: .NET
> Fix For: 2.1
>
>
> When exception occurs in user callback on .NET side we should serialize it
> entirely, send to caller node and re-throw there (wrapping in
> IgniteException) to preserve original stack trace.
> We already do this for Compute (see
> {{UnmanagedCallbacks.ComputeJobExecute}}), but not for ScanQuery, where only
> exception message gets propagated (which complicates debugging).
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
