You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tinkerpop.apache.org by "stephen mallette (JIRA)" <ji...@apache.org> on 2015/06/17 22:14:00 UTC
[jira] [Commented] (TINKERPOP3-733) Gremlin driver deserialization
exception does not bubble up
[ https://issues.apache.org/jira/browse/TINKERPOP3-733?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14590488#comment-14590488 ]
stephen mallette commented on TINKERPOP3-733:
---------------------------------------------
So I did have tests for bad serialization that occurred server-side. Those exceptions properly report in the client.
It wasn't clear to me the other day or today when i read this that the problem is that if the client-side experiences the serialization error - i didn't have tests for that scenario. I assume that your workaround involved registering a custom {{MessageSerializer}} - if so, then that is the solution. We probably have zero documentation on that anywhere - so that's yet another issue related to this.
> Gremlin driver deserialization exception does not bubble up
> -----------------------------------------------------------
>
> Key: TINKERPOP3-733
> URL: https://issues.apache.org/jira/browse/TINKERPOP3-733
> Project: TinkerPop 3
> Issue Type: Bug
> Components: driver
> Reporter: Adrian Gonzalez
> Assignee: stephen mallette
> Fix For: 3.0.0.GA
>
>
> I ran into this issue trying to use titan 0.9.0-M2 with gremlin-server 3.0.0.M9-incubator. Titan was returning a serialized RelationIdentifier which understandably could not be deserialized on the client side using the driver. I have managed to work around this by using the IoRegistry, however the SerializationException thrown in GryoMessageSerializerV1d0.deserializeResponse() does not bubble back up to the CompletableFuture<ResultSet> returned from the Client.submitAsync() nor the CompletableFuture<List<Result> returned from ResultSet.all(). A warning is logged in DefaultChannelPipeline.exceptionCaught(), but that is as far as it goes. Or am I missing something?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)