[infinispan-issues] [JBoss JIRA] (ISPN-8798) ByteString places too strict a constraint on cache name length

[Tristan Tarrant (JIRA)]

    [ https://issues.jboss.org/browse/ISPN-8798?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13536632#comment-13536632 ] 

Tristan Tarrant commented on ISPN-8798:
---------------------------------------

I'd rather see the use of hashes instead, to represent the long identifiers. The cache name is included in every RPC, and the bigger it is, the larger the impact on performance.

> ByteString places too strict a constraint on cache name length
> --------------------------------------------------------------
>
>                 Key: ISPN-8798
>                 URL: https://issues.jboss.org/browse/ISPN-8798
>             Project: Infinispan
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 9.1.5.Final, 9.2.0.CR2
>            Reporter: Paul Ferraro
>            Assignee: Paul Ferraro
>            Priority: Blocker
>             Fix For: 9.2.0.CR3, 9.1.6.Final
>
>
> The hibernate 2nd level cache code can easily generate cache names that exceed the byte length restriction enforced by ByteString.  Cache names are generated using the fully qualified class name of the entity, which are user defined, and in the case of WildFly, also includes the deployment name, which can be particularly long for EARs.
> At the very least, we should use an unsigned byte which doubles the allowed size.  Why not also consider using  https://github.com/infinispan/infinispan/blob/master/commons/src/main/java/org/infinispan/commons/io/UnsignedNumeric.java to marshal/unmarshal the size of the byte[] containing the name?  At worst, this costs an extra byte, but solves the size problem for even sizes larger than 255.



--
This message was sent by Atlassian JIRA
(v7.5.0#75005)