[jira] [Closed] (BVAL-114) ApacheValidatorFactory.getValidator() caching

["Matt Benson (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/BVAL-114?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Matt Benson closed BVAL-114.
----------------------------
    Resolution: Won't Fix

BVal 2.0.0 includes a complete rewrite of BV metadata which is now cached in association to the validation factory.

> ApacheValidatorFactory.getValidator() caching
> ---------------------------------------------
>
>                 Key: BVAL-114
>                 URL: https://issues.apache.org/jira/browse/BVAL-114
>             Project: BVal
>          Issue Type: Improvement
>          Components: jsr303
>    Affects Versions: 0.4, 0.5
>            Reporter: Jarek Gawor
>            Priority: Major
>         Attachments: BVAL-114.patch
>
>
> We have a piece of code that's essentially doing:
> ValidatorFactory factory = ....;
> for (...) {
>     Foo foo = ...;
>     factory.getValidator().validate(foo);
> }
> With this pattern the Foo class is scanned for annotations on each loop iteration. The problem is that each getValidator() invocation on the same factory instance create a separate ApacheFactoryContext instance with own MetaBeanFinder. Therefore, the cache in MetaBeanFinder doesn't really get reused between loop iterations.
> One obvious solution is to use the same Validator object instance in the loop. However, in certain situations it might not be easy to change the code pattern in that way. So, I'm wondering whether getValidator() can be optimized in such a way that the MetaBeanFinder (and therefore its cache) can be reused between calls. From JavaDoc it sounds like some sort of caching might be allowed.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)