You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@ignite.apache.org by "Pavel Tupitsyn (Jira)" <ji...@apache.org> on 2025/12/19 07:57:00 UTC

[jira] [Commented] (IGNITE-27361) Reintroduce check for POJOs without default constructors in MapperBuilder

    [ https://issues.apache.org/jira/browse/IGNITE-27361?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18046515#comment-18046515 ] 

Pavel Tupitsyn commented on IGNITE-27361:
-----------------------------------------

https://github.com/apache/ignite-3/commit/63c8f7f31b180ef3004d60f13dbbf7ca66ac509c

> Reintroduce check for POJOs without default constructors in MapperBuilder
> -------------------------------------------------------------------------
>
>                 Key: IGNITE-27361
>                 URL: https://issues.apache.org/jira/browse/IGNITE-27361
>             Project: Ignite
>          Issue Type: Improvement
>          Components: thin clients ai3
>    Affects Versions: 3.2
>            Reporter: Tiago Marques Godinho
>            Assignee: Tiago Marques Godinho
>            Priority: Major
>              Labels: ignite-3
>          Time Spent: 50m
>  Remaining Estimate: 0h
>
> In IGNITE-26985 the check for requiring default constructors in POJOs was removed.
> Previously, the check would produce an IllegalArgumentException for POJOs without default constructors in MapperBuilder(MyPojo.class) calls. In the new version, the user will not receive any feedback when creating the mapper, the error will be delayed until the first operation.
> I'd prefer having the check in the Mapper constructor as previously, but what I really dislike, is that we have a bunch of checks there (NativelySupported, Arrays, Enums etc) but not the check for default constructors.
> It's also true that we don't explicitly say in the docs that these preconditions will be checked, but sort of seems like an inconsistent behaviour to me anyway. Specially because we had it, and now we removed it, for no obvious reason (at least we should able check if the class is a record).
> Notes:
> * PR Link: https://github.com/apache/ignite-3/commit/958f1ccfdd9abca26b3895ae27ed81fa7520773d#diff-9265c221dc86711da6417d66dc81bcf9e923c31dda0b7e2161c0341e208f6244
>  
>  
>  



--
This message was sent by Atlassian Jira
(v8.20.10#820010)