[jboss-jira] [JBoss JIRA] (WFLY-3612) Implement missing transformers for new bean validation subsystem

[Brian Stansberry (JIRA)]

     [ https://issues.jboss.org/browse/WFLY-3612?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Brian Stansberry updated WFLY-3612:
-----------------------------------
    Description: 
The initial work on WFLY-3551 will be merged without needed transformers, because the transformer support we have won't allow the needed transformation. This JIRA is a reminder to add it once the required work in WFCORE-291 is done.

>From my comment on the PR for the WFLY-3551 work:

There needs to be a discard of this resource if a legacy slave is involved. The presence of the resource means BV is wanted, and by default the legacy slave will have BV, so if this resource exists, the legacy slave is compatible. The resource just needs to be discarded so the legacy slave doesn't fail when it sees resources/ops for this subsystem it can't understand.

If this subsystem is omitted in a config with ee, the legacy slave will be out of sync with newer ones running the same profile, because the old ones will have BV and the new ones won't. But

1) it's a corner case
2) it may not matter anyway in terms of behavior of an app
3) the transformer logic we have doesn't allow us to do anything about it anyway.


Re: that last part, when doing this work double check on skipping that, in case the WFLY-3563 work makes it real easy. Unlikely though.

  was:
The initial work on WFLY-3551 will be merged without needed transformers, because the transformer support we have won't allow the needed transformation. This JIRA is a reminder to add it once the required work in WFLY-3563 is done.

>From my comment on the PR for the WFLY-3551 work:

There needs to be a discard of this resource if a legacy slave is involved. The presence of the resource means BV is wanted, and by default the legacy slave will have BV, so if this resource exists, the legacy slave is compatible. The resource just needs to be discarded so the legacy slave doesn't fail when it sees resources/ops for this subsystem it can't understand.

If this subsystem is omitted in a config with ee, the legacy slave will be out of sync with newer ones running the same profile, because the old ones will have BV and the new ones won't. But

1) it's a corner case
2) it may not matter anyway in terms of behavior of an app
3) the transformer logic we have doesn't allow us to do anything about it anyway.


Re: that last part, when doing this work double check on skipping that, in case the WFLY-3563 work makes it real easy. Unlikely though.



> Implement missing transformers for new bean validation subsystem
> ----------------------------------------------------------------
>
>                 Key: WFLY-3612
>                 URL: https://issues.jboss.org/browse/WFLY-3612
>             Project: WildFly
>          Issue Type: Bug
>          Components: Domain Management, EE
>            Reporter: Brian Stansberry
>            Assignee: Eduardo Martins
>            Priority: Blocker
>             Fix For: 10.0.0.Final
>
>
> The initial work on WFLY-3551 will be merged without needed transformers, because the transformer support we have won't allow the needed transformation. This JIRA is a reminder to add it once the required work in WFCORE-291 is done.
> From my comment on the PR for the WFLY-3551 work:
> There needs to be a discard of this resource if a legacy slave is involved. The presence of the resource means BV is wanted, and by default the legacy slave will have BV, so if this resource exists, the legacy slave is compatible. The resource just needs to be discarded so the legacy slave doesn't fail when it sees resources/ops for this subsystem it can't understand.
> If this subsystem is omitted in a config with ee, the legacy slave will be out of sync with newer ones running the same profile, because the old ones will have BV and the new ones won't. But
> 1) it's a corner case
> 2) it may not matter anyway in terms of behavior of an app
> 3) the transformer logic we have doesn't allow us to do anything about it anyway.
> Re: that last part, when doing this work double check on skipping that, in case the WFLY-3563 work makes it real easy. Unlikely though.



--
This message was sent by Atlassian JIRA
(v6.4.11#64026)