[infinispan-dev] wf config

[Mircea Markus]

I guess this is a question for Paul.

On Jan 29, 2014, at 12:08 PM, Ales Justin <ales.justin at gmail.com> wrote:

> I'm looking at current WildFly integration.
> 
> In CacheAdd I see this code:
> 
>        if ((lockingMode == LockingMode.OPTIMISTIC) && (isolationLevel == IsolationLevel.REPEATABLE_READ)) {
>            builder.locking().writeSkewCheck(true);
>        }
> 
> but then locking has this validation:
> 
>   public void validate() {
>      if (writeSkewCheck) {
>         if (isolationLevel != IsolationLevel.REPEATABLE_READ)
>            throw new CacheConfigurationException("Write-skew checking only allowed with REPEATABLE_READ isolation level for cache");
>         if (transaction().lockingMode != LockingMode.OPTIMISTIC)
>            throw new CacheConfigurationException("Write-skew checking only allowed with OPTIMISTIC transactions");
>         if (!versioning().enabled || versioning().scheme != VersioningScheme.SIMPLE)
>            throw new CacheConfigurationException(
>                  "Write-skew checking requires versioning to be enabled and versioning scheme 'SIMPLE' to be configured");
> 
> Yet there is no versioning handling in WF subsystem.
> (just listing what's supported)
> 
>    private void parseCacheElement(XMLExtendedStreamReader reader, Element element, ModelNode cache, List<ModelNode> operations) throws XMLStreamException {
>        switch (element) {
>            case LOCKING: {
>            case TRANSACTION: {
>            case EVICTION: {
>            case EXPIRATION: {
>            case STORE: {
>            case FILE_STORE: {
>            case STRING_KEYED_JDBC_STORE: {
>            case BINARY_KEYED_JDBC_STORE: {
>            case MIXED_KEYED_JDBC_STORE: {
>            case REMOTE_STORE: {
>            case INDEXING: {
>            default: {
>                throw ParseUtils.unexpectedElement(reader);
>    }
> 
> How do you expect the user to get pass the validation,
> where you magically enable writeSkewCheck?
> 
> -Ales
> 
> 
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev

Cheers,
-- 
Mircea Markus
Infinispan lead (www.infinispan.org)