[infinispan-dev] Stuck on old configuration API
Sanne Grinovero
sanne at infinispan.org
Tue Mar 6 10:51:49 EST 2012
Great, thank you so much!
My fault, shouldn't explore unknown APIs at late night.
On 6 March 2012 11:22, Pete Muir <pmuir at redhat.com> wrote:
> I'm not sure what you are trying to do? You can't alter the Transport or TransactionManagerLookup at runtime, you must set them before the cache is created.
>
> If you need this functionality, and it's technically supported by Infinispan, then we can easily add a mutator for these items on the GlobalConfiguration object.
>
> BTW thanks for reminding me to clean up the adaptor stuff for 6.
>
> On 6 Mar 2012, at 00:10, Sanne Grinovero wrote:
>
>> Hi,
>> I'm confused on the new configuration API; while it's definitely an
>> improvement to understand the XML parsing, it seems functionality is
>> reduced too much.
>>
>> Previously I could set the instance of an existing (already
>> instantiated) service, especially Transport and
>> TransactionManagerLookup which I need to override/reuse instances.
>>
>> Now if I look into the .transport() section I see no setters at all,
>> unless I chain a *second* .transport() which reveals some unintended
>> public details [and looks awkward]:
>>
>> cacheManager.getCacheManagerConfiguration().transport().transport().setConfiguration(
>> GlobalConfiguration gc );
>>
>> but no way to override the instance.
>>
>> Regarding the TransactionManagerLookup, I just realized that this is
>> the third rewrite of the whole configuration read/build and that it's
>> now living in flux between THREE generations in the same codebase,
>> with adapters from new, to fluent (2nd gen), to "legacy" which I
>> happen to still need as even using 2nd gen it delegates to it.
>>
>> I had added a test to Infinispan core to make sure I would not end up
>> stuck on this, but all my tests are using the "old" 2nd generation API
>> from Infinispan 5.0 and were not migrated forward, so they didn't
>> highlight this problem.
>>
>> I've secured a written authorization to shoot if someone attempts to
>> remove one of those deprecated methods.. sorry but now that 5.1 was
>> released it's too late to cleanup.
>> _______________________________________________
>> infinispan-dev mailing list
>> infinispan-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>
>
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
More information about the infinispan-dev
mailing list