On 16 févr. 2011, at 16:19, Sanne Grinovero wrote:
Hi,
from my part I'd give priority to fieldcache and massindexer features.
The benchmark was blocked by Infinispan issues, all solved yesterday.
So now I'd like to shape the FieldCache feature:
* Implicitly using it on classname and id
+1 we can even have
* looking how it could be convenient to enable wider usage on demand
planned for 3.5 if we don't feel confident.
Having the benchmark I can now verify possible improvements on the
wikipedia index on all supported directories, and at same time proof
usefullness of massinder new fetures.
as goal for Alpha1 I'd set at least
1) implicit usage of FieldCache (HSEARCH-531)
2) HSEARCH-499 Pass a "filter" query to MassIndexing: index changed
entity based on timestamp
I hope it's fine if for now the benchmark project continues to live
outside Search, the setup scripts need ANT and won't be easy to port
into Search.
I don't think that's an issue.