[
https://issues.apache.org/jira/browse/OPENJPA-2760?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Krzysztof Nienartowicz updated OPENJPA-2760:
--------------------------------------------
Description:
We are trying to circumvent N+1 query problem by making cache more 1-m
relations aware. We have a working solution, but it would be much more elegant
if the datacache classes were much more expandable, i.e. is
DataCacheManagerImpl very private and this with a specialized code in the
AbstractDataCache that is supposed to be the base for cache expandability does
not help, i.e. we have there:
...
*if*(manager *instanceof* DataCacheManagerImpl){
...
Could you please refactor so all cache fields to be protected, especially in
this class? We can send a patch as well.
Once we have a working solution for N+1 problem we can share it, it looks
promising, but we are still unsure about certain corner cases.
Thanks
was:
We are trying to circumvent N+1 query problem by making cache more 1-m
relations aware. We have a working solution, but it would be much more elegant
if the datacache classes were much more expandable, i.e. is
DataCacheManagerImpl very private and this with a specialized code in the
AbstractDataCache that is supposed to be the base for cache expandability does
not help (i.e. we have there:
...
*if*(manager *instanceof* DataCacheManagerImpl){
...
Could you please refactor so all cache fields to be protected, esp in this
class? We can send a patch as well.
Once we have a working solution for N+1 problem we can share it, it looks
promising, but we are still unsure about certain corner cases.
Thanks
> Cache classes, esp DataCacheManagerImpl should be less private
> --------------------------------------------------------------
>
> Key: OPENJPA-2760
> URL: https://issues.apache.org/jira/browse/OPENJPA-2760
> Project: OpenJPA
> Issue Type: Improvement
> Components: datacache
> Reporter: Krzysztof Nienartowicz
> Priority: Minor
>
> We are trying to circumvent N+1 query problem by making cache more 1-m
> relations aware. We have a working solution, but it would be much more
> elegant if the datacache classes were much more expandable, i.e. is
> DataCacheManagerImpl very private and this with a specialized code in the
> AbstractDataCache that is supposed to be the base for cache expandability
> does not help, i.e. we have there:
> ...
> *if*(manager *instanceof* DataCacheManagerImpl){
> ...
>
> Could you please refactor so all cache fields to be protected, especially in
> this class? We can send a patch as well.
> Once we have a working solution for N+1 problem we can share it, it looks
> promising, but we are still unsure about certain corner cases.
> Thanks
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
