[9:49]<hannesw> the patch for the rhino snapshot in helma-ng just got significantly smaller: [9:49]<hannesw> http://dev.helma.org/trac/helma/changeset?new=helma-ng%2Ftrunk%2Flib%2Frhino-patch.diff%409725&old=helma-ng%2Ftrunk%2Flib%2Frhino-patch.diff%409373 [11:31]<simono> bug 664, one more thing i don't like about oracle [14:20]<hannesw> simono: when is this triggered? in HopObject.list(start, length)? or in HopObject.prefetch()? [14:22]<simono> in list(), i don't call prefetch [14:22]<hannesw> ok [14:23]<hannesw> i could use offset/limit queries here... the idea to do a id in (...) is to don't fetch objects already in cache. [14:24]<hannesw> but maybe that's just stupid, and the db is happier if i just do an offset/limit query. [14:27]<simono> okay. tom, our oracle expert, told me that IN with that many ids is not good for performance and in such cases a temp table would be appropriate [14:28]<simono> also a quickfix might be to do "SELECT ... id IN (1,2,3) OR id IN (4,5,6)" ... this is probably even worse performance wise [14:33]<hannesw> thanks for the info. [14:33]<hannesw> i'll probably switch to offset/limit queries anyway.