[Maria-developers] WL#156 New (by Dreas): Add mysqlcheck option to directly repair

2010-10-10 Thread worklog-noreply
--- WORKLOG TASK -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- TASK...: Add mysqlcheck option to directly repair CREATION DATE..: Sun, 10 Oct 2010, 12:58 SUPERVISOR...

Re: [Maria-developers] Implementing new "group commit" API in PBXT?

2010-10-10 Thread Kristian Nielsen
Paul McCullagh writes: >> In particular this, flushing the data log (is this flush to disk?): >>if (!thread->st_dlog_buf.dlb_flush_log(TRUE, thread)) { >>ok = FALSE; >>status = XT_LOG_ENT_ABORT; >>} > > Yes, this is a flush to disk. > > This could be done in the

[Maria-developers] Performance problem: need index-based access when have only non-sargable conditions

2010-10-10 Thread Sergey Petrunya
Hi! I was asked about this performance problem: consider a query select * from tbl where tbl.key like '%foo%' Table records are big, much larger than the length of the key. LIKE condition is very selective, however its pattern starts with '%', so we can't construct range access (and they ar