Jorge Bastos wrote: > Ok I'm testing in my test machine, no more tests on production machine since > last situation :)
Testing is good. Doing so on a test server is even better. But don't file bugs against the state of the trunk please. You know the drill. > Where does live this "ZDB" library? > I'm searching debian packages and don't see it. You can find the upstream tarbal at http://tildeslash.com/libzdb There's also debian packages on deb http://debian.nfgd.net/debian unstable main > > >> -----Original Message----- >> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On >> Behalf Of Jorge Bastos >> Sent: sábado, 29 de Março de 2008 14:37 >> To: 'DBMail mailinglist' >> Subject: RE: [Dbmail] New threading model >> >> Oh, >> It's just that I like to try new things!! >> >> >> >> >>> -----Original Message----- >>> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On >>> Behalf Of Paul J Stevens >>> Sent: sábado, 29 de Março de 2008 14:19 >>> To: DBMail mailinglist >>> Subject: Re: [Dbmail] New threading model >>> >>> Jorge Bastos wrote: >>>> Hi Paul, >>>> >>>> I've seen that there's some work on the new threading model in the >>> trunk. >>>> Is it in a mature state for use in production, even knowing that >>> trunk Is >>>> not for production? >>> You are joking, right? It's nowhere near production ready. I've been >>> playing with it, slowly working out a workable threading pattern. >>> >>> Thread per client is what's in the trunk right now. Connection >> pooling >>> works really great. Libzdb kicks ass. For ldap it's a different >> story. >>> I've played with a singleton ldap connection which worked pretty >> good, >>> since ldap queries are fast and not used that often compared to sql. >>> I've also been experimenting with thread-local ldap connections, >> which >>> is what's being used right now in the trunk. That works even better. >>> >>> But as soon as I start to pump up the concurrency I'm still hitting >>> threadsafety bugs. I discovered one likely source of it only >> yesterday, >>> and a few more are likely still hanging around. But I'm almost there >> in >>> that respect. >>> >>> So at a tangeant I've also begun working on short lived worker >> threads >>> to disengage from longer running database interactions. But that's >>> really new stuff, and I havent yet hit on a good pattern to use. More >>> study is required. >>> >>> -- >>> ________________________________________________________________ >>> Paul Stevens paul at nfg.nl >>> NET FACILITIES GROUP GPG/PGP: 1024D/11F8CD31 >>> The Netherlands________________________________http://www.nfg.nl >>> _______________________________________________ >>> DBmail mailing list >>> [email protected] >>> https://mailman.fastxs.nl/mailman/listinfo/dbmail >> _______________________________________________ >> DBmail mailing list >> [email protected] >> https://mailman.fastxs.nl/mailman/listinfo/dbmail > > _______________________________________________ > DBmail mailing list > [email protected] > https://mailman.fastxs.nl/mailman/listinfo/dbmail > -- ________________________________________________________________ Paul Stevens paul at nfg.nl NET FACILITIES GROUP GPG/PGP: 1024D/11F8CD31 The Netherlands________________________________http://www.nfg.nl _______________________________________________ DBmail mailing list [email protected] https://mailman.fastxs.nl/mailman/listinfo/dbmail
