openjpa-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Patrick Linskey" <plins...@gmail.com>
Subject Re: uuid-string not generating unique ids
Date Wed, 25 Jul 2007 19:28:16 GMT
That makes some sense -- I bet you're using multiple cores. Our tests
of the UUIDGenerator are all single-threaded.

-Patrick

On 7/25/07, Elias Torres <elias@torrez.us> wrote:
> It's just my laptop (IBM ThinkPad Core Duo 2.16Ghz).
>
> -Elias
>
> Patrick Linskey wrote:
> > Great! What type of hardware are you running on, btw? We were already
> > doing some testing in a pretty tight loop, but maybe Hotspot was being
> > clever and optimizing it away.
> >
> > -Patrick
> >
> > On 7/25/07, Elias Torres <elias@torrez.us> wrote:
> >> FYI, the patch worked.
> >>
> >> -Elias
> >>
> >> Patrick Linskey wrote:
> >> > There have been a lot of changes since then, but things should be
> >> > pretty compatible.
> >> >
> >> > If you've got a good test suite, I'd recommend switching to trunk and
> >> > seeing what happens.
> >> >
> >> > -Patrick
> >> >
> >> > On 7/25/07, Elias Torres <elias@torrez.us> wrote:
> >> >>
> >> >> Thank you Patrick and Mark. I'm using 0.9.7, so I'm either patching
it
> >> >> or switching to trunk. Do you think it'd be safe (for an agile
> >> >> development group) to switch to trunk from 0.9.7 or were there many
> >> big
> >> >> changes since then?
> >> >>
> >> >> Anyways, thanks again. I'll ask a couple of other things later.
> >> >>
> >> >> -Elias
> >> >>
> >> >> Patrick Linskey wrote:
> >> >> > I just committed a change that will maybe resolve the issue. Please
> >> >> > let us know if you see it come up again.
> >> >> >
> >> >> > I also logged the issue as OPENJPA-296.
> >> >> >
> >> >> > -Patrick
> >> >> >
> >> >> > On 7/25/07, Marc Prud'hommeaux <mprudhom@apache.org> wrote:
> >> >> >> Patrick-
> >> >> >>
> >> >> >> I bet you are right. Your analysis of the problem looks correct
> >> to me.
> >> >> >>
> >> >> >>
> >> >> >> On Jul 25, 2007, at 9:12 AM, Patrick Linskey wrote:
> >> >> >>
> >> >> >> > Hi,
> >> >> >> >
> >> >> >> > Based on a quick read, it looks like maybe the problem
is
> >> >> >> > UUIDGenerator.java:178. I suspect that we should be changing
> >> that to
> >> >> >> > obtain another newTime inside the if block after incrementing
the
> >> >> >> > sequence. However, I'm not an expert in this algorithm;
does
> >> anyone
> >> >> >> > have a better grasp on it than do I?
> >> >> >> >
> >> >> >> > -Patrick
> >> >> >> >
> >> >> >> > On 7/25/07, Elias Torres <elias@torrez.us> wrote:
> >> >> >> >>
> >> >> >> >> Hi,
> >> >> >> >>
> >> >> >> >> First off, I want to say thank you for the hard work
this
> >> >> >> >> community has
> >> >> >> >> put on this JPA implementation. I'm extremely eager
in
> >> getting to
> >> >> >> >> know
> >> >> >> >> it better and hopefully put it to good use in some
of the
> >> projects
> >> >> >> >> I'm
> >> >> >> >> currently working on at IBM. If all goes well, I'll
be asking
> >> >> some of
> >> >> >> >> the harder questions later and possibly discuss new
features not
> >> >> >> >> currently implemented.
> >> >> >> >>
> >> >> >> >> Now if you could please help me out with a small
issue I'm
> >> having
> >> >> >> >> with
> >> >> >> >> uuid generation.
> >> >> >> >>
> >> >> >> >> @Entity
> >> >> >> >> public class Foo {
> >> >> >> >>
> >> >> >> >>   @Id
> >> >> >> >>   @GeneratedValue(strategy=GenerationType.AUTO, generator="uuid-
> >> >> >> >> string")
> >> >> >> >>   private String id;
> >> >> >> >>
> >> >> >> >> }
> >> >> >> >>
> >> >> >> >> CREATE TABLE Foo (
> >> >> >> >>   id VARCHAR(16) NOT NULL
> >> >> >> >> }
> >> >> >> >>
> >> >> >> >> If I do one em.persist() it works just fine. However,
if I do
> >> a a
> >> >> >> >> whole
> >> >> >> >> list of them, I get exceptions (duplicate). If I
enable logging
> >> >> I can
> >> >> >> >> see they are all using the same id during INSERT.
I stepped a
> >> bit
> >> >> >> >> through the code and noticed that uuid-string doesn't
get
> >> generated
> >> >> >> >> while em.persist(), but much later while committing/flushing.
I
> >> >> then
> >> >> >> >> looked a bit (just a bit) at the UUID generator code
and noticed
> >> >> that
> >> >> >> >> the bytes are not changing much (maybe only the first
two), but
> >> >> >> >> again,
> >> >> >> >> that was while debugging, so I'm assuming I have
a supa-fast
> >> >> machine.
> >> >> >> >>
> >> >> >> >> Anybody running into the same problem? Any suggestions?
> >> >> >> >>
> >> >> >> >> -Elias
> >> >> >> >>
> >> >> >> >
> >> >> >> >
> >> >> >> > --
> >> >> >> > Patrick Linskey
> >> >> >> > 202 669 5907
> >> >> >>
> >> >> >>
> >> >> >
> >> >> >
> >> >>
> >> >
> >> >
> >>
> >
> >
>


-- 
Patrick Linskey
202 669 5907

Mime
View raw message