3:37 a.m.
it only took me a week. the weirdest sequence of phenomena that i ever saw.
eventually one of the obstacles in finding the problem was that the database returned a
list sorted on a timestamp. in most cases that gives a precise order... unless when the
multiple process instances have a start date within the same millisecond. in that case,
the order becomes random.
along the way i tried to upgrade hibernate. (i won't commit those changes) but i
want to mention that doing so was very hard. it seems we cannot rely on the automagical
third party dependency resolution in mvn. i assume that the dependencies are not always
correctly specified in the public repositories.
also i saw way too many dependencies. this was really frustrating when cleaning the repo.
it took me 1 hour to rebuild the repo. And during that one hour, i couldn't do
anything else because once and the while a jar wasn't found in a public repo and I
just had to retry.
maybe we should consider to have our own maven repo for our dependencies. this is
considered good practice with ivy. having a single repo for all of jboss is good, but
making sure that all the 3rd party dependency resolutions happer correctly might be easier
if we have full control. i'll add this as a topic to discuss on the meeting.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4183419#...
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...
4:06 a.m.
On Tue, 2008-10-21 at 04:37 -0400, tom.baeyens(a)jboss.com wrote:
eventually one of the obstacles in finding the problem was that the
database returned a list sorted on a timestamp. in most cases that
gives a precise order... unless when the multiple process instances
have a start date within the same millisecond. in that case, the
order becomes random.
does this mean you solved the problem?
4:17 a.m.
yes. in my branch it is solved. now i'm fighting with svn switch, maven updates,
merging and the likes to get the changes into trunk.
regards, tom.
Heiko Braun wrote:
On Tue, 2008-10-21 at 04:37 -0400, tom.baeyens(a)jboss.com wrote:
> eventually one of the obstacles in finding the problem was that the
> database returned a list sorted on a timestamp. in most cases that
> gives a precise order... unless when the multiple process instances
> have a start date within the same millisecond. in that case, the
> order becomes random.
does this mean you solved the problem?
_______________________________________________
jbpm-dev mailing list
jbpm-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/jbpm-dev
--
regards, tom.
4:31 a.m.
don't worry, you'll get used to that.
On Tue, 2008-10-21 at 11:17 +0200, Tom Baeyens wrote:
yes. in my branch it is solved. now i'm fighting with svn
switch, maven updates,
merging and the likes to get the changes into trunk.
regards, tom.
Heiko Braun wrote:
> On Tue, 2008-10-21 at 04:37 -0400, tom.baeyens(a)jboss.com wrote:
>> eventually one of the obstacles in finding the problem was that the
>> database returned a list sorted on a timestamp. in most cases that
>> gives a precise order... unless when the multiple process instances
>> have a start date within the same millisecond. in that case, the
>> order becomes random.
>
> does this mean you solved the problem?
>
> _______________________________________________
> jbpm-dev mailing list
> jbpm-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/jbpm-dev
4:35 a.m.
for some reason, maven seems to invalidate or block all the good things from eclipse.
before, when i worked with CVS and ant, eclipse always behaved very well as you can
do that with the basic, well tested eclipse parts.
now, with svn and maven, we need plugins that are slow and crappy. and we need to
switch to the command line regularly, and then messing up the synchronization between
filesystem and eclipse.
that is a mess. i am really in doubt if maven and svn are worth that kind of
trouble... maybe it is different with intellij. do they have better svn & mvn
integration?
regards, tom.
Heiko Braun wrote:
don't worry, you'll get used to that.
On Tue, 2008-10-21 at 11:17 +0200, Tom Baeyens wrote:
> yes. in my branch it is solved. now i'm fighting with svn switch, maven
updates,
> merging and the likes to get the changes into trunk.
>
> regards, tom.
>
>
> Heiko Braun wrote:
>> On Tue, 2008-10-21 at 04:37 -0400, tom.baeyens(a)jboss.com wrote:
>>> eventually one of the obstacles in finding the problem was that the
>>> database returned a list sorted on a timestamp. in most cases that
>>> gives a precise order... unless when the multiple process instances
>>> have a start date within the same millisecond. in that case, the
>>> order becomes random.
>> does this mean you solved the problem?
>>
>> _______________________________________________
>> jbpm-dev mailing list
>> jbpm-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/jbpm-dev
--
regards, tom.
6:18 a.m.
Hi.
What maven plugin are you using? The m2eclipse does relatively good work
managing dependencies: http://m2eclipse.sonatype.org/update/
I have like 40 projects in the workspace (modules) that depend on each
other in various ways, and it m2eclipse manages them pretty nicely.
One thing worth noting - in my experience, the versions of the plugin,
that are later than 0.9.4-20080603-0114 doesn't work well.
So if you have any struggles, try this version.
Also, I suggest - from my hard experience - it is worth considering
alternatives.
cvl
Tom Baeyens wrote:
for some reason, maven seems to invalidate or block all the good
things from eclipse.
before, when i worked with CVS and ant, eclipse always behaved very
well as you can do that with the basic, well tested eclipse parts.
now, with svn and maven, we need plugins that are slow and crappy.
and we need to switch to the command line regularly, and then messing
up the synchronization between filesystem and eclipse.
that is a mess. i am really in doubt if maven and svn are worth that
kind of trouble... maybe it is different with intellij. do they have
better svn & mvn integration?
regards, tom.
Heiko Braun wrote:
>
> don't worry, you'll get used to that.
>
> On Tue, 2008-10-21 at 11:17 +0200, Tom Baeyens wrote:
>> yes. in my branch it is solved. now i'm fighting with svn switch,
>> maven updates, merging and the likes to get the changes into trunk.
>>
>> regards, tom.
>>
>>
>> Heiko Braun wrote:
>>> On Tue, 2008-10-21 at 04:37 -0400, tom.baeyens(a)jboss.com wrote:
>>>> eventually one of the obstacles in finding the problem was that the
>>>> database returned a list sorted on a timestamp. in most cases that
>>>> gives a precise order... unless when the multiple process instances
>>>> have a start date within the same millisecond. in that case, the
>>>> order becomes random.
>>> does this mean you solved the problem?
>>>
>>> _______________________________________________
>>> jbpm-dev mailing list
>>> jbpm-dev(a)lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/jbpm-dev
>
6229
days inactive
6229
days old
5 comments
4 participants
participants (4)
-
Heiko Braun -
Tom Baeyens -
tom.baeyens＠jboss.com -
Vytautas Civilis