IDempiere/FullMeeting20130109

⇐ Table of Contents | Full Meeting Minutes | Full Meeting 2013-01-09

CarlosRuiz: Good Morning
nmicoud_: Bonjour
fcojvanninir: Hello
fcojvanninir: and good morning
CarlosRuiz: will start checking the peer review queue - but please feel free to open any topic :-)
nmicoud_: I've started http://jira.idempiere.com/browse/IDEMPIERE-531 ; wondering if there is a fast way to integrate modification from adempiere into idempiere
CarlosRuiz: checking some of the code
fcojvanninir: I would like to work on adapting this functionality: https://groups.google.com/forum/?hl=es&fromgroups=#!topic/idempiere/bRkhh-ZnYKI to any combobox and not just the main menu lookup combobox. How difficult would this be ?
nmicoud_: it has been done for swing client (http://jira.idempiere.com/browse/IDEMPIERE-4) ; maybe that could help
CarlosRuiz: we used that AutoComplete class when capturing a city on location dialog
CarlosRuiz: JanThielemann, AYT?
CarlosRuiz: fcojvanninir, I think you would need to explore something like changing the component Combobox by AutoComplete on the class WTableDirEditor
CarlosRuiz: there is also a feature when defining a String column, you can define it as autocomplete - and it will show a combobox of previous values
fcojvanninir: Ah, thats good to know !
fcojvanninir: What about the classes that govern the behavaior of the location dialog ?
fcojvanninir: I also noted that the city combo box was using the autocomplete
fcojvanninir: @nmicoud thanks for the link. I looked into and will take a closer look later.
CarlosRuiz: nmicoud_, I'm looking the class zkwebui/WEB-INF/src/org/adempiere/webui/trees/WTreeModel.java
fcojvanninir: It does look to be the same goal
CarlosRuiz: on the adempiere branch feature/ADEMPIERE-171-Lazy-Loading-Trees
CarlosRuiz: and found line 642 using LIMIT feature present on postgresql but not on oracle
CarlosRuiz: so - that's why I was looking if JanThielemann was here - to check the approach and if it's tested on oracle
nmicoud_: ok ; i have only look at swing code, didn't notice that
CarlosRuiz: so, probably the implementation for swing use a different approach?
nmicoud_: i think that (or i did not notice the LIMIT keyword)
nmicoud_: maybe that could be asked to jmpiloq ? that's him who make the commit
fcojvanninir: Have a good one fellows. Thank you for the information.
CarlosRuiz: nmicoud_, if I prepare an idempiere patch for the lazy load - can you test it ?
nmicoud_: yes
CarlosRuiz: every day is harder to integrate things from there :-)
nmicoud_: hope there won't be a lot of things to integrate in the future
CarlosRuiz: :-) do you think
nmicoud_: it will be done in the other way :)
jmpiloq_: Hi CarlosRuiz
jmpiloq_: HI nmicoud_
nmicoud_: hi
CarlosRuiz: Hi Jens
jmpiloq_: we did not test on Oracle
jmpiloq_: but LIMIT should also work on Oracle
jmpiloq_: Jan did some minor update the other day. We will push tomorrow
jmpiloq_: BTW: I tried to integrate idempiere back into adempiere without OSGI
jmpiloq_: you definitely have done an amazing peace of work...
jmpiloq_: piece
jmpiloq_: ok, just checked, we need to modify a bit for oracle
jmpiloq_: but should be easy...
jmpiloq_: I will have a look at the queries and come back with an oracle version
nmicoud_: thanks !
CarlosRuiz: I'm afraid because of the major refactoring that we made to the master-detail the lazy load cannot be integrated straight - it would need some extra-effort to make it work - speciall on the window/tab/panel classes
jmpiloq_: just need someone who can test that...
CarlosRuiz: jmpiloq_, why swing is not using that limit but zk does?
jmpiloq_: CarlosRuiz: I don't know, will ask Jan tomorrow, when he is back
CarlosRuiz: implementing limit on oracle is tricky
CarlosRuiz: you must use rownum on an ordered subselect
jmpiloq_: yes
jmpiloq_: select * from (select whatever, rank() over (order by bla) limit) from table) where limit <=10;
jmpiloq_: select * from (select whatever, rank() over (order by bla desc, rownum asc) limit) from table) where limit <=10;
jmpiloq_: (I love SQL)
CarlosRuiz: jmpiloq_, all the window part was big refactored in idempiere
CarlosRuiz: abstractadwindowcontent - adtabpanel - etc
CarlosRuiz: and is not easy to guess what Jan changed on those classes - so I think definitely in that part must be a work for you/Jan
hahmed: Hello everyone
hahmed: Hi CarlosRuiz, you here?
CarlosRuiz: Hi hahmed
hahmed: just got your message regarding IDEMPIERE-543
hahmed: the issue is that the current init script kills the idempiere-server.sh script on RedHat
hahmed: but does not kill any java processes it spawns
hahmed: to test do this: start idempiere-server on RedHat/CentOS, do ps -ef | grep java
hahmed: you will see a few Java processes that it starts (org.adempiere.application mostly)
hahmed: now stop the server using init script
hahmed: it will kill idempiere-server.sh but will not kill the java processes
CarlosRuiz: ah, interesting on Debian is using the "o pgid,command"
hahmed: yes that is the right way, pgid will kill the idempiere-server.sh plus anything in its group
CarlosRuiz: ah yes - I see that was fixed by hengsin here
CarlosRuiz: https://bitbucket.org/idempiere/idempiere/commits/e7affeaa7620/
CarlosRuiz: for Debian - but not for RedHat
hahmed: yeah, same should be done for RedHat (thats what my commit does actually)
CarlosRuiz: I think we would also need a "sort -u" at the end?
CarlosRuiz: the final kill looks something lik
CarlosRuiz: kill -15 -3395 3395 3395
hahmed: well since it kills all processes, so sort is not really needed
CarlosRuiz: yes, just seems redundant
CarlosRuiz: thanks hahmed for your explanation - just committed your fix
hahmed: Welcome CarlosRuiz
CarlosRuiz: ok guys - thanks a lot, I need to move to another meeting - thanks for attending
buildmaster: Project iDempiere build #654: SUCCESS in 8 min 41 sec: http://jenkins.idempiere.com/job/iDempiere/654/
buildmaster: hahmed: IDEMPIERE-543 RedHat Init Script does not stop iDempiere Server