IDempiere/FullMeeting20130417

⇐ Table of Contents | Full Meeting Minutes | Full Meeting 2013-04-17

CarlosRuiz: Good Morning
nmicoud: Bonjour
nmicoud: Carlos, could please look at : https://groups.google.com/forum/#!topic/idempiere/UrgFACYq2yg ? This is a very strange thing. And also this ticket which is very fast to fix : http://jira.idempiere.com/browse/IDEMPIERE-864 ; and nice logo btw ;)
CarlosRuiz: checking
nmicoud: thanks
AlexYang: Hi Carlos, hi everyone
nmicoud: Hi
AlexYang: What's topic talking now?
nmicoud: those 2 things : https://groups.google.com/forum/#!topic/idempiere/UrgFACYq2yg / http://jira.idempiere.com/browse/IDEMPIERE-864
AlexYang: I see it's a little silent in forum last week
nmicoud: Carlos: i'm using oracle; don't know it could matter
CarlosRuiz: :-( google apps mail and drive are having problems
CarlosRuiz: Nicolas
nmicoud: yes
CarlosRuiz: your InfoWindow seems different - line 666 in mine is a closing bracket
CarlosRuiz: are there steps to reproduce the issue?
nmicoud: it's around line 666
nmicoud: my message was not clear
nmicoud: you should have
nmicoud: if (label != null) {
nmicoud: if (infoColumn.getQueryOperator().equals(X_AD_InfoColumn.QUERYOPERATOR_Gt) ||
nmicoud: i've added those logs between those 2 lines
nmicoud: line 659 and 660
nmicoud: And QueryOperator is not a mandatory column and its value is tested. Why don't we test if it is null ?
nmicoud: before comparing it ?
CarlosRuiz: I see it mandatory
nmicoud: in ad or base ?
nmicoud: because i have some null values
CarlosRuiz: checking "Business Partner Info" on https://demo.idempiere.com/
CarlosRuiz: on Info Window as System
nmicoud: ok
CarlosRuiz: in column definition the query operator is red when empty
nmicoud: but it works well in demo.idempiere.com ; what is also weird is that a direct sql return 'something' whereas the get_Value return null
CarlosRuiz: ah - I see - there are some nulls - seems like mandatory is defined at window level
nmicoud: because the field is not always shown
nmicoud: only when you tick a checkbox
CarlosRuiz: ah- mandatory logic
nmicoud: 'criteria something'
nmicoud: that's why i would have seen something like "if infoColumn.getQueryOperator != null { ... } " ; no ?
CarlosRuiz: yep - those if must be written inverse if the value can be null
nmicoud: but that doesn't explain the difference between get_Value and sql ? i've never noticed that kind of problem
CarlosRuiz: cannot reproduce it in postgres - will try oracle
JanThielemann: hi carlos and alex
CarlosRuiz: Hi JanThielemann
CarlosRuiz: Hi AlexYang - sorry, I forgot to answer alex greeting :-)
JanThielemann: do you know if it is possible to observe value changs from a WSearchEditor? In my custom form i use one and i want to get notified after the user selected an entry via the search editor
AlexYang: Carlos, is ready to release? I see recent commits replaced all adempiere source to idempiere :)
CarlosRuiz: JanThielemann, I think you can add an event listener there
JanThielemann: which event do i have to use?
CarlosRuiz: AlexYang, yes, I think we're almost ready - I was replacing logo
CarlosRuiz: there is a value change event if I recall correctly
JanThielemann: i'll investigate it, thx
CarlosRuiz: Nicolas - cannot reproduce with oracle either
nmicoud: ah...
JanThielemann: got to go, bye@all
CarlosRuiz: 10 or 11?
nmicoud: 11
CarlosRuiz: will try in oracle 11g - meanwhile I import DB I'll check the second issue you reported
nmicoud: ok ; i will try to refresh/clean/restart eclipse
Edwin_Ang: hi everyone
CarlosRuiz: Hi Edwin_Ang
nmicoud: Hi Edwin
Edwin_Ang: what's the current discussion? :)
nmicoud: those 2 things : https://groups.google.com/forum/#!topic/idempiere/UrgFACYq2yg / http://jira.idempiere.com/browse/IDEMPIERE-864
AlexYang: I have tried compiere web client this week on their demo site: http://demo.compiere.com
AlexYang: I found it's much faster than http://demo.idempiere.com
AlexYang: Carlos & Hengsin, what do you think about this? I found compiere use GWT. does it mean gwt is faster than zk?
AlexYang: And I also found the UI style of compiere is much tidy than idempiere. it wastes so much space in idempiere's webui
AlexYang: I am not sure if my feeling is only personally or wide feeling about zk webui:)
CarlosRuiz: that was discussed recently on facebook forums :-)
nmicoud: found something : POInfo : m_columnNameMap doesn't have the QueryOperator column !
nmicoud: checking the sql that fills it
AlexYang: Carlos, are you saying that we are talking about zk and gwt in facebook group?
AlexYang: I am sorry I seem can't find relate discussion their. could you kindly share a link to me?
CarlosRuiz: I just said that was discussed before and we found kind of boring that discussion :-) you'll always find fans of gwt blaming zk and the opposite
CarlosRuiz: and you can find enough arguments for both sides if you search on google
CarlosRuiz: so - if I remember correctly the conclusion of the facebook discussion was - we have a nice and fast zk version - if you want a gwt version please go ahead
AlexYang: thanks Carlos
AlexYang: you are right. we inherit swing but gwt from compiere
CarlosRuiz: is like python vs java - useless discussion
AlexYang: currently, most important thing is that we already have a good zk version
CarlosRuiz: you're right - this compiere is faster than last time I tested it
CarlosRuiz: probably as this is a demo site they can avoid some parts that make it slow
CarlosRuiz: Nicolas - oracle 11 working fine here too
nmicoud: ok - need to find out what's wrong here ; thanks for checking
CarlosRuiz: in my m_columnNameMap the QUERYOPERATOR is there
nmicoud: i think i just forgot to create missing trl
nmicoud: checking that...
nmicoud: yiipiie !!!
nmicoud: a nice newbie error for me :))
nmicoud: Sorry for wasting your time ; i'll update the google group thread
nmicoud: And so, should we test if infoColumn.getQueryOperator() is null or it is ok ?
AlexYang: should go now. have good day. bye
CarlosRuiz: I checked that
CarlosRuiz: seems fine as the tree is created just for the query criteria columns and those have mandatory the queryoperator
nmicoud: ok -so no problem !
nmicoud: what is the best practice to return an error message to user in a beforeSave ? yesterday, i noticed this : https://groups.google.com/forum/#!topic/idempiere/b0DDmTJ9Q00 ;
CarlosRuiz: Edwin_Ang, IDEMPIERE-796 is ready to review - I see two commits in your repo
Edwin_Ang: yes
CarlosRuiz: does it have integrated Armen patches too?
Edwin_Ang: pls review and give feedback if there's anything that need to be improved
Edwin_Ang: yes
CarlosRuiz: gr8
Edwin_Ang: the only thing that i don't take from Armen's patch is setting error import to 'E'
Edwin_Ang: cos i don't really get what's the point of setting it to 'E' and then change it back to 'N'
CarlosRuiz: nmicoud, on the saveError - I don't get clear what is happening - is not showing correctly the translated message=
CarlosRuiz: ?
nmicoud: yes ; its show the translation of the SaveError msg and {0} whereas it should show the translation of the SaveError msg and the additional message defined in the beforeSave
nmicoud: log.saveError("Error", "bla bla bla); doesn't produce the same result as log.saveError("SaveError", "bla bla bla); when logged in French (i suppose it apply for all translated session)
Not-001: [iDempiere] 1 commit https://bitbucket.org/idempiere/idempiere/commits/
Not-001: [iDempiere] globalqss ae0f0a47e1b0 IDEMPIERE-864 oracle/unix scripts RUN_DBRestore.sh and RUN_DBRestoredp.sh are the same
CarlosRuiz: it sounds inconsistent - can you please open a jira ticket for that
nmicoud: yes - i will
buildmaster: Project iDempiere build #1038: SUCCESS in 6 min 24 sec: http://jenkins.idempiere.com/job/iDempiere/1038/
CarlosRuiz: Edwin_Ang, first script is erroring here with entitytype FA missing - is it ok to change it to D?
Edwin_Ang: oh
Edwin_Ang: it's ok
Edwin_Ang: just change it to D
CarlosRuiz: second script is also showing 3 errors
CarlosRuiz: first is that column C_DocType_ID already exists - checking
CarlosRuiz: ah yes - is creating it twice
CarlosRuiz: this is sometimes caused because the migration script log even failing inserts
Edwin_Ang: it's weird
Edwin_Ang: cos i work on it using a clean new db
Edwin_Ang: with the latest mig scripts
CarlosRuiz: throwing also errors because of window names dups
CarlosRuiz: I think is similar issue - you try to set it up - the UI shows an error - but anyways the update is logged - something to improve maybe on the migration script logger
Edwin_Ang: yo alexas
alexas: where's the party? =)
Edwin_Ang: north
Edwin_Ang: at alexis
Edwin_Ang: :D
CarlosRuiz: ok, finished some fixes for migration scripts
nmicoud: http://jira.idempiere.com/browse/IDEMPIERE-865 created (problem with log.saveError)
CarlosRuiz: thanks Nicolas
alexas: ok, monday after lodan meeeting.. sounds good =D
alexas: nice to see the new idempiere logo at jira.. dashing..
nmicoud: wondering : if i have to modify a class that i can't put in a plugin ; how could i export it to integrate into adempiere-server ?
nmicoud: should i generate the full installer ?
CarlosRuiz: which class for example?
nmicoud: APanel
nmicoud: or the equivalent for web
CarlosRuiz: well - the ideal is to check why you need to modify a core class - to see if we need a new extension point
nmicoud: means that everything could be put in a plugin if an extension point exists ?
CarlosRuiz: I mean - this is not easy to achieve - but the ideal is that core classes must not need modifications
CarlosRuiz: so - in our thinking on iDempiere
CarlosRuiz: if a core class needs a modification - we want to analyze the case to see if it points to a lack of extensibility
nmicoud: i see
nmicoud: thanks
Not-001: [iDempiere] 3 commits https://bitbucket.org/idempiere/idempiere/commits/
Not-001: [iDempiere] edwin_ang fd2f747951ca IDEMPIERE-796 All Fixed Asset Related Document Must Use New DocBaseType 'FAA'
Not-001: [iDempiere] edwin_ang 47d63dae0d28 IDEMPIERE-796
Not-001: [iDempiere] globalqss b5de46a6832e IDEMPIERE-796 - peer review migration scripts
Edwin_Ang: thanks for integrating, Carlos :)
CarlosRuiz: thanks to you
Edwin_Ang: CarlosRuiz: btw, i want to discuss about IDEMPIERE-480
Edwin_Ang: i need to fix this bug for a client
Edwin_Ang: i think it can be a global fix
Edwin_Ang: not customer specific
CarlosRuiz: is it not solved?
Edwin_Ang: i haven't tested it yet
Edwin_Ang: in demo.idempiere.com
Edwin_Ang: last time i tested was when i created IDEMPIERE-813
Edwin_Ang: it is related to IDEMPIERE-480
Edwin_Ang: and i don't seen any commit msg related to those two tickets yet
CarlosRuiz: 480 is solved - Daniel stated he needed something different - but IMHO is solved
CarlosRuiz: 813 is different
CarlosRuiz: I haven't analyzed but I think even if the date line is filled different - idempiere is using the journal date
CarlosRuiz: so - maybe is annoying - but if that's the case it must be behaving correctly - at least not corrupting data as 480 was
Edwin_Ang: well.. you guys are discussing about dateacct in document lines in 480
CarlosRuiz: maybe is worthy a beforeSave on JournalLine similar to what we added on statementline
Edwin_Ang: that's why i assume they are related
buildmaster: Project iDempiere build #1039: SUCCESS in 5 min 50 sec: http://jenkins.idempiere.com/job/iDempiere/1039/
Edwin_Ang: yes
Edwin_Ang: i am thinking of validating the statementline and journalline on beforesave
Edwin_Ang: to validate the inserted date is within the period specified in their header table
CarlosRuiz: in statement line there is no validation - it is a direct assignment of the header date
Edwin_Ang: what do you think?
Edwin_Ang: but user can change the date
CarlosRuiz: on statement line he cannot - it's overwritten on beforesave
Edwin_Ang: well.. i haven't tested statementline in current trunk yet
Edwin_Ang: testing it now
CarlosRuiz: just updated demo with latest fixed assets - if you want to test there too
Edwin_Ang: ok
CarlosRuiz: ah - dateacct on statement line is read-only - and set equal to header on save
Edwin_Ang: yep
Edwin_Ang: just realize it
Edwin_Ang: i think it is ok that way
Edwin_Ang: but gl journal line can use the same treatment
CarlosRuiz: as I see dateacct in gl journal line is not used anywhere in code
Edwin_Ang: cos when analyzing it, i found out that if it is allowed to have different date for multi currency gl journal transaction
Edwin_Ang: sorry, i mean it can't use the same treatmen
Edwin_Ang: just confirmed.. still have that same bug in gl journal line
Edwin_Ang: hi @red1
CarlosRuiz: Здравствуйте товарищ
red1: hi
red1: power blackout here in Siberia
CarlosRuiz: Edwin_Ang, testing on demo with GL Journal 1000 - it's behaving weird - posting with zero
red1: luckily I have sim card for web in phone
Edwin_Ang: i am testing with journal 1001
Edwin_Ang: CarlosRuiz: i think your issue is because no currency rate has been set for currency COP
Edwin_Ang: @red1: so how's siberia so far? :)
CarlosRuiz: Edwin_Ang, on Journal Line - the field rate is set according to the dateacct
CarlosRuiz: I think maybe is better to have a validation - you cannot save a multi-currency line if the rate is zero
red1: very nice friendly people, i just cooked tour fish kicap recipe, they liked it. can't find serai here though
CarlosRuiz: the document is posted with the header date - but the rate is calculated with the dateacct line
Edwin_Ang: CarlosRuiz: i think it is better to use another date column for controlling the currency rate
red1: when Yan we release DEWINA?
Edwin_Ang: dateacct is always related with journaling
CarlosRuiz: I think we're ready - hengsin wanted FA fixes integrated - and that was just done
red1: great! Russia's waiting too. There id someone groom Ukraine that was reviewing Heng Sin 's work and i told him to publish it
red1: they are shy of their English
red1: he can't make Wednesday meeting but he read the minutes
Edwin_Ang: we should make some noise
Edwin_Ang: :)
red1: i m very hopeful of Russia
red1: they call me red the first looks Peter the first the old king
Edwin_Ang: hahaha
Edwin_Ang: one issue while testing fixed assets on demo.idempiere.com
Edwin_Ang: field gl category is missing
CarlosRuiz: ¿?
CarlosRuiz: in script 3
CarlosRuiz: there was this instruction
CarlosRuiz: DELETE FROM AD_Column WHERE AD_Column_ID=55559
CarlosRuiz: that column could not be deleted because there were like 4 fields
CarlosRuiz: so I inactivated all the fields and the column
Not-001: [iDempiere] 2 commits https://bitbucket.org/idempiere/idempiere/commits/
Not-001: [iDempiere] richard1988 5b26a3e6bb77 IDEMPIERE-856 the broadcast message window is clipped on the bottom
Not-001: [iDempiere] globalqss 9264ea676878 IDEMPIERE-294 BroadCast Message / Several fixes: - Clients <> System cannot expire their messages - Reorganize window and set displaylogic for buttons - Fix names of buttons - Display just messages of client logged in - Display non-expired messages without expiration date
buildmaster: Project iDempiere build #1040: SUCCESS in 6 min 47 sec: http://jenkins.idempiere.com/job/iDempiere/1040/