IDempiere/FullMeeting20150318 - Revision history

CarlosRuiz: drop JIRA notifications from log

2015-03-19T00:45:42Z

drop JIRA notifications from log

CarlosRuiz: full meeting

2015-03-19T00:45:12Z

full meeting

New page

⇐
[[IDempiere|Table of Contents]] |
[[IDempiere/Full Meeting Minutes|Full Meeting Minutes]] |
Full Meeting 2015-03-18


'''''ocurieles''''': Good Morning guys 
'''''nmicoud''''': Hi 
'''''aguerra''''': hello everybody!!!! 
'''''Deepak_''''': Hello All 
'''''CarlosRuiz''''': Good Morning 
'''''nmicoud''''': Bonjour 
'''''aguerra''''': hello CarlosRuiz 
'''''Deepak_''''': Hello Carlos 
'''''Deepak_''''': Carlos, In pricelist, currently vendorbreak supports break per BPartner only 
'''''Deepak_''''': As BPartner is mandatory field 
'''''Deepak_''''': Do you know why qty break for all BPartner not supported? 
'''''CarlosRuiz''''': Deepak_, AFAIK not supported with price list break 
'''''Deepak_''''': See pricelist break work for a BPartner 
'''''Deepak_''''': But not for all BPartner by providing null BPartner 
'''''Deepak_''''': BPartner is mandatory field on price list break 
'''''octavian_''''': Hello. Is it possible to run process from model validator ? 
'''''Deepak_''''': CarlosRuiz, and following question is it good idea to make BPartner non mandatory and supporting qty break from price list break? 
'''''Deepak_''''': octavian, Yes it is possible 
'''''Deepak_''''': You can look at how InOutGen.generate method launch process 
'''''octavian_''''': Thank you 
'''''red1''''': Hello everyone.. 
'''''red1''''': hi Hideaki 
'''''Hideaki''''': Hi red1 
'''''red1''''': Konnichiwa 
'''''Hideaki''''': Hello :D 
'''''tbayen''''': Daarestiet from Krefeld! :-) 
'''''aguerra''''': buenos dias red1 
'''''red1''''': buenos aires aguerra 
'''''CarlosRuiz''''': Hi Thomas / Hi Alejandro 
'''''hieplq''''': hi CarlosRuiz, i and nmicoud test IDEMPIERE-2244 and see some issue, also fixed it. please help me review it. 
'''''CarlosRuiz''''': ok hieplq - testing 
'''''Deepak_''''': Sorry may my session got Zombie 
'''''Deepak_''''': CarlosRuiz, and following question is it good idea to make BPartner non mandatory and supporting qty break from price list break? 
'''''Deepak_''''': If you answered above question, please repost 
'''''Deepak_''''': As I can't see 
'''''CarlosRuiz''''': Deepak, for me sounds like a good idea - maybe check in forums what others think 
'''''Deepak_''''': ohk 
'''''Deepak_''''': We already started work on same 
'''''Not-311d''''': [IDEMPIERE] ocurieles updated IDEMPIERE-1604 description set to "Extend the features on Window Customization / Tab Customization * Add Displayed Yes/No * Display Logic * Sequence * Where and Order By * Tab Customization " 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-1604 
'''''tbayen''''': Hi all! We think about an extension of the login. We want to allow authentication through BasicAuth and through client certificates. If anyone is interested we can discuss that. 
'''''CarlosRuiz''''': Hi tbayen sounds interesting 
'''''CarlosRuiz''''': I saw something on Chuck's side about connecting BasicAuth with AD_User - but don't know the details 
'''''CarlosRuiz''''': Another topic that I see interesting would be to authenticate using google apps 
'''''CarlosRuiz''''': I was wondering about that - and I think it can be good to make the "authenticator" pluggable 
'''''CarlosRuiz''''': and maybe the LDAP implementation on idempiere would be moved as a plugin 
'''''tbayen''''': One of the ideas is to install a client certificate on an phone and give the user access without a need to give him a password. You can allow access to a single device (not a user). 
'''''CarlosRuiz''''': but - in my research it didn't see as an easy task - there are several services to authenticate and manage things in different ways 
'''''CarlosRuiz''''': if we make the authentication pluggable - then you can have a plugin that checks the certificate and correlates it with a user 
'''''CarlosRuiz''''': or maybe the certificate brings the user information within and you just need to move ahead in the login process as authenticated 
'''''CarlosRuiz''''': I think if we isolate the LDAP implementation to be a plugin (in beta status) we could evolve the plugin interface to cope with other authentication services 
'''''CarlosRuiz''''': tbayen, there was also something tricky in the research I made for multi-tenant environments 
'''''tbayen''''': My idea was that the certificate is checked by an apache proxy. It adds a new header that contains the user name. 
'''''CarlosRuiz''''': if different tenants wants different authentication services - then you would need a way to identify the tenant on the login page 
'''''tbayen''''': Then my idea was that the standard login process can use that header information to automatically fill the user and password field. 
'''''CarlosRuiz''''': not password - just user - and a way to know that the user was authenticated with the certificate 
'''''tbayen''''': Yes, that was the idea. 
'''''tbayen''''': I hope with a right proxy configuration I can guarantee that the header is guaranteed to be only there when the user is authenticated. So the pure existence of the header is like having a password. 
'''''hieplq''''': please investigate, we can rewrite login module with http://shiro.apache.org/ 
'''''CarlosRuiz''''': hieplq, that looks great 
'''''CarlosRuiz''''': at the moment we have two ways of authentication - ad_user and LDAP - seems like shiro can make it pluggable 
'''''hieplq''''': seem shiro is skeleton, but it's good design for pluggable other authenticate method 
'''''Not-311d''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-1604 
'''''Not-311d''''': [IDEMPIERE] Maybe AD_Tab.AD_Process_ID can be also a good candidate to customize for tenant. 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-1604 
'''''Not-311d''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-1604 Fix Version set to "None" 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-1604 
'''''Not-311d''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-1604 
'''''Not-311d''''': [IDEMPIERE] tip: I think the change is just on class GridTabVO.loadTabDetails Also, IsSingleRow and IsReadOnly must be changed to be Yes/No list instead of a checkbox - to allow nulls and change the code accordingly to check for null there. Also I see in the GridTabVO code the IsReadOnly is not being used And I think the Multi Row Only can be hidden as is not used. 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-1604 
'''''tbayen''''': My thought was that the apache authentication is used by very many people and is very stable. And apache can use many different auth modules too. I have a greater trust to apache. 
'''''Not-311d''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-2244 Attachment set to "IDEMPIERE-2244_fordev.patch" 
'''''Not-311d''''': [IDEMPIERE] Thanks [~hieplq], first patch I was able to integrate successfully, but second failed. I'm attaching for you review a patch called IDEMPIERE-2244_fordev.patch. Minor fixes there: - the names of the migration scripts were wrong - must be yyyymmddhhmm - there was an update to c_acctprocessor in the script which I deleted - the patch for VM_ARGUMENTS didn't apply on second patch - please check if this patch do what you wanted there - also, I 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2244 
'''''Not-311d''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-2516 Attachment set to "IDEMPIERE-2516_add.patch" 
'''''Not-311d''''': [IDEMPIERE] [~nmicoud], there is still a case (tested) if GardenWorld set the SYSTEM_IN_MAINTENANCE_MODE=N and System set SYSTEM_IN_MAINTENANCE_MODE=Y in this case GardenUser still can login. The attached patch IDEMPIERE-2516_add.patch solves that - I think it would be required to complete your approach - WDYT? 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2516 
'''''Not-311d''''': [IDEMPIERE] nmicoud updated IDEMPIERE-2516 
'''''Not-311d''''': [IDEMPIERE] Yes, i haven't though of this possibility, but in this case the System value must override the tenant value. Thanks ! Nicolas 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2516 
'''''norbertbede''''': hi all 
'''''CarlosRuiz''''': Hi norbertbede 
'''''hieplq_''''': hi CarlosRuiz, i will test IDEMPIERE-2244 when i has strong network. wdyt about IDEMPIERE-2523? 
'''''CarlosRuiz''''': hieplq, if you and nmicoud tell me that 2244 is ready I can integrate it / so, please make your tests thorough :) 
'''''CarlosRuiz''''': ah 2523 - :DDD the lazy patch 
'''''norbertbede''''': in interesting issue we found - related to peformance 
'''''hieplq__''''': sure, next day i will test script, and material to sure parameter update to eclipse 
'''''norbertbede''''': we have locally PG91 on amazon 93 same task tooks locally 3 sec on amazon 90 sec 
'''''norbertbede''''': we found a difference setup - parameter - enable_material 
'''''norbertbede''''': this was locally disabled on server 93 enabled 
'''''norbertbede''''': lot discussion arounds 
'''''norbertbede''''': will test night change parameter 
'''''norbertbede''''': any experience ? 
'''''norbertbede''''': more info: http://www.postgresql.org/docs/9.3/static/runtime-config-query.html 
'''''CarlosRuiz''''': nope - not experience with that parameter here 
'''''norbertbede''''': your setup is true or false, maybe just for curious 
'''''CarlosRuiz''''': it says there the default is on 
'''''CarlosRuiz''''': let me check - I haven't touched defaults 
'''''norbertbede''''': sure. thanks 
'''''norbertbede''''': suggest this tool http://explain.depesz.com/ for everybody 
'''''norbertbede''''': for analysing queries 
'''''CarlosRuiz''''': in my postgresql - checked with "show all" command 
'''''CarlosRuiz''''': is not showing any value for enable_material 
'''''CarlosRuiz''''': not in 8.4 neither in 9.4 
'''''norbertbede''''': this "explain" us materialisation tooks on server 1. 0sec from query time AND second server 47% from query time 
'''''norbertbede''''': strange 
'''''norbertbede''''': but thanks. i will share our experiences with that 
'''''Not-311d''''': [IDEMPIERE] tsvikruha updated IDEMPIERE-2437 
'''''Not-311d''''': [IDEMPIERE] I will attach next patch after changing it into Preference Dialog - seems it logical... Currently I am wondering of usability this theme concept per Client. Thats mean login page will have some general UI loaded from default theme defined in system config. After successful login, default theme for client will be loaded (theme which is defined in AD_Theme with IsDefault=Y). I've just tested this use case and looks like it is on a good way. 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2437 
'''''norbertbede''''': CarlosRuiz, @tomassvikruha update our favorite topic :) theme "Make iDempiere themes definable on different levels" 
'''''norbertbede''''': we move forward 1 step 
'''''norbertbede''''': means we want to find out how we can switch client theme when login 
'''''norbertbede''''': tomas made some java early stage classes but he describe our problem 
'''''norbertbede''''': would be helpful if you can advice Tomas direction 
'''''norbertbede''''': thanks well 
'''''norbertbede''''': we are online now both so can interact if required 
'''''CarlosRuiz''''': thanks norbertbede / tomassvikruha - I was struggling here with an eclipse issue 
'''''CarlosRuiz''''': I don't know about the resources not refreshed - sounds like some cached resources issue 
'''''CarlosRuiz''''': but on the issue about login page with a default system-wide theme 
'''''CarlosRuiz''''': in several tickets we've found the need to have a way to identify a tenant from the login page 
'''''CarlosRuiz''''': maybe something like http://url/webui?AD_Client_ID=1000000 could make the trick 
'''''tomassvikruha''''': Maybe its related to cached resource, or login page is loaded with header css styles, and after login only body of page is changes…WDYT? Because when I inspect page in chrome there was still linked old css styles 
'''''CarlosRuiz''''': I think if you push F5 it will reload the correct 
'''''norbertbede''''': can be this forced by login without F5 ? 
'''''norbertbede''''': by code i mean 
'''''tomassvikruha''''': yes, F5 will reload browser page 
'''''tomassvikruha''''': after it is everything correct 
'''''Not-311d''''': [iDempiere] CarlosRuiz_globalqss pushed 1 commit to release-2.1 [+0/-0/±3] https://bitbucket.org/idempiere/idempiere/commits/ 
'''''Not-311d''''': [iDempiere] globalqss 033c01e - IDEMPIERE-2516 iDempiere monitor shows inactive clients (maintenance mode / active sessions) / thanks to Nicolas Micoud (nmicoud) 
'''''Not-311d''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-2516 status set to "Resolved" -assignee set to "Nicolas Micoud" -resolution set to "Fixed" 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2516 
'''''CarlosRuiz''''': don't know if that refresh can be forced programatically 
'''''norbertbede''''': ok thanks well we will try again with hieplq 
'''''norbertbede''''': to sort out 
'''''Not-311d''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-2523 Attachment set to "IDEMPIERE-2523_alt.patch" 
'''''Not-311d''''': [IDEMPIERE] [~hieplq], I reviewed the patch and I think the modification to MSysConfig is not needed, uploading a patch IDEMPIERE-2523_alt.patch suggesting using Query to get the MSysConfig object. Now, instead of this patch I think it can be worthy to work on making WPreference extensible (or pluggable). I think this is one of the things that every implementation would like to add their own preferences for users. WDYT? 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2523 
'''''Not-311d''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-2298 status set to "Resolved" -resolution set to "Won't Fix" 
'''''Not-311d''''': [IDEMPIERE] Thanks [~hahmed], I'm not integrating this patch as r3 moved to jetty, but I think it deserves a wiki page explaining how to secure the default tomcat and pointing to the patch here - and explaining to change server.xml in a running server 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2298 
'''''Not-311d''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-2298 labels set to "Security" 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2298 
'''''Not-311d''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-2298 Version set to "2.1" 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2298 
'''''Not-311d''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-2361 
'''''Not-311d''''': [IDEMPIERE] Yes, sounds like Env.getURL maybe can do the trick as you point here - using the sysconfig key we can put it anywhere, not based on zkwebui 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2361 
'''''Not-311d''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-2389 
'''''Not-311d''''': [IDEMPIERE] Thanks [~hieplq] Some things noticed in the peer review: - script 201502041515_IDEMPIERE-2389 on postgresql has an additional insert into ad_element than the oracle version - script 201502161604_IDEMPIERE-2389 has wrong register filename on postgresql, and it has a sysconfig update for the developer password leaked - script 201513031604-IDEMPIERE-2389 has wrong name - the format must be yyyymmddhhmm - serialVersionUID of ProcessInfo chang 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2389 
'''''Not-311d''''': [IDEMPIERE] tsvikruha updated IDEMPIERE-2478 
'''''Not-311d''''': [IDEMPIERE] Hi @Hiep, after merging last version I noticed that this patch produces one bug. Busy Dialog is not opened when process is fired from Info Window. Process runs without busy dialog - that is very strange for users. Can you reproduce it? 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2478 
'''''Not-311d''''': [IDEMPIERE] tsvikruha updated IDEMPIERE-2478 
'''''Not-311d''''': [IDEMPIERE] Hi [~hieplq], after merging last version I noticed that this patch produces one bug. Busy Dialog is not opened when process is fired from Info Window. Process runs without busy dialog - that is very strange for users. Can you reproduce it? 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2478 
'''''Not-311d''''': [IDEMPIERE] tsvikruha updated IDEMPIERE-2478 status set to "Reopened" -resolution set to "None" 
'''''Not-311d''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2478

@@ Line 45: / Line 45: @@
 '''''Deepak_''''': ohk<br>
 '''''Deepak_''''': We already started work on same<br>
 '''''tbayen''''': Hi all! We think about an extension of the login. We want to allow authentication through BasicAuth and through client certificates. If anyone is interested we can discuss that.<br>
 '''''CarlosRuiz''''': Hi tbayen sounds interesting<br>
@@ Line 69: / Line 67: @@
 '''''CarlosRuiz''''': at the moment we have two ways of authentication - ad_user and LDAP - seems like shiro can make it pluggable<br>
 '''''hieplq''''': seem shiro is skeleton, but it's good design for pluggable other authenticate method<br>
 '''''tbayen''''': My thought was that the apache authentication is used by very many people and is very stable. And apache can use many different auth modules too. I have a greater trust to apache.<br>
 '''''norbertbede''''': hi all<br>
 '''''CarlosRuiz''''': Hi norbertbede <br>
@@ Line 114: / Line 95: @@
 '''''norbertbede''''': strange<br>
 '''''norbertbede''''': but thanks. i will share our experiences with that<br>
 '''''norbertbede''''': CarlosRuiz, @tomassvikruha update our favorite topic :) theme "Make iDempiere themes definable on different levels"<br>
 '''''norbertbede''''': we move forward 1 step<br>
@@ Line 135: / Line 113: @@
 '''''tomassvikruha''''': yes, F5 will reload browser page<br>
 '''''tomassvikruha''''': after it is everything correct<br>
 '''''CarlosRuiz''''': don't know if that refresh can be forced programatically<br>
 '''''norbertbede''''': ok thanks well we will try again with hieplq<br>
 '''''norbertbede''''': to sort out<br>