PMC QA Meeting 20100428

From ADempiere
Jump to: navigation, search
This Wiki is read-only for reference purposes to avoid broken links.

Date: 2010-04-28
Time: 8AM GMT
Venue: irc #adempiere-team
Support Spreadsheet: Adempiere PMC QA
Chat times in GMT-5


(03:04:09) interopen: hi all
(03:04:14) trifon: viola: should hte Mobile app be abel to do?
(03:04:51) viola: 1st stage: show important BP infos, ring up contact, show some BIRT reports
(03:04:53) interopen: Hi milos, thank you for join QA
(03:05:17) interopen: we let architecture finish first...
(03:05:25) milos_t: Hi interopen
(03:05:38) viola: well architecture is over - just spoking around ;-)
(03:05:53) hengsin: probably a web app is a good option, otherwise you end up doing multiple client - the mobile platform is very fragmented.
(03:06:05) hengsin: hi Ivan
(03:06:19) viola: nononono we want it native - more fun!
(03:06:53) CarlosRuiz: great to have Paul here! if you want we can start QA meeting - sorry for my absence on Architectural meeting
(03:07:01) viola: ok bye guys - happy QAing
(03:07:05) trifon: it native you can track phone call and record int in the ERP :)
(03:07:09) trifon: fun fun fun :)
(03:07:18) viola: yeah i see - you get it!
(03:07:47) trifon: i have callTrack for Android and it records all my call into my goolge Calendar :)
(03:07:54) interopen: hi hengsin
(03:08:09) trifon: no privacy anymore :(
(03:08:16) viola: :-)))
(03:08:36) interopen: ok, so we start with QA: propose list
(03:08:54) interopen: 1. Review Sahi fix done by carlos
(03:09:11) interopen: 2. Review with Milos task list
(03:09:39) interopen: 3. Some organization about Security testing?
(03:09:43) interopen: 4. Ci server
(03:10:00) hengsin: for the ID generator, we can use the getUuid() method in AbstractComponent, it should be safer.
(03:11:15) hengsin: i.e, in IDGenerator.nextComponentUuid, return component.getId() + "_" + ((AbstractComponent)component).getUuid();
(03:11:50) hengsin: I believe that should be safe for use in production too.
(03:12:56) CarlosRuiz: BTW - partial match of the ID can be managed by Selenium?
(03:13:12) viola has left the channel.
(03:13:15) interopen: we not have time to research it yet
(03:13:33) hengsin: carlos, if my memory don't fail me, that should be yes
(03:14:21) interopen: also in mine... sunny have it in her research task list....
(03:15:57) hengsin: Ivan, how's the sahia test going ? I guess if it really went well, we can stick with it for now.
(03:16:04) CarlosRuiz: ok, just in case we could manage a SysConfig key to state how to obtain the ID - for production use or for testing server use
(03:17:26) interopen: hengsin, we are adding documentation about it, some problems are solve, now just some small ones with buttons that should be solve this week
(03:17:42) hengsin: carlos, I guess the question here is the accuracy of the test is in doubt if the production and test environment is not the same ...
(03:18:00) interopen: we are focusing a lot this week on try to have the Ci server so all test 'testing' can be faster and more clear
(03:18:35) hengsin: I haven't investigate much on this yet but this is something that need consideration
(03:18:40) interopen: hengsin, that is what we want to achieve with sahi and if possible later with selenium, with carlos code there is an option now
(03:19:34) interopen: all test we need to develop them and deploy them in any server, there can be small diferences of course... but not always
(03:21:02) interopen: milos, how about if you have time also review and research test cases with Sahi, all doc is in http://www.adempiere.com/index.php/PMC:QA it can be a good way to get you started...
(03:22:06) milos_t: I have time, but I must became familiar with Sahi
(03:23:18) interopen: sure, we try to make the documentation step by step so is not hard to become familiar with the tools
(03:23:36) phib has left the channel (quit: Quit: Leaving).
(03:25:54) interopen: interopen, going back to your suggestion to use + getuid() i also suggest it to Carlos, but he manage to add a unlimited generator, any clear advantage carlos?
(03:26:22) interopen: sorry i mean hengsin in the previous message
(03:29:08) CarlosRuiz: the last POC didn't use "random" ID / it generated the id of a field based on columnname+tab+windowno guaranteed to be unique
(03:29:08) CarlosRuiz: not sure if it can represent issues - I didn't make heavy tests with several users
(03:32:21) [1]interopen [~interopen@220.181.110.46] entered the channel.
(03:33:16) CarlosRuiz: ah, [1]interopen - you got disconnnected - AYT?
(03:33:20) [1]interopen: sorry, back again.... base on zk documentation we can have issues with this if two user use the same window as ZK is event base as a client-server architecture
(03:34:18) [1]interopen: we will try to test this issue of several users today
(03:34:34) [1]interopen: if not we have to go back to version two with the random id + prefix
(03:34:44) interopen has left the channel (quit: Ping timeout: 259 seconds).
(03:34:45) [1]interopen now is known as interopen
(03:34:57) CarlosRuiz: ok
(03:35:12) interopen: so i see we have the two solutions quite clear
(03:35:38) interopen: just the question from hengsin, for version 2 use getUid() or not?
(03:39:55) milos_t: I see you made a lot of documentation, it is easy to follow. Are there any plans for next steps? Which tests to make, in what order?
(03:41:37) interopen: milos, for now we focus on research, document and create POC ( proof of concept), until we have the CI testing server ready in adempiereqa.com and there we will ask community to join to create tests
(03:43:22) milos_t: ok
(03:44:59) interopen: now there are three testing methods you can see here http://www.adempiere.com/index.php/PMC:QA:Compare_testing_tools_junit%2Cfitnesse%2Csahi
(03:45:19) interopen: junit and fitnesse are stable, sahi works but need more work on it
(03:48:53) interopen: Carlos_out_of_topic, we have add this documentation mayeb useful for funtionality or usability teams? http://www.adempiere.com/index.php/Adempiere_Chinese_Translation
(03:51:04) CarlosRuiz: sure - I suppose translations are in the scope of "local" usability - great tool what launchpad is providing
(03:51:12) interopen: milos, any question as you review the documentation please send me an email, and if you can help with POC this week will be great
(03:52:13) interopen: Carlos, ok then you can bring it in the group meeting and review with red1, it may save time and improve our localizations
(03:53:26) interopen: Carlos, about security testing, how about if i create a front page as we have in QA for Security group, so we can see if others want to join?
(03:53:38) hengsin: Ivan, isn't prefix + uuid a better option ? I didn't see any issue going with that, correct me if I'm wrong here.
(03:53:44) milos_t: interopen: will you create google doc for POC?
(03:55:37) CarlosRuiz: Ivan, sounds good the idea of front page for security
(03:56:01) interopen: milos, we document all in the wiki and code add to svn ( still on work...)
(03:56:51) interopen: hengsin, i not see any possible issue with this, you are more expert with ZK, what issue with can have with the inique id for each field and window open for a user?
(03:57:05) interopen: Carlos, ok, i add it to the task list.
(03:58:49) hengsin: Ivan, don't get you here. AFAIK, it just have to be unique.
(04:00:27) CarlosRuiz: hengsin - in the latest POC - do you mean replacing these three lines on method nextComponentUuid
(04:00:27) CarlosRuiz: // int i = Integer.parseInt(desktop.getAttribute("Id_Num").toString());
(04:00:27) CarlosRuiz: // i++;// Start from 1
(04:00:27) CarlosRuiz: // desktop.setAttribute("Id_Num", String.valueOf(i));
(04:00:27) CarlosRuiz: with just this one?
(04:00:27) CarlosRuiz: String i = comp.getUuid();
(04:00:34) interopen: hengsin, the id have to be unique, no mater what user or session we are in
(04:01:21) interopen: in last Carlos solution the id is only unique for the window the fields are in
(04:01:55) interopen: if a user open more than one window the id is unique, if the user re-login the id is the same as past login
(04:02:04) hengsin: carlos, no, I means replace with String i = comp.getId() + "_" + comp.getUuid();
(04:02:28) hengsin: oops, sorry, carlos, yes, that'right.
(04:03:16) CarlosRuiz: ok - so I'll send a version 4 of the POC  :-)
(04:03:50) CarlosRuiz: the POC is just covering fields - but it must be extended for Tabs and toolbar buttons
(04:04:14) interopen: Carlos, thanks we will try to move faster this week on complete the POC testing and research
(04:04:36) hengsin: carlos, I think that can goes into trunk now.
(04:04:37) interopen: Carlos, yes we have some issues also with buttons, and will have it with tabs
(04:04:41) interopen: also with the menu....
(04:06:51) interopen: but this work we can do it here in QA group
(04:07:10) CarlosRuiz: ok hengsin - then I'll commit if the POC v4 works
(04:07:11) interopen: Carlos, the POC you are doing are more than useful for us
(04:07:40) interopen: Carlos, ok we can test it easier this way
(04:08:07) CarlosRuiz: BTW - apart from size of the page code - is there any issue with length of the ID?
(04:08:14) interopen: so, if no other we can finish here and continue next week, thanks all for the help
(04:08:36) hengsin: carlos, don't aware of any such constraint
(04:08:46) interopen: milos, thanks for join QA group, if any question, suggestion, comment let me know by email, we are in early stages and lot of help is needed
(04:09:18) milos_t: ok, we will be in touch interopen
(04:09:57) hengsin: it always amaze me that oracle identifier is still max 30 after so many years and people have build all sort of workaround over the years.
(04:10:31) CarlosRuiz: yes - indeed we suffer with that limitation here
(04:10:31) CarlosRuiz: thanks to all for attending - and again, sorry for my absence on architectural
(04:11:20) hengsin: bye carlos. btw, is there anything to be further discuss here for the OI stuff ?
...
(04:22:02) CarlosRuiz: parenthesis - v4 of the POC didn't work - I need to go back to a mix of v2 and v3
...
(04:45:15) CarlosRuiz: another parenthesis [committed the zkwebui id generator POC on trunk]
...
(04:45:52) interopen: Carlos, thanks we test it....
...
(04:49:04) hengsin: carlos, just saw your commit - getuuid() doesn't work ?
(04:49:40) CarlosRuiz: nope - didn't work - getuuid calls nextComponentUuid - so it became recursive
(04:54:47) hengsin: ok, carlos, have just read more of the zk source for that now, your approach should be ok.

Links