CC Meeting Full 20070306

From ADempiere ERP Wiki

Jump to: navigation, search

hengsin: hi victor.
hengsin: hi all
CarlosRuiz: Hi Hengsin, Hi teo, Hi Victor
vpj-cd_: Carlos can we begin?
CarlosRuiz: yes, of course, the CC is well represented :-) 4/5
vpj-cd_: how are not here?
red1 [n=red1@] ha entrado en la sala.
vpj-cd_: Trifon?
red1: trifon is in Sarajevo
vpj-cd_: haaaa
vpj-cd_: hi red1
red1: hi.. i just informed trifon
croo [n=crooney@] ha entrado en la sala.
CarlosRuiz: ok, we can start if you want, the theme today is about the release strategy [ proposed by Hengsin ]
vpj-cd_: ok what is the stategy
CarlosRuiz: current strategy, feature freeze and wait for bugs to be fixed
CarlosRuiz: Hengsin think we can change strategy
vpj-cd_: ok what new strategy
vpj-cd_: what is new strategy Low?
CarlosRuiz: Victor, what's your opinion about testings by partners - while you was Compiere partner?
CarlosRuiz: I remember a compiere version very buggy where Compiere tried to put the testing efforts on partners
hengsin: fast track to 3.2 to get more real life testing and maintain that as a stable branch.
vpj-cd_: I think is unique way to
vpj-cd_: test reality the adempiere
vpj-cd_: if idalica, gobla era , e-evolution, qs, obf ,etc
hengsin: and to all, what is your assessment of 315 stability compare to 314 and earlier and also compare to whatever compiere version that you have use before ?
kontro [] ha entrado en la sala.
Mike_Judd [] ha entrado en la sala.
croo: well loking athe bugs I was thinking the quality was less but I think with the freeze proposed a lot of things were rushed in to get it into the stable release
croo: but also many bugs nehiond the scenes are now addressed!
vpj-cd_: if we make the inbtensive test
vpj-cd_: then we will have a stable version
vpj-cd_: I think the commnunity is second filter
croo: but this doesn't use the community.... why not freeze now - intensive test - then release as a RC to the community
vpj-cd_: but the enterprise consulting should give some time to this process
croo: when all priority bugs from that are addressed call it 3.2 stable
croo: yes ... intensive test by the enterprise consulting I meant
vpj-cd_: Croo the problem is we need a user manual to test adempiere to commnunity
vpj-cd_: but this we have not
CarlosRuiz: Colin, that's the plan, but the "intensive test" part can be failing :-)
CarlosRuiz: Victor, a user manual is not the solution, if that was, we can buy manuals from Compiere and give it to community to test ;-)
vpj-cd_: here are all development and we working with force to have stable to adempiere
croo: intensive test failing .. well I think because we pushed for as much NEW functioanlity rather than on the initial goal of a stable comiere version
vpj-cd_: in spacial Carlos , Teo and Law
croo: but it's there now so lets test it!
croo: a user manula is a good point ... in truth it's not a complete system without it!
vpj-cd_: I think now is the moment to enterprise consulting test
croo: but writing user manuals is a skill in itself!
CarlosRuiz: yes, but a user manual don't guarantee testings
vpj-cd_: adempiere
vpj-cd_: yes Carlos then only we have
croo: no but these are both tasks which need to be addressed
vpj-cd_: the test need make to somebody that know adempiere
CarlosRuiz: yes, of course, but that's another problem, how about release strategy?
vpj-cd_: this know will should in detail
vpj-cd_: I think
vpj-cd_: we need show that will begin the QA to 3.2
hengsin: we need to create a stable branch to get more testing by commnunity
croo: well my feeling is that the stabilty is less because we put som much extra functionality in ... we have called a freeze on everything except bugs and that should stabilise the code
vpj-cd_: next we divide the test to business process
croo: the biggest issue I see in the forums is people struggling with installing on postgresql
CarlosRuiz: yes, I would prefer people to test in Oracle XE, but obviously people prefer postgresql
vpj-cd_: yes :-)
vpj-cd_: then we begin a QA cicle
vpj-cd_: where the enprice consulting make all the tes of bisness process
vpj-cd_: that they choose
croo: yes but I think we do not QA to 3.2... but to 316 which is a RC1!
CarlosRuiz: Colin, my feeling is that stability was compromised with integrating Compiere changes, the work I did
CarlosRuiz: really the add-ons were around Adempiere, not touching the core
hengsin: one of the issue is many potential partner/user is not testing until we declare a stable branch ...
CarlosRuiz: maybe is not unstable, just we catched 3-4 bugs and fixed, and now is stable
vpj-cd_: Croo the QA test is to get 3.2
vpj-cd_: we can release 3.1.5 3.1.6
vpj-cd_: until we have the real stable
vpj-cd_: but this very import tell that we are find get 3.2
hengsin: or do a more frequent maintenance release for the 'current stable branch', stable as in we are committed to fixed bug fast :-)
croo: victor I say to not to 3.2 because I think only the community can really determine the quality ... if the goal is an intensive test to produce something that the community can easily test then I say that version is not fully testsed hence not 3.2 .. which we want to be rock solid
hengsin: so for instance, we can now release or 3.1.6 to incorporate recent bug fixes and at the same time provide a upgrade pack for 3.1.5
croo: yes hengsin people have suggested nightly builds ... that would be a very good idea
vpj-cd_: croo the community are all
vpj-cd_: also enterproce consulting
croo: yes victor but many in the community do not have the skills to compile & test ... so only the likes of us test!
CarlosRuiz: hengsin, we have provided "upgrade pack" always
vpj-cd_: but the real bug are show when you have good group of test
vpj-cd_: or when you are in implementation
vpj-cd_: but test the any software is neessesary know to detail the software
hengsin: another alternative to encourage more installation/testing from user, is to fast track to release as 3.2 to give user a stable feeling and we declare that as a long life version where we will release frequent maintenance/patch
CarlosRuiz: Victor, what happened with the matrix test in wiki?
CarlosRuiz: I think that's more important than user manual
vpj-cd_: I think the way to give a good suport , fast , easy is via 2pack
croo: hengsin, well being sour first stable version and being our goal was the bazaar is the best way to quaklity I think we need to pay special attention to this first release .. many will qtch and judge according to it!
trifon [n=chatzill@] ha entrado en la sala.
hengsin: carlos, we don't, we only provide a db upgrade pack, not program upgrade pack :)
vpj-cd_: if we set a
trifon: hi, everyone i will be just for 4-5 minutes, sorry
croo: qtch=watch
vpj-cd_: and make easy the work to pach your code then the people can continue with your test
CarlosRuiz: program upgrade pack is - download - uncompress - RUN_Setup
croo: carlos, yes it can be done in 10-15 minutes ... if you know what you are doing!
croo: but they'll want to test the migration too .. which can still be a little tricky
croo: i also think companies are reluctant to upgrade their system .... if it's working don't touch it is typical
CarlosRuiz: "make easy the work to pach your code" -> patches.jar - is really light and easy
Mike_Judd ha salido de la sala (quit: Read error: 104 (Connection reset by peer)).
croo: for a big org the risk that there is a problem is just too great to take
croo: easy to do carlos but the risk is the problem ... if it doesn't work or worse klooks like it works but does something esle
CarlosRuiz: Victor, what happened with the matrix test in wiki?
croo: the risk is the same
CarlosRuiz: I think that's more important than user manual
Mike_Judd [] ha entrado en la sala.
croo: i think in the end the matrix would be one of people using which models with which env .... in the end was it not?
CarlosRuiz: no, it's a functional matrix
hengsin: carlos, I would prefer each patch should change the version number to keep things straight forward.
CarlosRuiz: hengsin, currently the "release" work is hard
croo: hengsin we could add the reveision to the version number now?
hengsin: croo, not at the moment.
CarlosRuiz: yes, we could
vpj-cd_: yes I think we can add to release
CarlosRuiz: we could patch always, or put the release number in a different java file
croo: we don't but we could say 315_1816... right?
vpj-cd_: but I think we can use 3.1.5 to a big bug
CarlosRuiz: vpj-cd_ what happened with the matrix test in wiki?
croo: carlos, i didn't think that was agreed???
vpj-cd_: and to each small bug
hengsin: yes, we could or it can be for e.g
vpj-cd_: Carlos I set today
croo: Carlos I thin the test matric would be a monster ... today I test the Sales Process ... and it works ... 30 minutes later someone submits a patch and perhaps it doesn't work now but the wiki says it works
croo: and perhaps it works with 10g but not XE or postgresql
hengsin: I would suggest if the current trunk is of sufficient stability, fast track to 3.2 so that we can attract more testing and implementation from the community, it doesn't seems to happen with the current strateyg.
CarlosRuiz: Colin, but it will be a guide to testings, and to know what parts of the system are tested
croo: or a false promise to people!
CarlosRuiz: Victor, how detailed will be the matrix?
hengsin: and we declare that as a long life version to give confidence and also commitment.
croo: i think we should pay special attention that 3.2 really is stable ... not something to be tested by the community but smoething that is production ready as it is
vpj-cd_: the detail is a list of bisness process
CarlosRuiz: I think if we organize enough the testing work, people will know how to contribute
CarlosRuiz: that's the importance of the matrix testings
vpj-cd_: and idea is that we have change to testing
vpj-cd_: tester
CarlosRuiz: obviously better if we have a leader for such task, and the maintenance of the wiki matrix
CarlosRuiz: but if we at least have the matrix is ok
vpj-cd_: yes I think =
CarlosRuiz: ok, the matrix will have every event/doctype?
croo: well I originally sugegst a matrix but I don't believe it will work ... but instead become a maintenance nightmare
CarlosRuiz: every window/process/report/task ?
red1 ha salido de la sala (quit: ).
vpj-cd_: each bisness enterprise can set people to charge
vpj-cd_: and make your test
vpj-cd_: also any of commnunity is welcome
vpj-cd_: no Carlos
vpj-cd_: adempiere is focus in bisness process
vpj-cd_: no in modules or windows or task
vpj-cd_: we need test ie cicle to quotation
vpj-cd_: sales
vpj-cd_: ship
CarlosRuiz: ok, thanks that Colin is here :-) can we tell a little about testings?
vpj-cd_: invoice
vpj-cd_: cash
vpj-cd_: we have 2 type of result
vpj-cd_: 1 some bug in the process
vpj-cd_: and is sure that we have sugestion to new inproves
CarlosRuiz: invoice -> you can test - new, modify, delete
CarlosRuiz: but you can also test complete/reverse/void/post
CarlosRuiz: and you can test complete with every different type of invoices
CarlosRuiz: but if you put "sales proces" in the matrix
CarlosRuiz: this is very big to help people in testings
CarlosRuiz: I'm thinking on a matrix test with little functionalities
vpj-cd_: yes Carlos is good idea
vpj-cd_: then I fist se the structure to tests
vpj-cd_: and set more detail to test
vpj-cd_: in each process
CarlosRuiz: at least we will know what is tested and what not
vpj-cd_: I know what is the problem in proces
vpj-cd_: ie need add more parameter to standard report
CarlosRuiz: and all community will know in the best openness of this project
croo: carlos, only in our small subset of processes sytems & dbs
vpj-cd_: need review if reserve qay is working
vpj-cd_: qay qty
CarlosRuiz: look, stable doesn't mean NO BUGS - stable means enoughly tested with the most common cases
CarlosRuiz: with three databases supported, and every parameter changing the internal behavior is impossible to test all possibilities
hengsin: or else we will never release and all interest gone ...
CarlosRuiz: but we can ensure at least that common processes works well in a normal environment
croo: exactly .. so why say THIS works ...
croo: I say we say there are no serious bugs reported in THIS
croo: not the same thing!
CarlosRuiz: we can think "normal environment" is GardenWorld and make the testings with such company
trifon ha salido de la sala (quit: Read error: 110 (Connection timed out)).
CarlosRuiz: agree Colin - or "tested in commonest case"
croo: tested by the community!
vpj-cd_: mmm I have doubts
vpj-cd_: that commnunity
vpj-cd_: test
croo: we can do our tests locally how however we do .. when we are done yes we shuold have picked up many problems ...
CarlosRuiz: I think there are people eager to help, but don't know how
vpj-cd_: generality the ppl wait to have release stable
croo: but it's only when the coommunity is tested by using with no bugs reported that we say stable!
vpj-cd_: only all that are interst in the project
croo: and if the community doesn';t test then the bazaar doesn't work and we can all sign up with compiere!
vpj-cd_: only they can test the adempiere
CarlosRuiz: :-D
croo: :)
hengsin: croo, that is too idealistic
croo: well none of us have worked ina situation like this before ... i think whats' proposed is ghow it is done in a closed house ... but we are trying to do somethng different
croo: so we need a different approach
croo: but i see I am fighting 1 ti many here :)
croo: you guys are the CC ... i'm just heer to stir it up :)
CarlosRuiz: :-) Colin, you're the QA
croo: i guess i can summarise as follows ...
hengsin: croo, it is because we are doing it in the bazaar way, we need to also strategize how to get more feedback/testing from the community. I
CarlosRuiz: ok, what if we summarize the matrix test
vpj-cd_: I think the best way to motivate the community is give tools
vpj-cd_: to poach your code faster
CarlosRuiz: a "brute force" testings is not really needed
vpj-cd_: then the commnunity feal is listen
CarlosRuiz: maybe we can do "full/complete testings" in just one of the windows
croo: carlos, hengsin is QA too!
CarlosRuiz: most complex window
CarlosRuiz: i.e business partner or product
hengsin: croo, thanks for remining me :)
CarlosRuiz: and test complete functionalities
CarlosRuiz: insert/update/delete/zoom/preference/chat/attachment/report/print format/archive ...
CarlosRuiz: in just ONE REPRESENTATIVE window
CarlosRuiz: or 2, ok to test children table :-)
croo: ok but I guess you will weant to test all processes anyway before you implement for your customer right?
croo: and so will the rest of us!
croo: so I what I am suggestnig is we do that
CarlosRuiz: not all processes, significant processes
croo: log the bugs
croo: release to the community as a hopfully very stable rc1
vpj-cd_: I think we need focus to process
vpj-cd_: main process
vpj-cd_: no only in windows
croo: we wait and when bugs are reported fixe them ... then when we reach a point of no "serious" [to be defined] bugs we simle call it stable
CarlosRuiz: there are 172 processes (50 in menu and the rest embedded in windows)
CarlosRuiz: Victor, both
vpj-cd_: the mani bugs are find when you try make some diffrent
vpj-cd_: or need solve a spesifc problem
CarlosRuiz: Colin, we're in such point, and people are not reporting
CarlosRuiz: yes Victor, but we need to release a stable version, and stable means that you guarantee normal behavior
vpj-cd_: exist some way the get aventage the log the each test
croo: i haven't even started testing yet!
vpj-cd_: we can collect all the log and review
vpj-cd_: to find bugs
CarlosRuiz: don't exist company or strategy that can guarantee no bugs
croo: no I agree
croo: but you can have aproactive approach or reactive approach .. look for the bugs or fix those found
croo: i suggest with a app this big with this many variable it's not possible to find them all by simply lokoing and testing
croo: so we play a reactive role
CarlosRuiz: or a medium point
croo: but we don't want to say stable until it's been through the process ... so a few implementation ... the bugs we be found before go live I'm sure
CarlosRuiz: we can make/ask for basic testings (with a summarized matrix)
croo: well Joel suggested a good middle ground
croo: we post or for example I post
CarlosRuiz: I mean, we can be proactive with basic testings - declare it stable enough, and be reactive with the rest of platforms
hengsin: croo, you see part of the problem is the contributor pool is growing at a very slow rate, we need a strategy to grow that faster
croo: I etsted this in an Irish business env with sales, pos, shipping & distribution
CarlosRuiz: I would feel comfortable with basic testings on Oracle XE / 2 windows to test all window functionality / and complete 16 types documents
croo: I thought this was a good middle ground .. and we could great a much smaller less probelematic matrix!
croo: to give people some idea of what does and doesn't work!
hengsin: a prolong qa process will not help and a potential bottleneck for development resources too .
CarlosRuiz: look
CarlosRuiz: what I'm proposing is a summarized matrix that at least guarantee that system will work with Oracle-XE and most common processes
vpj-cd_: also postgresql
CarlosRuiz: we're not going to guarantee that jasper works
vpj-cd_: I think the test also use postgresql
croo: re: development bottleneck ..w ell that's why I suggested the RC1 be a new branch ... then development could continue in the trunk .. but only patches would be applied to the RC1 branch
CarlosRuiz: if there are testers for postgres, ok, but main database supported is oracle
vpj-cd_: mmm
vpj-cd_: Carlos I think the community prefered postgresql vs postgresql :-)
CarlosRuiz: is our inherited reality, we're not going to liberate from oracle so easily
CarlosRuiz: it's not what community thinks, is our reality, main database supported is oracle
vpj-cd_: postgres vs oracle
hengsin: croo, the bottleneck i mean is limited number of active committer :)
CarlosRuiz: if community wants postgres, ok, then community must come and help stabilize the port
CarlosRuiz: but I know for sure, that JJ approach of database independence is really loose
vpj-cd_: the data base is important in base who or the comunnity use
CarlosRuiz: I don't think we can achieve a really good port on any database without dropping the Convert layer
vpj-cd_: then I use postgresql is more importnat to me :-)
vpj-cd_: if you use oracle xe this is more importat to you
CarlosRuiz: no Victor, database is a commodity, ERP is supporting enterprise processes
vpj-cd_: then the comunnity have the free to test in the database they wnat :-)
CarlosRuiz: of course!!! I'm not stopping any effor on such direction
croo: yes ... it's evolutionary rather than revolutionary
CarlosRuiz: but I clearly now that supported database is oracle
CarlosRuiz: and postgres is a patche
CarlosRuiz: a bad patch
vpj-cd_: ie is is sure Usman and me meked the test use postgresql
croo: i think after postgresql we cshould concentrae on jpa of some sort .. but that's another story :)
vpj-cd_: because we are veri interst the pach any bug find in postgresql :-)
vpj-cd_: we are working with postgresql in production
CarlosRuiz: yes Victor, with the help of Hengsin postgres port is very stable at this moment
vpj-cd_: the no problem
vpj-cd_: :-)
CarlosRuiz: but personally, I always recommend my customers oracle
croo: i am interested in testnig postgres too but only after I complete my current implemenmtation with oracle :)
CarlosRuiz: if customer wants postgres, I advise him of the risk, and go ahead
hengsin: my primary test environment so far is postgresql but carlos is right, oracle will be inherently more stable and faster for adempiere.
croo: but I would be very happy to see the app 100% postgres and less on oracle!
CarlosRuiz: Colin, for that we need to drop Convert layer
croo: yes I think what you guys have been doing is great ...
croo: but like you say it's a patch
vpj-cd_: I have good example with my customer current
vpj-cd_: we take the leave with postgresql
vpj-cd_: this desition is togeter e-evolution and sismode
vpj-cd_: now he are happy because he before test with oracle
vpj-cd_: now he have not pay 12000 usd to your server
CarlosRuiz: yes, I have a customer also that preferred postgres, I said him the risk, I committed to help them with any problem that arise, and we're going to install with postgres
CarlosRuiz: it's his decision supported by us
vpj-cd_: and he could set this money in consulting :-)
croo: :)
CarlosRuiz: but we know that postgres is a patch, is not well supported
CarlosRuiz: that must be clearly explained to customers
vpj-cd_: yes is support to e-evolution
vpj-cd_: :-)\
croo: yes it's their risk to take
vpj-cd_: is the warranty
vpj-cd_: I now am invoice
vpj-cd_: take sales order
CarlosRuiz: I thought was supported by Hengsin  :-D
vpj-cd_: generate accounting
vpj-cd_: etc
vpj-cd_: and without some problem
vpj-cd_: yes Low maked a great work
CarlosRuiz: just joking, and acknowledging real good work made by Hengsin to stabilize the convert layer for postgres port
vpj-cd_: but we take the risk :-)
CarlosRuiz: agree Victor, and as I said you yesterday, we're trying to help you with your installation, is strategic
CarlosRuiz: I think we must put top priority to bugs you catch
vpj-cd_: yes, we get the a bazaar
vpj-cd_: Low and me work to postgresql in adempiere is reality
vpj-cd_: now we are take the risk to leave in production
vpj-cd_: with adempiere
vpj-cd_: I know that I have the support the bazaar
CarlosRuiz: yes, and we're supporting your project, Victor
vpj-cd_: this the great diffrent vs Compiere :-)
hengsin: thats my point, we fast track to release 3.2 and declare that as a top priority version. it is only through feedback from more implementation that we can get a real stable version.
vpj-cd_: ok
CarlosRuiz: and I also want a stable postgres, just doing a reality check here ;-)
CarlosRuiz: but every day I feel closer to a stable postgres
vpj-cd_: ok I need are out
vpj-cd_: I return in 10 minutes
CarlosRuiz: ok, what do you think on a summarized matrix?
hengsin: victor is in CC, understand the situation and implement using 315, many other from the community won't do that. What happen now is we are actually getting very limited testing by the community if we don't fast track to 3.2 and declare that as a priority/long life version.
CarlosRuiz: yes, also Victor is confident enough on having the resources in e-evolution to face problems, but other companies don't have that
CarlosRuiz: ok, what do you think on a summarized matrix?
hengsin: to me, it is also a way to pull more people to jump to the ship, hopefully that will grow the contributor pool and make the platform better!
vpj-cd_: thanks every body
CarlosRuiz: Hengsin, what I don't want (and understood that neither Colin) is to declare a version as stable, and then confront critical bugs there
croo: i think a summarized matrix that allow peopel to indicate the env & business processes the use anmd work with is good
vpj-cd_: I will set in wiki the matrix test
vpj-cd_: bit I divide in bisness process
vpj-cd_: and set subgestion to test in each process
croo: i don't think we should call an untested version a stable 3.2
hengsin: but it might be a risk that we have to take to grow the community ... we can't be of 100% sure anyway :-)
vpj-cd_: I think that 3.1.5 is more stable that any of Compiere :-)
croo: i think one obstacle to testing is the version changes with lots of new functionality every month ....
croo: theis freeze to test is a very good idea
hengsin: croo, it is not untested ...
CarlosRuiz: Victor
CarlosRuiz: one questions
CarlosRuiz: is 315 tested in your project?
CarlosRuiz: enoughly tested to declare it stable?
vpj-cd_: yes I am use
vpj-cd_: 3.15
vpj-cd_: and util now is working fine
vpj-cd_: but we are begin
vpj-cd_: with invoice
croo: there are a whole lot of open bugs for a stable version! some quiet serious!
vpj-cd_: shipm,ent
vpj-cd_: payment
vpj-cd_: purchace
vpj-cd_: etc
vpj-cd_: we find some error
vpj-cd_: but we have in control
CarlosRuiz: I would feel confident with the QA cycle made in Ecuador in this stage of the project
CarlosRuiz: are you testing trunk or 315?
CarlosRuiz: or 315 modified?
vpj-cd_: I am use
vpj-cd_: the las 315
vpj-cd_: last 315 with some modify
vpj-cd_: to reverse docuemnts
vpj-cd_: and print document the copy info
CarlosRuiz: it's ok, those are small patches not touching the core
vpj-cd_: and with seed to libero :-)
vpj-cd_: no
vpj-cd_: the core is without cahnge
vpj-cd_: change I have make some improve
vpj-cd_: in form as add category
vpj-cd_: and group
vpj-cd_: in product and bpartner info
CarlosRuiz: ok, Colin, Victor, Hengsin, can we change the question?
CarlosRuiz: what do we need to feel confident and declare a version as stable?
vpj-cd_: I think the key
vpj-cd_: is have a adempiere 3.1.5
vpj-cd_: in production :-)
CarlosRuiz: Victor is making me more confident on the stability of 315
vpj-cd_: then we can tell is stable
vpj-cd_: :-)
CarlosRuiz: maybe as Hengsin said, we can release 316 as Release Candidate 1 - very close to stable
hengsin: another food for thought, for outsider or non-active community member, it give the impression compiere is more stable, they declare 263c as a production release and we only have 315 which is a version release for 'testing'
croo: i would say first [and again] that given we split on a issue of openess whcih effected quality we need to make a evry stable release
croo: I like victors suggestion that some 315in production or about to go live
croo: PLUS
croo: I would say we go through the bugs reported and set the priorities right
croo: if it corrupts the data it's very serious
CarlosRuiz: Hengsin, Compiere will never declare a version unstable, they have marketing worries that we don't have
croo: if it's a UI annoyance less so
croo: when we have no serious bugs open woudl be a good point for me!
CarlosRuiz: we have openness commitment
CarlosRuiz: honesty commitment
croo: ok I get your point now ... yes I agree with must be honest and not say stable just to sell it!
croo: or anyone else it seems looking at their forums these days :)
vpj-cd_: ok I return
CarlosRuiz: ok, are we ready to make conclusions? not yet?
croo: well I've said my piece
CarlosRuiz: I think we can give Victor another week for testings in Ecuador and then release 316 as release candidate
croo: saying it's ready for implementation!
hengsin: actually, what is important is to declare a long life/supported version
hengsin: that's what potential customer want
croo: well by the time popel get to the implementation point I believe 3.2 ... lo=ife long supported version should be ready
croo: their implemanttaion will ahve tested?
CarlosRuiz: what do you mean with long life/supported version?
croo: get to go live that is
CarlosRuiz: can we commit with that?
croo: sorry ... a mistype I mean what hengsin said ... long life/supported
hengsin: you see the problem with compiere is that they release 260a then 260b then 260c ... there isn't a version that is maintained for a long period of time with just bug fixes and no enhancement.
croo: well I expect we who implement will be offer the commitment .. so we can support and patch if needed
croo: yes hengsin I agree
croo: we need stable and for a lonmg time .. 2 years
CarlosRuiz: me too, and as Colin said, it's a matter of who implement
CarlosRuiz: to support a version 2 years, very probably implementor must charge that for the project
CarlosRuiz: but my point is
CarlosRuiz: can we (as Adempiere commit with that)? I think not
croo: yes of course ... a % of original oproject is typical!
CarlosRuiz: we can commit to give the implementor the playground to support that
hengsin: and that's what is the most important, a long life version for implementor/service provider. For a complex system like Adempiere, you need that as artificial testing would not be sufficient.
croo: well we will maintain 3.2 on the svn with the ability to patch it .. perhaps SP type reelases every 6 months?
CarlosRuiz: what's SP type release?
croo: like windows .. does their SP1 or SP2 ... so I mean we packages the bugs
CarlosRuiz: ah, ok, service pack :-)
CarlosRuiz: but again
CarlosRuiz: I think we can't commit with that
CarlosRuiz: not at this moment
CarlosRuiz: we can commit to allow implementors tools / mechanisms to provide that service
hengsin: yes, periodic maintenance release for the last production or 'ready for implementation' release.
CarlosRuiz: do you get my point?
teo_sarca ha salido de la sala (quit: Read error: 104 (Connection reset by peer)).
teo_sarca_ [n=teo_sarc@] ha entrado en la sala.
teo_sarca_ ahora se llama teo_sarca
croo: ok carlos
CarlosRuiz: so, I don't know if we must declare a long/life version
CarlosRuiz: it will be if there are maintainers
croo: i guess those who use a release can produce the patch releases if need be
croo: well i guess what we want is we decalre this as the version that shoudl be implemented
hengsin: ok, maybe 2 year is too long for the bazaar, maybe at least for 6 months or 3 maintenance release.
hengsin: beyong that, it is the implementation vendor's responsibility.
croo: well I imagine that when I implement that nobody will want to upgrade for along time
croo: the whole cost of the prohject is now the implementation
hengsin: or alternatively, we can ask the community member to sponsor the bazaar to support that for 2 years ( might be difficult to implement )
croo: so people won;t wantt to be paying that every few years!
CarlosRuiz: Hengsin, we're very busy trying to keep the pace on trunk. I think we can't commit to that, maybe as a paid service
croo: or I mean every 6 months
CarlosRuiz: what we (as CC) can do is to provide implementors the mechanisms to support older versions and to release service packs
croo: i think this is something that will just happen ... in a few years there wikll be people on 3.2 3.4 or 3.6 and who knows what!
croo: supplied of 3.2 will provide the support! and their customers will pay thema yearly support charge for it ...
hengsin: carlos, but otherwise, this artificial declaration of a stable or production ready release if of little meaning
croo: that's how it works here anyway!
hengsin: I mean is of little meaning.
CarlosRuiz: Hengsin, that's what we're discussing, we're not going to do an artificial declaration
CarlosRuiz: or I misunderstood you
hengsin: well, to me, declaration base on artificial testing is an artificial declaration but that is just my opinion.
CarlosRuiz: I don't understand
CarlosRuiz: what's the meaning of artificial?
hengsin: testing based on a test plan/test script
hengsin: not field testing
CarlosRuiz: ah, ok, artificial has another meanings that I don't like :-D
CarlosRuiz: almost in every software QA cycle is made with test plan/test script, not in real field
vpj-cd_ ha salido de la sala (quit: Read error: 113 (No route to host)).
hengsin: yes and we all know the limited realibility that will provide and that's why the important of a supported version.
hengsin: only after feedback from field testing a software can achieve real realiability which typically means after 2 or 3 maintenance release.
hengsin: and most sucessfullt opensource project DOES provide maintenance release.
hengsin: ok, maybe this is something beyong the scope of CC, rather should be discuss by BDC or the bazaar councile.
CarlosRuiz: Hengsin, I'm not against providing that
CarlosRuiz: obviously the better will be to have SP's
CarlosRuiz: just another reality check, is that achievable with current resources?
hengsin: to me that is more important than having a fast moving trunk, compiere have a fast moving trunk and like what colin have say earlier, we can all sign up with compiere if that is what we want :-)
CarlosRuiz: ok, conclusions?
CarlosRuiz: 1 - Victor is making testings in Ecuador on 315 version in a real implementation
CarlosRuiz: 2 - We're going to wait for another week of bug reporting
CarlosRuiz: 3 - then release 316 as release candidate - safe for implementations with commitment of quick bug fixings
CarlosRuiz: what do you think?
hengsin: ok, agree
Mike_Judd: I have one - I've just gone live on 314 ;)
Mike_Judd: It would be good if we could get a few live implementations on the same version - that way we could bug fix together ...
CarlosRuiz: that's good Michael
CarlosRuiz: are we going to keep the feature freeze?
crooney [n=crooney@] ha entrado en la sala.
Mike_Judd: was that a question for me ? or for the group carlos ?
hengsin: carlos, it is the same supported version thing
crooney: hmm I didn't get a disconnect message but it all went quiet there!
CarlosRuiz: for the group Michael
Mike_Judd: thank ;) - didn't feel qualified to answer ...
vpj-cd_ [n=Horus@] ha entrado en la sala.
CarlosRuiz: copying for Colin:
CarlosRuiz: ok, conclusions?
CarlosRuiz: 1 - Victor is making testings in Ecuador on 315 version in a real implementation
CarlosRuiz: 2 - We're going to wait for another week of bug reporting
CarlosRuiz: 3 - then release 316 as release candidate - safe for implementations with commitment of quick bug fixings
CarlosRuiz: what do you think?
crooney: thx
CarlosRuiz: Michael, if we declare 316 as enough stable you're going to upgrade your customer to 316?
Mike_Judd: yes - happy to do that .....
Mike_Judd: it makes sense .....
crooney: what about the existing bugs? there are a lot and some serioud .. should be saw not a week but when eh serioud ones are gone
CarlosRuiz: that can be our common point, Victor-316, Michael-316, and I have three projects here, I can put them in 316 too
CarlosRuiz: Colin, like?
crooney: it takes me longer than a week to test! but I will restart my implementation and test as 316 be fully expect to actually go live with 32!
crooney: hold on I'll look
crooney: theer are product costing bugs...
CarlosRuiz: or we can just release 316 this week, 315 has a bad bug related with Jasper that can't be patched with patches.jar
crooney: incorrec tosts on shipment & MM
croo ha salido de la sala (quit: Read error: 110 (Connection timed out)).
crooney: phew .. I'm crooney now .. don;t worry I'm still here!
CarlosRuiz: 106 open bugs
hengsin: yes, I created that, to be verify again after carlos have imported some patch from compiere ( it have been fixed in compiere ).
crooney: anyway I just say that we usethe bug list to determine ... perhaps some can be marked down in priority .. if they do not corrupt data or stop a process
crooney: they they are not that serious?
hengsin: many open bug in the SF tracker is not verify
crooney: really? I thought they were... it's mostly well known names logging them!
CarlosRuiz: hmmm, there are duplicated, there are also opened by phib brought from compiere 253b, not tested currently
crooney: i just think if I was thinking of implementing I would check the bug list!!!!
CarlosRuiz: maybe a cleanup work in those 106 bugs will be good
Mike_Judd: is there any way we can allocate off the bug list to people (like me) to verify and rate them ?
crooney: we could assign them to you? or if people review them and choose ones they think they can do and assign them to themselves
crooney: then if I find one with say mikes name and thinkn I can help I can contact mike to co0ord with him
CarlosRuiz: Michael, that will be a relief if you can help us
CarlosRuiz: maybe we can split the list between volunteers
hengsin: also some reported without sufficient detail ...
crooney: what about first agreeing how to prioritize? the priority is very important
CarlosRuiz: the first must be review if the bug is applicable, if it's put priority, if not close it
Mike_Judd: yeah - I'm not ableto be allocated bugs at the moment .... but I'd happily go through a few and close them out - I agree that outsiders look at the bug list as a indication of quality of the product ...
CarlosRuiz: we can include Michael in the list to be assigned
Mike_Judd: so 1. assess bug, 2. rate bug, 3. allocate bug ?
crooney: we have 9 levels of prioirity ... but I think perhaps only 4 or so are needed
Mike_Judd: 3 - to best person to fix it - or perhaps queues ?
CarlosRuiz: I think initially just assess and rate
CarlosRuiz: assess and rate/close
CarlosRuiz: Colin, what's the proposal?
crooney: ok mike I added you to technician and admin
Mike_Judd: thanks .... I think ;)
crooney: well as mike says 1. assess the current buys
crooney: assign priority
crooney: highest is the system won't work .. i.e. it cannot be started 9
karsten_thiemann [] ha entrado en la sala.
CarlosRuiz: 9 - system stopper
CarlosRuiz: 1 - presentation bug
crooney: next is a process won't complete or data corruption
hengsin: croo, data corruption is a show stopper - 9
CarlosRuiz: 9 - system stopper / data corruption
crooney: depends on what is being corrupted, and if it can be easily corrected
CarlosRuiz: another variable -> it has workarounds
crooney: yes ...
CarlosRuiz: but I agree, data corruption is a stopper
teo_sarca ha salido de la sala (quit: Read error: 110 (Connection timed out)).
CarlosRuiz: if a process corrupting data has a workaround, then we must disable the process until fixed
crooney: a single process might cause corruption but you can still run the rest of the system ... so I do not see corruption as ashow stopper per se
crooney: yes again carlos
CarlosRuiz: security issues?
crooney: yes also high
crooney: not 9 though .... the system can still be used
CarlosRuiz: performance problems -> lower
karsten_thiemann: if you know the risks
hengsin: maybe 6 or 7
CarlosRuiz: and there are reported bugs that are feature requests or contributions
CarlosRuiz: they must be moved
CarlosRuiz: 9 - system stopper - data corruption without workaround
CarlosRuiz: 7 - security issues
CarlosRuiz: 3 - problems with workarounds - performance problems
CarlosRuiz: 1 - presentation problems
CarlosRuiz: what do you think?
vpj-cd_: Carlos I think shoukd exist to bad manage transaction
crooney: yeah that looks good ... can we set the default priority to be somethnig other than 5?
hengsin: ok for me.
vpj-cd_: exist some code here exted of compiere
crooney: say 1 by default :)
crooney: transactions problems could be considered corruptions?
CarlosRuiz: maybe it's ok, 5 is "not prioritized"
crooney: hmm yes
crooney: good thinking batman!
hengsin: carlos == batman ?
crooney: or we just have status for that ... open, confirmed, assigned
crooney: just an expression .. well here it is!
CarlosRuiz: no, ok, if its <> 5 it's prioritized
vpj-cd_: the problem and bad manage transacion
vpj-cd_: is he stop system
***CarlosRuiz wondering where Robin is
vpj-cd_: with a idle transaction
vpj-cd_: and may also generate corrupt data
vpj-cd_: or inconsistences
CarlosRuiz: trx problems can corrupt data or just don't save
CarlosRuiz: so it must be different depending on the case
vpj-cd_: yes
CarlosRuiz: ready, so that's the list, can I put again all trackers in 5 and start the work
crooney: ok
hengsin: ok, agree.
CarlosRuiz: done, I suppose we're going to receive lots of notification e-mails
crooney: so next question is ... at what elevel of bug do we say must be fixed before we are stable?
crooney: obviously 9
crooney: 7 & 3 too?
CarlosRuiz: I think just 9
crooney: ok
CarlosRuiz: opened 7 is like karsten said, known risks
hengsin: only 9 should be a MUST fix.
crooney: ok
CarlosRuiz: ok, how can we divide the work, to avoid two people working on prioritization on the same bug?
Mike_Judd: yes - that would be good ...
Mike_Judd: how to ? on functional areas ?
hengsin: by process or module ?
CarlosRuiz: who are the volunteers?
CarlosRuiz: 1 - MichaelJudd, 2 - CarlosRuiz
Mike_Judd: I'm volunteering ....
crooney: well this is just a quick review to assign the correct priority right?
CarlosRuiz: yes
crooney: but youy can count me too
teo_sarca_ [n=teo_sarc@] ha entrado en la sala.
teo_sarca_ ahora se llama teo_sarca
CarlosRuiz: colin, some must be tested if the apply
hengsin: ok, I can help too.
CarlosRuiz: if don't apply close it
crooney: teo logged most of the bugs :) so he help too :)
crooney: only joking
CarlosRuiz: ok, 4 volunteers for 106 bugs
hengsin: teo, can u help the review too ?
karsten_thiemann: maybe devide the work by ending no xx1 / xx2 for carlos xx3 /xx4 for colin ...
crooney: good idea karsten ... but what if someone doesn't know the area? just leave it 'til the end and hand it to another?
karsten_thiemann: hand it to another - but by that way it's clear which bug belongs to whom
Mike_Judd: functional area ?
CarlosRuiz: Michael, that can be for solving, but prioritizing can be done as Karsten proposed
crooney: ok well we could ay the accounting are for mike and just split the rest between us?
Mike_Judd: thats not really that fair - are there that many in accounting ?
crooney: loads! - lol
Mike_Judd: can we prioritise and allocate at the same time ?
CarlosRuiz: just 3 categorized as accounting
crooney: hhinestly I don;t know
CarlosRuiz: no, I don't think so
CarlosRuiz: possibly we need to allocate just priority 9
crooney: well I think it's best if people chose the bugs they will solve
CarlosRuiz: but this is volunteer also
hengsin: lets start with what karsten have proposed and we can discuss on IRC, forum or mailing list for bugs that we are not sure of.
crooney: by assign it to themselves
CarlosRuiz: ok ending numbers
crooney: ok
Mike_Judd: ok ....
CarlosRuiz: ending from 00-25 - Michael
CarlosRuiz: 26-50 - Carlos
CarlosRuiz: 51-75 - Colin
CarlosRuiz: 76-99 - Hengsin
CarlosRuiz: do you agree?
Mike_Judd: ok ....
crooney: ok
CarlosRuiz: maybe the distribution will be not exact, but I think is ok
hengsin: ok
CarlosRuiz: target date?
crooney: asap :)
crooney: i guess it depends on the bugs we meet ... i have no idea myself
CarlosRuiz: ok, do you agree target date march-7 ?
hengsin: :)
crooney: tomorrow
CarlosRuiz: hehehe, in Malaysia is today
crooney: well that's only a few hours away now
CarlosRuiz: ok, 48 hours from now
hengsin: yes, it is already due here :)
crooney: :)
crooney: ok
crooney: at least by then we will know what's involved to get to the end!
hengsin: ok
CarlosRuiz: good
CarlosRuiz: any other theme? is late in Kuala Lumpur
Mike_Judd: no probs - will do mine tonight ...
karsten_thiemann: just to let you know
karsten_thiemann: the university agreed to give us the rooms for the conference
crooney: hip hip....
crooney: hooray! :)
karsten_thiemann: so we will check them on friday
hengsin: which mailing list we shall use to discuss any bugs that we are not sure of ?
CarlosRuiz: I think the tracker itself
karsten_thiemann: the prof said they like to cooperate with us
crooney: nice
kontro ha salido de la sala ("Leaving").
crooney: i have one point to raise
crooney: earlier in the irc huehner suggested some svn command to assign execute status & eol change s to files
hengsin: I means to post to so that the other 3 person can take a look
crooney: so we don't have to run the find | chmod
CarlosRuiz: yes Colin, I'm asking there
crooney: ok sorry
CarlosRuiz: I didn't ignore him :-) just that we were busy here
crooney: ok I didn't notice ... no problems I told him I'd say it here
CarlosRuiz: Hengsin, I think comments can be put in the tracker, so it will notify subscribed list
hengsin: ok.
vpj-cd_ ha salido de la sala.
CarlosRuiz: ok, thank you very much, I think we can finish this meeting, very productive again, thanks
crooney: yes thx all
Mike_Judd: thanks ...
hengsin: thx all
karsten_thiemann: bye

Personal tools