meetings/irc: opensync-20081211.log

[Thu Dec 11 2008] [11:02:19] Topic      dgollub sets the channel topic to "OpenSync Project IRC Meeting | please not support question or bug reporting during this time.. thanks".
[Thu Dec 11 2008] [11:02:29] <dgollub>  let's start...
[Thu Dec 11 2008] [11:02:34] <dgollub>  whos present?
[Thu Dec 11 2008] [11:03:11]     * dgollub present
[Thu Dec 11 2008] [11:03:22]     * bricks present
[Thu Dec 11 2008] [11:03:26] <dgollub>  1. Action Items from previous Meetings
[Thu Dec 11 2008] [11:03:31] <fm__>     fm
[Thu Dec 11 2008] [11:03:33] <dgollub>    * AI dgollub: find OpenSync-Doxygen Maintainer
[Thu Dec 11 2008] [11:03:33] <dgollub>    * AI bricks: move API documentation to headers
[Thu Dec 11 2008] [11:03:33] <dgollub>    * AI bricks: porting frontend/consumer API examples to 0.40 API
[Thu Dec 11 2008] [11:03:33] <dgollub>    * AI dgollub: introduce simple & advanced plugin example
[Thu Dec 11 2008] [11:03:35] <dgollub>      (simple example: without hashtable, ...)
[Thu Dec 11 2008] [11:03:38] <fm__>     i am here too
[Thu Dec 11 2008] [11:03:57]     * ianmartin present
[Thu Dec 11 2008] [11:04:01] <dgollub>  reg. OpenSync-Doxygen Maintainer ... i guess paul is already doing a great job
[Thu Dec 11 2008] [11:04:16] <dgollub>  so we just put him in this role ;)
[Thu Dec 11 2008] [11:04:24] <bricks>   dgollub: would be nice if he can do this
[Thu Dec 11 2008] [11:04:48] <dgollub>  i guess he's just shy :P .. so applaus everyone for our new doxygen maintainer: Paul ;)
[Thu Dec 11 2008] [11:04:50] <bricks>   i made many changes for doxygen but don't want to maintain it
[Thu Dec 11 2008] [11:04:57] <bricks>   :)
[Thu Dec 11 2008] [11:05:20] <bricks>   i started to change the layout
[Thu Dec 11 2008] [11:05:22] <dgollub>  i committed last night a "simple plugin" and stripped glib direct-depenency from all example plugins
[Thu Dec 11 2008] [11:05:25] <bricks>   which should be nicer now
[Thu Dec 11 2008] [11:05:52] <dgollub>  bricks: reg. the fronetnd/consumer API example... needs this still to be done?
[Thu Dec 11 2008] [11:06:15] <bricks>   dgollub: yes didn't start yet sorry
[Thu Dec 11 2008] [11:06:29] <dgollub>  bricks: AI bricks: move API documentation to headers - this already sees nice progress.. not only from you - thanks a lot for that
[Thu Dec 11 2008] [11:06:36]     * ChrisH present ... sorry phone call
[Thu Dec 11 2008] [11:06:51] <bricks>   made today more changes
[Thu Dec 11 2008] [11:06:52] <dgollub>  bricks: no problem.. we just do a ticket.. so we don't have to shuffle this in a AI .. it's not right the highest prio. the api docu comes first
[Thu Dec 11 2008] [11:06:55] <dgollub>  Savago: you're crazy
[Thu Dec 11 2008] [11:06:59] <bricks>   which i will commit after the meeting
[Thu Dec 11 2008] [11:07:06] <Savago>   dgollub, just a little bit...
[Thu Dec 11 2008] [11:07:16] <dgollub>  what time is it? 3am?
[Thu Dec 11 2008] [11:07:37] <Savago>   5AM. But I'm planning to workout later... so no big deal.
[Thu Dec 11 2008] [11:07:39] <Savago>   :-)
[Thu Dec 11 2008] [11:07:52] <dgollub>  hehe - ok.. so we just finsihed with #1 .. previous AI
[Thu Dec 11 2008] [11:08:10] <fm__>     welcome Savago!
[Thu Dec 11 2008] [11:08:11] <dgollub>  Savago: we quickly inject your topic now... since we put infrastracture to the end of the meeting
[Thu Dec 11 2008] [11:08:38] <dgollub>  2. Topic: define the priorities of bugs/issues/features based on user stories
[Thu Dec 11 2008] [11:09:03] <dgollub>  bellmich: (we're no at #2 which is " define the priorities of bugs/issues/features based on user stories" - infra. is last)
[Thu Dec 11 2008] [11:09:36] <Savago>   fm__, hello there.
[Thu Dec 11 2008] [11:09:39] <dgollub>  Savago: i like this idea a lot... i guess this could help to make things more clear for lot of us
[Thu Dec 11 2008] [11:09:51] <bellmich> dgollub: Thanks.
[Thu Dec 11 2008] [11:10:22] <fm__>     i want applaud to everybody making it happen. Since yesterday I can sync my contact again for the first time in more than six months. thanks. this is a successfully user story ;)
[Thu Dec 11 2008] [11:11:02] <dgollub>  we should define such user-stories and "rate" them .. developers of engine/plugins can then comment on diffculties and what needs to be done .. and which milestone thi is
[Thu Dec 11 2008] [11:11:04] <Savago>   dgollub, cool. I described the basics on that previous message... but maybe this user stories could help us to define priorities.
[Thu Dec 11 2008] [11:11:40] <dgollub>  Savago: exactly ... i meant this by saying "rate" .. not quite sure if "rating" is the right term
[Thu Dec 11 2008] [11:11:54] <dgollub>  so i like the idea a lot.. how should we collect this stories?
[Thu Dec 11 2008] [11:11:58] <dgollub>  ticket system?
[Thu Dec 11 2008] [11:12:21] <bricks>   is there a voting system for trac like in bugzilla?
[Thu Dec 11 2008] [11:12:31] <fm__>     not too long ago replying to bricks mail. I suggested to get all of the developers phones working, as it is way easier to quickly debug something in our community than to wait for external reporters. are all your phones working? ;)
[Thu Dec 11 2008] [11:12:35] <dgollub>  should we introduce for this a new "ticket type" (beside defect, task, enhancment)
[Thu Dec 11 2008] [11:12:41] <dgollub>  or shouldwe file those as "enhancment"
[Thu Dec 11 2008] [11:12:49] <dgollub>  bricks: reg. voiting .. maybe there is a trac-voting-plugin
[Thu Dec 11 2008] [11:13:04] <Savago>   I'm not sure... I think that tickets should be used only for issues (being bugs, etc).
[Thu Dec 11 2008] [11:13:21] <Savago>   Maybe a wiki page could do the trick... with a set of user stories.
[Thu Dec 11 2008] [11:13:23] <dgollub>  bricks: there are vote plugins for trac...
[Thu Dec 11 2008] [11:13:27] <Savago>   What do you think?
[Thu Dec 11 2008] [11:13:39] <dgollub>  Savago: wiki is harder to maintain then ticket i guess
[Thu Dec 11 2008] [11:13:46] <fm__>     Savago: dgollub but we have to trac them. can i get an mail on every change of a wiki page?
[Thu Dec 11 2008] [11:13:48] <dgollub>  Savago: we can do quite powerful quieries
[Thu Dec 11 2008] [11:14:20] <fm__>     trac = track ;)
[Thu Dec 11 2008] [11:14:21] <dgollub>  fm__: if we would use the ticket-system we could easily trac them .. and you could also introduce new ticket queries
[Thu Dec 11 2008] [11:14:22] <Savago>   ah, ok. I'm not that much aware of what is possible to do using trac.
[Thu Dec 11 2008] [11:14:33] <dgollub>  Savago: http://opensync.org/report
[Thu Dec 11 2008] [11:14:48] <dgollub>  Savago: "repot" 10, 11 and 12 are custom queries we created
[Thu Dec 11 2008] [11:14:59] <dgollub>  it's pretty cool .. it's actually a SQL query
[Thu Dec 11 2008] [11:15:09] <fm__>     dgollub: yes that is why i would vote for the tickets. the problem is every user will want his phone to be the sucess sotry ;)
[Thu Dec 11 2008] [11:15:28] <dgollub>  fm__: sure - i don't hestiate to give this pretty low-prio. ;)
[Thu Dec 11 2008] [11:16:02] <dgollub>  ok - so anyone disaggree with "user-stories" for feature request/tracking and using the ticket-system?
[Thu Dec 11 2008] [11:16:11] <bricks>   no
[Thu Dec 11 2008] [11:16:20] <Tuju>     i'm kind of same with savago....
[Thu Dec 11 2008] [11:16:30] <Savago>   I agree (but I'm somewhat biased...).
[Thu Dec 11 2008] [11:16:52] <dgollub>  i would suggest that fm__ and Savago team-up and setup our ticket-sytem and Savago introduces some initial features as example for more users.. and maybe write an wiki-page as introduction what those feature reuqst should look like...
[Thu Dec 11 2008] [11:17:13] <dgollub>  Savago: how biased?
[Thu Dec 11 2008] [11:17:31] <dgollub>  Tuju: are you biased as well .. do i miss something? do you disagree with the ticket system?
[Thu Dec 11 2008] [11:17:32] <Savago>   Since I gave the initial suggestion...
[Thu Dec 11 2008] [11:17:34] <Savago>   ;-)
[Thu Dec 11 2008] [11:17:51] <dgollub>  Savago: oh - now i see.. sorry coffee is not yet working for me :P
[Thu Dec 11 2008] [11:17:58] <Tuju>     well, a bit felt that we're abusing the ticket system.
[Thu Dec 11 2008] [11:17:59] <Savago>   heheheheh... ok.
[Thu Dec 11 2008] [11:18:09] <dgollub>  Tuju: ticket != bug
[Thu Dec 11 2008] [11:18:13] <Tuju>     but if there is high need for this and nothing better, then go for it.
[Thu Dec 11 2008] [11:18:23] <dgollub>  Tuju: better then introducing yet another system
[Thu Dec 11 2008] [11:18:23] <Tuju>     dgollub: i know.
[Thu Dec 11 2008] [11:18:37] <dgollub>  ok let's do AIs
[Thu Dec 11 2008] [11:18:39] <bellmich> dgollub: Only a note - we should not focus too much on trac until we discussed the infrastructure in general.
[Thu Dec 11 2008] [11:18:51] <dgollub>  bellmich: ok - good point
[Thu Dec 11 2008] [11:19:43] <dgollub>  AI fm__: evaluate setup for "user-stories" depending on infra. decission made later and mail suggestion to opensync-devel@ for further comments
[Thu Dec 11 2008] [11:20:12] <dgollub>  AI Savago: draft "user-stories" guideline for "users" and for the "board" which should rate them
[Thu Dec 11 2008] [11:20:25] <dgollub>  everyone fine with this - next topic?
[Thu Dec 11 2008] [11:20:32] <_fm_>     yes
[Thu Dec 11 2008] [11:20:50] <Savago>   dgollub, cool, I'm on it.
[Thu Dec 11 2008] [11:20:58] <dgollub>  3. Doxygen API documentation
[Thu Dec 11 2008] [11:20:58] <dgollub>    * status
[Thu Dec 11 2008] [11:20:58] <dgollub>    * kudos
[Thu Dec 11 2008] [11:20:58] <dgollub>    * introduction of Doxygen API Guidelines
[Thu Dec 11 2008] [11:20:58] <dgollub>    * Doxygen API Section/Chapter for Plugins by instrumenting plugin code?
[Thu Dec 11 2008] [11:21:00] <dgollub>      && How?
[Thu Dec 11 2008] [11:21:11] <dgollub>  ok let me do the kudos thing first
[Thu Dec 11 2008] [11:21:24] <dgollub>  AWESOME!!!!!!!!!111111111111 i want to thank everyone getting their hands dirty on this
[Thu Dec 11 2008] [11:21:39] <dgollub>  especailly paul, ianmartin and bricks .. and of course  everyone i missed ;)
[Thu Dec 11 2008] [11:21:50] <dgollub>  this is a very very very very important thing
[Thu Dec 11 2008] [11:22:06] <dgollub>  reg. status ... i see very nice progress
[Thu Dec 11 2008] [11:22:27] <dgollub>  i hope this continues... anyone can give more detailed status to doxygen api?
[Thu Dec 11 2008] [11:22:42] <dgollub>  (exclude the whitespace - tab-intenstion discussion after the meeting)
[Thu Dec 11 2008] [11:23:02] <bricks>   ;)
[Thu Dec 11 2008] [11:23:17] <dgollub>  * introduction of Doxygen API Guidelines
[Thu Dec 11 2008] [11:23:32] <bricks>   the doxygen stylesheet isn't ok currenlty
[Thu Dec 11 2008] [11:23:46] <bricks>   it contains many style which are for trac i suppose
[Thu Dec 11 2008] [11:24:14] <dgollub>  yes - i didn't managed yet to follow the suggestion from this ticket how to fix the sytling issue
[Thu Dec 11 2008] [11:24:24] <bricks>   i started to clear the file but it contains to many different things
[Thu Dec 11 2008] [11:24:25] <dgollub>  btw. i hope you all now http://opensync.org/docs/
[Thu Dec 11 2008] [11:25:11] <dgollub>  so in meantime we can still usr ./docs .. but the API navigation bar is _really_ obivous... the company i talked to recently hit this button immedetially and were asking several question reg. the status of  the api
[Thu Dec 11 2008] [11:25:23] <bricks>   currently i am moving the doxygen annotation from the c files to the headers
[Thu Dec 11 2008] [11:25:40] <bricks>   and cleaning up the module structure
[Thu Dec 11 2008] [11:25:51] <dgollub>  bricks: do you need help on this? maybe we should assign some modules? so the work is shareable?
[Thu Dec 11 2008] [11:26:00] <bricks>   would be great
[Thu Dec 11 2008] [11:26:22] <dgollub>  anyone who want to move doxygen-API to headers? .. it's really a trivial thing?
[Thu Dec 11 2008] [11:26:23] <bricks>   i spend today 1 1/2 to move the format headers :(
[Thu Dec 11 2008] [11:26:34] <dgollub>  i now .. it's really i time wasting thing :(
[Thu Dec 11 2008] [11:26:38] <dgollub>  -now +know
[Thu Dec 11 2008] [11:27:19] <dgollub>  note to the potentional volunteers: it's just  about moving the doxygen description .. not documetnation itself
[Thu Dec 11 2008] [11:27:20] <ChrisH>   I have no svn accout nor any svn knowledge. But that might be a point for me.
[Thu Dec 11 2008] [11:27:30] <dgollub>  ChrisH: just send patches to opensync-devel@
[Thu Dec 11 2008] [11:27:48] <dgollub>  ChrisH: the only thing you need to know.. once you're done do: "svn diff > opensync-$module.diff"
[Thu Dec 11 2008] [11:27:49] <bricks>   btw. i try to fill this http://www.opensync.org/wiki/devel/api/0.40/review with comments which modules need more comments
[Thu Dec 11 2008] [11:28:02] <dgollub>  bricks: great!!!
[Thu Dec 11 2008] [11:28:36] <dgollub>  so what about ChrisH stats at the top of the module moving the headers... and bricks from the bottom? ;)
[Thu Dec 11 2008] [11:28:46] <dgollub>  ls -1d ./trunk/opensync/*
[Thu Dec 11 2008] [11:29:18] <bricks>   started with the top already
[Thu Dec 11 2008] [11:29:19] <bricks>   ;)
[Thu Dec 11 2008] [11:29:24] <dgollub>  hm it's mre like: find opensync/ -type d -maxdepth 1
[Thu Dec 11 2008] [11:29:29] <dgollub>  ok then lets turn it around...
[Thu Dec 11 2008] [11:29:30] <bricks>   currently at format
[Thu Dec 11 2008] [11:29:57] <ChrisH>   bricks: we should do a short phone call afterwards.
[Thu Dec 11 2008] [11:30:01] <dgollub>   find trunk/opensync/ -type d -maxdepth 1 | sort
[Thu Dec 11 2008] [11:30:09] <bricks>   ChrisH: ok
[Thu Dec 11 2008] [11:30:15] <dgollub>  ok - awesome
[Thu Dec 11 2008] [11:30:27] <dgollub>  i would suggest to commit for each module once you're ready...
[Thu Dec 11 2008] [11:30:47] <dgollub>  AI ChrisH: help rbicks moving doxygen API docu to header files
[Thu Dec 11 2008] [11:30:56] <bricks>   dgollub: right. that's the best solution
[Thu Dec 11 2008] [11:31:06] <dgollub>    * Doxygen API Section/Chapter for Plugins by instrumenting plugin code?
[Thu Dec 11 2008] [11:31:07] <dgollub>      && How?
[Thu Dec 11 2008] [11:31:26] <dgollub>  should we do "plugin writing introduction" by writing Doxygen-code?
[Thu Dec 11 2008] [11:31:39] <bricks>   dgollub: could you explain this in detail?
[Thu Dec 11 2008] [11:31:55] <dgollub>  the idea is to do "plugin writing introduction" inside the plugin-code it self
[Thu Dec 11 2008] [11:32:12] <dgollub>  so the guide gets rendered by doxygen...  and it's very easy to maintain
[Thu Dec 11 2008] [11:32:17] <bricks>   dgollub: ah ok. could be helpful
[Thu Dec 11 2008] [11:32:29] <dgollub>  since once you change the code.. you directly change the doxygen documetatnion as well
[Thu Dec 11 2008] [11:32:55] <dgollub>  not quite  sure how this is done in detail.. but Qt and KDE often do very extentinsive example sections per "class" .. thats pretty helpful
[Thu Dec 11 2008] [11:32:59] <dgollub>  i want to do the same for plugins
[Thu Dec 11 2008] [11:33:00] <bricks>   if anyone will write to guide i'll do the doxygen part
[Thu Dec 11 2008] [11:33:17] <dgollub>  bricks: i'll do the  guide
[Thu Dec 11 2008] [11:33:18] <bricks>   -to+the
[Thu Dec 11 2008] [11:33:42] <dgollub>  AI dgollub: prepare plugin guide for doxygen description and sent it over to bricks
[Thu Dec 11 2008] [11:34:00] <dgollub>  maybethis can be done with a "new group" ... but i leav this up to you
[Thu Dec 11 2008] [11:34:06] <dgollub>  ok ... anyhting else to the doxygen api?
[Thu Dec 11 2008] [11:34:11] <dgollub>  or next topic?
[Thu Dec 11 2008] [11:34:22] <dgollub>  oh wait... * introduction of Doxygen API Guidelines
[Thu Dec 11 2008] [11:34:58] <dgollub>  is there intersting of introduce something like this in CODING? and make this kind of a commit-policy for future implemetinos?
[Thu Dec 11 2008] [11:35:16] <dgollub>  (i would love to see that cdash summaries doxygen build warnings)
[Thu Dec 11 2008] [11:36:01] <dgollub>  right now the CODE DOCUMETNANION section in CODING is pretty brief... maybe we could quickly explain the doxgen basic to apply when writing opensync API?
[Thu Dec 11 2008] [11:36:09] <bricks>   dgollub: ok. then no code should be accepted which isn't documented
[Thu Dec 11 2008] [11:36:14] <dgollub>  and not quite sure if stuff like "Add FIXME, TODO and XXX comments appropriately."
[Thu Dec 11 2008] [11:36:23] <dgollub>  bricks: right .... anyone disagrees whit that?
[Thu Dec 11 2008] [11:36:43] <Savago>   Documentation definitely is part of coding.
[Thu Dec 11 2008] [11:36:47] <dgollub>  for sure
[Thu Dec 11 2008] [11:37:18] <dgollub>  reg. "Add FIXME, TODO and XXX comments appropriately." afaik doxygen is able to track also "TODO" ... but this is a special syntax in the API isn't it?
[Thu Dec 11 2008] [11:37:22] <dgollub>  \TODO or @tood?
[Thu Dec 11 2008] [11:37:50] <dgollub>  anyway .. this would be cool if we write-up some more detailed code guidelines..
[Thu Dec 11 2008] [11:37:54] <Savago>   @todo.
[Thu Dec 11 2008] [11:38:00] <bricks>   @todo should work because javadoc uses this also
[Thu Dec 11 2008] [11:38:03] <dgollub>  but less define that new implementation requires for ./trunk/ API docuemtino
[Thu Dec 11 2008] [11:38:07] <Savago>   I mean, \todo.
[Thu Dec 11 2008] [11:38:42] <dgollub>  we don't revert commits directly.. we contact the commiter via mail an CC' to opensync-devel@ .. so this is kind of a public blame...
[Thu Dec 11 2008] [11:38:55] <bricks>   :)
[Thu Dec 11 2008] [11:38:57] <dgollub>  if someone keeps committing without doxygen API .. we just disable SVN write-access
[Thu Dec 11 2008] [11:39:14] <dgollub>  to be clear: this is about new API interfaces or changes to the API
[Thu Dec 11 2008] [11:39:33] <dgollub>  bugfixes should be documentend as well.. but need no heavy doxygen inline coding - right?
[Thu Dec 11 2008] [11:39:43] <Savago>   What about plugins?
[Thu Dec 11 2008] [11:39:47] <dgollub>  (i have no idea about doxygen in-code documetatino)
[Thu Dec 11 2008] [11:40:27] <Savago>   I think bugfixes should not clutter the code.
[Thu Dec 11 2008] [11:40:27] <dgollub>  Savago: i leave plugins about to the plugin-maintainers
[Thu Dec 11 2008] [11:40:36] <Savago>   This are what svn log is for.
[Thu Dec 11 2008] [11:40:36] <Savago>   are/is.
[Thu Dec 11 2008] [11:40:36] <dgollub>  ok
[Thu Dec 11 2008] [11:40:37] <_fm_>     if it is commented in the commit there is svn blame ;)
[Thu Dec 11 2008] [11:40:38] <dgollub>  _fm_: and svn praise ;)
[Thu Dec 11 2008] [11:40:58] <dgollub>  anyhting to add to this? we can keep discussing this on opensync-devel@ .. we run out of time
[Thu Dec 11 2008] [11:41:33] <dgollub>  anyone would be intrested in drafting more detailed "CODE DOCUMETATNION" guidelines for ./CODING?
[Thu Dec 11 2008] [11:41:54] <Savago>   I can do that.
[Thu Dec 11 2008] [11:42:28] <dgollub>  AI Savago: draft new/more detailed "CODE DOCUMENTATION" chapter for ./CODING and send it to opensync-devel@ for feedback
[Thu Dec 11 2008] [11:42:32] <dgollub>  next topic...
[Thu Dec 11 2008] [11:42:43] <Savago>   (by the way, emacs has a doxymacs package, pretty useful for doxygen stuff).
[Thu Dec 11 2008] [11:42:47] <Savago>   </off-topic>
[Thu Dec 11 2008] [11:42:49] <dgollub>  4. Testsuite Status * monitoring of functional/performance regression * how to deal with random-failing tests * how to deal with uncompleted (not ported) testcases * renaming of unittests
[Thu Dec 11 2008] [11:43:09] <dgollub>  "*" seperated each item ...
[Thu Dec 11 2008] [11:43:29] <dgollub>  how should we monitor functional/perforamnce regresion of the testsuite?
[Thu Dec 11 2008] [11:43:54] <dgollub>  cdash is able to send out mail ... but you have to register for this... i got them - but i'm pretty busy and can't immedetially review all
[Thu Dec 11 2008] [11:44:06] <dgollub>  .. the problem is also i get per each commit a mail since we have lots of constant failing tests
[Thu Dec 11 2008] [11:44:21] <dgollub>  this brings me to another question .. what should we do with random-failintings tests...
[Thu Dec 11 2008] [11:44:24] <dgollub>  .. sure fix them.
[Thu Dec 11 2008] [11:44:44] <dgollub>  "talk is cheap, show me the code" ... it's not tirival to fix them just in one day
[Thu Dec 11 2008] [11:45:49] <dgollub>  so may suggestion would be to _disable_ those test which are known to fail constnetnly because they're not yet ported (sync-error) or known to fail random due to race-conditions which need to get fixed ASAP (check not-threadsafe -> ipc "IPC bad message", ...)
[Thu Dec 11 2008] [11:46:00] <dgollub>  the advantage of this is .. we only get mails of _NEW_ regressions
[Thu Dec 11 2008] [11:46:22] <dgollub>  does this sound fine to you?
[Thu Dec 11 2008] [11:46:29] <dgollub>  for sure we fail for each testcase we disable one  _blocker_ bug
[Thu Dec 11 2008] [11:46:31] <bricks>   yes
[Thu Dec 11 2008] [11:46:33] <bellmich> I do not like to disable tests because they will never ported/fixed after this point.
[Thu Dec 11 2008] [11:46:38] <dgollub>  bellmich: wrong
[Thu Dec 11 2008] [11:47:03] <dgollub>  bellmich: we file blocker bugs.. one year ago i fixed/ported the entire testsuite for 0.3x .. since there was no porting done
[Thu Dec 11 2008] [11:47:05] <bellmich> sync-error is good example and the wrong name of engine-error too.
[Thu Dec 11 2008] [11:47:06] <dgollub>  i'll do this again
[Thu Dec 11 2008] [11:47:23] <dgollub>  bellmich: renaming of test i want to discuss after the disabling of testcases
[Thu Dec 11 2008] [11:47:33] <dgollub>  .. ah btw. i'm talking about _disabling_ "testcases" not unittests
[Thu Dec 11 2008] [11:48:01] <dgollub>  btw. i have 242 unread cdash mails for "OpenSync" (not libsyncml)
[Thu Dec 11 2008] [11:48:12] <dgollub>  each of them tells me that at leat one test failed
[Thu Dec 11 2008] [11:48:33] <dgollub>  the regression notifiaction doesn't work if we have constatinly failing stuff
[Thu Dec 11 2008] [11:48:41] <bellmich> Therefore I think that mail is not always the perfect solution.
[Thu Dec 11 2008] [11:49:01] <dgollub>  and random stuff confuses people - e.g. Fritz was asking which tests are working at all on linux.. when he was doing hte windows port
[Thu Dec 11 2008] [11:49:29] <dgollub>  bellmich: whats the perfect solution then ... so far i think thats the best solution
[Thu Dec 11 2008] [11:49:41] <bellmich> Puh, some errors will only crash the suite if one error occurs.
[Thu Dec 11 2008] [11:49:51] <dgollub>  bellmich: nobody started scream last week when 4 additional tests started failing constnaly
[Thu Dec 11 2008] [11:50:04] <dgollub>  bellmich: what?
[Thu Dec 11 2008] [11:50:27] <dgollub>  bellmich: what do you mean by "suite"?
[Thu Dec 11 2008] [11:50:39] <bellmich> dgollub: fail_unless misuse ...
[Thu Dec 11 2008] [11:50:56] <Savago>   yep... its a noise. Should be turned off (and a ticket open for opensync 0.43?).
[Thu Dec 11 2008] [11:50:56] <dgollub>  bellmich: thats a different topic though
[Thu Dec 11 2008] [11:51:07] <dgollub>  Savago: nope - it's showstopper for 0.40
[Thu Dec 11 2008] [11:51:09] <bellmich> Yes, I know.
[Thu Dec 11 2008] [11:51:38] <dgollub>  bellmich: file bug for those.. but the notification doesn't work proparly unimplemented tests like sync-error
[Thu Dec 11 2008] [11:51:56] <dgollub>  the idea of the regression notifiaciotn is to quickly react on bad commits
[Thu Dec 11 2008] [11:51:57] <bellmich> Okay, I file a bug.
[Thu Dec 11 2008] [11:52:37] <dgollub>  e.g. the commit on last thrusday could be ruled out in the same error the mail got send out ... now we have to dive in the archive and have to find out which commit broke those tests.. and cause this g_parse_long_long warning
[Thu Dec 11 2008] [11:53:11] <dgollub>  bellmich: do you still disagree with disabling "testcases" which are not completed/ported or have race-conditions?
[Thu Dec 11 2008] [11:53:33] <dgollub>  (reg. race-condition.. there  is still wrong use of "check" in some tests because check is not thread-safe)
[Thu Dec 11 2008] [11:53:51] <dgollub>  which weren't fixed for a year or so
[Thu Dec 11 2008] [11:53:54] <bellmich> dgollub: I have no feeling for the source code base of OpenSync. So I think it is better to ignore me at this point.
[Thu Dec 11 2008] [11:54:20] <azeem>    I think it's somewhat the duty of whoever introduces new code to also fix up the test suites, in general; however, if there are some which have been broken at some time, it's probably better to ignore and file ticket
[Thu Dec 11 2008] [11:54:24] <bellmich> dgollub: My major focus is on libsyncml where I don't accept disabled tests if I discover them.
[Thu Dec 11 2008] [11:54:25] <dgollub>  bellmich: kk - jfyi, i hve 570 unread cdash mails for libsyncml .. not quite sure if you registered yourself for this on cdash
[Thu Dec 11 2008] [11:55:12] <dgollub>  arg.. cdash mail notifaciotn stopped working with the service-outage :(
[Thu Dec 11 2008] [11:55:25] <bellmich> No, I don't registered because some problems cannot be fixed but I still want to see that they are still present.
[Thu Dec 11 2008] [11:56:18] <dgollub>  bellmich: i agree on this.. but sync-error never got ported - if you like dive back into svn history exactly 11.5 month ago ... i spent 4 weeks porting _all_ expect sync-error and engine-multisync .. i run out of time to fix those
[Thu Dec 11 2008] [11:56:21] <bellmich> Examples are the Solaris trouble with libsoup 2.2 and the Debian stable trouble with the old check library.
[Thu Dec 11 2008] [11:56:44] <dgollub>  bellmich: i don't plan to disable such failing testcases
[Thu Dec 11 2008] [11:57:25] <dgollub>  maybe i should define more detailed which testcases to disable:
[Thu Dec 11 2008] [11:57:46] <dgollub>  - uncompleted testcases (e.g. sync-error, sync-mutlisync)
[Thu Dec 11 2008] [11:57:48] <bellmich> dgollub: Therefore I recommend to ignore me here because libsyncml is quite different here.
[Thu Dec 11 2008] [11:57:56] <dgollub>  kk
[Thu Dec 11 2008] [11:58:24] <dgollub>  - testcases with race-conditions with regard to "check" (e.g. ipc testcases)
[Thu Dec 11 2008] [11:59:00] <dgollub>  everything else are testcases which are considered to have real bugs and need to get fixed
[Thu Dec 11 2008] [11:59:26] <azeem>    can those maybe marked as "expected failures" temporarily, or is that abuse?
[Thu Dec 11 2008] [11:59:42] <azeem>    that way, testers would not be confused
[Thu Dec 11 2008] [11:59:42] Away       You are now marked as being away.
[Thu Dec 11 2008] [11:59:51] <ChrisH>   dgollub: since the update of trac I did not get any mail from cdash as before.
[Thu Dec 11 2008] [11:59:59] <dgollub>  azeem: i don't see the benefit for this... - disabling them would also not confuse the testers ;)
[Thu Dec 11 2008] [12:00:30] <dgollub>  ChrisH: cdash is indpeent of trac.. it's the platform change - i updated also cdash .. maybe something is missing will check  that later
[Thu Dec 11 2008] [12:00:45] <dgollub>  next otpic.. we'Re already overdue
[Thu Dec 11 2008] [12:00:46] <azeem>    dgollub: but then less people might see "oh, there seems to be some work needed here"
[Thu Dec 11 2008] [12:00:58] <azeem>    ok, anyway
[Thu Dec 11 2008] [12:01:15] <dgollub>  azeem: the ticket-system show that there is enough to do .. and testcases have high-prio
[Thu Dec 11 2008] [12:01:45] <dgollub>  should we do the infra. topic now?
[Thu Dec 11 2008] [12:01:46] <dgollub>  bellmich: ?
[Thu Dec 11 2008] [12:02:11] <dgollub>  the other topics i would say have less prio.
[Thu Dec 11 2008] [12:02:12]     * bellmich Present.
[Thu Dec 11 2008] [12:02:24] <dgollub>  2. Infrastructure * libwbxml/libsyncml * sf.net/freedesktop.org * Armin Bauer registered parts * scaling (trac, high load because of announcements) * robustness * work load of project admins
[Thu Dec 11 2008] [12:02:37] <dgollub>  bellmich: do you want to start on this?
[Thu Dec 11 2008] [12:02:56] <bellmich> 1. Status
[Thu Dec 11 2008] [12:02:56] <bellmich> 2. Requirements
[Thu Dec 11 2008] [12:02:56] <bellmich> 3. Future
[Thu Dec 11 2008] [12:03:13] <bellmich> So let's start with a short status ... dgollub please?
[Thu Dec 11 2008] [12:03:32] <dgollub>  the apache issue we appeared one week ago and yesterday is fixed
[Thu Dec 11 2008] [12:03:41] <dgollub>  apache was wrong configured
[Thu Dec 11 2008] [12:04:07] <dgollub>  the MaxClient was set to 150 .. which is too much when 150 proccess try to run Trac within apache via mod_python
[Thu Dec 11 2008] [12:04:18] <dgollub>  since this consums way to much memory
[Thu Dec 11 2008] [12:04:27] <dgollub>  this is not about "annoucments" on KDE bugs or something like that
[Thu Dec 11 2008] [12:04:41] <_fm_>     ;)
[Thu Dec 11 2008] [12:04:47] <dgollub>  search-engine robots don't wait if trac takes 5 minutes..  say just start simultateons requests
[Thu Dec 11 2008] [12:05:19] <dgollub>  unfortunley the robots.txt was also misconfigured .. so the robots kept indexing the entire trac-archive (source history, wiki history, timeline history, ....)
[Thu Dec 11 2008] [12:05:33] <dgollub>  the server is STABLE now
[Thu Dec 11 2008] [12:05:40] <dgollub>  we don't see _this_ issue again
[Thu Dec 11 2008] [12:06:01] <dgollub>  trac 0.11 is still the curlpit on this
[Thu Dec 11 2008] [12:06:03] <_fm_>     is there a need for some more ram? I guess we could collect some euros to buy it? Or is there not slot available at all?
[Thu Dec 11 2008] [12:06:24] <dgollub>  _fm_: no slots available.. but it's just a software issue
[Thu Dec 11 2008] [12:06:26] Away       You are now marked as being away.
[Thu Dec 11 2008] [12:06:36] <dgollub>  we surived the last 2 years with 512mb and trac 0.10.x
[Thu Dec 11 2008] [12:06:38] Away       You are no longer marked as being away.
[Thu Dec 11 2008] [12:06:55] <dgollub>  updating to trac 0.11 and updating to  1GB wasn't enough
[Thu Dec 11 2008] [12:07:00] <dgollub>  so i srhinked the apache config
[Thu Dec 11 2008] [12:07:19] <dgollub>  and we surived the robots last night - they don't reload robots.txt that  often
[Thu Dec 11 2008] [12:07:36] <dgollub>  reg. SVN revision number drift - fail diagnose.. everyhthing is ok
[Thu Dec 11 2008] [12:07:49] <dgollub>  Cdash mail notifiaciton is gone... fix TBD
[Thu Dec 11 2008] [12:07:55] <dgollub>  thats  it so far
[Thu Dec 11 2008] [12:08:01] <dgollub>  for the status
[Thu Dec 11 2008] [12:08:11] <dgollub>  (we updated doxygen on the server to fix this td width=100% issue=
[Thu Dec 11 2008] [12:08:11] <bellmich> 2. Requirements
[Thu Dec 11 2008] [12:08:23] <dgollub>  someone needs to fix trac 0.11
[Thu Dec 11 2008] [12:08:27] <bellmich> The question is what do we need?
[Thu Dec 11 2008] [12:08:33] <dgollub>  a new location
[Thu Dec 11 2008] [12:08:54] <dgollub>  Tuju: around?
[Thu Dec 11 2008] [12:08:57] <azeem>    is it possible to migrate from trac to bugzilla without losing all tickets, state and history?
[Thu Dec 11 2008] [12:09:06] <bellmich> Perhaps more general first - wiki, svn repository and ticket system.
[Thu Dec 11 2008] [12:09:06] <Tuju>     somewhat
[Thu Dec 11 2008] [12:09:35] <azeem>    I think wiki, lists and svn are less of a problem migration-wise
[Thu Dec 11 2008] [12:09:39] <dgollub>  wait wait wait.. could we first  discuss the location issue..  since any furhter descion on svn ticket sytem and  so depends on my time if we don't change the location
[Thu Dec 11 2008] [12:09:44] <bellmich> Additionally we need robustness, scalability and enough admins.
[Thu Dec 11 2008] [12:10:09] <bellmich> The question for me is, does a location change solve our problems?
[Thu Dec 11 2008] [12:10:19] <dgollub>  i spent yesterday from 9am to today/2:30am in front of the PC  .. so it would be nice if we would discuss the location problem
[Thu Dec 11 2008] [12:10:23] <dgollub>  YES
[Thu Dec 11 2008] [12:10:32] <dgollub>  someone else can do the administration
[Thu Dec 11 2008] [12:10:40] <dgollub>  and we can share the administration repsoinilitis
[Thu Dec 11 2008] [12:10:56] <dgollub>  i can't give access to many people onn the current location ..
[Thu Dec 11 2008] [12:10:59] <bellmich> The question is, must we administrate this?
[Thu Dec 11 2008] [12:11:19] <dgollub>  thats a good question .. but i would say for the next few month: yes
[Thu Dec 11 2008] [12:11:37] <dgollub>  migrating to different infra. systems (e.g. changin SCM, ticket system, ...) could kill the project
[Thu Dec 11 2008] [12:11:37] <bellmich> I know that tuju does not like sf.net but there are others like freedesktop which is hosted by universities.
[Thu Dec 11 2008] [12:12:01] <Tuju>     it's not up to me what i like or not. majority's vote counts.
[Thu Dec 11 2008] [12:12:15] <bellmich> My problem is that we have a real scaling issue.
[Thu Dec 11 2008] [12:12:27] <azeem>    fd.o and sf.net don't do trac AFAIK, so if there's no way to migrate, this is a no-starter
[Thu Dec 11 2008] [12:12:28] <dgollub>  bellmich: in which area?
[Thu Dec 11 2008] [12:12:42] <bellmich> Tuju: If I can find a solution which is suitable for all core developers then I would strongly prefer this.
[Thu Dec 11 2008] [12:12:49] <dgollub>  bellmich: human-scaling? platform-scaling? xyz-scaling?
[Thu Dec 11 2008] [12:13:14] <dgollub>  i guess it  just comes down to one thing - we need a location where we can have several server-admins
[Thu Dec 11 2008] [12:13:33] <bellmich> dgollub: If we are part of KDE and distros default installation  then we have enough to do with support and development.
[Thu Dec 11 2008] [12:13:57] <dgollub>  and have full remote-access .. i don't have full-remote access.. just ssh - if the box goes down due to swapping-to-death.. i have to kidnap someone from IT dep. to get in the jail where the server is locked up
[Thu Dec 11 2008] [12:14:14] <dgollub>  bellmich: we're already part  of default installations...
[Thu Dec 11 2008] [12:14:20] <bellmich> dgollub: We have not enough admin which are not key members of the project.
[Thu Dec 11 2008] [12:14:41] <bellmich> dgollub: ... but we are not fully usable and promoted ...
[Thu Dec 11 2008] [12:14:47] <dgollub>  agree
[Thu Dec 11 2008] [12:15:11] <dgollub>  i guess we should  give priorities to the different infra. compoonents...
[Thu Dec 11 2008] [12:15:23] <dgollub>  and try to move one by one to different locations
[Thu Dec 11 2008] [12:15:38] <dgollub>  or different solutions... so we can rely on institutions like fd.o or sf
[Thu Dec 11 2008] [12:15:52] <bellmich> Yes, this is what I started with libwbxml.
[Thu Dec 11 2008] [12:15:53] <dgollub>  the most important thing is the SVN repo
[Thu Dec 11 2008] [12:16:23] <bellmich> I saw yesterday that freedesktop supports SVN too.
[Thu Dec 11 2008] [12:16:46] <Savago>   What about googlecode?
[Thu Dec 11 2008] [12:16:46] <_fm_>     i just want to say that for me personally the trac integration of the wiki the tickets and the timeline is great.
[Thu Dec 11 2008] [12:16:48] <dgollub>  the question is if freedesktop is really working more reliable? ;)
[Thu Dec 11 2008] [12:16:50] <bellmich> Would this be more acceptable for you, Tuju?
[Thu Dec 11 2008] [12:17:11] <dgollub>  i know of more serve outages then the one we used ot have the last weeks
[Thu Dec 11 2008] [12:17:48] <dgollub>  also getting accepted as "new" projet is not easy... we're already there with OpenSync - not quite sure how  quickly the repsonse .. since they seem to lack resource as well
[Thu Dec 11 2008] [12:17:51] <Tuju>     bellmich: i've only stated that i don't like the sf interfaces. that's all. i don't know nothing about fd.o or i'm not anxious to get more responsibilities even i've offered to take them.
[Thu Dec 11 2008] [12:18:50] <bellmich> sf.net is quite stable and the users do not have to use the interfaces.
[Thu Dec 11 2008] [12:19:05] <Tuju>     so i'm happy with solution that *works*, if we cannot find other than own server, we keep our own server - wherever that is.
[Thu Dec 11 2008] [12:19:28] <_fm_>     is there any integration of the ticket system and websvn on sf.net??
[Thu Dec 11 2008] [12:19:37] <dgollub>  we already overdue 20minutes - i guess this is a really big discussion point - i highly recommend to continue this on the mailing-list
[Thu Dec 11 2008] [12:20:39] <bellmich> dgollub: I agree on this because I have no real idea how the final solution could be.
[Thu Dec 11 2008] [12:20:40] <dgollub>  anyone ok with continue this on the list? since this is really an extensive discussion
[Thu Dec 11 2008] [12:20:54] <dgollub>  ok .. lets move this on opensync-devel@
[Thu Dec 11 2008] [12:21:15] <Tuju>     bellmich: but again, choose what you like most. I'm not getting pissed off for that. but sf is not going to make me more to participate, it just sucks. (mostly their big need to get sponsor visibility and complex web interfaces)
[Thu Dec 11 2008] [12:21:19] <bellmich> bellmich: I send a topic starter in five minutes.
[Thu Dec 11 2008] [12:21:33] <dgollub>  OpenSync 0.40 API clean up (trivial tasks to give a way)
[Thu Dec 11 2008] [12:21:34] <dgollub>  * following API clean ups are proposals and look for volunteers to provide patches
[Thu Dec 11 2008] [12:21:53] <dgollub>  i listed several posiblities what kind of clean-up needs to be done
[Thu Dec 11 2008] [12:22:22] <Savago>   dgollub is there a ticket about this task?
[Thu Dec 11 2008] [12:22:24] <dgollub>  if you want to work on some of the clean-up suggestion please reply on opensync-devel@ to the "minutes" mail to avoid duplicate work
[Thu Dec 11 2008] [12:22:33] <dgollub>  no - good point... i'll do some
[Thu Dec 11 2008] [12:22:41] <dgollub>  and put them in the minutes
[Thu Dec 11 2008] [12:22:44] <Tuju>     i'm off for work. 
[Thu Dec 11 2008] [12:22:51] <dgollub>  please help on this .. some of htem are really trivial
[Thu Dec 11 2008] [12:22:54] <dgollub>  second-last topic:
[Thu Dec 11 2008] [12:22:59] <dgollub>  Bug priorities
[Thu Dec 11 2008] [12:22:59] <dgollub>      * top 5 show stoppers
[Thu Dec 11 2008] [12:22:59] <dgollub>            o symbol mix up
[Thu Dec 11 2008] [12:22:59] <dgollub>            o objtype mixing (e.g. contact <-> data doesn't work)
[Thu Dec 11 2008] [12:22:59] <dgollub>            o ... yours ...
[Thu Dec 11 2008] [12:23:16] <dgollub>  since we're overdue...
[Thu Dec 11 2008] [12:23:41] <dgollub>  ... those two and the testcases i guess are the "hottest" bugs right now
[Thu Dec 11 2008] [12:23:57] <dgollub>  if i'm wrong please correct me on the opensync-devel@ list with reply to the minutes
[Thu Dec 11 2008] [12:24:05] <dgollub>  last item:
[Thu Dec 11 2008] [12:24:10] <azeem>    Savago: symbol mixup is the google-calendar issue you commented on opensync-devel
[Thu Dec 11 2008] [12:24:15] <dgollub>  Next IRC meeting
[Thu Dec 11 2008] [12:24:53] <Savago>   azeem, yep! I'm trying to reproduce the bug here with latest trunk.
[Thu Dec 11 2008] [12:25:08] <dgollub>  Next IRC Meeting: 2008-12-17T19:00:00Z
[Thu Dec 11 2008] [12:25:19] <dgollub>  it's a WedendsdaY!
[Thu Dec 11 2008] [12:25:33] <azeem>    Savago: you might need to -DOPENSYNC_DEBUG_MODULES=ON when building opensync, not sure
[Thu Dec 11 2008] [12:25:34] <dgollub>  who wants to do the minutes next time?
[Thu Dec 11 2008] [12:25:40] <dgollub>  who is the backup?
[Thu Dec 11 2008] [12:25:56] <dgollub>  anyone?
[Thu Dec 11 2008] [12:25:59] <ChrisH>   I am on duty next meeting, I will not participate
[Thu Dec 11 2008] [12:26:08] <dgollub>  kk - anyone else? ;)
[Thu Dec 11 2008] [12:26:18] <Savago>   azeem, thanks for the tip. I will try it. :-)
[Thu Dec 11 2008] [12:26:54] <azeem>    Savago: the issue is quite clear now, but the gdata format and the syncml plugin have a symbol called "finalize"
[Thu Dec 11 2008] [12:26:57] <azeem>    namespace clash
[Thu Dec 11 2008] [12:27:03] <dgollub>  anyone?
[Thu Dec 11 2008] [12:27:04] <dgollub>  anyone?
[Thu Dec 11 2008] [12:27:04] <azeem>    s/but/both/
[Thu Dec 11 2008] [12:27:17]     * dgollub has to run to a minute in another floor in less then 1 minute)
[Thu Dec 11 2008] [12:27:27] <Savago>   azeem, if this is the bug I change google-data plugin right now.
[Thu Dec 11 2008] [12:27:41] <dgollub>  ok - i take the minutes next time........... please contact me if you want to be backup next time
[Thu Dec 11 2008] [12:27:46] <dgollub>  coudl some one restore /topic?
[Thu Dec 11 2008] [12:27:47] <dgollub>  bye
[Thu Dec 11 2008] [12:28:06] <Savago>   azeem, I thought that I was using gc_ as 'namespace' on the plugin...
[Thu Dec 11 2008] [12:28:09]     * Savago checking...
[Thu Dec 11 2008] [12:28:18] <Savago>   dgollub, see you.
[Thu Dec 11 2008] [12:28:30] <azeem>    Savago: mostly, yes, not for this one
[Thu Dec 11 2008] [12:28:40] <azeem>    maybe it was some copy&pasted code from the example plugin
[Thu Dec 11 2008] [12:28:40] <Savago>   ouch!
[Thu Dec 11 2008] [12:30:21] <Savago>   Ah... I got it.
[Thu Dec 11 2008] [12:30:48] <_fm_>     Savago: did you just solve the symbol mix up problem?
[Thu Dec 11 2008] [12:32:18] <jku_>     should I be able to use the new data sync api over http? like trying syncml-ds-tool with scheduleworld.com?
[Thu Dec 11 2008] [12:34:17] <bellmich> Who restores the IRC topic?
[Thu Dec 11 2008] [12:34:19] <_fm_>     jku_: bellmich will know most
[Thu Dec 11 2008] [12:34:39] <Tuju>     meeting is closed
[Thu Dec 11 2008] [12:35:08] Topic      You set the channel topic to "http://opensync.org/ 0.38 | http://libsyncml.opensync.org/ 0.5.0 | http://libwbxml.opensync.org/ 0.10.0 | http://www.bluez.org/openobex/ 1.4 | http://opensync.org/wiki/faq | http://pastebin.com/".