Unterhaltung mit #deegree

(14:14:01) Das Thema für #deegree ist: Welcome to deegree, an OSGeo project. Visit the main project page at http://deegree.org and our wiki at http://wiki.deegree.org with lots of extra info. Check out a running system at http://demo.deegree.org and follow us on twitter @deegree_org.
(14:29:39) tfr42 [574f5926@gateway/web/freenode/ip.] hat den Raum betreten.
(14:30:17) markusschneider: hi everybody!
(14:30:40) markusschneider: let's start with the agenda, shall we?
(14:30:41) markusschneider: http://wiki.deegree.org/deegreeWiki/TmcMeeting/TmcMeeting20140221
(14:30:59) markusschneider: if you have anything to add, just shout it out ;-)
(14:32:00) copierrj: hi
(14:32:03) markusschneider: as you see, we have quite a list of pull requests, so i suggest to do this with some concentration and focus
(14:32:06) copierrj: agenda is fine with me
(14:32:24) tfr42: let's start
(14:32:37) markusschneider: first item: TMC chair / help with releases
(14:33:09) markusschneider: i have to admit that i didn't have enough time in the last months
(14:33:29) markusschneider: this led to postponed meetings and releases
(14:33:40) markusschneider: actually, no releases for about 2 months
(14:33:46) markusschneider: sorry for that
(14:34:05) markusschneider: i would like to ask you for ideas to improve this
(14:34:19) markusschneider: i believe, the whole process shouldn't be dependent on my person 100%
(14:34:28) copierrj: we should have a backup process for dealing with pull requests
(14:34:33) copierrj: outside of TMC meetings
(14:34:51) markusschneider: would that actually help?
(14:34:54) copierrj: at the very least, the pull requester should get a timely response
(14:35:13) markusschneider: i don't have the feeling that tmc members have been following the mailing list
(14:35:27) copierrj: it would help to give the project a more active profile
(14:35:46) tfr42: and I would try to automate as much as possible.
(14:35:49) copierrj: i was also busy the past months, but i'm reading the list
(14:36:22) copierrj: what would you like to automate in this regard?
(14:36:27) copierrj: accepting pull requests??
(14:36:57) tfr42: if there is a pull request pending and automated test does not fail. the CI creates a release
(14:37:24) tfr42: yes, perhaps even this task can be automated
(14:37:26) markusschneider: this is nice and dandy, but with the current amount of automated testing we have, this is not feasible
(14:37:28) copierrj: what if there are no test failures because the pull request eliminates all tests?
(14:37:52) copierrj: we need complementary integeration tests for this to work
(14:37:54) tfr42: I don't now how but we could check if the OSS community has something in place
(14:38:04) tfr42: This is correct
(14:38:13) tfr42: here we have to do some work - by hand!
(14:38:13) copierrj: don't have them at the moment
(14:38:50) tfr42: well, then let's start with small steps
(14:39:00) markusschneider: what do you have in mind?
(14:39:09) copierrj: and we also need additional safeguards against people breaking stuff on purpose
(14:39:26) tfr42: I propose that Reijer and myself can host a TMC meeting
(14:39:48) tfr42: but at least 2 TMC members must be present to decide on open pull requests
(14:40:07) markusschneider: +1
(14:40:10) copierrj: How about utilizing the GitHub faciliies more extensivly?
(14:40:26) copierrj: +1 (2 tmc members)
(14:40:33) tfr42: what do you mean? which services?
(14:40:37) markusschneider: i believe, these are two different aspects.
(14:40:43) copierrj: comment on pull request when you have the time
(14:40:56) markusschneider: i believe, my original proposal has been answered.
(14:41:00) copierrj: other tmc members don't have to be available at the same moment
(14:41:13) markusschneider: considering technical improvements is another matter IMHO, albeit very important
(14:41:49) markusschneider: in order to keep the timeframe, i suggest to discuss this on the next meeting. ok?
(14:42:16) copierrj: ok
(14:42:42) markusschneider: one thing: it would be important that you can also do the pre-releases without me
(14:43:19) tfr42: regarding further automation of the release process I will add issues in our ticket system
(14:43:22) copierrj: I think any GitHub project admin is technically able to do that
(14:43:36) copierrj: pre-releasing
(14:43:43) markusschneider: tfr42: good idea
(14:43:57) markusschneider: copierj: it's about the following:
(14:44:05) markusschneider: 1. pull requests (fine)
(14:44:16) markusschneider: 2. build releases on hudson
(14:44:21) markusschneider: 3. fix issues, if any
(14:44:30) markusschneider: 4. update website
(14:44:42) markusschneider: 5. update documentation with latest version
(14:44:46) markusschneider: 6. write release email
(14:45:11) copierrj: do we really need all that for pre-releases?
(14:45:37) markusschneider: well, we did in the past
(14:45:53) markusschneider: i am not sure, but i believe it helps to make the project look active quite a bit
(14:46:14) markusschneider: steps 1-3 are needed in any case
(14:46:33) copierrj: hmm, sounds somewhat too heavy for just pre-releases
(14:46:45) markusschneider: sorry, misunderstanding
(14:46:50) copierrj: 2 could be done on a dev environment
(14:46:51) markusschneider: this is not about pre-releases
(14:46:57) markusschneider: it's about point releases
(14:47:02) tfr42: OK, I will put it into a ticket and make a proposal how to automate this.
(14:47:20) markusschneider: great!
(14:47:31) tfr42: For today let's decide that a TMC meeting shall take place every 2 weeks
(14:47:44) markusschneider: as long as we don't have automation, i would like to make sure that you guys can do without me... manually
(14:47:50) copierrj: when | TMC members | >=2
(14:47:58) tfr42: We three are responsible to keep that schedule
(14:48:09) tfr42: everyone can invite to the TMC meeting
(14:48:22) markusschneider: tmc chair is not better than the others in that regard
(14:48:26) tfr42: but at least 2 TMC members must be present
(14:48:30) markusschneider: agreed
(14:48:34) copierrj: ok
(14:48:36) markusschneider: +1
(14:48:39) tfr42: good
(14:48:41) copierrj: one thing:
(14:48:57) copierrj: do we really have to a point release every meeting?
(14:49:01) copierrj: to do
(14:49:22) copierrj: wouldn't accepting / declining pulls enough ocasionally
(14:49:24) markusschneider: torsten, would you be available after the meeting to communicate the tasks i do? also good to see, how we can automate it, i guess.
(14:50:18) markusschneider: when we pull, it's not really much effort to build, do you think=
(14:50:19) markusschneider: ?
(14:50:29) tfr42: sure, a quick voice call is possible
(14:50:33) copierrj: don't know, never did that myself
(14:50:41) markusschneider: did what?
(14:50:47) tfr42: let's do it once together with markus
(14:50:48) copierrj: build
(14:51:03) copierrj: tfr42: good idea
(14:51:04) markusschneider: reijer, are you available for about 30 mins as well?
(14:51:05) tfr42: than we can see if this is an easy job or not
(14:51:17) copierrj: hmm, perhaps, have another meeting too
(14:51:37) markusschneider: ok, otherwise i'll just do it with torsten
(14:51:41) copierrj: if we are able to wrap up this meeting in a timely fashion
(14:51:44) markusschneider: thank you.
(14:51:46) markusschneider: next item
(14:51:52) markusschneider: :Checklist for pulls
(14:52:29) markusschneider: i believe we agreed that a list for a consistent checklist for pull acceptance is needed
(14:52:39) markusschneider: contributors should now what to expect
(14:53:10) markusschneider: i suggest to start such a checklist. or is there a draft already?
(14:53:39) tfr42: Not that I am aware of.
(14:53:41) markusschneider: here are some easy proposals, which we already had. i'd like your vote, please
(14:54:25) markusschneider: 1. a pull for stable must have a corresponding pull for development. or otherwise guarantee that we don't loose a bugfix/improvement
(14:54:34) copierrj: +1
(14:54:39) tfr42: +1
(14:55:17) markusschneider: 2. new features that add configuration files / options need to include documentation updates
(14:55:29) tfr42: +1
(14:55:47) copierrj: +0.5 (depends imho)
(14:56:01) markusschneider: depends on what?
(14:56:18) copierrj: kind of added option
(14:56:31) copierrj: if only relevant for devs, it's imho a different matter
(14:56:43) markusschneider: i mean: add new user-controllable options
(14:56:56) markusschneider: user=deegree administrator
(14:56:59) copierrj: stuff like custom Java classes
(14:57:04) copierrj: refs
(14:57:08) markusschneider: ok
(14:57:13) markusschneider: agreed
(14:57:36) markusschneider: 3. new pulls must not break any unit or integration tests
(14:57:44) tfr42: +1
(14:57:47) copierrj: + 10^6
(14:58:09) copierrj: btw: how do we test this?
(14:58:20) markusschneider: that's a good one
(14:58:31) copierrj: is the build server capable of doing this?
(14:58:34) copierrj: somehow?
(14:58:37) markusschneider: currently, we test it by pulling all requests and letting the build server do the job
(14:58:45) tfr42: a committer shall prove that the pull request does not break the build by setting up a ci job
(14:59:02) markusschneider: i believe, this may be a bit much for most people
(14:59:04) copierrj: not everyone is able to do that
(14:59:07) markusschneider: except us
(14:59:07) tfr42: travis-ci.org is great for this
(14:59:12) tfr42: it is simple
(14:59:19) markusschneider: how does it work?
(14:59:31) tfr42: next time
(14:59:32) copierrj: perhaps there are some git/github plugins for hudson to do something like this automatically
(14:59:43) tfr42: that is travis-ci
(14:59:44) markusschneider: let's postpone that. ok?
(14:59:47) copierrj: oik
(14:59:48) copierrj: ok
(14:59:50) tfr42: yes please
(15:00:01) tfr42: but your vote for my proposal?
(15:00:06) copierrj: +1
(15:00:10) markusschneider: which proposal?
(15:00:33) markusschneider: "a committer shall prove that the pull request does not break the build by setting up a ci job"?
(15:00:36) copierrj: but we should be lenient to outsiders
(15:00:38) tfr42: correct
(15:00:59) markusschneider: we could also merge new pulls automatically and run tests on that branch
(15:01:02) tfr42: this is not binding but recommended
(15:01:25) markusschneider: i'd like to discuss the details before giving my vote
(15:01:31) markusschneider: can we postpone that?
(15:01:44) tfr42: sure
(15:01:48) copierrj: another idea:
(15:02:00) copierrj: a staging branch that is build by ci
(15:02:08) markusschneider: that's what i meant ;-)
(15:02:12) copierrj: ok
(15:02:13) markusschneider: "we could also merge new pulls automatically and run tests on that branch"
(15:02:25) markusschneider: maybe this even the better option
(15:02:27) copierrj: it's what they do in the Linux project
(15:02:42) markusschneider: anyways: next item:
(15:02:44) markusschneider: deegree3 stable pulls (3.3.9 + 3.4-pre10)
(15:02:52) markusschneider: keep the list in mind already, please
(15:03:01) markusschneider: https://github.com/deegree/deegree3/pull/189
(15:03:10) copierrj: +1
(15:03:11) tfr42: next?
(15:03:11) tfr42: yes, go ahead please
(15:03:42) markusschneider: reijer, is the 3.4 functionality guaranteed
(15:04:01) copierrj: oops, did check that
(15:04:13) copierrj: 0
(15:04:21) tfr42: +0 since not reviewed
(15:04:23) markusschneider: from my point of view, this is -1
(15:04:27) copierrj: postpone until there is a 3.4
(15:04:30) copierrj: pull
(15:04:33) markusschneider: ok
(15:04:51) markusschneider: https://github.com/deegree/deegree3/pull/199 and https://github.com/deegree/deegree3/pull/198
(15:05:02) copierrj: we need a standard text to append to such an pull request
(15:05:04) markusschneider: there's quite some comments already in https://github.com/deegree/deegree3/pull/198
(15:05:08) markusschneider: agreed
(15:05:21) copierrj: -1
(15:05:34) markusschneider: i can prepare a checklist page
(15:05:54) markusschneider: -1
(15:06:03) tfr42: -1
(15:06:04) Dirk_ [574f5926@gateway/web/freenode/ip.] hat den Raum betreten.
(15:06:40) tfr42: postpone to next meeting and we should give feedback
(15:06:42) markusschneider: https://github.com/deegree/deegree3/pull/205 and https://github.com/deegree/deegree3/pull/207
(15:07:13) copierrj: +1
(15:07:14) tfr42: +1
(15:07:22) markusschneider: cannot tell if it works: 0
(15:07:43) tfr42: we got feedback from users that this issue is solved
(15:07:46) tfr42: two instances
(15:07:51) tfr42: of deegree
(15:08:04) markusschneider: oops
(15:08:15) markusschneider: the 3.4 pull cannot be automatically merged
(15:08:50) markusschneider: what shall we do?
(15:09:05) markusschneider: btw, that's another requirement
(15:09:42) markusschneider: postpone it?
(15:09:55) markusschneider: and ask for an updated patch?
(15:09:56) tfr42: yes, and first review the changes and then do the merge manually.
(15:10:14) copierrj: we could ask them this fix it imho
(15:10:24) tfr42: ask diego for help
(15:10:37) markusschneider: agreed
(15:11:02) markusschneider: Next
(15:11:12) markusschneider: https://github.com/deegree/deegree3/pull/209 and https://github.com/deegree/deegree3/pull/211
(15:11:31) copierrj: can't merge
(15:11:44) copierrj: 3.3
(15:11:54) markusschneider: hmm
(15:12:11) copierrj: not rebased / cherry picked properly
(15:12:13) markusschneider: 3.4 can be merged. plus it includes the other merge
(15:12:26) markusschneider: shall we merge 3.4 then?
(15:12:39) copierrj: hmm
(15:12:41) markusschneider: i suggest to ask for a fix
(15:12:50) copierrj: i don't like that the pull requests are broken
(15:13:10) markusschneider: torsten is already commenting. let's postpone it
(15:13:16) copierrj: ok
(15:13:25) markusschneider: Next?
(15:13:43) tfr42: let's schedule a merge session with Diego
(15:13:50) tfr42: yes
(15:13:53) markusschneider: https://github.com/deegree/deegree3/pull/213
(15:14:12) markusschneider: -1 (no 3.4 pull)
(15:14:37) copierrj: is this behavior part of the standard?
(15:15:08) copierrj: -1 (3.4 pull missing)
(15:15:10) markusschneider: don't know. let's ask for clarification?
(15:15:21) copierrj: like the improvement though
(15:15:56) markusschneider: torsten? what do you think?
(15:16:17) tfr42: wrote a comment
(15:16:27) tfr42: +-0
(15:16:31) tfr42: next?
(15:16:44) markusschneider: https://github.com/deegree/deegree3/pull/215
(15:16:56) ndrs: l.
(15:17:20) tfr42: +1
(15:17:29) markusschneider: what about 3.4?
(15:18:05) markusschneider: sorry to say that again...
(15:18:07) markusschneider: :-(
(15:18:37) markusschneider: -1
(15:18:54) tfr42: wrote a comment
(15:18:59) markusschneider: thx torsten
(15:19:05) markusschneider: next
(15:19:08) markusschneider: ?
(15:19:20) markusschneider: https://github.com/deegree/deegree3/pull/220 and https://github.com/deegree/deegree3/pull/221
(15:19:52) copierrj: 221??
(15:19:58) markusschneider: sorry
(15:20:00) markusschneider: https://github.com/deegree/deegree3/pull/219
(15:20:03) copierrj: +1
(15:20:11) tfr42: stop
(15:20:20) tfr42: #220 or #219 ?
(15:20:29) markusschneider: #220 and #219 (3.3 + 3.4)
(15:20:50) markusschneider: sorry for the confusion
(15:21:20) markusschneider: +1
(15:21:34) tfr42: does this patch changes the config?
(15:21:38) copierrj: yes
(15:21:43) tfr42: no unit tests?
(15:21:49) tfr42: no documentation?
(15:21:54) tfr42: no ticket in trac?
(15:21:55) markusschneider: well, it changes the (not really documented) support for SLD
(15:21:55) copierrj: impossible (Java2D stuff)
(15:22:08) tfr42: 0
(15:22:19) markusschneider: is this part of the SLD standard?
(15:22:23) copierrj: no
(15:22:37) copierrj: just something I implemented a while ago to make a customer happy
(15:22:54) markusschneider: then, it actually needs to be documented, i am afraid
(15:23:02) markusschneider: otherwise, no one will know it's there...
(15:23:05) copierrj: ok
(15:23:10) markusschneider: -1 (sry)
(15:23:20) copierrj: i'll push an additional commit to those branches
(15:23:24) markusschneider: thx
(15:23:29) tfr42: please add a ticket in trac
(15:23:29) copierrj: keep pull open please
(15:23:33) markusschneider: sure
(15:23:41) markusschneider: ok. now 3.4-only pulls
(15:23:50) markusschneider: https://github.com/deegree/deegree3/pull/204
(15:24:00) copierrj: +1
(15:24:12) markusschneider: no documentation: -1
(15:24:36) markusschneider: but definitely a good thing to have.
(15:24:40) markusschneider: i want it :-)
(15:24:58) copierrj: do we also (strictly) require that for stuff in unstable?
(15:25:21) tfr42: sorry, but no ticket in trac, no documentation
(15:25:21) markusschneider: i am all for it. otherwise, it's going to be lost
(15:25:21) tfr42: -1
(15:25:35) markusschneider: i will add some documentation
(15:25:40) copierrj: thx
(15:25:47) tfr42: great
(15:25:49) markusschneider: postponed
(15:25:58) markusschneider: another thing:
(15:26:20) markusschneider: torsten proposed to have "bigger feature additions" to have a corresponding announcement on the ml
(15:26:31) markusschneider: i second that
(15:26:44) markusschneider: it's very good for visibility
(15:26:52) markusschneider: i will also send an email
(15:27:03) tfr42: thanks
(15:27:05) copierrj: agreed, would be nice
(15:27:09) copierrj: but, do we require that?
(15:27:37) markusschneider: well, i'd say it's most welcome and should definitely be the best practice
(15:27:47) copierrj: definitely
(15:27:51) tfr42: either a ticket in trac or an announcement to ml
(15:28:04) markusschneider: +1
(15:28:11) copierrj: +1 (ml)
(15:28:19) tfr42: personally I do document everything in trac
(15:28:51) markusschneider: ok. next?
(15:28:54) markusschneider: https://github.com/deegree/deegree3/pull/208
(15:29:09) tfr42: last one for today, please
(15:29:21) markusschneider: really?
(15:29:35) markusschneider: i'd definitely appreciate if we could get in the last two pulls
(15:29:39) markusschneider: or at least discuss them
(15:29:55) markusschneider: I have a problem with 208
(15:30:27) copierrj: me2
(15:30:31) tfr42: ok, then let's postpone
(15:30:34) markusschneider: fine
(15:31:00) markusschneider: https://github.com/deegree/deegree3/pull/225
(15:31:11) tfr42: guys, since I have to leave in time - right now - can you continue without me?
(15:31:15) markusschneider: yes
(15:31:22) copierrj: yes (there are still 2 of us...)
(15:31:24) markusschneider: thanks
(15:31:38) markusschneider: +1
(15:31:42) copierrj: +1
(15:31:53) markusschneider: merged
(15:32:00) tfr42: +1
(15:32:06) tfr42: talk to you soon
(15:32:11) copierrj: cu
(15:32:12) tfr42: bye bye
(15:32:13) markusschneider: cu
(15:32:15) markusschneider: https://github.com/deegree/deegree3/pull/212
(15:32:34) copierrj: +1
(15:32:36) markusschneider: +1
(15:33:01) markusschneider: merged
(15:33:09) markusschneider: https://github.com/deegree/deegree3/pull/214
(15:33:10) copierrj: +1
(15:33:13) markusschneider: +1
(15:33:30) markusschneider: https://github.com/deegree/deegree3/pull/216
(15:33:42) copierrj: +1
(15:33:51) markusschneider: Does this change the configuration?
(15:34:00) copierrj: adds to minor things:
(15:34:19) copierrj: 1) custom serializer
(15:34:32) copierrj: 2) enableDefaultFormats attribute
(15:34:37) markusschneider: hm
(15:34:40) markusschneider: +0.5 then
(15:34:58) markusschneider: merged
(15:35:06) markusschneider: https://github.com/deegree/deegree3/pull/217
(15:35:36) copierrj: nice,
(15:35:56) markusschneider: I believe the diagram needs some explanation...
(15:35:57) copierrj: could an image like that be generated automatically somehow?
(15:36:05) markusschneider: some text
(15:36:12) markusschneider: to describe how to read it
(15:36:20) copierrj: maintainability
(15:36:28) tfr42 hat den Raum verlassen (quit: Ping timeout: 245 seconds).
(15:36:30) copierrj: need source
(15:36:35) copierrj: !png
(15:36:41) markusschneider: yes, the source should be included
(15:36:43) markusschneider: agreed
(15:37:00) copierrj: add recommendation to pull?
(15:37:07) markusschneider: let's ask for that and some "how to read this" text
(15:37:10) markusschneider: yes
(15:37:20) markusschneider: for now: -0.5
(15:37:21) copierrj: i'll do that
(15:37:28) markusschneider: thx
(15:37:56) markusschneider: https://github.com/deegree/deegree3/pull/218
(15:38:34) copierrj: comment added to 217
(15:38:44) copierrj: +1
(15:38:47) markusschneider: thx
(15:38:48) markusschneider: +1
(15:38:57) markusschneider: merged
(15:39:30) markusschneider: https://github.com/deegree/deegree3/pull/221
(15:39:42) markusschneider: +1
(15:39:54) copierrj: +1
(15:40:20) copierrj: i would like to scan deegree for every 'null' occurrence someday....
(15:40:29) markusschneider: :-D
(15:40:31) markusschneider: merged
(15:40:53) markusschneider: https://github.com/deegree/deegree3/pull/223
(15:41:01) copierrj: needs testing
(15:41:02) markusschneider: -1 (verify first)
(15:41:12) markusschneider: https://github.com/deegree/deegree3/pull/224
(15:41:25) copierrj: could you clarify your vote
(15:41:25) markusschneider: that's a mean one ;-
(15:41:33) markusschneider: on 223?
(15:41:36) copierrj: yes
(15:41:38) markusschneider: -1
(15:41:41) markusschneider: needs testing
(15:41:46) copierrj: agreed
(15:41:58) markusschneider: https://github.com/deegree/deegree3/pull/224
(15:42:03) copierrj: +1
(15:42:18) markusschneider: -1 (no documentation)
(15:42:24) copierrj: hmm
(15:42:28) copierrj: that sucks
(15:42:33) markusschneider: this is a tie
(15:42:34) copierrj: really like this one
(15:42:41) markusschneider: sure, me too
(15:43:02) markusschneider: but we probably need to keep to our own standards
(15:43:08) copierrj: agreed
(15:43:11) markusschneider: but if you keep +1, i will merge
(15:43:15) markusschneider: as it's 0
(15:43:33) Dirk_ hat den Raum verlassen (quit: Ping timeout: 245 seconds).
(15:43:51) copierrj: hmm
(15:44:12) markusschneider: or just add a comment that i have to provide a handbook section
(15:44:27) copierrj: we could add your examples to the handbook
(15:44:31) markusschneider: right
(15:44:37) markusschneider: but they do need to be added
(15:45:50) copierrj: merge as soon as this is added to the handbook?
(15:45:55) markusschneider: ok
(15:46:01) copierrj: without additional TMC vote?
(15:46:23) markusschneider: hmm, i believe this would violate our standard procedure.
(15:46:41) markusschneider: and it's not a good thing to make exceptions when you want people to feel invited
(15:46:50) copierrj: as does merging new functionality without docs...
(15:47:12) markusschneider: from my side, it's definitely -1 (sorry)
(15:47:20) copierrj: well, i can wait two more weeks
(15:47:30) markusschneider: fine, then
(15:47:47) markusschneider: that's it. thanks for your attendance
(15:47:57) copierrj: you're welcome,
(15:48:00) markusschneider: let's schedule the next meeting
(15:48:02) copierrj: ok
(15:48:19) markusschneider: tuesday, 4th of March
(15:48:35) markusschneider: 14:30
(15:48:40) copierrj: ok,
(15:48:50) markusschneider: thx
(15:48:52) markusschneider: c u
(15:48:53) copierrj: not entirely sure yet, but we can always postpone later
(15:48:57) copierrj: ok, cu
(15:48:59) markusschneider: right