PSC Meeting 7 Mar 2014

Proposed meeting time:

Friday, Mar 07, 1500 UTC

previous meeting

next meeting

Members Present:

Paolo Cavallini
Anita Graser
Otto Dassau
Tim Sutton
Marco Hugentobler
Jürgen Fischer
Richard Duivenvoorde
Sandro Santilli lurking



15:51 [Users #qgis_meeting_140307]
15:51 [@jef] [ anitagraser] [ dassau] [ duiv] [ timlinux] 
15:51 -!- Irssi: #qgis_meeting_140307: Total of 5 nicks [1 ops, 0 halfops, 0 voices, 4 normal]
15:51 -!- Channel #qgis_meeting_140307 created Fri Mar  7 15:22:48 2014
15:51 -!- Irssi: Join to #qgis_meeting_140307 was synced in 1 secs
15:51 < timlinux> hi all
15:51 < duiv> hi timlinux 
15:51 < dassau> hi tim
15:51 < anitagraser> Hi
15:52 < duiv> hi all
15:52 < timlinux> hi duiv
15:52 < timlinux> and everyone :-)
15:53 < timlinux> I guess we are waiting a few minutes
15:53 < duiv> agreed
15:54 < timlinux> I also dont have too much time
15:54 < duiv> maybe create an agenda?
15:54 < duiv> and if there is nothing to agree, or talk about, we are finished :-)
15:55 < timlinux> yeah it sounds like the only item on the agenda is bugfix releases 
15:55 < timlinux> which I am pretty sure no one will agree about :-)
15:55 < duiv> ouch, that can be a loooonnnggg one
15:56 < duiv> timlinux: any experience with paver?
15:56 < timlinux> only by seeing that victor uses it
15:56 < duiv> yeah, and tomkralidis for metasearch
15:56 -!- strk [~strk@unaffiliated/strk] has joined #qgis_meeting_140307
15:56 < timlinux> I dont have a use case for it yet - or I dont understand it well enough to know that I should have a use case :-P
15:57 < duiv> so I thought to start using it for the website/docs building
15:57 < duiv> hoping that more people would chime in to build it, and hopefully help out with writing
15:57 -!- pcav [] has joined #qgis_meeting_140307
15:58 < pcav> hi all
15:58 < duiv> hi
15:59 < timlinux> duiv as I understand it it is a make like replacement
16:00 < timlinux> duiv which I generally use fabric for already
16:00 < timlinux> Marco said he will come too right?
16:00 < duiv> exactly, but I hope then windows/mac user will help
16:00 < timlinux> yeah
16:00 < timlinux> thanks for makeing hte manual pdf btw
16:01 < timlinux> I need to get an artist type to make it a bit less vanilla on the front page
16:01 < timlinux> and put qgis sponsors etc into the front
16:01 < timlinux> there is still a bunch of linfiniti blah blah in there which we can mostly remove
16:01 -!- mhugent [] has joined #qgis_meeting_140307
16:02 < timlinux> hi mhugent 
16:02 < mhugent> hi all
16:02 < timlinux> who are we waiting for ... paolo?
16:02 < strk> can I attend ? 
16:02 < timlinux> strk: sure anyone can lurk
16:02 < duiv> pcav: was here already
16:03 < timlinux> ah sorry I am blind :-)
16:03 [Users #qgis_meeting_140307]
16:03 [@jef        ] [ dassau] [ mhugent] [ strk    ] 
16:03 [ anitagraser] [ duiv  ] [ pcav   ] [ timlinux] 
16:03 -!- Irssi: #qgis_meeting_140307: Total of 8 nicks [1 ops, 0 halfops, 0 voices, 7 normal]
16:04 < timlinux> I think we lost duiv?
16:04 < pcav> duiv: ?
16:05 < duiv> nope
16:05 < duiv> I'm still here
16:05 < duiv> anybody chairing?
16:05 < duiv> or how do you say this
16:06 < pcav> I suggest timlinux as chair
16:06 < timlinux> OK lets go
16:06 < timlinux> any more items for the agenda?
16:06 -!- anitagraser1 [] has joined #qgis_meeting_140307
16:07 < duiv> not on
16:07 < timlinux> Ok so it really only bugfix releases to discuss
16:07 < duiv> maybe if we should arrange somthing for Vienna?
16:08 -!- anitagraser [] has left #qgis_meeting_140307 ["Leaving"]
16:08 < timlinux> ok
16:08 < timlinux> we seem to be having a lot of channel churn here
16:08 < timlinux> can we start with the bugfix releases
16:08 < timlinux> I think it has largely been discussed on the mailing list
16:09 < timlinux> and honestly I don't know if we can make an instant decision here
16:09 < timlinux> but
16:09 < duiv> yeah, I have the feeling that a lot of users want it, but nobody (can) do it
16:09 < timlinux> it would be good to try to establish a roadmap for what is needed
16:10 < timlinux> my feeling is that we should be looking for someone to sponsor e.g. jef to spend a day a month applying all fixes that seem reasonable and producing 3 bugfixe releases per release cycle
16:10 < timlinux> 1 per end of month
16:10 < timlinux> (assuming he is willing)
16:11 < timlinux> at the same time I kick myself in the arse to get the test suite into a good state
16:11 < pcav> I think we can find the money
16:11 < timlinux> so that we can use that to validate the releases (such as the test coverage will do for us)
16:12 < timlinux> pcav: I have a feeling the money will come too
16:12 < timlinux> I think there is a lot of demand for bugfix releases
16:12 < timlinux> even if purely on a pschological basis
16:12 < duiv> I think it will make stuff more complex (issues going over several branches etc)
16:13 < timlinux> duiv yes I think there is no dispute
16:13 < duiv> and besides this: that is where Boundless and Sourcepole try to make money with, isn't it?
16:13 < timlinux> yeah but those are not public products
16:13 < timlinux> well marco could speak to that better than I
16:14 < timlinux> but personally I prefer we have official QGIS bugfix builds
16:14 < duiv> I know, but users can choose for theirselves
16:14 < timlinux> duiv there is already a model to deal with the complexity
16:14 < timlinux> its gitflow
16:14 < mhugent> if you speak about QGIS Enterprise, we try to make money with the support contracts
16:15  * strk rounds eyes (bad experience with gitflow)
16:15 < timlinux> mhugent: yeah which is not the intention with our bugfix releases
16:15 < timlinux> we should leave the support to commercial providers like sourcepole, boundless etc
16:15 < mhugent> I think those are really two different things
16:16 < timlinux> strk: we started using it on some projects - so far no big issues
16:16 < timlinux> mhugent: agreed
16:16 < duiv> for me support could also be providing bugfixes for latest stable
16:16 < timlinux> strk: but we dont need to use that if there is something better
16:16 < timlinux> duiv but in that model someone may pay e.g. sourcepole who pushes it to master
16:17 < timlinux> and e.g. backports it to current release
16:17 < strk> +1 -- the "stable branch" would be an asset for anyone offering commercial support (a place to push deliverables to)
16:18 < duiv> but then not put the burden of it on them too?
16:19 < timlinux> <-- duiv there is a ticket I wrote for inasafe explaining how gitflow can help us
16:19 < strk> I don't think anyone should be _forced_ to do anything, in general. This is _free_ software after all :)
16:19 < timlinux> in many respects we use it anyway - PR's for peer review of patches
16:21 < timlinux> could we just take a straw poll, who is in favour of putting out bugfix releases say once a month?
16:21 < timlinux> +1 from me (logistics of who will do it and what the workflow will be aside)
16:22 < mhugent> 0 for me
16:22 < timlinux> mhugent: I think everyone else has fallen asleep with this boring topic :-P
16:22 < strk> the problem I had with gitflow model was that it basically changes the meaning of a well-estabilished name for a "trunk" branch (from "master" to "develop")
16:23 < mhugent> he, he
16:23 < strk> + it turns "master" into a _single_ stable branch, making it unheasy to have (say) multiple stable branches (one for each minor release, for example)
16:23 < duiv> -1 as I think average joe will not be able to handle creating issues for two branches
16:24 < duiv> so we will have a lot of issues in which we have to ask: "which version??" 
16:24 < timlinux> strk in our current model we have one branch per stable release but they are basically unmaintained so it buys us nothing
16:24 < strk> I wouldn't say so
16:25 < timlinux> duiv I still have to ask 'which version' most of the time anyway so 
16:25 < timlinux> I dont think that will change much
16:25 < strk> I think there might be some patches in 1.7 which never got into a release, and I know for sure there are some in 2.0 which hadn't got into a release
16:25 < strk> making the current obstacle a "packaging" one
16:26 < strk> that's of course my personal view on the matter, that is... I'm not asking anyone to _fix_bugs_ or _backport_ them
16:26 < duiv> pcav? dassau? anitagraser1? jef?
16:26  * anitagraser1 looking through aas sponsorship agreement
16:27 < strk> only to, eventually, consider packaging a new dot release in the (unlikely?) case anyone pushes fixes into it (support providers ? others caring about it ? myself ?)
16:27 < timlinux> anitagraser1: yeah we have an agreement to backport there iirc
16:27 < pcav> timlinux: confirmed
16:27 < anitagraser1> well, if we agreed ... then shouldn't we do it?
16:28 < timlinux> strk: well if no bugfixes come into the master / release branch we make no point release
16:28 < timlinux> anitagraser1: yes but who, how and with what frequency?
16:30 < strk> timlinux: right, that's my proposal: only cut a point release after 1 week of NO activity _after_ any activity in the stable branch (for any stable branch)
16:30 < timlinux> sounds reasonable
16:30 < timlinux> assuming someone can make a little script to send us a note on twitter that its time for a release or whatever
16:30 < strk> that'd mean _no_ fixed frequency. only burden on packaging and peer-reviewing for accepting or not actiity in the stable branches
16:30 < anitagraser1> timlinux: my two cents: one month after release and maybe 2,5 months after release; done by Jürgen or whoever agrees and is capable of doing it; paid for by sponsors
16:31 < timlinux> anitagraser1: y
16:31 < anitagraser1> timlinux: if there is no sponsor money left, we have to communicate the issue 
16:31 < timlinux> I dont have string feelings about the frequence but rest I agree with
16:31 < anitagraser1> (me neither)
16:32 < strk> the sponsored backport would basically trigger "stable branch activity", and one week after it's over.... cut!
16:32 < timlinux> yup
16:33 < strk> if anything takes more than 1 week to backport... it'll still be a good idea to cut one release and the rest will come out later (least changes per release, more stability...)
16:33 < timlinux> paolo if we committed ourselves to paying for someone to sit one day a month to apply all bugfixes to the release branch do we have enough funds to cope with that?
16:33 < timlinux> I am sure we would have a successful funding drive if we let it be known we plan some formal stabilisaiton work on the release branch
16:33 < timlinux> and we could limit work to current release only
16:34 < pcav> timlinux: yes
16:34 < timlinux> I dont think there is benefit in maintaining e.g. 6 releases into the past
16:34 < pcav> we do have the money
16:34 < pcav> and yes
16:34 < anitagraser1> +1 for not bothering with old releases
16:34 < pcav> I'm pretty sure we can raise more with a targeted campaign
16:34 < pcav> anitagraser: +1
16:35 < pcav> just supporting the current one
16:35 < timlinux> we really need jef in the discussion
16:35 < strk> +1, I'm sure people using 1.7.3 will be able to surive and finally decide to upgrade when 2.0.4 is out :P
16:35 < anitagraser1> timlinux: should we postpone and continue next week?
16:36 < timlinux> my understanding from him is that he is not fundamentally opposed to bug fix releases but has concerns about complexity etc
16:36 < timlinux> and no good test platform to verify releases on
16:36 < timlinux> anitagraser1: yes
16:36 < timlinux> I have to cut & run
16:36 < timlinux> anyone want to add a last word? We can mark this item as unconcluded for now
16:36 < pcav> ok
16:37 < timlinux> duiv you wanted to say something about the hackfet?
16:37 < duiv> is everybody in Vienna? We can discuss there too
16:37 < anitagraser1> I should write stephan about our decission for the evening
16:37 < anitagraser1> tuesday + thursday until 22:00 
16:37 < anitagraser1> ?
16:37 < timlinux> yeah sounds good to me
16:37 < duiv> for me too
16:39 < anitagraser1> duiv: i think everybody will be there
16:39 < timlinux> We will hold a PSC meeting in Vienna I presume
16:39 < timlinux> we can thrash out bugfix release discussions there
16:39 < duiv> by the way: I have a free bed in my hotelroom
16:40 < timlinux> and let jef get his say in
16:40 < timlinux> duiv maybe advertise it on the list? 
16:40 < duiv> yep will do
16:40 < timlinux> Ok I will call this meeting to an end if there is nothing further?
16:40 < pcav> ok
16:41 < timlinux> will see you in Vienna!
16:41 < anitagraser1> pcav: we'll need a billing address for the evening security
16:41 < anitagraser1> timlinux: see you!
16:41 < duiv> bye
16:41 < timlinux> bye
16:41 < pcav> anitagraser: ask Horst
16:41 < pcav> he is handling the money
16:41 -!- strk [~strk@unaffiliated/strk] has left #qgis_meeting_140307 []
16:41 < anitagraser1> pcav: ok thanks
16:42 < pcav> it is the Swiss association however
16:42 < pcav> anitagraser: thank you
16:42 < duiv> anitagraser1: there are some details here:
16:43 < duiv> (near Bank Transfer) if that is what you need
16:44  * jef is catching up
16:47 [Users #qgis_meeting_140307]
16:47 [@jef         ] [ dassau] [ mhugent] [ timlinux] 
16:47 [ anitagraser1] [ duiv  ] [ pcav   ] 
16:47 -!- Irssi: #qgis_meeting_140307: Total of 7 nicks [1 ops, 0 halfops, 0 voices, 6 normal]
16:49 <@jef> anyone still here?
16:50  * duiv waves
16:50 < duiv> by the way: we are on the same planes :-)
16:50 <@jef> nice :)
17:06 -!- dassau [] has quit [Quit: Verlassend]
17:13 < anitagraser1> hi jef. we postponed the discussions concerning backports and bugfix releases to next week
17:13 < anitagraser1> jef: will you be able to join next friday?
17:18 <@jef> anitagraser1: noted