Web   ·   Wiki   ·   Activities   ·   Blog   ·   Lists   ·   Chat   ·   Meeting   ·   Bugs   ·   Git   ·   Translate   ·   Archive   ·   People   ·   Donate

#sugar-meeting meeting, 2011-08-22 16:20:40

Minutes | Index | Today     Channels | Search | Join

All times shown according to UTC.

Time Nick Message
16:20 meeting Meeting started Mon Aug 22 16:20:40 2011 UTC. The chair is silbe. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:20 Useful Commands: #action #agreed #help #info #idea #link #topic #endmeeting
16:20 silbe 18:17 < m_anish> silbe, do you want to talk abt dx3 releasing first :)
16:20 18:20 < m_anish> I think we had covered all bases, right? /me : frozen rpm, /you : resource indicator
16:20 18:20 < m_anish> can we test one last time and release?
16:21 m_anish: can you remind me what the state of NetworkManager was. Does your latest build include the fixed version?
16:22 m_anish I think it did, since I added the nm(fixed) version to the dx3 repo...and it picked that...
16:22 silbe m_anish: I won't push my patch to add ShareTerm, BTW. It's broken and at least one of the issues doesn't have an obvious solution.
16:22 m_anish i'll double check it by testing again
16:22 silbe m_anish: ok, good.
16:23 let me check the bug tracker. I think there was another major bug...
16:24 m_anish: do we have sources for the accessibility related rpms by now? Or do we consider breakage of accessibility support not to be a release blocker?
16:25 m_anish I don't consider it a release blocker
16:25 silbe ok
16:25 m_anish wasn't aware of the missing rpms needed by accessibility, though
16:27 silbe m_anish: then you didn't test thoroughly enough ;)
16:27 m_anish didn't test accesibility :/
16:28 silbe #link https://bugs.sugarlabs.org/ticket/2802
16:28 you did test it once, though
16:29 m_anish yah, 4 months ago, acc to the bug report :)
16:30 ok, so I'll test the NM/mesh issue and if it works, i'll report it to you so that we can take a decision to release?
16:30 silbe ok. let's get this over with, even though we have known major bugs.
16:31 m_anish +1
16:33 okay, moving on beyond dx-3-alpha-1...
16:33 #link http://wiki.sugarlabs.org/go/Dextrose/3/Roadmap dx3 roadmap
16:33 #link http://wiki.sugarlabs.org/go/D[…]e/3/FeatureFreeze dx3 feature freeze
16:33 #link http://wiki.sugarlabs.org/go/D[…]se/3/SugarVersion dx3 sugar version
16:33 alsroot, ^^
16:33 also /me sent emails to silbe and alsroot reg. the above
16:35 silbe, alsroot we need to also run through this some time to filter out what we could fix before sep 19(dx3-alpha-2) but i'm very exhausted today :( http://dev.laptop.org.au/issue[…]ity_id%5D%5B%5D=4
16:35 &f%5B%5D=assigned_to_id&op%5Bassigned_to_id%5D=​%3D&v%5Bassigned_to_id%5D%5B%5D=me&v%5Bassigned​_to_id%5D%5B%5D=97&v%5Bassigned_to_id%5D%5B%5D=​95&v%5Bassigned_to_id%5D%5B%5D=43&v%5Bassigned_​to_id%5D%5B%5D=91&f%5B%5D=fixed_version_id&op%5​Bfixed_version_id%5D=%3D&v%5Bfixed_version_id%5​D%5B%5D=23&f%5B%5D=&c%5B%5D=project&c%5B%5D=tra​cker&c%5B%5D=status&c%5B%5D=priority&c%5B%5D=su​bject&c%5B%5D=assigned_to&c%5B%5D=updated_on&
16:35 c%5B%5D=cf_3&group_by=
16:36 wow, that's a long link :( , not good redmime :(
16:36 silbe if we decide to revert to the 0.88 collab code (which buggy as well), the only sensible way to do that would be to base on 0.88 and cherry-pick non-collab patches. The changes are too invasive for a revert.
16:37 m_anish silbe, that's what I had thought too, and we also lose out on the work you did porting dx2 patches on top of dx3
16:38 silbe we should also remember that these changes were done for a reason. Telepathy will drop some of the code Sugar < 0.90 relies on in the future. Any work we do on the 0.88 collab code is only for a limited amount of time.
16:39 so we should carefully consider whether the differences between those versions are big enough to warrant a huge investment into a short time solution.
16:40 as you might already guess, I'd prefer to go with 0.94 and continue fixing it.
16:41 it's simply too late to switch horses
16:43 m_anish silbe, got disconn
16:43 <m_anish> silbe, +1, the blocker though is, the readyness of new collab by the time dx3 is to be relased by dec
16:43 <m_anish> readiness=regressions fixed (atleast 2 releases before, by beta-1) and tested extensively
16:44 silbe, hmm, i'm of the opinion that we could perhaps discuss this a bit more, perhaps with erikos as well, and figure this out together...
16:44 silbe m_anish: do we have a list of actual regressions? I.e. cases where 0.88 can be shown to work reliably, but 0.94 does not?
16:45 m_anish silbe, there is a list here http://wiki.sugarlabs.org/go/D[…]e/3/FeatureFreeze
16:45 i'm not sure if it covers all cases
16:45 garycmartin has quit IRC
16:45 m_anish s/cases/regressions/
16:45 silbe m_anish: feel free to discuss some more. I'm unlikely to change my mind (discussed it with a friend as well), but I wouldn't block your efforts, of course.
16:47 bernie sorry to break into the discussion. i'm just curious: does 0.88 collaboration really work better than 0.94 at this time? i remember it was quite buggy too...
16:48 i haven't tested it in 0.94 at all, though
16:48 m_anish silbe, but if we think that fixing collab is too much work seriously affecting (1) its stability, workability (2) takes away from many other bugs/features we need to work on, i'd have to think otherwise :)
16:49 bernie, hi :-) silbe considers 0.94 collab broken atm, there are many regressions listed against it :)
16:50 yama` has quit IRC
16:51 silbe m_anish: Sugar lives in a software ecosystem. It isn't easy to roll back just individual pieces. A newer version of Fedora means newer versions of Telepathy, evince, etc. All pieces must match for Sugar to work as intended.
16:51 yama <yama!~yama@124-149-41-246.dyn.iinet.net.au> has joined #sugar-meeting
16:51 yama has quit IRC
16:51 yama <yama!~yama@ubuntu/member/yama> has joined #sugar-meeting
16:52 silbe m_anish: i.e.: be careful not to underestimate the amount of work for reverting to the pre-0.90 collab code.
16:53 bernie m_anish: ok, if there are serious *regressions*, then it will be really hard to convince users to upgrade
16:53 m_anish silbe, there are three options (1) base on sugar-0.88 (2) base on 0.94 but with reverted to collab (3) fix collab
16:53 bernie silbe: yeah, my sense was that the new collaboration code was *very* invasive
16:53 m_anish (3) is the ideal soln (2) doesn't seem reasonable (1) is the fall-back
16:54 silbe, did i get my understanding of the problem right?
16:54 bernie silbe: aren't erikos and godiard working on (3)? do they have an ATE for the code will stabilize?
16:54 a/ATE/ETA/
16:54 silbe m_anish: I'm not sure that Sugar 0.88 would work on F14.
16:55 bernie ouch
16:55 m_anish bernie, i asked erikos this question via email today, ccing you
16:55 alsroot silbe: 0.88 is pretty stable, I run it in the same env as in 0.9x
16:55 silbe bernie: they were working on it for Sugar 0.92 already, but it's so broken that time wasn't enough.
16:55 bernie m_anish: yes, i've seen it.
16:56 silbe alsroot: and it does work on Fedora 14? That's the part I'm not sure about (I don't run Fedora myself except for building DX3 images).
16:56 bernie silbe: do you know if some of the bugs caused by the new telepathy stack? or is it mostly bugs in sugar and sugar-toolkit?
16:56 m_anish bernie, ah i cced it, :)
16:57 alsroot silbe: didn't test exactly on fedora, but my assomtion is that PS is much more bullet proof (keeping in mind -1 dep in comaring w/ new collab) than 0.9x
16:57 m_anish silbe, FWIW, sugar-emulator using jhbuild with 0.88.1+uy patchset runs on f14 for me, obv not very thouroughly tested
16:58 silbe another point I'd like to mention is that somebody will eventually need to fix 0.92+ collab (for the reasons pointed out above). I don't think anyone besides OLPC and AC would spend time on it, so if we work on reverting to 0.88 collab instead of fixing 0.94, we rock our own boat.
16:58 ok, that's anecdotal evidence, but nothing we should really base such a major decision on.
16:59 m_anish silbe, +1, but the bigger priority is that dx-3 has a working collab.
17:01 silbe m_anish: then we should work on it. the first step would be to identify the actual regressions and prioritise them.
17:01 m_anish silbe, FWIW, i'm willing to spend time on fixing collab, but i'm not willing to spend _all_ my time fixing collab :-). if we can determine with reasonable certainty that new collab code can be fixed with our constraints, i'm ok with that
17:02 silbe while I consider 0.94 collab broken, the 0.88 one was rather buggy for me too. we should focus on what actually works worse in 0.92.
17:02 m_anish: I'm not certain of either option (0.88 or 0.94).
17:04 m_anish silbe, bernie i'd just like to point one more thing I learnt at py (and what yama explicitly told me too). people learn to live with old bugs and make their peace with them, but when there are new bugs, it really pisses them off
17:05 silbe, okay, so we need to discuss, perhaps with erikos, gonzalo too
17:05 alsroot silbe: for me regression is entirely collab, it is about programming culture, if such critical part was done w/o any regression tests (I mean sustainable ones in auto mode, not just ruinging it in dv env and see if if works), the entirely feature can't be landed to the trunk
17:06 and trying to "fix" this regression by doing the work that needs to be done out of trunk, it pretty wrong way for me
17:07 silbe m_anish: I see quite a few bugs are about Salut / link-local collab. One option would be to declare link-local collab to be unsupported for now. It's been buggy forever anyway and we could focus on fixing Jabber based collab (Gabble) instead. Salut has it's own special problems (because it uses unreliable multicast messaging) that are not easy to fix.
17:08 alsroot silbe: also, thats not only about having shell, we will get/already-getting bunch of bug reports and bad thinking about sugar if, eg, Chat will just stop working in "stable" sugar version
17:08 silbe alsroot: I agree this piece of code should have never landed. But ripping it out _now_ doesn't really solve anything.
17:09 alsroot silbe: my major idea is not about reverting new collab, not not saying that current >0.88 sugar is stable
17:09 silbe alsroot: we already have several "stable" versions now with unreliable collab :-/
17:09 m_anish silbe, would it (declaring link-local) cause regressions over previous versions, if that is so, its unacceptable as well
17:10 silbe alsroot: FWIW, I still consider Sugar merely a technology preview, a demo. Unfortunately that doesn't change the fact that ~ 2 million kids are using this demo :-/
17:10 alsroot and we need to fix this really big issue before realeasing any stable versions. but that can't be don in a mounth or so. at least 6m
17:10 with having decent auto testing for collab
17:10 silbe m_anish: depends on what you consider a regression. As mentioned Salut has never worked reliably for me. But of course that's just anecdotal evidence again.
17:11 alsroot for really stable sugars, I can't see any sugar versions more then 0.88
17:11 ..for now
17:11 m_anish silbe, regressions=what worked (and was possibly used) in 10.1.3-au and doesn't in dx3
17:12 silbe alsroot: once we got rid of hippo-canvas and I land a few patches that are almost ready in my drawer, we can do automatic UI tests. Doing automatic collab test will require some more work (since you need a seperate unix account for each simulated user), but doable.
17:13 m_anish: then we should ask yama & co what they consider as working in 10.1.3-au.
17:13 m_anish silbe, that won't be in the 0.94 timeframe i guess?
17:13 silbe, ok
17:13 alsroot silbe: tp testing is much more doable (if testing was kept in mind while coding it, ie, not for current impl) than UI
17:13 ..for me
17:15 m_anish silbe, note: i'm not suggesting at all that we should stop contributing to fixing collab. what i'm saying is that we need to see if we can do that within dx3 timeframe, if not, we leave it to be included in later versions
17:15 silbe m_anish: 0.94 is already in feature freeze. Automated UI tests is something for 0.96+. The changes would be too invasive for DX-3 anyway. We have enough on our hands fixing collab.
17:15 m_anish silbe, +1
17:17 silbe alsroot: if you have a patch set for better collab integration, I'm all ears ;)
17:19 alsroot silbe: I can't see how a word "path" might be related to the current collad code
17:19 *patch
17:20 m_anish silbe, alsroot ok, i think we've put our point on table :), and thanks a lot for that, atleast we know our position when we discuss with erikos, gonzalo
17:20 silbe, next topic?
17:20 wanted to go through the bug list link mentioned as todo in featurefreeze page, but too tired :/
17:20 silbe alsroot: If you have code that's a) better than what's in 0.94 and b) will continue working with future versions of Telepathy (which rules out 0.88), I'd be happy to merge it.
17:22 m_anish: ok. what's the next topic?
17:22 m_anish silbe, /me doesn't have any :)
17:22 alsroot, hows your work on SSK coming along :-) do you see yourself as free in coming few days to contrib to dx3-platfm
17:24 alsroot silbe: you got me entirely wrong, what you are saying is good eg for having new toolbar design in sugar. collab is a critical part (even of sugar philosophy), if it donesn't work, it affect sugar very drastically (including me and my projects, thats why I talking about it). the current sitation is that new collab was implemented in very not reliable way *from beggining* and needs to be rehashed by
17:24 devs who will take care of this code in the future
17:25 silbe alsroot: my point is that it never worked reliably anyway (for me at any rate). That's one of the reasons I still consider it just a demo.
17:26 alsroot silbe: well, such demo is nemed as stable sugar by olpc this..
17:26 and, if I got it right, we are talking about having it in new stable dx
17:26 silbe alsroot: and, even worse, used by ~ 2 million children :-/
17:27 alsroot: I'm talking about all versions of Sugar, not just > 0.88.
17:28 alsroot m_anish: I need to compose XO-1 image w/ SSK packages and in hope to release today/tomorrow
17:29 silbe: well, if thing having more demo for 0.9x is a good way, then we need to continue to do the same (patcing demo step by step)
17:29 *if you think
17:30 ie, for hte begging landing raw code to the trunk, then patching it during the next 3y
17:32 silbe alsroot: I think we should work on making it more than just a demo in the future. I don't think reverting to code that will stop working in the future is a good step in that direction.
17:32 m_anish alsroot, +1 on  ssk
17:33 silbe and unless somebody steps up to write new collab code, the one in 0.94 is the best we have, despite the bugs.
17:33 alsroot silbe: once more, I'm not about reverting on its own. I'm for sustainable coding (and reverting might not worst way for that, eventually)
17:34 silbe alsroot: I'm all for that as well. That's why I want to rip out as much as possible and replace it with something maintained by larger communities like Gnome.
17:34 m_anish alsroot, silbe i think were diverting a little here... the discussion here is what version of collab is the best bet for dx3.
17:34 silbe (while keeping the special features of Sugar)
17:35 m_anish silbe, (1) i'm not saying that we won't work on collab, i'm saying that we need to figure out if it'll be ready by dx3, that's it, and take decisions accordingly
17:36 alsroot, (2) the point w.r.t dx3 is not whether we followed bad coding practice while we were rewriting collab, the point is whether its better than 0.88/collab.
17:37 silbe, alsroot, (3) i'm all for following good coding practice while doing stuff, but we have to work with what we have right now, and decide accordingly :)
17:37 silbe m_anish: +1
17:37 m_anish alsroot, figuring out how to code collab is a bigger and 'different' issue for me
17:38 alsroot m_anish: well, for me it is clear. we need to continue support dx2
17:39 as a stable dextrose version
17:40 0.9x and dx3 might exist but it needs to be not declared for using(and supporting) in the field
17:40 m_anish alsroot, that;s only if we work with partners which use dx2. right now there is one 'uy' which uses something resembling dx2 (though af might join)
17:41 alsroot, dx3 will be the next stable version which will be used by au when we release it
17:42 alsroot m_anish: but it is feature freeze time for dx3, right?
17:42 m_anish alsroot, i don't see dx3 as needing to be 0.94 (if we decide to go with 0.88)
17:43 alsroot, i'm willing to extend that to till end of this week
17:43 (fri, 26th aug)
17:43 alsroot ie, after that we need to only polish it to cook ready-to-use-in-the-filed. in that case I don't see any possibility to have 0.94
17:45 m_anish alsroot, something like that, but its not a very easy decision, staying with 0.88 would mean a lot of other things (losing sync with olpc, having to cherrypick a lot of patches et al) :/
17:45 alsroot m_anish: well, it is only mine, pure tech, vision
17:46 thought, I don't have any other visions
17:46 m_anish :)
17:47 alsroot, ok
17:48 alsroot, silbe so the next step is to discuss this with erikos and figure out their stand, i'd imagine :)?
17:49 silbe m_anish: certainly
17:50 m_anish silbe, ok
17:51 ok, i don't have much else to discuss, alsroot silbe dirakx ?
17:52 silbe m_anish: nothing else for today I can think of
17:53 m_anish silbe, ok (you started the meeting :P)
17:54 silbe #endmeeting
17:54 meeting Meeting ended Mon Aug 22 17:54:19 2011 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot. (v 0.1.4)
17:54 Minutes: http://meeting.sugarlabs.org/s[…]-22T16:20:40.html
17:54 Log:     http://meeting.sugarlabs.org/s[…]11-08-22T16:20:40

Minutes | Index | Today     Channels | Search | Join

Powered by ilbot/Modified.