[Wednesday 24 October 2012] [19:02:15] so: meeting [Wednesday 24 October 2012] [19:02:25] aseigo: i might participate ^ PA4 [Wednesday 24 October 2012] [19:02:40] huzzzah :) [Wednesday 24 October 2012] [19:02:45] siddvicious: open to everybody ;) [Wednesday 24 October 2012] [19:03:06] right now since morning i am trying for features what i want [Wednesday 24 October 2012] [19:03:11] i might be selfish :D [Wednesday 24 October 2012] [19:03:56] notmart: any page for TODO ? [Wednesday 24 October 2012] [19:03:57] siddvicious: well, a part of the deal tough is helping implementing them.. just that you know :p [Wednesday 24 October 2012] [19:04:06] not yet [Wednesday 24 October 2012] [19:04:30] Join colomar has joined this channel (~colomar@frnk-590d9fcd.pool.mediaWays.net). [Wednesday 24 October 2012] [19:05:52] hi [Wednesday 24 October 2012] [19:06:14] well, i dunno if this is the right time to ask but it would be great if the kernel config could be changed a little: compile usb_serial as module and add generic usb serial support [Wednesday 24 October 2012] [19:06:24] Has the meeting already started? [Wednesday 24 October 2012] [19:06:40] ok, so, somebody wants to start? well i think i should by defining a bit better the scope [Wednesday 24 October 2012] [19:06:46] colomar: about to [Wednesday 24 October 2012] [19:06:50] ok [Wednesday 24 October 2012] [19:07:31] so, basically is how to proceed for the future [Wednesday 24 October 2012] [19:08:01] Hm, maybe that's a bit too broad? [Wednesday 24 October 2012] [19:08:19] first, do we want to even keep a release schedule like that? an idea floating some time ago was to have always releasable master and release would be == device release [Wednesday 24 October 2012] [19:08:33] second, narrwing it down a bit [Wednesday 24 October 2012] [19:09:04] ie... the bad news [Wednesday 24 October 2012] [19:09:37] Even with an always releasable master, we'd probably still want to have intervals when we do the big announcement and media buzz stuff, right? [Wednesday 24 October 2012] [19:10:10] Join pvdm has joined this channel (~pvdm@2001:980:16f0:1:6267:20ff:fe50:9be). [Wednesday 24 October 2012] [19:10:12] yep, that is probably necessary [Wednesday 24 October 2012] [19:10:24] So I think even if we decide for alwqys releasable master, we can still plan ahead until the next "release" [Wednesday 24 October 2012] [19:10:46] another thing important: since the base team simply doesn't scale, there won't be new apps developed as the ones now [Wednesday 24 October 2012] [19:11:08] anybody can volunteer for writing a new app and we should really encourage and push that tough ;) [Wednesday 24 October 2012] [19:11:23] at least with "always stable" we end up decoupling release from development [Wednesday 24 October 2012] [19:11:27] for the media buzz (and for media folks to find press releases), you should decide what's your official website now -- the topic here does not mention plasma-active.org for instance [Wednesday 24 October 2012] [19:11:57] * aseigo isn't sure journalists come here to irc, but ok :) [Wednesday 24 October 2012] [19:12:10] aseigo, the LWN guys do [Wednesday 24 October 2012] [19:12:44] Wait, maybe we should do some formalisms first, like 1. Determining who's actually here and available 2. Create at least some kind of agenda [Wednesday 24 October 2012] [19:13:04] what's wrong with having no formalism? [Wednesday 24 October 2012] [19:14:04] Things tend to be a bit too chaotic and we may forget to discuss important topic [Wednesday 24 October 2012] [19:14:16] notmart: want me to chair this meeting? [Wednesday 24 October 2012] [19:14:34] Join bolo has joined this channel (~quassel@p5DCC7B1E.dip.t-dialin.net). [Wednesday 24 October 2012] [19:14:40] I don't think we need a very strict agenda, but at least collecting which topics we'd like to cover before going into them might make sense [Wednesday 24 October 2012] [19:14:52] there should be probably an interval for announcements until we can really do release==device.. [Wednesday 24 October 2012] [19:15:06] Join pinheiro has joined this channel (~pinheiro@bl15-4-250.dsl.telepac.pt). [Wednesday 24 October 2012] [19:15:27] well, I am here for Maliit and for Wayland; both goes hand in hand, for me (and both seems to require migrating to Qt5) [Wednesday 24 October 2012] [19:15:28] aseigo: if you wish would be great ;) [Wednesday 24 October 2012] [19:15:46] cool .. ok :) meeting in session. first off -> who is taking notes for us? [Wednesday 24 October 2012] [19:16:00] * aseigo can't do both :) [Wednesday 24 October 2012] [19:16:45] Well we have a log anyways, so I guess summaizing it afterwards would be okay. I can do that [Wednesday 24 October 2012] [19:16:45] heh.. crickets... [Wednesday 24 October 2012] [19:16:46] not sure you'd find my notes useful, but I can do it [Wednesday 24 October 2012] [19:16:59] colomar: cool .. consider yourself beknighted ... [Wednesday 24 October 2012] [19:17:01] hah, and the prize goes to colomar [Wednesday 24 October 2012] [19:17:07] *lol* [Wednesday 24 October 2012] [19:17:20] and speaking of getting things moving, yes .. who is here for the meeting? [Wednesday 24 October 2012] [19:17:23] * aseigo is here. [Wednesday 24 October 2012] [19:17:44] * mikhas is [Wednesday 24 October 2012] [19:17:47] * colomar is here [Wednesday 24 October 2012] [19:17:52] * notmart is here [Wednesday 24 October 2012] [19:18:20] i'll also as usual post on the ml log and little bullet points main things [Wednesday 24 October 2012] [19:19:13] So these are the ones from the doodle which have not spoken up yet: [Wednesday 24 October 2012] [19:19:19] notmart: great :) [Wednesday 24 October 2012] [19:19:21] sorry.. i need to leave in about ten minutes.. if it is of interest for you ;) [Wednesday 24 October 2012] [19:20:54] So these are the ones from the doodle which have not spoken up yet: Giorgos Tsiapaliokas, mdfe, Bogdan Cristea, Shantanu [Wednesday 24 October 2012] [19:21:09] and sebas [Wednesday 24 October 2012] [19:21:35] ah wait, strike Giorgos, he didn't indicate he had time [Wednesday 24 October 2012] [19:22:18] hm.. ok, well, we can push on and hopefully they'll show up at some point :) [Wednesday 24 October 2012] [19:22:36] the agenda for the meeting is really in two parts (pretty simple really) -> [Wednesday 24 October 2012] [19:22:39] a) schedule [Wednesday 24 October 2012] [19:22:41] b) scope [Wednesday 24 October 2012] [19:22:50] both for PA4, of course [Wednesday 24 October 2012] [19:23:13] schedule was already started by notmart .. and we have the divide between release schedule and development cycle [Wednesday 24 October 2012] [19:23:38] when we make the leap to "always stable" in master, i think that will be good, but also disruptive in its own way [Wednesday 24 October 2012] [19:23:50] not really [Wednesday 24 October 2012] [19:23:59] all you'd need would be feature branches, no? [Wednesday 24 October 2012] [19:24:07] which get merged once they are stable enough [Wednesday 24 October 2012] [19:24:08] mikhas: it'll change what people are doing ... :) [Wednesday 24 October 2012] [19:24:20] yes, and we already use feature branches extensively (thankfully we're not completely insane ;) [Wednesday 24 October 2012] [19:24:36] yeah feature branches plus integration branch as well [Wednesday 24 October 2012] [19:24:38] and we have (as a group) gotten rather good at merging-when-ready [Wednesday 24 October 2012] [19:24:41] so then … always releasable just takes some discipline (and CI) [Wednesday 24 October 2012] [19:24:54] we just don't have the integration branch and the integrator selected to oversee that ... [Wednesday 24 October 2012] [19:25:26] one other wrinkle is that this would only cover the plasma-mobile repo; runtime and workspace would continue on as is, at least until we make the port to Qt 5 / Frameworks 5 [Wednesday 24 October 2012] [19:25:50] yeah, that may be a blocker [Wednesday 24 October 2012] [19:26:00] personally, i'd be very much in favour of taking that final step, and i'm even happy and willing to play the added role of integrator [Wednesday 24 October 2012] [19:26:37] notmart: as they are dependencies for plasma-mobile, and not the other way around, it should be ok.. [Wednesday 24 October 2012] [19:26:56] could we make that step only in part ie having integration branch and master stable by policy but still doing a freeze before a release date after which no merges will happen? [Wednesday 24 October 2012] [19:27:01] just.. confusing for new people perhaps (or we can just pretend with them that runtime and workspace are done that way too.. after all, it is nearly how we are running workspace these das) [Wednesday 24 October 2012] [19:27:14] notmart: yes, that's the role of the integrator [Wednesday 24 October 2012] [19:27:46] whenever we come up to a release announcement, the integrator just stops taking merge requests for whatever time is needed to branch master [Wednesday 24 October 2012] [19:27:48] +1 for a freeze. That's needed for beta testing anyways [Wednesday 24 October 2012] [19:28:10] and that can really be as short as a few minutes we branch instantly upon hitting beta / RC [Wednesday 24 October 2012] [19:28:17] er, if we branch instantly [Wednesday 24 October 2012] [19:28:32] makes sense [Wednesday 24 October 2012] [19:28:55] so if we have a rolled release in june, we freeze on, say, may 1 [Wednesday 24 October 2012] [19:29:01] * sebas is also there, sorry for being late, meeting slipped my mind [Wednesday 24 October 2012] [19:29:10] the branch is created right then and there, from master [Wednesday 24 October 2012] [19:29:16] sebas: np.. glad you could join us [Wednesday 24 October 2012] [19:29:23] \o/ [Wednesday 24 October 2012] [19:29:30] we build images from the branch and test those [Wednesday 24 October 2012] [19:29:50] aseigo: the merges of the branches in integration would be continuously done by the ci i guess? [Wednesday 24 October 2012] [19:29:53] Join AlmAck has joined this channel (~AlmAck@156-19.198-178.cust.bluewin.ch). [Wednesday 24 October 2012] [19:29:59] fixes go into master and branch (pref with into branch and merge into master) [Wednesday 24 October 2012] [19:30:11] Sounds good to me [Wednesday 24 October 2012] [19:30:18] notmart: yes, in a perfect world :) [Wednesday 24 October 2012] [19:30:35] these would only be branches considered ready for testing, of course [Wednesday 24 October 2012] [19:30:45] sounds good [Wednesday 24 October 2012] [19:30:48] not stable, but not utter crap either :) [Wednesday 24 October 2012] [19:31:03] especially with having a diligent and skilled integrator, sounds like fun actually [Wednesday 24 October 2012] [19:31:10] aseigo: we'll take one day when i'm up to play with a ci on a server that could be used ;) [Wednesday 24 October 2012] [19:31:12] i.e. more people who get to try your code [Wednesday 24 October 2012] [19:31:17] eah, i'm sure we'll end up with a bunch of nice little scripts too [Wednesday 24 October 2012] [19:31:21] gtg bbl [Wednesday 24 October 2012] [19:31:24] notmart: perfect [Wednesday 24 October 2012] [19:31:27] rcg: see you later :) [Wednesday 24 October 2012] [19:31:30] Quit rcg has left this server (Quit: Leaving). [Wednesday 24 October 2012] [19:31:55] so ... release timing ... (not devel cycle :) [Wednesday 24 October 2012] [19:32:23] how about 6 monthly, one or two months after Frameworks releases? [Wednesday 24 October 2012] [19:32:25] shall we continue on in 6 month blocks, or should we try and sync up with the KDE SC fanfare? [Wednesday 24 October 2012] [19:32:50] We did 6 months blocks? :D [Wednesday 24 October 2012] [19:32:55] well, not really :P [Wednesday 24 October 2012] [19:33:01] but close [Wednesday 24 October 2012] [19:33:14] it was a when is ready kind of thing.. [Wednesday 24 October 2012] [19:33:18] 3 months and 10 months, ... pretty close :P [Wednesday 24 October 2012] [19:33:36] sebas: i'm a statistician. that's 6.5 [Wednesday 24 October 2012] [19:33:38] ;P [Wednesday 24 October 2012] [19:33:48] sync with sc would actually be 6 months instead ;) [Wednesday 24 October 2012] [19:33:53] I guess somewhat syncing with KDE SC makes sense so we can safely use stable versions of the frameworks [Wednesday 24 October 2012] [19:33:57] shortly after a Frameworks release means we could depend on that, and base upon .1 or .2, so kind of a tested thing already [Wednesday 24 October 2012] [19:34:20] sebas: +1 [Wednesday 24 October 2012] [19:34:26] a pro for it is that it gets synced on what version of kde-workspace, kde-runtime etc you can use [Wednesday 24 October 2012] [19:34:31] disadvantage: features needed in Frameworks have to be in ~4months before release (but kdelibs is frozen anyway for now) [Wednesday 24 October 2012] [19:34:38] mm.. that really depends on how quickly we could port to Frameworks 5. and that may require a whole cycle on its own. [Wednesday 24 October 2012] [19:34:53] and knowing Murphy, may = will [Wednesday 24 October 2012] [19:34:58] yes, but I think the F5 jump is special anyway [Wednesday 24 October 2012] [19:35:22] at least spotting a release soon after Frameworks 5.0 gives us a good starting point in time to get that work done [Wednesday 24 October 2012] [19:35:25] we can always break our own rules for something like that [Wednesday 24 October 2012] [19:35:38] does F5 have a release date yet? [Wednesday 24 October 2012] [19:35:52] with "2 months after Frameworks", I actually meant two months after SC releases for now [Wednesday 24 October 2012] [19:36:00] aha.. :) [Wednesday 24 October 2012] [19:36:01] F5 has no release date yet [Wednesday 24 October 2012] [19:36:04] well, we need plasma2 anyways for that [Wednesday 24 October 2012] [19:36:19] I was being deliberately confusing to steal you all your time! ;) [Wednesday 24 October 2012] [19:36:29] *gg* [Wednesday 24 October 2012] [19:36:46] colomar: nice article, btw [Wednesday 24 October 2012] [19:37:03] sebas: thx [Wednesday 24 October 2012] [19:37:14] yep, indeed [Wednesday 24 October 2012] [19:37:34] Part starbuck has left this channel. [Wednesday 24 October 2012] [19:37:51] Quit siddvicious has left this server (Remote host closed the connection). [Wednesday 24 October 2012] [19:38:00] ok... so next release is 4.10 and scheduled for January 23 [Wednesday 24 October 2012] [19:38:15] that would give us a PA 4 in March with sebas' suggestion [Wednesday 24 October 2012] [19:38:24] sounds fine to me [Wednesday 24 October 2012] [19:38:27] i'm comfortable with that. [Wednesday 24 October 2012] [19:38:43] +1 [Wednesday 24 October 2012] [19:39:04] yay [Wednesday 24 October 2012] [19:39:07] ok.. let's run with that then.. March 31 is a sunday [Wednesday 24 October 2012] [19:39:12] as is the 24th [Wednesday 24 October 2012] [19:39:26] shall we shoot for the 24th with announcements on the monday of the 25th? [Wednesday 24 October 2012] [19:39:35] best day for releases are Tuesday [Wednesday 24 October 2012] [19:39:39] in the meantime we try to have already branches+integration to keep plasma-mobile unbroken all the way [Wednesday 24 October 2012] [19:39:54] right [Wednesday 24 October 2012] [19:39:57] Tue - Thu is good for press attention, Tue gives us two days to run beyond schedule [Wednesday 24 October 2012] [19:40:07] ok.. 25th for final, announcement on 26th [Wednesday 24 October 2012] [19:40:26] ok [Wednesday 24 October 2012] [19:40:30] Nick sujith_ is now known as sujith_h. [Wednesday 24 October 2012] [19:41:30] great... so devel cycle. with this rolling stability, i think it is nearly irrelevant. we'll just work as hard as we can and instead prioritize tasks? [Wednesday 24 October 2012] [19:41:55] sounds good [Wednesday 24 October 2012] [19:42:00] yeah [Wednesday 24 October 2012] [19:42:19] ok.. so .. that brings us to point (b) on the agenda [Wednesday 24 October 2012] [19:42:24] goals :) [Wednesday 24 October 2012] [19:42:45] what do we want to accomplish in the coming months, with what priorities [Wednesday 24 October 2012] [19:42:46] Hm I'd like to set a date for the beta as well [Wednesday 24 October 2012] [19:43:02] colomar: release - 1 month? [Wednesday 24 October 2012] [19:43:05] So that we don't end up with not enough testing time [Wednesday 24 October 2012] [19:43:05] colomar: sure ... i'd just as soon say 1 month [Wednesday 24 October 2012] [19:43:10] :) [Wednesday 24 October 2012] [19:43:27] since mikhas is here, perhaps we can cover keyboard / maliit issues first? [Wednesday 24 October 2012] [19:43:27] ok [Wednesday 24 October 2012] [19:43:32] :-) [Wednesday 24 October 2012] [19:43:48] Ok. I have quite some points for that [Wednesday 24 October 2012] [19:43:56] start? [Wednesday 24 October 2012] [19:44:03] colomar: ah, not to discount needing to schedule beta/rc ... i think we can just agree with a 1 month window for that [Wednesday 24 October 2012] [19:44:11] mikhas: yep ... [Wednesday 24 October 2012] [19:44:13] colomar: shoot [Wednesday 24 October 2012] [19:44:44] aseigo: yes, one month is okay since we have always releasable master anyway, I guess ;) [Wednesday 24 October 2012] [19:44:53] Ok, so about maliit [Wednesday 24 October 2012] [19:45:14] These are the things I noted so far: [Wednesday 24 October 2012] [19:45:20] - Internationalization [Wednesday 24 October 2012] [19:46:05] - Better terminal support (tab, cursor keys, more easily reachable special characters needed for the shell) [Wednesday 24 October 2012] [19:46:22] - Resizing of some keys (e.g. space should be wider) [Wednesday 24 October 2012] [19:46:45] - Generally it should use more of the available horizontal space [Wednesday 24 October 2012] [19:47:21] - Long-tap function cannot be used in conjunction with shift (so e.g. no capitaql umlauts) [Wednesday 24 October 2012] [19:48:17] re I18N: the issue can be split into two -- language layouts and word engines. Language layouts is easy, word engines (that take I18N into account) are seriously hard and almost impossible without help from vendors of proprietary word engines … this means you'd need to go for OEM deals eventually, because that's the prevalent business model in that area [Wednesday 24 October 2012] [19:48:24] - Currently you cannot select text while the keyboard is shown since averything else is non-interactive [Wednesday 24 October 2012] [19:48:35] a=e [Wednesday 24 October 2012] [19:49:14] mikhas: Oh you remind me: We currently have no autocomplete at all [Wednesday 24 October 2012] [19:49:18] Nick trueg_away is now known as s-trueg. [Wednesday 24 October 2012] [19:49:19] yep [Wednesday 24 October 2012] [19:49:24] last point is that the qml keyboards are actually full screen windows and that doesn't work that well [Wednesday 24 October 2012] [19:49:48] Nick mck182 is now known as mck182|afk. [Wednesday 24 October 2012] [19:49:50] notmart: Yes, that's basically what I meant with my last bullet, from a user's perspective ;) [Wednesday 24 October 2012] [19:49:53] aseigo had content-specific keyboards in mind [Wednesday 24 October 2012] [19:50:03] Oh yes, absolutely [Wednesday 24 October 2012] [19:50:25] we have support for that in the framework, and we recently patched GTK+ to have some content-type hints, too [Wednesday 24 October 2012] [19:50:39] so Qt, EFL and GTK+ now all offer something in that direction [Wednesday 24 October 2012] [19:50:43] E.g. a numeric pad, terminal, email... [Wednesday 24 October 2012] [19:51:01] nice [Wednesday 24 October 2012] [19:51:03] and maliit-framework supports content-specific keyboards … it's the plugins that need to implement it [Wednesday 24 October 2012] [19:51:36] yes, i think the #1 improvement to focus on in they keyboard is layouts support [Wednesday 24 October 2012] [19:51:42] it shouldnt even be too difficult, something I consider a low-hanging fruit → needs a bug report somewhere [Wednesday 24 October 2012] [19:51:50] from a technical point of view, all kindof depends from how plugins are done, ie the current qml ones are quite limited [Wednesday 24 October 2012] [19:51:56] (things like longtap + shift not working -> that's just a plain bug and needs fixing) [Wednesday 24 October 2012] [19:52:15] notmart: limited in which way? [Wednesday 24 October 2012] [19:52:16] right, let's treat bugs separate (just report them) [Wednesday 24 October 2012] [19:52:19] so is continuing in this direction vs waiting for qml inside the c++ plugin instead [Wednesday 24 October 2012] [19:52:26] the QML plugin API we have in maliit-framework is very poor [Wednesday 24 October 2012] [19:52:39] aseigo: the keyboard layout is actually hardcoded in the qml file [Wednesday 24 October 2012] [19:52:49] and maintaining two plugin API is just wasted effort [Wednesday 24 October 2012] [19:53:06] ah, yes.. well hardcoded is "fine", as long as we can switch between them [Wednesday 24 October 2012] [19:53:10] well, I think we will simply hide the C++ plugin API behind nice QML components [Wednesday 24 October 2012] [19:53:22] that's the plan for Maliit 1.0 at least [Wednesday 24 October 2012] [19:53:39] → one plugin API, but nice to re-use for QML [Wednesday 24 October 2012] [19:53:44] writing a QML generator from the layout description files is a couple days work at best at this point .. but i am interested in where maliit is currently going :) [Wednesday 24 October 2012] [19:54:00] mikhas: what's the timeframe for that, and how would it impact our usage of it? [Wednesday 24 October 2012] [19:54:08] One goal should be that layouts can be created easily (e.g. for i18n purposes) [Wednesday 24 October 2012] [19:54:13] I am supposed to rip out the event handling and rendering from the Maliit Keyboard anyway, for Qt5 [Wednesday 24 October 2012] [19:54:36] and I am going to study Nemo Keyboard for some API ideas, for my QML components [Wednesday 24 October 2012] [19:54:58] with some help from notmart, that direction should work out for everyone [Wednesday 24 October 2012] [19:55:59] summarized: I am still not convinced that throwing away all the C++ stuff is beneficial for us (but I have a different view of course, due to customer work) [Wednesday 24 October 2012] [19:56:00] Join rabuzarus has joined this channel (~rabuzarus@pD9E3B332.dip.t-dialin.net). [Wednesday 24 October 2012] [19:56:29] but we certainly want to get rid of QGraphicsView in any case, and target QtQuick2 in the mid- to long-term [Wednesday 24 October 2012] [19:57:16] anyway, generating QML files from XML: yes, sounds like a fun experiment. but I wont be the one doing it [Wednesday 24 October 2012] [19:57:35] next: better terminal support [Wednesday 24 October 2012] [19:57:53] the good thing is that we have this action and key event API now, in maliit-framework [Wednesday 24 October 2012] [19:57:59] action is something like copy+paste [Wednesday 24 October 2012] [19:58:09] and key events allows to send key sequences, too [Wednesday 24 October 2012] [19:58:16] so that helps with keyboard shortcuts [Wednesday 24 October 2012] [19:58:27] again, it's the plugins that need to implement it [Wednesday 24 October 2012] [19:58:36] I like the idea of treating terminal mode as a content type [Wednesday 24 October 2012] [19:58:54] none of the toolkit frameworks supports it that way, but it sounds like a good solution to me [Wednesday 24 October 2012] [19:59:03] For me ideally this would go hand in hand with the content-specific layouts: In bash, you need lots of special characters very often which you need rather rarely when typing text [Wednesday 24 October 2012] [19:59:07] and we can easily try it out with Qt + plasma active [Wednesday 24 October 2012] [19:59:21] looking at org/kde/active/keyboard/LandscapeVKB.qml, generating the layouts from xml looks like it should be quite simle. [Wednesday 24 October 2012] [19:59:39] (oh, yes, you already mentioned it mikhas) [Wednesday 24 October 2012] [19:59:54] mikhas: the current plugin arch alread has access to content types, correct? [Wednesday 24 October 2012] [20:00:09] resizing of keys → purely a design issue, not my area to care about [Wednesday 24 October 2012] [20:00:14] aseigo, correct [Wednesday 24 October 2012] [20:00:21] a problem that there will be in both cases (both using models and repeaters and pre generated files) [Wednesday 24 October 2012] [20:00:30] the plugin parts can probably be lifted from qt-components for a large part, I'm particularly thinking of the InputMethodHints (or how it's called there) [Wednesday 24 October 2012] [20:00:53] qt-components did a very poor reimplemention of Qt4 input method API [Wednesday 24 October 2012] [20:01:07] how can you adapt to many different resolutions? generating a qml file that scales well, generic enough that any layout can produce a thing good enough doesn't seem trivial [Wednesday 24 October 2012] [20:01:12] mikhas: do you recall what specific problems it has? [Wednesday 24 October 2012] [20:01:13] with Qt5, the gap luckily doesnt exist, because the Qt5 input method API is QML-friendly by default [Wednesday 24 October 2012] [20:01:22] sebas, no input method queries, for instance [Wednesday 24 October 2012] [20:01:39] (a hack, yes, but still the only way to implement custom behaviour) [Wednesday 24 October 2012] [20:01:49] things like a ".com" button or somesuch worked at least [Wednesday 24 October 2012] [20:02:08] notmart, I have an idea for that [Wednesday 24 October 2012] [20:02:27] scaling: query a style component that you feed with some key values, and it spits out the pixel values [Wednesday 24 October 2012] [20:02:39] in QML, this would all be property bindings [Wednesday 24 October 2012] [20:03:18] that's how the styling in the C++ plugin works already, even though it might not be visible as such [Wednesday 24 October 2012] [20:03:31] so the model directly gives pixel values? [Wednesday 24 October 2012] [20:03:36] sure, why not [Wednesday 24 October 2012] [20:03:49] as long as they can be adjusted dynamically … [Wednesday 24 October 2012] [20:04:27] sounds like something that should be done in .js though, as it's more on the representation side than on the data/model side [Wednesday 24 October 2012] [20:04:42] Yes, this would be very helpful, because both a too small and a too large keyboard are quite a problem [Wednesday 24 October 2012] [20:04:44] yeah, it's quite in between tough [Wednesday 24 October 2012] [20:04:48] notmart: i don't want to get bogged down in details, but i have thoughts on how to resolve those challenges [Wednesday 24 October 2012] [20:05:13] well, I am open for advice. I'll just go and try what I think could work, and then you tell me how to do it better ;-) [Wednesday 24 October 2012] [20:05:14] yes, I think a couple of work items for this topic is all we need from this meeting [Wednesday 24 October 2012] [20:05:21] since the model is also the key disposition, how the keyboard looks [Wednesday 24 October 2012] [20:05:35] aseigo: curious ;) [Wednesday 24 October 2012] [20:05:35] then: "Currently you cannot select text while the keyboard is shown since averything else is non-interactive" [Wednesday 24 October 2012] [20:05:40] that one is tough [Wednesday 24 October 2012] [20:06:01] yes; i don't think we need to put it on the table for this round [Wednesday 24 October 2012] [20:06:03] text selection in the N9 is where our team easily spend half a year to get it decent [Wednesday 24 October 2012] [20:06:06] I'm sure mgraesslin could come up with a hack ;) [Wednesday 24 October 2012] [20:06:12] aanyways, we can continue with that for some minutes, then summarize what is still needed/open to discussion and go on.. [Wednesday 24 October 2012] [20:06:16] i'd be happy with first picking off: [Wednesday 24 October 2012] [20:06:28] a) remaining bugs (e.g. the cap lock thing + long tap thing) [Wednesday 24 October 2012] [20:06:29] for text selection, I would suggest as "select" key + arrow keys work around, somewhere [Wednesday 24 October 2012] [20:06:37] (that's what keyboard like Onboard do) [Wednesday 24 October 2012] [20:06:58] b) implementing multiple layouts [Wednesday 24 October 2012] [20:07:36] c) react to input method hints (Qt::InputMethodHint values) [Wednesday 24 October 2012] [20:07:49] when we clear those, we can re-load with more keyboard TODOs [Wednesday 24 October 2012] [20:07:53] ok [Wednesday 24 October 2012] [20:08:00] sounds good [Wednesday 24 October 2012] [20:08:17] unless someone can suggest something with more priorit / user impact, of course :) [Wednesday 24 October 2012] [20:08:33] hm... I'd put a higher priority on the sizing [Wednesday 24 October 2012] [20:08:39] scaling [Wednesday 24 October 2012] [20:08:39] just one more point: Maliit 1.0 will be Qt5-only internally (that is, not compile with Qt4) [Wednesday 24 October 2012] [20:09:00] and Maliit 1.0 will most likely be one of the input methods that run with Wayland out of the box [Wednesday 24 October 2012] [20:09:04] mikhas: and when again is that due out? [Wednesday 24 October 2012] [20:09:15] 2013/Q2 [Wednesday 24 October 2012] [20:09:24] ok.. great :) so we can focus on that for PA5 [Wednesday 24 October 2012] [20:09:25] dont see how I would get it done earlier than that … [Wednesday 24 October 2012] [20:09:25] PA5 timeframe [Wednesday 24 October 2012] [20:09:30] sebas: lol :) [Wednesday 24 October 2012] [20:09:32] :) [Wednesday 24 October 2012] [20:09:57] mikhas: thanks for your time.. really, really appreciate you joining us tonight... feel free to stick around, of course, but i think we'll move on to other topics now [Wednesday 24 October 2012] [20:09:59] mikhas: we'll need some time to port to Qt5 and KDE Frameworks5 anyway [Wednesday 24 October 2012] [20:10:02] so, until then, keep our keyboard and fix issues like scaling/various bugs [Wednesday 24 October 2012] [20:10:09] not sure about layouts [Wednesday 24 October 2012] [20:10:25] well, they could be generated one-off from the layouts [Wednesday 24 October 2012] [20:10:34] but would be a temporary solution [Wednesday 24 October 2012] [20:10:47] Would it be possible to put in at least a tab key in general? [Wednesday 24 October 2012] [20:11:12] Because that's essential for comfortably filling in forms as well [Wednesday 24 October 2012] [20:11:15] other point that i'm not sure is fixable before maliit 1.0 is having a keyboard that doesn't block input [Wednesday 24 October 2012] [20:11:38] notmart: yes, i alread put that one off :) [Wednesday 24 October 2012] [20:11:45] Well, I don't think that's a huge problem [Wednesday 24 October 2012] [20:12:07] adding tab should be easy i guess [Wednesday 24 October 2012] [20:12:17] colomar: yes, tab is possible .. rather handy for certain things as you noted [Wednesday 24 October 2012] [20:12:21] It's a bit confusing at first, but once you get used to it, you just hide the keyboard if you have to select something different [Wednesday 24 October 2012] [20:12:31] rcg is doing some work on the keyboard [Wednesday 24 October 2012] [20:12:32] notmart, I will know more about input blocking once I get something to work with QtQuick2 and Qt5 … [Wednesday 24 October 2012] [20:12:40] he said especially to make it work better with konsole [Wednesday 24 October 2012] [20:12:51] we should check back with him about his plans, and incorporate that [Wednesday 24 October 2012] [20:12:56] agreed [Wednesday 24 October 2012] [20:13:06] MeeGo Terminal contains a lot of hints how to integrate Maliit with a terminal [Wednesday 24 October 2012] [20:13:08] aseigo: Ok. That also makes the input blocking even less of a problem since ou can just tab thorugh fields instead of selecting them on the screen [Wednesday 24 October 2012] [20:13:14] he was also on arrow keys, which are also quite useful sometimes [Wednesday 24 October 2012] [20:13:19] so .. other items in PA to focus on ... [Wednesday 24 October 2012] [20:13:38] * sebas has sync on the list [Wednesday 24 October 2012] [20:13:47] wanna continue with that, or something else first? [Wednesday 24 October 2012] [20:13:50] mikhas: in any way, with qt5 as well is always the same: issue of using a fullscreen window vs just as big as the keyboard (or an input mask applied to the fullscreen window) [Wednesday 24 October 2012] [20:14:10] sebas: no, let's cover that ground ... [Wednesday 24 October 2012] [20:14:20] +1 [Wednesday 24 October 2012] [20:14:24] ok, so my thoughts, plans... [Wednesday 24 October 2012] [20:14:58] the issue is multifold, it's about emails, contacts, messages in general (social networks), bookmarks and files [Wednesday 24 October 2012] [20:15:44] most of them are Akonadi things (we get caching, central store and retrieval for free), but the setup is a beyond horrible user experience [Wednesday 24 October 2012] [20:16:05] then files ... I've been working on an ownCloud client, which now mostly works [Wednesday 24 October 2012] [20:16:34] Quit arc_mat_ has left this server (Ping timeout: 240 seconds). [Wednesday 24 October 2012] [20:16:38] it has a daemon that controls the syncing with the server, an active settings module and a plasmoid to enable / disable folders [Wednesday 24 October 2012] [20:17:04] it's in a branch of Mirall on github, we'd need it packaged and tested [Wednesday 24 October 2012] [20:17:07] enable / disable folders? [Wednesday 24 October 2012] [20:17:12] Enabling/disabling folders won't be of much help for PA, since we're hiding the file sysdtem [Wednesday 24 October 2012] [20:17:18] ugh.. github. [Wednesday 24 October 2012] [20:17:23] yes, on the road you might not want to sync your huge photo collection [Wednesday 24 October 2012] [20:17:31] yes, ugh github [Wednesday 24 October 2012] [20:17:43] sebas: Then we have to go for file types [Wednesday 24 October 2012] [20:17:53] It's all Nepomuk on PA, remember? ;) [Wednesday 24 October 2012] [20:17:57] I need to rebase my branch on the latest Mirall stable [Wednesday 24 October 2012] [20:18:09] colomar: we need some integration tricks there [Wednesday 24 October 2012] [20:18:31] the synching lib used doesn't understand KIO unfortunately, so it needs physical folders on disk to work [Wednesday 24 October 2012] [20:18:36] not really doable without folders for file syncing [Wednesday 24 October 2012] [20:18:52] it uses the standard Picture / Music / Documents / etc folders in the homedir right now [Wednesday 24 October 2012] [20:18:53] even if it would probably [Wednesday 24 October 2012] [20:19:18] yes, that's probably fine. i don't know if we'll end up needing such fine-grained control, though. [Wednesday 24 October 2012] [20:19:23] so we have to make sure stuff ends up in the right directories [Wednesday 24 October 2012] [20:19:35] i'd imagine it could be more elegant to offer: [Wednesday 24 October 2012] [20:19:43] we first need to look at my current code, see what works and what doesn't [Wednesday 24 October 2012] [20:19:49] Join ingwa has joined this channel (~ingwa@static-88.131.62.36.addr.tdcsong.se). [Wednesday 24 October 2012] [20:19:51] * simple on/off (based on network location would be awesome) [Wednesday 24 October 2012] [20:19:59] it seems doable in the next release to me [Wednesday 24 October 2012] [20:20:12] aseigo: that's really easy to do with the current code [Wednesday 24 October 2012] [20:20:14] * bandwidth limit (based on network location would be almost mandatory there) [Wednesday 24 October 2012] [20:20:21] I've made it so one can easily drop in a policy [Wednesday 24 October 2012] [20:20:32] just haven't written any policy profiles yet [Wednesday 24 October 2012] [20:20:37] * aseigo nods [Wednesday 24 October 2012] [20:20:42] one thing important for syncing on devices is some quota, ie maybe between desktop, nas and laptop i'm syncing 200+gb of stuff... [Wednesday 24 October 2012] [20:21:08] the mobile devices would be the only ones that only get a small excerpt of this because well, they're tiny [Wednesday 24 October 2012] [20:21:13] upload is quite easy, but we will need to check for new stuff on the server once in a while, so can't be done entirely without wakeups :/ [Wednesday 24 October 2012] [20:21:15] and this requires mirall from git .. elsewhere. [Wednesday 24 October 2012] [20:21:17] What I think could be a typical setup is users having a local ownCloud server on their desktop machine with no access to the internet and just syncin when they're at home [Wednesday 24 October 2012] [20:21:46] sebas: what would you think about not including sync by default in the PA images but having it available via bodega? [Wednesday 24 October 2012] [20:22:01] it actually uses only part of mirall [Wednesday 24 October 2012] [20:22:08] colomar: yes, the concept of "home network" (plural,, though) needs to happen at some point [Wednesday 24 October 2012] [20:22:37] aseigo: it will need a server anyway, guess that's fine [Wednesday 24 October 2012] [20:22:55] this would accomplish a few things: [Wednesday 24 October 2012] [20:23:03] as long as we only support owncloud and no dropbox / google storage / skydrive crap, it won't be very generic [Wednesday 24 October 2012] [20:23:10] * note relying on packages outside of git.kde.org for the release images [Wednesday 24 October 2012] [20:23:15] s,note,not, [Wednesday 24 October 2012] [20:23:15] Join sujith has joined this channel (~sujith@122.178.205.201). [Wednesday 24 October 2012] [20:23:32] * ensuring we have a really nice use case for obs feeds through bodega [Wednesday 24 October 2012] [20:23:35] well, I can put a forked tree on git.kde.org, no problem [Wednesday 24 October 2012] [20:23:37] :) [Wednesday 24 October 2012] [20:23:45] aseigo: I'd prefer having syncing available by default, since it's a pretty basic function [Wednesday 24 October 2012] [20:23:59] it's not like anybody else works on my branch [Wednesday 24 October 2012] [20:24:34] colomar: to have it work by default one still needs a server setup, that either means setting one up yourself, buying an account or asking $sysadmin [Wednesday 24 October 2012] [20:24:55] I see bodega more for things that nobody misses if it's not there, but people are happy if they can get it in the store [Wednesday 24 October 2012] [20:24:58] yep.. and this is, as sebas noted, one syncing solution rather than a generic one... [Wednesday 24 October 2012] [20:25:08] the active settings ui could even be in by default.. [Wednesday 24 October 2012] [20:25:13] configure [Wednesday 24 October 2012] [20:25:15] colomar: that's a hard line to draw as it changes from preson to person [Wednesday 24 October 2012] [20:25:20] and installs the needed packages [Wednesday 24 October 2012] [20:25:26] aseigo: Sure [Wednesday 24 October 2012] [20:25:41] notmart: That would work as well, yes [Wednesday 24 October 2012] [20:25:42] Quit sujith_h has left this server (Read error: Connection reset by peer). [Wednesday 24 October 2012] [20:25:51] notmart: that turns settings into an add-ons getter too .. i'm not super keen on that [Wednesday 24 October 2012] [20:26:04] "here's lots of things you can configure your device to do but it doesn't do right now .. " [Wednesday 24 October 2012] [20:26:09] ".. just have to download more stuff" [Wednesday 24 October 2012] [20:26:18] yeah.. [Wednesday 24 October 2012] [20:26:20] that's going to translate into "settings has lots of pages for things i don't use" [Wednesday 24 October 2012] [20:26:25] hmm [Wednesday 24 October 2012] [20:26:25] another idea could be [Wednesday 24 October 2012] [20:26:29] on the addons app [Wednesday 24 October 2012] [20:26:30] we could integrate stuff from amazon, too!! [Wednesday 24 October 2012] [20:26:37] sebas: 4) profit [Wednesday 24 October 2012] [20:26:39] *lol* [Wednesday 24 October 2012] [20:26:43] having more entries for the sync software [Wednesday 24 October 2012] [20:26:47] "Features Apps" in the store? [Wednesday 24 October 2012] [20:27:00] and then we pimp all our cool personal crapware! ;) [Wednesday 24 October 2012] [20:27:06] sebas: yes, that's on the priority TODO (just finishd putting that togehter today actually) [Wednesday 24 October 2012] [20:27:07] like preconfigured with service1, preconfigured with service2 [Wednesday 24 October 2012] [20:27:12] notmart: yep [Wednesday 24 October 2012] [20:27:16] do it yourself for you home network [Wednesday 24 October 2012] [20:27:26] that's the idea i've been going on about for probably over a year now :) [Wednesday 24 October 2012] [20:27:44] the home network version could even launch the local area network scan automatically [Wednesday 24 October 2012] [20:27:48] Would I like about notmart's idea in this case is that I think pretty much every user will want some sort of syncing, but not all users use the same. So there is a syncing config for everyone which installs stuff specific to a certain type of syncing when it's chosen [Wednesday 24 October 2012] [20:27:57] sounds cool then [Wednesday 24 October 2012] [20:28:20] colomar: yep. and that's something bodega is good at doing [Wednesday 24 October 2012] [20:28:27] showing choices, letting you install them. [Wednesday 24 October 2012] [20:28:30] even abstracts what it uses [Wednesday 24 October 2012] [20:28:34] for the rest of $synching, we depend on Accounts (which then needs an Active UI) [Wednesday 24 October 2012] [20:28:44] (just to finish that non-specific part) [Wednesday 24 October 2012] [20:28:47] ok, so that's syncing .. sounds like good progress.. keep us informed :) [Wednesday 24 October 2012] [20:28:59] ie if some day dropbox or whatever will want to offer one, will just be listed together anything else even if installs different software [Wednesday 24 October 2012] [20:29:01] Accounts ... afiestas is meeting with the meego accounts framework people next week @ UDS [Wednesday 24 October 2012] [20:29:05] yes, I'll find some time to package and try what I currently have [Wednesday 24 October 2012] [20:29:05] notmart: yep [Wednesday 24 October 2012] [20:29:31] ah, good news for the accounts thing [Wednesday 24 October 2012] [20:29:34] i'm really hoping that we can just adopt that stack and throw our UIs on it [Wednesday 24 October 2012] [20:29:40] personal schedule-wise, I first want to get the containment stuff in before feature freeze, and then will likely have a bit more time for synching and F5 work [Wednesday 24 October 2012] [20:29:45] afiestas seems hopeful at this point (he was looking at some of it recently) [Wednesday 24 October 2012] [20:29:51] sebas: sounds good :) [Wednesday 24 October 2012] [20:29:52] (i.e. package async) [Wednesday 24 October 2012] [20:30:02] yeah, i am always happy when there is a nih less ;) [Wednesday 24 October 2012] [20:30:13] we'd need some serious sit together and hack up a critical mass for the accounts stuff [Wednesday 24 October 2012] [20:30:32] yep [Wednesday 24 October 2012] [20:30:34] there's enough people interested that we could probably muster up a few people for a bit [Wednesday 24 October 2012] [20:30:57] and it's awfully important for a *lot* of things on the roadmap [Wednesday 24 October 2012] [20:31:10] let's see how the meego framework pans out... if that's our direction then we can attack it directly and with focus [Wednesday 24 October 2012] [20:31:22] * aseigo needs to go soon... [Wednesday 24 October 2012] [20:31:22] right [Wednesday 24 October 2012] [20:31:33] other items .. well, i have a simple world view on them [Wednesday 24 October 2012] [20:31:38] so to recap sync: [Wednesday 24 October 2012] [20:31:46] we introduced a number of great new components and applications in PA3 [Wednesday 24 October 2012] [20:31:52] * aseigo waits for the recap [Wednesday 24 October 2012] [20:31:56] - contact / messages needs work for Accounts and Klapetek's social feeds [Wednesday 24 October 2012] [20:32:18] - files to be tested for basic file sync usecases (that's what we'll be able to deliver with PA4) [Wednesday 24 October 2012] [20:32:47] (done) [Wednesday 24 October 2012] [20:32:48] Oh, I just noticed another point for syncing: Is there _any_ way we can keep things like tags when syncing? Since we make metadata so important, It would be a shame if it would all be lost during the sync [Wednesday 24 October 2012] [20:33:19] eh, don't know about that [Wednesday 24 October 2012] [20:33:23] colomar: there's a way in principle, but the whole metadata synching thing is quite tricky, too tricky to prioritize right now [Wednesday 24 October 2012] [20:33:30] the way that *seems* possible is: [Wednesday 24 October 2012] [20:33:43] maybe by writing an owncloud plugin to store at least dome key/value data for synced files [Wednesday 24 October 2012] [20:33:49] nepomuk has sync and backup capabilities, which basically dump metainformation patches into files [Wednesday 24 October 2012] [20:34:12] those files *could* be synched to owncloud and dynamically be imported in the Nepomuk on other side [Wednesday 24 October 2012] [20:34:33] I've no idea if that works, and how well it does, it's one of those half implemented things [Wednesday 24 October 2012] [20:35:19] yes, that's another topic.. nepomuk sync ... definitely sth to get to later .. basic sync -> 1st priority [Wednesday 24 October 2012] [20:35:25] metadata sync -> next ball of wax [Wednesday 24 October 2012] [20:35:34] yep, let's go by dependency ;) [Wednesday 24 October 2012] [20:35:43] it already bears quite some value to see your photos transferred to your desktop automatically :) [Wednesday 24 October 2012] [20:35:47] so, move on? [Wednesday 24 October 2012] [20:35:47] ok [Wednesday 24 October 2012] [20:35:52] move on. [Wednesday 24 October 2012] [20:36:04] on my side: on plasma-mobile is mostly file browser: some bugfix and stuffs still missing, adapt parts of its ui for the add items stuff in the workspace, hopefully move the metadata models in nepomuk-core (and they would need quite some rework for it) [Wednesday 24 October 2012] [20:36:29] Files would also be next on my list [Wednesday 24 October 2012] [20:36:44] So should we talk about that now? [Wednesday 24 October 2012] [20:36:53] no, if that's alright :) [Wednesday 24 October 2012] [20:36:53] Join jpwhiting_ has joined this channel (~jeremy@67-2-193-173.slkc.qwest.net). [Wednesday 24 October 2012] [20:37:11] don't want to do big changes, mostly add some of the mostly annoying missing things.. file rename, tags delete.. [Wednesday 24 October 2012] [20:37:21] a) i need to go; b) i don't think discussing everything under the sun right here is a useful approach. we'll end up with a giant list and no focus. [Wednesday 24 October 2012] [20:37:31] Quit jpwhiting has left this server (Quit: Coyote finally caught me). [Wednesday 24 October 2012] [20:37:32] right [Wednesday 24 October 2012] [20:37:37] Quit lamarque has left this server (Quit: Konversation terminated!). [Wednesday 24 October 2012] [20:37:47] what might work nicer is to block our development into 2 or 3 week spans [Wednesday 24 October 2012] [20:37:56] and then pick one component (e.g. files) to work on [Wednesday 24 October 2012] [20:38:05] Quit jpwhiting_ has left this server (Client Quit). [Wednesday 24 October 2012] [20:38:12] so to summarize on my side: most work is needed for me on the file browser (and about 30% on the ui, 70% on the backend) [Wednesday 24 October 2012] [20:38:13] we would spend the first N days enumerating what works and what doesn't [Wednesday 24 October 2012] [20:38:18] Join jpwhiting_ has joined this channel (~jeremy@67-2-193-173.slkc.qwest.net). [Wednesday 24 October 2012] [20:38:23] Nick jpwhiting_ is now known as jpwhiting. [Wednesday 24 October 2012] [20:38:29] Sounds good. But maybe prioritize which blocks we want to work on beforehand? [Wednesday 24 October 2012] [20:38:38] Quit jpwhiting has left this server (Changing host). [Wednesday 24 October 2012] [20:38:39] Join jpwhiting has joined this channel (~jeremy@kde/developer/whiting). [Wednesday 24 October 2012] [20:38:42] and the rest of the cycle implementing what needs to be done to improve it. then move on to the next item. and really, really polish [Wednesday 24 October 2012] [20:38:48] colomar: yes, that we can do now. [Wednesday 24 October 2012] [20:38:52] i think we're all in agreement on Files :) [Wednesday 24 October 2012] [20:38:54] the workspace, instead is basically fine, so is quite at bottom of priority right now [Wednesday 24 October 2012] [20:39:04] Books + the reader also needs work [Wednesday 24 October 2012] [20:39:16] Alarms [Wednesday 24 October 2012] [20:39:20] SLC [Wednesday 24 October 2012] [20:39:21] notmart: agree, and that's a huge achievement in itself :) [Wednesday 24 October 2012] [20:39:22] Books is the "shelf"? [Wednesday 24 October 2012] [20:39:22] well, books is files right now ;) [Wednesday 24 October 2012] [20:39:38] Add Ons [Wednesday 24 October 2012] [20:39:51] ingwa: yes .. [Wednesday 24 October 2012] [20:40:08] notmart: i know :) but i think if we take time to focus on how it works in conjunction with the reader ... [Wednesday 24 October 2012] [20:40:27] addons.. ui also fine i think, it needs to reliably provide stuff from obs, or even other kinds of providers maybe [Wednesday 24 October 2012] [20:40:38] aseigo: I think the add Ons App itself works fine from a user perspective. What we need now of course is content ;) [Wednesday 24 October 2012] [20:40:44] we'll end up with some important points beyond what we'll get when looking at it as Files [Wednesday 24 October 2012] [20:40:56] aseigo: yep [Wednesday 24 October 2012] [20:41:30] one thing i was thinking about is if we can customize files more for different kinds of content.. [Wednesday 24 October 2012] [20:41:35] Can I inject automatic switch between landscape/portrait? Or is that out of line? [Wednesday 24 October 2012] [20:41:40] oh, the client needs some work too... http://community.kde.org/Bodega#Active_Client [Wednesday 24 October 2012] [20:41:44] ..or if we need even something different, that's a valid answer as well [Wednesday 24 October 2012] [20:41:46] but yes, it's in decent state now [Wednesday 24 October 2012] [20:42:14] ingwa: someone got that working somewhat with xrandr... the shell needs some adaptation for it indeed though [Wednesday 24 October 2012] [20:42:28] and it really sucks with x.org :/ [Wednesday 24 October 2012] [20:42:35] (in ways we can't fix) [Wednesday 24 October 2012] [20:42:41] How about the apps? Are they affected? [Wednesday 24 October 2012] [20:42:50] ingwa: yep, it works with xrandr and xinput on exopc, started working on a dataengine to react to QOrientationSensor but sensors are broken on exopc currently [Wednesday 24 October 2012] [20:42:55] apps work mostly ok [Wednesday 24 October 2012] [20:43:02] workspace has a couple minor bugs, but should be fixable [Wednesday 24 October 2012] [20:43:07] nice! [Wednesday 24 October 2012] [20:43:08] the apps basically would notice they are taller than wide all the sudden [Wednesday 24 October 2012] [20:43:16] that said, the device capabilities DataEngine can give you that information if it is tied into a working Qt Mobility Sensors pack [Wednesday 24 October 2012] [20:43:32] aseigo: ah, good to know [Wednesday 24 October 2012] [20:43:36] jpwhiting: ah, it was you.. i wasn't sure and didn't want to float your name in case i was wrong ;) [Wednesday 24 October 2012] [20:43:37] and i would prefer that way rather than the common way of the app explicitly drawing itself in a "vertical" mode, even rotated by itself [Wednesday 24 October 2012] [20:43:41] np [Wednesday 24 October 2012] [20:43:48] notmart: agreed [Wednesday 24 October 2012] [20:43:54] notmart: I was fine working with Files as it is with ebooks/PDFs for now. So yes, cutomizing the UI for selecting books might be nice, but I don't think it's high priority [Wednesday 24 October 2012] [20:43:57] Nick Praise- is now known as Praise. [Wednesday 24 October 2012] [20:44:00] notmart: definitely [Wednesday 24 October 2012] [20:44:02] jpwhiting: are you thinking of continuing to work on the rotation? [Wednesday 24 October 2012] [20:44:05] ingwa: with a shiny new kernel I built today sensors work again, but aren't exposed by QtMobility for some reason [Wednesday 24 October 2012] [20:44:12] aseigo: I plan to, yeah [Wednesday 24 October 2012] [20:44:17] jpwhiting: cool [Wednesday 24 October 2012] [20:44:18] jpwhiting: awesomesauce [Wednesday 24 October 2012] [20:44:25] jpwhiting: cool [Wednesday 24 October 2012] [20:44:32] haha, we all agree on that, it seems [Wednesday 24 October 2012] [20:44:34] Cool :) Rotation is something quite important, go jpwhiting :) [Wednesday 24 October 2012] [20:44:34] colomar: notmart: i have a use case that puts that on its head... periodicals [Wednesday 24 October 2012] [20:44:44] colomar: agreed [Wednesday 24 October 2012] [20:45:09] aseigo: sure, passing the device across the tablet [Wednesday 24 October 2012] [20:45:23] tablet [Wednesday 24 October 2012] [20:45:25] aseigo: Could you elaborate a bit? [Wednesday 24 October 2012] [20:45:26] books doesn't particularly work for magazines right now [Wednesday 24 October 2012] [20:45:26] darn, [Wednesday 24 October 2012] [20:45:27] table [Wednesday 24 October 2012] [20:45:31] sebas: lol [Wednesday 24 October 2012] [20:45:47] and we have people who would like to make their periodicals available [Wednesday 24 October 2012] [20:46:01] How easy/hard would it be to support running Android apps? [Wednesday 24 October 2012] [20:46:13] in any case ... i think we should look at it in a given dev cycle chunk .. [Wednesday 24 October 2012] [20:46:15] That would remove the one really big disadvantage with PA: The lack of apps [Wednesday 24 October 2012] [20:46:21] if we decide "it's perfect!" then we lose nothing :) [Wednesday 24 October 2012] [20:46:32] So what are the problems currently? I never read periodicals on a tablet, so I my imagination isn't that good here [Wednesday 24 October 2012] [20:46:47] there may be ways to make it work well [Wednesday 24 October 2012] [20:46:50] ingwa: needs a dalvik layer. it's possible, but we don't have the budget for that work right now [Wednesday 24 October 2012] [20:47:07] colomar: when we get to Books we can work it out :) [Wednesday 24 October 2012] [20:47:08] aseigo: yeah, that makes sense. It's encouraging that it's even possible. [Wednesday 24 October 2012] [20:47:17] like periodicals with their own tag (or other univoque metadata) [Wednesday 24 October 2012] [20:47:20] aseigo: sure [Wednesday 24 October 2012] [20:47:31] then extra qml somewhere [Wednesday 24 October 2012] [20:47:43] colomar: periodicals come in epub3 format which has many new formatting options in css. I don't think okular supports those yet. [Wednesday 24 October 2012] [20:47:50] but whatever, we can think about a good ui afterwards [Wednesday 24 October 2012] [20:47:59] ingwa: that said, android apps on tablets are not nearly as good as android apps on phones.. it's also a problem android faces. and we aren't, and can't, play the apps game. [Wednesday 24 October 2012] [20:48:00] Many of those were actually created for the use in periodicals. [Wednesday 24 October 2012] [20:48:09] ingwa: no, not yet... [Wednesday 24 October 2012] [20:48:21] but it's more than formats.. anyways.. no details! :) [Wednesday 24 October 2012] [20:48:27] yes [Wednesday 24 October 2012] [20:48:37] Okay, next block? [Wednesday 24 October 2012] [20:48:41] is everyone ok with the idea of chunking devel into 2 or 3 week bits and focusing on one thing at a time? [Wednesday 24 October 2012] [20:48:45] eBooks is close to ingwa's heart right now. [Wednesday 24 October 2012] [20:48:46] I'd say we should put a Plasmoids block in [Wednesday 24 October 2012] [20:48:48] with the first 3 week chunk being dedicated for files? [Wednesday 24 October 2012] [20:49:03] I'm ok with that [Wednesday 24 October 2012] [20:49:18] colomar: ah, yes... will add that... (plasmoid strip fits in there well too.. it's a great feature, but it's been bit rotting badly) [Wednesday 24 October 2012] [20:49:29] Does it make sense to have everyone working on files? [Wednesday 24 October 2012] [20:49:33] sebas: notmart aseigo once I'm back from UDS, accounts are going to be my main project, so it will advance fast [Wednesday 24 October 2012] [20:49:44] afiestas: awesome [Wednesday 24 October 2012] [20:49:48] afiestas: awesome sauce [Wednesday 24 October 2012] [20:49:53] :D [Wednesday 24 October 2012] [20:49:56] Part sankey has left this channel ("WeeChat 0.3.7"). [Wednesday 24 October 2012] [20:50:11] ok.. i think that's my time ... continue on without me, of course. . but i'd like to call a general end to the formal meeting :) [Wednesday 24 October 2012] [20:50:23] notes to the mailing list tomorrow? [Wednesday 24 October 2012] [20:50:56] * aseigo waves goodnight to the channel [Wednesday 24 October 2012] [20:50:59] ok [Wednesday 24 October 2012] [20:51:02] see you [Wednesday 24 October 2012] [20:51:28] * sebas waves [Wednesday 24 October 2012] [20:51:37] * notmart waves little hand [Wednesday 24 October 2012] [20:53:08] One thing we also definitely need to work on for PA4 is the browser. If you try to really use it, it shows that the surfing experience currently isn't good (due mostly to technical issues, the UI is okay) [Wednesday 24 October 2012] [20:54:04] sorry for being late in the meeting... [Wednesday 24 October 2012] [20:54:35] eh, that's quite a qtwebkit issue [Wednesday 24 October 2012] [20:55:06] ie i don't think is possible to do much on that [Wednesday 24 October 2012] [20:55:52] hm. But I have a much much fewer problems with e.g. rekonq on the desktop. Doesn't it use qtqebkit as well? [Wednesday 24 October 2012] [20:55:57] hopefully with qt5/webkit2 the result should be a tad less painful [Wednesday 24 October 2012] [20:56:18] yeah, webkit1 in qwidget seems to work a bit better than in qml [Wednesday 24 October 2012] [20:56:27] So for PA4 my first agenda will be to close the TODO's remaining in the calligra active wiki. [Wednesday 24 October 2012] [20:57:02] sujith: Could you post a link to those? [Wednesday 24 October 2012] [20:57:50] colomar: Here is the link http://community.kde.org/Calligra_Active#Planned_features [Wednesday 24 October 2012] [20:57:52] notmart: Oh :( So it's basically sit and wait on that front... I hope users won't get too angry till then. [Wednesday 24 October 2012] [20:58:48] sujith: I'd add "better consistency with Reader" to that, with high priority [Wednesday 24 October 2012] [20:59:19] yeah, i would really love to know what to do to it, but seems to be a bit of architectural problem [Wednesday 24 October 2012] [20:59:26] I'll write you a mail with details on that soon, or discuss it with ou directly [Wednesday 24 October 2012] [20:59:43] i could try with an old branch there should still be somewhere that changed a bit how scrolling was done [Wednesday 24 October 2012] [20:59:53] but didn't seem to do much difference in the end [Wednesday 24 October 2012] [21:01:06] colomar: By "better consistancy with reader", did you meant for doc page(s)? [Wednesday 24 October 2012] [21:01:10] notmart: If there's nothing we can do, than that's that. We should just make sure that we report all our problems to qtwebkit so they can make sure they're fixed with webkit2 [Wednesday 24 October 2012] [21:01:31] sujith: Mainly the .doc files, yes [Wednesday 24 October 2012] [21:02:20] Reader does the page selection thing pretty well and I'd like to keep these interactions consistent across applications (and across readers in particular) [Wednesday 24 October 2012] [21:03:23] Same goes for slide selection in ppts, of course [Wednesday 24 October 2012] [21:03:52] colomar: I hope you are not talking about the page preview selection of the reader (.doc files) [Wednesday 24 October 2012] [21:04:30] colomar: because the left tool bar which shows the page(s) for the .doc files has been removed. We now use page preview. [Wednesday 24 October 2012] [21:05:25] Nick bcooksley-away is now known as bcooksley. [Wednesday 24 October 2012] [21:05:26] By Reader, I mean our ebook/PDF reader, aka Okular Active. Sorry for the confusion ;) [Wednesday 24 October 2012] [21:05:59] sujith: Did you see slangkamp's work on providing pages to okular using calligra? [Wednesday 24 October 2012] [21:06:06] Perhaps that could be used here too... [Wednesday 24 October 2012] [21:06:31] As long as it's a reader only it doesn't add much extra value with a special Calligra UI. [Wednesday 24 October 2012] [21:07:24] ingwa: yes I saw [Wednesday 24 October 2012] [21:08:08] When it has editing features as well, then things become a bit different. [Wednesday 24 October 2012] [21:08:20] sujith: But for presentations you could use Okular Active's page selection component as well [Wednesday 24 October 2012] [21:09:09] See also the HIG for that:http://community.kde.org/Plasma/Active/Development/ActiveHIG/Drawers [Wednesday 24 October 2012] [21:09:24] The overlay drawer would be helpful [Wednesday 24 October 2012] [21:11:15] Ok Noted down. [Wednesday 24 October 2012] [21:12:23] It's pretty similar to your page preview, but using the same component would increase consistency even more [Wednesday 24 October 2012] [21:12:35] Join unormal_ has joined this channel (~fmario@adsl-89-217-77-94.adslplus.ch). [Wednesday 24 October 2012] [21:12:55] colomar: Yah, sounds good :) [Wednesday 24 October 2012] [21:13:12] * sujith takes little extra time to understand things :P [Wednesday 24 October 2012] [21:13:40] Quit [Nef] has left this server (Quit: Konversation terminated!). [Wednesday 24 October 2012] [21:13:47] Join linukiss has joined this channel (~linukiss@189.3.93.130).