**** BEGIN LOGGING AT Thu Mar 31 02:59:58 2011 Mar 31 03:24:56 gabrbedd: It's still listed as Tablet Developer Preview, and described as pre-alpha Mar 31 03:25:07 morning Mar 31 03:25:48 pixelgeek: We're talking about this http://repo.meego.com/MeeGo/builds/trunk/1.1.99.0.20110329.5/images/ Mar 31 06:14:38 ping Mar 31 06:14:51 anyone here on OS X try to use mxr recently? Mar 31 06:15:01 someone did a free text search for 'g++' but without a tree Mar 31 06:15:06 which doesn't lead to anywhere useful Mar 31 06:15:56 aha! Mar 31 06:15:59 * timeless_w7ip finds the bad link Mar 31 06:16:04 * timeless_w7ip thanks culprit Mar 31 06:25:59 fixed and pushed upstream w/ a bundle of other things which we've been staging Mar 31 06:43:18 i am going to buy idea pad for meego development does any one know place from where i can get discount Mar 31 06:43:21 :/ Mar 31 07:04:09 where can i upload the screen shots to.. if i need to share that in this IRC Mar 31 07:04:37 similar to pastebin... Mar 31 07:09:48 bkalinga, imagebin Mar 31 07:12:28 thanks andre_ Mar 31 07:22:01 andre__:do you know any script to upload image to this imagebin site and give the uploaded image link; the way pastebinit works Mar 31 07:22:48 bkalinga, sometimes just copy/paste the link is enough Mar 31 07:26:27 yes, luck. Mar 31 07:36:47 morn pohly :) Mar 31 07:40:33 Stskeeps: morning Mar 31 08:07:57 REMINDER: Nokia N900 DE Common Software (including hardware adaptation) meeting in #meego-meeting now Mar 31 10:16:31 \o MeeGoExperts wazd dneary Mar 31 10:16:40 hi lcuk Mar 31 10:17:13 Hi Everyone + lcuk :-) Mar 31 10:17:31 In fact, hi dneary Mar 31 10:18:15 meego-handset-ia32-qemu-1.1.99.0.20110329.5-runtime showing entire black screen and not moving further... Mar 31 10:18:27 is this a know issue Mar 31 10:18:55 *known Mar 31 10:21:05 bkalinga, /o\ oh cripes that does not sound great. especially since previous weeks qemu sdk images booted and got past there Mar 31 10:22:04 \o timeless_w7ip :P I was greeting people who I noticed entering on the page I saw Mar 31 10:24:52 this is true indeed: RT @mece66 @lcuk Me too, and I really like how fast things seem to be moving right now too! @jukkaeklund #meego #n900de Mar 31 10:25:17 lcuk: qemu-1.1.99.0.20110329.5-runtime i waited for 10minutes and its not moving from the initial black screen..where as qemu-1.1.90.3.20110215.10 showing Homescreen within a minute Mar 31 10:28:19 hmm Mar 31 10:28:46 how do I search bugzille for something numeric (20110329) without it trying to think it is a bug number? Mar 31 10:29:04 nm Mar 31 10:31:47 hi MeeGoExperts Mar 31 10:34:25 luck: Try https://bugs.meego.com/query.cgi Mar 31 10:35:50 for 20110329 there is no entry.. i searched Mar 31 10:37:18 bkalinga, is it showing x11 cursor Mar 31 10:38:12 ie bug 9850 Mar 31 10:38:14 <_MeeeGoBot_> Bug https://bugs.meego.com/show_bug.cgi?id=9850 nor, Undecided, 1.2, zhiyuan.lv, NEED, QEMU boots to X but no desktop UI appears Mar 31 10:38:54 although I do not see how older qemu builds work (as we know, you have been in the UI of earlier builds Mar 31 10:47:13 bkalinga, there are a few qemu related bugs https://bugs.meego.com/buglist.cgi?quicksearch=qemu+black Mar 31 10:47:26 ok Mar 31 10:47:27 but mostly old hence oddity with this Mar 31 10:47:43 pondering if it is a regression, certainly feels that way to me Mar 31 10:48:41 Jaffa, btw not sure you @mwkn this, but active open development of the dialer project is expanding even more: http://lists.meego.com/pipermail/meego-dev/2011-March/482339.html Mar 31 11:55:59 morning Mar 31 12:17:03 Hi all Mar 31 12:17:30 I wanted to know when does Qt call the void QGestureRecognizer::reset(QGesture *) function ? Mar 31 12:17:57 Is it when the QGesture change its status to Cancel, Finished or whatever ? Mar 31 12:22:52 re Mar 31 12:23:12 lcuk: I did indeed catch that, thanks Mar 31 12:23:36 \o Mar 31 12:27:58 too bad intel's components depend on various mtf libs Mar 31 12:28:40 like which? Mar 31 12:28:43 MIM? Mar 31 12:29:02 I mean the qt components stuff Mar 31 12:29:24 what libs does it rely on? Mar 31 12:29:26 they are linking to meegotouch libs directly, and using at least MGconfItem (trivial wrapper) Mar 31 12:29:38 villev: MGConfItem is in Mlite instead Mar 31 12:29:50 and contextkit which is not really in mtf but part of the mtf "family" Mar 31 12:29:55 let's see what else they use Mar 31 12:30:11 (mlite is useful stuff from mtf) Mar 31 12:30:55 contextkit is used in most qt apps too in meego, kinda essential for proper running Mar 31 12:34:38 villev: no Mar 31 12:34:43 they aren't using mtf, they use mlite Mar 31 12:35:03 admittedly, the distinction isn't that big - they both have 'm' in the name.. :) Mar 31 12:35:26 w00t_: but they have CONFIG += meegotouch in the src.pro Mar 31 12:35:41 villev: which package is this? Mar 31 12:35:53 meego-ex-components from git Mar 31 12:37:19 Stskeeps: intel components are using it for status bar indicators Mar 31 12:37:32 contextkit that is Mar 31 12:38:49 villev: I suspect that may be leftover Mar 31 12:39:07 the packaging doesn't refer to meegotouch, and having unused/invalid stuff in CONFIG won't cause errors Mar 31 12:39:21 w00t_: right, so it's probably just cruft Mar 31 12:41:26 I wonder, perhaps the status bar stuff shouldn't be in these components anyway Mar 31 12:42:12 MDeclarativeStatusBar is in public code Mar 31 12:43:35 yeah, early days, i think Mar 31 12:52:50 also the stuff in 'models' seem to be out of place in qt components Mar 31 12:54:41 villev: fwiw, they're very receptive to patches Mar 31 12:54:57 so if you find itches that annoy you, go wild Mar 31 12:55:09 the idea was to be 'api' compatible with qt components, right? Mar 31 12:55:39 hint taken. esp. stuff that makes it easier to build would be good candidates... Mar 31 12:55:50 Stskeeps: that's the mid-long term plan yes Mar 31 12:55:54 as I understand it Mar 31 12:57:37 villev: i'm just trying to increase the hands on it.. many hands, light work ;) Mar 31 12:58:51 got it w00t_ Mar 31 13:52:06 morning Mar 31 14:10:08 Does anyone here know why YAML is a good idea? Mar 31 14:11:36 gabrbedd: it really saves me a lot in general packaging work, personally Mar 31 14:12:15 gabrbedd: i didn't like it too much at first either but it really improves workflow Mar 31 14:12:44 as well as allows anyone to pick up the maintaining of the pacage Mar 31 14:16:25 YAML is the specify thing ? Mar 31 14:16:35 yeah Mar 31 14:16:48 the tool that's not available on fedora 14 ? Mar 31 14:17:01 install from source? Mar 31 14:17:02 that's what i do Mar 31 14:17:03 :P Mar 31 14:17:24 i thought it was a python 2.7 issue Mar 31 14:18:31 Stskeeps: How does it save you work? Mar 31 14:18:33 berndhs: http://repo.meego.com/MeeGo/tools/repos/fedora/14/noarch/ Mar 31 14:18:59 Whenever I use it, I spend half my time fighting YAML to generate a proper .spec file. Mar 31 14:19:31 gabrbedd: as a starter, i don't need to manually sort out patches, i just add them to the Patches: list and they're applied. when i have to package something, it's trivial. new version of sw, trivial, etc.. Mar 31 14:21:54 so this is practical now for any distro using rpm ? Mar 31 14:22:14 Stskeeps: Hmmm... that's true. I've flubbed that one up before. Mar 31 14:22:33 Stskeeps: And the builders are nice (when there is one). Mar 31 14:22:52 in addition to that, it really aids the transformation from being a debian person to doing meego packaging Mar 31 14:22:53 Stskeeps: Does Specify support quilt series fiels? Mar 31 14:22:55 :P Mar 31 14:23:04 gabrbedd: not sure, send a patch? Mar 31 14:23:04 :P Mar 31 14:23:29 Well, I'm a debian person... and so far I've really enjoyed RPM packaging. Mar 31 14:23:35 yeah, me too :P Mar 31 14:23:39 YAML... not so much. Mar 31 14:23:40 :-) Mar 31 14:23:58 yaml is like wine, really good when you've used it a bit Mar 31 14:23:58 :P Mar 31 14:24:42 Stskeeps: Well, after running into trouble, I noticed that a lot of packages have abandonded their YAML files. Mar 31 14:24:50 Stskeeps: So, I figured I would ask. Mar 31 14:25:02 I'll give it another try. :-) Mar 31 14:25:05 gabrbedd: or never had them Mar 31 14:26:04 Yes, but having one and abandoning was confirming my opinion "this is just wasting my time." Mar 31 14:26:07 :-p Mar 31 14:26:22 give it some time :) Mar 31 14:27:00 Yep, I'll give it another shot. Now that I'm dealing with more patch sets... that part makes a lot more sense, now. Mar 31 14:27:25 how reliable is the %{_vendor} macro ? will that stay as "meego" ? Mar 31 14:27:37 it used to be MeeGo until it broke gcc Mar 31 14:27:40 now it's meego Mar 31 14:27:40 :P Mar 31 14:27:52 ... Mar 31 14:27:54 do I want to know that story? Mar 31 14:28:00 no Mar 31 14:28:17 i was thinking that when apple see the light and starts using meego, they will change %{_vendor} Mar 31 14:28:53 * gabrbedd always removes %{_vendor} Mar 31 14:30:02 there are packages that dont use pkgconfig, so I still need to differentiate some names Mar 31 14:31:20 bergie: You're using it in the package name ?? Mar 31 14:31:41 thp about? Mar 31 14:31:41 berndhs: ^^^ (sorry, bergie!) Mar 31 14:32:35 i'm using it as %if X%{_vendor} == Xmeego Mar 31 14:34:31 and if that stays Xmeego its bad enough, but if I have to conside Xhtc and Xchinamobile its worse Mar 31 14:35:51 berndhs: I suspect meego defines a meego version or release macro Mar 31 14:36:29 berndhs: I just had M4 flashbacks.... Mar 31 14:36:59 similar to fedora and suse_version Mar 31 14:37:30 berndhs: Are you using it with YAML... or just SPEC files? (just curious) Mar 31 14:38:37 just spec Mar 31 14:39:07 meego version has some issue, I think it has a dot . in it Mar 31 14:40:01 actually I generate spec files, I write them in pieces Mar 31 14:44:35 Stskeeps: Do you know if there are plans to add window frames in the Tablet UX ? Mar 31 14:44:56 frames around what? Mar 31 14:45:07 they already have a titlebar Mar 31 14:45:23 farmes take up space :) Mar 31 14:45:33 Well, a title bar I guess is what I mean. Mar 31 14:45:43 And no, I don't see title bars. Mar 31 14:46:28 berndhs: Yes, but there is a "frameless window hint" that apps can set to tell the WM "I'll draw my own title bar, thank you." Mar 31 14:46:43 they aren't WM drawn Mar 31 14:46:46 screenshot please? Mar 31 14:46:51 or photo etc Mar 31 14:47:19 gabrbedd: yes but then you have to do your own moving and such. Also the hint is ignored on some systems Mar 31 14:47:34 w00t_: Bug 2953 has a similar screen shot. Mar 31 14:47:37 <_MeeeGoBot_> Bug https://bugs.meego.com/show_bug.cgi?id=2953 nor, High, ---, jesse.barnes, REOP, Many third-party Apps' window (i.e. glxgears, xterm) is not working in the fullscreen mode Mar 31 14:47:54 That bug is now resoved on Handset UX, but not for Tablet UX. Mar 31 14:48:02 gabrbedd: ok, if they're not using the infrastructure, then they won't have frames Mar 31 14:48:24 I don't imagine that will be fixed unless someone works on fixing it Mar 31 14:48:26 w00t_: What infrastructure? Mar 31 14:48:35 QML, etc Mar 31 14:49:21 So all apps need to be QML to be MeeGo? That's BS. Mar 31 14:49:31 that's the supported developer story, yes Mar 31 14:51:56 supported developer story? Mar 31 14:56:27 gabrbedd: my point is, Qt is the framework of choice for meego, so, while anything else not working properly is a valid bug, I don't imagine it will get fixed unless someone actually steps up to get it fixed (either as a hobby, or for whatever reason, investing commercial developers' time into it, presumably because they want things other than Qt) Mar 31 14:56:45 (this isn't really news, or anything - stuff doesn't get fixed unless someone wants to fix it) Mar 31 14:57:26 w00t_: And patches don't get accepted unless someone wants the patch. Mar 31 14:57:50 gabrbedd: I personally have had no problems getting my patches integrated; so I'm not sure what point you're trying to make Mar 31 14:58:14 gabrbedd: are you saying you have fixed this, and your patch hasn't been accepted? if so, I'd like to know more Mar 31 14:58:37 w00t_: The point is "people have strong opinions on this... are there any current plans?" Mar 31 14:58:55 gabrbedd: given that it hasn't been (properly) fixed for ~a year, I'd say no Mar 31 14:59:04 which bug are we speaking about again? Mar 31 14:59:07 w00t_: Also, bugs against the Tablet UX are being rejected still... because they don't apply. Mar 31 14:59:20 gabrbedd: I know, I'm subscribed to the bug to get categories set up Mar 31 14:59:29 w00t_: And the bug is FIXED in the Handet UX. Mar 31 14:59:32 but that is correct -- you can't just file bugs in the wrong place Mar 31 15:00:19 ok, so go look at the handset ux, see how it was fixed, and work on transplanting the fix to tablet, and getting it accepted? Mar 31 15:00:27 w00t_: Anyway, I've got no problems working on an issue like this... but I don't want to spend hours doing work that will be spurned. Mar 31 15:00:42 (it won't be easy, though, there's a lot of complex stuff there) Mar 31 15:01:02 is it the fullscreening of non-mtf apps stuff? Mar 31 15:01:06 Stskeeps: yes Mar 31 15:01:09 w00t_: Right, and if Joe is already wokring on it... all well and good. Mar 31 15:01:19 that broke recently again because of the tablet ux guys breaking mdecorator Mar 31 15:01:22 Stskeeps: No title bars for apps. Mar 31 15:01:34 btw, just adding a titlebar won't really solve it anyway Mar 31 15:01:56 it'll just give you a rather out of place looking titlebar :P Mar 31 15:02:11 (i.e. there's no close button on tablet or anything like that) Mar 31 15:02:20 * gabrbedd is getting aggravated... will go get some coffee. Mar 31 15:04:30 good morning DawnFoster Mar 31 15:04:40 o/ Dawn Mar 31 15:04:40 afternoon Stskeeps Mar 31 15:04:49 hey w00t_ Mar 31 15:05:19 how's #meego's favorite community manager? Mar 31 15:05:20 :P Mar 31 15:06:53 ha! Still waking up *tea in hand* Mar 31 15:23:02 is zypper able to install debuginfo rpms? Mar 31 15:24:47 tomaw: Yes, but you have to add (or enable) the "debug" repos. The package and debug repos are separated. Mar 31 15:25:08 tomaw: Package names have `*-debuginfo` at the end. Mar 31 15:25:30 ah, think I know how to add that repo Mar 31 15:26:57 worked, maybe it's time to make my first wiki contribution Mar 31 15:27:35 * TSCHAKeee hugs everybody... tablet UX source opened!!! :) Mar 31 15:31:05 TSCHAKeee: \o/ Mar 31 15:33:00 if anybody wants to check: http://wiki.meego.com/index.php?title=Zypper&diff=35356&oldid=33450 Mar 31 15:33:50 hey tomeu o/ Mar 31 15:34:17 hi w00t_ :) Mar 31 15:43:36 Robot101: we handle actions at the UI layer calling apps or using dbus calls. I don't know that if it was provided by libfolks I would use it since most of our apps launch at the UI layer for now. Mar 31 15:46:51 mnementh: telepathy has a mechanism for launching requests at the framework layer Mar 31 15:46:59 mnementh: which we'd like to add support for in libfolks Mar 31 15:47:07 other stuff like mailto: and http: we can do with MIME handling Mar 31 15:47:27 what other things need doing from an address book context do you think? Mar 31 15:48:22 drilling down to social networking statuses I guess, but it depends if the social things support per-person views or if they just show up aggregated in the panels Mar 31 15:53:24 Robot101: let me get back to you on that I think qt-folks provides a lot of what I need for online status, i need to plug in recent communication like call history, recent social networking updates, designers are still debating if that would duplicate panels too much Mar 31 15:53:55 cool - well we've got a Qt API for the Telepathy logger which has our IM and VOIP call history Mar 31 15:54:09 but of course, you don't store cellular there, so you need to query two places... boo hoo Mar 31 15:54:49 (I wonder if we could use telepathy-ring just for call and message logging purposes? :P) Mar 31 15:55:57 do we need a libcalllog which reads and aggregates multiple sources of call logs...? :D :( Mar 31 16:09:06 Hi Robot101 Mar 31 16:36:36 sigh Mar 31 16:38:27 lcuk: my sentiments exactly Mar 31 16:38:35 ??? Mar 31 16:38:44 did I miss something? Mar 31 17:06:01 * npm fixes netbook 1.2 network horkage by installing latest connman,bluez,ophono from http://download.meego.com/live/devel:/connectivity:/latest/Trunk/i586/ Mar 31 17:08:24 revels in a wired connection that works Mar 31 17:09:35 this fixes bug 12795 ? Mar 31 17:09:37 <_MeeeGoBot_> Bug https://bugs.meego.com/show_bug.cgi?id=12795 nor, Medium, ---, sameo, ASSI, connman crashes during suspend with manual IP Mar 31 17:11:45 perhaps... it fixes https://bugs.meego.com/show_bug.cgi?id=9007 https://bugs.meego.com/show_bug.cgi?id=13778 https://bugs.meego.com/show_bug.cgi?id=14822 and a whole bunch of others Mar 31 17:11:48 <_MeeeGoBot_> Bug 9007 nor, High, ---, martin.xu, RESO WORKSFORME, [REG]network notification pop up about every 5 minutes Mar 31 17:11:50 <_MeeeGoBot_> Bug 13778 maj, Medium, 1.1.90.4, martin.xu, ASSI, [upstream 0.71 has fixed]connection problems, page allocation failures Mar 31 17:11:52 <_MeeeGoBot_> Bug 14822 nor, Medium, ---, martin.xu, ASSI, [REG]Wifi list will refresh in few minutes Mar 31 17:12:21 i got tired of waiting to see connman 0.71 ... so i google searched for an RPM and found the repo above Mar 31 17:12:44 i'll update the bugs once i see it working stable for a whole day Mar 31 17:12:55 sounds reasonable Mar 31 17:13:21 not that it matters since the bugs are closed as "worksforme" anyways Mar 31 17:13:36 i use https://bugs.meego.com/show_bug.cgi?id=13653 as a workaround Mar 31 17:13:39 <_MeeeGoBot_> Bug 13653 nor, Medium, 1.3, sameo, ASSI, ConnMan applet disconnect does not disconnect ipv6 Mar 31 17:13:46 they're all dups, along with several others. Mar 31 17:14:05 disconnect wired from ipv4, live happily with ipv6 connection Mar 31 17:14:55 yeah i thought about forcing a different pathway thru connman... then i realized i don't have an ipv6 net to connect to Mar 31 17:16:01 https://bugs.meego.com/show_bug.cgi?id=13778#c34 was the key to my searching for connman version >> 0.69 which is buggy and still in all latest meego releases Mar 31 17:16:04 <_MeeeGoBot_> Bug 13778 maj, Medium, 1.1.90.4, martin.xu, ASSI, [upstream 0.71 has fixed]connection problems, page allocation failures Mar 31 17:16:14 anyone can make a local ipvt net, and tunnels are free Mar 31 17:16:35 s/ipvt/ipv6/ Mar 31 17:16:35 berndhs meant: anyone can make a local ipv6 net, and tunnels are free Mar 31 17:16:46 if i wanted to make debugging the network stack of meego more of a distraction than it's already been.... Mar 31 17:16:57 now i can get back to work :-) Mar 31 17:18:12 hmm i can just add that repo Mar 31 17:18:43 that's prolly the easiest fix Mar 31 17:19:02 or you can say "zypper install http://......rpm Mar 31 17:19:33 giving a full path to the desired RPM's from http://download.meego.com/live/devel:/connectivity:/latest/Trunk/i586/ Mar 31 17:19:54 just make sure to satisfy all the dependencies in one "zypper in" Mar 31 17:20:20 its a small repo, so adding it shouldn't confuse things Mar 31 17:20:35 not like adding all kinds of under-development things Mar 31 17:20:42 true Mar 31 17:20:58 and then you should just be able to zypper up Mar 31 17:21:06 that too Mar 31 17:47:25 * CosmoHill stabs MSSQL Mar 31 17:51:03 * lcuk uses MIC2 to play around building an image Mar 31 17:52:10 * CosmoHill goes looking for a SQL programmer Mar 31 17:55:01 oh you cock end, at not point did MSSQL think to go "btw you need to prefix this line with SET" Mar 31 17:55:23 lcuk: So, are your sighs related to mic2? (Mine are...) Mar 31 17:56:23 yeah gabrbedd - i have to dig into ubuntu to solve a bug Mar 31 17:58:19 lcuk: Which bug? I'm on ubuntu, too. Maybe we can work together. :-) Mar 31 17:58:47 I am on 10.04 and it does not have qemu 0.13 backported Mar 31 17:58:48 * gabrbedd was up to 1AM futzing with failed mic2 builds. Mar 31 17:59:19 lcuk: doesn't meego tools provide the qemu we need? Mar 31 17:59:21 using the provided 0.12blah version it crashes after downloading data Mar 31 17:59:30 not in the ubuntu repos Mar 31 17:59:46 because when the tools was being made the versions matched Mar 31 18:00:48 ah, I see. hmmm... Mar 31 18:00:55 I can solve it Mar 31 18:01:08 but it is from an unofficial repository and probably would not be a long term solution Mar 31 18:01:37 the patches should be pushed upstream :) Mar 31 18:01:39 I remember in the 1.0 days having to install a special meego-supplied qemu for 10.04. Mar 31 18:01:45 or backported or wahtever is needed Mar 31 18:02:00 yeah gabrbedd for the other OS's in the tools repo there are rpms Mar 31 18:02:10 but it was a concious decision to not include it Mar 31 18:02:13 gabrbedd: I dealt with this last night - you can install two packages from natty manually to get qemu 0.14 Mar 31 18:02:14 OK. Mar 31 18:02:35 special: Actually, this isn't what I was dealing with last night. Mar 31 18:03:00 I meant that I was having issues with qemu 0.12.x last night Mar 31 18:03:07 special: I was building b0rked images. I think it's because I stopped using `--pkgmgr=yum` Mar 31 18:05:17 use qemu 0.13 Mar 31 18:05:28 and yeah, zypp backend is quirky Mar 31 18:05:49 ...not to mention zypp backend is default Mar 31 18:05:51 :-) Mar 31 18:05:52 Stskeeps, yeah that is the ubuntu problem, in the 10.04 repository it is only 0.12 Mar 31 18:06:05 and in 10.10 Mar 31 18:06:38 special, you have the same problem Mar 31 18:06:42 or had Mar 31 18:06:54 lcuk: had, I installed the natty packages manually and that fixed it. Mar 31 18:24:40 gabrbedd, http://permalink.gmane.org/gmane.comp.file-systems.ceph.devel/1573 Mar 31 18:25:49 hmm that might not be it either Mar 31 18:25:57 special, which repo line is the new one on? :$ Mar 31 18:26:25 lcuk: :-) Mar 31 18:26:36 lcuk: So... are you building MeeGo images from within Qemu? Mar 31 18:26:42 lcuk: I wasn't in the mood to wrestle with repos, I just installed the .debs. Mar 31 18:27:08 lcuk: http://packages.ubuntu.com/natty/qemu-kvm-extras-static http://packages.ubuntu.com/natty/qemu-user-static download links are on there Mar 31 18:28:09 thanks special Mar 31 18:28:24 gabrbedd, just following the image building stuff on the wiki Mar 31 18:31:57 lcuk: Been a while since I read the wiki on that stuff. I've only recently started having troubles with mic Mar 31 18:32:09 lcuk: well... /different/ troubles that is... Mar 31 18:34:40 heh there are always different troubles. Mar 31 18:34:46 hopefully we get past them. Mar 31 18:34:48 :) Mar 31 19:06:38 hi Mar 31 19:07:30 evening Mar 31 19:08:03 finally got my n900 ^ Mar 31 19:08:15 :D Mar 31 19:08:21 :) Mar 31 19:10:01 rustylynch: ping Mar 31 19:10:11 (used) unfortunately it has a small very small chip in the display =/ Mar 31 19:10:29 other than that seems fine Mar 31 19:33:19 android 3.0 has 20 apps. http://mobile.slashdot.org/story/11/03/31/182237/Android-30-Is-Trickling-In-But-Are-the-Apps Mar 31 19:33:56 w00t_: i'm back Mar 31 19:34:34 lcuk: android 3.0 has me beat, I only have 7 apps Mar 31 19:35:31 well with some standardness around meego we could have 1000s per day easily Mar 31 19:35:56 yes probably Mar 31 19:36:02 afterall, definition of an app is something people can click and select to install Mar 31 19:36:32 open qt creator, choose nice picture, write a mesage on it and put it on the app mall Mar 31 19:36:41 or some other ide Mar 31 19:37:09 berndhs, which are your 7 apps? Mar 31 19:37:30 * lcuk could probably muster 10 usable apps and about 100 fragments Mar 31 19:37:39 listed here http://berndhs.users.sourceforge.net/ Mar 31 19:38:15 ahh yes, seeing the page reminds me! Mar 31 19:39:01 the link to sourceforge is 404ing Mar 31 19:39:06 for notably Mar 31 19:39:15 let me check that Mar 31 19:39:38 hm Mar 31 19:39:40 works from here Mar 31 19:39:41 http://notably.sourceforge.net/ Mar 31 19:40:03 worked 2nd time, yeah I see Mar 31 19:40:27 the bottom one is QML, the rest are vanilla Qt Mar 31 19:40:58 berndhs, yeah Mar 31 19:41:05 I see notably is using 4.6.1 Mar 31 19:41:13 I gather that would work on the n8x0 as well as n900? Mar 31 19:41:19 i forget Mar 31 19:41:30 heh Mar 31 19:41:31 i havent tried them on N900s Mar 31 19:41:52 saw the newsreader yesterday I tihnk Mar 31 19:42:02 the build in the OBS, screens layouts are not right for really small displays i'm sure Mar 31 19:43:24 berndhs, yeah, always difficult to make apps fit into small screens Mar 31 19:43:58 it is a tough thing and needs reminding people regularly to get it right. Mar 31 19:44:30 especially if you want to provide context for the main thing you're showing, where do you put it Mar 31 19:44:45 i dont have a device to try it on Mar 31 19:45:08 well with Qt, you can technically run it on your desktop Mar 31 19:45:17 and have added flexibility os sizing the window Mar 31 19:45:31 resolution is very different Mar 31 19:45:35 any app that dynamically handles rsizing should work well Mar 31 19:45:47 as long as it starts <=800*480 Mar 31 19:45:48 ;) Mar 31 19:45:48 so stuff that is 3 inches on a desktop isn't the same Mar 31 19:45:50 rustylynch: so am I now Mar 31 19:46:00 rustylynch: so I have an idea for our booster dilemma Mar 31 19:46:36 rustylynch: we can't use booster, because it relies on -fPIC and loading the app as a shared library, but we're looking at this all wrong.. we can instead use meego-qml-launcher as our 'booster' Mar 31 19:47:02 prelaunch it, listen for a dbus message from the launcher saying 'start this', and when it recieves that signal, do all the necessary legwork Mar 31 19:47:19 (i.e. loading the right QML) Mar 31 19:47:22 yea, that was my original idea Mar 31 19:47:34 uh oh Mar 31 19:47:35 mikeleib: you listiening Mar 31 19:47:40 that makes me think you've already ruled it out :) Mar 31 19:47:42 i know there were complications Mar 31 19:48:58 uh oh :-) Mar 31 19:49:04 well... mikeleib and tripzero they can describe some of the complications.... i just got the ball rolling, but didn't keep up with the details Mar 31 19:49:15 i think he is still out for lunch Mar 31 19:49:30 sure, ok Mar 31 19:49:43 I'll roll it past them, I didn't know it had been thought of Mar 31 19:49:50 * w00t_ -> dinner Mar 31 19:49:58 w00t_, if all processes are inside qml-launcher then everything is in process Mar 31 19:50:37 (basically identical to liqbase or hildon-desktop) and if apps crash it crashes the others Mar 31 19:51:38 when I make the ovi suite full screen, why is it like 10 pixels taller than it should be Mar 31 19:51:38 so do that only if you arereasonably sure apps cannot crash. Mar 31 19:51:42 they aren't all in process, and that isn't what I'm suggesting... Mar 31 19:52:05 but dinner Mar 31 19:52:22 sending dbus message into meego-qml-launcher which does tha parsing and running (faster) would mean they were? Mar 31 20:03:57 w00t_: you'd still have to fork though? Mar 31 20:11:43 one of the issues was how to deal with dbus.... it's not fork safe Mar 31 20:16:53 Stskeeps: no Mar 31 20:17:47 load QML, shutdown dbus, pre-exec new qml-launcher which restarts dbus Mar 31 20:24:55 what isn't fork-safe? Mar 31 20:25:00 the connection? Mar 31 20:25:51 if so, yes, the connection isn't fork-safe Mar 31 20:26:11 no socket activity is. The file descriptor is inherited by the child process and they're tied. Mar 31 20:26:30 [this is why I'd not fork to do it] Mar 31 20:26:41 are you talking about the booster? Mar 31 20:26:47 thiago_home: not exactly Mar 31 20:26:52 a mechanism *like* the booster Mar 31 20:27:18 a fork+dlopen mechanism? Mar 31 20:27:26 booster isn't really useful because the Intel guys want to avoid -fPIC (and the performance hit that comes with it), plus having a booster boosting a shared process seems a bit strange Mar 31 20:27:48 so what I'm suggesting is to prestart that shared process and have some way of saying 'load this application's QML and go about your business' Mar 31 20:27:56 james told me the performance penalty with -fPIC Mar 31 20:28:03 but remember that there are a LOT of libraries loaded Mar 31 20:28:13 the amount of code in libraries dwarfs the amount of code in the application Mar 31 20:28:21 yup Mar 31 20:29:33 thiago_home, so, how do you avoid that situation? In Windows, once a library is loaded afaik the code memory space is available for all apps Mar 31 20:29:45 if you profile an application running, I'll bet that the highest uses are in libraries anyway Mar 31 20:29:47 it just reinstantiates the data portion per binary Mar 31 20:30:01 lcuk: Windows uses prelinked dirty text Mar 31 20:30:32 if the DLL is loaded at its prelinked address, the text pages are identical to the disk, with no PIC register Mar 31 20:30:36 thiago_home, within Qt or all libraries? Mar 31 20:30:47 thiago_home: preloading the launcher app will preload a great proportion of those libraries, though Mar 31 20:30:55 not all of them, admittedly, as some of them are loaded via QML Mar 31 20:30:58 if it needs to be relocated, then the text pages will be dirty Mar 31 20:31:00 but does that have the effect of making app startup time better Mar 31 20:31:02 thiago_home: I'm going to (attempt to) get meego running on my ideapad to profile exactly that, anyway Mar 31 20:31:16 lcuk: I meant in general, for any modern app. I'm not thinking specifically of Qt here. Mar 31 20:31:19 w00t_, sure, but they will be in process Mar 31 20:31:26 ok thiago_home good Mar 31 20:31:38 * w00t_ would still give an arm and a leg for a way to generate FTC traces Mar 31 20:31:42 w00t_, in process *can* work Mar 31 20:31:51 and work well, but the code has to be bulletproof Mar 31 20:31:56 lcuk: which this code is not Mar 31 20:31:57 which is where Qa comes in Mar 31 20:32:01 both in terms of memory leaks and crashes Mar 31 20:32:06