| *** niqt has joined #harmattan | 00:04 | |
| *** piggz has quit IRC | 00:36 | |
| ajalkane | Anyone know if I have to specify some Aegis magic to change Flight mode on N9 from application? Seems not to work at the moment even as the function returns success... | 00:41 |
|---|---|---|
| javispedro | what's the problem? | 00:51 |
| javispedro | either way, you can trigger a "Do you want to get out of Flight mode" dialog by trying to make a connection using the qt bearer/mobility api | 00:51 |
| javispedro | but that's about it | 00:51 |
| ajalkane | javispedro: basically the problem is that even as I call MeeGo::QmDeviceMode.setMode(MeeGo::QmDeviceMode::Flight), the device doesn't go to flight mode | 00:52 |
| ajalkane | Oh, qtMobility API has flight mode setting too... I have missed that. | 00:53 |
| javispedro | according to my sources (aka intuition) that requires mce::DeviceModeControl | 00:53 |
| ajalkane | javispedro: okay thanks! I'm really virgin on everything according aegis. I'll look into how I can use that. | 00:55 |
| *** lardman_ has joined #harmattan | 00:55 | |
| javispedro | ajalkane: the problem is, you can't | 00:55 |
| javispedro | ajalkane: it's on the list http://wiki.meego.com/User:Javispedro/Activities_blocked_by_the_N9_and_N950_security_policy | 00:55 |
| ajalkane | Aww... shite. This is one of those forbidden things? | 00:55 |
| javispedro | yep | 00:56 |
| javispedro | also, my guess was correct: http://harmattan-dev.nokia.com/docs/library/html/qmsystem2/classMeeGo_1_1QmDeviceMode.html (setMode requires mce::DeviceModeControl) | 00:57 |
| ajalkane | fuckity-doo. | 00:57 |
| javispedro | yeah :( | 00:57 |
| ajalkane | I guess I'll dispatch tomorrow a message to Nokia's developer forum about Flight mode. Hopefully there's some Nokia engineers that could give some light to that... | 00:58 |
| *** lardman_ is now known as lardman | 00:58 | |
| *** lardman has joined #harmattan | 00:58 | |
| javispedro | good luck... | 00:58 |
| ajalkane | Kinda annoying to spend the evening coding flight mode stuff to be stabbed in the back by policy framework | 00:59 |
| lardman | ~lart qml | 00:59 |
| * infobot nukes qml with a single large nuke | 00:59 | |
| lardman | argh, bloody delegates, I'm at my wit's end | 01:03 |
| ajalkane | Seems like there's already been some discussion on this http://www.developer.nokia.com/Community/Discussion/showthread.php?229903-Package-x-denied-mce-DeviceModeControl-origin-does-not-allow-it | 01:04 |
| ajalkane | I better go sleeping. It's the best time for it. Being bloody and beaten. Crying in my own vomit and tears. | 01:05 |
| *** ajalkane has quit IRC | 01:05 | |
| javispedro | lardman: pfft, I'm trying to make a ListView where the items have varying heights | 01:06 |
| javispedro | lardman: and among other things ListView.contentHeight property is going crazy, its value changing depending on the selected item (!!!) | 01:06 |
| lardman | javispedro: I was trying for fixed height, but fixed to height of largest content | 01:06 |
| javispedro | (docs say it should be the sum of all the items' heights) | 01:06 |
| lardman | my opinion, is that it's not really very logical | 01:07 |
| javispedro | woah, that sounds even harder | 01:07 |
| lardman | e.g. Text items should scale to their container afaiu, but that doesn't seem to work, but then I can't actually obtain the pixel size of the contents | 01:07 |
| lardman | s/scale/wrap if wrap mode is set | 01:08 |
| javispedro | I found after some trial&error than you can do height: childrenRect.height | 01:08 |
| javispedro | so that the container of the Text element grows in height to contain all the height | 01:08 |
| javispedro | *all the Text element | 01:08 |
| lardman | yeah that doesn't work for me strangely, in portrait I see all of my listview items, but in portrait I see them all overlaid | 01:09 |
| javispedro | otherwise, it won't. It'll look like it works but in truth it's only working because for some reason QML doesn't default to clipping | 01:09 |
| lardman | ah apparently there's a clip property, according to some docs, but doesn't seem to work for me | 01:09 |
| lardman | might be clipping, /me would have to check again | 01:10 |
| lardman | either way, didn't work | 01:10 |
| javispedro | it just clips painting, does nothing regarding positioning | 01:10 |
| javispedro | I suggest you start with the delegate being directly a Text element, see if it works and move from there | 01:11 |
| lardman | tbh it's a pita, I'd prefer a graphical editor and to drag and drop the items as it's not logically consistent, or at least we don't know the rules and their precendence | 01:11 |
| lardman | my delegate is an image with a column of 3 text elements that move either below or to the side depending on the screen orientation | 01:12 |
| lardman | s/with/plus | 01:12 |
| *** Sazpaimon__ has joined #harmattan | 01:12 | |
| javispedro | both ListViews and Models are a pita | 01:13 |
| javispedro | positioning has always been IMHO Qt's achillees heel | 01:13 |
| javispedro | and models are just a pain to write. Can't even believe I had to write a function that converts a QList<Something*> to a QList<QObject*> | 01:14 |
| javispedro | (by casting each element) | 01:14 |
| *** Sazpaimon_ has quit IRC | 01:15 | |
| w00t | assuming you're talking about exposing C++ objects to QML, you wouldn't, if you'd used qmlRegisterUncreatableType | 01:18 |
| lardman | javispedro: quite | 01:18 |
| lardman | I've wasted so many hours arsing about with models and bloody qml (especially delegates) | 01:19 |
| javispedro | w00t: for a model, you can't, only recognized type is QList<QObject*>, period. | 01:19 |
| lardman | w00t: never even heard of that | 01:19 |
| javispedro | well, and QAbstractItemModel, but that'd probably be even more code | 01:19 |
| * lardman has QAbstractItemModel | 01:20 | |
| lardman | not much code, works out of the box pretty much, quite good, though the docs are very sparse | 01:20 |
| w00t | I'm pretty sure that there's a better way, but I don't use QList models, so I'm not certain | 01:20 |
| w00t | take a look at http://qt.gitorious.org/qt-labs/qml-object-model/blobs/master/qobjectlistmodel.h anyway | 01:20 |
| w00t | it's easy to use, and for most cases will probably make your life pretty easy | 01:21 |
| * lardman 's problem is QML delegates | 01:21 | |
| w00t | lardman: what about them? | 01:21 |
| *** zk8 has quit IRC | 01:21 | |
| lardman | I can't get them to work | 01:21 |
| w00t | that's.. not really the most helpful problem description ;) | 01:22 |
| lardman | http://pastebin.com/Pwr2HMSv | 01:22 |
| lardman | one gets that way after messing about with them for weeks on end after work ;) | 01:23 |
| w00t | ok, what am I looking at? | 01:23 |
| lardman | a ListView delegate | 01:24 |
| lardman | but yeah, I understand that it;s not that easy to understand a piece of code on its won | 01:25 |
| lardman | own | 01:25 |
| w00t | well, what about it, for instance | 01:25 |
| lardman | perhaps /me should accept that implementing UIs isn't his thing, and that he should just hard code the sizes and crack on with the interesting stuff (i.e. the functionality).... | 01:28 |
| w00t | I can't really just conjure a solution out of thin air from "it doesn't work" (with no symptoms or explanation as to what doesn't work) and a 153 line pastebin | 01:29 |
| lardman | of course not, I wasn't actually asking you too, just letting off steam | 01:29 |
| w00t | point being, describe the problem | 01:30 |
| lardman | there are too many deps to sort out to put up a repo quickly | 01:30 |
| w00t | I might be able to help | 01:30 |
| w00t | when I have a problem like that, I usually start cutting things down until I find something that does work | 01:30 |
| lardman | yeah I know, but I need the functionality | 01:31 |
| w00t | that either makes me find what the problem was, or gets me to a small example that demonstrates something being broken | 01:31 |
| javispedro | lardman: I built a separate project with a handwritten ListModel to test this stuff | 01:31 |
| lardman | but, yes I should cut it down to nothing, etc. | 01:31 |
| javispedro | lardman: so that it could be run from qmlviewer | 01:31 |
| lardman | that sounds like lots of extra effort and let's be realistic this is scraping the barrel as it is | 01:32 |
| lardman | babies and all | 01:32 |
| javispedro | hey, I've been with this since early past month =) | 01:32 |
| lardman | w00t: the problem, in a nut shell is this: in landscape I only see one item, in portrait I see all of them | 01:32 |
| lardman | aside from that my mouse click area thingie doens't work, but I;ve not looked into that yet | 01:33 |
| lardman | javispedro: I've been doing pretty similar stuff since Sept, the main code was almost finalised then ;) | 01:33 |
| w00t | lardman: you set no height/width on the root Item which probably won't help that | 01:33 |
| w00t | (why have you wrapped it in a Component, by the way?) | 01:33 |
| lardman | following the examples... | 01:34 |
| w00t | which? | 01:34 |
| w00t | it is necessary in some situations, but not when you're already putting it into a seperate file (and thus, component) | 01:34 |
| lardman | I'd have to search, first started this code back in Sept | 01:34 |
| lardman | ok fine, noted | 01:34 |
| javispedro | all of the examples define the delegate in a component | 01:35 |
| lardman | even if I don;t set the height/width of the root item, why does it work in one orientation and not the other | 01:35 |
| lardman | and in fact it doesn't overwrite itself, it only shows one item, I've no idea why that would happen | 01:35 |
| javispedro | it's not easy to follow that the Component { } is not necessary when it's the only element in a file or it's defined inline as delegate: property value | 01:35 |
| w00t | magical wizards? who knows | 01:35 |
| w00t | anyway, first thing I'd try would be to set a height/width | 01:36 |
| lardman | would be nice for the qml runtime to be able to print out its tree of size relations as it calculates them so I can work out what's messing things up | 01:36 |
| *** niqt has quit IRC | 01:36 | |
| lardman | perhaps, but it doesn't have, afaict, a logical reason | 01:37 |
| lardman | I do actually set the height/width in the states at the bottom btw | 01:37 |
| lardman | or do you mean of the Component? | 01:38 |
| * lardman has no clue which "items" default to 0 height/width and which change to fit their content | 01:39 | |
| w00t | none change to fit their content | 01:39 |
| javispedro | except Column/Row ;) | 01:40 |
| w00t | unless you explicitly set height: childrenRect.height; width: childrenRect.width | 01:40 |
| lardman | Text {} ? | 01:40 |
| w00t | Text has no content, it does have paintedHeight and paintedWIdth | 01:40 |
| lardman | ? | 01:40 |
| w00t | unless you put something in it | 01:40 |
| lardman | well exactly, but if I put sometihng in a Rectangle does that change size? | 01:40 |
| w00t | no | 01:40 |
| lardman | but Text does....? | 01:40 |
| w00t | nope | 01:41 |
| * javispedro dies | 01:41 | |
| lardman | but all qml items can render beyond their width/height I've noticed? | 01:41 |
| w00t | yes | 01:41 |
| lardman | very odd | 01:41 |
| lardman | and confusing | 01:41 |
| w00t | set clip: true if you don't want it to happen | 01:42 |
| javispedro | w00t: do you mean Text does not change its height even if you set text to something? or that it does not change its height if you add other children to it? | 01:42 |
| w00t | it's off by default for performance reasons | 01:42 |
| lardman | sure, I read about that yesterday, but that's pretty late in the day | 01:42 |
| lardman | +1 to javispedro's question | 01:42 |
| w00t | javispedro: Text will have an implicitHeight/mplicitWidth, but unless you *set* a height/width, it won't have one of its own | 01:42 |
| lardman | what does Text actually do then? | 01:42 |
| lardman | oh dear | 01:43 |
| lardman | so how can one obtain the height/width that will contain some text? | 01:43 |
| w00t | paintedWidth/paintedHeight | 01:43 |
| w00t | (will be the exact area it needs) | 01:43 |
| w00t | implicitHeight/Width are probably also useful | 01:43 |
| w00t | this, btw, is actually how clipping works | 01:43 |
| w00t | sorry | 01:44 |
| w00t | eliding, not clipping | 01:44 |
| w00t | you set the size smaller than the size it requires | 01:44 |
| lardman | so how come I can attach e.g. 3 Text items top to bottom and they don't overwrite one another? I've not set heights, so presumably that should be the outcome...? | 01:45 |
| w00t | implicitHeight | 01:45 |
| lardman | but that makes no sense | 01:46 |
| *** etrunko has quit IRC | 01:46 | |
| w00t | http://doc.qt.nokia.com/4.7-snapshot/qml-item.html#implicitWidth-prop | 01:46 |
| w00t | it does | 01:46 |
| w00t | implicitHeight/Width is the size that text *needs* to take | 01:46 |
| lardman | I understand that it works | 01:46 |
| w00t | size is the size you *set* it to take | 01:46 |
| w00t | a Rectangle does not *need* to take any room | 01:46 |
| lardman | just that if you don;t alter the height/width of the item, what use are they? | 01:46 |
| w00t | if it has an implicitHeight/Width, it'll use those (as it says) | 01:46 |
| lardman | so why does Rectangle not support that functionality> | 01:47 |
| lardman | ? | 01:47 |
| w00t | because a rectangle, as I just explained, doesn't need to take any room | 01:47 |
| lardman | we're talking about content though aren't we? | 01:47 |
| w00t | not necessarily | 01:47 |
| w00t | rectangles don't need to have contents | 01:47 |
| lardman | surely a rectangle containing a Text should have an implicit* set to contain whatever the Text needs | 01:48 |
| w00t | and also, calculating the size of the contents of something is not exactly straightforward | 01:48 |
| lardman | by rights.... | 01:48 |
| lardman | I understand it's not straight forward, but nor are the docs regarding these things imo | 01:48 |
| * javispedro also thinks it is counterintuitive | 01:48 | |
| javispedro | but hey, I already found that happens to Rectangle a few days ago.. | 01:49 |
| javispedro | (thus width: childrenRect.width) | 01:49 |
| w00t | javispedro: yeah, I mentioned that a (lot) earlier | 01:49 |
| lardman | yeah I knew it happened to a Rectangle but jnust assumed that Text was different and set its height/width, but I was wrong about the property names | 01:49 |
| w00t | lardman: it makes sense when you stop thinking of QML as being a system for layouts, because that's not what it does | 01:50 |
| lardman | that isn't afair explained anywhere in the docs | 01:50 |
| lardman | w00t: what does it do then? | 01:50 |
| javispedro | render particles! | 01:50 |
| w00t | it gives you a scene, and lets you position and anchor stuff wherever you want | 01:50 |
| lardman | my particles are Text shaped! | 01:50 |
| w00t | this is both a strength and a weakness | 01:50 |
| * lardman sees scene and it washes over him | 01:51 | |
| w00t | a strength, because it's a lot faster, proper layouts are hard (and slow), and a weakness because it means a bit more thinking about how to do things | 01:51 |
| w00t | anyway, a good rule of thumb is that you should probably always set anchors/sizes unless you're very sure you don't need them | 01:51 |
| w00t | anchors.fill: parent is your friend for simple cases | 01:51 |
| * lardman hopes that thjs stuff will be added to the intro | 01:52 | |
| lardman | w00t: that breaks delegates though I think | 01:52 |
| lardman | and setting fixed height/width is nasty | 01:52 |
| w00t | btw, a useful debugging trick related to size problems: Item { id: iWantToFindTheSizeOfThis; Rectangle { color: "red"; anchors.fill: parent; } } | 01:52 |
| w00t | it doesn't | 01:52 |
| w00t | and note that I said anchors/sizes, implying, one or the other | 01:53 |
| lardman | sure | 01:53 |
| w00t | (or both, if you need them) | 01:53 |
| lardman | I did try to use anchors only, but no no avail | 01:53 |
| lardman | coloured squares work to an extent, but when they are simply not shown it's hard to know why | 01:53 |
| lardman | again some output from the qml runtime would be a Godsend | 01:53 |
| w00t | if it's not shown, either something is obscuring it, you forgot a height/width, your positioning of that item is wrong, or you set it invisible somewhere | 01:54 |
| javispedro | I had a button that when pressed would output x,y,w,h | 01:54 |
| javispedro | via console.log | 01:54 |
| w00t | (opacity: 0, visible: false, one or the other or both) | 01:54 |
| javispedro | pain to write too, but helped. | 01:54 |
| lardman | w00t: indeed, but it would be nice to know which in particular | 01:54 |
| lardman | if qml simply output a list of objects and their x,y,width,height that would help | 01:55 |
| lardman | plus anchors | 01:55 |
| *** vladest has quit IRC | 01:55 | |
| w00t | javispedro: Rectangle { MouseArea { anchors.fill: parent; onClicked: console.log("x y w h" parent.x ...); } } was hard? | 01:55 |
| w00t | (substitute parent with any id, if you want to get information about something else for a more general button) | 01:56 |
| w00t | one thing I do wish that was easily possible is getting an item by id, though that's tricky due to multiple instantiations of the same component | 01:56 |
| w00t | (though I suppose getting a list of those would be ok) | 01:56 |
| lardman | w00t: can be done from C++ afaiu | 01:57 |
| w00t | it sort of can, using objectName | 01:57 |
| w00t | unless there's a trick I don't know about | 01:57 |
| * lardman throws away his delegate and goes back to a single coloured square, great.... | 01:57 | |
| lardman | on second thoughts, enough for the week | 01:58 |
| w00t | lardman: I honestly don't mind if you pester me with questions, btw | 01:58 |
| javispedro | w00t: for a few objects | 01:58 |
| *** vladest has joined #harmattan | 01:58 | |
| lardman | thanks w00t, not you, just taking so long to debug this qml is rather disheartening | 01:58 |
| w00t | I'm not always around, but usually am, european day/nighttimes, so if you need to bounce an idea off someone, just ask | 01:58 |
| lardman | cheers | 01:58 |
| javispedro | including some that were spawned by the listview element | 01:58 |
| w00t | javispedro: yes | 01:58 |
| lardman | anyway really must his the sack, night w00t javispedro | 01:59 |
| lardman | s/his/hit | 01:59 |
| javispedro | cya lardman | 01:59 |
| lardman | see what i mean | 01:59 |
| lardman | ;) | 01:59 |
| w00t | that's why I mentioned the getElementById thing - it would be handy for those sort of cases | 01:59 |
| *** lardman has quit IRC | 01:59 | |
| javispedro | ah, indeed. | 01:59 |
| *** crevetor has quit IRC | 01:59 | |
| w00t | TextArea { ... }, enter text, press button, get info | 01:59 |
| javispedro | though I would prefer a pointy-clicky "object-tree" explorer | 02:00 |
| javispedro | (lots of no-id objects) | 02:00 |
| javispedro | hmm, that could probably be done.. | 02:01 |
| w00t | i'd think so | 02:01 |
| w00t | introspection has to be possible for stuff like accessibility | 02:02 |
| *** javispedro has quit IRC | 02:21 | |
| *** crevetor has joined #harmattan | 02:24 | |
| *** Natunen has quit IRC | 02:28 | |
| *** diggy128 has joined #harmattan | 02:29 | |
| *** diggy has quit IRC | 02:31 | |
| *** diggy128 is now known as diggy | 02:31 | |
| *** JackaLX has quit IRC | 02:32 | |
| *** NIN101 has quit IRC | 02:40 | |
| *** hardaker has quit IRC | 02:51 | |
| *** JackaLX has joined #harmattan | 02:53 | |
| *** M4rtinK has quit IRC | 02:56 | |
| *** arcean has quit IRC | 03:12 | |
| *** daniel_o has quit IRC | 03:16 | |
| *** b3ll has joined #harmattan | 04:42 | |
| *** vladest has quit IRC | 05:05 | |
| *** vladest has joined #harmattan | 05:09 | |
| *** adlan has joined #harmattan | 05:43 | |
| *** Natunen has joined #harmattan | 06:09 | |
| Sazpaimon__ | I wiuld kill for nokia maps public transit to support philadelphia | 06:21 |
| Sazpaimon__ | then I can completely abandon google | 06:21 |
| *** hardaker has joined #harmattan | 06:34 | |
| *** pinheiro__ has joined #harmattan | 06:39 | |
| *** pinheiro_ has quit IRC | 06:43 | |
| *** vladest has quit IRC | 06:50 | |
| *** b3ll has quit IRC | 06:50 | |
| *** vladest has joined #harmattan | 06:52 | |
| *** b3ll has joined #harmattan | 07:05 | |
| *** lfranchi has quit IRC | 07:25 | |
| *** DocScrutinizer has quit IRC | 07:29 | |
| *** DocScrutinizer has joined #harmattan | 07:29 | |
| *** lfranchi has joined #harmattan | 07:30 | |
| *** b3ll has quit IRC | 07:38 | |
| *** b3ll has joined #harmattan | 07:42 | |
| *** sebas has quit IRC | 07:50 | |
| *** sebas has joined #harmattan | 07:50 | |
| *** b3ll has quit IRC | 07:58 | |
| *** npm_ has joined #harmattan | 08:14 | |
| *** npm has quit IRC | 08:15 | |
| hiemanshu | is there a GUI for apt-* yet? | 08:15 |
| *** hardaker has quit IRC | 08:43 | |
| *** crevetor has quit IRC | 08:43 | |
| *** crazypenguin_ has joined #harmattan | 08:52 | |
| tomma | apt-*? | 09:05 |
| luke-jr | the package manager for Maemo/Harmattan (since Nokia never noticed Debian moved to aptitude) | 09:06 |
| Stskeeps | they moved to aptitude? | 09:07 |
| Stskeeps | :P | 09:07 |
| *** psycho_oreos has quit IRC | 09:18 | |
| hiemanshu | is there a way to run an app as root? If not, then a GUI Package Manager would be useless | 09:20 |
| *** zk8 has joined #harmattan | 09:47 | |
| Hq` | en nyt laskenu mitään, mutta yleisesti jännitteen lasku vastuksen avulla ei välttämättä oo voittava idea | 09:54 |
| Hq` | wrong channel, sorry | 09:54 |
| *** veli has quit IRC | 10:01 | |
| *** trx has quit IRC | 10:09 | |
| *** trx has joined #harmattan | 10:13 | |
| *** CapsuleNZ has joined #harmattan | 10:14 | |
| *** piggz has joined #harmattan | 10:21 | |
| *** veli has joined #harmattan | 10:29 | |
| *** piggz has quit IRC | 10:46 | |
| *** adlan has quit IRC | 10:54 | |
| *** NGNUton-BC has joined #harmattan | 10:55 | |
| *** GNUton-BNC has quit IRC | 10:55 | |
| *** NGNUton-BC is now known as GNUton-BNC | 10:55 | |
| *** rcg has joined #harmattan | 11:26 | |
| *** mosty has joined #harmattan | 11:27 | |
| mosty | i have a regular qt app (using qwidgets) that i have setup to rotate based on the orientation of the device. but how can i tell the system to change the orientation of swipe gestures and system notifications? | 11:30 |
| djszapi | mosty: may I ask why you use qwidgets ? | 11:35 |
| mosty | djszapi, because i have a complex qt app that runs on multiple operating systems and devices and i don't want to rewrite it in qml | 11:36 |
| *** crazypenguin_ has quit IRC | 11:36 | |
| djszapi | mosty: you do not need to rewrite it, only the ui part. You will suffer much more with qwidgets "porting" than rewriting the ui in qml | 11:36 |
| mosty | djszapi, if i did that then i'd have a whole new codebase to maintain, which is not fun for a single device | 11:37 |
| djszapi | huh ? | 11:37 |
| djszapi | we have 60.000 LOC, and the pages for Harmattan (2-3 simple qml files). | 11:38 |
| djszapi | this is what qml was designed for. | 11:38 |
| djszapi | Moreover, pure qml usage is probably also wrong. Using platform components make more sense. Either you will have an ugly Ui or you will be reinventing the look'n feel which is already there. | 11:39 |
| mosty | i have everything working with qwidgets already though | 11:39 |
| djszapi | so what is the problem ? | 11:40 |
| djszapi | if everything works. | 11:40 |
| hiemanshu | djszapi: its not so easy to just port an app from QWidgets to QML | 11:40 |
| hiemanshu | Writing QML, easy, Writing QWidgets, easy, Porting not so easy | 11:41 |
| djszapi | hiemanshu: it is super-easy. That is the purpose of the qml, that is what it was designed for | 11:41 |
| mosty | the problem is that the system notifications and swipe gestures think the phone is always in landscape when my app is running | 11:41 |
| djszapi | I wrote a Ui in one day for a codebase with 60.000 LOC | 11:41 |
| hiemanshu | djszapi: I have tried, 4 times, so no | 11:41 |
| djszapi | pebkac, it works for many people. | 11:41 |
| djszapi | and porting qwidgets are probably ugly as many people said (also on this channel) since I do not think anybody has alone enough man power to reinvent the native look'n feel. | 11:43 |
| djszapi | is* | 11:43 |
| ZogG_laptop | и luke-jr ping | 11:44 |
| djszapi | the only problem I can think of if the codebase is really bad-structured, and the qwidget ui is not separated from the functionality. That is a really bad code infrastructur ethough. | 11:44 |
| djszapi | such a code should be refactored anyway, and not built more on top of such a codebase. | 11:45 |
| djszapi | mosty: so do you have well-separated ui and functionality ? | 11:48 |
| djszapi | (backend-frontend) | 11:48 |
| mosty | djszapi, not perfect but not terrible. but i want to use a single codebase on multiple platforms | 11:49 |
| djszapi | you can do that by using qml | 11:50 |
| mosty | we don't use qml on any other platform, this would just add maintenance cost | 11:50 |
| djszapi | qml is writable by even designers | 11:51 |
| djszapi | that is the whole point. It is much simpler to maintain than porting qwidget. | 11:51 |
| mosty | well when another platform appears that requires qml i will consider it | 11:51 |
| mosty | but i can't justify it for one device | 11:52 |
| *** piggz has joined #harmattan | 11:52 | |
| djszapi | that is a very bad approach imho, sorry for saying this | 11:52 |
| djszapi | nowadays, mobile platforms keep rocking, I am pretty sure people have just more and more device targetted. | 11:52 |
| mosty | i can't afford to use nokia's framework-of-the-month | 11:52 |
| djszapi | and if you now refuse qml, you will do at the next and so on. | 11:52 |
| mosty | i've seen it change too often | 11:53 |
| djszapi | so ... ? | 11:53 |
| djszapi | They do not break what is already available, they just improve it further.. | 11:54 |
| mosty | that wasn't the case for maemo (gtk?), nor for meegotouch | 11:55 |
| djszapi | qml is not gtk, not even meegotouch | 11:56 |
| mosty | no, those were the previous recommended frameworks, now there is another incompatible one | 11:56 |
| djszapi | huh ? | 11:56 |
| djszapi | I wrote a qml app one and half a year ago, and runs today too. | 11:57 |
| djszapi | The only problem is that, there are still missing bits, but as far as I know, they do not break the functionality with a QtQucik bump. | 11:57 |
| djszapi | it is not an alpha release. | 11:57 |
| mosty | my qwidget app runs today too | 11:58 |
| mosty | i'm just lacking one piece of platform integration code | 11:58 |
| djszapi | yeah, I see... basic functionality missing :) | 11:58 |
| djszapi | out of the curiosity, can you show me your code ? I got really interested in this.. | 11:59 |
| mosty | i can't unfortunately, i don't own the code | 12:00 |
| djszapi | so it is not even open source ? | 12:00 |
| mosty | correct | 12:00 |
| mosty | otherwise i would be glad for the pair of eyes | 12:00 |
| *** psycho_oreos has joined #harmattan | 12:02 | |
| ZogG_laptop | anyone here with gentoo? | 12:06 |
| ZogG_laptop | luke-jr: now i understand why i didn't want to use portage version, coz it doesn't have qmake for harmattan, i ended with only desktop target, no harmattan target | 12:11 |
| *** Venemo_N950 has joined #harmattan | 12:12 | |
| Venemo_N950 | hi | 12:13 |
| djszapi | w00t: ping | 12:13 |
| djszapi | w00t: have you read the pim feedback from Friedrich on the qt project mailing list (development) ? If yes, what is your opinion ? | 12:14 |
| Venemo_N950 | hey djszapi, how was your birthday party? :) | 12:20 |
| djszapi | Venemo_N950: super-crazy :P | 12:21 |
| Venemo_N950 | djszapi :) | 12:22 |
| djszapi | Venemo_N950: thanks for asking, so how are you ? | 12:24 |
| Venemo_N950 | djszapi, well as always. :) had 10+ hours of sleep, so I'm in a good mood | 12:25 |
| *** BluesLee has joined #harmattan | 12:32 | |
| *** BluesLee has quit IRC | 12:34 | |
| *** GNUton-BNC has quit IRC | 12:50 | |
| *** GNUton-BNC has joined #harmattan | 12:50 | |
| *** mardy has joined #harmattan | 12:55 | |
| *** CapsuleNZ has quit IRC | 12:58 | |
| *** CapsuleNZ_ has joined #harmattan | 12:59 | |
| *** arcean has joined #harmattan | 13:00 | |
| *** M4rtinK has joined #harmattan | 13:01 | |
| *** zz_gri is now known as gri | 13:04 | |
| *** psycho_oreos has quit IRC | 13:06 | |
| *** mosty has quit IRC | 13:14 | |
| *** diggy is now known as diggy128 | 13:17 | |
| *** diggy128 has quit IRC | 13:17 | |
| *** CapsuleNZ_ is now known as CapsuleNZ | 13:18 | |
| *** zk8 has quit IRC | 13:26 | |
| *** gri is now known as zz_gri | 13:31 | |
| arcean | ~malf | 13:40 |
| infobot | rumour has it, malf is http://maemo.cloud-7.de/Aegis-kills-device.jpg | 13:40 |
| *** diggy has joined #harmattan | 13:51 | |
| Hq` | miks tarvii laskee sille jännitettä, onks se jenkkisähkölle tehty? | 13:51 |
| Hq` | perkele | 13:51 |
| Hq` | wrong channel again | 13:52 |
| *** leinir has joined #harmattan | 14:06 | |
| *** Arkenoi has quit IRC | 14:18 | |
| ZogG_laptop | lol | 14:24 |
| *** diggy has quit IRC | 14:24 | |
| *** diggy has joined #harmattan | 14:26 | |
| *** smoku has joined #harmattan | 14:29 | |
| *** Venemo has joined #harmattan | 14:31 | |
| *** Venemo has joined #harmattan | 14:31 | |
| ZogG_laptop | Venemo: hooray | 14:32 |
| ZogG_laptop | you are from camputer? | 14:32 |
| ZogG_laptop | right? | 14:32 |
| ZogG_laptop | =\ | 14:35 |
| Venemo | hey ZogG_laptop, yes, but from windows unfortunately :) | 14:40 |
| Venemo | ZogG_laptop, but good news for you: irc-chatter is evolving once again :) I'm hacking on it on my ideapad | 14:40 |
| *** Venemo_N950 has quit IRC | 14:41 | |
| *** pawky has quit IRC | 14:42 | |
| djszapi | arcean: aegis is not the culsprit there. | 14:53 |
| djszapi | arcean: the problem was that the system services and dsme guys handled error retvals like MALF in each case in that mood. | 14:55 |
| djszapi | so after they fixed their error handling, that was not a problem anymore. Aegis was not changed in that regard, and actually it happened with many softwares underneath, not just aegis. The monitor software behaved weirdly, that is. So titling of that jpg that way is a bit misleady. | 14:56 |
| ZogG_laptop | Venemo: can you upload it to testing thru OBS to apss.formeego.org | 14:57 |
| djszapi | Venemo: long life to irc-chatter ! ;) | 14:58 |
| ZogG_laptop | king is dead, long live king | 14:58 |
| Venemo | ZogG_laptop, I didn't say it compiles yet | 14:59 |
| Venemo | :P | 14:59 |
| djszapi | everybody is welcome for the New Year's kick off Qt and Linux Mobile meetup in Helsinki: http://www.meetup.com/QtEverywhere/Helsinki-FI/533242/ | 15:00 |
| ZogG_laptop | damn i still can't set env for programming | 15:04 |
| ZogG_laptop | as i listned to luke-jr and got QtSDK from ebuild and not one from site and i have no target for mobile | 15:04 |
| ZogG_laptop | getting scratchbox now | 15:06 |
| ZogG_laptop | maybe it would help | 15:06 |
| ZogG_laptop | ok i need help | 15:15 |
| djszapi | w00t: you got an answer for your regexp post: http://blogs.kde.org/node/4510 | 15:16 |
| djszapi | ZogG_laptop: what help ? | 15:16 |
| ZogG_laptop | MADDE is what makes me able to build straight on n9 right? | 15:16 |
| RST38h | No | 15:16 |
| djszapi | no, you build on the host | 15:17 |
| Venemo | just uninstall that crap, download the Qt SDK and be done with it | 15:17 |
| ZogG_laptop | so i need scratchbox and MADDE? | 15:17 |
| djszapi | no | 15:17 |
| ZogG_laptop | Venemo: i want to install it properly on my system | 15:17 |
| Venemo | ZogG_laptop, http://qt.nokia.com/downloads | 15:18 |
| ZogG_laptop | so what is /home/zogg/QtSDK/Madde/targets/harmattan_10.2011.34-1/bin/qmake | 15:18 |
| RST38h | ZogG: You need one of two | 15:18 |
| ZogG_laptop | RST38h: one of two what? | 15:18 |
| RST38h | ZogG: Either 1) QtSDK with Harmattan target or 2) Scratchbox on Ubuntu Linux (aka Platform SDK) | 15:19 |
| ZogG_laptop | RST38h: no ubuntu here | 15:19 |
| RST38h | What OS do you have? | 15:19 |
| ZogG_laptop | RST38h: i have scratchbox ebuilds, madde ebuild and sdk ebuild | 15:19 |
| RST38h | what is ebuild? | 15:19 |
| ZogG_laptop | RST38h: funtoo/gentoo | 15:19 |
| RST38h | Ok, good luck on your adventure. | 15:20 |
| RST38h | A Windows machine with QtSDK would probably be your best bet though | 15:20 |
| Venemo | ZogG_laptop, just download the latest Qt SDK installer from my link and install the damn thing. in the installer, select the harmattan target, and it'll set it up for you | 15:21 |
| RST38h | Given that QtSDK works on Windows (my copy failed to package and run apps) | 15:21 |
| ZogG_laptop | i use funtoo as primary os, Venemo and i did that, it half worked as it has problems with libpng | 15:27 |
| ZogG_laptop | have 1.5 it wants 1.2 for no reason | 15:27 |
| Venemo | wtf, that's weird | 15:29 |
| ZogG_laptop | yeah i can have both, but still that's why i preffer ebuilds and OS env to install SQK and not from site to keep system clean and not to make zoo | 15:30 |
| Termana | damn it. My N950 won't let me access the storage and it's telling me it's in use. I tried umount -l /home/user/MyDocs but even with develsh/devel-su it says the operation is not permitted. Anyone have any other ideas? | 15:38 |
| jonni | go to running apps grid and close all applications and replug the cable | 15:40 |
| Termana | They are all closed | 15:40 |
| *** berndhs has joined #harmattan | 15:41 | |
| jonni | and reboot and try again :) | 15:41 |
| Termana | I tried rebooting. I guess I'll give it another shot | 15:42 |
| jonni | usually its just some application keeping filehandle open which prevents the umount, or having terminal open in that dir. | 15:43 |
| jonni | (which also prevents the umount) | 15:43 |
| Termana | Nope still not working after another reboot | 15:44 |
| *** piggz has quit IRC | 15:44 | |
| jonni | maybe youve corrupted filesystem then, or have some application in autostarting on boot | 15:46 |
| dm8tbr | try checking with 'fuser' or 'lsof' | 15:46 |
| *** rcg has quit IRC | 15:47 | |
| *** NIN101 has joined #harmattan | 15:50 | |
| Termana | dm8tbr, nothing | 15:54 |
| Termana | seems like its mounted read only on the device as well | 15:55 |
| *** JackaLX has quit IRC | 16:00 | |
| *** b3ll has joined #harmattan | 16:26 | |
| *** Venemo has quit IRC | 16:28 | |
| *** JackaLX has joined #harmattan | 16:28 | |
| DocScrutinizer | too bad, sounds like corrupted then | 16:30 |
| DocScrutinizer | check syslog | 16:31 |
| *** Venemo has joined #harmattan | 16:32 | |
| ZogG_laptop | jonni: is madde fro dev.nokia.com/madde is the smae used for harmattan and maemo? | 16:34 |
| RST38h | Madde is no longer a standalone project afaik | 16:34 |
| RST38h | It is now part of QtSDK | 16:35 |
| ZogG_laptop | damn | 16:35 |
| RST38h | Haven't you got a windows pc there? | 16:38 |
| ZogG_laptop | i don't like and don't want use it | 16:43 |
| *** trx has quit IRC | 16:46 | |
| *** trx has joined #harmattan | 16:53 | |
| *** zarlino has joined #harmattan | 16:56 | |
| *** pawky has joined #harmattan | 16:58 | |
| pawky | do mameo apps run on a N9 out of the box? | 16:58 |
| *** oscoder has joined #harmattan | 16:59 | |
| qronic | no | 17:00 |
| *** psycho_oreos has joined #harmattan | 17:00 | |
| pawky | not even non gui apps? | 17:01 |
| *** rcg has joined #harmattan | 17:01 | |
| Venemo | pawky, depends. | 17:10 |
| pawky | Venemo: on? | 17:10 |
| TSCHAKeee | why do you ask? | 17:11 |
| Venemo | pawky, softfp/hardfp differences | 17:11 |
| pawky | Just to check how much broader the software sphere is. There are quite some cool apps for the N900 that would be nice to use on th N9 as well :-) | 17:12 |
| *** crevetor has joined #harmattan | 17:13 | |
| Venemo | pawky, they will need a recompile to say the least | 17:13 |
| pawky | Venemo: I thought it was all Strong ARM .... | 17:14 |
| pawky | Venemo: ohh.. ah, yea libraries.. | 17:14 |
| Venemo | pawky, Fremantle doesn't use hardfp or thumb | 17:14 |
| Venemo | pawky, also, library versions differ | 17:14 |
| DocScrutinizer | I am using quite some N900 cmdline apps on N950/HARM | 17:14 |
| Venemo | pawky, also, Harmattan doesn't have GTK | 17:14 |
| Venemo | pawky, also, QWidget-based apps won't work well on Harmattan. | 17:14 |
| Venemo | pawky, etc. | 17:14 |
| pawky | DocScruitinizer:Is there any list anywhere? | 17:14 |
| DocScrutinizer | nope | 17:15 |
| DocScrutinizer | just dpkg -i and test ;-D | 17:15 |
| DocScrutinizer | either WFM or SEGV | 17:15 |
| DocScrutinizer | cmdline only programs seem rather safe, as long as not using any floating point arithmetic | 17:16 |
| pawky | DocScruitinizer: well... My initial idea is to keep this phone a tiny bit more clean. I will never forget how long it took me for getting rid of some CIFS half install on my N900... | 17:16 |
| pawky | DocScruitinizer: Floating point? Why would that differ? | 17:17 |
| DocScrutinizer | hmm, reflash is always like all new clothes :-D | 17:17 |
| DocScrutinizer | hardfp on HARM | 17:17 |
| pawky | DocScruitinizer: That's for wimps ;-) | 17:17 |
| DocScrutinizer | ? | 17:17 |
| Venemo | pawky, as I said, Harmattan uses hardfp and Fremantle uses softfp... | 17:18 |
| pawky | Venemo:ahh.. ok. | 17:19 |
| *** hardaker has joined #harmattan | 17:19 | |
| pawky | Venemo: Kind of cool... but I believe floating point is for wimps as well ;-) 64bit integers is the way to go :-D | 17:19 |
| DocScrutinizer | the problem more is like "which programs may run under paegis?" | 17:19 |
| Venemo | pawky, I won't comment on that. | 17:20 |
| DocScrutinizer | pawky: I'll comment on that: YES! | 17:20 |
| pawky | Venemo: I presume the idea for hardware floating point is for games right? | 17:20 |
| DocScrutinizer | gfx, yeah | 17:20 |
| DocScrutinizer | mainly | 17:20 |
| DocScrutinizer | while I can't see why gfx *needs* fp | 17:21 |
| pawky | DocScruitinizer: has anybody ever compared the fp, to some integer arithmetic, speed wise on these phones? :-) | 17:21 |
| Venemo | pawky, both softfp and hardfp use hardware for floating point calculations. the difference is that softfp passes floats in the integer registers, while hardfp passes them in the floating point registers. | 17:21 |
| Stskeeps | aw, i've taught venemo something | 17:21 |
| Stskeeps | :P | 17:21 |
| Venemo | yes! I learned this (and a lot more) from Stskeeps :) | 17:22 |
| pawky | Venemo: I see.. sounds like passing in integer registers would be the faster option, or? | 17:22 |
| Venemo | pawky, please think before you say something like that. | 17:23 |
| Stskeeps | no, it's faster in floating point, as you don't need to do float -> integer -> float conversions always when calling a function | 17:23 |
| DocScrutinizer | pawky: luke-jr started something over at #maemo, which seemed to indicate fixed is 3 times faster than fp, but the whole test was abit incoherent on results on a second closer glance, as his results supposedly meant for hardfp were identical to mine for softfp | 17:25 |
| DocScrutinizer | wow, should've read the next two posts - thanks Stskeeps & Venemo | 17:26 |
| DocScrutinizer | this explains a lot then | 17:27 |
| DocScrutinizer | still it's silly to use fp when you could get away with (long) int | 17:27 |
| DocScrutinizer | in my book gfx is such a case | 17:27 |
| Stskeeps | yeah... i guess it matters with GLES | 17:28 |
| *** crevetor has left #harmattan | 17:29 | |
| DocScrutinizer | so except for CAD and orrery/stellarium I don't see a use for fp in real life programming | 17:30 |
| DocScrutinizer | even for CAD it's questionable | 17:30 |
| *** pawky has quit IRC | 17:31 | |
| DocScrutinizer | for audio it's frequently utterly insane | 17:31 |
| *** adlan has joined #harmattan | 17:32 | |
| *** adlan has quit IRC | 17:32 | |
| *** b3ll has quit IRC | 17:55 | |
| *** pinheiro__ has quit IRC | 18:03 | |
| *** MikaT has quit IRC | 18:15 | |
| *** MikaT has joined #harmattan | 18:17 | |
| *** smoku has left #harmattan | 18:19 | |
| *** b3ll has joined #harmattan | 18:21 | |
| *** zarlino has quit IRC | 18:23 | |
| *** b3ll has quit IRC | 18:39 | |
| *** oscoder has quit IRC | 18:41 | |
| ZogG_laptop | luke-jr: PM me as soon as you here | 18:42 |
| ZogG_laptop | jonni: ping | 18:43 |
| *** NIN101 has quit IRC | 18:57 | |
| *** NIN101 has joined #harmattan | 18:57 | |
| *** crevetor has joined #harmattan | 19:20 | |
| *** zarlino has joined #harmattan | 19:20 | |
| djszapi | interesting, Ovi store asks this: "To install this software | 19:21 |
| djszapi | you need to install its dependencies, do you want to do it?" | 19:21 |
| Venemo | yeah, what's new? | 19:22 |
| djszapi | so Ovi store supports dependency installation ? | 19:22 |
| djszapi | it is really news to me. | 19:22 |
| Venemo | I dunno | 19:24 |
| *** deimos has joined #harmattan | 19:29 | |
| djszapi | Venemo: will ask internally :) | 19:30 |
| Venemo | djszapi, ok, thanks :) | 19:31 |
| ZogG_laptop | any nokia guys here? | 19:39 |
| djszapi | external matters ? | 19:42 |
| M4rtinK | Well, would be really nice if it did - we are in the year 2011 already after all :) | 19:42 |
| djszapi | it is not about "2011". | 19:43 |
| *** deimos_ has joined #harmattan | 19:43 | |
| djszapi | it is about policy, and they do with their app stores whatever you want. If they would like you to use their platform only, they can. Moreover, you are still not restricted enough since they do not check external libraries you can put into the package. | 19:44 |
| djszapi | X-Fade ping | 19:44 |
| djszapi | will apps4meego support depdendencies ? | 19:44 |
| djszapi | dependencies* | 19:45 |
| dm8tbr | IIRC: currently no, but it might in the future | 19:45 |
| M4rtinK | well, yeah everything is much simpler if you don't have to do dependency handling, but quite a few use-cases suffer | 19:45 |
| djszapi | M4rtinK: it is not about simple, you do not see the point | 19:46 |
| *** deimos has quit IRC | 19:46 | |
| djszapi | it is about policy, as in: "I want you to use my platform, and not something else". | 19:46 |
| M4rtinK | I don't see this dependency handling conflicting with store policies. | 19:47 |
| djszapi | ok, put it clear: the problem is not that for Intel, Nokia, Google, and others: they are weak to implement dependency handling as such big companies... It is clean policy, it is not about technical implementation and "2011". | 19:48 |
| *** diggy128 has joined #harmattan | 19:49 | |
| *** diggy has quit IRC | 19:49 | |
| M4rtinK | I would say their policy is set like this also because it's simpler like this. :) | 19:49 |
| djszapi | no it is not, again it is clean policy. I discussed it more times with one of the Ovi maitainers... | 19:50 |
| djszapi | but I am sure you know better for sure ;-) | 19:50 |
| djszapi | if they would like to have simple things, they would not have cared about security analyzes at all. | 19:51 |
| djszapi | since that is the most difficult part. | 19:51 |
| M4rtinK | lets say that If I built a store/repository myself, it would support dependency handling :) | 19:51 |
| djszapi | but it is only in your dreams, and you did get nothing ready yet, right ? | 19:52 |
| M4rtinK | of course not :) | 19:52 |
| M4rtinK | but I still don't like artificial limitations like this :) | 19:53 |
| djszapi | why ? Why is it artificial to have an opinion what they would like to do with their money ? | 19:53 |
| djszapi | and it is of course not any Nokia specific, butevery bigger companies agreed upon on this philosophy. | 19:54 |
| djszapi | so yes sure ... you can beat the dead horse ;) | 19:54 |
| M4rtinK | it is their decision of course - but in my opinion it is still wrong :) | 19:55 |
| djszapi | oh well | 19:55 |
| M4rtinK | or should every package that uses a not-in-SDK lib_example have it statically compiled inside ? | 19:56 |
| *** deimos__ has joined #harmattan | 19:56 | |
| djszapi | that is the worst idea every :) | 19:57 |
| M4rtinK | this wasted bandwith & storage space | 19:57 |
| djszapi | I wonder actually how you got to the conclusion :) Sounds really really nasty ;) | 19:57 |
| M4rtinK | and when there is a security issue discovered for the library - good luck finding & fixing all such packages :) | 19:57 |
| djszapi | I do not know how it is relevant | 19:58 |
| djszapi | your idea brings actually the security issues in... | 19:59 |
| M4rtinK | well with the current policy, static compilation fro libraries that are not in the SDK is the only possibility, isn't it ? | 19:59 |
| djszapi | say, someone submitted a library to the store, and I need to depend on that since I cannot bring my trusted version...I am kinda in a situation where I /must/ trust. | 20:00 |
| djszapi | no, it is not. That is a sad conlusion | 20:00 |
| djszapi | and a bit of insane | 20:00 |
| *** deimos_ has quit IRC | 20:00 | |
| djszapi | please read my example I wrapped up for the KDE Harmattan wiki page :) | 20:01 |
| djszapi | actually static linking is the only option which does not play for Qt. | 20:02 |
| djszapi | with gpl version for instance | 20:02 |
| djszapi | but at least static linking has all the troubles and that jazz: you basically need to maintain shared and static builds which is well ... insane for kde since even one build type maintenance is a huge time sink :) | 20:03 |
| luke-jr | static linking is bad. | 20:03 |
| M4rtinK | well, of course | 20:03 |
| M4rtinK | so say you want to submit Konqueror to a "Store" without dependency handling | 20:04 |
| M4rtinK | how would you do it ? | 20:04 |
| djszapi | 20:01 < djszapi> please read my example I wrapped up for the KDE Harmattan wiki page :) | 20:05 |
| M4rtinK | http://userbase.kde.org/Kontact_Touch/Harmattan ? or something else | 20:06 |
| djszapi | though, it is not any Harmattan specific; we basically do the same for qt apps on Android. | 20:06 |
| djszapi | M4rtinK: something else.. | 20:06 |
| djszapi | you need to read a bit, but you can learn a lot of things: starting with this note of mine: http://community.kde.org/KDE_Mobile/Sprints/November2011-Planning#Open_Questions | 20:06 |
| djszapi | http://community.kde.org/KDE_Mobile/Harmattan -> and here is a practical example of doing this. | 20:06 |
| *** tomyri has quit IRC | 20:08 | |
| djszapi | mmm, I have not documented everything, but if you do not like the rpath change trickery, you can just set the ld library path fine inside the desktop file. | 20:08 |
| *** tomyri has joined #harmattan | 20:09 | |
| *** Venemo_N950 has joined #harmattan | 20:09 | |
| Venemo | hm | 20:09 |
| *** Venemo_N950 has quit IRC | 20:10 | |
| M4rtinK | so if I understand this correctly, you take the compiled library files and just add them to your package | 20:12 |
| djszapi | yeah | 20:12 |
| *** zarlino has quit IRC | 20:12 | |
| M4rtinK | well, yeah also a possibility | 20:12 |
| djszapi | more or less. It is not that simple, though.. | 20:13 |
| djszapi | putting the so into your package is not enough | 20:13 |
| M4rtinK | but I still see a few issues - binary compatibility, wasted bandwidth/space/RAM due to not sharing with other packages | 20:14 |
| djszapi | of course you share | 20:14 |
| djszapi | that is the whole point | 20:14 |
| djszapi | binary compatibility, eh ? | 20:15 |
| djszapi | actually we are the only with along with debian solved it | 20:15 |
| djszapi | and nobody else cares about this on the market. | 20:15 |
| M4rtinK | you are including the binaries aren't you ? | 20:15 |
| djszapi | I do include of course. | 20:15 |
| djszapi | since it does not work the other way around anyway :) | 20:16 |
| *** Venemo_N950 has joined #harmattan | 20:17 | |
| M4rtinK | well if it were a source package it would fail after being rebuild for other architectures than the included binary libraries support IMO | 20:17 |
| *** Venemo_N950 has quit IRC | 20:17 | |
| M4rtinK | but I'm not sure what one actually submits to OVI & similar | 20:17 |
| djszapi | you do not of course put Mac binary into a linux package. | 20:18 |
| djszapi | but I fail to see the problem to be honest, since you need to rebuild your app on different platforms anyway... | 20:18 |
| *** Venemo_N950 has joined #harmattan | 20:19 | |
| M4rtinK | hmm | 20:20 |
| djszapi | M4rtinK: sorry, sauna time, be back in a bit :) | 20:20 |
| luke-jr | M4rtinK: "Stores" suck | 20:20 |
| *** Venemo_N950 has quit IRC | 20:20 | |
| M4rtinK | yeah, that's the point :) | 20:21 |
| luke-jr | that being said, a Linux store, properly designed, would be interesting | 20:21 |
| luke-jr | ie, developers submit source, and the store builds binaries for every participating distro, etc | 20:22 |
| luke-jr | dynamically linked using the pkg manager of course | 20:22 |
| M4rtinK | "depends: lib_example >= 1.0" feels righ, including binary files in a package by hand feels wrong :) | 20:22 |
| M4rtinK | yep, exactly | 20:22 |
| leinir | luke-jr: welcome to Project Bretzn ;) | 20:23 |
| M4rtinK | well, OBS makes this possible (provided it works as expected) :) | 20:23 |
| berndhs | OBS seems to work fine at the opensuse site, it builds stuff for several distros | 20:24 |
| M4rtinK | IMO apps formeego does this already (without dependency handling so far) | 20:24 |
| *** lmoura has quit IRC | 20:26 | |
| *** TSCHAKeee has quit IRC | 20:27 | |
| *** TSCHAKeee has joined #harmattan | 20:27 | |
| *** BluesLee has joined #harmattan | 20:37 | |
| jonni | ZogG_laptop: pong | 20:38 |
| *** ajalkane has joined #harmattan | 20:40 | |
| *** lmoura has joined #harmattan | 20:40 | |
| *** BluesLee has quit IRC | 20:44 | |
| *** diggy128 is now known as diggy | 20:46 | |
| *** diggy is now known as diggy128 | 20:50 | |
| *** Venemo_N950 has joined #harmattan | 20:52 | |
| ajalkane | anyone know what are these kinds of lines in restok.conf: | 20:52 |
| ajalkane | Recover: +mce::DeviceModeControl +mce::TKLockControl | 20:52 |
| *** Venemo_N950 has quit IRC | 20:52 | |
| ajalkane | What's the Recover thingy? I know we can do Requests in aegis.manifest. | 20:52 |
| *** Venemo_N950 has joined #harmattan | 20:53 | |
| *** Venemo_N950 has quit IRC | 20:54 | |
| djszapi | M4rtinK: apps4meego is not a business thing. | 20:54 |
| M4rtinK | and ? :) | 20:55 |
| djszapi | and do not confuse apps4meego which for the community, and it would not make sense to not support as much as possible. | 20:56 |
| djszapi | M4rtinK: Nokia will obviously not support non-Nokia things for fun and more resource. | 20:56 |
| djszapi | so does Intel not do (Tizen/Qt). I think you fail to respect what Nokia would like to do with their business. I personally understand the point of these app stores, and also a community store. They are way not exclusive, and should never ever be. | 20:57 |
| berndhs | Nokia supports many things that are not directly generating revenue | 20:57 |
| djszapi | M4rtinK: back to the technical discussion, I see no problem in putting the shared libraries into the same package. | 20:58 |
| M4rtinK | well yeah | 20:58 |
| djszapi | M4rtinK: as I said, you are lucky Nokia does not filter that out you did put something into the package ;-) | 20:59 |
| M4rtinK | I just want to distribute my applications to users in a clean and efficient way without being hampered by missing libraries :) | 21:00 |
| M4rtinK | and I find the current system a bit lacking in this regard, that's all :) | 21:00 |
| djszapi | because you only see community, and you do not see business, that is all :) | 21:00 |
| M4rtinK | well I'm a community developer after all :) | 21:01 |
| djszapi | you do not need to refuse business decisions to be a community developer. | 21:02 |
| *** zarlino has joined #harmattan | 21:03 | |
| *** zk8 has joined #harmattan | 21:04 | |
| *** berndhs has left #harmattan | 21:04 | |
| *** Venemo_N950 has joined #harmattan | 21:04 | |
| djszapi | M4rtinK: anyway, you are not even obligated to use ovi, and you can help to X-Fade with apps4meego, right ? Start it ;-) | 21:04 |
| *** Venemo_N950 has quit IRC | 21:04 | |
| djszapi | M4rtinK: I am personally happy long time contributors were interested in discussing the publish story at the KDE Harmattan sprint, and we got something done. | 21:05 |
| djszapi | M4rtinK: you were mentioning the binary compatibility, well that is what I solved in Harmattan and probably no other mobile solutions care about this. :) Adopted the strict debian policies about it in KDE at least. | 21:07 |
| *** zarlino has quit IRC | 21:07 | |
| djszapi | M4rtinK: it took me 13 KB to solve, so no biggie :) | 21:08 |
| *** Venemo has quit IRC | 21:09 | |
| M4rtinK | I plan to submit my applications to every viable repository available - which means both OVI and AFM | 21:11 |
| *** NIN101 has quit IRC | 21:11 | |
| djszapi | so do as documented ? :) | 21:12 |
| M4rtinK | the first in the queue is Mieru, which currently has no external dependencies | 21:12 |
| *** jluisn has joined #harmattan | 21:12 | |
| M4rtinK | but modRana would be more difficult, even when using a Qt based GUI | 21:12 |
| djszapi | why is it any difficult ? | 21:12 |
| M4rtinK | espeak for TTS and probably also Monav for offline routing in the future | 21:13 |
| *** ZogG_laptop has quit IRC | 21:13 | |
| djszapi | what is the issue precisely ? | 21:14 |
| ajalkane | Is AFM a similar source to OVI for installed applications, so that all same credentials are allowed as from OVI installed apps? | 21:14 |
| djszapi | nope | 21:14 |
| djszapi | that is why I do not prefer apps4meego for Harmattan, if the package can be put into Oiv. | 21:15 |
| djszapi | Ovi* | 21:15 |
| M4rtinK | djszapi: how to get get both modRana and espeak on the device without a need for user intervention :) | 21:15 |
| djszapi | also, the qa is even pre-mature | 21:15 |
| djszapi | (not offense, it is just that many people put a lot of energy into Ovi) | 21:15 |
| ajalkane | That's unfortunate, but understandable. But then AFM must differentiate in some other ways, like supporting dependencies. | 21:16 |
| ajalkane | Although having automatic update notifications for AFM applications is great. | 21:16 |
| djszapi | ajalkane: agree, if it does not support dependency it is a fail imho | 21:16 |
| M4rtinK | or not having 10^7 RSS readers :) | 21:16 |
| M4rtinK | so much for QA :) | 21:16 |
| ajalkane | martink: haha, yeah that's a plus too | 21:17 |
| djszapi | M4rtinK: could you please rephrase (or put more information into) your question ? | 21:17 |
| M4rtinK | djszapi: lets have two applications - modRana, a navigation system written in Python, thus architecture independent | 21:18 |
| M4rtinK | and Espeak a an utility/library written in C and accessed through CLI | 21:18 |
| M4rtinK | modRana uses espeak for TTS through CLI | 21:19 |
| djszapi | yeah .. ? | 21:20 |
| * djszapi is listening | 21:20 | |
| djszapi | ajalkane: Yes, it is understandable. Otherwise Ovi would abandon itself. | 21:21 |
| M4rtinK | the user installs modRana | 21:21 |
| M4rtinK | espeak should be installed together with it automatically | 21:21 |
| M4rtinK | somehow :) | 21:21 |
| djszapi | tried my documentation ? | 21:22 |
| M4rtinK | well, yeah I can just showe the espeak binary to my no longer architecture independent package and it would work | 21:22 |
| djszapi | I mean what is the point ? I thought we discussed this business decision question. :) | 21:23 |
| M4rtinK | well it seems wrong to bundle software like this | 21:24 |
| djszapi | so everything technical coming from that is effected. As I mentioned a few above there are cons too, not just pros. | 21:24 |
| djszapi | no, it is the only way. | 21:24 |
| ajalkane | djszapi: No, I mean, it's understandable because if Nokia wants to guarantee some sort of security it has to retain control of some features for services it has direct control over. | 21:24 |
| djszapi | ajalkane: it has nothing to do with security things and origin source | 21:25 |
| ajalkane | djszapi: then what is it about? | 21:25 |
| djszapi | about that, Nokia would not like to abandon itself. | 21:25 |
| M4rtinK | ?? | 21:26 |
| ajalkane | djszapi: You mean they crippled other sources just to ensure people use Ovi and not other sources | 21:26 |
| djszapi | security issues would be up to any externals stores and their qa processes. | 21:26 |
| ajalkane | ? | 21:26 |
| djszapi | ajalkane: ok, put it this way to understand a bit better: | 21:26 |
| djszapi | Nokia could have said, okay, apps4meego gets all the credentials ever, but the only trusted origin is Nokia (as in words) | 21:27 |
| djszapi | that way would mean to that they would lose many users, right ? | 21:27 |
| djszapi | probably not all, but vast majority. | 21:27 |
| djszapi | and this is obviously not the purpose of Nokia =) | 21:28 |
| ajalkane | They might lose some, but vast majority use just the default apps bundled in the phone | 21:28 |
| djszapi | the statistics show differently though. | 21:28 |
| djszapi | I think Kate Alhola was actually commenting on the opposite. | 21:28 |
| djszapi | and I agree: I mean it makes sense to not abandon yourself to any other app stores. | 21:29 |
| djszapi | if you invested the lot of money and energy. | 21:29 |
| ajalkane | Sure... I can understand that kind of reasons too. But I bet anyone higher up in the hierarchy would claim security reasons instead of what you said. Regardless what is the truth. | 21:30 |
| djszapi | no, not at all | 21:30 |
| ajalkane | Then do you know why we can't develop apps using for example mce::DeviceModeControl credential, it cripples even OVI store compared to competitors? | 21:31 |
| djszapi | that is a simple question, the manager layer refused our proposal for no real reasons. | 21:32 |
| djszapi | "complete open mode proposal" that we did first. | 21:32 |
| djszapi | we had the technical solution almost in place | 21:32 |
| djszapi | so that is not because of security, that is because of some insane reasons | 21:32 |
| *** zarlino has joined #harmattan | 21:33 | |
| ajalkane | I don't even need complete open mode, although I'd want some, but for my application I would want to control Flight Mode and Battery Saving states. Competiting products manage that. | 21:33 |
| ajalkane | But Nokia's management is famously crazy so I'm not surprised them shooting down your proposal without any reason | 21:34 |
| djszapi | if you check out my comment on those topics. I said many times here and also on the public bug tracker, I agree. | 21:34 |
| M4rtinK | yay for artificial limitations | 21:34 |
| ajalkane | djszapi: do you think we, as application developers, can anyway help in that by bringing concrete examples to bug tracker or somewhere to change their minds at least about some of the security policy decisions? | 21:35 |
| djszapi | ajalkane: however I know the history of that decision making (ovi source policies that is). Unfortunately, I cannot talk about it | 21:35 |
| djszapi | ajalkane: do not think community caught it, but internals were dumb to get it in time... :P | 21:35 |
| *** Natunen has quit IRC | 21:35 | |
| djszapi | the things community caught, we proposed like 1.5-2 years ago :) | 21:36 |
| ajalkane | djszapi: No I have no doubt many internals caught these things. But I also know how management thinks. | 21:36 |
| djszapi | they think it is not a developer device | 21:37 |
| ajalkane | They don't take things seriously until a concrete application can not be realized because their idiotic decisions. Well, many times they won't even then. | 21:37 |
| *** NIN101 has joined #harmattan | 21:37 | |
| djszapi | and if they think, they will think this even if we vomit blood | 21:37 |
| ajalkane | A consumer device is supposed to run applications done by developers. It's that whole ecosystem bullshit. | 21:37 |
| djszapi | ajalkane: we told the use cases, they do not wanna realize certain things | 21:37 |
| djszapi | but it is not Nokia specific, it is the habit of being an owner, decision maker and so on in most cases. :) | 21:38 |
| ajalkane | Okay... well... management is dumb. | 21:38 |
| ajalkane | If only every manager was a former coder, things might be better :) | 21:39 |
| *** ZogG_laptop has joined #harmattan | 21:39 | |
| ajalkane | Or not... but probably not a whole lot worse either. | 21:39 |
| ajalkane | djszapi: you might know... if I've put some requests for credentials that aegis/security framework will not grant for my application, where would I see some log information about that? | 21:42 |
| ajalkane | For example: I get credential for "Cellular" and "UID::user" and "UID::users", but not for ""mce::DeviceModeControl" | 21:43 |
| ajalkane | Didn't see anything in syslog | 21:43 |
| *** Almehdi has joined #harmattan | 21:44 | |
| *** natunen has joined #harmattan | 21:46 | |
| *** Almehdi has quit IRC | 21:49 | |
| *** Almehdi_ has joined #harmattan | 21:49 | |
| Almehdi_ | Does anyone have a clue why this doesn't work: devel-su -c 'gst-launch-0.10 ximagesrc ! videoscale ! videorate ! video/x-raw-rgb,framerate=5/1,width=854,height=480 ! ffmpegcolorspace ! dsph264enc ! avimux ! filesink location=/home/user/MyDocs/Movies/test.avi' as user.. if i golog in as devel-su it works. | 21:49 |
| Almehdi_ | 's/golog/log in/' | 21:50 |
| ajalkane | Wow... I got a Recover line into restok.conf with a request entry in aegis manifest | 21:54 |
| Almehdi_ | It writes the file but it is 0.00 seconds.. and if run it gives a green screen. If i do devel-su and the run the same command it plays as it should | 21:54 |
| ajalkane | So I guess those are something to do with credentials that you can get if you install from OVI, I guess | 21:54 |
| ajalkane | But still, mce::DeviceModeControl seems to be totally inaccessible. | 21:55 |
| ajalkane | So I guess I'll give up on that for now. | 21:55 |
| Almehdi_ | shouldnt "devel-su -c" and log in as devel-su be the same? | 21:55 |
| ajalkane | Gone to Windows, cya later! | 21:57 |
| *** ajalkane has quit IRC | 21:57 | |
| *** piggz has joined #harmattan | 22:08 | |
| *** diggy128 is now known as diggy | 22:16 | |
| *** zarlino has quit IRC | 22:30 | |
| DocScrutinizer | Almehdi_: I could only speculate that devel-su doesn't correctly pass parameters to shell when it execve()s shell | 22:39 |
| DocScrutinizer | after all devel-su is no alias for shell, but rather a wrapper around it | 22:40 |
| *** jluisn has quit IRC | 22:40 | |
| *** diggy is now known as diggy128 | 22:41 | |
| * DocScrutinizer could even figure that's again an intentional thing rather than a bug, due to some weird security considerations that normal minds can't even think of | 22:42 | |
| *** berndhs has joined #harmattan | 22:50 | |
| *** berndhs has left #harmattan | 22:53 | |
| *** diggy128 is now known as diggy | 23:01 | |
| Almehdi_ | DocScrutinizer: You wouldn't happen to know a way around it? | 23:01 |
| Almehdi_ | I am using it in a shell script | 23:01 |
| Almehdi_ | I have tried using adding "exec" and 'aegis-exec -a "GRP::video" -a "GRP::pulse-access"' without success | 23:03 |
| DocScrutinizer | sorry, no idea | 23:05 |
| DocScrutinizer | the correct way was to build a pkg that requests for the needed permissions via aegis manifest, so you wouldn't need devel-su | 23:07 |
| DocScrutinizer | I think, that is. If I'm not mistaken how all this aegis stuff is meant to work | 23:08 |
| *** Venemo has joined #harmattan | 23:08 | |
| Almehdi_ | hmm... | 23:09 |
| Almehdi_ | to bad... | 23:10 |
| Almehdi_ | Was hoping to give my n9tweak users a way to record their desktop | 23:11 |
| *** diggy has quit IRC | 23:15 | |
| *** Arkenoi has joined #harmattan | 23:16 | |
| *** diggy has joined #harmattan | 23:16 | |
| *** diggy is now known as diggy128 | 23:18 | |
| *** rlinfati has joined #harmattan | 23:18 | |
| rlinfati | Hi, where i can download the firmware for n950 beta 1 or beta 2 ? | 23:18 |
| Venemo | rlinfati, why would you want to download old firmware? | 23:19 |
| DocScrutinizer | o way to downgrade anyway | 23:19 |
| DocScrutinizer | no* | 23:19 |
| *** Venemo_N950 has joined #harmattan | 23:21 | |
| rlinfati | only i wanna extract some files | 23:21 |
| Venemo_N950 | rlinfati, I don't think you can do that. | 23:22 |
| *** Venemo_N950 has quit IRC | 23:23 | |
| *** Dotti has joined #harmattan | 23:27 | |
| *** diggy128 is now known as diggy | 23:31 | |
| *** diggy is now known as diggy128 | 23:32 | |
| *** diggy128 is now known as diggy | 23:43 | |
| *** Venemo_N950 has joined #harmattan | 23:48 | |
| *** Venemo_N950 has quit IRC | 23:49 | |
| *** diggy is now known as diggy128 | 23:50 | |
| *** Venemo_N950 has joined #harmattan | 23:52 | |
| *** Venemo_N950 has joined #harmattan | 23:58 | |
| djszapi | 21:42 < ajalkane> djszapi: you might know... if I've put some requests for credentials that aegis/security framework will not grant for my application, where would I see some log information about that? -> he left, but the answer is command line for instance during the installation | 23:58 |
| *** Venemo_N950 has quit IRC | 23:59 | |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!
