IRC log of #harmattan for Friday, 2011-07-22

*** wazd has joined #harmattan00:05
*** rlinfati has quit IRC00:06
*** CaCO3 has joined #harmattan00:13
*** piggz has joined #harmattan00:14
*** CaCO3 has quit IRC00:18
*** djszapi has joined #harmattan00:26
*** gri_ has quit IRC00:27
* rm_work just read all of the backlog00:31
rm_workGeneralAntilles: did you see what I was trying to do with
*** javispedro has quit IRC00:36
*** djszapi has left #harmattan00:39
GeneralAntillesAh, I see.00:39
ScifigAnybody managed to sync a google public calendar with CalDav?00:41
DocScrutinizerrm_work: I noticed it shows only 2 repos in the dropdown list on device. Maybe I missed sth00:45
rm_workit shows like 13 to me00:45
DocScrutinizeraaah, it only doesn't scroll easily00:45
DocScrutinizerusual crappiness of c-ts00:46
rm_workman, fms had some really "elegant" and "dynamic" perl code behind Packrat.... too bad I barely know perl and am ripping it to shreds and making it completely hardcoded and static to make this work quickly :P00:46
rm_workI should probably just rewrite the whole thing in python00:47
rm_workit's seriously a shame...00:47
rm_worki shed a single tear00:47
DocScrutinizerI'll shed a tear of joy when it starts to work :-)00:49
*** crevetor has left #harmattan00:49
GeneralAntillesCat mutilated my hand last night, now it's sore. :(00:49
DocScrutinizerdamn, what sorts of cat you keep?00:50
rm_workhe has a collection of Ligers00:50
GeneralAntillesI didn't wake up for the mutilating.00:50
rm_workor Toygers?00:51
GeneralAntillesI'm assuming I either rolled over on her00:51
GeneralAntillesor she was momentarily taken by The Darkness.00:51
DocScrutinizerligers o.O00:51
rm_work... is your girlfriend named Catherine? :P00:51
GeneralAntillesDeep canine puncture wounds, though.00:51
rm_work<GeneralAntilles> Cat mutilated my hand last night00:51
GeneralAntilless/Cat/cat/ :P00:51
DocScrutinizerthose are eating a cow for breakfast I heard00:51
rm_workit was just ambiguous enough :P00:51
rm_worktaken in that context, that conversation was hilarious :P00:52
npmis there a specific ifdef I can use to say I'm on harmattan, like Q_WS_MAEMO_501:15
*** aapo has left #harmattan01:16
*** djszapi has joined #harmattan01:17
*** djszapi has left #harmattan01:17
*** djszapi has joined #harmattan01:17
npmi've been using (defined(Q_WS_X11) && defined(QT_OPENGL_ES)) as a workaround.01:17
npmany better suggestions?01:17
djszapifor what ?01:20
deimosthere is also __arm__ , but  I seen somewhere Q_WS_MAEMO_6 ?01:20
djszapiI think it will be available soonish, as if someone had told that few days ago01:20
*** djszapi has quit IRC01:31
*** meegoexperts_mob has quit IRC01:39
leinirhmm... wonder what's causing the N950 SDK installer to say "Could not fetch Updates.xml: Error downloading - server replied: Bad Request"01:47
MohammadAG<npm> is there a specific ifdef I can use to say I'm on harmattan, like Q_WS_MAEMO_501:52
MohammadAG<npm> i've been using (defined(Q_WS_X11) && defined(QT_OPENGL_ES)) as a workaround.01:52
MohammadAGfinal SDK afaik01:52
*** eman has joined #harmattan01:52
*** rm_work has quit IRC01:53
MohammadAGnpm, use for now01:53
npmyes the final SDK will have  MEEGO_EDITION = harmattan... thanks for the link!01:53
MohammadAGFYI I currently define Q_WS_MAEMO_6, gives a standardized look ;)01:55
*** qgil_ has quit IRC02:05
*** TSCHAKeee has quit IRC02:05
*** TSCHAKeee has joined #harmattan02:06
*** achipa has quit IRC02:09
*** rcg has quit IRC02:11
*** Scifig has quit IRC02:22
*** andre__ has quit IRC02:35
*** smoku has left #harmattan02:52
*** deimos has quit IRC02:59
*** npm has quit IRC03:14
*** npm has joined #harmattan03:26
*** npm has quit IRC03:32
rm_yougraaaah someone needs to port pandora03:33
rm_youmaybe i'll do it <_<03:33
rm_youstill busy fighting with QML tho <_<03:34
*** npm has joined #harmattan03:44
*** wazd has quit IRC03:47
*** lcuk2 has joined #harmattan03:47
*** eman has quit IRC03:47
*** lcuk has quit IRC03:50
*** hardaker2 has joined #harmattan03:56
*** hardaker has quit IRC03:58
*** willer_ has quit IRC04:00
GeneralAntillesLion broke the Qt SDK.04:01
MohammadAGit's a conspiracy!04:05
*** leinir has quit IRC04:12
*** npm has quit IRC04:13
*** seif has joined #harmattan04:13
seifhey guys04:14
*** _NIN has quit IRC04:14
seifare there any community repositories for harmattan04:14
seifor somethign similar04:14
TronicHow many gigabytes does scratchbox require?04:14
TronicI am trying to install on a laptop with small SSD :/04:14
rm_youTronic: 5-6 IIRC04:15
TronicHmm, look like I'll have to use an external HDD.04:15
*** npm has joined #harmattan04:27
*** npm has quit IRC04:30
*** npm has joined #harmattan04:30
*** Ken-Young has joined #harmattan04:35
Ken-YoungI just received my N950 loaner, and am trying to install the preloaded SDK on an Ubuntu machine.   When I execute the installer, it states "Retrieving information from remote installation sources..." which produces an error popup window stating "Could not fetch Updates.xml" .  I know my network connection is good  - does anyone know what's going wrong?04:38
GeneralAntillesRepo could be down.04:41
seifhey guys quick question again04:42
seifi installed my own version of telepathy instead of the old version shipped with harmattan04:42
seifhowever now it tells me that around 100 packages are not needed04:42
seifcan i know what the main metapackage for harmattan is04:43
GeneralAntillesprobably has "pr" in its name somewhere, though.04:53
*** meegoexperts_mob has joined #harmattan04:54
Ken-YoungOh well, I should have done some searching.   It looks like it is a known problem that the SDK which comes loaded on the N950 is defective.04:57
MohammadAGat least you got your N950 :P04:59
Ken-YoungYup, and I'm a very happy guy.04:59
seifGeneralAntilles, i know i uninsalled it05:01
seifbut i couldnt find it in the repos afterwards05:01
Ken-YoungHow may one become root within the N950's terminal app?05:09
seifGeneralAntilles, you got an idea what else could be named in it05:09
macmaNrun may-i-become-root-please05:09
macmaNor devel-su as an alternative05:10
Ken-YoungmacmaN, I should have guessed!05:10
macmaNif you ask it so politely, it cant refuse!05:10
macmaNeven with aegis05:10
macmaNmaybe aegis will drop its guard, huh DocScrutinizer05:10
macmaNif ur polite05:10
GeneralAntillesseif, may not be available in the repos.05:10
seifthinking of getting it form the scratchbox if its there05:11
GeneralAntillesseif, given there hasn't been an SSU yet.05:11
GeneralAntillesScratchbox's is probably not the same.05:11
Ken-YoungmacmaN, Thanks - devel-su seems to work.05:11
macmaNgoogle n900 n950 migration05:12
seifGeneralAntilles, its not there -.-05:12
seifGeneralAntilles, ok any idea how i can backup my n950 contact list :P=05:12
GeneralAntillesseif, could extract it from the flashable image.05:12
GeneralAntillesDunno about backups05:12
seifGeneralAntilles, yeah05:12
seiflet me try that05:12
MohammadAGafaik there's also develsh05:13
seifGeneralAntilles, do you have an image05:14
seifbecause for some reason its taken ages to download here05:14
seifi am downloading the one click flash thingie05:14
GeneralAntillesseif, not handy, no.05:15
GeneralAntillesIt's a big image.05:15
TronicWhat do I need to get the harmattan targets enabled within scratchbox?05:46
TronicIn particular, do I need to fetch autotools-legacy from somewhere on the host system manually, or perhaps some other files too. I have only run the admininstall script so far.05:47
Tronic(and enabled sudomode + slightly modified the login script so that I can get to sb shell)05:47
seifguys how do i add this repo to my device05:48
Tronicseif: In /etc/apt/sources.list.d/, I suppose.05:50
Tronicseif: Create new file, use the other files as reference.05:51
seifTronic, what about the pgp keys06:07
seifElleo, how do i add your repository to my list06:12
seifdoes it need a pgp key06:12
*** macmaN has quit IRC06:20
GeneralAntillesHrm, where the hell do you turn off all of the input auto-bullshit.06:20
*** macmaN has joined #harmattan06:27
GAN900I really hate feeling clumsy with a new input device.06:29
GAN900Stupid N950 keyboard.06:29
TronicYou mean the feeling of the keys on N950?06:30
TronicI cannot quite put a finger on it (no pun intended) but somehow the N900 keys feel a lot better.06:31
TronicI suppose it might be due to N950 having too flat keys. On N900 they are nicely curved, on N950 it is all flat, except for the little notches on F and J.06:33
GAN900I can hit 60wpm on the N900 touch typing06:56
GAN900Dunno if I'll even get there with the N950.06:56
GAN900frals or somebody, would you please file a bug internally that you can't tell chat windows apart in the small task switcher overview? . . .06:56
*** meegoexperts_mob has quit IRC06:57
rm_youGAN900: yeah i know, the keyboard feels a bit weird. almost like it didn't get a whole lot of attention. i mean, it's like the main thing about the new device, right!?  -- oh wait.07:08
rm_youand on that note, bbl, sleep :P07:08
*** antman8969 has joined #harmattan07:16
*** apol_ has quit IRC07:19
*** DocScrutinizer has quit IRC07:34
*** DocScrutinizer has joined #harmattan07:34
*** smoku has joined #harmattan07:59
*** hardaker2 has quit IRC08:26
*** macmaN has quit IRC08:52
*** macmaN has joined #harmattan08:57
*** smoku has left #harmattan09:15
TronicThe icon templates offered seem to be in 80x80 pixel size but there is no icon folder on the SDK for that.09:19
TronicShould I scale it to 64x64 or use some folder that doesn't exist by default?09:19
TronicThis is for the application icon on the launcher.09:20
ieatlintthe standard deployment icon target from the qt sdk is the 64x64 dir09:24
ieatlintso i'd say you need to probably aim for that, but that's a guess09:25
TronicThe scratchbox template project also had targets for xpm and various sizes smaller than 64x64.09:25
ieatlintnot sure... qt sdk just has the 64x64 png09:26
TronicGot it working after a lot of attempts. The working recipe: no file extension in myapp.desktop, install myapp.png into $$DATADIR/icons/blanco/80x80/apps (with 80x80 dimensions).09:36
*** smoku has joined #harmattan09:36
TronicThe menu actually uses 80x80 icons. Allows for a 64x64 logo to fit nicely inside that rounded corner box.09:37
TronicIcon templates can e found at google://harmattan+icon+template09:37
*** rcg has joined #harmattan09:50
*** vgrade has quit IRC09:51
*** vgrade has joined #harmattan09:53
*** smoku has left #harmattan10:07
*** wazd has joined #harmattan10:08
*** cpscotti has quit IRC10:10
*** Khertan has quit IRC10:11
*** aleksander_m has joined #harmattan10:13
TronicYay, my first MeeGo app live on the the device \o/10:24
antman8969just read10:28
antman8969tronic, you should use a .svg and put it in /usr/share/icons/scalable10:29
antman8969despite what the default is10:29
antman8969if you use a png you may notice that your app icon is blurry and or your install icon is blurry10:29
*** CaCO3 has joined #harmattan10:29
antman8969and with a svg the dimensions don't matter10:29
antman8969but 44x44 is what I've seen10:30
*** 36DABX8UX has joined #harmattan10:30
*** 45PAACNFX has joined #harmattan10:30
Mekuhm, the install icon is a seperate one taken from the base64-encoded one in your control file10:30
antman8969sry, 80x8010:30
Mekso you can have a 64x64 for that and an 80x80 for the menu10:30
Mekno need to use svg (scalabable graphics are still nice of course)10:30
*** CaCO3 has quit IRC10:31
*** 45PAACNFX has quit IRC10:31
*** 36DABX8UX has quit IRC10:31
antman8969yea the install icon was mispeak, but for whatever reason any png I used was not rendered correctly on the app screen10:31
wazdso they are not svgs after all?10:31
antman8969optionally. I am using svgs10:32
wazdantman8969: can you try this? :)10:32
antman8969try making it an icon on the app screen?10:32
antman8969looks perfect10:35
antman8969did you make it?10:35
wazdawesome news, thanks a lot :)10:36
antman8969lol np, gj10:36
Summeliwazd: you're making a snes emu too?-)10:37
wazdSummeli: no but javispedro does I think ;)10:37
SummeliI have my antsnes already ported10:37
Summeliruns pretty well on N950 too10:37
wazdSummeli: oh :)10:37
antman8969source available?10:38
antman8969 / package made10:38
wazdSummeli: maybe I can create another one for you? :)10:38
Summeliantman8969: source is available, but I don't have the packaging working yet :(10:38
wazdSummeli: SNES controller has different color variations fortunately :)10:38
SummeliI'm using qtcreator to create the deb-packages, so I don't know how to integrate to the OBS etc.10:39
antman8969lol yea I havn't really done too much in OBS yet either.. I've been using creator and scratchbox10:39
antman8969I"d like to start seeing what people are doing10:39
Summeliit's basically same as this:
Summelithe qtcreator had somekind of "publish" interface for the maemo, so I might just wait them to add that to the harmattan too10:40
Summelior learn to actually use the obs10:40
antman8969lol well you could do scratchbox in the meantime10:41
wazdSummeli: <- this one? :)10:41
Summeliwazd: yeah, I have controller like that :)10:41
antman8969it's nice having a capacitive for these kind of things now lol10:42
wazdSummeli: so be it :)10:43
SummeliI have the gpsp also almost working.. The audio is good, but the video is failing badly10:43
*** andre__ has joined #harmattan10:50
*** andre__ has joined #harmattan10:50
*** antman8969 has quit IRC10:51
wazdSummeli: <- there you go :)11:05
Summelithanks :)11:05
Elleoseif: as root simply do: cd /etc/apt/sources.list.d/ && wget && apt-get update11:07
*** leinir has joined #harmattan11:09
wazdSummeli: small update:
wazdSummeli: I've been wrong bout shape of the buttons11:17
Summelithat looks nice11:17
SummeliElleo: ok, i can find my own (non-working package) already with that :)11:22
*** smoku has joined #harmattan11:32
*** cpscotti has joined #harmattan11:39
*** piggz has quit IRC11:42
*** achipa has joined #harmattan11:57
*** achipa has quit IRC11:57
*** achipa has joined #harmattan11:57
hiemanshumorning Termana12:05
*** piggz has joined #harmattan12:05
* cpscotti learns that inheriting your "MainWindow" from QDeclarativeView is not such a nice idea.. booster FTW! :D12:18
*** M4rtinK2 has joined #harmattan12:36
*** cpscotti_ has joined #harmattan12:43
*** cpscotti has quit IRC12:46
*** cpscotti_ is now known as cpscotti12:46
cpscottiHey all.. any ideas on this: (Using booster + QDeclarativeView + reimplementing QWidget.closeEvent)12:48
*** cpscotti has quit IRC12:55
*** cpscotti has joined #harmattan12:57
*** deimos has joined #harmattan12:57
tommacpscotti, how about using eventfilter?13:13
cpscottitomma, hmm dunno. Never tried..  how does that work?13:14
cpscottitomma, that looks good!13:15
cpscottitomma, thanks man! That looks pretty good!13:16
*** hulkkii_ has joined #harmattan13:19
*** piggz has quit IRC13:24
*** NIN101 has joined #harmattan13:28
*** piggz has joined #harmattan13:29
*** lcuk2 is now known as lcuk13:31
*** lcuk has quit IRC13:31
*** lcuk has joined #harmattan13:31
cpscottitomma, perfect! works like a charm.. Just one other bit, at installEvent's doc it says that returning true would "stop"/filter the event but I had to use even->ignore() to actually stop the window from being closed immediately. Does that sound right?13:37
*** slaine has joined #harmattan14:03
slaineright, just opened my dev kit box14:03
slainewondering of my n900's A-10E charger will charge the n950 ? Manual seems to indicate yes14:04
leinirslaine: people say yes :)14:05
slainegood to know, didn't want to blow anything up :)14:06
leinirat least people have been reporting on forums etc that it works - i've not tried myself, i so rarely use it :)14:09
slaineWhat's MOSLO ?14:09
leiniras i'm by the computer anyway, i just use that ;)14:09
leinirmore slow in jive?14:09
slaineleinir: see the topic14:09
leiniraah right ;)14:10
slainelol, I got sent here to talk about harmattan and moslo is now being discussed in #meego-arm14:11
Mekmoslo == boot loader iirc14:11
MohammadAGNOLO = bootloader14:14
MohammadAGmoslo is something for dual boot14:14
wazdobvious names indeed14:15
cpscottihey, anyone knows where's the source for the Boosted Widget Gallery? That one's startup is really boosted but if you just follow the booster docs you I can't achieve the same :/14:26
*** smoku has quit IRC14:37
MohammadAGcpscotti, are you using MTF?14:37
*** smoku has joined #harmattan14:37
Damion3gah, do I also need to sit on #meego-arm to pick up stuff?14:42
cpscottiMohammadAG, what's MTF?14:43
andre__meego touch framework14:45
cpscottihmm.. no.14:45
cpscottiandre__, I'm using harmattan-platform-api14:45
cpscottiis that it?14:46
tommaqml components?14:47
cpscottitomma, yes14:49
cpscottitomma, andre__ MohammadAG  .. so? any idea on where's that source? is that one using mtf?15:03
Mekyes, the widget gallery ones use mtf, not qml15:05
DocScrutinizera pity about the NICE dat/time-picker of MTF and the unbearably butt ugly ones of QML15:05
*** cpscotti has quit IRC15:05
Mekyeah... the qml components are lacking quite a few nice things from the MTF widgets...15:06
DocScrutinizerthough even the MTF datepicker *could* be smarter: could be just calendar app called with a special parameter, but showing the whole calendar - with appointments and all15:08
DocScrutinizerso picking a date from week or month view doesn't open the day's detail view in calendar app but rather returns the picked date to caller15:09
DocScrutinizerTHAT would be really awesome smart system design15:09
*** wazd has quit IRC15:10
DocScrutinizerwell, with QML we can forget about such fancy stuff anyway15:11
DocScrutinizerall hail to slotmachine :-/ without even day-of-week display15:12
DocScrutinizerI couldn't resist to open a ticket on nokia bugtrac for it, feel free to vote for it - or whatever it is you could do there15:14
*** Ken-Young has quit IRC15:20
DocScrutinizerbtw who's to pester about mask design? the one-line text-edit for comment is an offense15:28
DocScrutinizerI'm tempted to send 5..20 "sequel" comments, each with just the portion of text visible for proofreading in that one-liner text field15:30
DocScrutinizer...which btw is less than the previous one sentence post15:31
tommait does allow multiple lines... maybe your browser just doesn't know how to scale text areas ? =)15:31
DocScrutinizerumm, possible. Seems Nokia has "optimized for MSIE" all over the place15:32
DocScrutinizeranyway I usually don't have similar problems on other bugtrackers15:32
DocScrutinizerusing Konqueror just in case it matters, and NO this won't change just for Nokia15:33
*** razvanpetru has joined #harmattan15:36
DocScrutinizerbtw the main problem with a lot of Nokia webpages seems JS to me15:37
DocScrutinizernot checked if this applies to bugtracker as well15:37
razvanpetruanyone else had problems updating a button's imageSource?15:37
razvanpetruI see that the property is not an alias to the actual image source for some strange reason...15:37
*** hardaker2 has joined #harmattan15:38
DocScrutinizerravfrom some 5 years ago and plain Qt I know that images for widgets usually are included to source, you had to do some tricks to actually use an external file for source at runtime15:39
DocScrutinizerrazvanpetru: ^^^15:39
DocScrutinizers/tricks/dynamic loading of the image data in plain hard code/15:40
razvanpetruiconSource: (qmlBridge.isActive(itemId) ? "image://theme/icon-m-common-remove" : "image://theme/icon-m-common-add")15:41
razvanpetruisActive from C++ is not getting called and I get a add icon :)15:41
razvanpetruso what is the JS engine doing, evaluating the first part of the ternary OP as a variable? :P15:42
razvanpetruDocScrutinizer: for Button.qml I think it was designed to be updateable, the iconSource property is public...15:43
razvanpetruanyone interested in creating a gtd app?15:46
DocScrutinizerrazvanpetru: NFI about JS and QML :-)15:49
DocScrutinizersorry, my comment probably wasn't really helpful15:50
*** seif has quit IRC15:51
razvanpetruhm, so what happens is this: the function IS getting called, but before the sheet is even created. of course it returns false because there's not data at that point in it :) I think I'll have to use a Loader for this one...15:51
DocScrutinizerhehe, the joy of OO15:52
*** gri has joined #harmattan15:53
DocScrutinizerafter 10+ years I'm still undecided if I consider OO a good or a hindering thing, as you have a really hard time to keep up with what's *really* going to happen on your CPU, and *when*15:53
razvanpetruit's good :P15:53
razvanpetruit's just a way to manage complexity though, nothing more15:54
razvanpetrujust like functions, files and so on15:54
*** khertan_ has joined #harmattan15:54
razvanpetrubut it's quite easier to screw up the design with it15:54
DocScrutinizeryup, basically you're right15:54
DocScrutinizerthat's why I'm undecided15:54
*** khertan_ is now known as Khertan15:55
DocScrutinizerI tend to think for GUI and similar stuff it's usually the right paradigm15:55
DocScrutinizerunless you run into nice little gotchas like yours15:55
razvanpetruI think I ran into a JS gotcha myself - if it's typed in it's loaded, parsed and run :)15:56
TSCHAKeeetoo bad C++ OO is passed off as real OO...15:56
TSCHAKeeeat least Qt makes it closer to something sane.15:56
Mekthis gotcha has nothing to do with OO, and much more with declarative/functional code15:56
DocScrutinizerif you do, then you have a few funy hours digging through all the object sourcecode15:56
razvanpetruas opposed to fake OO? :P15:56
TSCHAKeeerazvanpetru: Stroustroup was so obsessed with compiler abstract data type resolution that he put his blinders on to every other aspect of OO.15:58
TSCHAKeeeI still think he's a nut.15:58
razvanpetruI kind of like the guy, he always seems relaxed when speaking/writing15:58
razvanpetruhe's attained zen...15:58
TSCHAKeeehe's had to. he constantly get pelted by the Smalltalk/ObjC/Ruby/Late Binding community (of which, I am one.) :P :)15:59
infobotTSCHAKeee meant: he's had to. he constantly gets pelted by the Smalltalk/ObjC/Ruby/Late Binding community (of which, I am one.) :P :)15:59
TSCHAKeeealways startles me when an infobot does that.15:59
razvanpetruthanks infobot15:59
DocScrutinizerinfobot: shame on you!16:00
infobotShame on you on you!!16:00
DocScrutinizerthanks apt16:00
DocScrutinizeroooh even this doesn't work anymore - good16:00
razvanpetrulate binding is good, but I'd like it to be explicit. e.g: I decide when to use it :)16:00
razvanpetrukind of like C# dynamic16:00
TSCHAKeeeinteresting approach.16:01
*** fiferboy has joined #harmattan16:04
razvanpetruwow, using a Loader is pretty easy, much easier than lazy-loading widgets. on Symbian I saw big issues with app startup time from QWidget creation and the workaround was not nice at all...16:07
*** rm_work has joined #harmattan16:07
*** rm_work has joined #harmattan16:07
*** hulkkii_ has quit IRC16:08
Khertanrazvanpetru, you will see later that Loader isn't perfect for everythings :)16:10
griwhen it loads dll/so files, it blocks ui :(16:10
*** cpscotti has joined #harmattan16:12
razvanpetruwhy/how load a dll with Loader?16:13
gricom.meego -> so file16:13
razvanpetrumuch to learn I still have :P16:14
*** lcuk has quit IRC16:15
*** apol_ has joined #harmattan16:26
*** lcuk has joined #harmattan16:33
*** crevetor has joined #harmattan16:34
griIs there a way to install a .deb package by ignoring it's architecture? The harmattan beta sdk directory only contains libqt4-meegographicssystemhelper-dev for armel but it's independent and should also work on i38616:35
*** seif has joined #harmattan16:35
piggzhas anyone noticed dissapearing digits from the lock screen clock?16:37
Kaadlajkit is fixed in later firmware16:38
DocScrutinizeryes, 9 and 016:38
razvanpetruit's a battery-saving feature16:38
griI just had it twice: After first boot and after a boot where battery got completely empty16:39
*** lardman|gone is now known as lardman16:39
DocScrutinizerI have it all the time now16:39
DocScrutinizerI have not all the time, all the time now :-D16:40
lcukhi lardman16:40
lardmanhey lcuk16:40
lcukenjoying your friday?16:40
griI have another nice feature: The phone application and messages application never shows the name of my girlfriend, only the number - all other contacts work. This may be a sign she is evil16:40
DocScrutinizerlardman: hi16:40
lcukand are you doing anything nice this weekend?16:40
lardmanlcuk: been to midwife appointment, so only just got into work16:41
lardmanhi Doc16:41
lardmanlcuk: more unpacking I think!16:41
lcukgri, the phone application shows me your girlfriends name ok!16:41
lcuklardman, how long to go now?16:41
lardmandue 14/816:41
lardmanyeah, getting scarily close16:41
lardmanbbiam, have to pop to bank16:42
fiferboylardman: I don't think I have congratulated you yet16:42
fiferboylardman: Congrats!16:42
lcukfiferboy, mrs lardman is going the hard work!16:42
fiferboylcuk: Nah, the husband is the most important part :)16:43
*** razvanpetru is now known as razvaway16:45
slainejust fired up my n950 for the first time16:47
slainefirst impressions are very good16:47
slaineI like the hardware and the software so far, not done much beyond noodle around though16:47
lcukfiferboy, if I was starting a qml app, which database storage would be recommended?16:48
mgedminsqlite, lcuk?16:49
*** seif has quit IRC16:49
lcukmgedmin, are there sqlite bindings for qml?16:51
mgedminno clue16:52
mgedminfiferboy, when I finally get a harmattan scratchbox, I'll be interested in the steps you took to build your fbreader .deb16:52
mgedminlcuk, google hints that uses sqlite as a backend16:53
*** achipa has quit IRC16:54
crevetorlcuk: there is16:55
crevetorlcuk: hold on a sec16:55
mgedmincrevetor, did you see my gitorious merge requests?16:56
lardmanfiferboy: thanks :)16:56
crevetorlcuk: actually I used what mgedmin sent16:56
*** seif has joined #harmattan16:57
*** smoku has quit IRC16:59
*** razvaway has quit IRC17:00
lardmanright, parsers....17:00
kimju - anyone want to test (or do a code review for the pre/post inst/rm scripts)?17:01
lardmanRST38h: you suggested something a while back, any links to good docs to read on how to write parsers?17:01
*** seif has quit IRC17:01
lcuklardman, yacc and bison17:01
lardmanI need to parse command strings in a Qt/C++ app17:01
lcukcommand line arguments, or interactive?17:02
lardmanarbitrary numbers of brackets for both grouping and function args17:02
*** seif has joined #harmattan17:02
lardmancommand-line-ish, they will be Proximus rule commands17:02
* lardman wonders to create own grammar or use xml or something like that which already has a parser17:03
*** smoku has joined #harmattan17:04
lardmane.g. sorts of commands in here:
lardmanany thoughts or ideas gratefully accepted17:05
lcuklardman, does proximus already include a parser for itself?17:06
fiferboymgedmin: Sure, let me know when you are set up and I'll send you the source17:06
lardmanlcuk: no, I'm writing it now17:07
lcuklardman, ahh I tohught proximus was an existing app with parser etc17:07
lardman"for itself"? It's a C++ app, so it only needs the parser to determine the tree structure of rules passed in17:07
lardmannah, just something Jaffa and I were thinking about and wrote a wiki page about17:08 has download icons with the arrows pointing *up*17:08
lardmanand which I'm now writing17:08
* lardman should push the location manager updates he started writing yesterday17:09
mgedminkimju, I'm looking at the postrm, and I don't understand why you remove the diversion only on purge17:09
mgedminoh, the file you're installing is a conffile!17:11
mgedminso it doesn't get removed on remove17:11
mgedminI'm not sure I like that; apt-get remove won't restore the layout to the default one, seems counter-intuitive to me17:11
lardmanhmm, I think perhaps writing it in XML would be easier, at the expense of making it harder to read the rules17:12
lcuklardman, you can always make a visual editor for the rules17:12
lcukand I agree, xml would be reasonable storage17:13
kimjumgedmin, problem is that the diverted file is not the configuration file, it's the datafile for definitions. and the conffiles are under /etc/meego-input-method-configuration/hw-keyboard17:15
lardmanlcuk: yeah, well people won't be looking at the rules and they will most likely be created by specific apps anyway17:16
kimjumgedmin, and if the diverted file gets removed before the config-files, update-hwkb-config will fail to set any reasonable layout.17:16
lardmanso xml seems reasonable17:16
mgedminkimju, I see17:16
kimjumgedmin, I don't like that either, but that was best i could come up with.17:16
lcuklardman, do we have always on gps and other nice things?17:17
lardmanwith Proximus?17:17
lcuki mean on device in general17:17
kimjuand it seems that the purge doesn't work either, as the postrm purge gets called before the conffiles are removed.17:17
lcukdoes it drain battery17:17
mgedminkimju, and now I realize I misunderstood how this works -- the user chooses a 'developer' layout in Settings, right?17:17
lardmanwoth n950 yeah I have it set to be on, battery seems good17:17
mgedminthe package doesn't silently override the default layout17:17
lcukgood stuff17:17
lardmanbut it's not always running as I don't have the map on all the time17:17
mgedminso there's no harm in leaving the extra layouts about; if the user doesn't want them, she can choose the default layout in Settings17:18
kimjumgedmin, there is no UI for selecting hwkb layout.17:18
mgedminoh? oh! aaaaugh17:18
kimjumgedmin, see the readme in that package and the /usr/bin/update-hwkb-config script..17:18
lcuklardman, well how will that work17:19
lcuksince you dont put the map on all the time17:19
mgedminthere was a readme?  I missed it17:19
lardmanlcuk: it only switches on when the map is on, like on the N90017:19
lardmanor are you talking about something else now?17:19
lardmanProximus will use passive location methods until it needs gps accurary17:19
lardmanthen it is switched on17:20
lcuklardman, I am talking about the "shopper" rule knowing you are near the supermarket17:20
lardmanah ok, use cellid with offline lat/lon db to work out general area17:20
lardmanthen gps when you're close enough17:20
lardmanthat is my plan anyway17:20
mgedminkimju, I think it would be nice to also briefly mention the need of apt-get purge in the Description field of debian/control17:20
mgedmin"For technical reasons apt-get remove is insufficient to disable these additional layouts, so if you want to be rid of them use apt-get purge xkeyboard-config-addons-rm680"17:21
kimjuok, will add.17:21
*** TSCHAKeee has quit IRC17:25
*** qgil has joined #harmattan17:37
*** willer_ has joined #harmattan17:39
*** razvaway has joined #harmattan17:41
razvawayharmattan has  a WebkitView but no WebView!?17:41
DocScrutinizerlardman: you ever had a look at dwimd, and the way it is using python for these "config" things?17:42
DocScrutinizerlardman: for creating complex rules python for sure is a better "config syntax" than any random xml makeup17:45
crevetorrazvaway: in QML ? I use WebView on Harmattan17:49
razvawayyea, PEBKAC17:49
*** CaCO3 has joined #harmattan17:51
*** aleksander_m has quit IRC17:53
*** apol_ has quit IRC17:56
lardmanthanks Doc, I'll have a look see17:56
lardmanThe only problem there is that you need to write Python scripts, which will be difficult to do automatically17:58
DocScrutinizernot really18:01
DocScrutinizerit seems fairly easy to insert templates and edit them18:01
DocScrutinizerhpwever there you *could* use "advanced" coding superpowers to tweak the "config" directly if you feel like doing that18:02
DocScrutinizeraiui you can even have multiple concurrenttly active configs in /etc/dwimd/dwimd_plugins/configs18:04
lardmanCertainly I can understand the utility of python scripts that abstract the things to be done and checked18:05
lardmanand it would be a good way to provide scripting18:05
DocScrutinizermeans you simply throw in another file containing exactly one complete config for one behaviour. You'd just edit some vaiables in that template prior to copy to the config directory18:05
lardmanbut I'm not so keen on the executing each script every 30min, etc18:06
DocScrutinizerlardman: 30s18:06
lardmanyes, typo on my part18:06
lardmaneven worse though18:06
lardmanI'm doing the same though, multiple rules, the question is how to make it easily expandable18:07
lardmanwhich may be using Python scripting to allow the addition of more complex checks18:07
DocScrutinizerand it's just the script will run all the time basically, it's just it should trigger events on certain conditions, so the sensors, if designed properly, will "block" and not constantly hog CPU and sensors18:07
lardmanI thought that each of the script rules is run every 30s?18:08
lardmanI'll have a look at the source and see how it's done there18:08
DocScrutinizerthat's an implementation detail I had no look into. My idea is the whole thing has an event loop that blocks and only triggers for certain sensors firing18:09
DocScrutinizerwe don't need to copy the implementation, but the design is kinda nice and sane18:09
lardmanyep, and in that case you need to keep track of which rules need to know about certain events occuring18:09
DocScrutinizerwell, aiui python does threading18:10
lardmanbut you may have more than one condition that can mean something should happen within a single rule18:10
DocScrutinizerso each sensor was a thread and would block until a IRQ, timer, or whatever fires/expires18:11
lardmanI've not really looked at how to handle sensors that don't generate their own events18:12
lardmanbut I guess you would need to support polling them, so that you can then signal the rule that something has happened18:12
lardmanideally you sidestep that by checking the condition tree to see whether such polled sensors/events can trigger the tree, and if not you only check their status if other callback-type events are triggered18:13
lardmanwon't work all the time though18:13
DocScrutinizeryes, but that's always true, independent of implementation. So for certain sensors you will want to run some "script" on a regular schedule18:13
DocScrutinizerack for your last comment18:14
lardmanyou'll want to check the sensor status on a regular schedule, but only if it's being used in a rule and only if the sensor can trigger the rule18:14
lardmanotherwise it can just wait to be asked for status once18:14
lardmanthat will cause a step change in power usage though, depending on how often the check is performed18:14
DocScrutinizersee, things like bme do polls several times a minute or even second, without too much impact on standby time. It's all about how much you actually do aka how much CPU load you create during the scheduled "poll"18:16
lardmanthough perhaps I underestimate how often the device wakes up for other reasons anyway18:16
lardmansure, so I'm underestimating18:16
DocScrutinizerideally the device wouldn't wake up at all. powertop however shows maemo5 wakes up several times per *second*18:17
DocScrutinizerTotal wakeups  4306, 143.5/s18:19
lardmanwell polling things like the proximity sensor would probably only be wanted if the device is active and doing something anyway18:19
lardmanwell sounds bad, I've no clue what is bad18:19
DocScrutinizerthat's quite normal18:19
DocScrutinizerso don't worry about ne wakeup every 30s ;-)18:19
*** smoku has quit IRC18:20
lardmanthe other question is whether 30s is often enough for some sensors18:20
DocScrutinizerworry about keeping CPU and peripheral load low while doing it18:20
DocScrutinizeryes, I think the dwimd concept is missing a dynamically adjustable poll time18:20
DocScrutinizeraiui each sensor shall have its own poll pace18:21
DocScrutinizersome never polling, others maybe once a second18:21
lardmanI'd also like to provide a service to other apps, so that other people can create rules for whatever they want, and pass them to e.g. proximus to service18:21
DocScrutinizermost sensors *should* work IRQ driven18:22
slaineDoes the Harmattan SDK / Qt Creator pack work ok on OS X 10.7 ?18:22
lardmanDocScrutinizer: just depends whether that's exposed for use I guess18:23
DocScrutinizerlardman: e.g accelerometer doesn't need 5/s sample rate do make sure you detect a movement. It is IRQ driven and thus could trigger a "script execution" when sensor is first level of a "if moving"...18:23
rm_youholy backlog18:24
rm_younot reading all that18:24
rm_youwell anyway18:24
lardmanrm_you: lol18:25
razvawayis there any hope to have a content area as big as the entire screen in a Dialog?18:25
rm_youdoes anyone know how to install a repo + package on N950? like what a .install file did on N900? or is it un-possible?18:25
lardmannot possible afaiu18:25
lardmanthough needs to be made so18:25
rm_youi fixed up Packrat for Harmattan repos yesterday18:26
rm_youbut the current .install format not only doesn't work, it is blocked by the browser18:26
DocScrutinizerlardman: I used a ISDN python based framework some 10 years or so ago. It had nice implementation of events. Lemme just check if I can find it18:26
DocScrutinizerlardman: you had a python "script" and in this you defined what's going to happen when e.g inbound-call event happened18:27
DocScrutinizerthe python interpreter was inside the ISDN "driver"18:27
lardmanMy issue with that is that you may want to combine more than one event18:28
DocScrutinizerand called the according function in script on events18:28
lardmanmBarocde uses Python scripting within the C++ app, so that's fine18:28
DocScrutinizerlardman: you always have just ONE event as trigger, everything else are conditions checked inside that event18:28
DocScrutinizers/ONE/ONE arbitrary/18:29
lardmanbut a rule can easily be more complex than that18:29
infobotDocScrutinizer meant: lardman: you always have just ONE arbitrary event as trigger, everything else are conditions checked inside that event18:29
lardmanequally, if some of the other conditions are callback rather than poll based, you don't even need to check them, just see what their last state was18:29
lardmanwhich I'm guessing should be "cheaper"18:29
DocScrutinizeryes, exactly. that's duty of sensor object implementation18:30
lardmanquite, so Proximus currently has manager classes for various things, e.g. location, which handles the callbacks from e.g. liblocation, then sends the updates to rule conditions that need it18:31
DocScrutinizerneeds to be carefully and individually crafted for each sensor18:31
lardmanyes indeed18:31
*** achipa has joined #harmattan18:32
*** achipa has joined #harmattan18:32
*** javispedro has joined #harmattan18:35
* lardman heads home18:42
lardmancatch you chaps next week18:42
*** lardman is now known as lardman|gone18:42
*** slaine has quit IRC18:43
DocScrutinizerlardman|gone: HEY!!!
* javispedro installed portmap again and no sign of aegis self-destruction so far. knock on wood.18:52
*** TSCHAKeee has joined #harmattan18:56
*** javispedro has quit IRC19:00
*** qgil has quit IRC19:09
*** razvaway has left #harmattan19:09
*** javispedro has joined #harmattan19:12
DocScrutinizerjavispedro: WB19:13
DocScrutinizer13min for reflash? :-P19:13
javispedroyes, reflashing already ... hrmpf.19:13
DocScrutinizer~curse aegis19:14
infobotMay you be reincarnated as a Windows XP administrator, aegis !19:14
DocScrutinizerthis is a MEGA security vulnerability *introduced* by aegis19:15
javispedroI didn't reflash yesterday so I had to either way, to see if it fixes the nokia account thing19:16
DocScrutinizeractually it seems aegis is turning arbitrary good-minded apps into timebombs - a REAL SUCCESS for security of a user device19:17
javispedroand I'm also playing with fire, as I set portmap to start on boot via upstart19:17
DocScrutinizerhow can you *dare*??? :-P19:17
javispedrook, something is wrong19:18
javispedrohe remembers my locale19:18
infobotjavispedro meant: it remembers my locale19:18
javispedroaaaaaah, sim in.19:18
DocScrutinizerdo you think any noob can come and start random crap at system boottime? like dbus-scripting daemons, or wharever? C'MON!19:19
*** MeeGoExperts has quit IRC19:19
DocScrutinizerthis is a *user device*!19:19
DocScrutinizerI notice I'm starting to get bitter about aegis crap19:20
* DocScrutinizer ponders to write and publish his first harmattan app: "Nuke Me!"19:21
DocScrutinizeryou can activate the random mode on install time, for even better thrills19:22
javispedrothe truth is, I do not think you can19:22
javispedrooh well19:22
DocScrutinizeron default mode clicking the icon will instabrick your device19:22
javispedroyes, you can.19:22
DocScrutinizerand HEY it even is *useful* XP - use it last second prior to evil police or other staff is snatching away your device to scrutinize it and download all your contacts -HAHAHA19:24
DocScrutinizerthe most simple most silly most awesome app ever written, even much better than *#-ena ;-P19:25
DocScrutinizerone line in a script:  echo >>$019:26
DocScrutinizeroh I forgot shebang, so actually a 2-liner19:26
* javispedro ponders if you need to spawn develsh19:27
javispedroand if aegis will brick after or before that =)19:27
DocScrutinizerseriously, this is *unbearable*, completely unacceptable behaviour of this aegis crap19:28
DocScrutinizerI'd consider it a showstopper for whole platform19:29
DocScrutinizeryou can't rollout a device with a system that bricks on $random19:29
DocScrutinizerI mean sony got bashed for their P1(?) *rebooting* randomly19:30
DocScrutinizerwhat do you think engadget et al will comment on a device that *bricks* randomly19:31
Summelihow can I get log from obs? I get that my package has been "exclued" but it would be nice to know why19:31
Summeliexcluded even19:31
javispedroNokia Account NOW works, first try.19:31
DocScrutinizerreflash HELPS19:32
DocScrutinizerprobably I should add my NukeMe! app to crontab, so you MUST reflash once a day :-P19:32
javispedroDocScrutinizer: fun for though: I was stracing the binary, and it does talk with /sys/kernel/security19:33
DocScrutinizerwhich binary?19:33
javispedrono, the nokia account thingie19:34
DocScrutinizererr portmap19:34
DocScrutinizerand you straced that? duh!19:35
DocScrutinizerevil hax0r19:35
DocScrutinizerwell, somebody (frals?) suspected reflashing would break login to nokia account on "our" devices. Seems it's the other way round, though I still miss to get it why it is broken at all19:37
DocScrutinizerI still wonder if it might store up some cookie whenever you enetr incorrect credentials for account on first try19:39
rm_youso what is the state of this: ?19:42
rm_youis there a deb somewhere i can just install, or is it not done yet at all?19:42
DocScrutinizerat least for login to on desktop PC this seems to be the root cause of all problems - I had to tag the "Rememebr me on this computer" checkmark so the login would actually let me in19:43
DocScrutinizerotherwise just trows me backon same login screen19:43
javispedrorm_you: there's no backend, so.19:43
rm_youis it really just the UI? or does it actually let you browse, but there's just no way to install?19:44
javispedroSummeli: you have selected the wrong repository19:44
Summelijavispedro: ok, how do I change it?-)19:45
DocScrutinizer FTW19:45
rm_youDocScrutinizer: :P19:45
rm_youDocScrutinizer: so you think it's useful so far? what changes do i need to make? obviously some cleanup, but...19:46
rm_youi am home sick from work so i can mess with it today19:46
Summeliah, found it19:46
DocScrutinizerwell, it obviously needs to expose .deb to click on, rather than installfiles19:46
Summeliwhat's the correct repository for harmattan?19:46
javispedrowith only armv7el as arch19:47
DocScrutinizerother than that it's just fine for me, YMMV regarding gfx design19:47
DocScrutinizeraka cleanup19:47
Summelijavispedro: ok, and that can be added in advanced interface i guess19:48
javispedroSummeli: ah, yes19:48
javispedroSummeli: put MeeGo:1.2:Harmattan in project19:48
javispedroand Maemo.org_MeeGo_1.2_Harmattan_standard in repo19:48
Summelijavispedro: thanks, I'm going forward now :)19:50
Summeli"no build results avaiable"19:50
Summeliok, now rebuild from somewhere19:50
rm_youDocScrutinizer: ok, installer link fixed19:51
javispedroSummeli: you only depend on libqt4-dev? No GLES headers?19:51
Summelijavispedro: ok, there's a bug, it should have core, gui, and the opengl19:52
javispedrosadly, I do not think you are able ot depend on GL :P19:52
javispedroI wasn't able to few days ago because seemingly the nokia propietary packages have not been imported yet19:52
DocScrutinizermeh, trying to get mc from - not available and mc-dbg_4.7.0.9-2_armel.deb link on still times out19:53
Summelitoo bad :(19:53
Summelijavispedro: but I should be able to at least try to build it? yes? :)19:53
javispedrooh well, try if you wish :)19:53
Summelimaybe I'll just wait19:54
SummeliI got the source structure working for deb and ocs, so I can wait for a while19:54
javispedroso, how does antsnes work? :)19:55
Summelijavispedro: pretty well actually19:55
DocScrutinizerrm_you: transfer started! \o/19:55
Summelithere's little things missing, like adjusting the audio (I'm stil with pulse-audio simple)19:55
DocScrutinizer"Can't install \n Invalid installation package"  SH*T same as with bash19:56
javispedropulseaudio simple is fine19:56
Summelithat was easy to use19:56
rm_youDocScrutinizer: so this is not ever going to work because of ... aegis?19:56
javispedroin fact it is probably better than the alternative unless you want to do wacky stuff such as changing rate on the fly19:56
DocScrutinizerrm_you: nfc what this error message means, in Web browser19:57
Summelijavispedro: volume control would be nice19:57
DocScrutinizerrm_you: I've seen other pkgs install flawlessly19:57
DocScrutinizerI suspect the source isn't trustworthy or sth19:58
DocScrutinizerJul 22 18:55:55 (none) tracker-extract[5955]: GLIB MESSAGE Tracker - Setting up monitor for changes to config file:'/home/user/.config/tracker/tracker-extract.cfg'19:59
DocScrutinizerJul 22 18:55:55 (none) tracker-extract[5955]: GLIB MESSAGE Tracker - Loading defaults into GKeyFile...19:59
DocScrutinizerJul 22 18:55:57 (none) pkgmgrd[5961]: GLIB MESSAGE default - ASSUME NETWORK 019:59
rm_youso good, it's just the one package?19:59
DocScrutinizerI'd need to test with another preferably known-good .deb20:00
DocScrutinizerany suggestion on what that known-good pkg should be?20:01
*** javispedro has quit IRC20:03
rm_youno :/20:03
rm_youmaybe memory-game?20:03
rm_youi assume that works20:03
rm_you(if the page breaks for you at any point, it's because i'm messing with it live right now :P)20:04
DocScrutinizeryup that works20:04
rm_yourepos list works20:05
rm_yousections works20:05
rm_youfinally got the system name to print "Harmattan"20:05
rm_youk, basic tweaks fone20:06
rm_youholy @*%& my monitors have completely different color profiles20:07
DocScrutinizerI wonder what's the issue with bash and mc .deb20:09
DocScrutinizerobviously both need some permissions they don't get, and this gets detected when browser "transfer" tries to start installation on me clicking the downloaded .deb20:10
DocScrutinizerI bet I could wget && dpkg -i the mc.deb, and on trying to run it I'd get same "operation not permitted" error20:12
DocScrutinizeras I get with bash20:13
rm_youi assume so20:13
*** seif has quit IRC20:13
DocScrutinizerok, so the only two .deb I was actually interested in installing so far, both failed20:14
DocScrutinizerI declare this platform unusable for development20:14
DocScrutinizerhonestly it seems the only projects you could develop there are wallpaper changer class, or my NukeMe! app20:16
*** TSCHAKeee has quit IRC20:17
*** TSCHAKeee has joined #harmattan20:17
* TSCHAKeee has been using Qt Creator solely for dev, and it seems to be behaving, with the exception of it trying to load file paths for relative paths out of /home/developer, instead of the app path.20:18
achipaDocScrutinizer: send me the source package20:19
DocScrutinizeryeah, I heard on Android it's also quite easy to "develop apps" as long as you stay inside the sandbox they define there20:19
DocScrutinizerachipa: ???20:19
rm_yousooo who wants to make a QML frontend for Pianobar? :P20:19
achipabtw wget you should have out of the box ?20:19
achipaDocScrutinizer: to check the manifest20:19
achipafor mc20:19
*** crevetor has quit IRC20:20
TSCHAKeeeseriously, what do you expect? they've laid out what to use to develop for a phone that will be targeted for consumers.20:21
DocScrutinizerTSCHAKeee: yes, QML20:22
* TSCHAKeee is using a hybrid of STL C++, Qt C++, and QML20:22
DocScrutinizeryou obviouly don't need mc on target device to click together some QML in SDK20:22
TSCHAKeeetrust me, it's not just qml20:23
DocScrutinizeryou don't need bash for that either20:23
TSCHAKeeebut whatever.20:23
TSCHAKeeeqml is just the icing.20:23
DocScrutinizerTSCHAKeee: fine, so you say the sandbox has many toys for you and fits your needs. Not mine though, I don't want a sandbox20:24
*** seif has joined #harmattan20:26
TSCHAKeeein the end, Nokia saw fit to send me a device to create an app that will be run on the N9. So that's what I'm doing.20:26
TSCHAKeeewe shouldn't lose sight of that fact.20:26
rm_youout of curiosity20:27
rm_youwill an app that uses Python-Qt work?20:27
DocScrutinizerfor the things I'm interested in, I need a proper toolbox *on device* starting with a real shell like bash, and not ending on I2Ctools. And I'm not going to build each and every of those tools native, with each single build taking 2 days until you figured how to get it right for aegis20:27
TSCHAKeeedoesn't madde provide what you need for that?20:27
rm_youanyone have a SB environment they want to let me SSH to? :P20:27
achipaErrorcode: unknown20:28
achipa I <3 OBS20:29
DocScrutinizerTSCHAKeee: me? I2Ctools, bash, on MADDE? lol!20:29
TSCHAKeeeno, smartass20:29
TSCHAKeeejust to build the packages you need20:29
TSCHAKeeeso you can do dev on device20:29
achiparm_you: define python-qt20:30
TSCHAKeeei mean, we've got OBS.... MADDE.... Qt Creator....20:30
TSCHAKeeethat covers a lot of use cases.20:30
DocScrutinizerI'm not inclined to build my own mc, just because aegis doesn't want me to use the one djszapi built20:30
achipaDocScrutinizer where is libslang2 supposed to come from ?20:31
DocScrutinizerachipa: how am I supposed to know about the answer to that question?20:32
DocScrutinizerare you asking me to run a find across my local mirror of nokia repo?20:32
rm_youachipa: Depends:python2.5, python2.5-qt4-dbus, python2.5-qt4-gui, python2.5-qt4-webkit, python2.5-qt4-network, python2.5-gobject20:33
DocScrutinizerachipa: please notice I'm not the one who built mc20:33
achiparm_you: nope, rinse and repeat in pyside20:33
rm_youachipa: i'm looking at this package for fremantle:
DocScrutinizerthat was djszapi20:33
rm_youachipa: considering attempting to port, just trying to determine initial complexity. if it was just building a few packages for harmattan, i'd do it right now, but if it'll require coding and redoing the UI, it'll have to wait20:34
achipa(08:10:46 PM) DocScrutinizer: obviously both need some permissions they don't get, and this gets detected when browser "transfer" tries to start installation on me clicking the downloaded .deb20:34
achipaso this is basically guesswork ?20:35
achipabecause the package fails because of a fscked dependelncy20:35
achipanothing to do with aegis or permissions or anything20:35
DocScrutinizerachipa: for bash it's result of investigations javispedro did20:35
achipagood, well now you know that mc fails for another reason20:36
achiparm_you: not redoing, just packaging and import statement mods, mostly20:36
rm_youachipa: ok, i'll give it a shot20:36
DocScrutinizerachipa: what I don't get though is how it can build with a missing dependency20:37
rm_youanyone out there want to give me a shell with Scratchbox? :P20:37
*** apol_ has joined #harmattan20:37
achipaDocScrutinizer: it's not a build-dep, he might have just left it in the control file20:37
achipahm, it does have buil-depends... but where is it then...20:38
* DocScrutinizer sighs20:38
achipaI this turns out to be some OBS buggery...20:39
DocScrutinizerohwell, leave me a note when this platform is ready for *serious* development20:40
DocScrutinizermeanwhile I'm going to add some chrome and bells and icing to my NukeMe app20:42
*** qgil has joined #harmattan20:42
achipawhat's the platform got to do with a random package a random dude built on a random OBS build ?20:43
achipa(random ranting)20:43
DocScrutinizergo fix then again ask me about state of platform20:45
DocScrutinizerscratc ramndom from $package, $dude, and $OBS for ^^^ that one20:46
rm_youDocScrutinizer: still, one or two packages don't mean the whole system is broken20:48
rm_youand obvious it is very beta, that's been well known from the start20:48
DocScrutinizerjust it's the first and only two packages I really tried so far20:48
achipawell if it's a Nok package go file a bug...20:48
rm_yousmall sample size20:48
achipaslang is not part of the API, it's only in the SDK repo20:49
DocScrutinizera samples size of two negatives out of a pool of NNN is no random result20:49
achipait is an unknown result20:49
achipa(if we are in the game of nitpicking :) )20:49
DocScrutinizeradd in javispedro20:50
DocScrutinizerhe nuked his device two times now, err 3 times. 2 times at random moments when portmap did sth aegis didn't like, and one time when he tried to use a self compiled stock kernel20:51
DocScrutinizerso for me it's NukeMe app for now, all beyond that is too tedious20:53
rm_youwhat does NukeMe do? :P20:53
DocScrutinizerecho >>$020:53
achipayeah, I'm totally sending back this thinkpad to Marc Shuttleworth20:54
rm_yourandom question, is there any way to just DISABLE the VKBD? it's really annoying when it pops up randomly20:54
achipacompiled a stock kernel, entered a line or two and now it doesn't boot20:54
DocScrutinizerplease elaborate on >entered a line or two<21:00
achipasudo rm -rf /21:01
achipait all went downhill from there21:01
achipaand they keep telling me that sudo is there to protect from unauthorized access21:02
achipaso anyway21:02
achipaI took a look at the mc package21:02
achipait lacks a manifest21:02
achipanot even a default one21:02
DocScrutinizerso anyway building and flashing a STOCK kernel is supposed to not work, yes??21:02
DocScrutinizerno "entering rm -rf" :-P pfff21:03
achipahey, what do I know, ask 'keeps21:04
DocScrutinizersorry, got some useful things to do, bbl21:04
achipaand tell djszapi to build with a manifest21:04
fiferboyachipa: I thought there is at least a default manifest by virtue of building in Harmattan environment?21:11
fiferboyIs that not ture?21:11
achipafiferboy: it was on OBS, not the 'real' SDK21:12
achipa(i.e. he didn't build with the SDK)21:12
fiferboyachipa: So the OBS implementation doesn't bestow the default manifest?21:13
*** antman8969 has joined #harmattan21:14
achipaor does it in a was that didn't go through in this particular build21:14
fiferboyI'll have to check my packages from OBS, but I /think/ they have them...21:15
achipaDocScrutinizer: I tried to install the bash package but it won't brick, dammit... what am I doing wrong ?21:17
*** CaCO3 has quit IRC21:18
DocScrutinizerwho said dpkg -i bash.deb would brick the device?21:18
achipaso you did what exactly ?21:18
DocScrutinizerwhat you're doing wrong? you don't listen to what I say21:19
achipa(08:54:38 PM) DocScrutinizer:
achipahow did you get here21:19
*** qgil has quit IRC21:20
achipaokay, so you DID do a rm -rf /21:20
DocScrutinizerI didn't install a kernel though, please don't arbitrarily mix up things21:21
*** antman8969 has quit IRC21:21
achipathat was a joke21:23
DocScrutinizer>>After installing the package, do NOT modify the installed files if they request a token. Security FW will discover an unexpected change in the file and lock the device (ops! reflash)<<
achipathat's what I'm saying21:24
DocScrutinizerthat's about python files, perfectly sane to edit21:24
achipayou did a rm -rf / equivalent21:24
achipait's a sane command21:24
achipabut also deadly21:24
DocScrutinizerallowing aegis to brick the device is deadly - for N9 merchantability21:25
achipayes, that's why there are a couple of obstacles in terms of other sources, developer mode, etc, etc21:26
DocScrutinizerif it's not allowed to edit a python file, then it's aegis' damn duty to forbid editing, not to brick device21:26
DocScrutinizerside note: I didn't do that21:27
achipaI mean, I can brick pretty much ANY OS in two or three command line cmds21:27
achipaDocScrutinizer: then at best it's a packaging bug21:27
achipathere is no protection against bricking on ANY OS21:28
achipa(if you go into admin/developer modes and throw a wrench in there)21:29
DocScrutinizerthere's a difference though between exposing /dev/hda and plastering the whole place with invisible AP mines21:29
GeneralAntillesrm_work, any better progress today?21:29
DocScrutinizerand honestly aegis enforcing a reflash of a perfectly sane system is outright malicious21:30
achipaDocScrutinizer: not invisible, but I fully agree that there is not enough docs to deal with them21:31
rm_youGeneralAntilles: progress on what? :P on my QML app, no21:31
GeneralAntillesrm_work, OK, anything else then? :P21:31
rm_youGeneralAntilles: cleaned up a bit21:31
achipait's not about sanity, it's about being tainted21:31
DocScrutinizera device in developer mode is tainted by definition21:32
achipadeveloper mode is a flexible term21:32
DocScrutinizerindeed, and Nokia's definition is so utterly odd I don't find terms to describe it21:33
*** vgrade has quit IRC21:33
achipayou should see the Microsoft definition21:34
DocScrutinizerI'm not going to touch this thing anymore prior to Nokia explaining how I get $random binary running and *working* in DEVELOPER mode (my definition of developer mode, I don't care what's the term in Nokia speak, may it be "open mode" or "tainted" or "developer" or whatnot)21:35
*** antman8969 has joined #harmattan21:36
DocScrutinizerI got thid device to mess with kernel and low level stuff, that'S been my application. The system doesn't allow21:36
DocScrutinizerI'm waiting for Nokia to fix the issue21:37
DocScrutinizerI don NOT NEED aegis in what I think is a proper developer mode21:37
DocScrutinizerand I don't care if half the system refuses to run in devel mode as Nokia thinks it's too dangerous to expose IP by allowing it to run21:38
DocScrutinizerthat's what TC concept is meant to define developer mode like21:39
DocScrutinizerNOT aegis forcing a reflash, WTF?!21:39
DocScrutinizerwhich gashead even dared to think about such nonsense21:40
*** crevetor has joined #harmattan21:44
lcukDocScrutinizer, is the cable detecter really a viable thing now we have magnetometer?21:44
SpeedEvillcuk: yes - if it can sample fast enough21:46
lcukwell how fast  might it work we have qmlcompass somewhere?21:46
crevetorlcuk: what's up with qmlcompass ?21:47
SpeedEvilYou're looking for signigicant 50/60hz variation in readigs21:47
crevetorThe rate is supposedly user-definable21:47
lcukcrevetor, nothing up, but wondering about using it to prototype one of DocScrutinizer's device app suggestions21:49
lcukie testing one of the simpler non aegis items21:50
crevetorlcuk: the sources are there fyi :
*** antman8969 has quit IRC21:52
lcukcrevetor, is it on a repository for install?21:52
crevetorOups sorry about that, wrong window21:52
SpeedEvilcrevetor: it's not so much thr rate that's interesting - it's if the device has a sampling period minimum of under a few milliseconds.21:52
SpeedEvilIf it's possible, you will see - wafter ignoring the static fiel - a vector changing at a frequency related to the samplerate of the matetometer and the mains21:53
crevetorlcuk: not a repo but you can download the .deb here :
crevetorSpeedEvil: you mean like changing at 50Hz in europe ?21:54
SpeedEvilSo average over a second (say), and see what the residual data looks like after you remove that average21:54
SpeedEvilcrevetor: Sort-of.21:54
crevetorSpeedEvil: ok21:54
lcukcrevetor, and will this be installable on DocScrutinizer's n950 without any aegis drama?21:54
SpeedEvilcrevetor: The magnetic field changes twice per cycle - or 100hz in europe21:54
crevetorlcuk: yes21:54
crevetorlcuk: or is DocScrutinizer's N950 different from the others ?21:55
SpeedEvilIf you can sample at 50Hz, then the resultant frequency you will get will be something like 10hz21:55
lcukcrevetor, not sure21:56
crevetorlcuk: well in installs without drama on other N95021:57
lcuki am installing it now21:58
lcukthe webbrowser on n950 is very slick21:58
crevetorit is :) Sometimes it explodes on mine though : it says it crashed21:59
*** antman8969 has joined #harmattan22:03
GeneralAntillesWe need an N950 icon22:04
*** crevetor has quit IRC22:05
*** crevetor has joined #harmattan22:06
lcukcrevetor, does the compass just directly read via qt mobility?22:07
*** gri has quit IRC22:11
mgedminlcuk, AFAIU yes, check out
antman8969so does the obs actually let you download a package?22:21
kimjuyes, it just takes a while to get published.22:21
lcukantman8969, normally yes22:21
antman8969is the package located on my project page?22:21
kimjuand unless you have disabled the publishing22:22
lcukit is a building factory, from the source files you provide it generates packages in each selected repository and build type out the other end22:22
lcukantman8969, the links to the destination repositories is22:22
* mgedmin once again wonders about convincing Qt Creator to build FBReader22:22
*** vgrade has joined #harmattan22:23
antman8969was that you looking for a link lcuk?22:23
antman8969or just stopping typing lol22:23
lcukmgedmin, which books are you reading atm?22:24
* lcuk has not read many recently22:24
mgedminare you interested in recommendations, or what I'm currently reading?22:24
lcukmgedmin, conversation22:24
lcukso either :)22:24
mgedminif the latter, David Weber's latest series, because I'm bored and out of better books to read22:24
mgedminif recommendations, I've received most fun from Lois McMaster Bujold, Steven Brust, Jim Butcher (Dresden Files, starting with book 2; Codex Alera was a bit dull in comparison)22:25
mgedminCharles Stross22:25
crevetorlcuk: yes22:27
DocScrutinizerlcuk: the magnetometer can sample up to 1000/7.5 s^-122:28
crevetorlcuk: it uses the QtMobility QML bindings22:28
crevetorI don't treat the values at all : each time the compass sends me an update  I update the position of the arrow22:28
DocScrutinizerlcuk: it can get read out by simple script22:28
DocScrutinizer#!/usr/bin/env python import struct f = open("/dev/ak89740", "rb") while True:        data =*2)        x, y, z, valid = struct.unpack("hhhH", data)        print "x=%d y=%d z=%d valid=%d"%(x,y,z,valid)22:29
DocScrutinizershould run @ 100Hz I'd guess22:30
lcukDocScrutinizer, the qml one looks pretty though :P22:31
DocScrutinizerlcuk: you need to run it by feeding script name to python as parameter, as shebang doesn't work, once again thanks aegis22:31
* lcuk never needed a shebang to run the qmlcompass22:32
lcukor a snake for that matter22:32
lcukbut it does appear not so quick, hopefully that can be rectified22:32
DocScrutinizerI consider eventually running the script >log.txt and feed log.txt to a fft after that22:33
*** CaCO3 has joined #harmattan22:35
DocScrutinizerthe fun part will be to run a realtime 50/60Hz filter against the data from /dev/ak89740 and see which axis shows which value and phase for the 50Hz, then deteriming where the source of the magnic field is and output some user feedback about it22:35
DocScrutinizerand magnetic* MEH22:36
DocScrutinizerI expect sme aliasing for 60Hz grids like USA as well, but that shouldn't be a problem, as we know we are looking for 60Hz, so 100Hz sampling rate is just fine to detect 60Hz22:38
DocScrutinizerthe wire is 90° rotated to the one axis with signal, assuming the other two axis got no signal at all22:41
DocScrutinizerwell 90° from axis with signal, and NOT at 90° of the axis with no signal22:43
mgedmindoesn't Nyquist Theorem or something say that you've got to sample at 120 Hz to detect a 60 Hz signal?22:47
DocScrutinizermeh, I'm fuzzy. The axis that could wrap around the wire if you'd bend it is the one that gets signal, I.E. it's at 90° to the direction of the straight wire, and a tangent on a circle around the wire22:48
DocScrutinizermgedmin: otherwise you get aliasing, yes22:48
mgedminaliasing means something like you detect a 10 Hz signal instead of 60 Hz?22:49
DocScrutinizerbut if you *know* you're looking for 60Hz, the aliasing doesn't matter, you just look for the aliasing products rather than the original 60Hz22:49
* mgedmin feels all clever for understanding this now :)22:49
DocScrutinizerwell, the sensor itself has a probing delay of 7.5ms, the rest is sw to run it as fast as possible22:52
DocScrutinizernote it has no sample rate, it just starts a new probe each time you read from /dev/ak8974022:52
DocScrutinizerI'd guess we could beat this up to >60Hz22:53
DocScrutinizererr 120Hz22:53
*** CaCO3 has quit IRC22:54
mgedminwhoa, infobot is a calculator too!22:54
DocScrutinizereven a somewhat smart one22:55
mgedminyou crashed it!22:55
* SpeedEvil remembers another channel that had a bot that was an interface to bc.22:56
infobotsomebody said 55! was 12,696,403,353,658,275,925,965,100,847,566,516,959,580,321,051,449,436,762,275,840,000,000,000,00022:56
mgedminit was a logical paradox, you know that's a sure way to break computers22:56
SpeedEvilAnd an author that didn't understand that you can write infinite loops in bc.22:56
DocScrutinizerwell, anyway you need to add to this 7.5ms the time to read and write registers via I2C at very least22:59
*** apol_ has quit IRC23:01
DocScrutinizerbut as mentioned, as long as the sampling window is significantly shorter than 1/(2f)23:01
DocScrutinizerand the sampling rate has a sufficently jitter free pace23:01
DocScrutinizeryou would even get away with 2 samples / s23:02
DocScrutinizerthough this would increase detection time quite a bit :-)23:02
lcukDocScrutinizer, have you installed the qmlcompass?23:03
DocScrutinizerbut compass is doing a lot of averaging etc23:04
lcukit looks more like qt mobility is just jittery23:04
lcuklike unsmoothed accelerometer was on n90023:05
DocScrutinizerI think 40Hz was the highest in this QMLsensors app you could set for magnetometer23:06
lcukDocScrutinizer, thanks for highlighting the /dev node though, will have a peek at that next time I open scratchbox23:06
DocScrutinizeralso the Qmagnetometer was broken as in it never yielded a OVL for me, it rather just stopped delivering values23:07
*** Scifig has joined #harmattan23:08
*** GAN900 has quit IRC23:10
DocScrutinizerlcuk: magnetic overflow, sensor saturated23:10
DocScrutinizersee the example on that webpage :-) (how handy) it shows some accuracy or error value as 4th value for QMagnetometer23:12
DocScrutinizeralways been 1 for me23:12
mgedminsmall UI fail23:12
DocScrutinizeryou'd think it should become 0 when chip asserts error, rather it stopped yielding values then23:12
mgedminI swipe a fullscreen app just a little bit to see the status line -- battery status, clock23:12
mgedminonly the information is outdated there23:12
mgedminit shows what the time was when I last switched tasks23:12
mgedminand the battery animation is frozen23:13
mgedminI even thought my USB cable fell out or something23:13
DocScrutinizerbtw you are aware you'd need to run such a electric current detector app with realtime schedule, to make it keep the 7.5ms pace? Wonder if that's possible without another barf up of aegis23:17
*** javispedro has joined #harmattan23:17
DocScrutinizerjavispedro: hi!23:18
DocScrutinizerjavispedro: read last few hours backscroll, you missed all the fun ;-P23:18
DocScrutinizercheers, bbl23:18
javispedrowhat can I say, other than... *sigh*23:22
javispedroI am doing that a lot lately...23:22
achipafor those that care, mc can be build with literally <50 bytes of package-file editing23:25
DocScrutinizerI'm just back again to throw in another remark: N9 magnetometer allegedly is better than N950's23:25
achipa(I just took the source file from maemo, shoved it in SB, adjusted package names and that was it)23:26
mgedmin*sigh* I need to install scratchbox23:26
DocScrutinizerachipa: so it's available for th enoob masses (i.e. me :-D) now, as a .deb?23:26
achipaand I didn't even have to touch anything aegis-related23:26
DocScrutinizercool, that's what it should be :-D23:27
mgedmindo I dare run sudo python admininstall?23:27
* MohammadAG wonders when his N950 will ship23:27
DocScrutinizerI'd like to install it right away and see it run on my N950, would probably make my day23:27
achipaalways dare, a day without reinstalling is a day not lived23:28
javispedrowhat was the problem with the original packaging then?23:28
achipajavispedro: stupid dependency settings ?23:28
DocScrutinizerjavispedro: seems djszapi missed to include manifest23:28
DocScrutinizeror that23:28
MohammadAGwhat manifest23:29
javispedrobut then it was not installable?23:29
achipaI didn't put in a manifest23:29
lcukDocScrutinizer, I installed .deb package built in fremantle sdk which I doubt has manifest file23:29
lcukand it installed/runs correctly23:29
DocScrutinizerjavispedro: indeed browser said "wrong packagin format" OWTTE23:29
javispedrolcuk: NO. WAY.23:29
javispedrolcuk: no digsigs, so you had to run it as user in developer-mode23:29
lcukjavispedro, I have fremantle scratchbox only23:29
GeneralAntillesachipa, can you mention to somebody UI-side that it's impossible to tell chat windows apart in the small thumbnail task switcher?23:29
MohammadAGpython package lcuk?23:29
lcukand just grabbed one of the liqbase .debs I had23:29
lcukno, libliqbase23:30
achipaDocScrutinizer: because it depended on a lib you didn't have23:30
MohammadAGerr, hardfp?23:30
lcukand liqcalendar23:30
lcukidk what harpfp is23:30
lcukhardfp *23:30
DocScrutinizerachipa: I think I got that part :-)23:30
MohammadAGhardware floating point23:30
achipaGeneralAntilles: I can try...23:30
MohammadAGno binary from the N900 can run on the N950 afaik23:30
lcukjavispedro, idk what dev mode is23:30
lcuki just sshed in as root23:30
lcukand dpkg -i23:30
mgedmindev mode is ON by default on our n950s23:30
mgedminyou can go turn it off in Settings, if you dare23:31
javispedronot relaxed-exec.23:31
javispedroit wasn't of mine.23:31
javispedro*on mine.23:31
lcukMohammadAG, well it does23:31
lcukand did23:31
javispedroI keep it: NO WAY.23:31
MohammadAGit shouldn't23:31
DocScrutinizerachipa: so is the mc.deb you built available somewhere?23:31
lcukjavispedro, *shrug*23:31
MohammadAGI second javispedro's saying23:31
lcukwhat I could *not* do23:31
achipaI can upload somewhere23:31
lcukwas scp the later inary over the top23:31
achipathe .desktop file is prolly broken23:31
achipaI really just 1:1 took the maemo sources and compiled23:32
DocScrutinizerachipa: I don't need any .desktop :-D23:32
lcukoh I had to move the .desktop file manually23:32
lcukfrom /usr/share/applications/hildon back a folder23:32
*** wazd has joined #harmattan23:32
mgedminscratchbox installing ...23:32
* lcuk will install another app from .deb now 23:32
* mgedmin wants vim!23:32
DocScrutinizerachipa: actually I prefer it comes without23:32
GeneralAntillesctrl-q will quit MeeGoTouch applications, by the way.23:33
achipathough if you want to sniff around with this you probably want a manifest...23:33
DocScrutinizerGeneralAntilles: :-D see my wiki note on kb shortcuts23:33
*** Scifig has quit IRC23:34
mgedminharmattan sdk installer uses qt and assumes my charset is an ancient Latin-123:35
mgedminthe embedded terminal window shows malformed localized apt messages23:36
lcukok confirmed, n900 binary .debs install fine for me23:36
javispedrolcuk: link.23:36
lcuklibliqbase has a few wonks as per usual23:36
DocScrutinizerlcuk: but do they RUN? (err does it blend?)23:36
lcukjust grab any of mine from maemo23:36
lcukDocScrutinizer, yes23:36
lcukbecause I am working on calendar23:37
MohammadAGhardfp is an ABI break23:37
DocScrutinizerwell, probably fair enough for cmdline stuff23:37
lcukn950 is on the left23:37
MohammadAGlibliqbase is not cmdline23:37
lcukrunning calendar23:37
mgedminMohammadAG, does it affect apps that do not use floating point?23:37
lcukand I am not skilled enough to install multiple sdks23:37
MohammadAGmgedmin, afaik it affects everything23:37
DocScrutinizerMohammadAG: is harmattan hardfp?23:37
lcuki have not touched my scratchbox install in months23:37
DocScrutinizermeh, c'mon23:38
DocScrutinizerwhere's a friggin fremantle .deb to install?23:38
MohammadAGIf MeeGo 1.2 and Harmattan are API compatible (and ABI compatible), it should be hardfp23:38
* lcuk mehs and goes playing in visual studio 23:38
RST38hthe heretic!23:38
javispedroand MohammadAG^^23:39
*** fiferboy has quit IRC23:39
lcukjavispedro, that is osso23:39
lcukdifferent issue23:39
javispedroalso aegis complaining23:40
lcukdoesnt stop it installing though23:40
lcukit got past those and found the dependency23:40
mgedminwas aegis _complaining_?23:40
lcuklibliqbase can be built without needing osso23:40
mgedminit said it was installing something23:40
MohammadAGI'd guess proper installation is $? = 023:40
lcuki had to remove the dependency years ago23:40
mgedminchecksums for the binaries in the .deb, I assume23:40
javispedromgedmin: no digsigs23:41
lcukjavispedro, there is a screen resolution bug which might show up bug 1308423:41
MeeGoBotBug maj, Medium, ---, carsten, ASSI, [n900] Horizontal tearing with xvimagesink23:41
mgedminoh, hey, are there any tools to see which binaries are signed/checksummed/whatever-it-is-that-aegis-does?23:41
lcukbut I worked around that easily by forcing xvideo to use 800*48023:41
javispedromgedmin: in a package, or in the system?23:41
mgedminI'm thinking a command-line tool23:42
lcukbut the point was, fremantle scratchbox built .debs actually install happily23:42
mgedminsomething e.g. aegis-info /path/to/file23:42
lcuksame way as I used to use diablo scratchbox on fremantle23:42
javispedromgedmin: but for the files inside a package, or for all of the files in the system?23:42
mgedminso I can know if I'm allowed to tamper with that file23:42
MohammadAGdiablo and fremantle were both softfp23:42
mgedminwithout trying it and bricking my n95023:42
lcukMohammadAG, well this is softfp then too23:42
MohammadAGjavispedro, does liqcalendar run?23:42
javispedromgedmin: ah, they are stored in /var/lib/aegis/refhashlist23:42
mgedminhow can one distinguish hardfp from softfp?23:43
javispedroMohammadAG: not found23:43
mgedminanything in /proc/cpuinfo? uname -m? dpkg --print-architecture?23:43
MohammadAGlcuk, then how is MeeGo 1.2 hardfp23:43
javispedroI've noted anything though; aegis created digsigs for the package.23:43
infobotjavispedro meant: I've noted something though; aegis created digsigs for the package.23:43
lcukMohammadAG, I don't know23:43
lcukI build from source on meego proper23:43
lcukas in "make; sudo make install"23:44
lcuki have not managed to wrangle obs enough yet23:44
* lcuk likes simplicity23:44
lcuki have build-essential on my n95023:44
lcukI was going to rummage through and use same build commmands on that too23:44
lcukdirect .tar.gz source based distribution would be awesome for liq* ;)23:45
lcuksimple app running on any device to download/make/make install23:45
lcukand not an autobuilder or complex packager around23:45
DocScrutinizerRM680-22-6_PR_RM680:~# dpkg -i findutil.deb23:46
DocScrutinizerfile /var/lib/dpkg/ cannot be opened23:46
DocScrutinizerNo digsigsums file23:46
DocScrutinizerSelecting previously deselected package findutils-gnu.23:46
* lcuk did not think his installing .debs was a big deal23:46
lcukmgedmin, I have learnt a lot since me, you and Gosia were sat in that cafe23:46
mgedmingood times, eh? :)23:47
lcukwell everytime I think of open source, it is your direct conversation sat there that I think back to23:47
DocScrutinizerRM680-22-6_PR_RM680:~# find / -name find23:47
mgedminhelp does anybody know perl?23:47
mgedminI don't, any more, apparently23:47
DocScrutinizerRM680-22-6_PR_RM680:~# /usr/bin/gnu/find --version23:48
DocScrutinizersh: /usr/bin/gnu/find: Operation not permitted23:48
mgedminhm, crap23:48
DocScrutinizerRM680-22-6_PR_RM680:~# ls -l /usr/bin/gnu/find23:49
DocScrutinizerlrwxrwxrwx    1 root     root            27 Jul 22 22:45 /usr/bin/gnu/find -> /opt/maemo/usr/bin/gnu/find23:49
DocScrutinizerRM680-22-6_PR_RM680:~# ls -l /opt/maemo/usr/bin/gnu/find23:49
DocScrutinizer-rwxr-xr-x    1 root     root        192192 Mar 21  2010 /opt/maemo/usr/bin/gnu/find23:49
DocScrutinizersame as it ever was23:49
DocScrutinizerbut yes, it installs23:50
javispedroaegis shows here the maemo findutils-gnu use setgid&setuid23:50
DocScrutinizerhehe, as user it runs23:51
DocScrutinizerprolly related to develsh or aegis-mode --relaxed or whatever23:51
Damion3most of the binaries I'be tried were from the n90023:51
achipagenerally root is not what you want23:51
javispedroI have not enabled it yet23:51
achipathink about it as sudo23:51
achipabut anyway23:51
DocScrutinizerachipa: I'm usually root when I log in via ssh23:52
achipasu - user ?23:52
javispedroor aegis-exec -u user -l ;)23:52
DocScrutinizerthat's what I did23:52
achipasame thing - almost :)23:52
achipaI couldn't be bothered to make a proper manifest23:52
DocScrutinizersu - user23:52
achipaso ignore the errors23:53
achipastart as user23:53
javispedroachipa: manifest for what -- what does mc use?23:53
achipaand that's it23:53
achipajavispedro: nothing. but I made it setuid root. just for kicks23:53
javispedroyou made the mc binary setuid root, and it _launches_?23:53
TSCHAKeeeit shouldn't, unless you've deliberately relaxed privs23:54
Damion3setuid as an aegis capability or just 4755 ?23:54
* javispedro wishes for some of your aegis taming skills23:54
achipadidn't say it works ;)23:54
lcukhas anybody worked out how come maemo5 .debs work?23:54
javispedroDocScrutinizer: did find work as user then on N950 then?23:54
Damion3lcuk: I'm not sure why they shouldn't23:55
achipaDocScrutinizer: DO start it as user23:55
mgedminok I've got scratchbox23:55
javispedroDamion3: hard vs softfp23:55
Damion3oh I've also used armel debian debs of perl binary libs23:55
Damion3the linux kernel does it soft if your cpu doesn't have hardfp23:55
Damion3that's been normal behaviour for decades23:56
javispedrothis is abi we're talking about23:56
achipaalso, for aegis younglings, when launching stuff, check syslog, usually has some enlightening "credential not present" msg here and there23:57
javispedroachipa: which most of the time are setuid/setgid, which we are not allowed to request in a manifest23:58
achipatechnically, it's not that they are not allowed to request... but rather that they don't have the right source...23:58
Damion3surely if libc uses syscalls the kernel can deal, if it's instructions then the kernel can mark as ones to trap on and deal with them.  assuming the abi concerns are in what binaries have with their dealins with libc?23:58
javispedro(not from ""/unknown origin that is)23:58
javispedroconsidering that there are different sets of requestable privileges in developer-mode unknown origin and "normal" unknown origin, would it be too much asking for setuid/setgid?23:59
achipahave fun, off to have fun in the city !23:59

Generated by 2.15.1 by Marius Gedminas - find it at!