IRC log of #harmattan for Saturday, 2011-07-16

*** pabs3 has joined #harmattan00:01
*** spenap has quit IRC00:13
*** NIN101 has quit IRC00:15
*** NIN101 has joined #harmattan00:16
*** mzanetti has joined #harmattan00:17
*** conny has quit IRC00:29
*** deimos has left #harmattan00:43
*** deimos has quit IRC00:43
* javispedro thanks god the n950 seems more robust than n900, because stock cable plug is quite... hard to unplug.00:47
javispedro*the n950 usb port..00:47
*** pabs3 has left #harmattan00:49
SpeedEvilI'd be tempted to do the file thing.00:49
SpeedEvilUse a file to remove the sticky-up bits on the plug00:50
javispedrothe cable seems el cheapo.00:51
javispedrowere the n900 has four "sticky-up bits" this has two00:51
*** Scifig has joined #harmattan01:40
mzanettiModifying a PageStack while it is busy freezes the complete application. Are there already any known workarounds for that?01:53
mzanettiDepending on the property "connected" I replace() the page on the pageStack. If I change a connection the property goes to disconnected and then to connected again.01:54
mzanettiThis calls pageStack.replace() twice and freezes the app01:55
mzanettiis there a way to queue the property change as long the pagestack is busy?01:56
DocScrutinizerSpeedEvil: javispedro: it's pretty unclear what *are* the sticky-up bits, so I'd not do this. Also it "fixes itself" really fast - at least did for me02:03
*** Scifig has quit IRC02:03
*** Scifig has joined #harmattan02:03
SpeedEvilI mean carefully remove after examining closely, and ensure no spurs are left that could stick up02:04
*** DocScrutinizer has quit IRC02:05
*** DocScrutinizer has joined #harmattan02:06
*** DocScrutinizer has quit IRC02:06
*** DocScrutinizer has joined #harmattan02:06
DocScrutinizerSpeedEvil: javispedro: it's pretty unclear what *are* the sticky-up bits, so I'd not do this. Also it "fixes itself" really fast - at least did for me02:08
DocScrutinizersome graphite applied by a smooth pencil may do miracles though02:09
DocScrutinizerNOT on the inside of the plug!02:09
DocScrutinizeranyway I concur the CA-179 is kinda crappy and I prefer the CA-101 each day02:13
DocScrutinizergot 4 of them ca-101 now, should suffice for regular use02:13
DocScrutinizerand yes, none of the 4 CA-101 (un-)mates that hard as the CA-17902:14
*** rcg has quit IRC02:23
* SpeedEvil ponders claiming first 'harmattan game' by adding a link to a locally stored web-game.02:29
*** willer_ has quit IRC02:33
DocScrutinizerWOOOW the proximity sensor on N950 does real distance02:38
DocScrutinizerI mean distance readings02:38
DocScrutinizerstarting with 27.06932 when hand is ~3 fingers distance (4cm)02:40
ieatlintwhat'd you use to access it?02:40
DocScrutinizergoing up to 55.66716 when touching surface02:40
ieatlintnice, yeah, i just went and looked at QProximitySensor and see that it will only return a boolean indicating if something "is close"02:42
DocScrutinizerand QCompass sensor widget isn't as broken as the compass app there02:42
DocScrutinizerieatlint: most of the sensors are abysmally poor02:42
DocScrutinizerby design02:42
DocScrutinizerin QtMobility02:43
ieatlintyeah, it also has instructions on how to get the real reading02:43
ieatlintwith notes that the values are device specific02:43
DocScrutinizerQOrientation: no hysteresis at all, just braindamaged 1..6 for "current orientation is more loke"02:43
DocScrutinizer~dict grue02:47
infobotDictionary 'grue' n. [from archaic English verb for `shudder', as with fear] The grue was originated in the game {Zork} (Dave Lebling took the name from Jack Vance's "Dying Earth" fantasies) and used in several other {Infocom} games as a hint that you should perhaps look for a lamp, torch or some type of light source. Wandering into a dark area would cause the game to prompt you, "It is very dark. If you continue you are likely to be eaten by a grue." ...02:47
DocScrutinizerWTF is GrueSensor?02:47
javispedroomg, that is quite an invention02:48
javispedroprobably the KILLER application of this device02:48
javispedro/me wants02:48
ieatlint"The Grue Sensor informs you of your chance of being eaten by a Grue."02:49
javispedroiPhone does not have that!02:49
javispedroneither crapdroid!02:49
DocScrutinizeryeah but QtMobility has QGrueSensor and I wanna know what it's *supposed* to show02:49
ieatlintit's a demo on how to create your own sensors02:50
javispedro" A probability of 1 means you are currently being eaten."02:50
ieatlintit's attached to the light sensor02:51
ieatlintcover it and your chance increases02:52
ieatlintfor every second it's dark, you have a .1 chance more of being eaten, apparently02:52
DocScrutinizernot where I am - the chance there'snothing left for the next grue is >102:52
DocScrutinizeri.e. the ALS won't notice much of a difference when I cover it02:53
DocScrutinizerI found it rather erratic this sensor behaves like Heissenberg: it always starts at 0 as soon as you look at it02:54
javispedroobviously, because you're alive.02:54
DocScrutinizerthat's rather unusual for a sensor to influence the sensored value02:55
javispedroah, you mean the ALS, not the grue one =)02:55
DocScrutinizerunless you go to quantum mechanics02:55
javispedrobtw the sensors guys also can't properly create harmattan packages02:56
DocScrutinizernah, I mean that gruesensor that always starts counting up from zero as soon as you start it02:56
DocScrutinizerthat's quite obviously nonsense for virtually any kind of sensor02:56
DocScrutinizerthat's why my first question was "what it's *supposed* to probe"02:57
DocScrutinizeras it was instantly clear this sensor is broken02:57
javispedroomg it installed02:58
DocScrutinizeranother broken by design QtMobility sensor02:58
javispedromore like, it started installing, then it hung.02:58
DocScrutinizernot much worse than the others02:58
ieatlintooh, there's a battleship game up there with multiplayer via bluetooth02:59
javispedroI just had that idea today.03:01
javispedroieatlint: where?03:01
javispedroonly harmattan though?03:02
javispedroUNLESS there's another meegoconf soon.03:03
ieatlintyeah, i wonder why... if it uses qt, it should work fine anywhere03:03
DocScrutinizerok, my magnetometer is definitely broken at best03:08
DocScrutinizercould it be? Nokia forgot to use a magnetically neutered speaker component - and vib motor as well?03:10
SpeedEvilhave you found the raw x/y/z outputts?03:10
DocScrutinizersure, javispedro provided a minimalistic python script some days ago, and sensorgallery has QMagnetometer03:11
DocScrutinizerboth showing same results basically: a gigantic offset, and really shaky values that don't strictly correlate with orientation of device wrt to earth03:12
SpeedEvilSeveral things occur.03:13
SpeedEvilYou're not ear any large metal objects - or is the building you're in made from concrete?03:13
DocScrutinizerI guess with massive integration over time, and offset calibration, you might get some meaningful result03:13
javispedroDocScrutinizer: QCompass seemingly tries to provide something useful03:14
SpeedEvilI'd try it in the garden, or somewhere outside03:14
javispedrohere it has around ~10º offset, but is consistent within itself03:14
DocScrutinizerthe building has rather few metal in it - it's 150 years old03:14
DocScrutinizerI'm talking about offset in the 3 axis raw values03:14
DocScrutinizerthe one does never go negative, for example03:15
javispedrobut that's how it should be aiui03:15
DocScrutinizerhow would one axis never go negative?03:16
javispedroah yes, you always do all orientations =)03:16
javispedrodepending on portrait/landscape qcompass has a different offset03:17
DocScrutinizermeans there's a bias on the field that's fixed to the device and way huger than earth magnetic field03:17
javispedroso it's useless as an absolute value but works well enough as a relative03:17
DocScrutinizerI nevertheless wonder if the shaky part in the values maybe already is the effect I'm looking for to base my cablefinder on it, and it was way stronger than I might have thought03:19
DocScrutinizerI might want to dump the data to a file and run some statistics over it03:21
DocScrutinizeresp filters03:21
DocScrutinizeror better even FFT03:21
DocScrutinizeror simply feed it to PA ;-P03:22
SpeedEvilI would try degaussing your device03:24
SpeedEvilAfter turning off ad removig the battery03:24
SpeedEvilDo not degauss the device when it's on03:24
DocScrutinizeranyway if vib motor has a moving permanent magnet inside, the whole compass thing is fsckd03:24
DocScrutinizerand vib motor has an extremely strong magnet03:26
DocScrutinizerat least in FreeRunner03:26
DocScrutinizeryou can *almost* pin the whole phone to a steel board at the wall03:27
javispedroI love how you need to frigging _pin_ the data rate slider to the _EXACT_ rate otherwise it goes back to 0Hz03:28
DocScrutinizerOK - speaker is it here03:28
javispedroin an animation that firstly looks cool but after the 37th makes you want threw phone.03:29
javispedroalso: set slider to 40Hz in landscape, switch to portrait: slider disappears.03:29
DocScrutinizerduh, I thought you simply can not set datarate03:29
DocScrutinizerwhich sensor?03:30
javispedroonly 40Hz works.03:30
javispedro(and 0Hz... which is not really 0...)03:30
DocScrutinizeryou can store steel nails at device backside where speaker sits03:30
javispedroah, and 2Hz.03:30
DocScrutinizerfriggin magnetometer chip has NO datarate at all03:31
DocScrutinizeralso I just pushed the slider to a random position and it worked, unlike all other sensors I tested03:32
javispedroand the random position was 40? =)03:32
DocScrutinizernope, more like 1703:33
javispedrobzzt! nope03:33
javispedro(goes back to 0)03:33
DocScrutinizerincredible, it was 40 ;-P03:33
DocScrutinizerfirst shot a hit03:34
javispedroworst slider widget ever.03:34
javispedro(but it's fun)03:34
DocScrutinizeroh, it slides back to 40 now when I set random 5403:34
javispedrothis sensorsgallery is the most fun app in the phone now.03:35
javispedrointerestingly enough, swiping out the camera app reveals it is not using an overlay color key like videos app03:36
javispedroand yet from a performance PoV I would not have realized.03:37
DocScrutinizeryou might have thought 4th parameter in magnetometer is bool valid, but no, it's always 1 and sensor overload results in readout stalling03:40
DocScrutinizerso I wonder hat that 4th parameer might be03:40
javispedroit is actually bool valid03:41
DocScrutinizerjust another bug. I wonder if there's *one* sensor in QtM that's not flawed or buggy03:43
javispedrodon't blame QtM, at least for the ones that provide raw data03:43
DocScrutinizerbtw if you wondered: the magnetometer seems to sit "upside left" - quite close to the earpiece transducer (unless kbd opened), and "other side" of  main camera03:45
javispedroseems that this is the same magnetometer as used on the playstation move03:46
javispedroso there's some RE around.03:46
* DocScrutinizer wonders if that's a bad sign when he starts to consider portrait the normal orientation of the device03:46
DocScrutinizerdatasheet for magmeter is available. Its actually a funny chip design03:47
DocScrutinizerthe chip needs power-on cmd, does one readout and auto-powerdowns. You nevertheless can readout registers when powered down03:48
DocScrutinizerthat's why I said "there is no datarate of magmeter chip"03:48
DocScrutinizerthe arbitrary restriction in QMagnetometer is random BS03:49
DocScrutinizeractually I guess sysfs node read() blocks until chip completed a probe cycle03:50
DocScrutinizernot as stupid as it sounds, as the chip is really fast03:51
DocScrutinizersome 7ms iirc03:51
DocScrutinizeror was it 0.7?03:51
DocScrutinizertime for measurement: 7.30ms03:54
*** Venemo has quit IRC04:08
*** NIN101 has quit IRC04:11
*** Scifig has quit IRC04:40
GAN900DocScrutinizer, hope we can fix that with some community-based improvements04:52
GAN900Portrait-primary is sucky.04:52
rm_youGAN900: still in UK04:53
rm_youhow do they fail that hard04:54
GAN900fiferboy's appears to have fallen into the ocean somewhere.04:54
GAN900I don't know.04:54
rm_youthey MUST be either failing to scan things in right, or having system trouble04:54
javispedroGAN900: rm_you: it could be worse like that guy in f.m.c who says his order just dematerialized.04:55
javispedrobelieve me when I say the device is not worth the wait!04:56
* javispedro hides while laughing04:56
* javispedro hits aegis again05:00
* javispedro types "ME. WANTS. RPCBIND. !!!!" on terminal to see if omniscient aegis answers05:04
*** antman8969 has joined #harmattan05:11
*** mzanetti has quit IRC05:12
rm_youi thought you could just turn off aegis on the dev machines05:19
*** infobot has quit IRC05:24
javispedrojust watch the twitter feeds.05:32
javispedronoone has so far utterly defeated aegis.05:32
DocScrutinizeraegis is a proper fully armed TC implementation. It's probably easier to install a new jkernel than to neuter aegis in this one05:33
javispedroPackage portmap: denied 'CAP::sys_rawio' -- origin '' does not allow it05:34
DocScrutinizergood one05:35
GeneralAntillesWhat happened to Open Mode?05:37
javispedroin docs dev mode and open mode are often confused05:41
javispedroDocScrutinizer: can you do uname -a on n950?05:55
javispedroor anyone05:55
* javispedro just built zImage- , pondering whether it is the correct one05:55
javispedronah, instant panic05:57
*** trx has joined #harmattan05:58
DocScrutinizerRM680-22-6_PR_RM680:~# uname -a05:58
DocScrutinizerLinux RM680 #1 PREEMPT Wed Jun 1 18:17:45 EEST 2011 armv7l GNU/Linux05:58
javispedroso version is right :(05:58
DocScrutinizerjavispedro: do I get it right you're already tackling kernel building and got to the point where it panics?06:02
DocScrutinizerjavispedro: that's really promising06:03
javispedronot really :P06:03
javispedrothere's a source, a default config, and I typed "make":06:03
DocScrutinizerand that's enoug to make it panic? you installed it with new modules? I guess the hashes might not match anyway06:04
javispedrothey shoulda.06:04
DocScrutinizerit's a new kernel, even if only uname -a has a different date, the hash will change06:05
DocScrutinizerfor a kernel with changed hash the modules won't match06:06
DocScrutinizermy 2 cent06:06
javispedrothere should be no hash06:07
javispedrothere's only version string06:07
javispedrounless it's enabled in config.. hm..06:07
javispedroCONFIG_MODULE_ELF_HASH=y (0 google hits)06:08
DocScrutinizerthat's a really ambitious project to get a new kernel running on that beast06:11
*** javispedro has quit IRC06:22
*** antman8969 has quit IRC06:28
GeneralAntillesDeparted Facility in EAST MIDLANDS - UK EAST MIDLANDS - UK 03:5306:35
GeneralAntillesSat there for 22 hours06:36
*** infobot has joined #harmattan06:49
*** ChanServ sets mode: +v infobot06:49
ieatlintheh, "Failed to connect to remote themedaemon. Retrying in one second." and then repeating indefinitely every second when trying to launch my app06:55
DocScrutinizerremote themedaemon? WTF!?06:55
ieatlintthe best part of working on new platforms is being one of the first people to get errors, so that google searches just give you the source repo as hits06:55
ieatlintno idea wtf that is06:56
ieatlintapparently part of libmeegotouch (this app is qml, qt and qt-mobility only)06:56
ieatlinti actually think it might be a permissions issue06:56
DocScrutinizerhey, never seen it this way - just the source repo, nice :-)06:57
ieatlinti ran into a qt-symbian error message one like that too06:58
ieatlintonly had hits in gitorious06:59
ieatlintyep, was a permissions issue06:59
ieatlinti was being lazy and trying to run it as root06:59
*** antman8969 has joined #harmattan07:00
*** antman8969 has quit IRC07:01
*** antman8969 has joined #harmattan07:01
*** DocScrutinizer has quit IRC07:34
*** DocScrutinizer has joined #harmattan07:34
*** antman8969 has quit IRC09:04
*** Scifig has joined #harmattan09:21
*** antman8969 has joined #harmattan09:21
dm8tbrgood moaning09:25
dm8tbrDocScrutinizer: btw: there is a lxr/mxr, it even features those beloved aegis things, but I think something is broken with the updating, so use with caution.09:36
*** Scifig has quit IRC09:41
*** rcg has joined #harmattan09:42
*** antman8969_ has joined #harmattan10:12
*** antman8969 has quit IRC10:12
*** antman8969_ has quit IRC10:14
*** antman8969 has joined #harmattan10:14
*** Mece_ has joined #harmattan10:32
*** Mece_ has quit IRC10:48
*** achipa has joined #harmattan11:00
*** achipa has quit IRC11:00
*** achipa has joined #harmattan11:00
*** SpeedEvil has quit IRC11:01
*** SpeedEvil has joined #harmattan11:01
*** SpeedEvil has quit IRC11:01
*** SpeedEvil has joined #harmattan11:01
*** laasonen has quit IRC11:01
*** laasonen has joined #harmattan11:01
lcukdm8tbr, timeless was on meego a couple of days ago11:16
lcukand him and lbt got it updating11:16
*** Venemo has joined #harmattan11:26
*** Venemo has joined #harmattan11:26
Venemomorning harmattanites11:28
*** spenap has joined #harmattan11:32
*** gri has joined #harmattan11:37
Venemoanyone wanna try my little app?11:50
Venemo -> says it all11:58
Venemoand here is how to keep aegis at bay:
macmaNnice, got blueman connected to N950 BT DUN almost plugnplay12:10
macmaN(imagine airplug)12:10
macmaNlinuxing teh internetz from teh car!12:10
VenemomacmaN :)12:13
macmaNblueman also worked doing PAN to n900 bnep0 so i gotta give thumbs up to the guy12:14
mecehey how do I autorotate stuff in Harmattan?12:17
*** antman8969 has quit IRC12:18
Venemomece, my app autorotates pretty automatically12:18
Venemowithout me needing to do anything about it12:20
dm8tbrlcuk: I know, I started the conversation12:27
*** gri has quit IRC12:29
meceVenemo, what kind of app is it?12:42
lcukdm8tbr, ++12:45
*** mzanetti has joined #harmattan12:48
macmaNhmm strange thing is n950 doesnt show internet connection when youre connected through bt dun13:20
macmaNin the status panel13:21
Venemomece, a memory game13:23
mzanettiVenemo: nice game :)13:46
mzanettiVenemo: the toolbar seems to be quite redundant...13:46
mzanettiVenemo: there are about 3 different actions but like 10 ways to access them13:46
mzanettiVenemo: and IMO the animation when you score should be tuned down a bit. It seems like flickering13:47
*** gri has joined #harmattan13:48
*** NIN101 has joined #harmattan13:52
rZrVenemo: European Union Public Licence, is this floss ?14:13
rZrVenemo: can i distribuate it in compiled form on my h rep ?14:13
mzanettirZr: see LICENCE file section 2.414:21
rZrlooks good14:22
mzanettirZr: you also need to provide "a machine-readable copy of the Souce Code [...] easily and freely accessible [...]"14:25
mzanettirZr: where can I find the repo to see what else you a re hosting?14:29
Venemomzanetti, thank you14:32
VenemorZr, yes, you can.14:32
VenemorZr, yes it is OSI approved14:32
rZrwhat about fsf ? :)14:35
rZri am trying to rebuild it14:35
griVenemo: What app do you have to test?14:45
Venemogri, [12:24] <Venemo>
VenemorZr, read the license, it's very short, and it's FOSS, yes14:46
Venemomzanetti, the only redundant things are what're in the menu... I'll change that :)14:50
mzanettiVenemo: it also feels strange to have close button in a harmattan app14:50
mzanettiVenemo: and I think other platforms add a close button to each app themselves14:51
Venemomzanetti, lol. I'm annoyed by the fact that there is no close button in apps14:51
Venemomzanetti, so I did the css hack trick and I have close buttons now in some MTF apps14:51
mzanettiVenemo: sure, its just my opinion. I for one like the swipe thing14:52
Venemomzanetti, me too, but I hate to be unable to close apps14:52
mzanettiVenemo: as far as I heard you should be able to close them by swiping from top to bottom soon14:54
griVenemo: My girlfriend loves memory games, I'll have to delete it before she notices it :P14:55
mzanettiVenemo: the row and column settings don't work here but I guess that is work in progress.14:56
griAlso in other applications the "more" button shows the menu and hides the toolbar, here the toolbar stays on bottom14:58
*** mece has left #harmattan15:03
Venemogri, hehehe15:14
Venemomzanetti, I heard, but currently I can't15:14
Venemomzanetti, they don't work if you are already in game, that is correct15:14
Venemomzanetti, stop the game, then set them15:14
VenemoI nead to tweak that in my QML15:15
*** trx has quit IRC15:17
*** Termana has quit IRC15:18
*** lcuk has quit IRC15:18
*** trx has joined #harmattan15:19
*** Termana has joined #harmattan15:19
*** lcuk has joined #harmattan15:19
*** trx has quit IRC15:47
*** trx has joined #harmattan15:52
Venemomgedmin, ping16:10
dm8tbr - would be kind of interesting if there would be a harmattan category or is that what they mean by "MeeGo"16:14
dm8tbrWTH, 2 bible readers on the first page of the MeeGo listing...16:15
Stskeepsyou'd be surprised at how big that market is16:16
Stskeepspeople generally don't want to drag around bibles, easier on the mobile16:17
Venemoisn't the bible obsolete yet?16:19
DocScrutinizerQ-ran readers also very popular16:19
TermanaVenemo, ha ha, going in for the kill there aren't you tiger? :p16:20
VenemoTermana :P16:20
VenemoI would have thought that after a few hundred years of doing that, people would be smarter than killing each other for some stuff in a book... but let's not go OT :P16:20
* DocScrutinizer is slightly puzzled on that common habit16:20
Venemosoo... how do I remap the hw keyboard16:21
VenemoI wanna reassing some keys that currently don't do anything when pressed with Fn16:21
DocScrutinizeresp denying the fact that books are *always* written by humans is something that bewilders me16:22
DocScrutinizerand some of the oldest books have undergone several severe edits and censoring, also by humans with not always honest intentions. This is a provable fact16:24
VenemoDocScrutinizer, yeah16:24
VenemoDocScrutinizer, so, do you happen to know how to remap N950's hw kbd?16:24
DocScrutinizersame way you do that for N900? wild guess16:25
DocScrutinizerwhy? aegis?16:27
Venemomgedmin told me something about it at a tile when I wasn't able to fully focus on IRC... the thing that I remember is one needs to tweak some XML file16:27
DocScrutinizeror have they massively changed the underlying gear16:27
Venemodunno about the details16:27
VenemoI tried the same stuff that worked on N900, that did not work definitely16:27
DocScrutinizerwell there are IRC logs all over the place16:28
Venemohe didn't tell me the details16:28
DocScrutinizerlinux and keyboard - a long confusing and annoying story basically. Each time I'm finished with reading the last manpage/source to find out about some deatils, I already lost the complete picture and have to start all over16:29
Venemothis is true for almost every aspect of Linux16:30
DocScrutinizerI guess input methods have more layers than OSI16:30
DocScrutinizersee [2011-07-16 14:32:05] <DocScrutinizer> 5.3 is epic16:32
*** NIN101 has quit IRC16:32
DocScrutinizerread -s -N2 -t 0.001; read -s -N1 -t 0.001  WOWHAA*cough*HA16:33
*** NIN101 has joined #harmattan16:35
Venemolcuk, have you tried my humble little app yet?16:35
*** javispedro has joined #harmattan16:35
lcukVenemo, i have not.  when i next ssh into n9 i will do16:36
Venemookay, no need to hurry16:37
javispedrovoice navigation in drive!16:37
* javispedro is happy16:37
Venemojavispedro, srsly?16:37
* lcuk is pondering a complex thingamibob on other machine16:37
Venemothat sounds good :)16:37
VenemoIf only we'd have landscape support16:37
javispedrowe have landscape support -- in drive16:38
javispedroand ovimaps fwiw16:38
VenemoI'm talking about homescreens and contacts, music, settings, etc. apps16:38
DocScrutinizermeh, my bash doesn't know read -N116:38
Venemomaps, messages, mail, etc. do support landscape16:38
* javispedro is already totally disregarding the hwkb, too cumbersome..16:39
lcuki think xv windows are being sent through compositor and using yuv->rgb on the powervr16:39
Venemojavispedro, srsly?16:39
dm8tbrcontacts has fun problems if you edit a contact and then edit an entry with the kbd slid out16:39
lcuksad panda because it means liqbase runs only as fast as on the n810!16:39
Venemojavispedro, it is a very nice keyboard... typing with vkb is a lot slower for me16:39
javispedrolcuk: there might be something about this, because videos are slower than on n90016:39
javispedroVenemo: not saying vkb is faster. it's just that hwkb requires two hands to slide16:40
lcukjavispedro, it also encounters bug 13084 when I use 854*480 video mode16:40
MeeGoBotBug maj, Medium, ---, carsten, ASSI, [n900] Horizontal tearing with xvimagesink16:40
Venemojavispedro, indeed.16:40
Venemojavispedro, maybe I used my N900 too much? :P16:40
javispedrohum, tearing.16:41
lcukjavispedro, gatepost alignment error16:42
javispedrolcuk: do you have a liqbase build for harmattan around?16:42
lcukon the n900-ce, it reads 801 pixels per row16:42
lcukjavispedro, only thusfar used the n900 builds16:42
javispedrothen that's quite worse than tearing16:42
lcukhappily they work16:42
javispedron900ce builds?16:43
lcukbut encounter problems16:43
lcukie this tearing16:43
javispedroso you do not have hardfp vs softfp problems?16:43
Venemojavispedro, if he doesn't use floats, why would he?16:43
lcuki did not try to use deb package on my n900-ce16:43
lcuki built ondevice directly16:43
javispedroah, ok :)16:43
lcukso would not know about hard/soft fp16:44
javispedroVenemo: dynamic loader is too stupid to know about whether a function has fp calls16:44
javispedrothat's all it sees, symbol and symbol names.16:44
javispedroso it can either refuse to bind ALL symbols from different FP ABI files, or it can just blindly accept them all16:46
javispedromy guess is that it will do the former :)16:46
VenemoI have no idea16:46
DocScrutinizeryeah, puzzling issue16:46
Venemohowever, I managed to get my game into a bugged state...16:46
DocScrutinizerso, FWIW - N950 charges reasonably "OKish" from non-D+-_short USB wallwarts16:48
javispedroWHAT THE HEL16:49
DocScrutinizercharging battery with 80some mA (when idle and screen blanked), to keep it at ~90%16:49
javispedro# portmap -h16:49
javispedrousage: portmap [-dfv] [-t dir] [-i address]16:49
javispedrorpcbind _works_ today, FOR NO REASON AT ALL!16:49
infobotmessy... err busybox is meant for lean scripting. Regarding all the missing options and immanent limitations (see su) it's not really the interactive shell of choice. A lot of people hate busybox because a lot of system integrators don't understand the difference between busybox and a decent user interactive shell plus unix utils16:49
javispedroI DID NOTHING!16:50
javispedroso it is true there's an omniscient Aegis god.16:50
DocScrutinizerjavispedro: today the day-of-month is a 2^x16:50
* javispedro makes a note to sacrificate some lambs so that God allows me to run more programs.16:51
GAN900Still stuck in the UK16:51
DocScrutinizerWTF GAN90016:51
javispedroGAN900: probably will miss Monday then :(16:51
DocScrutinizerthere are no volcanoes hindering airplane starts in EU right ATM16:52
DocScrutinizerGAN900: curse homeland security - I bet the device has to undergo in-depth inspection for potentially hazardous physical and logical content16:53
GAN900Doubt it.16:54
DocScrutinizerafter all it's addressed at *you* ;-P16:54
GAN900DHL incompetence seems the likelier.16:54
DocScrutinizerDHL incompetence is notorious too, yes16:55
* javispedro removes man of faith hat, puts on his researcher hat and tries to understand what the *** happened with aegis.16:55
VenemoDocScrutinizer, doesn't Aegis count as hazardous logical content?16:55
Venemojavispedro, didn't you set your Aegis to relaxed mode or whatever?16:55
javispedroI was trying to install rpcbind16:56
javispedrothe NFS portmapper16:56
javispedrothis must run as root because it opens ports < 100016:56
javispedroso of course aegis refuses it like there's no tomorrow.16:57
javispedrowell, yesterday. today it ran it.16:57
javispedroalbeit I'm noticing it still does not allow it to open listen sockets.16:57
Mekhmm, I managed to get my n950 in a reboot-loop...16:58
javispedronot hard.16:58
*** aapo has joined #harmattan16:58
javispedrooh, interesting.16:59
javispedroaegis removed portmap user and group from /etc/passwd16:59
Venemoaegis is pure evil17:00
Venemowhat's the whole point of it?17:00
javispedrobut this might be the reason it works17:00
aapoHello, I have N950, but I really do not know what I have signed about NDA. I saw one speculation thread on TMO, but is there any 'official' simple English (or even simple Finnish) list what I can publicly say about it?17:02
lcuk++ Nokia Design, N950 can stand up in portrait mode with the keyboard open17:03
Venemolcuk, lol.17:03
SpeedEvilaapo: you have the NDA - you just don't understand it, or you don't have the NDA.17:03
javispedrolcuk: :D17:03
lcuk-- for location of the USB port to allow making a proper charging dock17:03
Venemoaapo, you can retrieve it from Nokia site17:03
Venemolcuk, doesn't it annoy you to great length when a portrait-only app opens while the keyboard is slid out?17:03
Venemoit disturbs the flow of my work when it does that17:04
aapoSpeedEvil: I clicked same default long text than any others too, but I do not understand it enough to feel comfortable17:04
Venemoaapo, go to the site where you ordered the N950 and you will find it there17:04
aapoVenemo: yes, but I think I do not understand it17:05
DocScrutinizeraapo: not even Nokia understands it, evidently17:06
Venemoaapo, which part?17:06
Venemoaapo, there are two important points in it.17:06
Venemoaapo, -> you play €500 if you lose/destroy/etc it17:07
SpeedEvil= get phone insurance17:07
javispedroalso, look at this IRC log, if there's something we could not say we've probably already said it =)17:07
aapoa) if I get something running on it, can I take video and image and post it on my blog?17:07
Venemoaapo, -> you are forbidden to pubilicize any deficiencies of the device and its software which come from its pre-release nature17:07
Venemoaapo, you can publicize good points, you can take pics, etc.17:08
Venemoaapo, but you can NOT make blog post about how crap it is17:08
aapovenemo, ok. Can I report any challenges I encounter?17:08
SpeedEvilIn order for that clause to apply, wouldn't the device need to be planned to be released?17:09
aapocan I speak about bugs? e.g. in nano, which is not meant for end-users?17:09
Venemoaapo, I think so...17:09
VenemoSpeedEvil, the software will be.17:09
Venemoaapo, not sure. if I were you, I would be very light on mentioning bugs17:09
Termanajavispedro, and that allows anyone else who agreed to the agreement to talk about it, under the clauses that allow public information to be discussed17:10
Termana7. The foregoing obligations shall not apply to any Information which17:10
Termana(a) is publicly available at the time of disclosure or later becomes publicly available through no fault of the Receiving Party;17:10
aapovenemo: ok.17:10
aapowhat is place to report bugs (if they are meant to be hidden of public)?17:10
Termanaaapo, you don't report bugs.17:11
SpeedEvilIn general, they are reportedly not wanting reports.17:11
SpeedEvilAs the software is a _very_ old version.17:11
Venemoit is >5 weeks old17:11
Venemoso they probably fixed your bug by now17:11
aapotermana, how they get fixed then?17:11
javispedroat least for SDK bugs17:11
javispedroalso, device bugs: =)17:12
Termanaaapo, there are several ways to answer that question, but, the same way as if we weren't being given devices - the team that's working on it will need to discover it.17:13
TermanaIf they haven't already and fixed it17:13
DocScrutinizerVenemo: the fork happened quite some time before those 5 weeks you mentioned17:14
VenemoDocScrutinizer, oh, right. you already mentioned once or twice17:14
Venemomewants new sw image17:14
* javispedro thinks that generally, one should not worry much about the NDA unless one works for engadget, and that in fact, bugs might be filled.17:16
aapoanother topic: what is state of 'extras-devel for harmattan'? I understand concept of own PPA, but how about one single place for (needed) community made libraries?17:16
* DocScrutinizer thinks Nokia has no manpower to assign to triaging bugs filed against a months-old fork of an OS that seen ~3000 bugfixes since that fork17:17
javispedroDocScrutinizer: I think you're overstimating the amount of bugfixing work that can be done in a month.17:17
aapoI can see two packages on:
*** piggz has joined #harmattan17:18
javispedroDocScrutinizer: my current opinion is that whatever image is the most recent will not be that different from ours.17:18
DocScrutinizerjavispedro: no estimations in my rationale17:18
javispedroexcept of course it will not have all the "preparations" that have been done for ours17:18
javispedro(search for all those "-dadd" packages)17:18
SpeedEvilUnless they're doing more UI stuff17:18
SpeedEvilthat they don't want to leak to the proles.17:18
Venemojavispedro, "-dadd"?17:19
SpeedEvilWhich seems likely.17:19
piggzhi ppl....who has synced with google over mfe?  im not getting my contacts in17:19
javispedroso, to sum it up: I say file a bug if you want to see it changed before PR1.1 .17:19
javispedroand even if you do, you're probably late already.17:19
*** gri has quit IRC17:19
*** cpscotti has joined #harmattan17:20
javispedroVenemo: in dpkg -l, lots of dadd packages, which as I found while removing the developer edition branding means sth like "developer edition addin"17:20
DocScrutinizerI think there is a way more recent firmware image for N9 and for nokia internals, and we just may hope we eventually will participate17:21
VenemoDocScrutinizer, maybe if we bug some guys from Nokia on this channel, they may leak it to us17:21
javispedroDocScrutinizer: recent, sure. different, on the other hand...17:21
DocScrutinizerjavispedro: konttori clearly said "your version lacks the ~3000 bugfixes we have in the recent image since fork"17:22
DocScrutinizerjavispedro: check #maemo chanlogs17:23
piggzim not keen on just how tich the usb cable the port more secure than the n900?17:26
DocScrutinizeryeah, I linked to the original statement of konttori there - it puts my statement about fork happening much earlier a bit on question, but it says "3000 bugs fixed, so yours is far from reality"17:26
DocScrutinizerpiggz: yes, the port is a thru-hole component and won't come out unless you tear the PCB in two pieces17:28
DocScrutinizerat least that's what I think to have seen, with magnifying glass17:29
piggzDocScrutinizer: thank goodness :D ...... ive always been very careful with the n900, knowing about the problem17:29
piggzthe cable that came with the n900 is a slighly looser fit in the n950, compared to the one with the n95017:29
DocScrutinizeryes, the CA-179 is crap17:30
Venemohm, is it a problem if I've used N900's cable with N950 so far?17:30
DocScrutinizerI prefer the CA-10117:30
DocScrutinizerVenemo: the problem is you didn't notice how crappy CA-179 is17:31
* javispedro 's n950 hostname : "harmy"17:31
VenemoDocScrutinizer, that is right, I did not notice it is crappy17:31
DocScrutinizeryours, or stock?17:31
javispedroDocScrutinizer: yours, stock is the usual Nokia-etc17:31
DocScrutinizerjavispedro: how's kernel building coming along?17:32
javispedrowell ,maybe I should retry what I did yesterday today considering Aegis seems to be in a better mood.17:32
*** smoku has joined #harmattan17:34
VenemoDocScrutinizer, why is it crappy?17:34
javispedroaegis is today VERY workable17:35
javispedroit just granted me CAP::net_bind_service17:36
DocScrutinizerjavispedro: how did yu bribe it?17:38
aapohave somebody already bricked+flashed N950? Does it work? Is this right tool:
*** mzanetti has quit IRC17:41
*** mzanetti has joined #harmattan17:42
*** laasonen has quit IRC17:44
javispedroand... portmapper working!17:50
javispedroI wonder if there's a list around specyfing which credentials a "developer" package is able to request.17:51
DocScrutinizerjavispedro: where do you implant that magic spell?17:55
DocScrutinizeraapo: yes, and yes17:55
javispedro.deb package17:55
DocScrutinizeraapo: and works as advertised17:56
aapoDocScrutinizer: thanks, good to know17:56
DocScrutinizerjavispedro: good question17:56
javispedroDocScrutinizer: the infamous Nokia master public signatures are in /var/lib/aegis/domains17:57
SpeedEvilSomeone should make a distributed public key cracking app. :)17:58
javispedroyou can edit this file in develsh mode17:58
DocScrutinizerjavispedro: aiui there's absolutely no use in editing it, as it probably needs a signature from same private key as the root cert in ROM18:00
javispedroit's not in aegis hashed files list.18:00
javispedroeither way I do not know how to sign packages at all for the time being.18:01
javispedrohashed files list is in /var/lib/aegis/refhashlist18:01
javispedroaiui, this get loaded into the kernel during very early boot18:02
DocScrutinizerjavispedro: a mount is very enlightening18:02
DocScrutinizerrhe lines more to the bottom ;-)18:03
*** NIN101 has quit IRC18:03
DocScrutinizeraegisfs on /etc/ssl/certs18:03
javispedromy passwords are on there.18:03
Venemobut, srsly? why Aegis?18:04
Venemowhy isn't SELinux enough?18:04
DocScrutinizerVenemo: that's a question "beyond the scope of this channel"18:04
javispedrofound it18:04
javispedrofound the list of credentials granted to devs18:05
javispedroit's on /var/lib/aegis/restok/restok.conf18:05
VenemoDocScrutinizer, humm18:05
javispedrofull list:18:06
javispedroso that's all you can grant to a non-nokia package18:06
javispedroeverything else, SORRY!.18:06
javispedrospecially note no setuid/setgid, so forget about daemons that drop privileges that way.18:07
fluxjavispedro, hmm, so sshd works only because it's from nokia?18:09
fluxjavispedro, what does it mean to have ie. UID::user there?18:11
DocScrutinizerI just had a very comfortingly thought: Nokia *always* messed sth up with configuration/packaging/maintenance of their maemo system. So why should they get *everything* "right" with aegis? And it's an immanent property of such a chain-of-trust architecture that one single point of failure compromises the whole security. IOW find a single ill-configured manifest and you "rooted aegis" - once you fully understand how to mess with it in18:11
DocScrutinizerany way you like18:11
javispedroyes, sshd is ""18:11
javispedroand thus has chown, fsetid, setuid, setgid granted18:11
Stskeepsdoes any derived apps have the same? :P18:12
javispedroDocScrutinizer: there's plenty of stuff about R&D signatures, which I'm starting to guess it's the way to enable "open mode18:12
fluxfrom security point of view I guess it doesn't matter a lot. you can just drop all priviledges? or will you still be able to modify files as root?18:12
DocScrutinizereven better: aiui there's no revoke lists or anything like that18:13
javispedroStskeeps: I do not know how inheriting credentials works, but develsh has all of the credentials and I couldn't spawn from it a portmap that did setgid(daemon);18:13
Stskeepsis it just me or is the voices of Drive horribly distorted?18:15
DocScrutinizerhonestly this friggin aegis thing either has to vanish, or somebody @ nokia damn publish a howto for it18:15
javispedroDocScrutinizer: heh, I'm sure someone @ nokia is now saying "you should have payed MORE attention to aegis talks!" ;)18:15
javispedro"but I though you weren't that serious!"18:15
javispedro"YES WE WERE!"18:16
DocScrutinizerI opted for developing on linux, NOT on aegis18:16
Stskeepsdon't worry, we won't have aegis in CE18:16
Venemois Aegis a Nokia development?18:16
Stskeepswe'll just be using the TPM chip to randomly take pictures of you and your loved ones in compromised situations18:16
DocScrutinizerbasically yes18:16
Stskeepsto finance the development team ;)18:16
javispedrothat team will soon be psychopathic if you look at geek's "compromised" situations photos ;)18:17
DocScrutinizerVenemo: the aegis flavour seems a Nokia variant of the larger age old notorious security framework/TC concept18:18
javispedroflux: (UID::user) dunno so far, maybe accessing files owned by user or similar.18:19
javispedrowhat I know for sure is that CAP::* maps to linux capabilities 1:118:19
fluxso, does aegis help? I'm guessing for example the web browser or mail agents won't have sufficiently reduced access permissions to remove the ability to do some damage..18:19
DocScrutinizer""sorry team, no money here, no matter how unpleasant the photos you got. But I can share some of that pizza I stole, do you like?""18:19
VenemoDocScrutinizer, but what is the point of it?18:19
infobotfrom memory, aegis is , or "The purpose of this framework is: ... to make sure that the platform meets the requirements set by third party software that requires a safe execution environment.", or
piggzahh, atleast, that was a monumantal delay...i thought id proken the rss feeds app as it wasnt starting, but 30 seconds later it fired up18:20
* javispedro notes that you _COULD_ implement an android-like permission system -- "would you like to allow RandomHackerApplication to read your address book?" -- on top of aegis18:20
javispedrobut so far, I see no UI for that, and instead if (comes_from_nokia) allow_everything()18:21
DocScrutinizersure, you *could*18:21
javispedromaybe Ovi will have some of that.18:21
DocScrutinizerhonestly I'm NOT willing to become an aegis expert. Gimme a 2 pages mini-howto, with all the magic incantations listed there, and I may reconsider sending back this aegis coffing with the tag "USELESS CRAP"18:22
infobotDocScrutinizer meant: honestly I'm NOT willing to become an aegis expert. Gimme a 2 pages mini-howto, with all the magic incantations listed there, and I may reconsider sending back this aegis coffin with the tag "USELESS CRAP"18:23
piggzim surprised how much looks familar when compared to kinda is good to know some things are shared18:24
javispedrofamiliar wasn't a Gtk+ distro?18:24
javispedrosorry, misparsed.18:24
DocScrutinizeralternatively give me a true developer environment where aegis is *completely* eradicated18:24
Venemocouldn't one just remove it from the kernel and be done with it?18:25
lcuki don't mind aegis18:25
lcukonce things are installed it prevents random single file updates18:25
Venemolcuk, it prevents DocScrutinizer from developing usb hostmode for us18:25
javispedroI _want_ aegis, as long as I'm in control.18:25
Venemojavispedro ++18:25
DocScrutinizerlcuk: you obviously missed last 30min as well as previous chats about how aegis is spoiling every devel's day18:26
lcukDocScrutinizer, find out how to update the sha hash18:26
DocScrutinizerlcuk: NO I WONT!18:26
* Stskeeps yawns18:26
DocScrutinizerI am not in for learning/hacking FSCKNG aegis!18:26
lcuki enjoyed learning about it18:26
lcukand made a deb package in fremantle scratchbox to cure issue18:27
DocScrutinizerthat's your cup of tea18:27
Venemolcuk, maybe you could help DocScrutinizer with it then?18:27
lcukpackage up your file and see if aegis will let it update?18:27
javispedroDocScrutinizer does not even want to build a deb ;)18:27
DocScrutinizerjavispedro tried to build a *stock* unchanged kernel yesterday, flashed it ->instant FAIL, panic18:28
DocScrutinizerso WHAT THE F***!?18:28
Venemothere is a 110k+ LoC diff between Maemo 6 kernel and stock kernel with the same version18:28
fluxdocscrutinizer, so, it booted, but failed to mount root?18:29
javispedroit was stock harmattan kernel18:29
lcukwhat is this "kernel" thing anyway, do I need it, can't you just uninstall it and run harmattan?  :P18:29
Stskeepsjavispedro: deb or manual flashing?18:29
DocScrutinizerlcuk: FU18:29
fluxdocscrutinizer, I mean, if it starts, isn't that like that the game is won already?18:29
javispedroStskeeps: -k -l -b, but built from the src package.18:29
Stskeepsjavispedro: what message did you get?18:29
Stskeepsand were you in developer mode?18:29
javispedro(dev mode) yes, and no message at all other than rapidly blinking led then shutdown.18:30
javispedrothis was stock config, so not even console FB.18:30
*** rlinfati has joined #harmattan18:30
javispedroDocScrutinizer pointed that I did not rebuild the modules, I did not thought that the vermagic had changed,18:30
DocScrutinizerlcuk: sorry, but I'm losing my countenance18:31
javispedrobut seemingly due to aegis kernel has some more "strict" hasing (CONFIG_MODULE_ELF_HASH?")18:31
Stskeepsjavispedro: i'll ask around, fairly sure that's supposed to work18:31
Stskeepsjavispedro: well, at least booting your own kernel and rootfs18:32
javispedrothe key here is "your own rootfs"18:32
javispedroI was trying to boot Harmattan.18:32
*** djszapi has joined #harmattan18:32
Stskeepsright, but even that18:32
javispedro(because I bet N950ce works ;) )18:32
javispedroand that if you installed aegis there, accli -I would say "Device mode: OPEN" :)18:33
djszapiHi! Does anybody know where I could download the libncursesw5-dev package if it is possible ? -> I did not find it here.18:33
Venemoany news about FMTX/FMRX in N950?18:33
lcukDocScrutinizer, it was understood that aegis would be on the device18:33
Stskeepsi'd (personally) enable consolefb just to see what's going on18:33
lcukand the n950-ce build is coming18:33
lcukwhich has no such aegis18:33
lcukand is meego pure18:33
Venemolcuk, it seems that DocScrutinizer doesn't care about meego18:33
djszapiI do not think that is good if something has no security...18:33
lcukdjszapi, I agree, i found out about the install permissions thing quite early on18:34
djszapiThere was an aegis-like smack kernel implementation for meego, but they dropped it.18:34
lcukit is a nice thing  actually knowing that files are only coming from packages18:34
djszapiI think the Harmattan security is the best ever :)18:35
djszapiThat was also the feedback from Android guys.18:35
javispedroI'll say that when I see the "Application wants to open Address Book, allow?"18:35
javispedrountil that, this is more DRM-like than security-like.18:35
dm8tbrI think it has merits as long as you are able to control it if you really own the device18:35
dm8tbre.g. install an own signing certificate that you can use for development etc18:36
rlinfatithe flasher can reboot the n950 if it's in pc_suite (ovi) mode :S18:36
djszapijavispedro: DRM has been completely dropped...18:36
djszapiquite a while ago.18:37
fluxrlinfati, isn't that like a feature?18:37
Stskeepsheh, nice feature18:37
dm8tbrdrm in portable devices is dead18:37
rlinfatiflux, not for me.... i run the all-in-one flasher to extract the firmware.... and it's reboot my phone and flash-it !!18:37
fluxrlinfati, so how is flasher supposed to work, if you don't enter the flash mode?18:38
fluxrlinfati, hm, how is the firmware extracted? from the device?18:39
djszapidm8tbr: I agree with you about that the usability is also very important. However what is on the meego front, selinux, that is far away from usability pov.18:39
fluxsomeone should do a writeup on how selinux and aegis differ from each other :)18:39
dm8tbrdjszapi: selinux is a pain to configure I've heard, dunno if that would work well enough for embedded18:40
rlinfatirun Linux_OCF_22-6_EMMC_RM680-OEM1-916.bin and see /tmp/selfextracted....18:40
djszapiflux: there is18:40
fluxdjszapi, cool! you know the url?18:40
djszapidm8btr: yeah it has been very hard and Linus accepted a new kernel approach :)18:40
DocScrutinizerdjszapi: so WTF do *devels* need a "protected" devise??????18:40
djszapiis this question serious ?18:40
djszapiwell, I would not like to use my data and having the part of other cheats that happened quite soonish with windows systems.18:41
djszapi* to lose my data18:41
DocScrutinizerthere's a devel-mode and a "normal" mode - you'd thing devel mode should just completely KILL aegis18:41
piggzbecuase you need to devel/test on a framework that is similar to the consumer device too18:41
djszapieven if I develop an application.18:41
* rlinfati reading /usr/include/cellular-qt/CellularQt/networkcell.h18:42
djszapiI would never use devel mode that...18:42
djszapi* then18:42
djszapihowever it might kill most of it, though.18:42
DocScrutinizerpiggz: seen my comment one line above about that mode switching?18:42
djszapiWhat is wrong about the open mode ?18:42
VenemoSELinux _can_ be turned off...18:42
piggzDocScrutinizer: yeah, just after i wrote it ;)18:42
djszapithat is the best feature of it :)18:42
javispedrodjszapi: there's nothing wrong, but _how to enter open mode???_18:42
djszapithat is the default...18:43
djszapion the N950 devices.18:43
Venemoif this is open mode, what is the closed mode?18:43
javispedro# accli -I18:43
javispedroCurrent mode: normal18:43
javispedrothis is closed mode.18:43
djszapijavispedro: did you also check the settings ?18:43
Venemoisn't there some kernel parameter or something?18:43
javispedrothe settings is what seems to be called "developer mode", which grants a reduced set of permissions to the "" source (aka developer source)18:44
javispedrobut this set of credentials does not even include setgid.18:44
javispedroand obviously not a way to load kernel modules.18:44
javispedroie I cannot remove the "seal" bit from aegis, and put it in non-enforcing mode18:44
djszapiyou can grant that kind of permission via a manifest file.18:44
djszapiwhy not ? (I could)18:45
javispedrobecause it is "" source, and CAP::setgid is not granted to SRC:.18:45
*** NIN101 has joined #harmattan18:45
javispedroPackage portmap: denied 'CAP::setgid' -- origin '' does not allow it18:46
djszapijavispedro: is aegis-exec available ? (I do not have a device by hand)18:46
lcukone question, if I install a fart app and it gets restricted, does it become trapped wind?18:46
djszapiyou can grant that way permission to the processes.18:46
javispedrolast time I tried that we ended up bricking DocScrutinizer's device18:46
* DocScrutinizer shakes head, deeply depressed about the indiscriminate hype of aegis/SF that some people display here, esp those who obviously haven't really looked into how it works wnd what it means for developers18:46
fluxI think it was explained here that aegis-exec doesn't actually grant any priviledges, it rather drops them?18:47
lcukDocScrutinizer, mjority of developers can install/update/remove packages18:47
javispedrodjszapi: this is what happened when we tried to grant setgid to bash:
* djszapi is checking :)18:47
Venemolcuk, Aegis prevents the SDK from implementing the "skip packaging step" feature (best feature for Maemo5)18:47
Venemolcuk, and this makes debugging my app a very clumsy process18:48
djszapiyeah, you can remove the malfunction entry after reboot, but this is odd.18:48
lcukVenemo, i understand18:48
javispedroDocScrutinizer: btw, how did you install that bash package? from Nokia repo iirc?18:48
djszapiVenemo: not sure what you are doing, but I did not feel anything special while debugging my app.18:48
DocScrutinizerjavispedro: yup18:48
Venemodjszapi, do you know what happens when you click on the Run button in Qt Creator?18:49
javispedroeither way, the fact that we already have two persons that are surprised to see that Aegis is "that restrictive" to us makes me thing there's something that has not been de/activated on our images :S18:49
javispedro*think ;P18:49
javispedro(and should have)18:51
djszapiVenemo: nope, I do not use QtCreator, but someone was complaining about that it worked without QtCreator but not with that. Some dbus related application.18:52
lcukthe calendar whole screen kinetics is nudged by swipe18:52
Venemodjszapi, well, basically Qt Creator builds a .deb package every time I click on Run, which means that I basically lose a minute of my life every time I run my app.18:52
djszapijavispedro: well, meego was always much worse on N90018:52
Venemodjszapi, people claim it's because Aegis prevents just copying over the binary and executing it18:53
lcukVenemo, no you gain a minute to think of the next step18:53
djszapiDocScrutnizier: I think the worse hype is that it is sold as a complete meego system18:53
DocScrutinizerthere is *definitely* no use whatsoever in protecting developers from their own mess. Please stop even thinking about that - rather give us a developer- aka open-mode that's *really* _completely_ open. If anybody feels threatened by this, he's a) no developer, and b) not understanding that he can implement additional restrictions to his liking18:53
Venemolcuk, I know the steps.18:53
lcukVenemo, sure18:53
Venemolcuk, I usually do before I begin coding :)18:53
lcuki say the same thing and at some parts of development18:53
djszapiVenemo: that is actually true, you cannot run your own binary, that is why this SDK is existing....18:53
lcukthat delay between code/test/modify cycle is important18:53
djszapithere is a serious hash calculation (integrity protection) in the background.18:54
rlinfatiany tips to use maps ? i can not login (service unavailable )18:54
lcukand even extra seconds is bad18:54
lcukvs whole minutes18:54
* lcuk really does understand ;)18:54
Venemodjszapi, building a package takes a long time and it ruins the debugging experience18:54
fluxwhat the approach prevents as well if just having the .py on the device and developing it there18:54
fluxcompletely eliminating any delay induced by even copying the file18:54
Venemodjszapi, WP7 is an equally (if not more) closed platform, and still, VS can make my app run on the device instantly18:54
djszapiVenemo: well you should use cmake18:54
djszapiif it is that big issue for you.18:54
Venemodjszapi, while Qt Creator can't.18:55
fluxof course, any programs that can be run through an interpreter can avoid the signature checking..18:55
djszapiVenemo: complain to QtCreator18:55
javispedroflux: I use a NFS mount to mount from device my host computer, so it's like if both computers are the same ;)18:55
Venemodjszapi, I opened a bug. :)18:55
djszapiVenemo: you can also run an own script which updates the binary hash18:55
mzanettiVenemo: I just hit a floating point exception in your app when clicking the play button in your app.18:55
djszapiif it is /really/ that big problem for you.18:55
DocScrutinizerlcuk: who are you to teach us about important cycles in development?18:55
lcukdjszapi, Venemo is highly active hacker and has wide experience developing and packaging cross platform18:55
djszapiOther than that I guess the developer mode should enable you to update binary and run it without packaging18:55
lcukDocScrutinizer, ?18:55
djszapiVenemo: the problem is that you do not have proper source origin18:56
lcuki was talking with Venemo18:56
DocScrutinizer<lcuk> Venemo, no you gain a minute to think of the next step18:56
javispedroaegis-developer-mode --relaxed-exec does enable you to run unsigned binaries18:56
javispedrobut only as user18:56
djszapithis way, you should use the developer mode, but do not forget the image is slightly old on that device. Not sure how much it concern on the aegis implementation18:56
javispedronot root, for some reason (not totally understood yet)18:56
DocScrutinizer<lcuk> that delay between code/test/modify cycle is important18:56
Venemowell the point is18:56
djszapijavispedro: root would not make sense18:57
Venemowhen I make a single line modification in my QML, I want to know INSTANTLY the result18:57
Venemothe required compile time is 018:57
lcukDocScrutinizer, it was conversation and agreement with him, wassup with that?18:57
djszapiroot is not that "approach" like with the general POSIX capability approach18:57
Venemoso the packaging shit is an overhead18:57
Venemoand that's it.18:57
lcukgood for first install :P18:57
lcukand if things change in your package18:57
djszapiVenemo: no it is not18:58
djszapiVenemo: I would not like to have own installation and pollute my system actually. But you have been told how to do it anyway.18:58
Venemodjszapi, so, maybe ask MS how they manage to launch my app instantly and then find a way on the N950 to do the same19:00
Venemowhen I debug it, I don't care for the packaging. if I want to package it, I package it.19:01
aapojavispedro: is SDL on your PPA working on N950? I  just tested sdl-hello, but it is seg faulting for start. Is there something I should use for resolution and flags?19:02
DocScrutinizerdjszapi: who cares what you don't want? if you're not capable of coping with it and implement / follow your own restrictions defined to your liking, you suggest rather Nokia should implement those restrictions for all of us? Very polite, thanks for partonizing me.19:02
javispedroaapo: I cannot really completely build it until nokia-binaries is imported into COBS :(19:02
javispedroaapo: for now, get it from git and built it on your sbox19:03
DocScrutinizerdjszapi: we're still debating DEVELOPER mode here19:03
aapojavispedro: do you mean upstream sdl git or maemo specific?19:03
javispedroaapo: any will work unless you want multitouch or GLES19:04
javispedrothen you can install mine on device19:04
aapojavispedro: ok, very this moment I just want see something on the screen, later GLES+multitouch19:04
javispedroI am looking for some cool demo with GLES2 and multitouch19:05
javispedrolike a water ripple effect shader or sth19:05
javispedroif you find one tell me :)19:05
* javispedro crosses fingers and runs aegis-exec -a CAP::setgid portmap -v -f -d19:11
djszapiDocScrutinizer: what do you mean ? Cannot you turn into "proper" developer mode ? What is your exact issue ?19:12
javispedroDocScrutinizer: confirming, it was me who bricked your device, as running that has caused exactly the same big warning icon.19:13
javispedroharmy is dead19:13
DocScrutinizerHeh congrats, welcome to the club :-D19:14
djszapiVenemo: ask MS about Harmattan ? Sounds weird. Did you try what javispedro said ?19:14
javispedronow, how to unbrick this ? =)19:15
javispedrooh, harmy is alive!19:15
javispedroDocScrutinizer: triple reboot fixed it :P19:16
javispedrostill in closed mode19:16
DocScrutinizerhow cute and smart and cool this aegis is! really great19:16
javispedroseems everything is the same19:16
GAN900I think Amazon--ordered Friday--may beat out the N950.19:17
javispedroGAN900: hey, it did for me.19:17
* DocScrutinizer goes bathroom, puking a bit, then for some nicer times in beergarden. Feels sick with harmattan19:18
javispedrocya DocScrutinizer19:18
djszapijavispedro: I think you can remove the related entry, or did I not boot after that "brick" ?19:18
aapowhat is wrong with my root, it can't kill user's processes??19:19
javispedrodjszapi: for some reason, after running the command you get the big warning icon that was shown on that photo. then I rebooted it via powerbutton, and got big warning icon. Power button did nothing, so I had to hold it for a while until it shutted down19:20
DocScrutinizeraapo: haha, aegis19:20
djszapiaapo: if it does not have the proper credential, it cannot19:20
javispedrodjszapi: turning it back on this third time didn't show the big warning icon and proceeded to normal boot.19:20
aapoAnnouncement: initial version of Tuxpaint running on N950. Not usable yet, screen is flickering, but I can draw with it =)19:21
aapobtw: do we have official channel for announcements?19:21
djszapilet us say, you as a user would not like anybody make anything with your data for instance. Root could even do that,  but this way it cannot.19:21
djszapijavispedro: yes, 8+ secs is the hard reset.19:21
aapoI think I too hate aegis19:22
djszapiDocScrutinizer: why cannot you try to understand the security model ? A user device is not really like a Unix workstation...19:23
djszapiaapo: why do you hate aegis ?19:23
javispedrothose old unix bears then to like, you know, being able to kill everything they want when they become root ;)19:24
infobotjavispedro meant: those old unix bears tend to like, you know, being able to kill everything they want when they become root ;)19:24
aapodjszapi: a) I have used aegis-developer-mode --relaxed-exec, but still it sometimes whine, and I have rebooted phone to get it stopped19:25
DocScrutinizerdjszapi: on back of my N950 is a *large* printing "DEVELOPER DEVICE"19:25
Venemodjszapi, what I said, VS can deploy my app on a WP7 device instantly. so why can't Qt Creator do the same?19:25
Venemoanyway, I don't want to argue on this any further, it's pointless19:25
djszapijavispedro: I would not really like to have a root uesr doing anything with my user data....19:26
djszapijust if I grant permissions for that...19:26
djszapiaapo: please tell me your issue, I try to help if I can19:26
GAN900DocScrutinizer, polish it off yet?19:26
DocScrutinizerI didn't apply for the aegis-fieldtest program19:26
djszapijavispedro: it is not really a unix workstation, network where you have thousand users and you can do everything you want with those users19:26
DocScrutinizerdjszapi: again, who cares what you'd like root to do or not to do - this is DEVELOPER mode and DEVELOPER device19:26
djszapiVenemo: complain to QtCreator19:27
djszapiVenemo: you have been told more times how to do it from console. If that works, really complain to QtCreator.19:27
djszapiDocScrutinizer: Look. I complained about the image version weeks ago, it is really an old image for starter.19:27
djszapiDocScrutinizer: Second, I could run a binary, thus I do not understand the issue yet properly19:27
aapodjszapi: not much more to say. I done needed tricks to get unsigned binaries working, then I left phone for couple of hour and when I continue hacking, it sayd "Error: Permission denied ", Rebooted and it worked again19:28
javispedroaapo: lol, that's exactly what happened to me with portmap19:29
djszapiaapo: this sounds really weird...any scenario to reproduce it ?19:29
javispedroyesterday it did not even want to exec() it19:29
javispedrothis morning I found that it started runnign it atleast..19:29
SpeedEvildjszapi: The issue is that the role, policies, and operation of aegis is unclear.19:29
* javispedro had even thanked the gods for that19:29
SpeedEvildjszapi: Especially the future state for the n9 release to the public.19:29
djszapiSpeedEvil: I did not check the public documentation yet, but isn't there something written how the security model looks like ?19:29
aapodjszapi: I think I do not want reproduce it19:29
javispedroI at least can't reproduce it19:30
SpeedEvilWhat will it prevent doing at all - for example - can we ever put meego on it?19:30
djszapiI am pretty sure there is since Elena hold presentations at the meego conferences and also at the local meetup19:30
piggzhas anyone ocr'd/typed the code from the front of the box?19:30
djszapiaapo: not much to discuss, maybe pebkac then19:30
javispedronow I purge portmap, reinstall, whatever, and it always exec() it19:30
djszapiSpeedEvil: no, you cannot flash your own image with flasher in my understanding19:31
djszapibut this is something I am not sure about yet.19:31
djszapiI tried to update the image with some OEM1 internal image, but it did not work19:31
djszapiI had to switch back to N9, not sure about that bit. It might be that they provide a more advanced flasher or I do not know that yet19:31
SpeedEvilOr especially - given this is likely to get less support than the n900 post-release - how can we work around 'features' in the release image.19:31
Venemodjszapi, as I said, I already opened a bug :)19:32
djszapiVenemo: please do not blame aegis, if it is a creator bug.19:32
djszapiSpeedEvil: to be quite honest. MeeGo never worked for me on N900, not even this ce edition. I know people do not like hearing it, but this is the fact.19:33
Venemoworked for me last time I tried. it was incomplete, but hey.19:34
SpeedEvildjszapi: Meego-ce has had perhaps three orders of magnitudes fewer man-hours put into it.19:34
djszapiVenemo: we have few games and it was not enjoyable at any time, not even few weeks ago I last tried. I really spent serious time with meego testing on N900.19:35
SpeedEvilAnd I'm not talking about the state of meego at n9 releast time even.19:35
javispedroI want to try n950ce and see if the 1GiB RAM helps19:35
SpeedEvilI'm talking about it a year or so after that.19:35
djszapiSpeedEvil: How knows what XXX thinks in one year in this market ? :D :D19:36
djszapi* who19:36
SpeedEvildjszapi: I know that if I scrape up the cash for a n9, which is debatable - I am unlikely to change it in under a couple of years.19:36
SpeedEvildjszapi: I also know that meego will at least somewhat continue development, even if dropped by all corporate sponsors.19:37
DocScrutinizerdjszapi: the point is this is a DEVELOPER DEVICE, with two modes selectable from settings: developer mode ON|OFF. We don't need nor want nor see the use of aegis intercepting *anything* in DEVELOPER mode. We want to develop, not learn aegis - we'll cope with aegis and manifests and packaging .deb and whatnot, when we finished developing our app. We go switching developer mode=off in settings then, and do whatever needed to make aegis19:37
DocScrutinizerhappy. We really do NOT want to do this *first*, prior to even starting development of our app. Bottom line: the developer mode is way too restrictive, it ideally should just switch off the whole aegis thing19:37
lcukVenemo, did you say earlier you are using wp7 also? which device are you testing on?19:37
*** lawl0r has joined #harmattan19:38
Venemolcuk, I have no device currently, just the emulator19:38
lawl0rjust recieved my n950 :D19:38
lawl0rso awesome19:38
lcukyou said earlier that your code just runs almost instantly on device19:38
djszapiDocScrutinizer: do not replicate yourself, I understood your purpose. What is not working ?19:38
lcuk(comparison to n950 qt creator slow way)19:38
Venemolcuk, because it works with devices too.19:38
djszapiSpeedEvil: I agree with you, but this is seriously impossible to plan with one device for years19:39
lawl0rbut compared to the n900 the n950 is shit19:39
Venemolcuk, I did try WM6.5 dev with running my app on a device, and that was fast too19:39
SpeedEvildjszapi: Umm - what do you mean it's impossible to plan?19:39
Venemolcuk, maybe I'll borrow a device for testing19:39
lawl0rthere is only one cool thing about the n950: aegis19:39
DocScrutinizerdjszapi: *sigh* see backscroll of this very channel, you'll find a dozen things that didn't work same way they would without aegis19:39
djszapiSpeedEvil: Well, I do not think I can buy any good phone, that will be supported so long in the current state19:40
lawl0rnokia finally realized how to implement a proper security system19:40
lawl0rwas about damn time19:40
DocScrutinizerdjszapi: basically NOTHING works same way as on a system without aegis19:40
SpeedEvildjszapi: The key aspect - 'is it possible to install a new OS or drastically modify the OS on an end user phone' is answerable.19:40
djszapiDocScrutinizer: first, it was a QtCreator issue19:40
SpeedEvildjszapi: Expecting support from nokia - of course not.19:40
javispedrobtw, you also can't place files in upstart dir (/etc/init) unless they're digsig'd19:40
djszapiSecond, aegis blaming without proper steps how to reproduce, thoes things I saw19:40
djszapinone of them is serious for consideration19:40
DocScrutinizerdjszapi: killing user process when you're root is NOT a QtCreator issue19:40
Venemodjszapi, DocScrutinizer is not talking about my little "issue" here...19:40
SpeedEvildjszapi: I would be astonished  if there was more than one bugfix release of software - post release.19:41
djszapiDocScrutinizer: You are really not going toward the purpose. If you do not need aegis, turn it off. Doesn't that work or what is your issue ?19:41
djszapiWhy do you try to use aegis if you do not wanna instead of turning off telling me if it does not work ?19:41
SpeedEvilYou cannot turn it off.19:41
DocScrutinizerI wouldn't even know how to turn it off19:41
DocScrutinizerdjszapi: that's the whole fuss here19:42
DocScrutinizerPLEASE let us turn off aegis!19:42
djszapiSpeedEvil: I cannot seriously answer the question about the Nokia policy... I do not read hype either. Ask some manager :)19:42
SpeedEvilSuggesting 'turn it off' - when it doesn't seem possible to turn it off isn't especially helpful.19:43
SpeedEvilAnd I understand that policy questions are difficult.19:43
lcukcan it be apt-get remove'd ?19:43
DocScrutinizerdjszapi: I (or we) understood that developer mode was supposed to turn off aegis - obviously it does NOT19:44
lcukwhen was it documented that developer mode would do that?19:44
DocScrutinizerlcuk: when did you learn to post this type of trolling questions?19:45
lcukand do we actually know what effect it does have?19:45
lcukDocScrutinizer, i ain't trolling19:45
lcukjust trying to understand it19:45
*** rlinfati has quit IRC19:45
*** djszapi has quit IRC19:46
*** djszapi has joined #harmattan19:46
DocScrutinizerlcuk: ok then: no it's never been stated that developer mode would really do this, and no we don't know exactly what parts of the system will act up in a true "open mode" where aegis is down or neutered. As there's no howto about the whole friggin thing19:47
djszapilcuk: not sure you got all my messages. My mobile internet connection is really bad :(19:47
djszapiDocScrutinizer: You are really confusing the open mode with the developer mode19:47
DocScrutinizerlcuk: nevertheless e.g frals suggested there'S a way to switch off aegis that will not work on and will brick our "CE devices"19:47
DocScrutinizerdjszapi: maybe I'm confusing it, but that's due to missing docs about the whole topic. All i'm interested in is "how do I turn the system into a mode where things work like 'usual' 99.5% of the cases?"19:49
djszapiDoes anybody know how to get this package from the public Harmattan repository or from somewhere else ? libncursesw5-dev19:49
djszapiI do not see it here, at least:
djszapiDocScrutinizer: If you do not get answer, I try to come back to you on Monday. I cannot do more. If turning off aegis does not work, then it does not work...19:50
DocScrutinizerdjszapi: fair enough, thanks19:50
javispedrodjszapi: I believe it's the same src package as ascii ncurses19:51
javispedroaka "ncurses"19:51
djszapino, it is not19:51
* javispedro was 80% sure :P19:51
* javispedro checks19:51
djszapilibncurses5-dev and libncursesw5-dev are different packages19:52
javispedroyes, but build from the same source19:52
javispedroit's just a define19:52
Venemodjszapi, since it's a -dev package, could you not try to look in the sdk repo?19:52
djszapias far as ncurses and ncursesw are not the same.19:52
djszapiVenemo: I tried here:
VenemoI see19:52
djszapijavispedro: I do not really feel nice building from source, if it is not possible. Hence why I am asking first :p19:52
djszapi* if it is possible19:53
djszapicool thanks :)19:54
djszapiDocScrutinizer: have you checked with flasher the gadget is in developer mode ?19:54
djszapialso the developer flags ?19:55
javispedrohooooooow? :)19:55
javispedroit's the old plain rd mode?19:55
djszapijavispedro: ?19:55
javispedrohoow to check that19:55
djszapiuse flasher19:55
javispedroiirc, it was _not_ in R&D mode19:56
javispedrois R&D mode == aegis open mode?19:56
DocScrutinizerdjszapi: just note that I'm not exactly asking for "turning off aegis", I'm just interested in aegis allowing random stuff, like e.g. killing user processes when you're root, editing arbitrary files when you're root, generally doing things each devel is used to do since decades. And I understand that some parts of the system might cease to work when aegis is set to this operation mode, whatever it is called "aegis off" or "aegis has a19:56
DocScrutinizerpolicy to allow all in docS-mode" or no matter how it'S achieved19:56
djszapiDocScrutinizer: no, turning off as a root is against the aegis design19:56
DocScrutinizerdjszapi: no, I didn't know about that19:56
SpeedEvil'turning off when magic switch is pulled' then19:57
djszapijavispedro: no it is not as I said above...19:57
SpeedEvilI don't think anyone especially cares if it's as root.19:57
djszapiDocScrutinizer: you cannot edit files of course, that would result no integrity protection. What you want is to turn off aegis if this design is unacceptable for you.19:58
djszapior you canusethe aforementioned cli command for double checking.19:58
javispedrodjszapi: note that even the nokia wiki pages confuse open and developer modes, but I was talking about r&d mode here (which aiui is another "mode" not aegis related in any way)19:58
djszapiDocScrutinizer: However you can grant permissions from the manifest file. Again, if this granulraity design is unacceptable for you, please do not use security on your phone.19:59
djszapiSpeedEvil: yes I do.19:59
DocScrutinizerI'd happily follow your advice, just have no clue HOW to "not use security on my phone"20:00
djszapijavispedro: yes, that mode is related to the integrity protection imo.20:00
djszapiDocScrutinizer: write manifest files if you want to spread the credentials as you want.20:00
SpeedEvildjszapiI mean - I don't think anyone especially has the opinion that root should be able to do anything - and would mind if aegis was turn-offable with a magic incantation.20:00
javispedroDocScrutinizer: "please open nearest trash can, carefully drop phone into trash can device..." ;)20:00
djszapior there is also a package which auto-generates things for you.20:01
javispedroand that package is? =)20:01
javispedrobut but but.20:02
djszapinot sure it is enough for you.20:02
djszapifor fine granularity though20:02
javispedronope :(20:02
djszapibut for basic things, it is fine to forget the manifest writing things20:02
*** mzanetti has quit IRC20:02
javispedroe.g. it is not smart enough to map setgid() -> CAP::ssetgid20:02
*** mzanetti has joined #harmattan20:03
javispedrobut we could fix that20:03
djszapiit is very smart and overthought20:03
Venemois Aegis a kernel module btw? or how does it work?20:03
javispedroeither way that is still part of the packaging process which Doc et all want to avoid.20:03
DocScrutinizerVenemo: partially, yes20:03
djszapiVenemo: I do not know what information is public, what not...20:03
javispedrothe entire aegis is open source.20:03
javispedrowell, or most of it.20:03
djszapiVenemo: please read the documentation, that is the safest for me :)20:03
djszapiVenemo: but basically *in general* security frameworks, like smack, selinux, aegis, etc have kernel parts for sure20:04
Venemohow about 'modprobe -r aegis' then?20:04
djszapino, it is not open source....20:04
djszapinot even "most of it".20:04
javispedrodjszapi: ?20:04
djszapibtw, can you download the kernel ?20:04
djszapiI did not actually know20:04
Venemodoesn't that violate the kernel's GPL license?20:04
djszapibut that was my impression the kernel is internal and not open20:05
djszapiI might be wrong with this.20:05
javispedroyeah, kernel is open, they're not _THAT_ evil ;)20:05
SpeedEvilCan't (legally) be.20:05
djszapijavispedro: then again *sigh* it is a qtcreator bug20:05
djszapiyou /can/ run binaries....20:05
javispedrobut we have several problems here.20:05
djszapijavispedro: the kernel is open really ?20:06
javispedroprobably 5 or 6 "situations that were possible pre-aegis and now aren't" have been mentione20:06
djszapiI thought they will not publish it, just binary, mmm20:06
javispedrodjszapi: all of it that I can see.20:06
Venemodjszapi, that is against the kernel's license20:06
SpeedEvilYou're required under the GPL licence to release kernel source. In general this is dealt with by having an open-source kernel part that talks through a public API to a closed-source userspace part.20:06
*** NIN101 has quit IRC20:06
djszapiVenemo: is also the android kernel open ?20:06
SpeedEvilIn theory.20:07
javispedroAndroid manufactures are way more sloppy.20:07
djszapiSpeedEvil: there is no closed userspace part....20:07
javispedroand being based in $REMOTE_COUNTRY hardly sue-able.20:07
Venemodjszapi, yes.20:07
SpeedEvilSome vendors are better than others at keeping kernel sources up-to-date20:07
djszapiif the kernel is open, then 100 % open20:07
djszapiat least the aegis part of the platform.20:07
djszapiSpeedEvil: that is almost impossible.20:08
javispedroif kernel had been closed you'd been hearing from me. Probably with a fitchfork ;)20:08
SpeedEvildjszapi: What is almost impossible?20:08
djszapijavispedro: ok, I am not a lawyers, I trust you :)20:08
djszapiSpeedEvil: to be in sync with upstream.20:08
djszapiotherwise you could push your changes to upstream anyway...20:09
djszapiif Linus and others accept it for sure :)20:09
SpeedEvildjszapi: I mean keeping the released sources in sync with the shipped devices.20:09
SpeedEvildjszapi: Not getting 'vanilla' kernel to work - that's impossible in many cases.20:09
*** NIN101 has joined #harmattan20:10
djszapiSpeedEvil: we merged the upstream security fixes back as much as it was possible.20:10
*** smoku has left #harmattan20:11
SpeedEvildjszapi: yes - that's not what I mean - some vendors release source only months after - or not at all - they update the device kernel.20:11
djszapi20:06 < javispedro> probably 5 or 6 "situations that were possible pre-aegis and now aren't" have been mentione -> There is no 5-6 things here. 1) Turn off aegis 2) Not turn off and learn20:11
javispedrountil anyone knows how to _temporarily_ disable aegis, that is not a solution.20:12
javispedroand as for 2), well, someone needs to put on a wiki page to explain all of the "alternatives"20:13
djszapijavispedro: I am not speaking about implementation details. If something is missing or buggy, provide me steps how to reproduce. I got zero. I am more like speaking from the design pov.20:13
javispedronoone is saying they are bugs20:13
javispedro(well, save for maybe that weird stuff that required a reboot)20:13
djszapiDo not forget most of the people left Nokia, there is not much manpower over there right now though for fixes.20:13
javispedrobut still nuisances.20:14
javispedroand imagine now for a moment the amount of _users and developers_ that are going to leave Nokia if they suddenly come to the realitzation that you cannot have real root on the device, or that you cannot package a daemon that requires setgid without a signature.20:14
javispedrohave to say that includes me.20:15
SpeedEvilIndeed - it'd make me really hesitant to plop down cash for n920:15
SpeedEvilI question how many users that is though.20:15
djszapijavispedro: you do not still understand what two ways exist20:15
djszapiif you /want/ "real" root, you want to turn aegis off20:15
Jaffadjszapi: I think javispedro is very much aware.20:15
Venemodjszapi, so as I understand it, we basically need to learn how to write this "manifest file" or whatever thing to allow our packages to do certain things?20:16
djszapiback to the meego and CE thingy. I would really not recommend that way, if aegis causes issues for you guys. As for me, it seems Intel is on the way to ship selinux which is *MUCH* worse than aegis20:17
djszapiVenemo: nope20:17
djszapiaegis manifest is file based, not package-wise. Even more grnaularity20:18
Venemodjszapi, sure, but I need to package up that manifest, don't I?20:18
djszapiok put it mildly: with meego it will be way much more awkward and nightmare according to the principles.20:18
VenemoI'm currently talking about Harmattan20:18
VenemoDocScrutinizer, could you please elaborate on why the CA-179 is crap?20:18
djszapiI am just explaining why meego ce does not make sense for me. If they will really ship selinux, I am out of meego completely.20:19
djszapiVenemo: manifest is a simple xml format20:19
Venemodjszapi, could you please point me to a doc about it?20:19
Venemodjszapi, also, what do you have against SELinux?20:20
javispedromanifest only works if you want to request one of the following credentials:
javispedrootherwise, you're SOL.20:20
djszapiVenemo: as I said more times, nope. As said I have never read public documentation, I do not know what they published. It is not my call20:20
djszapijavispedro: not really, no20:20
javispedroso, how?20:21
javispedroI've been able to request any of the credentials from that list.20:21
javispedrothat includes net_bind_service so I can at least open ports < 100020:21
javispedrobut not setgid, or setuid, or ..20:22
djszapijavispedro: my impression was that you can pretty much provide and request custom credentials as well20:22
djszapihowever it might be turned on that level, it is just for the platform internally, this is something I am not sure about. Someone should write a test manifest for that20:22
Jaffadjszapi: Remember that what Aegis as a framework allows and the Harmattan implementation of Aegis aren't the same20:22
djszapiJaffa: who said it is the same ?20:22
*** mzanetti has quit IRC20:23
djszapiwhy would anybody say that ?20:23
Jaffajavispedro: Which makes sense - if you were given setuid/setgid rights, you could bypass it20:23
*** mzanetti has joined #harmattan20:23
DocScrutinizerdjszapi: >>if you /want/ "real" root, you want to turn aegis off<< Please give clear instructions on how to do exactly that20:23
Jaffadjszapi: "custom credentials" aren't relevant to the concrete realisation, right?20:23
djszapijavispedro: aegis-exec could grant me those, setuid, setgid20:23
djszapiDocScrutinizer: Have you checked the developer mode and flags ?20:24
djszapiJaffa: they are.20:24
djszapiThat is the whole idea about...20:24
djszapion the top of the general POSIX capabilities.20:24
DocScrutinizerno, I don't want to check anything and guess anything. I need a clear word "do this: x, then y"20:24
djszapiDocScrutinizer: I cannot help then, sorry (right now)20:25
javispedrodjszapi: remember that aegis-exec -a CAP::setgid results in a boot loop ;)20:25
GAN900Where is our Aegis expert, anyway?20:26
javispedroHave to say that GAN900 has said the smartest thing of the day.20:26
javispedrothat's what is missing here.20:27
javispedroso, we might have been ... not really concentrated during the aegis meegoconf sessions ... potentially because no one thought of the consequences. Bring back the lady and make her talk again.20:27
SpeedEvilIn a way, aegis is irrelevant20:28
djszapijavispedro: I cannot reproduce...20:28
SpeedEvilIt's the eagis policies that's important20:28
SpeedEvilSeemigly the CE releases differ.20:28
GAN900javispedro, well, haven't things changed a lot since then?20:28
javispedroSpeedEvil: indeed - but also generic stuff like develoepr vs closed vs open20:29
DocScrutinizerSpeedEvil: CE devices are missing a Nokia-developer-cert or sth20:30
djszapiWindows CE ? :)20:30
MekCE labelled devices20:31
djszapiI was just joking ;)20:31
javispedroSpeedEvil: by CE you mean DE?20:31
djszapiwell, in my opinion, the security has nothing to do with this issue. If Nokia managers thought they give out devices without documentation, proper images etc20:31
DocScrutinizerno, some obviously nokia internal term referring to "CE" cert of hw, which correlates with the missing cert20:31
javispedrothe nokia policy is theoretically described in /var/lib/aegis/restok/restok.conf20:34
DocScrutinizerdjszapi: well, yes. It seems like some of the higher-up managers at Nokia are doing the whole meego-N9 thing in a rather half-arsed way20:34
javispedrothere you can find stuff like what privileges can request.20:34
djszapiDocScrutinizer: I cannot help sorry. I will be asking my colleagues on Monday who are still at Nokia and not on vacation. Hope the best...20:35
djszapijavispedro: that is not just nokia policies, any credentials.20:35
djszapilike a "database".20:35
javispedroyes, aiui it is edited everytime dpkg-aegis parses a _aegis manifest20:36
djszapithat is a good debug interface whether or your manifest credential landed on the device properly.20:36
djszapialso accli though20:36
djszapiit is called "dpkg" and the original is "dpkg.real"20:37
djszapior did it change recently ?20:37
DocScrutinizerno, still there20:37
djszapiDocScrutinizer: btw, I know how we turned off the security, I am not sure it works publicly though20:38
djszapiecho 0 > /sys/kernel/security/digsig/enforce and echo 0 > /sys/kernel/security/digsig/enabled, but it seems to not work anymore20:39
DocScrutinizerdjszapi: that's probably what frals also told us. Echo 1 >/etc/die-aegis-DIE etc20:39
Mekyou can't do that in CE labelled devices anyway20:39
Mekyou can only do that if you already have put your devicein open mode20:39
Venemo:what is a "CE labelled device"?20:40
Mekand how to do that is completely secret and/or not possible with these CE labelled devices20:40
DocScrutinizerVenemo: a device with the letters "CE" written somewhere on the case20:40
VenemoDocScrutinizer, I don't recall seeing that on the N95020:42
djszapiDocScrutinizer: nope, "/etc" is integrity protected20:42
GAN900DocScrutinizer, high-up managers screwing things up has been the issue from the start.20:42
javispedrodjszapi: nope, I was just calling the aegis one "dpkg-aegis"20:42
MekVenemo: it's behind the screen I think20:42
djszapijavispedro: yeah, I have just double checked, it is still dpkg and dpkg.real20:42
*** vgrade has quit IRC20:44
VenemoMek, and what does it matter?20:44
DocScrutinizerVenemo: see my explanation above20:44
Mekwell, the label itself doesn't matter, but the label being there means that for example R&D certificates are not prsent20:45
DocScrutinizerVenemo: [2011-07-16 19:31:41] <DocScrutinizer> no, some obviously nokia internal term referring to "CE" cert of hw, which correlates with the missing cert20:45
Venemook, thx for explaining20:45
Mekso you can't flash R&D image (if you would have access to them), and do a couple of other things20:45
djszapithat is why it does not have newer images where serious certificates are needed, like skype, jokuspot, gtalk etc20:45
djszapiMek: you cannot even do that with production image.20:45
Mekwell, nightly build production images are available signed with the right key20:46
djszapiuntil they say, okay, this is good for shipping.20:46
Venemoso how does a CE label prevent me from flashing an image?20:46
djszapiMek: nope, for N950 you cannot use "simple" production image20:46
MekVenemo: images are signed with a key, if te key doesn't match a cert on the device, you can't flash the image20:46
djszapiit must be OEM1 certified etc etc etc etc20:46
Mekdjszapi: yes, OEM1 certified nightly builds for PR images are made20:46
javispedrofun stuff surely :)20:47
djszapithere is no available OEM1 nightly build, not even for internals when I last checked20:47
Mekdjszapi: there has been for a couple of days20:47
djszapiMek: for public ?20:47
Venemoso, can we get our hands on it?20:47
* javispedro removes setgid() calls from portmap and hopes noone decides to crack my n950.20:47
Mekdjszapi: no, internal; and only of PR images20:47
DocScrutinizerhmm, I always thought the purpose of TC/security-framework/aegis was to allow apps like skype to check if device is in open mode and refuse to run if it was. Never thought the purpose was to forbid open mode completely20:47
* djszapi is checking...20:47
djszapiahh I do not have access, well I will check that out on Monday.20:48
djszapiMek: I checked it after the meetup when the devices were given out, few weeks ago.20:48
djszapiweek 22 is a bit old without flash, skype, hotspot, gtalk etc20:49
*** antman8969 has joined #harmattan20:49
Mekdjszapi: yeah, they weren't available then, an interal blog post of earlier this week mentioned they would start making them (and are making them now)20:49
djszapiMek: mmh, they may have seen my screaming post about it on the forum xD20:49
djszapithat is a good progres, cool.20:49
djszapiMek: kdelibs compilation is 92% for harmattan :)20:50
Venemowhen can we get our hands on a new N950 image?20:50
javispedrodjszapi: well, many thanks for the conversation.20:51
DocScrutinizerVenemo: this question is utterly useless20:51
DocScrutinizerVenemo: Nokia NEVER promises any release dates20:51
djszapino problem, I do really understand it can be an issue. Not trivial framework, but it cannot be by purpose.20:51
djszapiI have never seen any trivial security framework. However turning off should be trivial and well documented, yeah. I will ask around on Monday.20:51
DocScrutinizerdjszapi: MUCH appreciates. Many thanks!20:52
* javispedro nfs-mounts his home on 95020:55
javispedroaaaaah, home sweet home.20:55
* javispedro is happy.20:55
VenemoDocScrutinizer, I don't need/want it to be an official release20:59
DocScrutinizerthere's no inofficial releases20:59
djszapiVenemo: I was complaining about it as well on the forum after the meetup, you can join crying there :p21:00
DocScrutinizergood advice :-)21:00
javispedrook, for some reason I just tried to swipe a terminal window. On my desktop.21:01
javispedroWith the mouse.21:01
javispedrolike Fremantle's tap outside dialog, swipe also grows on you seemingly...21:02
djszapijavispedro: how about zooming with two mouse like a pincharea ? :)21:02
javispedroI could use two mouses... :P21:03
Venemodjszapi, ok, I will, gimme a link21:04
* javispedro out -- cya!21:05
DocScrutinizerdjszapi: I tried to zoom pinch on N900 browse just 2 days ago :-)21:06
Venemodjszapi, already pressed the thanks button :P21:08
*** javispedro has quit IRC21:09
djszapiVenemo: :) What type of application do you develop, if I can ask ?21:17
Venemodjszapi, various; whatever comes to my mind21:20
Venemodjszapi, right now I'm making a memory game to learn QML with :)21:20
antman8969anyone know any designers interested in drawing me up an anna -style icon? ;)21:21
Venemodjszapi, wanna try?21:21
Jaffaantman8969: javispedro's icon generator to the resuce21:21
antman8969haha, do you have a link?21:22
antman8969well... that makes my life a lot easier21:23
djszapiVenemo: sure when I got kdelibs working :) Actually we have been doing a game development framework, but Nokia also started providing something. Not sure you need to write it from scratch.21:23
antman8969thanks Jaffa21:23
Venemoantman8969, also see
Venemodjszapi, it is a very simple app which I ported from QGV21:24
GAN900Still in the UK.21:25
GAN900Lame stuff21:25
antman8969are you from the US?21:26
antman8969me too... n950 has been in the UK for 3 days now21:30
antman8969I'm thinking there is a row boat that leaves the UK once a week, and thats what dhl is waiting on21:30
JaffaDHL aren't always 100% accurate about when stuff leaves, but I guess this is the US freight restrictions21:30
GAN900Never had issues before.21:34
GAN900Andrew's is stuck, too, so I'm thinking DHL incompetence.21:35
GAN900US really doesn't have a lot of restriction on imports.21:35
*** mzanetti has quit IRC21:36
*** javispedro has joined #harmattan21:37
javispedroyou are not going to believe what just happened21:37
javispedroso I get my n950 with portmap, which was working perfectly21:38
javispedroexit home21:38
javispedroand after 4 minutes or so21:38
javispedroI look at the screen and bang21:38
javispedroAegis security problem21:38
javispedrowith /sbin/portmap21:39
javispedrothis time no way to get out of it21:39
javispedroso I had to go back home and swap sim to n90021:39
* javispedro sighs21:39
* javispedro sighs loudly.21:40
* javispedro writes off harmy, needs a reflash21:41
javispedrocya again21:41
antman8969hey javispedro, I'm tryingto use your icon maker but getting some trouble...have a sec?21:46
*** javispedro has quit IRC21:46
*** rlinfati has joined #harmattan21:46
rlinfati2 => ProductionServer is         // hostname wrong (WTF)21:47
rlinfati:(21:47 vs
GAN900Jaffa and his murder of N950s.21:49
*** lcukn900 has joined #harmattan21:53
djszapilcukn900: got harmattan working on N900 ? :)21:59
*** harbaum has joined #harmattan22:02
lcukn900djszapi, no but an app grid 6*4 in liqbase looks awesome :p22:04
lcukn900as does swipe swiping :p22:04
harbaumlcukn900: do you consider rewriting your stuff in qml?22:04
lcukn900harbaum i do, and i try different things often22:05
lcukn900latest calendar prototype is qwidget based and so far seen working on c5-03 n8 and most impressively the qt simulator lol22:06
lcukn900err a qwidget based calendar22:07
lcukn900qml oes not offer scaleable vectors fast enough yet, i try advancing the rendering speed often with folks22:07
lcukn900the art of line drawing is gone if everything is a rectangle22:08
harbaumhow does the navi then work? It seems to be a qml app22:08
harbaumis the nav view itself a c++ widget?22:09
lcukn900that is 1 rectangle with a view i believe22:09
* DocScrutinizer considers boxing and suspending N950 til the day much more knowledgeable devels like javispedro managed to at very least get successfully accomplished usual everyday activities like setting up portmap or recompiling and flashing/using stock kernel22:09
lcukn900liqbase needs thousands of them22:09
harbaumliqbase just need a 2d canvas widget + everything qml already has22:10
lcukn900harbaum complex paint processors and caching would be required and extremely wastefu to render to bitmap first22:10
DocScrutinizerreally I'm not interested to "develop" the 97th wallpaper changer, and it seems harmattan is not yet ready for doing serious development22:10
lcukn900many 2d canvas widgets22:10
lcukn900one sketch per rectangle22:10
lcukn900the graffiti wall has currently almost 10000 sketches22:11
lcukn900live rendered as you pan22:11
harbaumthat's still one 2d canvas with 10000 instances22:11
lcukn900and can you not see a memory problem yet?22:12
lcukn900one heavy set canvas with all its fancypants configuration options is much more memory than a normal qml rectangle22:12
rZrVenemo: i build your game and it works22:14
VenemorZr, thanks22:14
rZri made a few changes , wanna merge ?22:14
rZr+Build-Depends: debhelper (>= 5), libqt4-dev22:15
VenemorZr, umm, what are the changes?22:16
rZr+       # Add here commands to clean up after the build process.22:16
rZr+       -$(MAKE) clean22:16
VenemorZr, you altered the packaging?22:16
rZra bit yet22:16
rZri can push my changes22:17
VenemoI've deleted the debian folder from the game becasue Qt Creator can't deploy when it's present22:17
Venemothe original ones do have a lot of stuff22:18
rm_youGeneralAntilles: 8Departed Facility in EAST MIDLANDS - UK EAST MIDLANDS - UK03:5322:19
VenemorZr, the harmattan branch has not yet been pushed to gitorious22:19
rZrk thx22:19
*** harbaum_ has joined #harmattan22:20
rZri made a puzzle qml game too22:21
rZri'll may commit it too22:21
VenemorZr, jigsaw puzzle?22:21
rZri saw you create one too22:22
rZri can test yours if it's ready22:22
Venemoyes, but it's not packaged/ported to Harmattan yet22:22
rZri can do that job if needed22:22
rZrmy repo is growing every day22:23
*** harbaum has quit IRC22:23
VenemoI understand, but I want to port it to Harmattan first :)22:23
djszapirZr: I have also been doing a lot of packages for Harmattan. It would be nice to avoid the duplication.22:26
*** harbaum has joined #harmattan22:26
rZrdjszapi: there is a wiki page for that ICCR, I'll sure share the job22:27
djszapirZr: we might be able to share a repository for that. Actually, I am not entirely sure how this sharing works on c-obs.22:28
rZri think there is already a repo for that but it's empty22:29
*** harbaum_ has quit IRC22:29
djszapirZr: it would make sense to sync up with repositories from trusted volunteers.22:30
djszapior just sync and fix if there is any breakge.22:30
rZrwe already overide22:30
djszapi* breakage22:30
*** harbaum_ has joined #harmattan22:34
*** rlinfati has quit IRC22:35
*** harbaum has quit IRC22:36
*** antman8969 has quit IRC22:41
GeneralAntillesrm_you, exactly what mine says.22:43
rm_youGeneralAntilles: same plane? :P22:43
rm_youfiferboy's ever land?22:44
GeneralAntillesCrashed into the Atlantic.22:47
* GeneralAntilles sighs at the lightbulb idiocy.22:48
*** lawl0r has quit IRC22:48
GeneralAntillesI bet the implementation of the efficiency standards generates more CO2 than the bulbs they're regulating.22:48
djszapirZr: I have kdelibs working. When we can get c-obs back I will be uploading :=)22:52
rm_youGeneralAntilles: lol23:04
rm_you(to both things, kinda the same amount of funny but sad)23:04
GeneralAntillesrm_you, well, you know why the industry backed the new regulations so hard, right?23:05
rm_youbecause insanity / lobbyists, same reasons as always?23:05
GeneralAntillesSo they could close down their incandescent manufacturing plants in the US and move everything over to the CFL plants in China.23:05
GeneralAntillesWhich don't have to comply with any efficiency or real environmental standards.23:05
GeneralAntillesAnyway, enough of my off topic political whining.23:06
GeneralAntillesWHERE IS MY N950?!23:06
rm_youi'll laugh if it passes fiferboy's23:07
ieatlintbut the CFLs give us a new source of mercury which has been sorely missed since the banning of mercury thermometers23:07
ieatlintwe just need to start a rumour that huffing the vapor inside of them gets you high and we can ruin an entire generation of kids23:08
GeneralAntillesI still have some old thermostats sitting in the garage.23:09
*** harbaum_ has quit IRC23:16
DocScrutinizerGeneralAntilles: (lightbulbs) CFL? forget it, the energy needed to produce all this hightech crap needed in those crappy energy save lamps defeats any reasonable purpose - esp given the promised lifetime usualy isn't even reached by a tenth23:16
*** djszapi has left #harmattan23:17
GeneralAntillesDocScrutinizer, yes, I agree.23:18
GeneralAntillesGovernments trying to pick technology (or encourage a direction) does way more harm than good.23:18
DocScrutinizerbtw now I have those shitty CFL lamps and ponder buying an electrical heater :-P23:20
DocScrutinizer(means the heat generated by incandescent is not always to be considered 'lost' - actually I actively plan with this heat)23:21
ieatlinti just run several computers23:22
ieatlintworks well23:22
DocScrutinizerI pondered building a shower heating from 150 dimmed incandescent lamps - moot project23:22
GeneralAntillesDocScrutinizer, same with my desktop during winter.23:23
DocScrutinizerthe result is I have to heat the bathroom all the time now, as the conventional heating isn't fast enough for a spontaneous shower23:23
ieatlintsomehow the electrical requirements of 150 incandescent bulbs in the shower seems like a hazard23:25
lcukworht a try, the results would be enlightening23:25
* DocScrutinizer just hopes for some clever manufacturer selling 100W incandescent in a black shirt now, labeled as heater unit, and you just rip off the black shirt23:25
GeneralAntillesI want to hook a towel warmer up to a relay in the shower valve.23:26
ieatlintsome showers do have a heating lamp in the ceiling that you can turn on23:26
DocScrutinizerieatlint: 150 * 20..30W seems just fine23:26
* SpeedEvil has 600W of solar cells in the kitchen.23:26
SpeedEvilPutting them together is the fun part. :)23:27
ieatlintDocScrutinizer: so turning on your bathroom light would be a mere 3-4.5kwh? haha23:27
ieatlintit's like running two microwaves23:27
SpeedEvilI have a garage light made from 20*22W Energy saving bulbs.23:27
SpeedEvilEvery time I turn it on, I save around 1.5Kw.23:27
DocScrutinizerit's like running 1.5 .. 2 conventional heaters which is just ok to heat up the bathroom in 3 minutes23:28
*** djszapi has joined #harmattan23:28
djszapiCould someone tell me where I can find the "python-setuptools" package here: ? There are more packages depending on that package in this repository, but I failed ot find it. I hope it is just my sleep deprivation :p23:29
*** spenap has quit IRC23:31
DocScrutinizerdjszapi: I'd not be surprised to find more flaws in that location/repo that just "my" bash .deb that refuses to run and then killed the system when javispedro and I tried to make it run nevertheless ( )23:33
DocScrutinizers/that j/than j/23:33
infobotDocScrutinizer meant: djszapi: I'd not be surprised to find more flaws in that location/repo than just "my" bash .deb that refuses to run and then killed the system when javispedro and I tried to make it run nevertheless (
DocScrutinizerdjszapi: anyway ->23:35
DocScrutinizer[jr@lagrange nokia]$ find . -name '*python-setuptools*'23:35
djszapiI think this python-setuptools has been forgotten ti upload.23:35
lcukJaffa, attitude with compass patch for the win please :$23:35
djszapithanks :)23:35
Jaffalcuk: Roger.23:37
lcukJaffa, your app would make the idea test for new functionality !23:38
DocScrutinizero/ Jaffa23:38
GeneralAntillesHow's the community repo situation?23:39
djszapiGeneralAntilles: what do you mean exactly ?23:40
GeneralAntillesIs there a repo? Where is it, if so? And does Harmattan want to accept it?23:41
djszapi and also some on -cobs23:41
JaffaGeneralAntilles: Apps-testing and Apps repos are being set up, but there are key issues being signed up. People's home repos (equivalent to multiple Extras-devel) are springing up23:41
djszapi* c-obs23:41
djszapiDocScrutinizer: plese gimme w3m :p23:45
DocScrutinizerfind '*w3m*' ?23:46
djszapiHave you added this repository to your /etc/apt/sources.list file ?23:46
DocScrutinizerdjszapi: sorry?23:46
DocScrutinizerdjszapi: no, copied to my box by wget23:47
DocScrutinizerdjszapi: you want to know about where '*w3m*'?23:47
djszapiwget -r ?23:48
DocScrutinizersmt like that, yes23:48
djszapiit would be nice if they put a search engine there.23:48
DocScrutinizerwell, now that I have the mass, an occasional update is painless23:49
djszapiyeah, right now I need to know ^wm3$23:49
DocScrutinizerwget is smart enough for this23:49
DocScrutinizerok, moment23:49
DocScrutinizerwm3 or w3m ?23:49
djszapisorry, my bad :P ^w3m$23:50
djszapithe text based browser23:50
DocScrutinizerdoesn't exist23:50
djszapiok, I am packaging then..23:52

Generated by 2.15.1 by Marius Gedminas - find it at!