IRC log of #maemo-ssu for Saturday, 2013-09-07

FatPhiljust for reference, DT will slow you down00:22
*** dhbiker has quit IRC00:25
*** sunny_s has quit IRC00:31
*** freemangordon_ has joined #maemo-ssu00:58
*** freemangordon has quit IRC00:58
sixwheeledbeastPali: will there be an update for worldclock in cssu-devel?
Palisixwheeledbeast: yes if ade push changed to gitorious and tell me that new version is ready for compiling :-)01:15
sixwheeledbeastok ta I'll ask him01:15
Palialso need to push new ke-recv to cssu-devel :-)01:15
*** Vlad_on_the_road has quit IRC01:16
*** NIN101 has quit IRC02:05
*** freemangordon has joined #maemo-ssu02:23
*** freemangordon_ has quit IRC02:23
*** dos1 has quit IRC02:58
*** Pali has quit IRC03:43
*** M4rtinK has quit IRC04:09
*** nox- has quit IRC04:16
*** LauRoman has quit IRC04:23
*** BCMM has quit IRC04:24
*** jon_y has quit IRC05:06
*** amiconn has quit IRC05:12
*** amiconn_ has joined #maemo-ssu05:12
*** amiconn_ is now known as amiconn05:12
*** jon_y has joined #maemo-ssu05:16
*** infobot has quit IRC06:45
*** oldtopman has quit IRC07:35
*** infobot has joined #maemo-ssu07:43
*** ChanServ sets mode: +v infobot07:43
*** FlameReaper has joined #maemo-ssu07:54
*** discopig has quit IRC07:59
*** discopig has joined #maemo-ssu08:00
*** LauRoman has joined #maemo-ssu09:08
*** sunny_s has joined #maemo-ssu09:36
*** Vlad_on_the_road has joined #maemo-ssu09:47
FatPhilwhat kind of fucknut uses grep -x "^...[elided]...$"?09:47
FatPhilwhat bit about ^ and $ wasn't clear?09:47
FatPhilright, let's flash this fscker, nothing of worth on it, clearly...09:48
freemangordonFatPhil: yep, just flash PR1.3 and eMMC09:48
FatPhilmaybe I should do it when I'm sober09:49
freemangordonhmm, isn't it too early in the morning to be drunk?09:49
FatPhilI've got the apt sources lists, and the dpkg -l, so I should be able to find any package that was on here in the past that still exists09:50
FatPhil"unfortunately", last night's tasting session consisted entirely of about 9% beers, so I still have a low density of blood in my alcohol stream09:51
freemangordonoh, I see :)09:51
FatPhilso, a 13-2 EMMC image, and either a 36-2 or 38-1 root image09:53
freemangordonboth roots will do it, 38-1 has some sertificates blacklisted, but this is in cssu too09:53
freemangordonFatPhil: BTW if you're going to jump on cssu-testing, you'd better jump on cssu-thumb09:54
freemangordonsame packages, but way faster09:54
FatPhilthis device is for being on the bleeding edge09:55
freemangordonFatPhil: then cssu-thumb is what you need09:56
infobotsomebody said cssu-thumb was <DocScrutinizer05> [thumb2 microb] indeed seems to render like mad, subjectively, or
FatPhilnope, a whole pint of pineapple and grapefruit juice is what I need09:56
FatPhilhaha, man, I feel so dumb, I've not used this flasher in about 2 years09:58
freemangordontry some yougurt, mixed with water and a bit of salt, this is what I use usually09:58
*** FlameReaper has quit IRC09:58
freemangordonunfortunately google translates the above mixture as "kefir" which for sure is not true :D09:58
FatPhilhmmm, salty lassi09:59
jon_yis there a way to lock only the hildon-desktop orientation?09:59
jon_yI was doing it with gconftool in earlier versions09:59
freemangordonthe same in newer09:59
jon_ybut now the option seems to be ignored09:59
infobotmethinks cssu is
freemangordonjon_y: gconftool-2 -s /apps/osso/hildon-desktop/desktop_orientation_lock -t bool true10:01
jon_yinfobot: desktop-lock is gconftool-2 -s /apps/osso/hildon-desktop/desktop_orientation_lock -t bool true10:02
infobotokay, jon_y10:02
jon_yfreemangordon: strange, I set it through the gui gconf editor10:03
jon_yit didn't work there10:03
jon_yah wait, I spoke too soon10:03
freemangordonjon_y: then check your gonf editor :)10:03
jon_yno, still going into potrait10:03
freemangordonjon_y: this locks only desktop, not tasknav or app menu10:03
jon_yyeah, I want to lock desktop only10:04
jon_yis it maybe because of turbo speed hack?10:04
freemangordonno, the same happens here10:05
jon_yok, yours is not locking either?10:05
freemangordonplease file a bug10:05
jon_yhow do I file bug?10:05
FatPhilis the website rather flakey?10:06
freemangordonFatPhil: BMO works just fine10:07
freemangordonjon_y: you need to register to enter a bug ofc10:08
FatPhilwiki wasn't10:08
freemangordonwiki is ok too10:08
freemangordonFatPhil: ^^^10:08
FatPhilit's my wifi then10:09
povbot_Bug 12707: /apps/osso/hildon-desktop/desktop_orientation_lock setting ignored10:17
jon_ybtw, is there no way to get around the red boxes in turbo mode?10:18
jon_yI remember it did not have red boxes when setting it for the first time10:18
jon_yonly appeared after the reboot10:18
jon_ydisabling the turbo hack removes the red boxes10:19
jon_yit is back reenabling it10:19
freemangordonshit, ke-recv-extra doesn't have dbg package :(10:19
jon_yI'm fine living with the red boxes though, no biggie10:19
jon_yfast menu makes n900 feels fast10:19
FatPhilstock nokia software is shit10:24
FatPhilI've just waited a minute for the browser to load the wiki page, so I could install the thumb repo10:24
FatPhilnot even briefly did the piece of shit browser think that it ought to ask me where the freaking internet is10:25
freemangordonhmm, it is possible your network connection to be crappy10:25
FatPhilwe interviewed someone from the browser team a few weeks ago10:25
FatPhilI forgot to punch him in the face10:25
FatPhilnormally it's amazing10:26
FatPhilI live on a university campus, we have more bandwidth than I know what to do with10:28
FatPhilexcept today, apparently10:28
FatPhilbut that doesn't stop the builtin browser from being a steaming pile10:29
freemangordontry thumb2 version, you'll be surprised10:30
FatPhilIf I can get the current browser pointing in the direction of that repo, I'll do that10:30
FatPhilit really does stink10:31
FatPhilI just disconnects from the wlan, thought I'd try again, and then I did a web search - again, it didn't ask me where the internet was.10:31
FatPhilbring back the nokia bugzilla, I want to get some serious swearing out of my system10:31
FatPhilMaybe they broke it in week 3810:32
FatPhilwhilst they were rejecting real bug fixes, they were  incorporating regressions instead10:33
FatPhilI think this is the first time I've used week 38's image, it's tragically fucked10:33
FatPhilmy daily device is working fine (the week 36 image)10:36
*** LaoLang_cool has joined #maemo-ssu10:37
FatPhilwoh, I can downgrade?!??!10:37
*** sunny_s has quit IRC10:38
FatPhilI didn't check, maybe this has an R&D certificate on it10:38
freemangordonFatPhil: wait, what are you doing?10:38
freemangordondid you run cssu-installer?10:39
FatPhilhaven't run anything on the PoS yet10:42
FatPhilI've just put 36-2 on it, as 38-1 simply didn't  work10:42
FatPhilOnly nokia could have a shutdown process that starts daemons...10:43
FatPhil(that's a lie...)10:43
jon_yfreemangordon: btw, is the n900 qemu a workable solution to test kernel modifications?10:43
freemangordonjon_y: ask Pali, I've never used it10:44
jon_ywill ping him when I see him10:44
FatPhilwhen you test kernel mods - do you have the ability to upload the kernel only to RAM, and run from there? or do you have to flash it?10:45
freemangordonflasher can be used to boot the kernel10:46
freemangordonbut modules should be on rootfs10:46
FatPhilno --initfs switch?10:47
jon_yalternatively I could just put it in /opt/boot and use uboot10:47
jon_ydon't think it can10:47
jon_yalso, what kind of dev environment do you guys use?10:48
jon_yat least gcc-4.7.x pls :)10:48
jon_ythe linaro gcc?10:49
FatPhilI build kernels with 4.4.1 apparently10:49
jon_yuboot and linux 3.1x apparently requires 4.7 at least10:50
FatPhilanything that requires a particular compiler version is broken, IMHO10:50
jon_yno, it needs some new features10:50
FatPhilthen it would be broken, see my statement 2 lines up ^^^10:51
*** FlameReaper has joined #maemo-ssu10:51
freemangordonFatPhil: do you remember having problems with gio channels and newer kernels?10:51
jon_yFatPhil: you backported gcc-4.7 features?10:52
FatPhiljon_y: what about what I've said lead you to that conclusion?10:52
freemangordonthere is gio set on /proc/mounts10:52
jon_yit does require some new features from later versions10:52
jon_yiirc some new asm instruction10:53
freemangordonthe callback is called every now and then, with GIOCondition == G_IO_IN10:53
freemangordonthis looks similar o the poll() case10:53
jon_ysomething to do with DSP instructions according to pali10:53
jon_ywell, here are the binaries I found if anyone is interested
*** Pali has joined #maemo-ssu10:54
freemangordonjon_y: 4.2.1 just can't parse some of the headers10:54
FatPhilGIO gives me nothing but glib hits10:54
jon_yhi Pali10:54
jon_yfreemangordon: ok10:54
freemangordonFatPhil: yeah10:54
FatPhilwtf's that to do with the kernel?10:54
freemangordonPali: hi, hulda hits a problem similar to the poll() one10:55
FatPhilif glib can't do something, that's a glib problem10:55
*** LaoLang_cool has quit IRC10:55
jon_yPali: what dev environment do you recommend to work with if I were to do some kernel hacking10:56
*** LaoLang_cool has joined #maemo-ssu10:56
FatPhilchrish, 36-2's horrifically broken too, for the same reason10:56
freemangordonPali: it does g_io_add_watch on /proc/mounts and the callback is non-stop called with GIOCondition == G_IO_IN10:56
Palijon_y: is some env needed? I'm using crosscompiler and that is all10:56
*** LaoLang_cool has quit IRC10:57
Palifreemangordon: ok, thanks10:57
jon_yPali: I see10:57
Palifreemangordon: glib does not support some flag?10:57
jon_yPali: these will do fine?10:57
freemangordonPali: the same happens for a NETLINK socket10:57
freemangordonseems like10:57
Palijon_y: which distribution do you have?10:57
Palifreemangordon: I'm sure you *can* wait for data on netlink10:58
jon_yPali: plain debian testing10:58
Palifreemangordon: also in my kernel branch there is already reverted sysfs patch...10:59
freemangordonPali: the problem is that the callback is called with cond == G_IO_IN|G_IO_ERR, which hulda doesn't process. I will fix that and see what happens10:59
freemangordonbecause someone never think that cond is a bitfiels11:00
Paliok, try to fix it11:00
*** discopig has quit IRC11:01
freemangordonPali: any clue what hulda is supposed to do?11:01
Palijon_y: looks like debian does not have arm cross comilers in repository11:01
Palitry to install some11:01
freemangordonI will try11:01
Palifreemangordon: no11:02
*** discopig has joined #maemo-ssu11:02
jon_yPali: also, what do you recommend I use if I am to rebuild your power kernel?11:02
jon_yjust cross compilers?11:02
Palijon_y: for kernel-power, you can use scratchbox or madde too11:02
FatPhilubuntu has xcompilers11:03
infobot[scratchbox] a cross-compiling system that uses binfmt_misc, rpc calls, and an nfs mount to make a cross-build appear to be 100% native, and is found at, or hosted by now, maintainer thedead144011:03
Palibut cross compiler is enough (if you do not want flasher package)11:03
jon_yI want the boot image for uboot11:03
Palifor zImage you do not needed nothing else11:03
jon_yok, just scratchbox11:04
jon_ysomething like this I guess
FatPhilnah, wifi just doesn't seem to work in this device :-(11:09
FatPhilgonna make installing CSSU rather hard11:09
kerioi have that problem too :c11:11
keriousb networking is still an option11:11
*** iDont has joined #maemo-ssu11:11
FatPhilthe bizarre thing is that I'm sure it worked before I flashed it11:13
FatPhilonly nokia could write software that shuts down in order to power up11:15
freemangordonPali: netlink socket problem solved11:16
Paliok :-)11:17
freemangordonPali: do you have rights to create repos on gitorious (CSSU that is)?11:18
infobotmethinks kp is
FatPhilwoh - I just had the blue-on-white nokia logo with the 5-dots animaation playing in front of it11:19
Palifreemangordon: no :-(11:19
Pali~kp is
infobotPali: i already had it that way11:20
infoboti guess kp is
Paliinfobot: no kp is
infoboti already had it that way, Pali11:21
infobotrumour has it, kp is
jon_ymy debian vm hasn't been powered up for nearly a year, time to update everything11:22
jon_yyeah, looks like linaro 4.7.2 is still there11:24
Palimerlin1991: ping11:31
FatPhilok, how to diagnose wifi issues?11:33
FatPhil[  497.545776] wlan0: authenticated11:34
FatPhil[  497.545806] wlan0: associate with AP 94:44:52:bb:52:7d11:34
FatPhil[  497.551605] wlan0: RX AssocResp from 94:44:52:bb:52:7d (capab=0x411 status=0 aid=3)11:34
FatPhil[  497.551635] wlan0: associated11:34
FatPhilwlan0     Link encap:Ethernet  HWaddr 00:26:CC:77:EE:BF   inet addr:  Bcast:  Mask:
freemangordondid you try to reboot the router?11:34
FatPhilthat's in another guy's room :-(11:34
PaliFatPhil: try to look in syslog, wlancond or icd2 sometimes write something...11:36
freemangordonPali: /proc/mounts problem fixed too, need a git repo to commit11:44
Palifreemangordon, create personal clone on maemo.gitorious.org11:45
*** arcean has joined #maemo-ssu11:45
FatPhilwoh, this thing doesn't have syslog!11:45
FatPhilI can ping my daily device over wifi11:47
FatPhilanother flatmate's skyping, so wifi works for him11:48
FatPhilit seems to be just this browser that's failing11:48
freemangordonFatPhil: add repos by hand11:48
freemangordondeb fremantle free non-free11:49
jon_yPali: this patch still needed?
freemangordondeb fremantle free non-free11:49
freemangordonFatPhil: both of them needed11:49
jon_yseems to be some trouble with newer binutils11:49
freemangordonFatPhil: also make sure you have all the nokia repos enabled11:49
Palijon_y: try to compile with or without that patch11:50
jon_yPali: I made this patch some time ago11:50
jon_yI'm srestarting my kp builds11:51
Palithis patch is needed in upstream 3.x kernel11:51
Palibut no idea if for 2.6 too11:51
freemangordonPali: another problem - if you lock the device, the leds got switched off, but digitizer is still active, any clue?11:51
jon_yPali: it was for kp5211:52
freemangordontouch screen works11:52
jon_yso yes, if you use newer linaro, I think11:52
jon_yit would be nice if you can accept it :)11:52
Palifreemangordon, mce problem?11:52
freemangordonseems like, just mentioning it here11:52
Palijon_y: but old gcc on autobuilder does not know anything about that sec11:53
Palifreemangordon: we need to patch MCE...11:53
jon_yI see, ok11:53
PaliI already have tree with trying to backport harmattan version for fremantle11:53
freemangordonPali: hmm, don;t sound good11:53
PaliI think now it compiling, but it was nvere tested11:53
freemangordondoesn't even11:54
*** NIN101 has joined #maemo-ssu11:54
freemangordonPali: you found the initial fremantle tree state?11:55
Palino, it does not exists11:55
Paliwhat we have is diablo MCE11:55
Paliand harmattan MCE11:55
Paliand fremantle dbus files11:55
freemangordonFatPhil: do you have fremantle mce git tree on your laptop by chance :)?11:55
Paliso with above data I tried to reconstruct fremantle MCE11:56
freemangordonok, no more hulda eating 100% CPU11:56
freemangordonPali: I see11:56
freemangordonPali: ke-recv will have to be fixed too :(11:59
freemangordonit does not automount uSD on backcover closed. and filemanager segfaults11:59
freemangordonhopefully we have all that12:00
freemangordonsource code I mean12:00
*** FlameReaper has quit IRC12:00
jon_ywhere can I get the in progress 3.x kernel?12:04
jon_yhow do I start playing with it? :)12:04
FatPhiloperation failed (no updates available)12:11
*** FlameReaper has joined #maemo-ssu12:15
Palikillall apt-get and apt-worker processes, close HAM and open it again12:15
*** infobot has quit IRC12:15
FatPhildone, even rebooted the wireless router12:25
FatPhillaptop's coping, daily device is working, samsung's working, n9 isnt, new n900 isnt.12:29
*** iDont has quit IRC12:31
FatPhilhow can this POS be 400MB into swap already, and I've not even done anything with it yet?!?!?12:36
*** discopig has quit IRC12:43
*** discopig has joined #maemo-ssu12:44
FatPhileven packed data isn't working12:51
FatPhilhmmm, this device ist kaput12:52
*** bsdmaniak has joined #maemo-ssu12:52
FatPhilit's so unresponsive too, takes several seconds to respond to a screen tap sometimes12:53
*** Vlad_on_the_road has quit IRC12:53
freemangordonFatPhil: 400MB of swap?!?12:55
freemangordonnever seen that12:55
FatPhilaccording to free12:56
freemangordonomg, what top has to say about it?12:56
FatPhiltracker's flattening my battery12:56
freemangordonbut who is using that memory?12:56
FatPhiltracker's using 107m12:57
freemangordondid you flash eMMC?12:57
freemangordonhmm, wait a second12:59
FatPhil13-2 EMMC12:59
freemangordonFatPhil: do
freemangordonbut still, 400MB of swap?!?13:01
freemangordonFatPhil: besides tracker, what else eats the memory?13:03
freemangordonjon_y: look in the chanlog from yesterday, Pali posted a link to his repo13:06
Palihere are all links13:07
FatPhilfreemangordon: nothing obviously enormous13:11
FatPhilmmcqd is using that's a bit of a lie - it all looks like bloatware to me13:12
FatPhilpulseaudo 80MB!?!?!13:12
freemangordon36-2 firmware?13:13
FatPhilsignond 23MB, maemo-xinpit-sounds 43MB, camera-ui 54MB13:13
FatPhilI put 38-1 back on, aftr 36-2 was equally broken13:14
FatPhilosso-connectivity-ui-dialogs 47MB13:14
freemangordonshe is dead :(13:14
freemangordonFatPhil: that looks impossible to me. Not that I don't trust you :)13:16
freemangordonFatPhil: you used the release firmware, ain't?13:16
FatPhilwhat we called the PR image, yes13:17
freemangordonFatPhil: is it the same ?13:19
freemangordonBTW 38-1 is not released13:20
FatPhilI can go back to 36-213:21
FatPhilliterally got nothing to lose13:21
freemangordonFatPhil: download eMMC and rootfs from, just in case13:21
FatPhilmy file sizes are identical13:22
jon_yfreemangordon: ok13:22
freemangordonFatPhil: hmm, ok13:22
freemangordonFatPhil: make sure to follow the "flashing emmc" sequence13:22
FatPhilI lie - what I've got named 38-1 is the same size as what you have as 36-2.13:23
jon_yPowerVR SGX530 :(13:23
jon_yhow is PowerVR supposed to work anyway?13:23
freemangordonFatPhil: well, download both images from the above link and make sure you follow the instructions on
freemangordonFatPhil: esp the part that you have to flasf rootfs after eMMC, without rebooting. iirc13:24
freemangordonjon_y: hmm?13:24
*** LauRoman has quit IRC13:25
jon_yfreemangordon: according to pali's link
freemangordonjon_y: 2.6.28 powervr driver is forward-ported to 3.x kernels13:25
jon_yoh, it works already?13:26
freemangordonthe point is that it is not upstreamed and will never be aiui13:26
jon_yok, it works but will not be upstreamed13:26
freemangordonjon_y: how do you think we boot maemowithout working graphics?13:26
jon_ylicense problems?13:26
FatPhilmd5sum differs...13:27
freemangordonjon_y: well, noone CBA to upstream it13:27
FatPhilEMMC is a match though13:27
jon_yI see13:27
freemangordonjon_y: not to say it is TI/Imagination's job to do that13:28
FatPhilooooh, I had a LEGACY image, maybe that's what's confusing it13:28
FatPhilreflashing with the skeiron images13:29
jon_yyeah, I know, I thought PowerVR hate linux13:29
freemangordonFatPhil: ^^^13:29
jon_yinfamous for not documenting their hw13:29
* freemangordon is afk13:30
DocScrutinizer05yoghurt, water, salt? AYRAN13:31
freemangordonDocScrutinizer05: exactly, I was not sure there is such word in english :D13:32
freemangordonthere is no better cure for hangover13:33
DocScrutinizer05I think there is such word in turkish only13:33
freemangordonthough we use it in bulgarian too13:33
*** FlameReaper has quit IRC13:37
FatPhilwhat's a hangover? they're for amateurs, I hear.13:37
DocScrutinizer05hangover? never happens when you take many minerals and one aspirin the evening before you would awake with hangover13:39
jon_yfreemangordon is bulgarian and DocScrutinizer05 is ...? :)13:40
jon_ythat too :)13:41
jon_yor maybe he meant Hanoverian13:41
DocScrutinizer05damn bot13:44
FatPhilright, not swapping frantically any more13:44
FatPhilhow bizarre13:45
*** infobot has joined #maemo-ssu13:45
*** ChanServ sets mode: +v infobot13:45
DocScrutinizer05~spell hangover13:45
infobot'hangover' may be spelled correctly13:45
DocScrutinizer05~spell hungover13:46
infobot'hungover' may be spelled correctly13:46
DocScrutinizer05~dict hungover13:46
infobotcould not find definition for hungover13:46
DocScrutinizer05~dict hangover13:46
infobotDictionary 'hangover' (2 of 5): \hangover\ n. 1. An unpleasant feeling, such as a headache, occurring as an aftereffect from the use of drugs (especially alcohol). [WordNet 1.5 +PJC] 2. an official who remains in office after his term. Syn: holdover. [WordNet 1.5] ;; an official who remains in office after his term .13:46
DocScrutinizer05~dict 2 hangover13:47
infobot[2/5] disagreeable aftereffects from the use of drugs (especially alcohol)13:47
FatPhilyay - internet!13:48
FatPhilso, off to CSSU/Thumb...13:50
FatPhilis it basically going to replace everything on the device?13:54
FatPhilwhere are "Nokia" and "Extras" catalogues?13:56
jon_ywill 0xFFFF be ever ported to windows?13:56
FatPhilah, "Extras" = ""13:58
jon_ywhat exactly does 0xffff require anyway?13:58
jon_yah libusb13:58
jon_ywindows has libusb13:58
jon_yuh oh, ioctl13:59
jon_yprobably relegated to cygwin now13:59
jon_yoh #include <linux/fs.h>13:59
jon_yforget cygwin either13:59
DocScrutinizer05FatPhil: it's just a supplementary repo on top of cssu14:00
DocScrutinizer05which in turn is a supplementary repo on top of the base nokia repo14:00
*** DrCode has quit IRC14:01
*** dos1 has joined #maemo-ssu14:01
DocScrutinizer05cssu-thumb need a thumb-enabled kernel though, or device will segfault like mad14:01
FatPhilbusy clicking "OK"...14:02
infobotmethinks cssu is
jon_ydoes it make sense to enable selinux on n900?14:02
jon_yyou know, because *secure*14:03
FatPhildoes it make sense to enable selinux anywhere?!??!14:03
jon_yI don't know, I know the text editor isn't supposed to run external programs14:03
infobotextra, extra, read all about it, combined is the rootfs fiasco image of maemo. For N900 latest (PR1.3.1) see  or
*** DrCode has joined #maemo-ssu14:05
DocScrutinizer05isn't this guy running a really helpful dude?14:06
FatPhilhmmm, app manager is saying liqtm-bearer conflicts with some of the stuff14:06
DocScrutinizer05QtM hmmmmm14:07
* DocScrutinizer05 would probably kill qtm with fire on his device14:08
FatPhilsaying "No"...14:08
DocScrutinizer05what are you doing? CSSU?14:10
FatPhilam mow14:10
FatPhils/ m/ n/14:10
infobotFatPhil meant: am now14:10
DocScrutinizer05~wtf mow14:10
infobotGee...  I don't know what mow means...14:11
* DocScrutinizer05 does a mtr from neocortex to dsl modem14:11
DocScrutinizer05oh, there's a few -ENOCAFFEINE in syslog14:12
DocScrutinizer05you're following the procedure in ?14:13
FatPhilI think so14:14
FatPhildownloading 50.0 MB ...14:14
DocScrutinizer05only update the CSSU system update14:14
DocScrutinizer05don't do an update-all14:14
DocScrutinizer05most of the stuff will get updated by cssu-update anyway14:15
DocScrutinizer05update-all time for after cssu got installed14:15
FatPhilUpdating 'Maemo 5 Community SSU Package(thumb)'14:17
FatPhilI need some freaking breakfast14:17
DocScrutinizer05I'm completely lost on the correct procedure to directly upgrade to cssu-thumb14:18
DocScrutinizer05h,, too early for breakfast here14:18
DocScrutinizer05hmm even14:18
*** BCMM has joined #maemo-ssu14:22
FatPhiloperating system successfully updated14:22
*** LauRoman has joined #maemo-ssu14:23
FatPhilwhere's the list of closed source packages?14:24
*** infobot has quit IRC14:37
FatPhilmany thanks14:37
*** jonwil has joined #maemo-ssu14:42
*** Vlad_on_the_road has joined #maemo-ssu14:43
DocScrutinizer05DMAN BOT!!14:55
*** bsdmaniak has quit IRC14:55
DocScrutinizer05hi jonwil14:56
DocScrutinizer05how's RE business?14:56
*** infobot has joined #maemo-ssu14:57
*** ChanServ sets mode: +v infobot14:57
jonwilHaven't really been inspired to do any N900 work lately, been busy with other stuff (including spending time trying to find someone who will pay me to write code since I am currently unemployed :)14:58
DocScrutinizer05same here14:58
jonwilalthough having read through most of the big Neo900 project, I might be inspired to jump back into reverse engineering of those bits that will be the most useful to the Neo900 project going forward from a software perspective (if one could figure out what those actually are :)14:59
FatPhilunemployment in 4 weeks and counting...14:59
DocScrutinizer05bot is really dead14:59
DocScrutinizer05jonwil: GREAT!15:00
*** infobot has quit IRC15:02
jonwilWhat I need to do is figure out who is actually thinking about Neo900 software (if anyone) and get a discussion going about just what blobs are going to need replacing to make Neo900 project possible15:02
jonwilAlso right now I need something good to sink my teeth into so I can completly avoid the TV and internet news (we just had an election in my country and it looks like the party I really really hate is going to win with a massive majority and bring in all these laws that are bad for the country and only good for the big end of town so reading news about it will just make me depressed)15:04
jon_yjonwil: replace proprietary software?15:05
jon_yis it possible to rework the sim interface code so I can save numbers into sim cards?15:06
jonwilI dont know enough about the SIM interface to answer that question15:07
jon_yok, what do you plan to work on?15:07
jonwilI dont know15:07
jonwilI need to talk to people who actually know things about the Neo900 and its planned software stack in order to figure out where my skills would be most useful...15:08
jon_yall I know that it is supposed to be compatible with the original n90015:08
jonwilyeah I read that too15:09
jonwilI seriously doubt they will be able to use the same cellular modem though so replacing the cellular services daemon will probably be necessary15:09
dos1jonwil: we plan to use FSO ( for the GSM middleware and implement some kind of bridge between it and isi so blobs can still access modem like they do on N90015:09
jon_yoh, make conversation recording less of a pain15:10
jon_yif the neo900 can record calls automatically and easily, I'd be raelly happy15:10
jon_yrecaller is a bit buggy at times15:11
dos1or reimplement GSM daemon to have the same API but use FSO internally15:11
dos1I'm not sure how GSM middleware works right now on Maemo (but I'm quite experienced with FSO)15:11
jonwilGSM on maemo is done by the cellular services daemon15:11
jonwiland by plugins for that daemon (e.g. csd-call, csd-gprs, csd-sms etc)15:12
jonwilmost of the talking to the next pieces up in the stack is done by dbus15:12
dos1so I guess we can try to reimplement its dbus API with a bridge that will call FSO dbus API15:13
jonwilI suspect a useful use of my time would be to document as much as possible of the places (dbus and otherwise) where the cellular services daemon and its plugins/related libraries talk to the outside world and the upper levels of the N900 software stack15:13
jonwilIts hard because the cellular services daemon dbus interface is undocumented :(15:13
dos1jonwil: yeah, that would be great15:16
jon_ymake a fake dbus server that pass stuff through?15:16
dos1what about hacking libdbus to log every call made via it? :D15:17
jon_yyeah, something like that15:17
jon_yyay printf debugging!15:17
keriojust use dbusmonitor15:18
jonwilThe issue isn't logging things (that's what dbus-monitor is for), its figuring out what stuff actually does and about making sure you have covered every possible interface (which involves reverse-engineering)15:18
jonwilIt is a pitty the N900 community does not have the kind of guru-level reverse engineering guys that other communities have15:24
*** NIN101 has quit IRC15:25
dos1jonwil: some an alternative to reimplementing csd api would be reusing csd and imitating isi modem (which seems to be documented pretty well)15:27
dos1but I guess getting rid of csd blob could be beneficial in long term15:28
FatPhilhmmm, so now I have thumb CSSU what should I do with it?15:29
jonwilisi modem isn't as documented as you might think15:29
Palifreemangordon, did you fixed hulda?15:29
dos1jonwil: there already is working ISI support in FSO15:31
dos1(but not sure if complete)15:32
dos1so it'd have to be "reversed", and I suspect imitating modem should be a bit easier than handling it15:32
DocScrutinizer05jonwil: (replacing cmt interface sw) indeed, that's the most demanding and painful/problemantic part15:37
DocScrutinizer05ooh, dos1 beat me to it15:38
*** Vlad_on_the_road has quit IRC15:38
jonwilparts of ISI are well documented (mostly though the work Nokia did when they added N900 support to ofono) but lots of other parts are not documented (for example there is a dbus call named get_service_provider_info and there is no documentation at all anywhere for the ISI calls that this dbus call ultimately matches to15:39
jonwilThere are parts we have documentation on from various sources15:40
jonwilParts we have some limited details of (e.g. header files with data structures) but no documentation15:40
jonwilAnd parts we have nothing for15:40
DocScrutinizer05I thought we might replace libisi?15:40
DocScrutinizer05isn't there such a thing that finally handles *all* communication 8minus sound) to the modem?15:41
DocScrutinizer05would be one (or more) layer *under* csd*15:41
jonwilBasically what you have is modem firmware15:42
jonwilthen kernel and drivers that talk to it15:42
dos1I'm not sure if such thing like libisi exists in Maemo15:42
DocScrutinizer05 /usr/lib/
dos1och, so it exists :)15:42
jonwilok, so basically what you have is libisi doing most of the communications to the modem (some other bits of the system may talk to kernel directly without going through libisi but I dont know about that)15:43
jonwillibisi though is a generic library for talking to the modem and passing data packets from higher level libs15:43
*** Vlad_on_the_road has joined #maemo-ssu15:43
jonwil'and its totally undocumented15:44
DocScrutinizer05there are more than one approach to this, we need to discuss the pros and cons of each15:44
dos1so confusing... library fso and ofono folks worked on was also named libisi15:44
jonwiland totally different to any isi code in fso (or in ofono where fso guys got lots of the N900/isi code/info from IIRC)15:44
DocScrutinizer05I'm sure we will be able to RE what libisi API offers15:45
dos1personally I'd prefer to see csd being replaced with fso bridge, such solution would be the sanest from my pov15:46
dos1but that does not take the amount of work needed for that into consideration15:46
DocScrutinizer05csd does a lot of high level protocol heavy lifting15:46
DocScrutinizer05IOW much of the every day phone stack stability of maemo is from csd15:47
DocScrutinizer05I _think_ we might go for a fake libisi _first_, then in step two of this enterprise create a much more ambitioned project that#s aiming at replacing csd by fso15:48
DocScrutinizer05I'm open to discussions on that15:49
jonwilBasically we have several theoretical options (some of which are probably not possible):15:50
jonwilIf we ignore the options that we cant do (i.e. using the same cellular chip set and firmware as Nokia) the options we do have are:15:51
jonwil1.Replace the kernel bits that talk to the cellular modem with something that exposes the same interface to userspace but which takes the N900-format cellmo data packets and converts them into whatever the cellmo we use is expecting15:52
jonwil2.Same as #1 but replacing libisi (low level library that talks to kernel and takes in data packets from next level up)15:53
*** Vlad_on_the_road has quit IRC15:53
jonwilProblems with #1 and #2 are all the isi data packets where no documentation exists (or where we only have a .h file with no info on what it actually means)15:54
dos1for me, we can forget about #1, which imo is basically #2 done in the wrong place :P15:55
jonwilyeah we can :)15:55
jonwilOption #3 is to forget about isi, reverse engineer the cellular services daemon and replace it with something that talks the same dbus (and other) interfaces to the next bits up but does things differently further down15:55
jonwilOption #4 is to identify the things that have dependencies on the cellular services daemon and rewrite them so they do things differently (e.g. replacing libicd-network-gprs with something that is functionally identical except that it talks to whatever GPRS interface is exposed by FSO rather than to what is exposed by the csd-gprs plugin)15:57
DocScrutinizer05dos1: I don't agree about wrong place. I completely agree about "forget it, can't be done"15:57
FatPhilso do I need to install a new kernel or anything?15:57
DocScrutinizer05for thumb?15:58
jonwilOption #5 is a combination of #3 and #415:58
DocScrutinizer05or on Neo900?15:58
dos1DocScrutinizer05: AT interpreter in kernel wouldn't be out of place? that would in fact transform into in-kernel fsogsmd with isi api instead of dbus15:59
FatPhilDocScrutinizer05: for thumb, yes15:59
jonwili.e. we identify all the bits of the system that are dependent on CSD and its plugins and depending on that we either replace it with something that talks to FSO (e.g. for the widget that displays cellular signal level we can replace with something that talks to existing FSO interface for signal level)15:59
DocScrutinizer05yes, you need thumb-enabled kernel and you're probably in trouble already if you don't have it installed15:59
jonwilor we implement the interfaces that are required for that blob inside FSOGSMD16:00
DocScrutinizer05you can flash thumb kernel, but you can't install the modules needed if you don't have already and your device blows up during boot16:00
DocScrutinizer05or at least it's for sure not easy to install those modules when the system blows chunks16:01
DocScrutinizer05dos1: there are ways to expose a phonet device but do the conversion to AT in userland16:02
DocScrutinizer05but I guess it would be a nightmare16:02
DocScrutinizer05since you not only need an AT interpreter but also a comprehensive ISI interpreter16:03
DocScrutinizer05I don't know if it would be easier to accomplish than replacing libisi with something that doesn't talk to phonet device but exposes similar compatible API to next upper layer16:04
DocScrutinizer05my general stategy would be: start as low as possible, evaluate feasibility. If not possible to get done, go up one layer and redo evaluation.16:06
DocScrutinizer05layer 0: phonet device. layer 1: libisi API16:06
DocScrutinizer05layer 2: probably csd replacement16:06
DocScrutinizer05layer 2.5: dbus translator/bridge to csd16:07
DocScrutinizer05layer 3: apps?16:07
DocScrutinizer05layer 3 == jonwil's option #416:08
*** oldtopman has joined #maemo-ssu16:20
*** Vlad_on_the_road has joined #maemo-ssu16:33
*** LauRoman has quit IRC16:35
jonwilthinking some more about closed bits on the N900, GPS shouldn't be an issue since we can use whatever open code exists for the GPS solution that ends up in the Neo900 and build a binary-interface-compatible replacement for liblocation (the library that all the userspace stuff uses to actually talk to the GPS)16:39
jonwilliblocation has a full dev and doc package which would help16:39
dos1what handles supl server on maemo?16:44
jonwillocation-proxy does it I believe16:45
*** LauRoman has joined #maemo-ssu16:49
DocScrutinizer05sounds good16:51
DocScrutinizer05we#ll most likely use the GPS that comes with Option 601/801 modem16:52
DocScrutinizer05saves us some chips and thus some money and particularly PCB real estate16:52
DocScrutinizer05sure a "free" GPS would be nicer, but sometimes you have to make compromises16:53
jonwilIs there any discussion of Neo900 other than the TMO thread?16:53
jonwilI would like to know the best place to throw some thoughts on what to do with different binary blobs16:53
jonwilone thing we definatly dont need to care about anymore is BME16:54
jonwilsince last I checked, someone had replaced all the bits of BME that actually mattered16:55
Paliyes, bme is not needed16:55
*** LaoLang_cool has joined #maemo-ssu16:56
jon_yis it possible to ship with selinux enabled with the new kernel? permissive of course :)16:56
jonwilso is there anywhere better to discuss Neo900 than the TMO thread?16:57
jonwilor is the TMO thread the best place for my thoughts?16:57
jon_ysome real security on a phone is nice to play with16:58
jon_ythough I don't expect it to go into enforced mode16:58
LaoLang_cooloh, neo900?16:58
jon_y*mandatory mode, not like the n916:58
jon_yLaoLang_cool: pali is working on 3.x kernel16:59
* jon_y wishes to play with selinux on a phone16:59
jonwilI will take the lack of response to mean that the TMO thread is the best place to discuss Neo90016:59
LaoLang_coolit would be slow on n900 I guess16:59
LaoLang_coolpoor hw of n900....16:59
jon_ywell, it is set to permissible17:00
jon_yso, play with it if you want17:00
jon_yit has no effect otherwise17:00
dos1jonwil: TMO thread, this channel and #maemo17:00
dos1jonwil: that's it for now; some other places probably will come later17:00
jonwilI will post my thoughts on what to do with the blobs in the TMO thread17:00
jonwilthen we can discuss it in IRC :)17:01
dos1sounds good ;)17:01
*** dos1 has quit IRC17:01
*** dos1 has joined #maemo-ssu17:02
*** dos1 has quit IRC17:03
*** dos1 has joined #maemo-ssu17:03
*** BCMM has quit IRC17:15
*** LauRoman has quit IRC17:30
*** LaoLang_cool has quit IRC17:38
*** dos1 has quit IRC17:41
*** Vlad_on_the_road has quit IRC17:42
*** NIN101 has joined #maemo-ssu17:42
DocScrutinizer05jonwil: probably creating a new thread for this discussion is best17:47
jonwilmakes sense17:47
DocScrutinizer05linking to new thread from
DocScrutinizer05I planned to do that since some days already, didn't get around to it17:48
DocScrutinizer05jonwil: I'll kick off the thread, you can join in and link to it by editing that post of yours17:51
*** dos1 has joined #maemo-ssu17:53
*** Vlad_on_the_road has joined #maemo-ssu17:54
jonwilbtw, is the plan to use same wifi and BT as N900 or to use something different (i.e. whatever GTA04 already has)?18:00
DocScrutinizer05jonwil: we won't be able to source the wl1251 I guess, and even if we could, it's not really free. I think we got nicer alternatives for wlan18:09
jonwilok, thats what I thought18:10
keriois there a wlan chip that doesn't require a binary blob?18:10
jonwilthe actual wlan/bt chips being used are disconnected enough from the rest of the system that its not an issue for us18:10
DocScrutinizer05but the rest of the chip can either be well documented or not at all documented18:11
DocScrutinizer05I don't know if there are even open firmwares for some chips18:11
DocScrutinizer05some chips don't need firmware upload each time you start them, since they got flash storage for "the blob"18:12
keriomh, i suppose that that's the same, reall18:13
kerioi was thinking of completely open chips18:13
DocScrutinizer05[2013-09-07 17:11:39] <DocScrutinizer05> I don't know if there are even open firmwares for some chips18:16
DocScrutinizer05iirc some guys worked on a hardmac for Prism-II18:17
DocScrutinizer05yeeears ago18:17
DocScrutinizer05even the website vanished meanwhile18:17
DocScrutinizer05prism-II or prism-5418:18
dos1kerio: well, RMS is fine with device with unchangable non-free blob in its internal storage, but he's not fine with device to which you have to upload non-free blob18:21
Palikerio: atheros chips have open also firmware18:22
PaliDocScrutinizer05: ^ (some) atheros are open18:23
DocScrutinizer05not that RMS is any relevant to me or Neo900 here18:23
dos1right, just pointing out that he sees some difference there18:24
DocScrutinizer05my take on that is: any blob you upload to a chip is initialization data, and thus not a subject to FOSS and openness of the system18:24
DocScrutinizer05dos1: I think RMS doesn't ,ake sense in that respect18:25
dos1yeah, I agree with you18:25
dos1for me uploadable blob can be even better, as you can play with it, RE and send your own one instead18:26
kerioDocScrutinizer05: the "issue" is that you're sending something that you don't really understand18:26
kerioinstead of a well-understood "initialize dat shit" command18:26
DocScrutinizer05kerio: you do that all the time18:26
keriowhen dealing with closed hardware18:27
kerioand you don't do that if you don't need to upload a firmware18:27
DocScrutinizer05even when dealing with open hardware18:27
DocScrutinizer05I don't see a difference between "write 0x34af2900 to rgister 4 for enabling XY mode" and "write those 4000bytes to register 9 for enabling wlan engine"18:28
DocScrutinizer05you NEVER have complete insight into what you're really doing, when dealing with peripheral hardware18:29
DocScrutinizer05unless the peripheral *hardware* is documented down to the last resistor18:30
jonwilonly 180 packages left to sort for my post18:31
dos1OT: just removed pulseaudio from my laptop once again18:31
dos1is this shit ever going to work properly?18:32
DocScrutinizer05in fact RMS only speculates that the blob must be some sort of software, and thus his claim that we shall have the source for that18:32
*** Vlad_on_the_road has quit IRC18:33
DocScrutinizer05thinking too much about that will leave you trapped in a catch2218:33
jon_ywell, you can't fault him for trying to look out for us lowly drones18:33
jon_ymore freedone the better18:33
DocScrutinizer05sure. But "we don't believe in Kings and Gods"18:34
DocScrutinizer05RMS has a notion, that's fine. And often his notion is valuable and quite to the point. sometimes not. anyway it's not an commandment for me18:35
dos1he always was and is taking his opinions to extremes, so listening to him literally probably is not the wisest thing to do18:35
dos1but he's still almost always pointing at least at right direction18:36
dos1so I see him as a perfect source of some "food for thought"18:37
jon_yhe's the closest thing to a pope that I have :)18:38
jon_ycertainly influential, but no absolute say18:39
dos1pope of the church of emacs ;)18:39
jonwilchurch of GNU more like :P18:40
jon_yyeah, more likely GNU18:40
dos1so st Ignucius then :)18:41
*** freemangordon_ has joined #maemo-ssu18:41
*** freemangordon has quit IRC18:41
jon_yhe needs a college of Cardinals too, for in case he bites the dust too soon18:43
freemangordon_Pali: yep, it seems to be fixed, will upload the source in my repo in a couple of minutes19:03
freemangordon_FatPhil: if you installed cssu-thumb, you already have the needed kernel19:03
*** freemangordon_ is now known as freemangordon19:03
jonwilthe list covers packages I think we can drop (either due to hardware changes or just stuff we dont need or care about), packages we have actual code (or useable replacements) for, packages we can probably use as-is without change and then various packages we need to make decisions on19:04
jonwili.e. drop, clone, modify, replace, re-use-as-is, whatever19:04
Palijonwil: I updated closed packages wiki page, so all working are now green19:05
FatPhilfreemangordon: are there n900-based build tools? (yes, it takes 2 hours to build the kernel, this I already know)19:07
jonwilso yeah feel free to comment on my list and get discussion going on where we need to focus efforts re cloning things/reverse engineering things/whatever :)19:09
sixwheeledbeastjonwil: your TMO post is lacking in [code] boxes or some sorta formatting ;)19:14
freemangordonFatPhil: afaik yes, but I've never tried it19:21
jonwilno-one has any useful comments on my list?19:22
jonwilthe Cellular Services Daemon certainly looks like the biggest missing piece of the jigsaw puzzle that needs solving19:24
jonwilits by far the biggest undocumented piece of maemo19:26
DocScrutinizer05jonwil: there are probably a truckload of comments to your list. I already posted19:28
DocScrutinizer05I think you can't handle all the possible comments on a single thread in tmo19:28
DocScrutinizer05we should find out about hierarchy of the libs related to CSD19:29
jonwilI already know heirerchy19:29
jonwillibisi and libisi-glib live at the bottom19:30
jonwilthen libraries like libsms0 sit on top of libisi19:30
DocScrutinizer05there must be some sort of hierarchy, like "all other libs base on either directly or via other libs"19:30
jonwilthen the CSD plugins sit on top of those libs19:30
DocScrutinizer05we need to document this19:30
jonwileverything that the CSD plugins do ultimately does end up in libisi19:30
DocScrutinizer05thus my suggestion to port it to a wiki19:31
DocScrutinizer05sorry folks, RL calling19:31
DocScrutinizer05oh, and welcome on board of FPTF, jonwil :-)19:32
dos1Fremantle Porting Task Force :)19:33
DocScrutinizer05Fremantle Porting Task Force19:33
DocScrutinizer05you did a nice and bold introduction19:34
*** arcean has quit IRC19:34
jonwilI think my #1 maemo wish would be to have the docs and details of all the dbus interfaces that are exposed by the cellular services daemon with #2 being some dev packages for a few libraries like libconnui and libconnui-cellular and a couple others :)19:34
jonwilThe good news is that we DO have docs for both the csd-gprs plugin and the csd-net plugin19:35
dos1oh, googled - in package itself :)19:38
DocScrutinizer05needs all to go onto that wiki page!19:38
jonwilWhat we dont have is any docs at all for the csd-call, csd-info, csd-sim, csd-simpb, csd-sms or csd-ss plugins (except for what little has been figured out via reverse engineering or pulled from open pieces of code like bluetooth stack)19:45
jonwilbluez in particular was nice for figuring out a few things :)19:47
DocScrutinizer05well, we can figure a lot of what they do. simpb==SIM phonebook access, ss==supplementary-services, info is a tad obscure. All probably have top end d-bus IF, and bottom end lib-xy (whatever or anything similar)19:49
jonwilyeah we can figure out in broad terms what they do19:50
jonwilbut we need to know what the various dbus interfaces they expose do19:50
DocScrutinizer05the state machines inside those is what we would need to re-create19:50
DocScrutinizer05it's not exactly trivila to send or receive a sms19:51
DocScrutinizer05for example19:51
jonwilwell ofono/fsogsmd presumably already does SMS sending, we would just need to make it expose the same interface as csd-sms (for example)19:51
DocScrutinizer05sned sms PDU, wait for response, handle error response, handle OK response, handle receipt notification, handle unsolicited interleaved responses19:52
DocScrutinizer05yup, basically that's it19:53
DocScrutinizer05handle all that configuration SMs et al cruft19:54
DocScrutinizer05sure FSO (and ofono) are supposed to already know how to handle all this19:55
DocScrutinizer05FSO even works on d-bus, just like CSD does19:56
dos1I'd be willing to work on recreating those APIs based on FSO19:56
dos1if we go that way19:56
DocScrutinizer05I still prefer looking into libisi and what are the methods exposed on it's API upside interface19:57
DocScrutinizer05only ONE blob to RE19:57
jonwilreplacing libisi wont help, we would still need to understand all the undocumented n900 cellmo packets that get sent down from above19:58
DocScrutinizer05everybody talking to libisi "modem do this, modem do that". We should be able to translate to another slang19:58
jonwillibisi is just transport library19:58
jonwilso stuff says to libisi, "connect to modem interface xyz"19:58
DocScrutinizer05umpf, then scrap that idea19:58
jonwilor "send this packet to isi"19:58
DocScrutinizer05I think suddenly replacing CSD by FSO looks _much_ more applealing than it did to me 10 min ago19:59
dos1damn it19:59
dos1is there some easy way to power N900 without usb and battery? :P20:00
DocScrutinizer05err, nope, not really20:00
DocScrutinizer05unless you got a lab power supply20:01
DocScrutinizer05or other *stabilaized* 4.0V power source20:01
DocScrutinizer05o/ bbl20:02
dos1or maybe would soldering loose cables to usb pins be easier than soldering whole socket? :P20:04
keriobuy something that can charge a bl-5j20:04
dos1I'm total noob in this regard, but I'd like to have N900 running even in "stationary mode"20:05
dos1kerio: with my luck and laziness I suspect that scenario when I'll find some time for playing with it I'll find out that battery is flat will happen often :P20:06
dos1but maybe with second battery it wouldn't be a problem...20:06
*** DrCode has quit IRC20:09
*** DrCode has joined #maemo-ssu20:18
FatPhilThe funny thing is that I love my n900, but I haven't done this whole app-installing frenzy that everyone else has done, it's unchanged in about 4 years. I'm now making loads of new discoveries, and I'm liking what I see!20:23
dos1yay, Nokia 5800 with broken screen for ~7 EUR together with shipping20:24
dos1smells like good BL-5J charger :)20:24
freemangordonFatPhil: I wonder why (you did not update your device with CSSU and new applications)20:25
FatPhilI got used to the device that I was working on. Never installed stuff as I knew I'd be reflashing it the next day!20:26
FatPhilmy g/f's put more stuff on hers.20:28
*** DrCode has quit IRC20:29
freemangordonOMG, 4 years after the lat update20:31
freemangordonPali: did not test on 2.6.28, won't be surprised if something doesn't work20:31
*** sunny_s has joined #maemo-ssu20:41
*** arcean has joined #maemo-ssu20:45
*** DrCode has joined #maemo-ssu20:46
*** NIN101 has quit IRC20:54
*** NIN101 has joined #maemo-ssu20:54
*** _ade_ has joined #maemo-ssu20:59
_ade_freemangordon: any idea why all the timezone files are put in the libc6 packages on Maemo?21:00
_ade_freemangordon: makes updating them in a proper way less convenient21:00
freemangordon_ade_: NFC, better ask aapo, he did a lot of libc stuff21:00
_ade_freemangordon: I guess most distros have a separate package for that. Nokia must have had it's reasons21:02
_ade_freemangordon: any news on the RE-ing? :)21:04
freemangordon_ade_: yes, but all of them are bad. It seems I have to spent a month or more on that just to implements an enhancement and to fix a bug which hits once a year. I don't think this is justified, sorry21:05
freemangordon_ade_: unfortunately I spend the time I planned to re that closed blob on fixing :(21:06
*** Sohil876 has joined #maemo-ssu21:09
_ade_freemangordon: that was no RE job in the end. Too bad it can't be done easily. But you should make your own choices anyway, you are not on my payroll :)21:10
freemangordon_ade_: it was not RE job, but this doesn;t mean it was easy :)21:11
freemangordon_ade_: keep in mind there will be lots of REing to happen in regard to Neo90021:11
freemangordonPali: hmm, seems hulda works on 2.6.28 too, at least no visible signs of something got broken because of my changes21:13
Palifreemangordon: ok21:13
freemangordonPali: going to pu it in cssu-devel21:13
freemangordonPali: in the repo, feel free to upgrade21:16
jonwilfreemangordon: What exactly do you want to RE?21:16
Palijonwil: that hildon library21:16
_ade_freemangordon: yeah, I saw the components list of Jonwil on TMO. We need more manpower for that I guess.21:16
freemangordonjonwil: :nod:21:16
freemangordonjonwil: just a second to tell you the name21:16
jonwilwhy do you want to reverse engineer the hildon library for?21:16
jonwilhildon-plugins-notify-sv I think it is21:17
freemangordonbecause it has a bug which sometimes leads to hildon-home being locked with alarm sound playing21:17
freemangordonand it ignores custom alarm sounds21:17
Palithat bug is really anoying...21:18
Palisometimes I'm not able to stop alarm clock21:18
*** NIN101 has quit IRC21:18
jonwilthat library has an x86 version, if I had more skills in the libraries its using (libsoundfile, libprofile, libplayback, pulseaudio, gconf, glib, dbus) I might be able to RE it21:21
*** nox- has joined #maemo-ssu21:21
*** iDont has joined #maemo-ssu21:22
jonwilWhat surprises me is that Nokia went out of their way to do extra bits so they could keep that library closed. I wonder just what that library is doing that Nokia considered important or sensitive21:22
*** DrCode has quit IRC21:22
freemangordonjonwil: yeah, it has x86 version, I did some stuff and can hand you .idb file along with 2 source files if you wish21:22
dos1spying! :D21:23
jonwilyeah that would be nice :)21:23
freemangordonjonwil: it is not that complicated, just that the binary is of vkbrender size and I still remember what it took me to RE vkbrenderer :D21:23
freemangordonjonwil: ok21:24
*** Sohil876 has quit IRC21:25
Palievery systemui package is closed because that SW is nokia legacy and comes from first maemo version on nokia 77021:29
Palionly some few dbus headers was opened (where dbus-monitor told you everything needed)21:30
freemangordonPali: systemui is easy21:30
freemangordonI can RE a plugin per day while I am having my morning coffee :D21:31
freemangordonmaybe the daemon itself could take me a bit more21:31
Palifreemangordon, btw in powerbutton plugin you do not using value which comes from dbus open/close function21:32
freemangordonPali: so?21:32
*** arcean has quit IRC21:32
freemangordonor rather - should I?21:32
PaliI'm trying to find out what original powerbutton plugin doing with it21:32
*** NIN101 has joined #maemo-ssu21:33
Palivalue is DBUS_TYPE_UINT32 and is either MODE_FLIGHT or MODE_NORMAL (depending on selected mode)21:33
*** DrCode has joined #maemo-ssu21:34
freemangordonPali: which argument is that?21:36
Palifirst argument for open method (also close)21:37
freemangordonPali: of open/close callbacks?21:37
freemangordonno :)21:37
freemangordonopen_handler(const char *interface, const char *method, GArray *args, system_ui_data *data, system_ui_handler_arg *out)21:37
PaliI mean first dbus argument of open method21:38
freemangordonoh :)21:38
freemangordonPali: I don;t thin systemui cares about flight/normal mode21:39
freemangordonpower key plugin that is21:40
freemangordonPali: it just passes args to check_plugin_arguments(), but before the call R@ contents is destroyed ( MOV R2, #1)21:42
freemangordonI do the same21:42
_ade_freemangordon: in cssu perspective, how important is it that timezones and summer/wintertime changes are up to date?21:43
Palitryint to decode it, I see that it checking signature of first argument21:43
Palibut it really only checking dbus singature and dropping value?21:43
freemangordon_ade_: I guess on par with certificates or slightly lower priority21:43
freemangordonPali: it drops the whole args parameter21:44
freemangordonPali: yeah, it only checks if the args is what the plugin expects21:44
freemangordonsee disasm I posted21:44
_ade_freemangordon: any plans in that direction? It is "only" a matter of updating the freely available timezone files21:45
*** mkaindl has quit IRC21:46
freemangordon_ade_: I don;t have plans, feel free to patch libc (once it finally enters cssu :) )21:46
_ade_freemangordon: I  rather not touch such an important package. That's why I asked about it. Could it split in an easy and safe way, or is that a nogo?21:52
*** infobot has joined #maemo-ssu21:53
*** ChanServ sets mode: +v infobot21:53
*** arcean has joined #maemo-ssu22:03
*** _rd has joined #maemo-ssu22:05
*** iDont has quit IRC22:09
*** rd_ has joined #maemo-ssu22:14
*** _rd has quit IRC22:18
*** jonwil has quit IRC22:46
*** DrCode has quit IRC22:46
*** arcean has quit IRC22:47
*** rd_ has quit IRC22:52
*** arcean has joined #maemo-ssu23:10
*** Fat_Phil has joined #maemo-ssu23:13
Fat_Philany idea why my custom xkb configuratio for shell-obsessives works in the shell, but not in this irc client?23:18
*** DrCode has joined #maemo-ssu23:23
*** arcean has quit IRC23:23
Fat_Philworks in every app apart from this client apparently.23:24
*** LauRoman has joined #maemo-ssu23:25
*** FlatPhil has joined #maemo-ssu23:26
*** dafox has joined #maemo-ssu23:27
*** Fat_Phil has quit IRC23:33
DocScrutinizer05FatPhil: [2013-09-07 22:33:31] <-- Fat_Phil has left this server (Remote host closed the connection).23:37
DocScrutinizer05alas my client doesn't show re-joins23:37
DocScrutinizer05for your client I either need a DCC VERION request, or you disclose delibaerately what you're using ;-)23:39
DocScrutinizer05why does *my* client still expand the nick, even when the user is gone?23:39
DocScrutinizer05ugh, no "_"23:40
DocScrutinizer05[2013-09-07 22:40:32] [CTCP] Received CTCP-VERSION reply from FatPhil: irssi v0.8.12.23:40
DocScrutinizer05so, NFC23:40
DocScrutinizer05should be plain shell23:40
FlatPhilthe client that didn't let me do {}%|`<>[] from the hw keybd was calles "communi"23:51
FlatPhiloh, not forgetting ~23:51
*** _ade_ has left #maemo-ssu23:53
FlatPhilmy g/f's put many of them on the toolbar for the x terminal, but I prefer them to be universal. hate the popup onscreen keyboard.23:54
DocScrutinizer05 FWIW23:55
DocScrutinizer05try to find out about VI_s awesome keymapping. It's burried deeeeeep inside the bot's factoids23:56
DocScrutinizer05~listvalues vi_23:57
infobotFactoid search of 'vi_' by value returned no results.23:57
DocScrutinizer05~dice 100d2023:58
infobotrolled 2,19,14,18,11,12,18,9,16,14,18,2,4,9,7,7,3,7,5,18,6,4,8,16,19,20,19,17,19,8,14,19,1,18,13,1,9,19,1,15,15,11,5,14,20,20,3,9,14,16,11,3,14,6,4,14,7,18,12,2,20,15,14,5,11,16,7,18,11,12,18,18,11,19,5,17,18,4,9,15,7,4,2,19,20,16,2,5,3,11,8,9,9,1,9,4,18,10,16,7 = 112023:58

Generated by 2.15.1 by Marius Gedminas - find it at!