**** BEGIN LOGGING AT Sun May 15 02:59:59 2011
May 15 03:15:33 <pronto> http://i.imgur.com/ddRQ0.jpg
May 15 04:24:18 <snappy> i'm very confused as to the naming of devices on the n900. Is /dev/mmcblk1 the external SD card in maemo?
May 15 04:25:18 <infobot> DocScrutinizer: infobot joined!
May 15 06:57:47 <kerio> ~botsnack
May 15 06:57:48 <infobot> kerio: thanks
May 15 06:57:54 <kerio> so cute ^_^
May 15 07:44:49 <habmala> Hi, I read this thread because i had LED-issues http://talk.maemo.org/showthread.php?t=55821&highlight=error+status%3A+65280+failed+set+red+led+current&page=2 it said to remove the simcard, wich I tried, and to my big suprise i worked. Does anyone know that's actually happening, a regular reboot doesn't fix it?
May 15 07:47:31 <MohammadAG> the lp5532 (I think) chip gets locked up
May 15 07:47:55 <MohammadAG> apparently, it keeps receiving power even when the phone is off (I'm guessing this is for the LED to show when you hold the power button)
May 15 07:48:06 <MohammadAG> the only way to reset it is to remove the battery
May 15 07:49:23 <MohammadAG> you don't have to remove the SIM btw
May 15 07:50:23 <robbiethe1st> MohammadAG: It doesn't happen to have a reset line, or reset command(via i2C or w/e) does it?
May 15 07:50:30 <lcuk> MohammadAG, reseating the sim is a really good way to ensure the battery also gets removed
May 15 07:51:22 <MohammadAG> lcuk, lies, I replaced my SIM with the device on!
May 15 07:51:25 <MohammadAG> :P
May 15 07:51:37 <MohammadAG> robbiethe1st, no idea, I'm not the best person to give out i2c commands
May 15 07:51:47 <robbiethe1st> heh
May 15 07:52:15 <MohammadAG> btw, the sim replacement part with the device on is actually true :p
May 15 07:52:43 <habmala> okey, thanks MohammadAG
May 15 07:53:26 <habmala> and yes its for lighting up on power up, but also notifications and the keyboard backlight..
May 15 07:55:16 <habmala> ooh, I just understood that last part MohammadAG.. hehe my bad ;) and yes that seems reasonable..
May 15 07:55:53 <MohammadAG> I meant the blue light at startup :)
May 15 07:56:37 <habmala> Yes, I re-read the message and realised I had misunderstood you ;)
May 15 07:57:00 <Evanescence> can I install hildon-desktop on my ubuntu system ?
May 15 07:57:16 <MohammadAG> yes, not directly though
May 15 07:57:22 <MohammadAG> see cordia's repos
May 15 08:01:53 <lcuk> MohammadAG, you did that because you are awesome :P
May 15 08:02:03 <lcuk> but most mere mortals would have trouble
May 15 08:03:28 <MohammadAG> offline mode, connect charger, stop bme, remove battery, replace sim in under 25s, replace battery in under 5s, unplug charger, online mode, start bme
May 15 08:03:42 <MohammadAG> easy as 123...45...
May 15 08:07:54 <lcuk> MohammadAG, but that requires dexterity - like the superspeed cup stacker people
May 15 08:09:15 <lcuk> MohammadAG, this is how I imagine you doing what you just said: http://www.youtube.com/watch?v=lY6fyRAGH78
May 15 08:11:28 <MohammadAG> that was fast o_O
May 15 08:17:01 <topro> Hi, anyone here knows what's happened to eCoach development? Website and SVN seems dead for more than one year now.
May 15 08:18:48 <lcuk> topro, Sampo has been doing work stuff
May 15 08:21:58 <topro> lcuk, I would like to implement some stuff into ecoach, but first of all I would need the latest version, qt based especially. Would like to have it working on maemo and meego
May 15 08:22:23 <lcuk> topro, then send sampo an email and ask him
May 15 08:22:47 <topro> kk, thanks
May 15 08:25:39 <robbiethe1st> Ok, f***. does anyone have a nice list of (py)QT4 widgets?
May 15 08:26:05 <robbiethe1st> Like, all of them, with pictures and such? I remember there being such a page, but I cannot find it anymore
May 15 08:26:41 <robbiethe1st> oh, wait. there we go
May 15 08:36:27 <zakkkkm> if i have power-kernel.. could i use uboot?
May 15 08:42:11 <MohammadAG> robbiethe1st, just use designer and look at the widgets there :)
May 15 08:43:25 <robbiethe1st> Tried that once; couldn't figure out the proper names of the widgets in my 5 minutes of fiddling
May 15 08:44:51 <RST38h> http://1.bp.blogspot.com/-pThp9ZQe0Jc/Tc-H1TMWe1I/AAAAAAAADa8/X2yTbfisB8g/s640/zombie-orig.jpg
May 15 08:45:13 <RST38h> Police vs Zombies
May 15 08:51:52 <habmala> hey, robbiethe1st whats that site? sounds interesting ;)
May 15 08:52:19 <robbiethe1st> http://doc.qt.nokia.com/4.5/gallery-cleanlooks.html
May 15 08:52:38 <robbiethe1st> Seems you have to search by "qt widgets" instead of "qt4 widgets" for some odd reason
May 15 08:52:49 <habmala> k, thanks
May 15 08:57:33 * MohammadAG wants a spinning circle Qt widget
May 15 09:05:28 <RST38h> Doc: Tokyo Electric Power Company, or TEPCO, says the meltdown at the No.1 unit
May 15 09:05:29 <RST38h> created holes in the reactor and damaged the containment vessel.
May 15 09:05:37 <RST38h> A large amount of highly radioactive water is believed to be leaking out, but it is
May 15 09:05:38 <RST38h> not known where it is flowing.
May 15 09:45:52 <DocScrutinizer> well, I always was sure there's been a meltdown, even though they claimed it's been not - fools
May 15 09:51:03 <RST38h> Doc: Meanwhile: http://img-fotki.yandex.ru/get/5604/bender008.20/0_66075_4d7424e4_orig
May 15 10:06:11 <DocScrutinizer> well, if they filled the groundwater with large amounts of caesium-137, then this will soon be a normal view in Japan
May 15 10:07:55 <DocScrutinizer> zakkkkm: yes, there's a pk+uboot pkg courtesy MohammadAG
May 15 10:18:59 <DocScrutinizer> ~uptime
May 15 10:19:18 <DocScrutinizer> ~botsnack
May 15 10:19:18 <infobot> thanks, DocScrutinizer
May 15 10:19:25 <RST38h> Doc: I prefer the cat-eared schoolgirls version =)
May 15 10:19:46 <DocScrutinizer> X-Fade: could you please set infobot +V
May 15 10:20:28 <ShadowJK> RST38h, seen the robotic cat ears that read your brainwaves to determine your mood and moves the cat ears accordingly?
May 15 10:21:01 <RST38h> ShadowJK: Yea, the technological progress is unbelieavable nowadays =)
May 15 10:21:13 <RST38h> ShadowJK: Now, of course, they will also have to make the tail...
May 15 10:21:19 <ShadowJK> lol
May 15 10:22:00 <MohammadAG> but it's power46 not power47
May 15 10:22:02 <maybeHere> so, how much longer until http://26.media.tumblr.com/tumblr_lhjkhtbRR31qh3bh2o1_500.jpg ?
May 15 10:34:20 <DocScrutinizer> MohammadAG: and it's probably still missing the iso9660.ko and/or other modules needed to access DVD/CD
May 15 10:35:52 <MohammadAG> DocScrutinizer, I don't handle modules in it
May 15 10:35:59 <MohammadAG> it's just a zimage
May 15 10:37:53 <Andy80> a simple question about packaging: I'm going to package a library that is available on gitorious. It doesn't have any version number, but dh_make want the folder to be in the versione "application-version", so... what version number should I use instead?
May 15 10:38:45 <RST38h> 1.0.0 ?
May 15 10:39:24 <Andy80> RST38h: but... using 1.0.0 looks like a stable version, while this one is far from being stable at the moment
May 15 10:39:29 <DocScrutinizer> 0.1.0?
May 15 10:39:33 <DocScrutinizer> 0.0.1?
May 15 10:39:50 <MohammadAG> 0.1 imo
May 15 10:40:08 <RST38h> Japanese government is extending the evacuation zone to the areas where annual radiation levels exceeded 20mSv
May 15 10:40:34 <Andy80> yeah... 0.1 or 0.1.0 would be better...
May 15 10:40:43 <RST38h> +7700 people to evacuate
May 15 10:41:10 <MohammadAG> strictly speaking, that's the package version rather than the library's
May 15 10:42:20 <Andy80> MohammadAG: oh, you're right!
May 15 10:44:12 <MohammadAG> :)
May 15 10:51:52 <MohammadAG> symbian.nokia.com sucks
May 15 10:51:59 <MohammadAG> The credentials you provided cannot be determined to be authentic.
May 15 10:52:43 <RST38h> They won't be around for much longer anyway
May 15 11:05:03 <DocScrutinizer> Andy80: http://www.tldp.org/HOWTO/Program-Library-HOWTO/shared-libraries.html plus/minus a few chapters
May 15 11:07:54 <MohammadAG> that's not related to debian packaging DocScrutinizer :p
May 15 11:08:09 <DocScrutinizer> but definitely to lib naming
May 15 11:08:32 <DocScrutinizer> thought it could help if somebody doesn't even know the diff
May 15 11:09:24 <Andy80> just a note, it's not a library I've written... I'm just packaging
May 15 11:10:03 <MohammadAG> yeah
May 15 11:10:21 <MohammadAG> besides rebuilding a kernel, is there a way to change the CMDLINE?
May 15 11:10:34 <MohammadAG> uboot isn't an option, it can't read eMMC
May 15 11:11:25 <MohammadAG> (or is there a way to decompress a zImage? I guess I could sed it)
May 15 11:12:01 <DocScrutinizer> MohammadAG: messing with libs is sth that needs quite a bit of expertise and taking responsibility, as other devs rely on your proper management of things like lib naming, ABI compatibility etc
May 15 11:12:25 <SpeedEvil> MohammadAG: Install meego on SD - then do whatever?
May 15 11:13:39 <DocScrutinizer> MohammadAG: of course you can decomrpess zimage, it's a normal gp/bz whatever
May 15 11:13:40 <MohammadAG> SpeedEvil, I don't have an SD card
May 15 11:13:44 <DocScrutinizer> gz*
May 15 11:13:57 <MohammadAG> I already have MeeGo on my eMMC, I just need to permanently flash the kernel
May 15 11:14:07 <MohammadAG> I tried gzip -d, no dice
May 15 11:14:16 <DocScrutinizer> try bzip
May 15 11:14:30 <SpeedEvil> MohammadAG: not even a teeny 1G one?
May 15 11:14:48 <DocScrutinizer> or try ark which has some heuristics to select the right decompressor
May 15 11:15:02 <MohammadAG> bzip2, tried that too
May 15 11:15:10 <MohammadAG> SpeedEvil, nope, had a 4GB one, it died
May 15 11:15:18 <MohammadAG> and the 16GB one isn't for testing
May 15 11:17:15 <DocScrutinizer> MohammadAG: http://www.google.de/search?q=decompress+zImage
May 15 11:18:25 <MohammadAG> http://www.codeguru.com/forum/archive/index.php/t-415186.html
May 15 11:23:37 <DocScrutinizer> MohammadAG: I'm sure there are tools that are aware of the properties of the auto-decompressor code that's heading the plain compressed kernel code
May 15 11:24:14 <DocScrutinizer> maybe even the decompressor itself in vmlinuz can get used to write the expanded linux kernel to somewhere
May 15 11:29:11 <MohammadAG> none of the methods are working
May 15 11:29:11 <DocScrutinizer> MohammadAG: I seem to recall there are scripts to `make vmlinuz` for maemo. The tools called there and files and filesizes used should straightly lead you to the answer how to decompress vmlinuz
May 15 11:29:23 <MohammadAG> I wonder if it's not a vmlinuz image and meego's just using a wrong name
May 15 11:30:32 <ShadowJK> I would have thought that since MeeGo uses uboot, it'd be a uImage?
May 15 11:30:36 <DocScrutinizer> given the limited amount of space on that kernel mtdpart, I bet it's a compressed image
May 15 11:31:12 <ShadowJK> Though there's a long standing tradition by various distros to name bzimage and zimage as "vmlinuz" on the filesystem
May 15 11:31:26 <DocScrutinizer> ShadowJK: the kernel vmlinu? image is a selfextracting binary basically
May 15 11:32:44 <DocScrutinizer> ShadowJK: so uBoot doesn't care, no bootloader does
May 15 11:33:28 <ShadowJK> dunno, for my Sheevaplug's uboot I have to make the kernel with "make uImage", iirc
May 15 11:33:32 <DocScrutinizer> load vmlinuz into ram, jump to startaddr
May 15 11:34:08 <DocScrutinizer> the kernel then basically decompresses itself
May 15 11:34:13 <ShadowJK> (and then the memory address in kernel .conf and uboot have to match)
May 15 11:35:39 <DocScrutinizer> of course you can 'easily' create a userland tool that does exactly the same, and stopps execution after decompressing
May 15 11:38:42 <MohammadAG> ShadowJK, I have both
May 15 11:38:49 <MohammadAG> (a uImage and a zImage)
May 15 11:40:45 <MohammadAG> I'm out of ideas :/
May 15 11:44:23 <DocScrutinizer> MohammadAG: learn to build a maemo z/uImage, check and understand the used tools, and you got your answer how to reverse the process
May 15 11:44:43 <MohammadAG> I know how to build a maemo image...
May 15 11:45:02 <MohammadAG> it uses simple tools like gzip
May 15 11:45:07 <DocScrutinizer> fine, so step 1 acomplished
May 15 11:45:16 <MohammadAG> but a zImage is self-decompressing, so the actual gzip file isn't at the start of the file
May 15 11:45:31 <MohammadAG> the gzip header is 8b1f
May 15 11:46:06 <DocScrutinizer> yes, obviously while building the zImage somebody gets a file of binary decompressor code and glues that at head of zImage
May 15 11:46:20 <MohammadAG> I found 008b1f0 2432 82c6 3474 32dd 96a8 8588 e377 c913
May 15 11:46:38 <DocScrutinizer> find out who's doing that and what size the used code is, and you're done
May 15 11:46:51 <MohammadAG> this isn't a maemo image, it's a meego image
May 15 11:47:13 <MohammadAG> and I can't find the kernel sources for that exact image, and chances are if i do find them, they'll be heavily based off RPM stuff
May 15 11:48:26 <DocScrutinizer> or you use the brute force approach and try to unzip from addr:0 to addr:30000 and see where unzip doesn't barf ;-)
May 15 11:48:38 <MohammadAG> I thought of that
May 15 11:48:44 <MohammadAG> is there a sane way to do it? :P
May 15 11:49:09 <DocScrutinizer> for ((i=0; i<30000; i++)); do...
May 15 11:49:34 <DocScrutinizer> (bashism!)
May 15 11:50:59 <DocScrutinizer> test you script with a dummy first! create a (small) .bz and add 20bytes of filler to head - feed that to your script
May 15 11:51:41 <MohammadAG> how do I break a loop in bash?
May 15 11:51:47 <DocScrutinizer> make sure your script prints "20" or even stops with uncompressed .bz file
May 15 11:52:43 <DocScrutinizer> help break
May 15 11:52:59 <DocScrutinizer> break: break [n]
May 15 11:53:01 <DocScrutinizer>     Exit for, while, or until loops.
May 15 11:53:09 <MohammadAG> this'll take a while :/
May 15 11:53:20 <DocScrutinizer> of course a exit works as well ;-)
May 15 11:57:32 <MohammadAG> hmm
May 15 11:57:42 <MohammadAG> technically, this'll take 112.6 days
May 15 11:58:03 <MohammadAG> each attempt takes 5s
May 15 11:58:04 <SpeedEvil> umm...
May 15 11:58:21 <SpeedEvil> read man hexdump
May 15 11:58:34 <SpeedEvil> find out how to get it to output a single line of hex digits
May 15 11:58:36 <MohammadAG> I tried hexdump and got 008b1f0 2432 82c6 3474 32dd 96a8 8588 e377 c913
May 15 11:58:41 <SpeedEvil> now, search for the magic string
May 15 11:58:49 <MohammadAG> that's it ^
May 15 11:58:59 <SpeedEvil> if this is at position 3993940, then divide this by two, and use dd
May 15 11:59:02 <SpeedEvil> job done.
May 15 11:59:29 <MohammadAG> didn't get the last part
May 15 12:00:12 <SpeedEvil> Hexdump is configurable. It can also be set to output a single unbroken line with two output bytes representing one input byte
May 15 12:00:22 <SpeedEvil> you can then simply text search for your desired string
May 15 12:01:15 <MohammadAG> isn't that the line I pasted above?
May 15 12:01:24 <MohammadAG> I'm not sure how to get the number to use with dd from that
May 15 12:01:26 <DocScrutinizer> MohammadAG: wut? on a beagleboard?
May 15 12:01:47 <MohammadAG> nope
May 15 12:01:51 <SpeedEvil> 2432 82c6 3474 32dd 96a8 8588 e377 c913
May 15 12:01:55 <SpeedEvil> remove the spaces
May 15 12:02:04 <SpeedEvil> then search for that string in the unbroken line from your kernel
May 15 12:02:18 <SpeedEvil> probably only the first several bytes
May 15 12:03:57 <DocScrutinizer> umm, what's that byte sequence?
May 15 12:04:43 <DocScrutinizer> some kinda global magic word?
May 15 12:04:55 <DocScrutinizer> or maybe it's x86 specific?
May 15 12:05:33 <MohammadAG> that shouldn't matter in this case though, I only want the gzipped image which is arch independent
May 15 12:05:38 <MohammadAG> (at least as a compressed image)
May 15 12:05:46 <DocScrutinizer> WUT?
May 15 12:05:56 <DocScrutinizer> ok, so it's a magic word then
May 15 12:06:18 <DocScrutinizer> I.E header of the (b)ziped kernel
May 15 12:06:38 <MohammadAG> yeah 8b1f
May 15 12:07:41 <MohammadAG> though I'm not sure what's meant by unbroken line
May 15 12:08:13 * DocScrutinizer wonders if unzip had a repair option that does all this magic by itself
May 15 12:08:36 <DocScrutinizer> MohammadAG: no linefeeds, CRs, or white spaces
May 15 12:08:53 <DocScrutinizer> which could break your search pattern
May 15 12:09:29 <DocScrutinizer> if you're searchin for FFFFFFFF, the a sequence like "FF FF\nFF FF" won't match
May 15 12:11:19 <MohammadAG> but I already found the magic header
May 15 12:11:21 <MohammadAG> http://pastebin.com/raw.php?i=P1ZVUP2u
May 15 12:11:22 <DocScrutinizer> MohammadAG: I also cloudly seem to recal meego changed compression algo (like from gzip to b2zip) eventually
May 15 12:12:32 <DocScrutinizer> that's a minor change as only the autoexec decompressor header of zImage is affected usually. Not in your case though
May 15 12:12:37 <DocScrutinizer> just take care
May 15 12:15:04 <DocScrutinizer> MohammadAG: 468b 1f2e
May 15 12:15:10 <DocScrutinizer> :-o
May 15 12:24:25 <DocScrutinizer> MohammadAG: od -t x1 -w30000 <file> might help
May 15 12:25:38 <MohammadAG> that displays one line, yeah
May 15 12:36:53 <DocScrutinizer> probably a veritable hexeditor is even better as it shows offset, and doesn't (shouldn't - i've seen others) care about "linebreaks"
May 15 12:39:21 <AcTiVaTe> Hey all, after a recent flash and restore I am having some problems. One of the biggest is the recent calls in my phonelist not showing any entries or updating anymore.
May 15 12:40:10 <AcTiVaTe> Any repair I have to do on any sqlite db's for this or do I simply have to recreate a db?
May 15 12:45:59 <AcTiVaTe> recreating the DB seems to be the answer
May 15 12:49:51 <DocScrutinizer> yep
May 15 12:50:39 <DocScrutinizer> seems the format of db changed from some PRn-1 to PRn
May 15 12:50:49 <DocScrutinizer> so restore breaks that db
May 15 12:52:26 <DocScrutinizer> mere handwaving, I couldn't share a pointer to some ticket or tmo thread about it
May 15 12:55:58 <n900-dk> Is there a way to see apps installation log? I guess I have installed something eating my batteri the other day, but can't remember the name of the app..
May 15 12:56:15 <n900-dk> s/batteri/battery/
May 15 12:56:15 <infobot> n900-dk meant: Is there a way to see apps installation log? I guess I have installed something eating my battery the other day, but can't remember the name of the app..
May 15 13:04:08 <n900-dk> or maybe a tool to see what is eating battery..
May 15 13:04:31 <SpeedEvil> powertop
May 15 13:10:14 * n900-dk puts on the glasses to interpret powertops output..
May 15 13:10:32 <SpeedEvil> See also http://images.4chan.org/b/src/1305463695361.jpg
May 15 13:10:36 <SpeedEvil> err - not that.
May 15 13:11:14 <SpeedEvil> http://wiki.maemo.org/N900_Software_Power_management
May 15 13:12:08 <n900-dk> heh, like the first one better :) But thanks
May 15 13:17:53 <rd> does anybody know which qemu version the autobuilder includes?
May 15 13:44:37 <merlin1991> SpeedEvil: lurks on b ?
May 15 13:45:44 <SpeedEvil> Sometimes.
May 15 13:45:56 <SpeedEvil> Just for the pictures of kittys.
May 15 13:46:57 <merlin1991> that wash-basin is defenitely awesome :D
May 15 14:22:57 <DocScrutinizer> merlin1991: ??
May 15 14:23:13 <merlin1991> DocScrutinizer: ??
May 15 14:23:26 <DocScrutinizer> which wash-basin
May 15 14:23:37 <merlin1991> SpeedEvils 4chan link
May 15 14:23:39 <merlin1991> prob 404 by now
May 15 14:23:43 <DocScrutinizer> yes
May 15 14:24:31 <DocScrutinizer> weird shit, I never even take notice of URLs containing /b/
May 15 14:24:51 <DocScrutinizer> or 4chan
May 15 14:25:08 <DocScrutinizer> tbh I have no idea what this is all about
May 15 14:25:12 <merlin1991> then you miss out on those 0,000001% of good links :D
May 15 14:25:24 <merlin1991> (within all the 4chan links)
May 15 14:27:28 <infobot> msg from chanop: pupnik, please stay on topic and DO NOT INSULT OR GO XXX, this isn't #politics and (usually) a worksafe chan
May 15 14:27:39 <pupnik> lol
May 15 14:28:00 <merlin1991> pupnik: ftw
May 15 14:28:12 <vldcnst> #politics part was right on!
May 15 14:28:19 * merlin1991 wonders what nsfw link pupnik posted
May 15 14:28:50 <pupnik> now playing http://dissentradio.com/radio/11_05_13_woods.mp3
May 15 14:29:02 <DocScrutinizer> no link, just plain insulting XXX
May 15 14:29:29 <vldcnst> DocScrutinizer: can I get a 'greeting' too?
May 15 14:29:31 <pupnik> it's hard to switch modes when coming off other channels
May 15 14:30:47 <n900-dk> How can I see, which processes is using swap mem?
May 15 14:31:00 <DocScrutinizer> htop
May 15 14:31:55 <DocScrutinizer> vldcnst: no
May 15 14:32:04 <pupnik> wanted to drop this:  if you want to du -s * in MyDocs, and want to see .sounds, .video, and you're using bash, type "shopt -s dotglob"  and bash will show you hidden directories with du
May 15 14:32:48 <DocScrutinizer> :-D
May 15 14:32:54 <DocScrutinizer> nice hint
May 15 14:35:25 <n900-dk> hmm, there is no swap column in htop
May 15 14:54:41 <compengi> when i install a new application, the new created shortcut icon in the main menu section isn't set. it's only updated after rebooting the device. does anyone knows why?
May 15 14:55:33 <trx> thats normal
May 15 14:56:17 <compengi> oh
May 15 14:56:18 <compengi> really?
May 15 14:56:19 <DocScrutinizer> seems this possibly is related to a buggy sequence of installing files. The desktop monitors the .desktop dir and recaches icon cache as soon as a new file appears there
May 15 14:56:31 <trx> you need to clear the cache
May 15 14:56:38 <trx> reload, etc
May 15 14:56:42 <DocScrutinizer> if a packet installs .desktop first, then the icon file
May 15 14:56:43 <trx> cant remember exactly
May 15 14:57:11 <DocScrutinizer> touch your.desktop should do the trick
May 15 14:57:30 <DocScrutinizer> if that hypothesis is true
May 15 14:57:32 <DocScrutinizer> :-D
May 15 14:57:35 <trx> :)
May 15 14:59:20 * DocScrutinizer wonders if that recaching process could use a 'sleep 5' after detecting inotify for *.desktop
May 15 15:01:51 <DocScrutinizer> as .deb are rarely built in a way so dependency files like an icon get installed prior to the files depending on them like .desktop
May 15 15:04:01 <DocScrutinizer> I only wish I could bother to investigate this enough for opening a ticket against it
May 15 15:06:40 <compengi> DocScrutinizer, yeah, it's annoying
May 15 15:06:55 <compengi> i don't remember this before cssu update
May 15 15:08:47 <DocScrutinizer> compengi: it's a race and always been there (always assuming my explanation holds some truth)
May 15 15:09:32 <DocScrutinizer> MohammadAG: nevertheless we might be able to fix ^^^this in cssu, no?
May 15 15:10:40 <DocScrutinizer> just add a usleep(5000000) after the inotify callback, in whatever does that icon cache update
May 15 15:11:00 <DocScrutinizer> hildon-desktop?
May 15 15:11:44 * GAN900 yawns.
May 15 15:11:51 <GAN900> Have I missed anything interesting?
May 15 15:11:55 * DocScrutinizer reyawns
May 15 15:12:10 <DocScrutinizer> no
May 15 15:12:35 <DocScrutinizer> unless I missed it too ;-D
May 15 15:15:01 * MohammadAG reads
May 15 15:15:24 <DocScrutinizer> well, to whom it may concern: SHR got call audio semi-working on N900
May 15 15:15:50 <MohammadAG> it's not related to the CSSU
May 15 15:16:04 <DocScrutinizer> it's actually a matter of days rather than months now
May 15 15:16:06 <MohammadAG> it happened ever since PR1.1 removed the gtk-icon cache script
May 15 15:16:21 <DocScrutinizer> MohammadAG: that's well understood
May 15 15:16:31 <MohammadAG> <compengi> i don't remember this before cssu update
May 15 15:16:35 <DocScrutinizer> MohammadAG: nevertheless we could *fix* it in cssu
May 15 15:16:42 <MohammadAG> true
May 15 15:18:40 <DocScrutinizer> make that a usleep(15000000)
May 15 15:19:03 <DocScrutinizer> just to be on safe side of fence
May 15 15:20:39 <DocScrutinizer> even better: "batch" it (means only start the icon cache update when CPU load drops to a level indicating HAM/apt is done)
May 15 15:21:41 <DocScrutinizer> hmm, the sleep is quite probably just 'nifty' enough
May 15 15:21:53 <DocScrutinizer> and a nice one-liner patch
May 15 15:22:46 <DocScrutinizer> while batching might get awkward to do
May 15 15:23:55 <n900-dk> How come Sociality can start so quickly, when I click the icon? It pops up immediately. When I start it from terminal it takes a little before I see it
May 15 15:23:57 <trx> it would be better to implement it in HAM
May 15 15:24:09 <DocScrutinizer> if update-icon-cache is a spwned process, you could lower prio (aka nice) of it to be below that of apt-worker
May 15 15:24:38 <DocScrutinizer> n900-dk: fake, and mirrors ;-P
May 15 15:24:42 <trx> n900-dk it loads an image first afaik
May 15 15:24:56 <trx> or w/e
May 15 15:25:15 <n900-dk> but it runs the same command from shortcut as from terminal, right?
May 15 15:25:45 <trx> probably
May 15 15:25:55 <trx> check the .desktop file for more info
May 15 15:26:20 <DocScrutinizer> that's this unbearable "show a screenshot" default behaviour of PR1.3 hildon, which I'd love to kill all together on my system
May 15 15:27:47 <DocScrutinizer> ask MohammadAG
May 15 15:28:13 <DocScrutinizer> he knows better about the right names, and how it's working exactly
May 15 15:29:09 <n900-dk> ok thanks, just wondering, maybe Sociality was running in the background or something :)
May 15 15:29:27 <MohammadAG> <n900-dk> How come Sociality can start so quickly, when I click the icon? It pops up immediately. When I start it from terminal it takes a little before I see it
May 15 15:29:30 <DocScrutinizer> no, hildon has a dir with screenshots
May 15 15:29:48 <MohammadAG> I fake startup, iPhone-style
May 15 15:29:48 <MohammadAG> well, hildon-desktop does, I just found a way to do it in Qt apps :p
May 15 15:30:16 <MohammadAG> DocScrutinizer, make /home/user/.cache/launch non-writable
May 15 15:30:24 <MohammadAG> then delete all pvrs in it
May 15 15:30:26 <DocScrutinizer> thanks :-D
May 15 15:30:39 <DocScrutinizer> pvr was the friggin buzzword
May 15 15:30:53 <trx> hmm
May 15 15:31:16 <trx> MohammadAG maybe that could be used as a splash screen?
May 15 15:31:34 <DocScrutinizer> it IS a splashscreen
May 15 15:31:51 <trx> oh, its like that
May 15 15:31:58 <DocScrutinizer> just it's a friggin "autogenerated" screenshot
May 15 15:32:48 <trx> makes more sense to why it exists :)
May 15 15:33:06 <trx> i allways found it rather stupid..
May 15 15:33:10 * DocScrutinizer is pondering to imagemagic -dim 50% *.pvr
May 15 15:34:09 <DocScrutinizer> so you at least can tell there's no use to click like a fool on settings screen, as it's only the splash and you need to wait another 3s
May 15 15:35:03 <DocScrutinizer> trx: it's definitely stupid
May 15 15:35:41 <trx> 50% dim is actually good idea
May 15 15:35:49 <trx> or whatever effect
May 15 15:36:23 <DocScrutinizer> superinpose a stinky finger ;-P
May 15 15:36:54 <MohammadAG> That's not a bad idea, except making it configurable would be better
May 15 15:37:08 <DocScrutinizer> also more robust as multiple processing doesn't change an already processed image
May 15 15:37:40 <n900-dk> Does powerkernel eat more battery, even when overclocking is not enabled?
May 15 15:37:42 <DocScrutinizer> while multiple dim50 would result in a black screen eventually
May 15 15:37:48 <DocScrutinizer> no
May 15 15:38:03 <DocScrutinizer> not per se
May 15 15:38:32 <n900-dk> cool
May 15 15:39:20 <DocScrutinizer> there are some quirks that may affect your battery management and cpu load, which show up as soon as you do certain things only possible on PK though
May 15 15:39:44 <DocScrutinizer> like modprobe bq27200
May 15 15:39:50 <n900-dk> IPv6 one of them?
May 15 15:39:56 <DocScrutinizer> dunno
May 15 15:40:00 <DocScrutinizer> potentially
May 15 15:40:17 <trx> or do dim 100%
May 15 15:40:27 <DocScrutinizer> yeah :-P
May 15 15:40:30 <trx> then do -10% over few seconds
May 15 15:40:48 <DocScrutinizer> umm
May 15 15:41:38 <DocScrutinizer> trx: you are thinking of pvr. and about dynamic effects. I suggested a onetime processing of *.pvr, as that's easy to implement with a initscript
May 15 15:41:46 <MohammadAG> <trx> MohammadAG maybe that could be used as a splash screen?
May 15 15:42:01 <MohammadAG> in PSFreedom I implemented it as "PSFreedom is loading, please wait..."
May 15 15:42:06 <MohammadAG> rather than a fake startup
May 15 15:42:37 <trx> ill have a look at that
May 15 15:42:47 <trx> DocScrutinizer yeah good point
May 15 15:42:50 <DocScrutinizer> could anybody find me a gfx tool available on maemo, to superimpose something into those *.pvr?
May 15 15:42:52 <MohammadAG> it's just a PVR I created externally
May 15 15:43:14 <MohammadAG> DocScrutinizer, http://talk.maemo.org/showthread.php?p=837973#post837973
May 15 15:44:10 <trx> lol @ post date
May 15 15:44:14 <trx> 10-10-10
May 15 15:46:29 <MohammadAG> says 10-11-10 for me :P
May 15 15:47:29 <trx> 23:48
May 15 15:47:44 <trx> i am in GMT+1
May 15 15:49:02 <MohammadAG> GMT +3 here
May 15 15:49:09 <MohammadAG> (well, +2 +1 DST)
May 15 15:49:26 <trx> well that explains it then :)
May 15 15:50:16 <DocScrutinizer> to the point (pvr): http://talk.maemo.org/showthread.php?p=1001784#post1001784
May 15 15:52:07 <compengi> anyone here uses maemo sdk on gentoo?
May 15 15:52:53 <BigWookie> hi, i know this sounds ridicolos but any advice to kernel debugging on a n900?
May 15 15:53:50 <jacekowski> don't do it
May 15 15:54:52 <BigWookie> I use uboot to boot from mmc and try to run webos, the kernel starts but at a certain stage it hangs and flashes the display
May 15 15:55:28 <DocScrutinizer> MohammadAG: how would I *edit* *.pvrs?
May 15 15:55:52 <jacekowski> BigWookie: well, you would have to connect to debug ports under the battery
May 15 15:56:20 <BigWookie> okay, so best thing is forget about the whole thing ;_;
May 15 15:57:10 <DocScrutinizer> kernel debugging without a serial console working from very start is rather hard
May 15 15:57:42 <DocScrutinizer> that's why OM shipped debug board with Neo
May 15 15:57:45 <BigWookie> :/ yeah
May 15 15:58:51 <DocScrutinizer> MohammadAG: what a format are those *.pvr?
May 15 15:58:58 <DocScrutinizer> special?
May 15 15:59:05 <rd> I compiled navit myself and the menu icon disappeared. Can somebody tell, where it should be located?
May 15 15:59:56 <MohammadAG> DocScrutinizer, PowerVR
May 15 16:00:08 <MohammadAG> I found no way to make them outside hildon-desktop
May 15 16:00:25 <MohammadAG> besides the iPhone SDK I think, though I never tried that
May 15 16:00:52 <DocScrutinizer> wtf PowerVR? o.O
May 15 16:00:54 <trx> rd opt/usr/share/icons/hicolor/*
May 15 16:01:08 <MohammadAG> PVR = PowerVR image
May 15 16:01:18 <DocScrutinizer> never heard of it
May 15 16:01:27 <MohammadAG> it's used extensively on iOS
May 15 16:01:38 <DocScrutinizer> meh!!! :-/
May 15 16:01:51 <MohammadAG> kinda like PNGs, except they're accelerated by the SGX
May 15 16:01:57 <DocScrutinizer> ouch
May 15 16:02:24 <MohammadAG> though the quality of them is like 24-bit BMPs tbh
May 15 16:02:31 <DocScrutinizer> so is there any converter pvr-> <any-sane-common-format> ?
May 15 16:03:23 <DocScrutinizer> I gather your pvr-creator is just converting png -> pvr, but not the other way round
May 15 16:04:06 <MohammadAG> no, it's just taking a screenshot of the window using libhildon's methods
May 15 16:04:20 <DocScrutinizer> well, whatever the source
May 15 16:05:01 <DocScrutinizer> could you add a function to manipulate the resulting pvr?
May 15 16:05:20 <DocScrutinizer> like superimposing some png?
May 15 16:05:38 <DocScrutinizer> or simply dimming
May 15 16:05:51 <DocScrutinizer> dimming should be rather simple
May 15 16:06:01 <MohammadAG> lemme point you to the function
May 15 16:06:11 <DocScrutinizer> sure, go ahead
May 15 16:06:24 <GAN900> Ah, gaffer's tape is the best
May 15 16:06:31 <GAN900> No more rattlin' Lenovo battery.
May 15 16:07:05 <DocScrutinizer> but then, wait a minute, the friggin pvr for e.g settings is autogenerated by hildon, no pvr-creator involved at all
May 15 16:08:01 <MohammadAG> https://gitorious.org/community-ssu/hildon-desktop/blobs/master/src/home/hd-home.c#line1031
May 15 16:08:13 <MohammadAG> DocScrutinizer, it's a hildon-desktop function
May 15 16:08:19 <MohammadAG> all devs can add it into apps
May 15 16:08:19 <DocScrutinizer> aah
May 15 16:08:27 <DocScrutinizer> yeah
May 15 16:08:29 <DocScrutinizer> sure
May 15 16:08:38 <DocScrutinizer> I'd like to tweak it
May 15 16:08:41 <MohammadAG> my app just sets the image you provide as a background then takes a screenshot
May 15 16:09:00 <DocScrutinizer> mhm
May 15 16:10:13 <DocScrutinizer> well, that all is awkward for my purpose. I simply want to *edit*  the *.pvr files, no matter who created them
May 15 16:11:07 <rd> why is /opt needed, why is /usr/share/icons/hicolor/* not ok?
May 15 16:11:17 <rd> http://pastebin.ca/2060493
May 15 16:11:28 <rd> seems marble uses just /usr, no opt...
May 15 16:12:20 <DocScrutinizer> ~optification
May 15 16:12:21 <infobot> from memory, optification is a inventive duct tape workaround to reclaim space in fs root, done due to the fact the partitioning is FUBAR, or http://wiki.maemo.org/Documentation/Maemo_5_Developer_Guide/Packaging,_Deploying_and_Distributing/Installing_under_opt_and_MyDocs, or ""OMG - I wish somebody had looked into FHS and moved /usr to eMMC"", or http://www.pathname.com/fhs/pub/fhs-2.3.html#PURPOSE2 bullet1,2 and fhs-2.3.html#PURPOSE16 sentence3
May 15 16:12:50 <MohammadAG> don't use /opt
May 15 16:13:04 <MohammadAG> use /usr/share/icons directly, the folder is optified anyways
May 15 16:13:13 <MohammadAG> DocScrutinizer, don't think it's possible
May 15 16:13:27 <DocScrutinizer> rd:
May 15 16:13:29 <DocScrutinizer> IroN900:~# ls -ld /usr/share/icons
May 15 16:13:30 <DocScrutinizer> lrwxrwxrwx 1 root root 20 2010-06-23 06:03 /usr/share/icons -> /opt/usr/share/icons
May 15 16:14:17 <rd> Thanks, DocScrutinizer
May 15 16:14:48 <DocScrutinizer> so aiui I need to exploit the hildon function that loads that friggin pvr to screen, then take a screenshot to get a proper format image, edit that with whatver tool I find, and feed the result to pvr-creator
May 15 16:15:33 <DocScrutinizer> *burrrp*
May 15 16:15:34 <MohammadAG> actually
May 15 16:15:50 <MohammadAG> there are apps to edit PVR images, but they seem to need parts of the iPhone SDK or Mac OS X
May 15 16:16:02 <DocScrutinizer> lol, sure
May 15 16:16:19 * DocScrutinizer ponders about gimp
May 15 16:16:36 <DocScrutinizer> if gimp can't edit pvr, then doomsday
May 15 16:19:06 <DocScrutinizer> MohammadAG: couldn't you add that 'tiny bit of additional functionality' to pvr-creator? :-D
May 15 16:21:27 <DocScrutinizer> quite a few users would adore that option to call `pvr-creator --edit "dim50" ~user/.cache/launch/*.pvr` on sysinit
May 15 16:23:37 <DocScrutinizer> pvr-creator --edit "exec:gimp --foo --bar %file%" ~user/.cache/launch/*.pvr
May 15 16:27:17 <MohammadAG> DocScrutinizer, I think you need the actual hardware
May 15 16:27:18 <DocScrutinizer> MohammadAG: I'd even contribute the script to avoid multiple processing of same *.pvg
May 15 16:27:30 <MohammadAG> ie a PowerVR
May 15 16:27:38 <MohammadAG> or not, it works in the maemo 5 sdk
May 15 16:28:06 <MohammadAG> DocScrutinizer, I'll rewrite it in Qt/C++, my Gtk skills suck
May 15 16:28:13 <DocScrutinizer> I never thuoght about running this on my laptop ;-D
May 15 16:28:48 <DocScrutinizer> it's a true maemo app
May 15 16:29:14 <DocScrutinizer> called pvr-splashscreen-stamper
May 15 16:29:22 <DocScrutinizer> ;-)
May 15 16:29:45 <MohammadAG> well, Qt allows easy image manipulation
May 15 16:29:51 <DocScrutinizer> aka settings-nosuck
May 15 16:30:15 <lcuk> MohammadAG, from which APIs, and are they available directly from qml?
May 15 16:30:44 <DocScrutinizer> well, provide whatever you find easy to implement: superimpose another png, blurr, dim...
May 15 16:31:49 <DocScrutinizer> I'm almost fine with all of that, as long as it makes for clearly distinguishable pvr "splashs"
May 15 16:33:18 <DocScrutinizer> my only alternative was to create a (identical) *.pvr from a "SUCKS" image, for all apps using that *.pvr shit
May 15 16:33:36 <DocScrutinizer> or completely killing it the way you suggested
May 15 16:34:59 <DocScrutinizer> s/"SUCKS"/"LOADING... please wait"/.
May 15 16:35:36 <DocScrutinizer> this screenshot fake is so utterly silly and annoying
May 15 16:35:55 <n900-dk> +1
May 15 16:36:32 <DocScrutinizer> and even worse, takes up some more time on app-start
May 15 16:37:26 <DocScrutinizer> I'd accept this additional 0.5s when I get a proper feedback. But definitely not for getting fooled by the settings-syndrome
May 15 16:39:31 <DocScrutinizer> MAAAN, this pissed me so many times. And now (http://talk.maemo.org/showthread.php?p=1001784#post1001784) I learn I'm not the only one
May 15 16:39:49 <n900-dk> Please fix DocScrutinizer ;)
May 15 16:40:06 <DocScrutinizer> I'm unfixable fubar
May 15 16:41:29 <GAN900> DocScrutinizer, they should do what Apple does.
May 15 16:41:49 <DocScrutinizer> uhuh, what is it that apple does?
May 15 16:42:19 <GAN900> What Jaffa proposed years ago.
May 15 16:42:35 <GAN900> When the application opens it shows the widget outlines.
May 15 16:42:44 <GAN900> Then fills them in when it actually opens.
May 15 16:42:44 <DocScrutinizer> what I proposed last millenium ;-P
May 15 16:43:29 <DocScrutinizer> GAN900: well, another nice way of saying "this is a splash screen"
May 15 16:43:49 <DocScrutinizer> I'd be happy with blurr or dim as well
May 15 16:44:27 <DocScrutinizer> way easier to implement, though of course is missing the coolness factor of apple
May 15 16:44:55 <GAN900> http://www.maemopeople.org/index.php/jaffa?blog=6&page=1&paged=5
May 15 16:45:03 <GAN900> Improving Application Startup Mockup
May 15 16:45:24 <DocScrutinizer> GAN900: my suggestion some 30 min ago been quite along that line
May 15 16:45:58 <DocScrutinizer> GAN900: and seems MohammadAG bitten
May 15 16:46:12 <MohammadAG> lcuk, not sure about QML, but QPixmap and QImage are very interesting
May 15 16:47:07 <MohammadAG> DocScrutinizer, are you sure it takes more time?
May 15 16:47:11 <MohammadAG> I'm pretty sure it doesn't
May 15 16:48:23 <n900-dk> hmm, wonders why I suddenly have no catalogs in HAM..
May 15 16:48:39 <DocScrutinizer> MohammadAG: loading an image from eMMC takes some CPU cycles that aren't available for app startup
May 15 16:49:38 <DocScrutinizer> probably not much, but definitely some
May 15 16:50:06 <DocScrutinizer> those 0.5s were straight from my ass though
May 15 16:50:24 <DocScrutinizer> you know my ass has good numbers most of the times
May 15 16:51:22 <MohammadAG> DocScrutinizer, there might be a way to dim the pvr, rather than edit each one
May 15 16:51:28 <MohammadAG> but this requires a modded hildon-desktop
May 15 16:51:40 <DocScrutinizer> YEAH!!! go for it!!
May 15 16:51:54 <MohammadAG> I'm not sure I'm C-skilled enough
May 15 16:51:58 <DocScrutinizer> that's a true cssu improvement
May 15 16:54:14 <DocScrutinizer> though.... It will take even more time, for doing something over and over again that could easily get applied to the *.pvr once and done
May 15 16:54:58 <n900-dk> Is there a way to prefer specific connections over another? I have an MMS APN, and the device shouldn't autoconnect to this APN
May 15 16:55:22 <DocScrutinizer> so while it sounds cute, it actually seems the inferior design from a optimization and system performance PoV
May 15 16:56:17 <lcuk> MohammadAG, C skills you have I am in no doubt
May 15 16:56:37 <DocScrutinizer> n900-dk: there is, but I lost the link. You need to edit some weird stuff in connection manager config files or sth
May 15 16:56:39 <lcuk> C in specific libraries and APIs are usually more difficult to remember and know
May 15 16:57:30 <DocScrutinizer> it's always the api and libs - c is simple
May 15 16:57:57 <MohammadAG> well, the image is a ClutterActor
May 15 16:57:57 <MohammadAG> so if there's a way to dim that...
May 15 16:58:09 <n900-dk> DocScrutinizer: ok, maybe I will find it at TMO
May 15 16:58:27 <DocScrutinizer> yeah n900-dk I think it's been there
May 15 17:01:03 <MohammadAG> think I found an easy way
May 15 17:03:52 <DocScrutinizer> I hope it's also a fast way :-/
May 15 17:04:12 <n900-dk> oh, I could just delete the MMS APN - fMMS don't need it anymore :)
May 15 17:04:17 <DocScrutinizer> unless you're talking about the pvr-creator improvement
May 15 17:04:21 <MohammadAG> DocScrutinizer, it uses existing code
May 15 17:04:25 <MohammadAG> from hildon-desktop
May 15 17:04:38 * MohammadAG compiles
May 15 17:05:16 <DocScrutinizer> still wondering if it's a patch to hildon-desktop
May 15 17:05:27 <DocScrutinizer> or to pvr-creator
May 15 17:05:47 <MohammadAG> h-d
May 15 17:05:52 <DocScrutinizer> I'd not really like the patching H-D isea
May 15 17:06:29 <MohammadAG> meh, my internal battery came loose again
May 15 17:06:41 <MohammadAG> date and time going back to 6:00 and 1/1/09 :/
May 15 17:07:35 <DocScrutinizer> it's the more intrusive patch, makes the feature non-optional to install for all those using cssu, and introduces additional potential bugs for a concept that's suboptimal
May 15 17:08:06 <MohammadAG> it can be made configurable
May 15 17:08:13 <MohammadAG> but before that, it needs to work
May 15 17:08:23 <MohammadAG> => Segmentation fault
May 15 17:09:07 <DocScrutinizer> wile a design along an augmented pvr-creator editing *.pvr is a plain user pkg you can install and purge any time, use it on all maemo systems, and it's more resource saving
May 15 17:10:06 <DocScrutinizer> all of these points are knockout arguments NOT to implement it into H-D
May 15 17:10:45 <MohammadAG> oh btw DocScrutinizer
May 15 17:10:59 <MohammadAG> the loop's still running
May 15 17:11:02 <MohammadAG> I forgot about that
May 15 17:11:08 <DocScrutinizer> hahaha
May 15 17:11:35 <DocScrutinizer> it's plain weird it takes 5s per iteration
May 15 17:11:50 <MohammadAG> well, dd, and it tries to gzip
May 15 17:13:05 <DocScrutinizer> oh hum, maybe you shoul've piped it, so unzip errs out on beginning of dd shuffling MBs, not first copying the whole crap then try to unzip
May 15 17:14:02 <DocScrutinizer> dd with a blocksize of 1, well that's for sure slow ;-)
May 15 17:14:31 <DocScrutinizer> 2 mio write()
May 15 17:15:00 <DocScrutinizer> I guess you can kill that loop
May 15 17:15:40 <DocScrutinizer> (I'd even unzip >/dev/null)
May 15 17:17:36 <DocScrutinizer> (H-D pvr patch) plus the idea of multiple times editing the same picture same way, for each start of app when you are interested in a fast execution and low overhead - that's really no smart design
May 15 17:18:35 <DocScrutinizer> (you can see how my mind sometimes is changing a 180° in a short while - not always first impression is the right one)
May 15 17:20:03 <MohammadAG> DocScrutinizer, it's a ClutterActor
May 15 17:20:17 <MohammadAG> you're not editing anything, you're just showing it dimmed
May 15 17:20:28 <MohammadAG> it's the same as dimming what's behind dialogs
May 15 17:20:28 <DocScrutinizer> that's also the reason why it takes me the longer to start on writing specs the more complex a task is - it just needs longer to evaluate all the aspects of each possible solution
May 15 17:22:46 <DocScrutinizer> MohammadAG: to get just in vicinity of a on-par with the editing-.pvr concept, this H-D patch would at very least need a per-app config in .desktop
May 15 17:23:15 <MohammadAG> DocScrutinizer, or a config where it belongs, in /usr/share/hildon-desktop/transitions.ini
May 15 17:23:26 <DocScrutinizer> or that
May 15 17:23:49 <DocScrutinizer> though I'm yet undecided on that
May 15 17:24:14 <DocScrutinizer> it's no general transition but a per-app startup detail
May 15 17:24:56 <DocScrutinizer> you don't want to have per-app settings in transitions.ini
May 15 17:25:58 <DocScrutinizer> per-app things in global files always cause lots of headache
May 15 17:26:24 <MohammadAG> why per app?
May 15 17:26:30 <MohammadAG> either it affects all apps or none
May 15 17:27:32 <DocScrutinizer> a user might want to have proper pvr for one app, dimmed pvr for second one, his own undimmed pvr-creator made custom splash on third, and plain no pvr on 4th
May 15 17:28:52 <DocScrutinizer> If I create my own custom splash with pvr creator then I don't want this to get dimmed, though I might want dimming for all other apps
May 15 17:29:38 <DocScrutinizer> you see the shortcoming adding up for the H-D patch concept
May 15 17:30:28 <DocScrutinizer> augment pvr-creator to pvr-editor, it's the way better approach
May 15 17:30:53 <DocScrutinizer> and works on plain maemo as well ;-D
May 15 17:32:15 <DocScrutinizer> never patch system bits if you can achieve same (or even better, here) effect with a plain user app
May 15 17:32:49 <DocScrutinizer> development is not about simplicity of implementation
May 15 17:33:30 <DocScrutinizer> or more precise, simplicity is a criterion after best approach has been selected
May 15 17:34:42 <DocScrutinizer> in fact a simple suboptimal approach has killed off any work on the optimal one so many times
May 15 17:34:49 <merlin1991> kulve: I'm going to reflash now, and then I'll retry ogg-support 1.0.7~rc1
May 15 17:37:27 <DocScrutinizer> for a recent example see how lis3lv02.ko module botch upstream completely blocks any proper implementation of a lis302dl driver going upstream, as "there is already a driver for that"
May 15 17:39:02 <DocScrutinizer> better partially implement a complicated but wel designed concept, than have a simple-to-implement botch nicely done so nobody ever will touch it again
May 15 17:41:30 * DocScrutinizer sighs
May 15 17:41:42 <DocScrutinizer> probably already too late :-/
May 15 17:44:02 <DocScrutinizer> ...and another bleeding edge patch to cssu that's neither needed nor tested nor discussed, and probably will give us another step away from cssu-stable
May 15 17:44:43 <DocScrutinizer> i.e. more possible bugs
May 15 17:45:02 <merlin1991> DocScrutinizer: cssu has no team that wants to push it towards -stable
May 15 17:45:24 <DocScrutinizer> evidently
May 15 17:45:29 <DocScrutinizer> too boring
May 15 17:45:34 <merlin1991> exactly
May 15 17:45:51 <DocScrutinizer> implemeting a patch to dim pvr splashscreens is way more fun, and exciting
May 15 17:46:12 <khertan> fixing that qt bugs should be the priority !
May 15 17:46:24 <khertan> the raster engine is too much bugged !
May 15 17:46:48 <lcuk> khertan, how many other qt bugs should be prioritised?
May 15 17:47:02 * DocScrutinizer declares cssu lethaly sick
May 15 17:47:08 <khertan> lcuk: i count 2
May 15 17:47:36 <merlin1991> hm how long does it usually take to boot after a flash?
May 15 17:47:42 <khertan> lcuk : the one which prevent qt hildon widget to works, and the qt18720
May 15 17:47:49 <merlin1991> I'm looking at the dots for qutie some time now and beginning to worry
May 15 17:47:57 <khertan> lcuk: at the least this are the both that i found blocking
May 15 17:48:09 <khertan> lcuk: and i didn' t found where it s come from
May 15 17:48:19 <DocScrutinizer> merlin1991: might take minutes
May 15 17:48:37 <khertan> lcuk: i ve even try to uncommit many git commit without success
May 15 17:48:53 <khertan> lcuk: at least for the shift click selection bug, we have a patch
May 15 17:54:12 <kuzew> Hello, has anyone attempted to mount maemo's rootfs on nitdroid?  I'm trying to fix a multiboot item so I can get back into maemo
May 15 17:54:33 <DocScrutinizer> meh *eg*
May 15 17:54:53 <kuzew> the multiboot rescue shell works, but I can't type in any special chars14]
May 15 17:54:56 <kuzew> [#maemo] ell works, but I can't type in any special chars
May 15 17:55:01 <kuzew> err
May 15 17:55:04 * DocScrutinizer wishes he got a € for each "multiboot issue"
May 15 17:55:19 <kuzew> the rescue shell works, sorta
May 15 17:55:30 <n900-dk> does nitroid have landscape mode?
May 15 17:55:41 <kuzew> n900-dk: yeah
May 15 17:55:52 <DocScrutinizer> fsck nitdroid
May 15 17:56:12 <kuzew> hey, I was only playing with it
May 15 17:56:29 <kuzew> I want to go back to maemo! :P
May 15 17:56:35 <DocScrutinizer> no, they were playing with you ;-P
May 15 17:56:39 <n900-dk> but no phone yet?
May 15 17:57:04 <DocScrutinizer> and they played evel prank on you, as it seems it's impossible to uninstall nitdroid
May 15 17:57:08 <kuzew> n900-dk: no phone yet, I can't even read the sim card correctly
May 15 17:57:18 <n900-dk> damit
May 15 17:57:31 <DocScrutinizer> reflash
May 15 17:57:55 <kuzew> yeah, I'm fearing I'd have to do that... :/
May 15 17:57:58 <DocScrutinizer> I haven't heard of a single recover from borked mutiboot
May 15 17:58:54 <kuzew> If Android has the jffs2 kernel module, I'd be able to mount the partition to fix my mutliboot item
May 15 17:59:23 <DocScrutinizer> nitdroid dudes are happy when you get THEIR system to run. Don't care at all about concurrent alternative OS or even uninstalling nitdroid (and multiboot)
May 15 17:59:25 <kuzew> it's funny, I can mount my home and opt (ext3 and vfat) fine
May 15 17:59:48 <kuzew> hmm, not cool
May 15 17:59:49 <DocScrutinizer> ubifs
May 15 17:59:54 <DocScrutinizer> not jffs2
May 15 17:59:58 <kuzew> oh?
May 15 18:00:21 <DocScrutinizer> afaik
May 15 18:01:07 <kuzew> it's ubifs+jffs2 I think
May 15 18:01:30 <DocScrutinizer> errr
May 15 18:01:44 <DocScrutinizer> how's that?
May 15 18:02:11 <kuzew> hmm, I'm confused now
May 15 18:02:24 <DocScrutinizer> yeah, like all nitdroid devels
May 15 18:02:27 <DocScrutinizer> :-P
May 15 18:02:29 <kuzew> haha
May 15 18:02:36 <kuzew> you're right DocScrutinizer
May 15 18:02:39 <kuzew> ubifs
May 15 18:03:33 <kuzew> the flasher doesn't have anyway to dump the rootfs image, right?
May 15 18:04:25 <DocScrutinizer> multiboot is *EVIL* and deprecated, as it breaks/conflicts with any other normal action of kernel, like installing powerkernel, updating stock kernel, adding uBoot for meego, you name it
May 15 18:05:03 <DocScrutinizer> no, you can't dump rootfs via flasher (unless there's a hidden cmdline option for that)
May 15 18:05:11 <kuzew> heh, right then
May 15 18:06:03 <DocScrutinizer> you *could* dump the raw NAND image under nitdroid
May 15 18:06:25 <DocScrutinizer> maybe even loopmount that dumpfile via ubifs later
May 15 18:06:54 <kuzew> yeah, I was just looking into that
May 15 18:07:36 <DocScrutinizer> of curse if friggin nitridiot supports mounting ubifs, that'd be even better
May 15 18:07:50 <DocScrutinizer> ooh, curse - yeah :-P
May 15 18:08:09 <kuzew> I'm working with the android debug bridge, likes to crap out a lot, I'm even surprised to see it working
May 15 18:08:26 <DocScrutinizer> nfi
May 15 18:09:09 <DocScrutinizer> you could as well tell me about win-P7 secret registry tunes ;-)
May 15 18:09:22 <kuzew> I don't run windows
May 15 18:09:31 <DocScrutinizer> I don't run google
May 15 18:10:31 <kuzew> heh, otherwise you could find out where I am and what I search for the most...
May 15 18:11:16 <DocScrutinizer> just heard they invented new craplocks there, while all the sane unix world tries to get rid of *locks as much as possible ;-D
May 15 18:12:03 <kuzew> hah
May 15 18:12:19 <DocScrutinizer> wait... wastelocks?
May 15 18:12:35 <DocScrutinizer> aaah wakelocks
May 15 18:12:53 <kuzew> oh now I get it
May 15 18:12:54 <kuzew> haha
May 15 18:13:15 <kuzew> yeah, the big kernel lock was removed a few revs ago
May 15 18:14:58 <kuzew> # cat /proc/wakelocks
May 15 18:15:02 <kuzew> yep! still there
May 15 18:15:57 <kuzew> now, time to figure out which /dev/mtd I can dump
May 15 18:16:56 <DocScrutinizer> http://lwn.net/Articles/319860/ (<-wakelocks)
May 15 18:17:04 <kuzew> yeah, I was just reading that :P
May 15 18:18:01 <merlin1991> kulve: I reflashed the n900
May 15 18:18:06 <merlin1991> installed extra decoders from -devel
May 15 18:18:14 <merlin1991> and 1.0.7 works just fine
May 15 18:20:55 <kuzew> damn, I wish the power47 kernel update didn't remove my uboot
May 15 18:23:48 <jacekowski> DocScrutinizer: you need some *locks
May 15 18:24:04 <jacekowski> DocScrutinizer: to make sure that other thread isn't messing up your memory
May 15 18:24:17 <DocScrutinizer> jacekowski: sure, like mutex. But I don't like fakelocks
May 15 18:24:38 <DocScrutinizer> and I don't like andridiot
May 15 18:24:42 <jacekowski> DocScrutinizer: well implementation is other thing
May 15 18:25:16 <DocScrutinizer> nah, adridiot's wakelocks are definitely wrong
May 15 18:27:04 <DocScrutinizer> if I want to protect my embedded device against runaway processes, I get some ulimits, quotas, and OOMd and busy-process-killer-daemon
May 15 18:28:33 <DocScrutinizer> we discussed that at Openmoko several times (particularly me and raster)
May 15 18:29:23 <DocScrutinizer> FSO has a resource "CPU" that you can allocate
May 15 18:30:03 <n900-dk> MohammadAG: How come Sociality isn't named Sociality?
May 15 18:31:11 <DocScrutinizer> if you think you need BS like dsme to auto-restart segfaulting processes, then that DSME as well needs to handle CPU resource, to avoid system suspending when such a supervised app segfaults
May 15 18:33:31 <DocScrutinizer> and maemo evidently obviously doesn't need any such things like wakelocks, CPU resource and suspend-to-ram, as SmartReflex and a proper driver design handle power just nicely
May 15 18:34:29 <kuzew> that I agree with
May 15 18:34:53 <kuzew> SmartReflex doesn't work for me though
May 15 18:35:09 <DocScrutinizer> so all it'd need in e.g. maemo to achieve a fool-app proof thing was a kind of automatic powertop killing runaway apps
May 15 18:35:12 <kuzew> then again, I just hit a checkbox, so I don't know what they are doing is proper
May 15 18:35:43 <DocScrutinizer> kuzew: what you think of is just a little part of the whole SR thing
May 15 18:35:57 <kuzew> SR?
May 15 18:36:13 <DocScrutinizer> smartreflex
May 15 18:36:20 <kuzew> ah
May 15 18:37:45 <DocScrutinizer> SR is a bunch of measures to minimize system power consumption. Most of them are on a chip level and can't get switched on/off
May 15 18:38:44 <kuzew> ah, I didn't read the ARM docs
May 15 18:38:59 <DocScrutinizer> others are for example suspending of particular subsystems when not in use, like the ULPI bus from CPU to MUSB core USB controller
May 15 18:40:12 <DocScrutinizer> the /sys/*/smartreflex is only about the voltage regulators *automatic* adjustment via a dedicated I2C connection to CPU
May 15 18:40:19 <kuzew> ah
May 15 18:41:13 <DocScrutinizer> you still can adjust the voltages 'manually', i.e. via software, which maemo (kernel) is doing. AFAIK
May 15 18:42:03 <DocScrutinizer> other aspects of whole SR are not at all affected
May 15 18:42:16 <kuzew> I was going to ask, but yeah
May 15 18:43:05 <kuzew> Hmm, ULPI is neat
May 15 18:44:26 <DocScrutinizer> yeah, and it's from MUSB to PHY, not to CPU. I was wrong there, that's in fact yet another bus that has its own suspend control
May 15 18:46:10 <DocScrutinizer> and this is one of the features of SR: fine grained suspend/power control for subsystems
May 15 18:46:49 <DocScrutinizer> so suspending whole system like adridiot does is a rather awkward thing
May 15 18:47:03 <kuzew> yeah, I see that a _LOT_ in dmesg
May 15 18:47:42 <kuzew> does the OMAP/DSP also have a sleep state you can put it in?
May 15 18:47:50 <DocScrutinizer> sure
May 15 18:48:08 <kuzew> rather, is it being used?
May 15 18:48:15 <DocScrutinizer> I'd bet it is
May 15 18:48:31 <SpeedEvil> _everything_ has a sleep state.
May 15 18:48:46 <DocScrutinizer> though that's quite probably burried inside the closed drivers for DSP
May 15 18:49:07 <kuzew> yuck, they are closed as well?
May 15 18:49:21 <DocScrutinizer> yes, DSP lib is closed
May 15 18:50:08 <DocScrutinizer> not the kernel/sys api though, I'd guess
May 15 18:50:34 <kuzew> yeah, I'd think that would be open, the drivers I would understand being closed
May 15 18:50:43 <GNUtoo|laptop> I saw on oe some dsp bios stuff, but I was told at the same time that the DSP side was proprietary
May 15 18:51:09 <GNUtoo|laptop> basically there are some CPU side  FLOSS stuff
May 15 18:51:13 <GNUtoo|laptop> for instance gst-dsp
May 15 18:51:28 <GNUtoo|laptop> but personally I prefer the CPU
May 15 18:51:35 <GNUtoo|laptop> the CPU reads more formats
May 15 18:51:49 <GNUtoo|laptop> maybe it consumes more battery tough
May 15 18:51:53 <SpeedEvil> And uses more power
May 15 18:52:08 <SpeedEvil> It's not - so - important if you have the screen at maximum.
May 15 18:52:12 <kuzew> DSP can be more efficient at times, depending what you are doing
May 15 18:52:18 <SpeedEvil> But at minimum brightness it makes quite a difference.
May 15 18:52:32 <DocScrutinizer> DSP bridge: http://wiki.maemo.org/Documentation/Maemo_5_Developer_Guide/Architecture/Multimedia_Domain
May 15 18:54:00 <DocScrutinizer> DSP is not per se better on power conservation. It's simply faster on some special tasks
May 15 18:54:19 <kuzew> yes
May 15 18:54:26 <DocScrutinizer> DSP easily might eat more power when idling than the CPU
May 15 18:54:52 <kuzew> you'd hope it be in a sleep state then
May 15 18:55:07 <DocScrutinizer> that's the topic we just discussed
May 15 18:57:15 <DocScrutinizer> you should't expect e.g. mp3-playback via DSP to be always better on power than doing same thing via CPU
May 15 18:57:33 <DocScrutinizer> should *not*
May 15 18:59:24 <SpeedEvil> IMO, you probably should, if you are just asking about mp3 decoding.
May 15 18:59:29 <SpeedEvil> And not streaming to the DAC
May 15 18:59:29 <kuzew> welll
May 15 18:59:31 <SpeedEvil> But...
May 15 19:00:17 <SpeedEvil> That reminds me - I meant to look at the functional block diagram, and find out if the DMA unit can stay awake without the CPU core, DMAing a decoded raw audio to the DAC
May 15 19:00:48 <kuzew> you mean DMAed into the DSP then to the DAC
May 15 19:01:13 <DocScrutinizer> well, the DSP can decode your whole song in 0.4s while CPU takes 3s for that. But then the DSP keeps eating power for the rest of the song, pushing bytes out the DAC, while CPU could sleep most of the time, or even offload to a DMA and sleep all of the time
May 15 19:01:56 <SpeedEvil> I mean - if the DSP can be used to decode - say - 3s of audio, and go back to sleep, and then the CPU setup DMA to push this to the speakers, go back to sleep, and wakeup when there is .5s later
May 15 19:02:11 <SpeedEvil> .5s left, and repeat
May 15 19:03:04 <DocScrutinizer> yes, that'd maybe the optimum, depending on how much better DSP is, compared to the overhead in CPU setting it up and moving data around
May 15 19:03:28 <SpeedEvil> where 3 may be 30
May 15 19:04:32 <DocScrutinizer> it's a non trivial evaluation
May 15 19:04:59 <DocScrutinizer> you'd finally have to implement both concepts and compare
May 15 19:05:35 <SpeedEvil> Indeed.
May 15 19:05:43 <SpeedEvil> Well - not really.
May 15 19:06:02 <SpeedEvil> The 'if it can DMA it out to the DAC' is a clear winner if it's the CPU or the DSP doing it
May 15 19:06:09 <SpeedEvil> the rest is a matter of question.
May 15 19:06:20 <DocScrutinizer> yep
May 15 19:06:33 <kulve> merlin1991: ok, good to know
May 15 19:19:56 <haj> annoying.. my N900 seems to loose contact with the SIM all the time... I tried cleaning the connectors but it doesn't seem to help...
May 15 19:27:55 <SpeedEvil> haj: Same here - I need to return it for warranty. It is a known failure mode apparantly.
May 15 19:29:19 <haj> uhm... mine is out of warranty... wonder if it's an expensive fix..
May 15 19:30:58 <SpeedEvil> Ah.
May 15 19:31:10 <SpeedEvil> I'm unsure that it's a simple fix.
May 15 19:31:24 <SpeedEvil> I would quite like to know what's wrong.
May 15 19:32:07 <haj> SpeedEvil: when did yours start doing it?
May 15 19:32:14 <haj> did you try another sim?
May 15 19:32:42 <SpeedEvil> haj: umm... hangon
May 15 19:35:34 <SpeedEvil> At least last august
May 15 19:38:38 <haj> annoying...
May 15 19:39:35 <davide> hello
May 15 19:40:27 <haj> SpeedEvil: does your phone also keep asking you for the PIN-code? Mine did that the other day, but now it seems it just looses contact for a few sec. and then it works again..
May 15 19:42:42 <SpeedEvil> No
May 15 19:42:48 <SpeedEvil> I've never set a pin.
May 15 19:55:50 <DocScrutinizer> haj: check syslog/dmesg
May 15 20:45:33 <wazd> woo, congats finns :)
May 15 20:47:09 <wazd> oooooh
May 15 20:49:00 <achipa> they're really on fire
May 15 20:50:28 <wazd> humiliation :)
May 15 20:52:18 <sp3000> well now I feel bad for fasth
May 15 20:52:28 <sp3000> that like totally messes up his avg
May 15 20:53:33 <moo__> good evening
May 15 20:53:37 <moo__> riddle me this
May 15 20:54:19 <moo__> why does power kernel cause 1000 wakeups/second
May 15 20:54:29 <moo__> where as stock only causes 300/s?
May 15 20:55:02 <moo__> this was tested on a freshly flashed device
May 15 20:57:57 <moo__> son of a bitch
May 15 20:58:01 <SpeedEvil> Pastebin.
May 15 20:58:03 <moo__> its blood!
May 15 20:58:18 <moo__> Who wrote the pastebinnit in devel?
May 15 20:59:17 <moo__> ok, scratch everything I just said
May 15 20:59:28 <piggz> does qputenv("N900_PORTRAIT_SENSORS", "1");  work for a qml app?  it doesnt seem to be
May 15 20:59:33 <moo__> just ran another powertop and got the longest sleep time I have ever seen
May 15 20:59:50 <moo__> 2957ms!
May 15 20:59:56 <moo__> not even offline mode!
May 15 21:01:44 <achipa> piggz: which mobility are you using ?
May 15 21:01:56 <piggz> achipa: 1.1
May 15 21:02:22 <achipa> i don't think anything prior to 1.2 final has that tweak...
May 15 21:02:56 <piggz> achipa: ok, i was going to update to 1.2 anyway
May 15 21:03:45 <SpeedEvil> moo__: yes - that should be about what you should get.
May 15 21:04:03 <SpeedEvil> moo__: In general - look at time in C4, rather than wakeups - at least in the first instance.
May 15 21:05:44 <DocScrutinizer> qt-mobility, sensors, portrait - MUHAHAHAHAHAAAA
May 15 21:06:03 <moo__> SpeedEvil: yeah
May 15 21:06:12 <moo__> SpeedEvil: this is the longest I have evr seen
May 15 21:06:32 <moo__> for some reason prior to flash I was getting >1000
May 15 21:07:00 <moo__> now im gonna have to CHECK every peice that I install
May 15 21:08:00 <moo__> to find the f***** that was causing all thos wkups
May 15 21:09:46 <DocScrutinizer> moo__: unplug charger, lsmod|grep bq2
May 15 21:11:32 <moo__> you are suggesting the bq module causes shitloads of WU?
May 15 21:11:48 <moo__> because I have rmmod it and tested
May 15 21:11:51 <moo__> loads of times
May 15 21:17:29 <moo__> sonovva bitch
May 15 21:17:34 <moo__> profilesx
May 15 21:17:38 <moo__> pile of shite
May 15 21:17:58 <moo__> causes an extra ~200 HW wakeups/second
May 15 21:18:22 <moo__> kernel wakeups
May 15 21:18:28 <moo__> interrupt 56
May 15 21:18:34 <moo__> gp timers i believe
May 15 21:19:35 <moo__> well...extra profiles are over-rated anyway
May 15 21:19:43 <SpeedEvil> Report it as a bug.
May 15 21:20:06 <moo__> maemo seems totally fucked
May 15 21:20:20 <moo__> there is complete fragmentation across all the repos
May 15 21:25:54 <moo__> yup, profilesx==gay
May 15 21:26:36 <moo__> did anyone figure out how to convert a backupmenu style opt+rootfs archive into a flash bin yet?
May 15 21:30:54 <kuzew> that would of been nice
May 15 21:31:00 <moo__> DocScrutinizer: has quit?
May 15 21:31:04 <moo__> wut?
May 15 21:32:14 <moo__> SpeedEvil: so now I have apt-get purged profilesx, the WU wont go away
May 15 21:32:19 <moo__> can you explain that?
May 15 21:32:45 <SpeedEvil> I ave no clue.
May 15 21:33:33 <moo__> SpeedEvil: is it possible to get rid of dsme?
May 15 21:33:49 <moo__> i.e. turn of wd timers first
May 15 21:34:09 <SpeedEvil> In principle yes.
May 15 21:34:14 <moo__> in practice?
May 15 21:34:17 <SpeedEvil> In practice, I question how much it'll help.
May 15 21:34:21 <moo__> ok
May 15 21:34:52 <SpeedEvil> In that it gets ~10 days of battery life offline.
May 15 21:35:15 <moo__> you said before
May 15 21:35:53 <SpeedEvil> I gorget stuff.
May 15 21:36:10 <moo__> SpeedEvil: have you installed the enhanced busybox package?
May 15 21:36:26 <SpeedEvil> No.
May 15 21:36:30 <SpeedEvil> I see little or no point.
May 15 21:36:36 <SpeedEvil> I have installed gnu tools though.
May 15 21:36:50 <moo__> what is included in the gnu tools package?
May 15 21:37:03 <moo__> I use the console a lot, is it wirth my investigation?
May 15 21:37:28 <SpeedEvil> gnu find, ...
May 15 21:37:32 <moo__> k
May 15 21:52:12 <moo__> SpeedEvil: do you use the BFS kernel?
May 15 21:53:04 <moo__> for that matter does anyone here use BFS kernel?
May 15 21:53:11 <moo__> does hawaii hang out here?
May 15 21:53:39 <piggz> achipa: i installed libqtm-12*, changed my qml to 'import QtMobility.sensors 1.2' and added the improt path in my main()...but it complains on startup that it cant find sensor 1.2
May 15 21:53:42 <piggz> ?
May 15 21:58:52 <SpeedEvil> moo__: no,
May 15 22:17:14 <DocScrutinizer> moo__: I sugest you stop posting for 300s after each line ending with a "?" or otherwise requiring any interaction. Your style is kinda spammy
May 15 22:18:14 <DocScrutinizer> pushing out more questions in fast sequence will NOT result in more or better answers - the contrary is the case
May 16 00:13:57 <Macer> are there any maemo4 repos?
May 16 00:17:03 <Macer> oh... nm
May 16 01:35:04 <djszapiN900> has anybody tried to do an internet connection sharing from fremantle to the laptop without customized kernel ?
May 16 01:53:28 <SpeedEvil> yes
May 16 01:53:59 <SpeedEvil> bring USB-network connection up manually, and then use socks-proxy through ssh
May 16 01:54:28 <SpeedEvil> Or just select 'pc-suite' mode when plugged in, and treat it as a normal USB modem
May 16 02:52:07 <JZA> hi how can I tether the n900 on Mandriva linux
May 16 02:53:30 <JZA> I do get the usb N900 as well as an alternative port but none of the get's it
May 16 02:55:08 <SpeedEvil> Plug it in, select 'pc suite' mode.
May 16 02:55:16 <SpeedEvil> then set it up as any other 3g modem
May 16 02:55:40 <JZA> SpeedEvil: great, I never set up a 3g modem
May 16 02:55:41 <JZA> :)
May 16 02:55:47 <SpeedEvil> No idea.
May 16 02:55:49 <JZA> like I said I found 2 interfaces
May 16 02:55:59 <JZA> usb0 and another one
May 16 02:56:02 <SpeedEvil> I did it last time with pppd, and /etc/
May 16 02:56:38 <SpeedEvil> Google to find out how to setup a 3g modem. It should appear as /dev/ttyACM0 I think.
May 16 02:56:49 <SpeedEvil> Though the name may vary
May 16 02:58:07 <JZA> SpeedEvil: will do, thanks
**** ENDING LOGGING AT Mon May 16 02:59:57 2011