**** BEGIN LOGGING AT Sun Feb 21 02:59:57 2010 Feb 21 02:59:58 and 'avo :) Feb 21 03:00:02 mate :) Feb 21 03:00:06 guess it's about themes Feb 21 03:00:28 my freerunners have been collecting dust for.. a long time Feb 21 03:00:32 i dont think they even turn on anymore Feb 21 03:00:35 batteries dead Feb 21 03:00:47 nasty boy Feb 21 03:00:53 so.., i wouldnt know :) Feb 21 03:01:29 need a batery? Feb 21 03:01:35 * raster slides is elementary list around silky-smoothly Feb 21 03:01:45 i have a few Feb 21 03:01:49 i just couldnt be bothered Feb 21 03:01:50 :) Feb 21 03:01:57 yeah Feb 21 03:02:05 same here, basically Feb 21 03:02:24 "lets go and use the c64 now... thats a good idea!" Feb 21 03:02:34 when u have a perfectly good core2-due desktop on your desk Feb 21 03:02:35 :) Feb 21 03:02:36 raster, ytube vid Feb 21 03:02:55 huh? Feb 21 03:02:58 JesusMcCloud: havent seen any youtube vide of efl stuff of late Feb 21 03:03:14 raster, was couple fo montsh ago Feb 21 03:03:18 so not sure i can be impressed Feb 21 03:03:21 oooh months ago Feb 21 03:03:50 thats like... pre-historic Feb 21 03:04:08 yeah, 15bp vs 32bp can do a fsckng lot on glamo even Feb 21 03:04:09 u told me to kick out images and use layers of rects instead... now i had the time to do it Feb 21 03:04:37 and? Feb 21 03:04:49 DocScrutinizer: 16bpp is so last week. Feb 21 03:05:04 WEEK? Feb 21 03:05:19 make that century Feb 21 03:05:20 its an expression Feb 21 03:05:21 :) Feb 21 03:05:22 raster, SOFTWARE seems now nearly as fast as SORFTWARE_16 Feb 21 03:06:12 scrolling in genlist... Feb 21 03:06:28 it shouldnt be too bad Feb 21 03:06:42 but still - glamo and freerunenr are a pretty awful platform Feb 21 03:06:49 yup Feb 21 03:07:03 * DocScrutinizer scanning the other zillion irc chan he got on screen lately Feb 21 03:07:04 * JesusMcCloud is waiting for shr on htc dream Feb 21 03:07:15 i can sit here and scroll - and my numbers say i'm doing 100fps... or therabouts.. with no theme fiddling Feb 21 03:07:19 480x800 Feb 21 03:07:20 32bpp Feb 21 03:07:24 here Feb 21 03:07:40 and thats WITH a compositor Feb 21 03:07:41 pfft, N900 can do that Feb 21 03:07:46 (also evas based) Feb 21 03:07:49 my point was simply that the default theme is not suited for the freerunner Feb 21 03:07:56 DocScrutinizer: i doubt it can do it while compositing Feb 21 03:07:57 :) Feb 21 03:08:05 it hits 200fps+ without Feb 21 03:08:24 with a full gles engine i'd guess it *can* Feb 21 03:08:32 i doubt it Feb 21 03:08:39 as this is an sgx540 - 200mhz Feb 21 03:08:45 your n900 is an sgx530 at 110mhz Feb 21 03:08:45 that's your playground Feb 21 03:08:55 u wont get the same speed Feb 21 03:08:57 not even close Feb 21 03:09:05 it'll still be better than freerunner Feb 21 03:09:12 tho my n900 is dead Feb 21 03:09:16 LOOOOOL Feb 21 03:09:16 i give up on that piece of shit Feb 21 03:09:43 gimme the 13' wrench Feb 21 03:09:51 * DocScrutinizer opens beerbottle Feb 21 03:09:54 it wont goot Feb 21 03:10:09 if i'm lucky i can plug it in and out 50 times and finally get a usb device to pop up on my pc Feb 21 03:10:11 and i can try and flash Feb 21 03:10:16 i managed to get it to do that once Feb 21 03:10:25 and.. flash failed 70% trhu Feb 21 03:10:30 its still bricked Feb 21 03:10:32 i give up Feb 21 03:10:32 :) Feb 21 03:11:01 raster: you got evil karma on electronics :-P Feb 21 03:11:10 i do Feb 21 03:11:10 :) Feb 21 03:11:14 tho my samsung stuff works Feb 21 03:11:22 seems my karma is good with sammy Feb 21 03:11:24 I'm really happy with my N900 Feb 21 03:12:19 (though for sure not with some of the braindamaged decissions of Nokia managers wrt e.g. partitioning) Feb 21 03:12:54 ~optification Feb 21 03:12:55 rumour has it, optification is a botch to reclaim space in fs root, done due to the fact the partitioning is FUBAR, or http://wiki.maemo.org/Free_up_rootfs_space Feb 21 03:13:02 as apt is trans-chan Feb 21 03:13:20 trust me Feb 21 03:13:22 not just nokia Feb 21 03:13:38 its a combo of insisting on doing usb mass storage Feb 21 03:13:56 and on some insane love of nand instead of using the mmc they have soldered on Feb 21 03:14:03 and idiocy Feb 21 03:14:05 TAsn, your latest convert script might be faulty Feb 21 03:14:17 ack Feb 21 03:14:40 1. get rid of mass storage export Feb 21 03:14:47 export usbnet only Feb 21 03:14:52 yep!!!! Feb 21 03:14:54 and use nfs/samba Feb 21 03:15:03 samba sufficient Feb 21 03:15:04 that will fix up a whole lot of caveats right there Feb 21 03:15:16 DocScrutinizer: yeah. tho i'd do nfs just tomake linux nerds happy Feb 21 03:15:16 :) Feb 21 03:15:22 u can do that one trivially anyway Feb 21 03:15:29 knfsd etc. Feb 21 03:15:40 noticed camera app borks when you scratch the VFAT? LOOL muhahahaha Feb 21 03:16:13 nfs? for nerds? idiots, should use fuse Feb 21 03:16:22 and 2. stop trying to cram an os into ths stupidly small nand u have.. give up on nand. stick a few mb of it in for a bootloader (ie bios equivalent of pc's) and lock that down so it cant be modified (easily) Feb 21 03:16:25 and then just boot off mmc Feb 21 03:16:32 if its solders on mmc or a card then becomes largely moot Feb 21 03:16:56 DocScrutinizer: fuse is not a network filesystem Feb 21 03:17:01 nfs is Feb 21 03:17:28 aerr huh? Feb 21 03:17:38 hat's a network fs then? Feb 21 03:17:42 raster: ditch nand completely and boot off the rom with mmc boos capabilities any decent soc should have these days :) Feb 21 03:17:43 What's Feb 21 03:17:56 hey wpwrak :-DD Feb 21 03:18:04 wpwrak: thats if u have reliable mmc boot ability in rom Feb 21 03:18:14 DocScrutinizer: nfs is a network fs Feb 21 03:18:17 raster: samsung does. now :) Feb 21 03:18:21 even random guests in my pup know you (they don't know andy :-P) Feb 21 03:18:26 wpwrak: depends on the soc :) Feb 21 03:18:51 raster: i think they pretty much all do these days. well, unless you get one of the ancient mistakes, such as the 2442 Feb 21 03:19:12 wpwrak: also invariably those rom boots come with some nasty caveats for the media... you are still going to be better to have a minimal BL on nand that actually understands real partitioning and fs's and will load a vmlinuz/zimage etc. imho Feb 21 03:19:35 loool acient mistake Feb 21 03:19:36 DocScrutinizer: *grin* they probably remember countless hours of fighting with lilo :) Feb 21 03:20:06 raster: yeah, qi in rom, that's what we want :) Feb 21 03:20:18 nah this particular random guest had a FR Feb 21 03:20:27 wpwrak: yup. i'm with u - just u are unlikely to get it on rom.. so a litle blob of nand will do the trick Feb 21 03:20:37 but yeah.. an intelligent bl is what is needed Feb 21 03:20:55 DocScrutinizer: must be quite a special pub then :) Feb 21 03:20:58 the omap bl is highly restrictive of the mmc setup and it a pita - it does work. if you appease it just right Feb 21 03:21:19 since u invariably get some nand on any pop/mcp setup Feb 21 03:21:21 may as well use it Feb 21 03:21:22 wpwrak: OM EE Germany official club pub ;-P Feb 21 03:21:44 raster: does it read a file system a la dos boot loader or does it rely on "magic partitions" like samsung's irom ? Feb 21 03:21:55 i'd use all that excess nand to hold other thigns that are possibly not able to be put on mmc as well Feb 21 03:22:01 DocScrutinizer: okay, that would explain it :) Feb 21 03:22:06 mostly for things that may not be redistributable license-wise Feb 21 03:22:13 eg gle-s libs for sgxz Feb 21 03:22:21 raster: (nand tax) that's unfortunately true. wish i could trade the nand for more ram Feb 21 03:22:38 wpwrak: nah, cause and effect are 180° inverse than what you think Feb 21 03:22:38 to allow peolpe to make their own os's without needing to re-distribute tem... put them on nand - your image can mount nand and symlink to these things Feb 21 03:22:58 raster: not redistributable = don't touch it :) Feb 21 03:23:26 TAsn, either convert script or whatever stuff you additionally changed is faulty Feb 21 03:23:29 wpwrak: the omap one requires you partiction your mmc with a specific # of cylinders+heads, you must use fat16 from memory and then the next state boot file must be the first file put int he fat table Feb 21 03:23:39 (NAND) use nand for very basic / and all the frequently used files that have really high impact on speed of system Feb 21 03:23:46 raster: btw, how do you like arm now doing their own graphics subsystem ? farewell powervr, may thy stay in hell be long and unpleasant ? Feb 21 03:23:48 i had to low-level fdisk many an sd card to make it happy Feb 21 03:23:54 use eMMC for all the rest of the crap Feb 21 03:24:14 and if u got unlucky with fat dir orde4.r. u were screwed Feb 21 03:24:23 thus my desire for a non-brain-dead bootloader Feb 21 03:24:25 as u said Feb 21 03:24:29 qi is decent enough Feb 21 03:24:34 raster: (omap boot) pheew. stiff requirements. not quite sure which one i should dislike more, omap's or samsung's :) Feb 21 03:24:37 something very much qi-like or just use qi Feb 21 03:25:30 wpwrak: hehehe. well better than nothing - it's a "save my brick" rom... thats good. but its not good enough Feb 21 03:25:44 apropos BL: Nokia plans to have a crypto-signature checking BL for N9xx++ Feb 21 03:25:54 ~nuke nokia Feb 21 03:25:54 thus.. u paid the nand tax already. u can stick quite a nice qi - even with a splashscreen on nand.. and then some./ Feb 21 03:25:55 * apt prepares 100 missle silos, and targets them at nokia ... B☢☢M! Feb 21 03:26:06 DocScrutinizer: cute :) Feb 21 03:26:18 YEAH REEEAALLLLYY Feb 21 03:26:19 as i said - nand can also be where u can stick other things u may not be able to put into os images - like non-redistributable things Feb 21 03:26:36 PLUS TPM aka fritzchip Feb 21 03:26:38 nothnig i can do about the legal nature there... just do my best to work with the situation as it stands Feb 21 03:26:39 :) Feb 21 03:26:55 and.. arm yes.. doing a gpu - which is just as closed up as the sgx's of this world. Feb 21 03:27:19 "we need DRM, uh really, we do..." bastards Feb 21 03:27:26 raster: i would just stay away from non-redistributable things. they just make your life painful to please some lawyers in the manufacturer's company Feb 21 03:27:58 wpwrak: talk to imgtec... they are the problem Feb 21 03:27:59 raster: (arm gpu closed) oh, really ? that's bad. Feb 21 03:28:02 nothing to be done there. Feb 21 03:28:05 and yeah Feb 21 03:28:14 the mali (arm's gpu) is closed up tight as a nun. Feb 21 03:28:22 i would even argue imgtec is more open Feb 21 03:28:24 idiots Feb 21 03:28:30 and thats saying something Feb 21 03:28:47 (at least we have the src - as the oem making products with their gpu inside) Feb 21 03:28:48 well, we still have samsung's 2d accelerator ... Feb 21 03:28:54 nup Feb 21 03:28:59 seems I seen that for glamo :-( Feb 21 03:29:00 not on c1xx... Feb 21 03:29:13 ah, didn't look at that one Feb 21 03:29:13 on s3c6xxx yes Feb 21 03:29:19 not the cortex stuff Feb 21 03:29:32 (s5pc1xx....) Feb 21 03:29:45 u do have samsung's media codec stuff Feb 21 03:29:49 ie for video decode Feb 21 03:29:50 etc. Feb 21 03:30:03 but gpu - all of it.. is imgtec Feb 21 03:30:16 well u have samsungs basic fb stuff Feb 21 03:30:19 like overlays/underlays Feb 21 03:30:22 yuv planes etc. Feb 21 03:30:24 * DocScrutinizer feels that strange nausea again. must be the month of year Feb 21 03:30:26 but no 2d accel ? Feb 21 03:30:26 but that doesnt count Feb 21 03:30:30 no Feb 21 03:30:34 sigh Feb 21 03:30:42 so either u blast it all with the cpu Feb 21 03:30:55 or... u use opengl-es(2) and the imgtec closed driver libs Feb 21 03:30:59 at lest the kernel code is open Feb 21 03:31:00 :) Feb 21 03:31:04 thats something... Feb 21 03:31:16 most of the work has been on the kernel side actually Feb 21 03:31:27 getting proper vsynced pageflipping to go right Feb 21 03:31:28 kick me DRM, crypto-BL, closed GFXaccel, and I go vomitting Feb 21 03:31:30 no binary-only kernel modules ? or intrusive "private" APIs ? Feb 21 03:31:38 and compositing via dri2/drm objects Feb 21 03:31:47 the kernel part is open Feb 21 03:31:48 gpl Feb 21 03:31:56 DocScrutinizer: yeah, all this sucks. let's hope wolfgang makes good progress with his gpu project :) Feb 21 03:32:06 its the userspace libs that are closed Feb 21 03:32:25 wpwrak: duh, he has a GPU project now as well?? Feb 21 03:32:44 raster: is the kernel part deemed to be mainlineable ? i.e., how long are the flames greg throws ? :) Feb 21 03:32:57 thought his GSM chip ideas are just more of a bite than he can swallow Feb 21 03:33:04 DocScrutinizer: that milkymist thing Feb 21 03:33:21 DocScrutinizer: he doesn't do gsm chips (yet) Feb 21 03:33:48 don't get me wrong, I hope all the best for him Feb 21 03:33:51 DocScrutinizer: other might, though. i've already heard from us miliary and yet another university with an smt line in brazil Feb 21 03:33:53 wpwrak: no idea Feb 21 03:34:05 but chip design really is another playground Feb 21 03:34:18 wpwrak: i doubt it will ever get mainlined as its kind of part and parcel with the closed userspace bits Feb 21 03:34:58 raster: it has that feeling. well, if you can tap communication in the kernel, you can also reverse engineer it :) Feb 21 03:35:39 and honestly I think it's much easier to design a new SoC than to create a semi-functional GSM chipset Feb 21 03:36:30 GPU is easy compared to GSM Feb 21 03:37:16 wpwrak: good luck... reverse engineering a wifi chip is one thing Feb 21 03:37:23 gpu... 10000x the complexity Feb 21 03:37:45 rememebr the gpu is a cpu too Feb 21 03:37:49 DocScrutinizer: i think it's just a matter of starting somewhere. the technology is old, so you have some flexibility now. and there should be enough people who understand the principles, particularly when it comes to analog and rf. protocol stacks are an entirely different matter. there's only a handful of people who even know the tools you need to have a chance of getting them right. Feb 21 03:37:51 ie a whole set of them (shaders) Feb 21 03:37:59 with their own instruction sets, registers and more Feb 21 03:38:17 u will nedd to reverse enigneer that - make a compiler for it that competes with the vendor closed one Feb 21 03:38:19 raster: well, you can just do the 2d ops. Feb 21 03:38:23 not to mention all the rest of the gpu stuff Feb 21 03:38:37 no one uses the 2d ops anymore Feb 21 03:38:39 :) Feb 21 03:39:00 its not interesting - as u just wont find anything using them Feb 21 03:39:03 wpwrak: I humply dare to disagree completely Feb 21 03:39:05 raster: let's bring back the good old times :) Feb 21 03:39:07 they either do it with the cpu Feb 21 03:39:10 or they are ding opengl Feb 21 03:39:24 so u are gong to be reverse engineering the 3d (opengl) gpu Feb 21 03:39:30 and as fo gl-es2.0 - nothng happens without a shader Feb 21 03:39:47 so u hneed to manage to reverse engineer the shader cpu's fully first Feb 21 03:39:59 also i cant actually tell u where the scheduler logic live Feb 21 03:40:10 i know the sgx has a scheduler there Feb 21 03:40:18 that handles thread and task flipping for the gpu units Feb 21 03:40:36 gpu units suport hw threads (zero-cost thread switching ibelieve) Feb 21 03:40:39 raster: why would i need the shaders if i want to do 2d ops to make my gui bearably fast ? Feb 21 03:40:50 and if that scheduler needs to be also created first before u can actually do anything Feb 21 03:40:54 each monkey can translate GSM specs to a stack. You never know what your analog ciruit will do, even if you built a thousands up til now Feb 21 03:40:59 because thats all the gpu is doing Feb 21 03:41:02 its not doing 2d ops Feb 21 03:41:05 its doing 3d Feb 21 03:41:09 there are no 2d ops i know of in it Feb 21 03:42:02 DocScrutinizer: yeah, you mentioned that a few times :) i still don't believe it can be all that hard. maybe you need to separate digital rf and anlog rf, use some discrete circuit for analog, and work on that later. Feb 21 03:42:03 esp if it's a GHz design Feb 21 03:42:37 wpwrak: that's a really decent plan Feb 21 03:45:34 DocScrutinizer: (protocols) yeah, i've seen my share of telephony docs translated into stacks. they tell you what your stuff is supposed to do, but they don't teach you the know-how to actually make it happen. that's why many implementationg don't. e.g., for linux-atm, i had an entire "cisco bug compatibility mode". fore systems had their share of problem too, although much much better than cisco. ascom really impressed me. their Feb 21 03:45:34 stuff worked flawlessly, even better than mine. that's probably why they abandoned their project ... Feb 21 03:45:51 wpwrak: told ya digital is a science, while analog is voodoo. You'll get to know once you tried to build a stable transmitter for such a simple thing like FM stereo Feb 21 03:46:41 raster: hmm, so the shader is also pushing rectangles (characters) around and such ? Feb 21 03:47:36 wpwrak: aaaah no. Feb 21 03:47:42 shaders come in 2 types Feb 21 03:47:47 vertex and fragment Feb 21 03:48:00 vertex shaders are passed verticies by the hw somehow Feb 21 03:48:08 (this is where scheduler etc. come in) Feb 21 03:48:17 how that is implemented depends on the gpu Feb 21 03:48:21 may be in silicon, firmware - who knows Feb 21 03:48:22 anyway Feb 21 03:48:25 raster: (vertex) sounds very 3d Feb 21 03:48:33 vertex shaders get to be passed data from input fromt he app Feb 21 03:48:39 that generally comes in arrays of data Feb 21 03:48:41 wpwrak: thanks Feb 21 03:48:43 no worries DocScrutinizer Feb 21 03:48:44 :-) Feb 21 03:48:46 vectors etc. Feb 21 03:48:54 DocScrutinizer: sure I agree SoC is much easier than GSM Feb 21 03:48:55 (as integers, floats and so on) Feb 21 03:48:56 so I start with the SoC Feb 21 03:49:01 LatticeMico32 core Feb 21 03:49:04 vectores can be of myltiple dimensions Feb 21 03:49:06 2,3, 4 etc. Feb 21 03:49:15 colors are also vectors Feb 21 03:49:19 vertexes are Feb 21 03:49:21 and so on Feb 21 03:49:27 DocScrutinizer: raster : nice video here: http://en.qi-hardware.com/wiki/Milkymist_One Feb 21 03:49:36 raster: the next generation will have fractional dimensions ? ;-) Feb 21 03:49:45 wolfyg: :-) wpwrak had a decent idea of completely separating analog and digital Feb 21 03:49:46 yoru vertex shader inveriably handles the matrix transform for projection (making things look 3d based on z value) Feb 21 03:50:05 and these values are passed onto fragment shaders and interpolated as thew surface is drawn Feb 21 03:50:13 fragmment shaders are passed the interpolated values Feb 21 03:50:33 that is interpolated texture coords (based on x, y and z position) interpolated color and so on... Feb 21 03:50:50 DocScrutinizer: Yes I am listening. No worries. I thought about what you said about ghz and operating at physical limits, and starting with FM radio, etc. etc. Feb 21 03:51:01 we get there... :-) Feb 21 03:51:25 fragment shaders then ge to swizzle this any way they like - fetch a texel (the textel fetch instructiosn should handle all the mipmap/asinotropic filtering etc. and linear interpolation from textures and conversion from texture format to an rgba vector) Feb 21 03:51:38 and then u can do what u like with this color and whatever other input u have Feb 21 03:51:43 wolfyg: really I don't want to scare you or demonstrate my age. It's just RF is *scary* Feb 21 03:51:45 I just wish I had some more money I could throw at guys like wpwrak and DocScrutinizer Feb 21 03:51:50 and blast it back out as a resulting fragment (pixel) Feb 21 03:52:00 wolfyg: sell more nanonotes :) Feb 21 03:52:01 thats also rgba - and then a compositor may write/blend that to the buffer Feb 21 03:52:04 yes, very appreciated Feb 21 03:52:04 not about age Feb 21 03:52:04 it's good feedback! Feb 21 03:52:14 of coruse fragments may get skipped based on z buffer tests etc. Feb 21 03:53:06 now precisely what is a silicon instructionwhena shader is compiled and what is thrown in by the compiler in generating code.. depends from gpu to gpu :) Feb 21 03:53:11 wolfyg: yo yo Feb 21 03:53:22 raster: hmm, so if i want to put a character cell on the display, can i tell the chip just to copy a block from A to B or do i have to express it in 4d space ? Feb 21 03:53:47 raster: yo yo what? Feb 21 03:53:47 wolfyg: if you manage to define a decent analog<->digital "API" you're halfway to success Feb 21 03:54:02 wpwrak: a character would either be a set of triangles u draw with a solid color - good luck with that... or a texture Feb 21 03:54:06 raster: did you look at that video? what do u think? Feb 21 03:54:17 uy generate the t6exture content from the char bitmap/alphamap (eg freetype2) Feb 21 03:54:19 DocScrutinizer: yes! :-) sounds like that's a good next step Feb 21 03:54:28 you then upload or map that directly as a whole texture or part of a larger texture Feb 21 03:54:44 as such the sgx has an mmu so it can mmap pretty much anything in ram Feb 21 03:54:48 DocScrutinizer: how does gnuradio's usrp's look ? Feb 21 03:54:51 but u generally write this as part of a texture Feb 21 03:55:02 a texture will probably have a format u are not familiar with - a toled format Feb 21 03:55:09 instead of linear (the good old framebuffer type) Feb 21 03:55:09 raster: i was afraid you'd say "texture" ... Feb 21 03:55:12 wpwrak: huh? Feb 21 03:55:17 u can have linear textures - but they take a performance hit Feb 21 03:55:22 anyway... u would fill a texture Feb 21 03:55:29 like a big pixmap/fb image with a bunch of character images Feb 21 03:55:44 then u will end up drawing a bunch of triangles Feb 21 03:55:47 2 per character Feb 21 03:56:29 DocScrutinizer: my strategy would be to try to build the full gsm stack with gnuradio. they already have some of the rf side of gsm working if i understand this right. Feb 21 03:56:33 that use parts of the texture containing the chars (you will need a nice table/structure to know what characger glyph maps to which geometry in the texture - also u'll need a 2d texture region allocation algorithm) Feb 21 03:56:42 and well. then u draw a whole chunk of triangles Feb 21 03:57:00 triangle in this case would be the vertex points, and color vectors (for color of the text) Feb 21 03:57:01 DocScrutinizer: then try to optimize for only gsm (if that makes sense) Feb 21 03:57:04 wpwrak: I still failed to have a close look yet Feb 21 03:57:19 if done right u can just prepare an array of vectors and blast them to opengl via gldrawarrays Feb 21 03:57:28 wpwrak: of course makes sense Feb 21 03:57:41 ort even better u get gl to allocate the buffer for thexe vertexes and use a vbo - if the content never changes u can re-use the vbo any time as its stored on the gpu Feb 21 03:57:52 (and u can apply vertex translations and so on to move it around) Feb 21 03:58:08 wpwrak: basicaly a chip isn't any different than a classical circuit Feb 21 03:58:09 tho invariably that doesnt work well with evas ad it generates these arrays/pipelines on the fly all the time Feb 21 03:58:18 anyway Feb 21 03:58:19 raster: so .. that should be reverse engineerable, no ? it's not that many operations. Feb 21 03:58:29 as such to just draw a character via the gpu is a farily involved task Feb 21 03:58:49 and you need to do a lot more infra if u want to use it for multiple things as u will be sharing the same pipe and rendeing path with multiple primitives Feb 21 03:59:02 wpwrak: just you can't probe, change, or quickfix. And you got a assload of new strange issues Feb 21 03:59:14 and so uneed the logic of setting up which shaders to use when, setting up state of the engine (blend func/mode, dest/src buffers etc.) Feb 21 03:59:14 DocScrutinizer: (make sense) yes, what i meant is i'm not sure you can "trim" the usrp or if it would be easier to design something specialized from scratch. Feb 21 03:59:32 wpwrak: it is possible - it is just a monumental task Feb 21 03:59:47 by the time u have figurted out half of he gpu.. its obsolete by 10 years and nothing has it anymore. Feb 21 03:59:53 its a losing battle :( Feb 21 04:00:04 DocScrutinizer: universal -> special would be very nice for development. replace one part, see if it works, try the next one, etc. if something regresses, you always have something to fall back to or to compare with. Feb 21 04:00:11 i dont think reversing a gpu is a game worth playing due to the complexity and nature of the beast Feb 21 04:00:20 yhe best bet is to convince someone with a gpu to go open Feb 21 04:00:24 wpwrak: step 1: use usrp, understamnd the issues. step2: go from scratch with what you've learnt Feb 21 04:00:27 wolfyg: will look in a sec Feb 21 04:00:59 DocScrutinizer: hmm, the harder path Feb 21 04:01:12 DocScrutinizer: no training wheels :) Feb 21 04:01:39 raster: sounds suckish. so we have to wait for wolfgang's chip then :) Feb 21 04:01:50 wolfyg: hmm video screws up 10% of the way thru Feb 21 04:01:51 wpwrak: you can't hope for evolution from a design not aimed to what you want. Feb 21 04:02:08 wpwrak: even then.. open gpu is fine.. but it will need to be within ballpark of its competitors Feb 21 04:02:12 :( Feb 21 04:02:18 if its 1/5th the speed Feb 21 04:02:21 its pointless Feb 21 04:02:26 if it uses 3x the power.. its pointless Feb 21 04:02:33 it needs to be within the same ballpark, or its not in the game Feb 21 04:02:54 as an fpga - uselss Feb 21 04:03:03 if it becomes a real asic... then there is something to talk about Feb 21 04:03:06 but.. its discrete Feb 21 04:03:12 raster: that's the point where (by my count) 15 prior projects stalled. And it fits on GSM chip as well Feb 21 04:03:16 whcih emans no one using n soc is interested.. unless its already on their soc Feb 21 04:03:17 :( Feb 21 04:03:41 ooh its blody java Feb 21 04:03:50 raster: video screws up? hmm Feb 21 04:03:54 yes Feb 21 04:04:02 it's a javascript player, ogg theora file Feb 21 04:04:09 which browser? Feb 21 04:04:13 its a java player Feb 21 04:04:14 there Feb 21 04:04:20 clicked on video itself and followed the links Feb 21 04:04:25 firefox Feb 21 04:04:32 i clicked on the play button below Feb 21 04:04:55 cant comment on the gpu Feb 21 04:05:03 what its drawing is rather simple Feb 21 04:05:23 I agree reversing a GPU is nonsense Feb 21 04:05:31 the whole reversing and anti-vendor stuff doesn't make sense tome Feb 21 04:05:37 its easy to do that in software Feb 21 04:05:41 it isnt accessing much source data Feb 21 04:05:46 hmm, just gassed about 20 gnats. will they ever learn that my tasty blood is not worth the risk ? :-( Feb 21 04:05:49 its interating a simple transform on 1 buffer Feb 21 04:05:56 and drawing some new buffer elements Feb 21 04:06:21 wpwrak: with 15 synapses? hardly Feb 21 04:06:29 raster: i would be happy with a decent 2d unit for now. maybe some video decoding help. Feb 21 04:06:54 wpwrak: to be honest.. for a modern soc. thats what your cpu is for Feb 21 04:07:01 as such the cpu can saturate the bus Feb 21 04:07:14 (if ua re not totally dumb in how you do things) Feb 21 04:07:37 in reality... the needs for uis are moving into the realms of needing 3d Feb 21 04:07:39 wolfyg: you still failed to invite me for occasional free revievs and comments Feb 21 04:07:42 so the reason I like Milkymist is because it will allow me to _SLOW DOWN_ SoC development Feb 21 04:07:49 they need blending, scaling, rotation ad any angle - hellp perspective too Feb 21 04:07:53 thats what the ui designers WANT Feb 21 04:07:54 raster: so you're saying you can do the 2d ops as efficiently with the cpu already. that's good then. Feb 21 04:07:55 wolfyg: (regarding "more money") Feb 21 04:07:56 DocScrutinizer: oh wow, that's an epic failure Feb 21 04:07:56 and they insist on having Feb 21 04:08:12 and as such the bar is being drawn much higher year after year Feb 21 04:08:25 wpwrak: for basic 2d - even things like alpha blending and so on... Feb 21 04:08:52 wolfyg: dunno - is that sarcastic? Feb 21 04:08:55 raster: good. that would cover most of the stuff you actually need :) Feb 21 04:10:20 DocScrutinizer: that's not sarcastic Feb 21 04:10:28 I always valued your input, and do. Feb 21 04:10:34 http://www.rasterman.com/files/cpuvsgpu.html Feb 21 04:10:39 for example Feb 21 04:10:41 the reason I wasn't coming up directly is that I have no money :-) Feb 21 04:10:43 only pennies Feb 21 04:10:53 (thats on a desktop - 3ghz core2 duo vs nvidia 8600gts) Feb 21 04:11:12 cpu and gpu are not far off Feb 21 04:11:14 let's see whether we can turn NanoNote into business, will be hard (but the right path I believe) Feb 21 04:11:19 cpu wins some, gpu, others Feb 21 04:11:25 and i've been optimising gpu a fair bit Feb 21 04:11:29 wolfyg: Your word you'l remember me once you got more bucks is enough for me Feb 21 04:12:08 oh wow, OK. thanks. let me write something... Feb 21 04:12:40 raster: nice. let's hope this is also true for the soc segment, with less beefy cores but also not quite so many afterburners attached to the gpu subsystems Feb 21 04:12:55 let me get u a software run on this device too Feb 21 04:13:03 i have results for soc too Feb 21 04:13:06 for gl that is Feb 21 04:13:11 will run a software test now Feb 21 04:13:26 also rememebr that evas is doing more complex ops too Feb 21 04:13:34 u can see - the simple stuff - software quals or can beat gl Feb 21 04:13:38 the comeplx stuff - gl wins Feb 21 04:14:22 so rectangles... software rumages along nicely Feb 21 04:14:35 anythign involving cpu generated data (pixels) sioftware can avoid copies Feb 21 04:15:05 and then "solid (non alpha blended stuff) software does well at too Feb 21 04:15:53 0polygon speed is my bad- my gpu version is brain-dead. Feb 21 04:15:55 wolfyg: I'm doing a lot for mere fun. And I'm a image whore, appreciating my name to show up in locations/contexts I feel ok with Feb 21 04:15:57 raster: okay, that's great Feb 21 04:16:12 as i dont tessalate - i literally scan-converty into rectangle spans and render them Feb 21 04:16:57 also the occlude tests actualyl do clipping Feb 21 04:17:02 i need to fine-tune that i think Feb 21 04:17:08 it could be better on the gpu side Feb 21 04:17:11 actually on all sides Feb 21 04:25:40 wpwrak: u ned up with simialr results Feb 21 04:26:18 raster: excellent. i like this :) Feb 21 04:26:26 end up Feb 21 04:26:32 ok cached things Feb 21 04:26:35 now runing real test Feb 21 04:26:46 this is an 800mhz cortex-a8 soc Feb 21 04:26:52 with sgx540 @ 200mhz on it Feb 21 04:26:54 fyi Feb 21 04:26:58 480x800 screen Feb 21 04:27:29 raster: so you got your "tall portrait" in the end ? :) Feb 21 04:27:43 * raster really needs to look into the occlusion handlign - gl does not like being thrown lots of geometry Feb 21 04:27:48 yes Feb 21 04:27:49 :) Feb 21 04:27:57 much better than vga Feb 21 04:28:08 i actually would prefer event taller Feb 21 04:28:16 but hey.. beggers cant be choosers Feb 21 04:28:18 *grin* i hope not too bulky. Feb 21 04:28:19 beggars Feb 21 04:28:27 its super-amoled :) Feb 21 04:28:36 (it means black is just that.. bitch bloody black) Feb 21 04:28:51 and whole screen technology is much thinner than lcd Feb 21 04:28:52 raster: I explained to tha random guest at pub some hours ago we (=gta02-core) are planning to do a better device *after* we "established the process" Feb 21 04:30:26 DocScrutinizer: make this "hoping to be able to plan" (as opposed to "phantasize"). still need to find that pot of gold. Feb 21 04:30:43 DocScrutinizer: or, better, to be found by it :) Feb 21 04:30:52 :-D Feb 21 04:31:37 DocScrutinizer: (alas, barristers of the late minister mugu don't count. there would be no shortage of them.) Feb 21 04:32:09 DocScrutinizer: as wpwrak said.. u need your pot of gold Feb 21 04:35:17 I for one still feel pessimistic on that Feb 21 04:36:14 DocScrutinizer: if they have billions to burn on banks, it can't be all that hard to find a few lousy millions for something real :) Feb 21 04:36:31 hehe Feb 21 04:37:19 probably raster chose the more efective way to change the world Feb 21 04:38:16 DocScrutinizer: we'll see how the final showdown goes :) Feb 21 04:38:19 fun is another separate topic Feb 21 04:38:19 i'd love to produce a totally open phone Feb 21 04:38:23 if only i had the money to do so Feb 21 04:38:31 so it's try to change things from the inside Feb 21 04:38:36 money problem solved Feb 21 04:38:38 hw problem solved Feb 21 04:38:43 production problems solved Feb 21 04:38:52 just need to solve the open bit Feb 21 04:38:57 and for me open doesnt include hw schematics Feb 21 04:39:10 i dont think that is part of it Feb 21 04:39:13 i feel guilty when i don't come here and chat. Feb 21 04:39:16 for me open is.. everything in my kernel and up Feb 21 04:39:35 raster: it's good to have. and it opens an almost virgin market Feb 21 04:39:35 (on my host cpu(s)) Feb 21 04:39:36 raster: I sypathize. Was my ife concept of 1980 Feb 21 04:39:40 Life Feb 21 04:40:51 http://www.hardwarezone.com/img/data/articles/2010/3147/Samsung_Wave_01.JPG Feb 21 04:40:52 there Feb 21 04:40:57 example of amoled vs super-amoled Feb 21 04:41:04 nice screen... i have to say Feb 21 04:41:13 lcd is just no comparison Feb 21 04:42:59 raster: does it age? (like e.g. plasma screens do, and OLEDS alwas have a big issue with oxidation from athmospheric O2) Feb 21 04:44:37 A nice brilliant screen isn't any more nice if it becomes dull after 6 months Feb 21 04:45:15 http://www.oled-display.net/files/u2/lcd.jpg Feb 21 04:45:16 lcd Feb 21 04:45:22 http://www.oled-display.net/files/u2/amoled.jpg Feb 21 04:45:24 oled Feb 21 04:45:34 http://www.oled-display.net/files/u2/super-amoled.jpg Feb 21 04:45:37 "supaaaa" Feb 21 04:45:52 ie basicalle less reflection, more transmittion less layers - thinner. Feb 21 04:46:00 oled ages Feb 21 04:46:01 always Feb 21 04:46:13 i'd just hope it ages slowly enough to "not matter" Feb 21 04:46:14 after how many thousands of operation hours? Feb 21 04:46:26 dunno Feb 21 04:46:33 http://www.youtube.com/watch?v=F-iP-4PzMS0&feature=player_embedded\ Feb 21 04:46:37 http://www.youtube.com/watch?v=F-iP-4PzMS0&feature=player_embedded Feb 21 04:46:38 oops Feb 21 04:46:40 extra \ Feb 21 04:46:45 anyway Feb 21 04:46:49 means its much thiner than lcd Feb 21 04:46:52 or even normal oled Feb 21 04:46:57 i guess it'll fade Feb 21 04:47:00 dont know how long Feb 21 04:47:03 but its a phone Feb 21 04:47:05 not a tv Feb 21 04:47:19 so really - hoe many 1000's of horus of display will u have.. on a phone? Feb 21 04:47:47 my N810 has more hours than my TV now (and TV is 24/7 on) :-P Feb 21 04:47:59 hegehehehehehee Feb 21 04:48:13 probably because u just leave it on - when it doesnt nedd to be Feb 21 04:48:19 but i'm being realistic Feb 21 04:48:20 :) Feb 21 04:48:20 and still works Feb 21 04:49:20 sure it's always on, as I know the LCD doesn't care and the LEDs have 100.000s of hours lifespan Feb 21 04:49:46 OLED though ages even when off :-S Feb 21 04:50:16 it will eventually fade i guess Feb 21 04:50:20 jhust a question of how long Feb 21 04:50:29 yep exactly Feb 21 04:50:40 doesnt matter i dont think Feb 21 04:50:44 mostly because ..l sales matter Feb 21 04:50:51 and peolpe are impressed with the thing n store Feb 21 04:50:59 not 5 years later Feb 21 04:51:09 as long as datasheet doesn't come up with numbers, I'm sceptic Feb 21 04:51:10 the people with it 5 years later who complain becuase its now 1/2 as bright Feb 21 04:51:16 because they had it on 24/7 Feb 21 04:51:20 well. thats very few peolpe Feb 21 04:51:34 and.. mostly will be dismissed as "what did u think would happen if u did that?" Feb 21 04:51:37 much like the battery Feb 21 04:51:38 :) Feb 21 04:52:43 24/7 doesn't care with LED (oxidation doesn't need electric current), and with 5 years I'm fine. Not with 5 months though Feb 21 04:53:05 OLED Feb 21 04:54:15 oxidation of the organic material is a funcion of perfect sealing and temperature Feb 21 04:55:00 wpwrak: http://www.rasterman.com/files/swvssgx.html Feb 21 04:55:14 software vs gpu rendering Feb 21 04:57:46 let me put it like taht: if they had decent figures, they for sure placed it in their datasheets Feb 21 04:57:52 quite good. same for many common ops, at worst about 50% gpu speed for slightly less common ones, and scores a few wins as well Feb 21 04:58:04 wpwrak: yeah... Feb 21 04:58:06 as i said Feb 21 04:58:07 it depends Feb 21 04:58:08 DocScrutinizer: maybe they don't quite know yet either ;-) Feb 21 04:58:14 some thigns the gpu races away at and is awesome at Feb 21 04:58:29 at others cpu rendering really works awesomly and can skip overhead the gpu has Feb 21 04:58:36 both are 32bpp Feb 21 04:58:42 gpu gets to page-flip Feb 21 04:58:49 software has to copy updtes to the fb Feb 21 04:59:03 sw gets to minimise updates, gpu has to re-render Feb 21 04:59:13 etc. etc. Feb 21 05:00:48 * DocScrutinizer dreaming strange dreams, totally OT Feb 21 05:01:32 raster: open kernel and up won't work, although I still wish you luck! Feb 21 05:02:14 things get integrated down, that's what makes a 100g smartphone that can run for days doing all sorts of things the way it is Feb 21 05:02:15 magnetron + 15.000uF@3KV Feb 21 05:02:53 plus a dish Feb 21 05:02:55 where do you think FOSS developed better? On the Linux kernel or the Window NT kernel? Why didn't people ignore the kernel and focused only on the apps? Feb 21 05:03:06 I like the gta02-core way of openness, start with the process & tools Feb 21 05:03:10 then up from there Feb 21 05:03:15 of course including ICs Feb 21 05:04:19 wolfyg: ack Feb 21 05:04:39 wolfyg: kernel and up will work... as such in reality gsm etc. stack is still on its own processor Feb 21 05:05:27 the rest is is pretty standard Feb 21 05:06:20 people bitch about maemo5/N900 for close BLOB for battery management Feb 21 05:06:29 thats bad Feb 21 05:06:33 it should be open Feb 21 05:07:30 yeah "warranty void if you treat battery nasty" Feb 21 05:07:41 thats fine by me Feb 21 05:08:03 not by JohnDoe NokiaBuyer though Feb 21 05:08:06 i'm happy for there to be some firmware that checks some signature Feb 21 05:08:08 u provide it Feb 21 05:08:12 esp if bat explodes Feb 21 05:08:14 only a signed kernel ca1n Feb 21 05:08:16 and if u dont Feb 21 05:08:24 it flips a bit with "u voided warranty" Feb 21 05:08:26 thats fine Feb 21 05:08:32 now its your problem - not the manufacturers Feb 21 05:08:50 so either u use signed and approved nokia kernels.. and keep warranty Feb 21 05:08:59 or.. u give up warranty and get to blow your own battery up Feb 21 05:09:08 hard to establish Feb 21 05:10:30 do we have any txqueuelen values suggested that yall have heard of ? for example: ifconfig usb0 txqueuelen 1000 Feb 21 05:11:05 errr...? Feb 21 05:11:30 calypso hw FIFO is 64bytes TX/RX Feb 21 05:11:59 ya, i meant ppp0 :) thanks Feb 21 05:15:21 raster: also people tend to feel really pissed about such approach Feb 21 05:18:02 raster: Nokia is worse with a checksumming cryptosignature BL for next device. Even when they promise it'll "only" deactivate the TPM fritzchip when loading unsigned kernel. You somehow feel like they stole your hw Feb 21 05:18:15 but nevertheless Feb 21 05:18:31 crypto-BL is *EVIL* Feb 21 05:19:35 and still Nokia had really a hard time to detect each violation of nontainted-kernel policy Feb 21 05:21:16 i also feel guilty about not learning the bitbake system and how to build an shr image. and not contributing to the community as well. Feb 21 05:21:26 i'll stop with that now. :) Feb 21 05:25:00 raster: and did you know that starting of just one unsigned app needs to kick all the TPM based infrastructure out of memory? Feb 21 05:25:28 DocScrutinizer: i'm not surprised Feb 21 05:25:37 any app might be a disguised gdb Feb 21 05:25:42 it makes sense as its probably fragile Feb 21 05:25:49 yup Feb 21 05:26:34 isn't DRM a great concept? :-o... hualp Feb 21 05:26:49 yay! Feb 21 05:27:12 i've been bad. tpm me harder ! :) Feb 21 05:27:25 :-D Feb 21 05:27:45 wpwrak: you always have a talent to brighten my day Feb 21 05:27:51 btw my personal opinion - drm can be as good as it can be bad Feb 21 05:27:58 so I think we should embrace it Feb 21 05:28:12 I actually also want drm, so I am sure only signed stuff from official debian repositories can run on my device :-) Feb 21 05:28:16 the freedom drm Feb 21 05:28:22 wolfyg: TPM is like polonium Feb 21 05:28:22 DocScrutinizer: heh, a good deed every day ... :) Feb 21 05:28:26 that makes sense Feb 21 05:28:55 I just upgraded my personal setup to route 100% out of the china crazy firewall via openvpn Feb 21 05:29:05 I see all sorts of crap here, man-in-the-middle attacks etc. Feb 21 05:29:10 I am tired of this. Feb 21 05:29:13 I want my freedom drm. Feb 21 05:29:31 wolfyg: drm is like door locks. few people mind to have lockable doors in their homes - as long it's them who hold the keys. if it's others that have the keys, that arrangement is usually called "prison cells" Feb 21 05:29:40 I totally think firefox should not accept the chinese NIC root certificate as a trusted source. it's a joke. Feb 21 05:29:47 it validates their man-in-the-middle shit Feb 21 05:29:59 wpwrak: right Feb 21 05:30:06 i am happy I have my openvpn here Feb 21 05:30:09 wpwrak: i think drm is actually ok if used wisely Feb 21 05:30:12 and a tunnel to good old bavaria Feb 21 05:30:16 eg if you buy a closed application Feb 21 05:30:23 u made the choise to buy it given those pre-conditions Feb 21 05:30:39 like I said on my notebook, I wouldn't mind at all to make it really hard/impossible to run any binary that is not a 'trusted' debian binary Feb 21 05:30:40 if that app uses drm to ensure that it is legally paid for according to its licensing - thats fine Feb 21 05:30:58 they can try to sneak in there as well, and will, but then they have the whole crowd watching over them, and someone will see the hole they are creeping through Feb 21 05:31:12 raster: the problem with such "choices" is that they're a way to trick the gullibles and then force the reluctant by sheer peer pressure Feb 21 05:31:43 not our problem if people are stupid and gullible Feb 21 05:31:47 raster: e.g., why do we have to put up with .doc and .xls if plain text would do in so many cases ? Feb 21 05:31:49 :) Feb 21 05:33:12 BAAAAHH Feb 21 05:33:16 raster: e.g., if the application to do your taxes only comes for the drm-enhanced system "everyone" uses, the task of doing your taxes all of a sudden acquires the extra dimension of holding back the police. Feb 21 05:33:22 24h disconnect Feb 21 05:33:32 LOL Feb 21 05:33:35 feed em man, feed 'em! Feb 21 05:33:39 only way to fight big brother is a DDoD Feb 21 05:33:43 err DDoS Feb 21 05:33:47 wolfyg: "impossible" isn't so good for developers Feb 21 05:33:47 wpwrak: then dont use it. Feb 21 05:33:47 OVERLOAAAAAAD Feb 21 05:33:50 you have a choice Feb 21 05:34:05 wpwrak: yeah but u get my point Feb 21 05:34:07 either tears down the net or the MitM Feb 21 05:34:09 I want freedom drm Feb 21 05:34:17 raster: like "you can always decline to get searched ... and not board the flight" ? Feb 21 05:34:17 I see mitm attacks here, for real Feb 21 05:34:20 it's happening Feb 21 05:34:31 wpwrak: thats a totally different matter Feb 21 05:34:47 I am in beijing, luckily I have openvpn right now. it will get harder in the future. Feb 21 05:34:49 raster: with ubiquitous drm, it wouldn't be Feb 21 05:34:51 wpwrak: "you can deny the police entry to your home if they dont have a warrant" is closer to a good comparison Feb 21 05:34:51 :) Feb 21 05:35:34 raster: how about "nobody will get near your house if you don't give the police a warrant" ? Feb 21 05:37:07 TPM and DRM (the same) is *BAD* Feb 21 05:37:17 you won't cjange that Feb 21 05:37:21 DocScrutinizer: what protects me from mitm attacks then? Feb 21 05:37:33 can't dance wit the devil Feb 21 05:37:47 wolfyg: (freedom drm) yeah, you can just make it optional. "do you really want to compromise your freedom for faster nvidia 3d ?" Feb 21 05:37:56 I absolutely need to establish safe connections to other people & organizations I trust, upon my choice. I may add or remove such organization as I want to. Feb 21 05:38:13 wolfyg: your kernel and Os built from source, and a openssh Feb 21 05:38:14 I would never trust the Chinese NIC root certificate. it's a real joke. Feb 21 05:38:24 the mozilla guys have no idea (of course all other browsers do it as well) Feb 21 05:38:31 the Chinese government _IS_ using this to spy on their own people Feb 21 05:38:42 wolfyg: so throw the certificate out ? Feb 21 05:38:47 yep Feb 21 05:38:48 of course Feb 21 05:38:52 wpwrak: exactly Feb 21 05:39:29 and in three weks you have the choice to either throw away your gadget, or ride the DRM train Feb 21 05:41:04 wolfyg: TPM *will NOt* save you from MitM Feb 21 05:41:33 au contraire Feb 21 05:42:48 you'll use ssh built by foobar government, and you'll never know what backdoor is in there Feb 21 05:43:21 as you kill your other apps as soon as you started a selfcompiled ssh Feb 21 05:43:47 it's not signed -> not trusted -> all trusted apps get kicked Feb 21 05:46:44 DocScrutinizer: DRM allows me to efficiently trust other groups of people (organizations) Feb 21 05:47:05 it's about speed. in or out. efficiency. I have a life too, cannot always thing about every library. Feb 21 05:47:10 no, networks of trust do that Feb 21 05:47:20 ok fine you are getting closer :-) Feb 21 05:47:25 see GPG Feb 21 05:47:48 nothing is perfect, sure not. but I want to have the (efficient) option of trusting (or not trusting) whole groups of people. Feb 21 05:48:22 so let's see how technologies develop, and which one will make this possible Feb 21 05:48:29 that's not related to kernel based (TPM chip based) DRM Feb 21 05:50:58 Digital Restrictions Management starts with your bootloader which refused to load unsigned kernels. It continues with kernels loaded via kexec voiding TPM. And it ends with *every* app needs to be signed, or you'll kick out of workspace *every* app/daemon/kernelmodule that relies on a trusted platform Feb 21 05:52:12 any app can be gdb! Feb 21 05:53:30 as soon as you start *one* app that's not signed, the platform is tainted and it'smandatory to purge *all* TPM-relying code Feb 21 05:53:50 *before* the unsigned app is even loaded Feb 21 05:54:29 TPM/DRM is *evil*. You can't play nice with this concept Feb 21 05:55:54 * juri_ ^5 doc. Feb 21 05:56:07 err... ^5 ?? Feb 21 05:56:15 high-five? Feb 21 05:56:23 aah :-) Feb 21 05:59:44 pheew. designing tables is tricky. Feb 21 06:04:12 tables are easy. relations arw tricky Feb 21 06:07:59 relations .. between the parts of the table Feb 21 06:08:46 a two-dimensional non-self-referring table is about as easy as it can get Feb 21 06:09:31 erm, i want mine in 3d ;) also, they're not entirely symmetrical Feb 21 06:09:46 if the table is self-referring though :-D See my proposal for opimd database of yesterday Feb 21 06:11:03 ouch - 3-d tables Feb 21 06:11:51 not many of database engines that can even handle this Feb 21 06:12:33 SQL can handle zillions of 2d tables in one databese Feb 21 06:13:00 referring back and forth like mad Feb 21 06:13:48 squillions! Feb 21 06:13:58 but things like irrational indezes... :-S Feb 21 06:15:26 DocScrutinizer: errr ... i mean real tables. the kind you put stuff upon ;-) Feb 21 06:15:45 you'd probably need something like prolog to even handle the API Feb 21 06:16:45 tzzz :-P Feb 21 06:17:50 DocScrutinizer: www.almesberger.net/misc/ndep/parts.ps Feb 21 06:18:03 (abusing fped a little :) Feb 21 06:21:01 dawn outside. and me missing the joke on my account. N8 Feb 21 06:22:07 s/account/debit/ Feb 21 06:22:07 DocScrutinizer meant: dawn outside. and me missing the joke on my debit. N8 Feb 21 06:22:27 sweet dreams ! Feb 21 06:30:03 wow, no postigs (except in #meego) for >1h Feb 21 06:30:20 goot time to say Feb 21 06:30:24 * DocScrutinizer waves Feb 21 06:30:59 yugo. meego to. Feb 21 06:31:13 meEgo Feb 21 06:31:30 meego pee go Feb 21 06:31:48 and off Feb 21 06:31:49 no .. meego pee Feb 21 06:32:38 yugo bed Feb 21 06:32:39 etc. Feb 21 07:11:30 I think using a trusted certificate for mitm is a real offense, worth reporting with actual data to mozilla, i'm sure they'll rip off the certificate. Feb 21 07:13:04 JesusMcCloud|zZz, how come? please elaborate, ciao. Feb 21 07:54:11 Folks, what wpa_supplicant version are you using on FR? Feb 21 07:54:27 lindi-: what about you? Feb 21 07:58:35 PaulFertser: shr wpa-supplicant_0.6.9-r5.1.4_armv4t.ipk Feb 21 07:58:55 JaMa: here it doesn't work while 0.6.4 does. Feb 21 07:58:58 (on debian) Feb 21 07:59:05 JaMa: have you tried it lately? Feb 21 07:59:12 yes every day.. Feb 21 07:59:30 well tested only on armv5 (spitz) Feb 21 07:59:42 FR is usually connected on USB Feb 21 08:01:31 JaMa: what matters is wpa_supplicant on FR. Feb 21 08:01:56 JaMa: i suspect it's doing something not well understood by the driver. Feb 21 08:02:29 hmm ubuntu version of util-linux-ng-2.16 or it's init script version stops booting when last mount time stored in superblock is in future (normaly I see this only as warning) :/ Feb 21 08:03:02 JaMa: also it'd be an interesting experiment if you (assuming it's not much of a hassle) could confirm that FR can connect using wpa/rsn to an ap with SSID=="Paul-g700af" but cannot to "Paul-g700ap". Feb 21 08:03:25 pity when time is not working properly and future is 1.1.1970 00:00:25 and now is 1.1.1970 00:00:17 :) Feb 21 08:05:33 PaulFertser: I'll retest with my ssid, but on my AP is more people connected so I won't change it now (maybe at night with less traffic) Feb 21 08:09:04 JaMa: i see. Feb 21 08:09:11 JaMa: thanks anyway Feb 21 08:17:53 JaMa: (newest wpa_supplicant) nevermind, it seems like i forgot to power-cycle the chip late evening yesterday. But ssid test is still worth a try. Feb 21 08:57:09 wpwrak: the problem with dhcp seem to be related to the fact wpa_supplicant never sees disconnect event on suspend/resume. And two CONNECTED in a raw has no effect for wpa_cli script. Feb 21 09:09:08 that is true, after suspend wpas connects successfully again but never disconnects explicitly - me too would be interested in getting to know how to react to that with a dhclient call Feb 21 09:12:01 josch: want a workaround? Feb 21 09:12:28 josch: http://paste.debian.net/60747/ Feb 21 09:18:32 SHR: 03mok 07phonefsod * rab089738b673 10/src/phonefsod-fso.c: handle lock = phone correctly - Patch by Joachim Ott Feb 21 09:30:45 Heh, appears i lack compiler to try my patched kernel :-O Feb 21 10:00:20 josch: wanna try a kernel patch? Feb 21 10:01:18 PaulFertser: wpasupplicant 0.6.10-1 Feb 21 10:01:51 PaulFertser: configuration files and log files are at http://iki.fi/lindi/eduroam/ Feb 21 10:01:59 lindi-: ok, i see. I did something wrong yesterday's night (not power-cycled the chip), so it works now. Can you try ssid test? Feb 21 10:02:02 Please :) Feb 21 10:02:36 i hope my AP is stable enough for it :) Feb 21 10:02:38 lindi-: i suppose that ar6000 can't connect to an AP with SSID "Paul-g700ap" with WPA/RSN but can connect to "Paul-g700af". Feb 21 10:02:49 At least that's a consistent behaviour here. Feb 21 10:02:55 my own AP is not using WPA Feb 21 10:03:05 does it happen without WPA? Feb 21 10:03:11 lindi-: i guess no Feb 21 10:03:21 i'd rather not mess with the AP right now :( Feb 21 10:03:25 lindi-: association completes but wpa negotiation times out. Feb 21 10:03:28 lindi-: ok, np Feb 21 10:03:33 it's openwrt svn head and barely stable Feb 21 10:03:46 (they haven't released a free version so i am forced to follow svn) Feb 21 10:03:46 lindi-: what hardware is that? Feb 21 10:03:53 linksys wrt54gl Feb 21 10:04:00 Oh, quite old? Feb 21 10:04:03 ye Feb 21 10:04:17 And still svn head? Weird. Feb 21 10:04:34 PaulFertser: the released version uses non-free driver Feb 21 10:04:40 http://lindi.iki.fi/lindi/openwrt.txt is the setup i have Feb 21 10:05:09 lindi-: i found another ar6000 gotcha: it doesn't send disconnection event and wpa_cli doesn't call my script to call dhcp because of that, i'm testing a fix atm. Feb 21 10:05:41 PaulFertser: can you tested if the essid problem occurs without WPA? Feb 21 10:05:46 s/tested/test/ Feb 21 10:06:08 lindi-: i guess i can do that a little bit later, yes Feb 21 10:06:39 PaulFertser: which one? Feb 21 10:07:01 PaulFertser: anyways, if you file a bug report about this then next time i am configuring wpa enabled AP I can test this if I can easily find the instructions from bugs.openmoko.org Feb 21 10:07:07 lindi-: but i doubt that, wpa_supplicant showed that association is ok and it received 1/4 message of 4-way handshake but got no response to its 2/4. Feb 21 10:07:13 lindi-: sure Feb 21 10:07:35 josch: http://paste.debian.net/60751/ Feb 21 10:09:11 I'll test it myself really soon, my gcc is being built atm. Feb 21 10:09:43 PaulFertser: okay then go ahead - i dont have much time atm :) Feb 21 10:09:49 PaulFertser: just out of curiosity, how do you build your cross-compilers? oe, debian packages, just manually? Feb 21 10:10:44 lindi-: i used to have OE on my laptop and i just used theirs. Right atm i just asked gentoo crossdev to build me one (and i do not care about armv4/armv5 difference because it doesn't matter for the kernel). Feb 21 10:11:10 PaulFertser: ok Feb 21 10:54:36 Heh, my patch doesn't work :/ Feb 21 11:04:55 yahoo Feb 21 11:06:02 wpwrak: here's the patch: http://paste.debian.net/60758/ Feb 21 11:20:41 PaulFert`: hmm, makes me wonder if wpa_supplicant is doing the right thing here. shouldn't it be enough to just tell it that we have something possibly new ? Feb 21 11:23:39 naw, others are doing it too. Feb 21 11:24:51 wpwrak: it's actually not wpa_supplicant, the logic to not call the script is inside wpa_cli. Feb 21 11:25:20 wpwrak: so, what's your resolution? Feb 21 11:26:27 you may want to set the address family, though, like in net/mac80211/mlme.c:ieee80211_sta_send_apinfo (now that mac80211 would be particularly consistent about either, e.g., ieee80211_sta_join_ibss in the same file) Feb 21 11:26:50 wpwrak: i checked wpa_supplicant sources, it doesn't pay attention to the address family it seems. Feb 21 11:27:12 wpwrak: but i agree, doing the things like mac80211 is reasonable. Feb 21 11:27:24 wpwrak: other than that, are you ok with it? Feb 21 11:27:34 your patch seems to make sense. oh. and add a blank line before the A_MEMZERO :) Feb 21 11:27:49 Sure Feb 21 11:27:58 yup, looks good Feb 21 11:28:45 nice find, by the way. now you have to solve the (E)SSID riddle ;-) Feb 21 11:29:33 wpwrak: there's no independent confirmation yet Feb 21 11:29:55 of the essid problem ? Feb 21 11:29:57 Yes Feb 21 11:30:38 wpwrak: also i do not know enough about that 4-way handshake process to be able to compare raw packets. Feb 21 11:30:39 maybe you found one of those one-in-2^80 hash collisions ;-) Feb 21 11:31:24 wpwrak: in fact now i need to do boring stuff Feb 21 11:31:38 : finally prepare the patch, write a decent mail blablabla Feb 21 11:32:56 there's a good book by o'reilly on wlan that describes most parts of the protocols in a decent enough way. not quite on par with "tcp/ip illustrated", but you can see where it took its inspirations from Feb 21 11:34:56 wpwrak: johill took a look at the dump and didn't find anything odd. Though he might have payed little attention to it -- he's always "too busy". Feb 21 11:37:33 guess you'll have to become an expert at wlan hermeneutics yourself :) Feb 21 11:38:39 Or just forget about odd cases with this stupid chip Feb 21 11:39:07 ;-)) Feb 21 11:41:22 wpwrak: what's more interesting about this "disconnected" issue is that when i tried to place the code in the proper place ("disconnect event" callback from the hardware layers) i couldn't connect at all since it emitted this signal right after it was asked to connect and hence confused wpa_supplicant. Feb 21 11:42:27 So i thought that it should at least signal disconnection on device removal. In the case of roaming the network id will be different so wpa_cli will notice it's not the same connected event and act accordingly. Feb 21 11:44:30 hmm yes, that bit isn't so nice indeed Feb 21 11:45:08 I do not have any better ideas Feb 21 11:45:14 And i'm not sure anyone should care anyway. Feb 21 11:47:11 :) Feb 21 11:47:34 well, ar6000 is a mess so if you make it a little messier, you don't to any real damage :) Feb 21 11:48:28 wpwrak: btw, remember our rfkill talks? linux-wireless guys told me that the modern notion is that maximum powersaving is obtained when all the interfaces are down, that's the established rule. Feb 21 11:48:54 "down" as in ifconfig eth0 down ? Feb 21 11:52:21 (oops :) Feb 21 11:56:06 wpwrak: yes Feb 21 11:56:27 wpwrak: new generation would call that "ip l s wlan0 down" Feb 21 11:56:29 ;) Feb 21 11:56:35 wpwrak: mail sent, thanks a lot! :) Feb 21 11:59:08 PaulFert`: thanks for making the ar6000 suck a little less :) Feb 21 12:30:33 could it be that one of the last upgrades introduced a speed regression? i loose nearly all my calls because the freerunner does not wake up fast enough and displays the call screen Feb 21 12:53:39 freesmartphone.org: 03mickey 07cornucopia * rafdd41a666e3 10/fsogsmd/src/plugins/dbus_service/plugin.vala: fsogsmd: dbus_service: start revamping the opening/closing logic Feb 21 13:07:18 freesmartphone.org: 03mickey 07specs * rae3965142f09 10/ (3 files in 3 dirs): Feb 21 13:07:18 freesmartphone.org: org.freesmartphone.Resource: add org.freesmartphone.Resource.UnableToEnable, raised when a resource can not Feb 21 13:07:18 freesmartphone.org: be enabled Feb 21 13:09:44 freesmartphone.org: 03mickey 07libfso-glib * r0ab6abf79ce5 10/src/freesmartphone.vala: regen Feb 21 13:13:56 freesmartphone.org: 03mickey 07specs * rdda3351ae0a7 10/ (2 files in 2 dirs): org.freesmartphone.Resource: and be able to raise the error from enable() as well Feb 21 13:14:32 freesmartphone.org: 03mickey 07libfso-glib * r64091b41e498 10/src/freesmartphone.vala: regen Feb 21 13:14:59 freesmartphone.org: 03mickey 07cornucopia * re46475f2dc7a 10/libfsoresource/fsoresource/dbusresource.vala: libfsoresource: enable now can raise a resource error as well Feb 21 13:20:53 freesmartphone.org: 03mickey 07cornucopia * r38363afe9c5d 10/fsogsmd/src/plugins/dbus_service/plugin.vala: fsogsmd: dbus_service: throw ResourceError, if modem can't be opened Feb 21 13:44:24 freesmartphone.org: 03mickey 07cornucopia * r8ba2a53d583a 10/fsogsmd/src/ (lib/modem.vala plugins/dbus_service/plugin.vala): Feb 21 13:44:25 freesmartphone.org: fsogsmd: serialize opening channels; open the main channel first, then the others. Feb 21 13:44:25 freesmartphone.org: If one channel can't be opened, power down the device. Feb 21 15:30:38 I've written some hints about making wifi work on FR, please someone repost this to shr-users (i guess i'm still not subscribed): http://lists.openmoko.org/pipermail/community/2010-February/060373.html Feb 21 15:32:14 interesting Feb 21 15:32:16 i should try that Feb 21 15:32:26 never been able to use WiFi on my FR Feb 21 15:32:31 doesn't like my router OS Feb 21 15:32:33 i guess Feb 21 15:32:59 but it might as well be the ssid as you found out Feb 21 15:34:47 mickey|sofa: i think the ssid thing is reproducible 100% here for me but i've no independent verification yet. Feb 21 15:35:41 *nod* Feb 21 15:35:54 mickey|sofa: i haven't tried yet but i guess that without encryption it'd work with any SSID because wpa_supplicant showed information about successful association and then timed out only during EAPOL exchange. Feb 21 15:36:28 right, but it doesn't Feb 21 15:36:32 it has never worked here Feb 21 15:36:39 no matter which encryption or without Feb 21 15:37:14 mickey|sofa: i remember your issue, it has a nice long discussion on trac with Werner looking at various bits of the dumps and the like :) Feb 21 15:37:22 yep Feb 21 15:37:49 in the end i just gave up Feb 21 15:37:59 <[Rui]> PaulFert`: I find that wifi works more or less reliably if I only use wpa Feb 21 15:38:22 <[Rui]> PaulFert`: I think I have had only about 1 attempt to successively use mokonnect (for instance) Feb 21 15:38:56 <[Rui]> I wrote a small shell script that just calls wpa_supplicant with the file sent by argument, then I call udhcpc and it usually works Feb 21 15:39:33 mickey|sofa: i got curious after i learned that my FR can work with exactly the same AP at university but fails to work at home, tried reproducing the configuration but what mattered was SSID that i changed after Werner's advice :) Feb 21 15:39:54 [Rui]: does it survive suspend/resume properly? With what i described it does. Feb 21 15:40:41 <[Rui]> PaulFert`: since the idle timeouts are very aggressive and changing them in the UI doesn't work reliably, I usually inhibit suspend and dimming when I want to be online Feb 21 15:41:20 <[Rui]> PaulFert`: I will, however, add your suggestions to my very small shell script (it saves the step of changing it on the setting sui) Feb 21 15:41:20 [Rui]: i take that as "no" ;) Feb 21 15:41:32 <[Rui]> PaulFert`: not exactly Feb 21 15:41:46 <[Rui]> PaulFert`: it's more that It's a PITA to wake up and reset everything Feb 21 15:41:54 [Rui]: i now realised there's a small typo, you should start it with "wpa_cli -a ..." the first time too, or the first instance won't get killed. Feb 21 15:42:05 <[Rui]> PaulFert`: you mean that with your instructions, even if it suspends when it resumes it will be working? Feb 21 15:42:13 [Rui]: of course Feb 21 15:42:20 <[Rui]> PaulFert`: ah... Feb 21 15:42:29 [Rui]: wpa_supplicant survives suspend/resume cycle and reestablishes the connection. Feb 21 15:42:51 But to get IP again you need something like wpa_cli -a script. Feb 21 15:43:10 <[Rui]> I'd like to understand why usb0 isn't working right after a reboot Feb 21 15:43:24 <[Rui]> I need to go into phone and do ifdown ifup Feb 21 15:46:04 <[Rui]> I just freed up 9MB of ram by resetting fsodeviced.log Feb 21 15:48:40 <[Rui]> PaulFert`: what do you mean with "there's a small typo, you should start it with "wpa_cli -a ..." the first time too, or the first instance won't get killed." ??? Feb 21 15:48:51 <[Rui]> all wpa_cli calls have -a Feb 21 15:51:27 [Rui]: pkill -f "wpa_cli -a" will kill only the one that has -a immediately after wpa_cli. Feb 21 15:52:02 [Rui]: killing all wpa_cli instances is impractical as you might have it running in a terminal to select networks or monitor events. Feb 21 15:52:04 <[Rui]> PaulFert`: ah ok, just a matter of reversing the argument order. np Feb 21 15:52:59 GNUtoo: if you use the touchscreen Feb 21 15:53:02 it gets broken Feb 21 15:53:10 nothing works anymore Feb 21 15:53:14 heh Feb 21 15:53:17 what did you do to the kernel? :) Feb 21 15:53:21 neither touchscreen nor trackball Feb 21 15:53:27 uhhm Feb 21 15:53:36 seems to be something in userspace Feb 21 15:53:44 the devices still give input -.- Feb 21 15:53:51 leviathan, image compiled and still touchscreen problem? Feb 21 15:53:56 jup Feb 21 15:53:58 ok Feb 21 15:55:19 and JesusMcCloud|zZz nEo-Theme aka gry fucks somehow up because of different resolution... Feb 21 15:55:20 hmm Feb 21 15:55:40 in short worts: SNAFU Feb 21 15:55:47 situation normal: all fucked up Feb 21 15:55:48 ;) Feb 21 15:58:08 good, i will not hurry moving into UI then but stay in my comfortable console-based environment for the time being ;) Feb 21 16:01:49 mickey|sofa: can you tell me why can it be the problem of frameworkd and fso-abyss on loglevels lower then DEBUG now? Not so long ago there were the same problems if set to anything lower than INFO Feb 21 16:02:44 i don't know. how does it fail exactly? do you have a log of a failed attempt? Feb 21 16:03:05 did you try without the upper layers, i.e. SHR's phonefsod and UI? Feb 21 16:03:16 GNUtoo: the error seems to lie in enlightement Feb 21 16:03:22 i should really resurrect zhone Feb 21 16:03:26 leviathan, ok Feb 21 16:03:28 if .e does not exist it recogniues all inputs Feb 21 16:03:32 we need an alternative stack for testing Feb 21 16:04:23 I tried evtest the other day and it seemed to work normally Feb 21 16:04:38 maybe X+Gtk? Feb 21 16:05:16 uhm Feb 21 16:05:17 yes Feb 21 16:05:28 maybe I should try building an opie image Feb 21 16:05:48 would be nice anyway to try it in combination with fso and lightphone ;D Feb 21 16:05:55 yeah Feb 21 16:06:03 mickey|sofa: this might be the log you need: http://trac.shr-project.org/trac/ticket/957 in attachment. Feb 21 16:06:20 as soon as I'm done with fsogsmd, I will cook a simple UI similar to zhone Feb 21 16:06:32 i need something simple to show off the middleware Feb 21 16:06:39 Q-Master: i'll take a look Feb 21 16:07:14 ok Feb 21 16:07:34 some people asked me, if we could integrate qt into SHR Feb 21 16:07:35 mickey|sofa: also you can read my comments in shr-devel and shr-user on that topic Feb 21 16:07:52 leviathan: there also are qt libs in SHR Feb 21 16:07:58 already Feb 21 16:16:04 yes Feb 21 16:16:10 thats what I told them too Feb 21 16:16:29 ok Feb 21 16:16:31 well Feb 21 16:16:37 now I'll build opie Feb 21 16:16:44 lets try if it boots ^^ Feb 21 16:16:59 <[Rui]> PaulFert`: wpa_cli doesn't seem to work here: Failed to connect to wpa_supplicant - wpa_ctrl_open: No such file or directory Feb 21 16:18:56 heh Feb 21 16:19:01 Opie... blast from the past... Feb 21 16:19:35 gpe is better and is fixed by #openpma people Feb 21 16:19:41 btw it seem to be xorg Feb 21 16:19:44 check abi etc... Feb 21 16:20:00 but it would tell normaly if there is an abi mismatch Feb 21 16:23:29 <[Rui]> fsoraw exit code is 2 Feb 21 16:24:18 JesusMcCloud|zZz, please elaborate concerning the script Feb 21 16:24:24 mrmoku|away, do you use the new opimd? Feb 21 16:24:29 Q-Master, does the new opimd work for yoU? Feb 21 16:26:02 leviathan, it's xorg.conf Feb 21 16:26:06 fix it Feb 21 16:26:14 touchscreen is now working for me Feb 21 16:27:09 uuhm Feb 21 16:27:13 can I see it? Feb 21 16:27:29 our can you push it into gitorious? Feb 21 16:27:39 ok not fully functional yet Feb 21 16:27:56 http://gitorious.org/shr-msm/shr-msm/commits/org.openembedded.dev Feb 21 16:28:01 hmm Feb 21 16:28:03 okee Feb 21 16:28:06 http://www.pastebin.com/m5970ec44 Feb 21 16:28:24 CoreSomething + the last line Feb 21 16:28:29 is the diff Feb 21 16:28:34 I'll try to remove last line Feb 21 16:29:33 ok Feb 21 16:32:42 <[Rui]> PaulFert`: nothing works here Feb 21 16:38:08 <[Rui]> PaulFert`: I get "enabling2 followed by "disabling" in fsodeviced.log Feb 21 16:40:41 <[Rui]> PaulFert`: duh! such an idiot that I am. no bash here. Feb 21 16:51:08 mickeyl: I've opie running on htcdream Feb 21 16:51:13 its AWSOME! Feb 21 16:56:43 <[Rui]> PaulFert`: the wpa_cli thing it's also not working very well Feb 21 17:01:31 leviathan, why not gpe that got some attention recently Feb 21 17:01:32 ? Feb 21 17:02:23 I could also build a gpe image Feb 21 17:02:37 I was just wondering how opie works with our new kernel Feb 21 17:02:53 ~62MB RAM free Feb 21 17:02:57 awsome Feb 21 17:02:59 ^^ Feb 21 17:03:11 used 13MB Feb 21 17:03:32 yeah Feb 21 17:03:35 it was quite ahead of its time Feb 21 17:03:47 unfortunately cooperation with trolltech didn't work out too well Feb 21 17:03:53 otherwise we had continued working on it Feb 21 17:04:14 AND its sooo fast Feb 21 17:05:56 aloha! Feb 21 17:06:47 JesusMcCloud: I've got opie running on htcdream Feb 21 17:06:54 its absolutly awsome Feb 21 17:06:55 ^^ Feb 21 17:07:34 TAsn, well i havent investiagted further, but using latest convert script and latest opimd, all default fields are generated. but the the segfault problem is there again) Feb 21 17:08:25 freesmartphone.org: 03tom 07opimd-redesign * re883710ad301 10framework/framework/subsystems/opimd/db_handler.py: Fixed a typo. Feb 21 17:08:37 leviathan, i know about resolution problmes, but i can'T fix it without the device Feb 21 17:08:49 uhhm Feb 21 17:08:59 hm.. mind sending me your db yet again? :P Feb 21 17:09:17 JesusMcCloud: what about simply taking the resolution from bb variables Feb 21 17:10:46 leviathan, some parts of the theme are hardcoded (flaunch bar at the bottom is hackish for example, clock size is hardcoded, or didi this change?!) Feb 21 17:10:53 [Rui]: for wpa_cli to work you need to specify ctrl_iface in config Feb 21 17:11:28 <[Rui]> PaulFert`: no... I think that since I'm trying to do the steps in a script, I just need to put a few more sleeps Feb 21 17:11:43 <[Rui]> but I still don't get IP Feb 21 17:12:07 [Rui]: check it works the manual way first... Feb 21 17:13:57 <[Rui]> http://pastebin.com/d314cb19a Feb 21 17:14:06 <[Rui]> PaulFert`: there must be a bug in wifi-wpa.sh Feb 21 17:14:24 <[Rui]> PaulFert`: I tried an alternative form by using explicit if condition, but it's just the same. Feb 21 17:14:55 <[Rui]> PaulFert`: no if address until I call udhcp manually. perhaps the $2 is not happening ok? Feb 21 17:15:33 uhm, you could use svg and compile pngs out of it while building the package in bb Feb 21 17:15:54 [Rui]: it's written dhclient but you use udhcp usually? And probably the script should be made executable. Feb 21 17:17:08 <[Rui]> PaulFert`: ok, I think I need coffee :| *feeling sleepy dumb* second obvious problem from a non existing program :D Feb 21 17:18:13 <[Rui]> got IP Feb 21 17:21:19 <[Rui]> I'm unplugging the phone, and will let it suspend, now. Feb 21 17:21:32 JesusMcCloud: uhm, you could use svg and compile pngs out of it while building the package in bb Feb 21 17:21:57 leviathan, neither buttons not vlock use images Feb 21 17:22:27 flaunch will need a varaible from bb in any case, like you suggested Feb 21 17:24:08 <[Rui]> PaulFert`: yup, seems to work Feb 21 17:24:22 [Rui]: cool, thanks for the confirmatio Feb 21 17:26:13 <[Rui]> PaulFert`: looking forward to a kernel with the patches :) (and a working kernel without debug) Feb 21 17:27:41 [Rui]: this one patch will go in shortly. As to working kernel without debug... I really hope larsc has something in the works. Feb 21 17:28:14 hm Feb 21 17:28:46 <[Rui]> these patches are going upstream, right? Feb 21 17:30:39 TAsn, if i am not completeley mistaken b6eaa81242850845a77287cb96088b4788d25c7b <- *works* but 3b867ce2400c2d1dc63a7f24ed054b03ff7fc994 <- is broken Feb 21 17:31:46 will take a look Feb 21 17:32:31 <[Rui]> PaulFertser: snet an email with this findings. Feb 21 17:33:04 leviathan, could you try bos1's niebiee theme? (or did u kick out shr in favor of opie?) he might have fixed the issues, since he knows edje better than me.. Feb 21 17:33:27 leviathan, or even better, could you upload sreenies to scap? Feb 21 17:33:35 <[Rui]> niebiee is unusable in dayligh Feb 21 17:33:38 <[Rui]> daylight Feb 21 17:33:51 hmm, I will try it Feb 21 17:33:58 just wanted to try opie ^^ Feb 21 17:35:45 [Rui]: you know that wpa_supplicant allows to specify any amount of different networks with different settings (including open and wep) and it'll choose the one based on the scan results. And also you can change to a particular network by using wpa_cli. Feb 21 17:36:10 <[Rui]> PaulFertser: yes, but that requires more typing :) Feb 21 17:37:19 [Rui]: usually you have only one network you connect to in any particular place anyway. Feb 21 17:37:41 <[Rui]> PaulFertser: what I fear is a network with the same name :) Feb 21 17:38:13 [Rui], you can prolly blame the poor contrast of niebiee on the weather where dos comes from :P Feb 21 17:38:55 <[Rui]> JesusMcCloud: yeah, with less sunlight than Portugal :) Feb 21 17:39:30 <[Rui]> JesusMcCloud: dark text on light background is what works best (the more extreme the contrast the better). Feb 21 17:39:33 I'm afraid FR's screen is unusable in strong light with any theme. Feb 21 17:39:41 TAsn: thanks. It works much better than previous. Feb 21 17:39:54 [Rui], i developed the first version of neo in summer, but still the results could be better Feb 21 17:40:01 TAsn: even name resolving works flawlessly! Feb 21 17:40:17 sweet :P Feb 21 17:40:27 Q-Master, tell JesusMcCloud about it, he's very skeptical :P Feb 21 17:40:28 <[Rui]> I'm using the default, gry, and I only hate to death the bubbles. Feb 21 17:40:30 ^ ^ Feb 21 17:40:48 TAsn: but you hadn't told me about that I need to swap an phonelog also. 8) Feb 21 17:40:55 hi mickeyl, how do you change font size in vala terminal by default,gtkrc seem to have some effect on midori but not on vala-terminal Feb 21 17:40:56 yeah :P Feb 21 17:41:00 forgot about it Feb 21 17:41:04 <[Rui]> it's better to have no background than 50/50 of two colors Feb 21 17:41:04 [Rui], what's wrong with bubbles? Feb 21 17:41:31 JesusMcCloud: Are you an author of a Neo theme? Feb 21 17:41:41 Q-Master, yup Feb 21 17:41:51 ah seem to work from commandline Feb 21 17:42:12 and seem hardcoded into the application => .desktop Feb 21 17:42:37 JesusMcCloud: there's a problem with your theme in contacts. Top part of the contact where Name Surname Phone is displayed black on black. 8( Feb 21 17:42:45 Q-Master, niebiee and neo used to be very similar (internal structure) Feb 21 17:42:52 Q-Master, that is a bug in phoneui Feb 21 17:42:58 8) Feb 21 17:43:07 <[Rui]> JesusMcCloud: the buttons suffer a bit of the same syndrome Feb 21 17:43:10 or to be more specific lobphone-ui-shr Feb 21 17:43:22 Q-Master, u can work around it though Feb 21 17:44:05 Q-Master, just cp /usr/share/libphone-ui-shr/neo.edj /usr/share/libphone-ui-shr/default.edj Feb 21 17:44:13 Q-Master, and tehn killall phoneuid Feb 21 17:44:22 aha Feb 21 17:45:02 Q-Master, you do have phone-ui-shr-theme-neo installed, do you? Feb 21 17:45:20 yes Feb 21 17:45:48 JesusMcCloud: why didn't you make a bootsplash also? Feb 21 17:46:40 Q-Master, because i dont know how (but you can activate the e17 splashscreen, i customized this one) Feb 21 17:47:50 JesusMcCloud: I mean shr-splash-theme-niebiee. Feb 21 17:48:55 Q-Master, i know what you mean, i just dont know how to make one (it is not a e issue) Feb 21 17:50:03 JesusMcCloud: It's just a bzipped copy of a framebuffer. Feb 21 17:50:26 GNUtoo: do you run the xserver with the correct DPI? Feb 21 17:50:39 JesusMcCloud: It is unbzipped to /dev/fb0 Feb 21 17:51:03 I don't know but the other app seem ok Feb 21 17:51:53 Q-Master, if you know hot to do it just cd /usr/share/enlightenment/data/themes and edje_decc neo.edj, then use the boot_image.png form the resulting neo directory Feb 21 17:53:50 JesusMcCloud: phoneui theme helped. thnx Feb 21 17:54:14 Q-Master, np Feb 21 17:54:34 JesusMcCloud: how can I enable e's boot splash& Feb 21 17:54:35 ? Feb 21 17:55:02 found it Feb 21 17:55:53 Q-Master, it will only give you a splashscreen upon starting X Feb 21 17:56:02 JesusMcCloud: I know. Feb 21 18:00:13 lindi-: mickeyl: without wpa it works even with "bad" ssid Feb 21 18:01:16 interesting, so my case is truly special Feb 21 18:01:22 Yeah :) Feb 21 18:02:43 By works i meant i got IP with dhcp. Feb 21 18:02:47 No further testing. Feb 21 18:03:08 leviathan, can you rsync the build system Feb 21 18:03:46 So, any volunteers to try the "bad" essid and WPA on their AP with FR? Feb 21 18:08:26 SHR: 03jesus 07shr-themes * r820f4d41531a 10/e-wm/e-wm-theme-illume-neo/ (7 files): altered bootscreen Feb 21 18:09:17 "bad" ssid? Feb 21 18:09:29 hi Feb 21 18:09:34 SHR: 03jesus 07shr-themes * rfb377425e69c 10/elementary/elementary-theme-neo/ (20 files): probalby fixed sth in elementary theme, but i dunno Feb 21 18:09:41 What have you discovered, Mr. PaulFertser? Feb 21 18:11:10 SHR: 03jesus 07shr-themes * r27e2247d7dce 10/phoneui-shr/phoneui-shr-theme-neo/ (31 files in 2 dirs): added idle screen. recipe needs to be adapted. does idle screen even belong here? also fixed sim auth (hopefully Feb 21 18:11:17 a quick one: is it still the way that after one defines bluetooth headset in ...conf/phone/default.yaml then all calls are routed there? Feb 21 18:11:53 or is there a way already one can keep the config but switch from bt headset to handset? Feb 21 18:12:45 i have been updating the shr manual and would like to straighten this one up.... Feb 21 18:15:12 GNUtoo: what? Feb 21 18:17:22 GNUtoo: rsync the builldsystem?? Feb 21 18:17:41 how do you mean that? Feb 21 18:18:15 mwester: may i direct you to read this: http://lists.openmoko.org/pipermail/community/2010-February/060373.html Feb 21 18:18:48 vanous123: yeah, still same, nothing's changed :( Feb 21 18:19:43 mrmoku|away, a present: Feb 21 18:19:57 PaulFertser: thank you, i cannot test, my bt is broken and i found the issue this week Feb 21 18:19:58 SHR: 03tom 07libphone-ui-shr * rf52c5b0d300c 10/src/ (util/ui-utils.c view/contact-view.c): fixed the value not getting removed issue in contact-view Feb 21 18:20:20 PaulFertser: (well, could plug another bt in but have no working headset) Feb 21 18:21:28 PaulFertser: the flat flexible wire loom going to the bt module is cracked at the bend. three wires broken. will still try to fix that ( later on this year :) Feb 21 18:22:00 vanous123: how did it happen? Feb 21 18:22:23 PaulFertser: have no clue Feb 21 18:22:33 PaulFertser: but, it appeared after 1024 fix Feb 21 18:22:53 PaulFertser: so perhaps disassembly/assembly... Feb 21 18:23:06 yep, might be Feb 21 18:23:18 PaulFertser: perhaps some sudden stops after free falls? anything.. Feb 21 18:23:31 vanous123: :( Feb 21 18:24:31 one more question here :) Feb 21 18:25:37 gms on latest shr-u still has startup issues - just doesn't come up. usb connection helps... is there any cure now? Feb 21 18:25:37 vanous123: you should replace it by a standard flat ribbon cable Feb 21 18:25:51 DocScrutinizer51: that is my plan Feb 21 18:26:38 TAsn, did u try converting an old db with the latest convert script? Feb 21 18:27:09 no, I fixed this bug in libphone-ui-shr (I just committed) Feb 21 18:27:22 sec, let me check the revisions you gave me Feb 21 18:27:30 PaulFertser, that's an amazing bug. :o Feb 21 18:27:58 Is there a pattern of any sort, any special characters, or bit pattern, or anything known on that? Feb 21 18:28:30 JesusMcCloud, no such revisions... Feb 21 18:28:54 TAsn, hmpf Feb 21 18:29:22 give me the correct revisions :P Feb 21 18:29:25 And I'll be able to check Feb 21 18:29:38 vanous123: wait. the BT has no real PCB, the whole module is soldered to the FPC :-( Feb 21 18:29:54 DocScrutinizer51: i see Feb 21 18:29:57 here is a pic Feb 21 18:29:59 http://vanous.penguin.cz/files/om/bt/2010-02-16-141159.jpg Feb 21 18:30:15 the trouble is, that the bend is hard to get under a micro to solder on it Feb 21 18:30:20 TAsn, i just downloaded form git and i still have the tarballs, shoudl the rev be part of the filename? Feb 21 18:30:35 probably Feb 21 18:30:36 but i am sure something will workout there... Feb 21 18:30:52 (no idea, but maybe) Feb 21 18:31:21 TAsn, well those were the strings appended to the tarball filename... Feb 21 18:31:31 weird Feb 21 18:31:59 vanous123: is the blue thing the main PCB? Feb 21 18:31:59 TAsn, how else can i find out which rev it was? Feb 21 18:32:17 JesusMcCloud, not even the slightest idea Feb 21 18:32:30 DocScrutinizer51: yes. is it possible to make a real soldering on the FPC? Feb 21 18:32:44 vanous123: YES Feb 21 18:32:48 oops Feb 21 18:34:41 DocScrutinizer: well, will see. i am not using bt for anything atm. thank you for the comments :) Feb 21 18:34:58 vanous123: I suggest you unsolder FPC at mainboard side, glue some support to the backside of the crack, bridge the cut traces, then resolder to main PCB Feb 21 18:36:25 DocScrutinizer: this makes sense. especially as getting the bend under the micro is impossible to focus on Feb 21 18:36:46 TAsn, just try the latest version and tell me if it works Feb 21 18:36:59 TAsn, including convert, preferably using my db Feb 21 18:37:07 so tell me your symptoms... Feb 21 18:37:52 TAsn, default fields were generated (with types, i checked that in contact screen) but messages and callscrenn were segfaultign again Feb 21 18:38:02 i c Feb 21 18:38:05 ok Feb 21 18:39:20 hehe. opensource GSM sw for TI Calypso. 8) Feb 21 18:42:33 Q-Master: pointer? Feb 21 18:42:46 DocScrutinizer: git://git.osmocom.org/osmocom-bb.git Feb 21 18:42:49 TAsn, do you have a new local build of libphone-ui-shr for me? Feb 21 18:42:56 no Feb 21 18:43:00 because there is one more issue Feb 21 18:43:03 so sec :P Feb 21 18:43:16 SHR: 03david.lanzendoerfer 07shr-themes * rb7ec69247ae0 10/alsa-scenarii-shr/htcdream/stereoout.state: Adding temporaraly stereoout.state for htcdream (fixing bb bildissue) Feb 21 18:43:27 SHR: 03david.lanzendoerfer 07shr-themes * r9508e9b149cc 10/ (60 files in 4 dirs): Merge branch 'master' of git+ssh://git.shr-project.org/shr-themes Feb 21 18:43:33 DocScrutinizer: http://lists.osmocom.org/pipermail/osmocom-announce/2010-February/000000.html Feb 21 18:44:59 mwester: i couldn't determine any Feb 21 18:45:10 How odd. Feb 21 18:45:11 JesusMcCloud, I can send you a buggy version :P (if you want) Feb 21 18:45:27 TAsn, define 'buggy' Feb 21 18:46:09 I shall try my other AP in WPA then; I have been using my "lightly-secured" WEP AP because i just assumed the problem was WPA was broken... Feb 21 18:46:40 JesusMcCloud, everything works except renaming a field (which copies a field instead of renaming) Feb 21 18:46:44 but you can always remove the old one Feb 21 18:46:47 and it'll just work. Feb 21 18:47:02 anyhow, testing db conversion: will do now. Feb 21 18:47:13 TAsn, any progress on the messages bug? Feb 21 18:47:24 what messages bug? Feb 21 18:47:42 TAsn, adding receipients Feb 21 18:47:48 no Feb 21 18:47:53 *no idea Feb 21 18:47:56 did you try it? Feb 21 18:48:16 nvm Feb 21 18:48:19 doesn't work :P Feb 21 18:48:36 TAsn, yup doesnt work Feb 21 18:49:37 sec, converting my db with latest script Feb 21 18:49:38 Q-Master: thanks Feb 21 18:50:21 JesusMcCloud, converting Feb 21 18:50:24 * TAsn is going to eat. Feb 21 18:50:25 Q-Master: I knew they were tackling that, but I'm amazed they got drivers and layer one already done Feb 21 18:51:19 Q-Master: Dieter made great progress at reverse engineering the friggin TI blobs it seems Feb 21 18:51:33 leviathan, I mean that I could opkg update && opkg upgrade Feb 21 18:51:37 Q-Master: full control over DSP, wow Feb 21 18:52:24 DocScrutinizer, ??? somebody has replaced OMAP DSP's OS ? Feb 21 18:52:33 Q-Master: that should mean we finally can shed some light on the AT%Nxxxx mystery Feb 21 18:52:49 ah Dieter sparr Feb 21 18:52:52 so telephony Feb 21 18:53:11 not Multimedia Feb 21 18:53:14 wow Feb 21 18:53:16 yep Feb 21 18:53:40 oh Feb 21 18:53:51 you mean update the repository? Feb 21 18:53:53 nope Feb 21 18:53:54 mom Feb 21 18:53:56 ^^ Feb 21 18:54:09 leviathan, indeed Feb 21 18:54:46 its updating Feb 21 18:54:57 will take some while Feb 21 18:55:37 GNUtoo: btw: you didnt push the new xorg.conf?? Feb 21 18:55:53 and my stomache makes noises Feb 21 18:55:58 leviathan, I thought you would test it and do it, I was eating Feb 21 18:56:04 I need to look for some nurations Feb 21 18:56:11 oh Feb 21 18:56:12 ok Feb 21 18:56:14 *nutrition Feb 21 18:56:14 ok Feb 21 18:56:18 uhm Feb 21 18:56:19 sry Feb 21 18:57:06 leviathan, I've an issue with my wifi antenna Feb 21 18:57:09 GNUtoo: but you're using fbdev again Feb 21 18:57:23 GNUtoo: laptop or dream?? Feb 21 18:57:24 leviathan, yes I needed it for testing gtk apps Feb 21 18:57:31 oh, ok Feb 21 18:57:36 leviathan, so adapt it Feb 21 18:57:54 ok Feb 21 18:58:06 ok, now I need to take food to me Feb 21 18:58:13 ??? Feb 21 18:58:13 ah wifi Feb 21 18:58:13 desktop Feb 21 18:58:13 need some soldering Feb 21 18:58:14 I'll be back in some minutes Feb 21 18:58:20 ah, ok Feb 21 18:58:22 looking forward to publish FREEMOKO01 eventually :-D Feb 21 18:58:33 :) Feb 21 18:58:45 DocScrutinizer: OpenMoko is free ;P Feb 21 18:59:01 MOKO11 not though Feb 21 18:59:22 leviathan, not its gsm baseband firmware Feb 21 18:59:49 JesusMcCloud, done Feb 21 18:59:53 damn Feb 21 18:59:54 lol Feb 21 19:03:01 * mwester gets the bar of soap ready Feb 21 19:10:07 mwester: is there any saying similar to "every joke has its part of a joke"? Feb 21 19:10:52 <|nschle85|> TAsn: which receipe builds libstdc++.so.6 ? is gcc the right one or is it too much ? Feb 21 19:18:38 DocScrutinizer: seems that that would be not so fast as wanted to... Feb 21 19:18:46 |nschle85|, sounds way too much, Feb 21 19:19:04 Q-Master: why? Feb 21 19:19:19 |nschle85|, but no idea Feb 21 19:19:55 try Feb 21 19:19:57 libstdc++6 Feb 21 19:20:03 this should be it :P Feb 21 19:20:28 Q-Master: Harald und Dieter are really familiar with that GSM OTA shit now, after openBSC Feb 21 19:21:31 DocScrutinizer: Heh. They're just 2. I don't think that GSM fw could be written in concious time by 2 ppl in a free time progging. Feb 21 19:21:33 <|nschle85|> TAsn: its too late for testing, mesa-dri now compiles :-) (used gcc receipe) Feb 21 19:21:44 DocScrutinizer: better me be wrong Feb 21 19:22:00 Q-Master: you'd be amazed Feb 21 19:22:19 Q-Master: but for sure it will take another few months Feb 21 19:26:41 Q-Master: also I bet there are more than just Harald and Dieter Feb 21 19:29:44 TAsn, i'll get me sth to eat now, tell me late about the convert test Feb 21 19:29:59 JesusMcCloud, worked. Feb 21 19:30:08 nothing more to say :P Feb 21 19:30:15 TAsn, coolio Feb 21 19:30:22 though it's maybe because I use an homebrew non-backward compatible libphone-ui Feb 21 19:30:27 braggard Feb 21 19:30:27 (backward = old opimd) Feb 21 19:30:47 And there's no reason why to check old broken libphone-ui Feb 21 19:30:57 anyhow, I'm going to watch tv Feb 21 19:30:58 ciao. Feb 21 19:31:02 TAsn, please also tell me about the progress of you local build linbphe-ui-shr Feb 21 19:31:07 enjoy :-) Feb 21 19:31:09 as I said Feb 21 19:31:10 * JesusMcCloud|afk off to eat Feb 21 19:31:10 it's done Feb 21 19:31:14 DocScrutinizer, thanks Feb 21 19:31:17 JesusMcCloud|afk, but there's that bug Feb 21 19:31:24 talk to ya later. Feb 21 19:32:42 <|nschle85|> TAsn: sucessfully compiled shr-unstable image from scratch, thanks to shr team Feb 21 19:33:01 GNUtoo: I'm back Feb 21 19:33:09 leviathan, ok Feb 21 19:33:33 hmm Feb 21 19:33:39 free baseband firmware?? Feb 21 19:33:45 how cool is that? Feb 21 19:33:46 ;D Feb 21 19:33:52 awsome Feb 21 19:34:05 |nschle85|, lol, np. :P Feb 21 19:34:06 ciao. Feb 21 19:34:09 now we can do monitoring, injection, filtering Feb 21 19:34:32 leviathan, you won't be able to operate it yet... Feb 21 19:34:33 hack the gsm station, switches, controllers Feb 21 19:34:42 hmm Feb 21 19:35:02 leviathan, http://bb.osmocom.org/trac/wiki/LegalAspects#Usingmodifiedphonesoncarriernetworks Feb 21 19:54:37 leviathan, did you fix atr btw? Feb 21 19:54:43 inherit gettext works? Feb 21 19:54:44 uhm Feb 21 19:54:50 leviathan: yeah, great ticket for a looong free holiday in a gvmnt facility, before you go payback the fee for rest of your life Feb 21 19:54:56 I needed to install gettext on my workstation Feb 21 19:54:59 emerge gettext Feb 21 19:55:02 then it worked Feb 21 19:55:14 ok Feb 21 19:55:38 DocScrutinizer: what do you mean? Feb 21 19:55:41 ohh Feb 21 19:55:53 leviathan: pointing with a laser to landing airplanes is much easier though Feb 21 19:55:54 this one with injection? xD Feb 21 19:55:55 DocScrutinizer, he doesn't seem to care much about legality....(hosting non-redistributable firmwares,illegal docs etc...) Feb 21 19:56:16 in german theres a sentense Feb 21 19:56:22 legal, illegal, scheiss egal Feb 21 19:56:27 xD Feb 21 19:57:29 ok Feb 21 19:57:38 scheiss egal doesnt exist in english Feb 21 19:58:00 means such as "it does absolutly not matter" :) Feb 21 19:58:21 like the citate in hackers, take down of thomas shimamura Feb 21 19:58:21 also doesn't exist in vocabulary of any random athorney Feb 21 19:58:27 xD Feb 21 19:58:52 anyway, I really don't care about such things Feb 21 19:59:10 not much anyway Feb 21 19:59:31 leviathan, if you play too much at gta(grand teft auto)02 you'll get caught... Feb 21 19:59:31 the main thing is, that I know, how it works Feb 21 19:59:51 :) Feb 21 20:01:15 Total lack of respect for the law: http://www.youtube.com/watch?v=HLzGj10fg2g Feb 21 20:01:52 GNUtoo: you can now do: opkg update && opkg upgrade Feb 21 20:01:56 ok thanks Feb 21 20:02:02 youre welcome ;) Feb 21 20:03:12 PaulFertser: cool ;D Feb 21 20:04:14 leviathan: do you cycle? Feb 21 20:04:32 hmm Feb 21 20:04:34 nope Feb 21 20:04:40 <|nschle85|> hello, have problem, clicking next button in shr-wizard (Language screen) have ssh connection to phone who can help me ? Feb 21 20:04:56 uhm Feb 21 20:05:02 which device? ;D Feb 21 20:05:10 we have encountered this issue on dream Feb 21 20:05:11 <|nschle85|> om gata02 Feb 21 20:05:14 hmm Feb 21 20:05:24 cd /home/root Feb 21 20:05:26 rm -rf .e Feb 21 20:05:29 <|nschle85|> i hav ethis issure since several days Feb 21 20:05:33 hmm Feb 21 20:05:46 you have now a serious problem Feb 21 20:05:49 BECAUSE Feb 21 20:05:52 leviathan: i do cycle but not as furiously, either lacking courage or have too much smartness. Feb 21 20:05:55 you dont have a trackball Feb 21 20:06:16 leviathan, maybe you want free phones for making criminals communicate with openssl + camera....like in movies or similar things Feb 21 20:06:27 PaulFertser: I got a ticket, because I overtapt the 30limit with inlineskates >_< Feb 21 20:06:38 <|nschle85|> really rm -rf .e ? Feb 21 20:06:42 hmm Feb 21 20:06:50 not Feb 21 20:06:57 if you have it in practical use Feb 21 20:06:59 ... mom Feb 21 20:07:12 http://www.pastebin.com/m5970ec44 Feb 21 20:07:18 try this one for xorg.conf Feb 21 20:07:19 ... Feb 21 20:07:26 leviathan: oh, really? Ticket? Did they measure the speed with some equipment? Feb 21 20:07:33 jup Feb 21 20:07:38 32.4 Feb 21 20:07:39 >_< Feb 21 20:07:47 my colluege had 33.2 Feb 21 20:08:14 you see me there, without any savities and a wine bottle in the hand Feb 21 20:08:18 ^_^ Feb 21 20:08:23 leviathan: is it legal to go on skates on the roads at all? Feb 21 20:08:30 nope Feb 21 20:08:34 ok Feb 21 20:08:42 it was 3 o'clock in the morning Feb 21 20:08:45 it was dark Feb 21 20:08:51 and no one on the street Feb 21 20:09:26 Uber-cool Feb 21 20:10:17 <|nschle85|> this xorg.conf is not for OM-GA02 Feb 21 20:10:30 DocScrutinizer, I've also a good video about breaking the law while driving: http://www.youtube.com/watch?v=PK6jVxd_o14 Feb 21 20:10:43 leviathan: do they fine cyclists often? Because 30 is damn easy to break on a bicycle. BTW, in russia it's not punishable if you do not have more than 10km/h over limit i guess. Feb 21 20:11:46 s/DocScrutinizer/PaulFertser Feb 21 20:12:13 sorry Feb 21 20:12:45 <|nschle85|> (EE) PreInit returned NULL for "Neo1973 Buttons" Feb 21 20:12:45 <|nschle85|> (EE) config/hal: NewInputDeviceRequest failed (8) Feb 21 20:12:45 <|nschle85|> (II) config/hal: Adding input device s3c2410 TouchScreen Feb 21 20:12:45 <|nschle85|> (II) LoadModule: "tslib" Feb 21 20:12:45 <|nschle85|> (II) Loading /usr/lib/xorg/modules/input/tslib_drv.so Feb 21 20:12:46 <|nschle85|> (II) Module tslib: vendor="X.Org Foundation" Feb 21 20:12:46 <|nschle85|> compiled for 1.7.4, module version = 0.0.1 Feb 21 20:12:47 <|nschle85|> Module class: X.Org XInput Driver Feb 21 20:12:47 <|nschle85|> ABI class: X.Org XInput driver, version 7.0 Feb 21 20:12:48 <|nschle85|> (**) s3c2410 TouchScreen: always reports core events Feb 21 20:12:48 <|nschle85|> (II) XINPUT: Adding extended input device "s3c2410 TouchScreen" (type: TOUCHSCREEN) Feb 21 20:12:49 <|nschle85|> xf86TslibControlProc Feb 21 20:12:49 <|nschle85|> xf86TslibControlProc Feb 21 20:12:50 <|nschle85|> (II) config/hal: Adding input device PCF50633 PMU events Feb 21 20:13:43 GNUtoo: what's special about it? :( Feb 21 20:13:49 |nschle85|: pastebin! Feb 21 20:13:52 the next time Feb 21 20:14:08 PaulFertser, he drives while playing with his touchbook....it's dangerous Feb 21 20:14:22 but I'm not shure it's illegal Feb 21 20:14:26 GNUtoo: not nearly as the stuff i showed :) Feb 21 20:14:30 indeed Feb 21 20:14:45 but it uses a touchbook... Feb 21 20:15:25 In fact when you're on a bicycle yourself, it doesn't feel that dangerous, as you see everything around, you have control over situation, etc. But those messengers are really playing with fire. Feb 21 20:18:00 TAsn: yeah, I'm using the new opimd Feb 21 20:18:25 Mad snowy these days, had troubles going only 3.5 km, funny thing that it's harder to move on the roads than on the pavement. Feb 21 20:22:49 PaulFertser: weather staying the same crazy? some thawing here in cz... will be going to Moscow next week... shell i dress up again? :) Feb 21 20:23:46 PaulFertser: kinda anticyclic winter this year, no direct proof of climate global warming. I'm fed up with it, honestly Feb 21 20:25:01 * mickey|sofa hates snow Feb 21 20:25:14 PaulFertser: Nope, I wasnt on byke Feb 21 20:25:20 I was on inline skates ;) Feb 21 20:26:01 vanous123: they say it'll be ~0C next week Feb 21 20:26:25 leviathan: i understood but i wonder whether they usually fine bikers on that road. Feb 21 20:26:25 PaulFertser: ok. will be continuing to Каза́нь though... Feb 21 20:26:39 vanous123: :) Feb 21 20:26:43 PaulFertser: not 3 oclock in the morning Feb 21 20:26:56 * PaulFertser can't on earth understand how the hell to ride a fixie in city traffic Feb 21 20:27:05 xD Feb 21 20:27:14 makes fun Feb 21 20:27:26 in barcelone I used cars for making moves Feb 21 20:27:39 hijacked a car to get speed and so on xD Feb 21 20:28:04 on skated of course xD Feb 21 20:28:16 PaulFertser: uff, -23, feels like -35 now... good i looked it up now, would freeze otherwise... Feb 21 20:28:27 vanous123: -23 in Kazan? :-O Feb 21 20:28:35 PaulFertser: unfortunately... Feb 21 20:43:50 hi there. I'm using an up-to-date shr teting. When I try to delete the last sms it disappears from the message list but when I restart the phone it is still there. I have try to do it over a dozen of times and it always return. Can you help me with this?, pelase Feb 21 20:45:01 GNUtoo: the kernel image size is to big to flash again >_< Feb 21 20:45:14 leviathan, ah? Feb 21 20:45:49 leviathan, mkbootimage compresses it Feb 21 20:46:58 stillt to big :) Feb 21 20:47:30 but opie on dream is hilarious Feb 21 20:47:40 it fits the screen Feb 21 20:47:44 click sound works Feb 21 20:47:53 only problem LEDs and so on Feb 21 20:47:57 thats working on fso Feb 21 20:48:06 but its a proof of concept Feb 21 20:48:23 our kernel is now suitable for multiple different desktops Feb 21 20:48:25 :D Feb 21 20:49:57 now lets finish it and make it stable :) Feb 21 20:50:48 freesmartphone.org: 03mickey 07cornucopia * r0b31a2deab1c 10/libfsoframework/fsoframework/ (plugin.vala subsystem.vala): libfsoframework: more string templates Feb 21 20:52:15 waaahrg, i cannot receive sms anain thanks to my carrier.. Feb 21 20:52:38 ~lart T-Mobile Austria Feb 21 20:52:38 * apt squishes T-Mobile Austria like a bug Feb 21 20:54:52 freesmartphone.org: 03mickey 07cornucopia * r2f45582c2171 10/libfsoframework/fsoframework/ (plugin.vala subsystem.vala): libfsoframework: don't try to unload a plugin that has never be loaded Feb 21 21:03:00 mrmoku, does it work? Feb 21 21:09:39 ok Feb 21 21:09:48 need to sleep now Feb 21 21:09:58 tomorrow I need to get up around 6°° Feb 21 21:10:13 and the day after tomorrow I need to go to military 2d Feb 21 21:10:15 -.- Feb 21 21:10:28 tell them, that I dont wanna have to do anything with them Feb 21 21:11:30 I do not take orders from creatures which are under me in the nutritions chain Feb 21 21:11:41 so Feb 21 21:11:44 now I lie down Feb 21 21:11:47 try to sleep Feb 21 21:11:59 to be fit for tomorrows analysis 2 lecture Feb 21 21:15:04 TAsn, could you send me your local build of libphone-ui-shr? (i dont care for the field editing bug) Feb 21 21:15:17 deletion is more important atm... Feb 21 21:15:41 btw recreating the db helped with my name-to-number resolving bug Feb 21 21:16:31 told you :P Feb 21 21:16:35 sending, sec. Feb 21 21:17:16 thx Feb 21 21:17:27 sent Feb 21 21:21:09 GNUtoo: I guess I'll sleep now Feb 21 21:21:10 gn8 Feb 21 21:21:57 leviathan, ok bye Feb 21 21:22:04 freesmartphone.org: 03mickey 07cornucopia * rb8db72b8cb73 10/fsodeviced/ (12 files in 3 dirs): fsodeviced: ship alsa state files for openmoko_gta family of devices Feb 21 21:41:24 Q: SIM PIN known and working, PUK unknown. Any way to recover PUK? Feb 21 21:42:27 freesmartphone.org: 03mickey 07cornucopia * r4b3edba8ea4e 10/libfsobasics/fsobasics/utilities.vala: libfsobasics: fix prefixForExecutable for sbin Feb 21 21:42:28 JaMa, ping Feb 21 21:46:18 pong Feb 21 21:47:04 JaMa, thing is: some of the neo theme still suggest libframeworkd-phonegu-efl theme... kick this out. Feb 21 21:48:29 also the lates commit for libphoneui-shr-theme neo includes an idle screen theme -> http://scap.linuxtogo.org/files/4888e5082bc37dc7e0dae004a8b068fd.png the files are ina subdirectory in libphone-ui-shr-theme neo and the resulting edj shoudl be placen in the same directory as the libphoneui-shr theme Feb 21 21:48:47 if you have the time... personally i'm in no hurry Feb 21 21:49:38 also the latest rev of phoneui theme fixes sim auth Feb 21 21:52:39 JesusMcCloud: thanks, probably later, today I'm somehow deadly tired.. Feb 21 21:52:57 JaMa, as i said i'm in no hurry Feb 21 21:53:08 JaMa, one last thing: Feb 21 21:53:55 the idle_screen has no proper theming mechanism, so the default edj (idle_screen.edj but not the idl_screen_wallpaper.png) is just overwritten Feb 21 21:55:56 hmm that would need u-a, aw overwritting is really bad habit :) Feb 21 21:57:33 JesusMcCloud: is there a metapackage that installs all components of the (neo) theme? Feb 21 21:57:57 vanous123: shr-theme-neo Feb 21 21:58:02 i see Feb 21 21:58:16 not really all as some are deprecated already Feb 21 21:58:29 like that libframeworkd-phonegu-efl-theme-neo Feb 21 21:58:38 and IIRC etk-theme-neo Feb 21 21:59:01 JaMa: ok, i am watching it pulling it now Feb 21 22:00:14 JaMa, do you want me to look into the recipe dependencies of the metapackage and into the SUGGESTS of the theme packages? Feb 21 22:01:01 vanous123, libphoneui-shr as abug regarding theming. ping me when you have installed it, there is a dirty workaround Feb 21 22:02:51 vanous123, also the curretn build of the libphoneui theme has issues (already fixed in git), i'll give you my local build when you have installed Feb 21 22:02:59 JesusMcCloud: no need to look into it.. I moved it from RRECOMMENDS to RSUGGESTS intentionaly.. but if you really want to remove all libphoneui-shr-theme traces.. I can remove it Feb 21 22:03:23 JesusMcCloud: but as RSUGGESTS aren't installed by default I expected that having it there won't harm Feb 21 22:03:31 JaMa, i just want to remove etk and libframeworkd Feb 21 22:03:38 COMPLETELY Feb 21 22:03:48 JesusMcCloud: ok Feb 21 22:03:52 the rest is correct Feb 21 22:04:27 JesusMcCloud: from my POV: more important would be implement theming to that idle screen.. Feb 21 22:04:38 JaMa, is already in git Feb 21 22:04:52 maybe some update-alternative way to avoid overwriting Feb 21 22:04:57 JesusMcCloud: ping. what is the issue there? Feb 21 22:05:20 JesusMcCloud: I mean implement theme selection as works for other phoneui apps Feb 21 22:05:42 vanous123, sim auth, and keypad Feb 21 22:06:19 * JaMa off Feb 21 22:07:21 JesusMcCloud: have no pin. keypad... right, but not bothering me now so much Feb 21 22:08:07 JesusMcCloud: but i would love to see the today screen themed :) as the default slider never moves for me Feb 21 22:08:07 vanous123, i'll send you the fixed edj in a sec Feb 21 22:08:21 vanous123, i'll send you idle screen aswell in a sec Feb 21 22:08:28 cool Feb 21 22:11:38 vanous123, http://jmccloud.jm.funpic.de/neo/phoneui/ <- copy the files int /usr/share/libphone-ui-shr Feb 21 22:12:48 vanous123, then killall phoneuid Feb 21 22:15:13 JesusMcCloud: worked flawlessly :) thank you Feb 21 22:15:40 nP Feb 21 22:16:22 JesusMcCloud: the keypad's #5 is white, is that for a reason? Feb 21 22:16:34 vanous123, yup Feb 21 22:16:59 to help orienting in the screen? Feb 21 22:17:26 kinda...like on a physical keypad where #5 has a little knob Feb 21 22:18:01 but i have to fiddle with keypad anyways... Feb 21 22:18:27 slider on today is better now, though still not as fast as without any theming long ago... but better, as i say Feb 21 22:19:00 vanous123, will look at that aswell (but i dunno if twaeking edj can do anything) Feb 21 22:20:34 plus i should've been working on sth for university since the beginnign of februaray ;-) Feb 21 22:20:35 i would doubt that... once is themed it seems to have it's way... Feb 21 22:22:01 vanous123, edje is very powerful and the idle screen uses scripts, so there might be a chance of optimization Feb 21 22:22:27 vanous123, you coudl also try to kick out the backgroudn and see if it makes a difference... Feb 21 22:24:16 where is the Todays config now? Feb 21 22:24:48 vanous123, i hardcoded the backgroudn iamge into the edj Feb 21 22:24:53 i see Feb 21 22:25:01 but you can easily kick it out Feb 21 22:25:04 used to be configurable, but this was with python Feb 21 22:26:57 JesusMcCloud: this is alright for now, thank you Feb 21 22:27:14 have corrected themes installing in the shr manual now too Feb 21 22:27:34 vanous123, thanks Feb 21 22:27:45 will head for the bedroom, have monday again tomorrow :( Feb 21 22:28:06 good night Feb 21 22:28:24 likewise Feb 21 22:30:30 hi there... Feb 21 22:30:45 is it possible to turn off the pin of my simcard with some kind of dbus message? Feb 21 22:31:00 since zhone doesn't support entering the pin, so i'm kinda stuck without gsm here Feb 21 22:33:08 farhaven: i think zhone supported entering pin Feb 21 22:33:15 hmm Feb 21 22:33:20 then something's off here Feb 21 22:33:29 because it doesn't ask for a pin on a fresh debian install Feb 21 22:33:36 farhaven: and you can use mickeyterm to communicate with the modem directly to try to disable the pin Feb 21 22:33:44 how would I do that? Feb 21 22:33:49 farhaven: but i do not remember the commands for that Feb 21 22:33:50 is there some kind of guideline i could read? Feb 21 22:34:08 farhaven: i guess your problem is something else, not the pin. Try to look at frameworkd log Feb 21 22:34:26 where is it supposed to be? i don't see it in /var/log Feb 21 22:34:46 it is running though Feb 21 22:34:47 farhaven: depends on /etc/frameworkd.conf Feb 21 22:36:50 ah, that's empty Feb 21 22:36:55 i guess that clarifies a few things :D Feb 21 22:37:30 farhaven: yeah, you should install some mach-specific package Feb 21 22:37:40 farhaven: and /j #openmoko-debian Feb 21 22:37:52 will do, thx Feb 21 23:05:59 freesmartphone.org: 03mickey 07cornucopia * rbf3970cebbaa 10/libfsobasics/fsobasics/utilities.vala: libfsobasics: add Utility.machineConfigurationDir() Feb 21 23:31:10 how correct is the toolchain in om wiki? Feb 21 23:31:51 freesmartphone.org: 03mickey 07cornucopia * r53881ae4e141 10/libfsobasics/fsobasics/ (inotifier.vala kobjectnotifier.vala netlinknotifier.vala): fsoframework: inotifier: use FsoFramework logging instead of GLibs Feb 21 23:32:17 *toolchain wiki page Feb 21 23:33:50 freesmartphone.org: 03mickey 07cornucopia * r8d98414173c8 10/ (3 files in 3 dirs): fsodeviced: use new device specific alsa data file path in /etc/freesmartphone/default/conf//... Feb 22 01:48:04 I've got a question regarding opkg, am I in the right place? Feb 22 02:06:51 basically yes Feb 22 02:07:10 right place, maybe wrong time Feb 22 02:08:30 oh, i see. too late for most people? Feb 22 02:09:02 well, i suppose it can't hurt to ask Feb 22 02:09:33 I've built Angstrom using openembedded, and now I'm trying to add some packages to the system Feb 22 02:10:03 I've built those also using OE, but when I try to install an app from a .ipk file, opkg complains about an md5sum mismatch Feb 22 02:10:50 I'm guessing it's because opkg expects those apps to be provided by the repositories, and my version had a different md5 but provides the same application Feb 22 02:11:14 so is there a way to tell opkg to ignore the md5sums? Feb 22 02:22:20 checked opkg --help ? Feb 22 02:22:58 yeah, everything seems to be about dependencies and the like Feb 22 02:23:32 right now i'm getting by by extracting data.tar.gz from the .ipk file Feb 22 02:25:04 hmm, I seem to remeber some similar issue where it's been adviced to remove the original repositories from the package manager settings, then do an update to clear the database, then install, and finally add the repositories to put everything back to normal. Dunno if that applies here Feb 22 02:26:00 ahh, that makes a lot of sense Feb 22 02:26:33 i wasn't sure if that would make opkg 'forget' about old packages, but that sounds like it would solve my problem Feb 22 02:30:38 yup, that solved it. thanks so much! Feb 22 02:32:59 yw **** ENDING LOGGING AT Mon Feb 22 02:59:58 2010