**** BEGIN LOGGING AT Tue Jul 26 02:59:56 2011 Jul 26 03:00:12 yes cehteh, just like 2.5 inch flash SSDs, performance usually degrades., controller dependant. Jul 26 03:00:22 good SSD's to idle time maintenance to regain performance .. but i doubt SD card firmware does that Jul 26 03:00:41 Kingston cards are so braindead they always run in slow mode :-) Jul 26 03:01:02 my laptops SSD is very consistent in speed even without TRIM because dmcrypt Jul 26 03:01:04 yes they are more primitive controllers in the SD, but do they also use some overprovisioning? Jul 26 03:02:58 Yeah, because they don't actually test the flash before or after they put it in a *SD card. Jul 26 03:04:15 Ah, so , for ability to reject bad sectors yet still work, thus increasing yield from production. I was thinking more for moving stuff around to minimise erases. Jul 26 03:04:27 It never moves stuff around Jul 26 03:04:31 Not quite. Jul 26 03:04:43 There's basically no wear leveling of static data Jul 26 03:04:43 There are always bad sectors in most NANDs Jul 26 03:04:49 there are expectedd to be Jul 26 03:05:04 they have ecc for that Jul 26 03:05:20 same for spinning platters .. errors are expected Jul 26 03:05:27 But when the capacity is full, I think it should be doing intelligent writing and erasing Jul 26 03:05:40 Primes, it doesn't know about capacity Jul 26 03:05:41 Indeed - the ECC isn't quite for that - it's picking up the errored sectors before they become problematic. Jul 26 03:05:46 So they can reallocate. Jul 26 03:05:52 they dont do write verification but they have enough error correction information to cover almost all expected errors Jul 26 03:06:38 nowadays the read signals (both, ssd and hdd) are quite weak Jul 26 03:06:55 All cards I've had that I've worn out have had nasty failure behaviours. One silently corrupted data. One just hangs indefinitely when touching the bad sector. One does the same, except the bad sector randomly moves around (wear leveling the dead sector) :-) Jul 26 03:07:17 for hdds its the magnetic impulse which must be over the noise floor for ssds and MLC its the resistance encoding the data Jul 26 03:07:48 And yet writing to other areas of it works full speed without problems, so it's not doing bad sector reallocation or wear leveling over the entire area Jul 26 03:08:00 They do it over regions Jul 26 03:08:01 technology is maxed out there is no marigin for error on the hardware level, that all needs to be compensated by clever software and error correction bits Jul 26 03:08:07 Of a thousand blocks or so Jul 26 03:08:57 cehteh, error correction is more overhead, detection can be done by some CRC on a large size area, if it mismatches, just invalidate the whole block I guess. Jul 26 03:09:04 At least according to stuff I've read. Jul 26 03:09:12 Not quite. Jul 26 03:09:19 Every block written has ECC along with it. Jul 26 03:09:38 If, on read, the ECC nearly fails to correct the block, then that block is marked bad Jul 26 03:09:38 Primes: when you expect errors to be seen at read time and dont do write verify then you rely on ECC Jul 26 03:09:47 otherwise the storage would be leaky .. Jul 26 03:10:11 and this is usually implemented in some hardwired circruits at least partially Jul 26 03:10:22 In any case, cards don't do any fancy wear leveling or bad block replacements. They're quite primitive, which only makes SpeedEvil's want for raw mode the more relevant, you could use the powerful host processor for doing something more clever, which would not only boost longevity, but also boost speed by a few magnitudes :/ Jul 26 03:10:24 It's expected there is some error rate - the key is that the ECC can 'always' catch it before it goes bad irrecovera bly Jul 26 03:10:43 cehteh, but ECC can only correct some limited errors, not many bit errors, so it's possible you still could lose data Jul 26 03:11:05 Primes: The ECC is designed so it has enough depth for worst-case degradation of the memory Jul 26 03:11:10 ShadowJK: well yes .. i sometimes wonder about all this wear leveling .. why dont you use blocks sequentially until they are worn and then replace them Jul 26 03:11:15 In principle Jul 26 03:11:35 cehteh: The hard part is remembering which block is which without doing a media-wide scan. Jul 26 03:11:40 cehteh, you would if you had the CPU and RAM :-) Jul 26 03:11:47 in some implementation, single bit error detection AND correction, it multibit only detect errors can't correct. like hamming coding used for teletext bit recovery Jul 26 03:11:51 Hence they 'zone' the disk into thousand block or so. Jul 26 03:12:17 Primes: The ECC can correct and detect multibit errors. Jul 26 03:12:28 Primes: advanced ECC is quite smart and people building this devices know what errors they expect and tie the algorithms to that Jul 26 03:12:53 cehteh, because the overwriting to those sectors will not be the same throughout. Hence a good controller may choose to move stuff around Jul 26 03:12:55 for example when you know that a worn out cell sticks to some bit or so Jul 26 03:13:05 cehteh, iirc I once calculated that for a sequential log structured filesystem like that, the intel M25 128 or 256 (I forget which) would use all or most of it's declared 64MByte ram "cache" for the datastructures required to keep track of it all :-) Jul 26 03:13:32 maybe :) Jul 26 03:13:49 I guess it can speedevil, but when we say multibit, is it 2 or is it 20 bit error run? Jul 26 03:13:57 Lots Jul 26 03:14:23 I vaguely recall it's around 1% ECC bits Jul 26 03:15:18 caching reads is fine, caching writes is dangerous unless you have a good UPS, battery backed raid controller, ideally battery backed cache ram by power or supercapacitors Jul 26 03:15:19 good morning Jul 26 03:15:47 Yeah that's what I'm saying, they're not using that massive ram for caching Jul 26 03:15:56 Caching writes being dangerous only makes sense if you have applications that do the right thing. Jul 26 03:16:32 i wonder if at least for SSD some big capacitor would work to buffer power glitches and purge some emergency stuff to the flash Jul 26 03:16:55 Otherwise, you can't assume that you can simply stop the CPU, and the drive will contain a sane filesystem or files. Jul 26 03:17:01 cehteh: sure Jul 26 03:17:10 well at the beginning of this year i made a SSD odyssey .. trying serveral ones each one broken in different ways Jul 26 03:17:21 until i finally found a reliable (and extra expensive) one Jul 26 03:17:31 cehteh, yes commercial products do this using capacitor, supercapacitor. You're holding up the ram long enough to dump the content to some flash Jul 26 03:17:39 cehteh: a 100uF at 350V on a SPSU input that works down to 100V Jul 26 03:18:03 i was more thinking about a goldcap directly inside the ssd Jul 26 03:18:18 (to dump its own buffers) Jul 26 03:18:41 but prolly that isnt feasible Jul 26 03:18:50 cehteh: won't work as those circuits are designed to work with a certain VDD Jul 26 03:19:02 cehteh I have some of those gigabyte i-ram drives, they are DDR ram, with battery backup. But batteries need replacing, whereas supercapacitors don't ;-) Jul 26 03:19:28 yeah Jul 26 03:20:07 cehteh: you could charge a goldcap with a step-up converter, and on power glitch enable the LDR to provide 5V Jul 26 03:20:43 something like that .. dunno at what voltage flash runs internally maybe as low as 3.3V nowadays? Jul 26 03:21:17 1.8 or lower Jul 26 03:21:27 you can shift the voltage pretty easily Jul 26 03:21:48 on whatever voltage it runs, it will get exactly that voltage via the power supply, as you don't have regulators on those components Jul 26 03:22:20 acard make a nice ANS-9010, it's DDR2 based SSD which is backed by battery. If power fails the battery keeps the content while it copies content onto a compact flash card. Jul 26 03:22:36 so no matter how large your capacitor, it will drop below the regular voltage in no time Jul 26 03:22:58 * SpeedEvil holds up a 330F cap. Jul 26 03:23:14 LOL Jul 26 03:23:40 use charge pumping to reach and maintain the target voltage ;-) Jul 26 03:24:25 also consider parallel/series configurations. Some people didn't want single point of failure Jul 26 03:25:12 ok, allow 2% (assuming the PSU has a precision of +-3% while component tolerates +-5%), then you can calculate how long it takes to make that capacitor drop by those 2% at a given current drawn Jul 26 03:25:23 DocScrutinizer, battery voltage varies over time too Jul 26 03:25:47 Battery voltage is decoupled from component voltages by internal PSUs Jul 26 03:26:02 so, why not just take the capacitor output and regulate it to the target? Jul 26 03:27:11 that's exactly what I suggested Jul 26 03:27:46 It's funny how a 3.3 or 5v powered device, can derive +12 and -12v to drive RS232. Maybe you could take a similar approach lol Jul 26 03:28:39 max232 Jul 26 03:28:43 :-P Jul 26 03:28:56 lol and it's variants ;-) Jul 26 03:29:33 and the N900 can produce 5V on USB from its 3.7V battery :P Jul 26 03:29:59 IF you hack it for usb host mode ;-) Jul 26 03:29:59 what?! Jul 26 03:30:00 yes, with a step-up converter Jul 26 03:30:17 jiero: what what? Jul 26 03:30:34 DocScrutinizer N900 can produce 5V on USB from its 3.7V battery Jul 26 03:30:40 yes Jul 26 03:30:47 yes it can ;-) Jul 26 03:31:16 jiero: how else should h-e-n work Jul 26 03:31:41 but they wanted to fasttrack the approvals to get the device to market so it is officially usb slave mode Jul 26 03:31:56 I have no knowledge about those... Jul 26 03:32:00 errr somewhat, yes Jul 26 03:32:01 Better just watch Jul 26 03:32:08 ~hostmode Jul 26 03:32:09 rumour has it, hostmode is http://talk.maemo.org/showthread.php?t=65232 Jul 26 03:33:18 * DocScrutinizer burps Jul 26 03:33:23 ~xyawn Jul 26 03:33:24 from memory, xyawn is sleep Jul 26 03:33:28 hmm Jul 26 03:33:32 gnite Jul 26 03:33:47 ~botsnack Jul 26 03:33:48 thanks, DocScrutinizer Jul 26 03:43:05 I have a couple more questions about my maemo5. First , on the status item just right of the green battery, I have a bluetooth icon, and TWO level meters that might be for wifi, was it some wifi measuring app I installed? Jul 26 03:43:42 probably Jul 26 03:43:50 Secondly, is it possible to get cell tower ID and strength measurements from the phone subsystem, without r and d mode? Jul 26 03:43:51 Or cpu-mem applet Jul 26 03:44:08 Primes: you can't get them for unconnected towers Jul 26 03:44:15 with or without rd Jul 26 03:44:20 Otherwise - install 'netmon' for connected Jul 26 03:45:09 I mean when I turn on GPRS or 3G data, then I'm connected to a cell tower, can't I get some info on it, maybe using a command or app? My 3G wifi mifi dongle gives this in the web interface Jul 26 03:45:20 see netmon Jul 26 03:46:38 Netmon is an app? Jul 26 03:46:44 yes Jul 26 03:47:04 http://maemo.org/packages/view/netmon/ Jul 26 03:47:59 "Displays Signal Strength in dBm, Cell ID, LAC, RNC, Network Code, etc." yup, that's the stuff Jul 26 03:48:39 I worked for several telecomms companies, I thought a hacker phone should have such capability Jul 26 03:50:49 the cpu-mem applet I loaded is like a widget showing 3% and 136MB. That seems different to these bars to the right of the battery status. Are those there on yours or is it unusual because I installed something Jul 26 03:51:20 no Jul 26 03:51:22 disk space? Jul 26 03:51:37 Go to app-installer then uninstall Jul 26 03:51:39 to see what's there Jul 26 03:52:00 there are two of these things. Each one is a grey dot, with two white dots to the right of it. Sometimes it might draw some red also? Jul 26 03:52:35 Don't know - look to see what you've installed. Jul 26 03:52:39 AFK - sleep Jul 26 03:53:00 I looked hard, still working on it Jul 26 03:53:38 Load applet is the name Jul 26 03:54:34 comes with screenshot in system status menu Jul 26 03:57:12 I always thought load applet should come per default, as it's as important to understand your phone as is the motor temperature and RPM display in your car Jul 26 03:57:13 cpumem-applet uninstalled. one of those bar things is gone now, one remains Jul 26 03:57:34 that's what htop is for lol Jul 26 03:57:34 load applet - keep it! Jul 26 03:58:15 I like the one showing the little microchip, it's those little mini bars I don't like Jul 26 03:58:22 BS, whe your device becomes unresponsive, the you won't start htop to find out why - you can't Jul 26 03:58:32 Lots of cars lack rpm display :) Jul 26 03:58:57 if I already had it running, I don't need to start it Jul 26 03:59:33 I drove the nissan leaf electric car recently, it can be remote startup heater and diagnostics from a smartphone Jul 26 04:00:27 My biggest issue with load applet was that it didn't distinguish between CPU use and iowait. People stare at it and said "ZOMG i need to overclock!!" Jul 26 04:06:15 maybe it was fixed Jul 26 04:08:38 ok I got rid of the second tickbar thing. It was the load applet that can do screenshots. I like the screenshot capability but don't want those bars things. Jul 26 04:18:29 ctrl-shift-p takes screenshot Jul 26 04:19:07 and unlike harmattan, it works any time Jul 26 04:20:37 I mean how terribly must they have messed up things to get a formerly global shortcut neutered to only work in MTF apps Jul 26 04:22:11 how do you ctrl-shift-p on a virtual keyboard? .. hey how does one work with emacs on a virtual keyboard? :) Jul 26 04:22:23 well keyboard shortcuts aren't that relevant on harmattan ;p Jul 26 04:22:36 BS Jul 26 04:23:22 One could even argue that any function or feature that is only sanely accessible from hwkb is a bug in harmattan Jul 26 04:23:24 ok thanks for the tip, I wrote down CTRL shift p Jul 26 04:23:38 bullshit!!! Jul 26 04:23:58 Maybe call it a "feature" rather than a bug Jul 26 04:25:10 As there wont be a keyboard, requiring a keyboard is a bug ;p Jul 26 04:25:12 next you'll declare HID profile of BT a bug Jul 26 04:26:30 seriously, sometimes it pisses me off Jul 26 04:26:35 gnite Jul 26 04:27:15 The lack of keyboard pisses me off more, and then I'll really get pissed off if there's lots of shit that needs the nonexistant keyboard Jul 26 04:28:10 oh yeah, so better keep the shit but disable the kbd shortcut - get a brain!! Jul 26 04:28:28 And dragging a table, phonestand and bt keyboard with you isn't only very slow way of accessing hwkb, it's also not very practical :) Jul 26 04:28:44 "to access advanced screen capture features using obscure keyboard combinations, please attach your usb or bluetooth keyboard" Simples Jul 26 04:29:39 those laser projection bluetooth keyboards do look rather cool though Jul 26 04:29:56 * ShadowJK would make a bet that bt hid wont be supported and Aegis will make sure it'll never be supported ;p Jul 26 04:31:03 who can tell me where in the app manager is located nano text editor I don't seem to find it now Jul 26 04:31:20 I think you need to apt-get it Jul 26 04:31:44 bin app manager I'd just view the everything category and start typing nano Jul 26 04:32:27 well shadow, it appeared before, but would not install said it was missing some file, which I guess I need apt get to obtain that file, but I can't say what file because now I can't locate the app install?? Jul 26 04:32:44 uh Jul 26 04:33:21 Ah it is under GNU nano Jul 26 04:34:36 Problems: Application packages missing: libncursesw5 (>=5.7+20081213-1.maemo1+0m5) Jul 26 04:35:02 I like nano, it's easy to use, unlike vim Jul 26 04:35:26 and whereas leafpad looks a little too simple for me Jul 26 04:35:57 there is a emacs for the n900 :P Jul 26 04:36:09 but you have to remap some keys to make it useable Jul 26 04:36:16 I remember the days of IBM Xedit under VM/CMS Jul 26 04:36:53 I remember the days of programming using diode arrays. Jul 26 04:37:04 (it was 1988, I had lots of diodes) Jul 26 04:37:29 I had a machine with 1Kbyte ram, in 1981 Jul 26 04:37:38 zx81 here too Jul 26 04:38:08 Two zx81s, and a 16k rampack. And a schoolfriend who built a zx80 from kit form Jul 26 04:38:20 I think the closest I have come to development for anything funky would be the time I did some 65816 ASM Jul 26 04:38:49 * jonwil waits for people to ask what a 65816 is :P Jul 26 04:38:51 I did z80A, 6502, 6809, 8086, PIC assembler Jul 26 04:39:11 The 65816 was the 16-bit sequal to the 6502 Jul 26 04:39:11 What is this number of which you speak? Jul 26 04:39:44 It was used in, among other things, the Apple IIGS Jul 26 04:39:47 I can haz Arduino? Jul 26 04:39:51 and the Super Nintendo Jul 26 04:39:59 which is where my interest came in Jul 26 04:40:27 right , I have an old super nintendo, NTSC version Jul 26 04:40:54 the 65816 is the CPU in the SNES Jul 26 04:41:12 I bought a PS3, as a "cheap" cell development system (cheaper than the 20K Mercury). Unfortunately Sony removed the other linux OS option. Jul 26 04:41:58 There are hacks out there that let you bring OtherOS back Jul 26 04:42:41 not authorized by Sony obviously Jul 26 04:42:42 but usable Jul 26 04:43:35 yes, but I miss out on other features maybe like gaming , 3d blueray playback. The better FP version of cell is now used in some IBM bladecenter blade Jul 26 04:43:50 No you dont miss out on the other features Jul 26 04:43:58 You can install the hack and keep your games and bluray Jul 26 04:44:05 but now, with the help of FPGAs and opencores, we can just make our own cpus Jul 26 04:44:37 I think recent games try to update/validate the firmware, or gaming on the PSN may also Jul 26 04:44:41 I think the 65000 is the nicest architecture I have experienced with Jul 26 04:44:47 and ARM is the worst Jul 26 04:45:22 transputer: ahead of its time Jul 26 04:46:00 clearspeed accelerators, nice Jul 26 04:46:30 * jonwil wishes IBM had picked the 65000 instead of the 8086 for the PC Jul 26 04:46:41 hey even Xilinx making an FPGA combined with a cortex A9. Jul 26 04:47:36 just thinking of the savings of documentation textbooks. If IBM chose the Z80A, I could have just kept with my Rodney Zacks bible Jul 26 04:49:01 I'm in the local linux user group, my preferred distro is opensuse where we have YAST. But can I install libncursesw5 from apt get? Jul 26 04:51:18 I blame Geohot as well as Sony for the removal of OtherOS. He is not innocent, he should have kept to hacking iphones if he could not keep his mouth shut and think of the consequences Jul 26 04:54:09 it'd be installed automatically if available Jul 26 04:55:17 This should be PR1.3 like most. Installing GNU nano says "Unable to install GNU nano. Some application packages required for the installation are missing. Jul 26 04:55:47 So does anyone have nano installed and working? If the ncurses is missing, can I get it? Jul 26 04:56:11 I'm not Debian person (yet) Jul 26 04:57:09 Try enabling sdk tools repo Jul 26 04:58:08 And this is done in app manager: application catalogs: NEW ..... Jul 26 04:58:30 Primes: n900? Jul 26 04:58:43 yes n900 Jul 26 04:58:53 i have nano installed and no problems Jul 26 04:59:17 since some days i also have a chroot on the device with the sdk .. nice feel :P Jul 26 05:00:03 nano is secondary editor after emacs :P Jul 26 05:00:17 http://repository.maemo.org/ fremantle/sdk free non-free Jul 26 05:00:23 for now I only have nokia apps, nokia system software updates, ovi, maemo.org catalogs Jul 26 05:01:14 you do it in a chroot right? Jul 26 05:01:44 sdk on the main installation likely will break because the sdk isnt optified Jul 26 05:02:19 you're "loosing" me Jul 26 05:02:40 uh oh :P Jul 26 05:02:43 cehteh, he just wants to have nano Jul 26 05:02:48 ah ok Jul 26 05:02:49 afaik Jul 26 05:03:00 Later I will try all the sdk wizardry Jul 26 05:03:00 wherent there 2 nano packages (or more) Jul 26 05:03:55 they all seem to depend on libncurses in sdk tools? Jul 26 05:03:56 I only see "GNU nano" in the app manager, with a missing dependency. Yes I want to fix this to get nano. It's a learning exercise. Something else had something missing too, so if I can learn how to find and get, then I can overcome that too Jul 26 05:04:08 nano-opt - GNU nano - an enhanced clone of the Pico text editor. Jul 26 05:04:27 yes I like pico and nano on desktop linux Jul 26 05:04:30 Primes: you need extras enabled Jul 26 05:04:52 http://wiki.maemo.org/Extras Jul 26 05:05:15 (i think?... dunno i always have :P) Jul 26 05:05:27 there where/is an older broken nano package Jul 26 05:06:55 and, when adding the catalogue (repository) what should I enter for the distribution name? Jul 26 05:06:58 he wouldn't see nano at all in ham if he didn't have extras enabled Jul 26 05:07:23 ShadowJK: there are about 3 or more nano packages Jul 26 05:07:39 i dont know from which repo they all comme Jul 26 05:07:47 extras testing tools .. wtf :P Jul 26 05:07:57 all the ones i see have same dependencies Jul 26 05:07:59 i just can assert that i have a working nano here Jul 26 05:08:08 no, that's not it, EXTRAS is already as one of my 4 catalogs Jul 26 05:08:44 thankyou for your assertion cehteh, it gives me hope Jul 26 05:08:59 (working since 1 year or longer ...) Jul 26 05:09:16 in the beginning there was a nano which segfaulted when you used the search function Jul 26 05:09:38 Primes: maybe enable testing and tools too Jul 26 05:09:39 lol but the option to show line numbers is nice and useful Jul 26 05:10:10 cehteh, so you have more than 4 catalogs configured? Jul 26 05:10:17 deb http://repository.maemo.org/community-testing/ fremantle free non-free Jul 26 05:10:24 i am just checking .. only 4 Jul 26 05:10:37 i once had more (wtf, where did they go) Jul 26 05:10:51 community-testing? wtf is that? Jul 26 05:11:13 http://wiki.maemo.org/Repository Jul 26 05:11:32 ShadowJK: somehow i ask myself how that made it into that Jul 26 05:11:39 I think it goes into devel the raw stuff then approved into testing, then migrates into the production version Jul 26 05:11:44 also i dont know if fapman using those Jul 26 05:11:55 lemme check fapman Jul 26 05:12:36 yeah in fapman i have extras-testing Jul 26 05:12:45 and devel Jul 26 05:12:51 extras-devel Jul 26 05:13:22 it's not listed on the wikipage Jul 26 05:14:10 strange .. i dont know where that comes from .. i rarely use HAM Jul 26 05:14:53 but it exists :P Jul 26 05:15:24 isnt that the cssu repo? Jul 26 05:15:31 this seems relevant http://wiki.maemo.org/Documentation/devtools/maemo5 Jul 26 05:15:32 could be Jul 26 05:15:39 the installer installs it in ham Jul 26 05:15:46 ah yes Jul 26 05:16:08 fap has seperate repo list, it dont use apt's Jul 26 05:16:17 ham uses apt's Jul 26 05:16:36 where does fapman keep its configs? Jul 26 05:16:53 in /root/.fapman Jul 26 05:16:56 do you have devtools repository loaded? Looks like nice things like traceroute there for me Jul 26 05:17:04 ah Jul 26 05:17:09 thanks Jul 26 05:17:21 traceroute is in sdk repo Jul 26 05:17:45 i have added sdk repo in fapman for stuff like it and gawk Jul 26 05:18:04 looks like this devtools repo is a sanitized version of sdk tools? Jul 26 05:18:10 if it isnt the same Jul 26 05:18:25 its a subset i think Jul 26 05:18:35 not sure Jul 26 05:18:52 i aded sdk and sdk tools in fap Jul 26 05:19:57 be careful with sdk .. there are huge not optified packages Jul 26 05:20:02 on 2nd thot i think devtools is sdk tools Jul 26 05:20:35 ceteh yeah i know Jul 26 05:20:58 i only use it for small things like gawk, tracert etc Jul 26 05:20:59 thats why i saied, put it in a chroot .. Jul 26 05:21:12 i have 44mb free rootfs Jul 26 05:21:22 well i just installed ful blown stuff Jul 26 05:21:42 git self-build, ccache, linux .git checkout ... Jul 26 05:21:42 i manually optified and MyDocsified tons of stuff Jul 26 05:22:37 i had gcc and tcc running without chroot and 34mb free rootfs before last flash Jul 26 05:22:57 havent reinstalled it yet tho Jul 26 05:23:04 how is tcc on the n900 Jul 26 05:23:09 I will be planning on gcc sooner or later, just taking baby steps now Jul 26 05:23:17 i tried once on the main but devel libs where missing :P Jul 26 05:23:18 tcc linker sux Jul 26 05:23:32 ok Jul 26 05:23:33 There is only one sane solution to cross-compiling the kernel. Jul 26 05:23:36 http://bellard.org/tcc/tccboot.html Jul 26 05:23:43 haha Jul 26 05:23:54 if u write stuff with math functions or so u need to link with gnu linker Jul 26 05:23:59 ok Jul 26 05:24:05 well i have gcc anways now Jul 26 05:24:12 tcc linker doesnt handle the arm fpu stuff Jul 26 05:24:22 also i tried to compile luajit which worked with little efforts Jul 26 05:24:35 next time i make that ready to push it upstream Jul 26 05:25:20 "it can compile and run a typical Linux kernel in less than 15 seconds on a 2.4 GHz Pentium 4", was the kernel less verbose back in 2004? Jul 26 05:25:21 luajit would be awesome .. its so much faster than python Jul 26 05:25:46 huh Jul 26 05:25:54 if i dont get my linux box back soon and with it scratchbox ill reinstall gcc too Jul 26 05:26:02 i can remember kernel compiles in 90 seconds on my dual p2 back in time Jul 26 05:26:52 i can remember kernel 2.0.29 taking 10min on a P66 Jul 26 05:27:19 err P60 Jul 26 05:27:32 heh my first pentium was a 133 or so Jul 26 05:27:41 before i had an amiga Jul 26 05:27:45 that was my second Jul 26 05:28:23 i had a 486/33 later upgraded to dx4-100 before the p60 Jul 26 05:28:31 before that amiga Jul 26 05:28:46 before that c64 and c128 Jul 26 05:28:48 what a shock .. amiga with a 8mhz cpu worked more responsive than windows nt 3.51 on a pentium1 with 133Mhz Jul 26 05:29:32 cant even remember how much ram my first pentium had .. 32MB? .. no idea Jul 26 05:29:56 i had nt4 for a while on a p2 300 @ 450 Jul 26 05:30:28 i had one of those 450 relabled as 300 due to 300 shortages, lucky :) Jul 26 05:31:04 i think my p60 had 32 gb Jul 26 05:31:26 my 486 had 16 and much later 64 Jul 26 05:32:53 i used that 486 til 2006 Jul 26 05:33:10 ran linux ipcop from 2000-06 Jul 26 05:33:25 but then my line got too fast for it Jul 26 05:42:15 well I think I got the devtools repo setup in app manager. Now to have the traceroute app, I need to xterm, sudo gainroot, apt-get install traceroute ? ? Jul 26 05:48:43 yep Jul 26 05:49:06 or simply download it and dpkg -i it Jul 26 05:49:13 when I apt-get install traceroute, I get warning: the following packages cannot be authenticated! traceroute. Install these packages without verification Y/N ....... would this be a good idea? Jul 26 05:49:45 its no prob Jul 26 05:49:59 so it is normal to say yes? Jul 26 05:50:05 yea Jul 26 05:50:24 is it result of some key or codesigning? Jul 26 05:50:32 yea Jul 26 05:50:49 and this is not result of some version mismatch? Jul 26 05:50:55 if u check fapman logs u see it all the time in extras-dev Jul 26 05:51:21 its some broken ssl certificate Jul 26 05:51:21 I never ever heard of fapman I have no clue what it does. I used yast in suse, occasionally stuff from rpm Jul 26 05:51:35 faster application manager Jul 26 05:51:55 apt-get seems closer to my experience Jul 26 05:52:00 btw i dont recommend leaving devtools enabled in ham Jul 26 05:52:15 just enable it when u need it Jul 26 05:52:37 I know to temporarily enable, load uboot then disable when heading towards meego dualboot Jul 26 05:52:59 just uncheck the enabled thing in ham Jul 26 05:53:09 after u installed tracert Jul 26 05:53:16 need to go bbl Jul 26 05:53:54 I now have a working traceroute command. This rocks Jul 26 05:55:06 cssu camera app always says there's no free space when trying to make a shot. any workarounds/fixes? Jul 26 05:56:07 but there is? Jul 26 05:56:11 sure Jul 26 06:04:22 Thanks for your help casketizer, I disabled the new repo for now. Jul 26 06:04:42 nano is working, that was my objective Jul 26 07:57:53 Ooh, cool! http://www.dealextreme.com/p/i-smartsim-2008-sim-card-unlock-attachment-for-nokia-and-most-other-cell-phones-12167 Jul 26 08:07:08 N95 was listed as unsupported (no signal/cellular connectivity), I'm guessing N900 would probably do the same Jul 26 08:08:00 huh Jul 26 08:08:05 are there any simlocked n900's in existence? Okay, there will probably be plenty of N9's simlocked Jul 26 08:10:43 the size of this sim thing is 'mini' so it won't work with 'micro' e.g. N9 Jul 26 08:10:48 have a look at some of the user's/reviewers comments on the same page :) Jul 26 08:11:47 ah yeah Jul 26 08:11:56 so seems rather useless for anything maemo :) Jul 26 08:12:56 what? Jul 26 08:16:56 http://cordiatab.com ! Jul 26 08:23:34 seems like there's a fair few areas of that wiki (belonging to cordiatab) site needs to be fleshed out :p there's SIM slot but no specifications on bands it could operate, and 802.11b/g ? wtf... Jul 26 08:28:02 RST38h: yeah, smoku's baby :) Jul 26 08:28:29 smoku is good, very good =) Jul 26 09:17:48 Hi Jul 26 09:18:30 Does anyone know how to create a custom image for easy debian? Jul 26 09:24:10 nicofs, qole does Jul 26 09:24:34 lcuk, how can i talk to him...? Jul 26 09:27:30 nicofs, he is asleep most likely now Jul 26 09:27:35 (based in Canada) Jul 26 09:27:43 email I guess would work Jul 26 09:27:44 :) Jul 26 09:27:59 or start a thread Jul 26 09:28:02 and mail him the link Jul 26 09:29:18 lcuk, that thread thing would be best in case others want to do the same... Jul 26 09:29:24 hmm, anyone used ramz on power48? Jul 26 10:01:49 i've just been reading my logs here... Load Applet displays cpu and mem usage on the left of the camera things. if i click that icon, i get some kind of menu in the background, with Item 1 to Item 4. this doesn't seem to work. anything i'm doing wrong, or it works this way for everyone else? Jul 26 10:02:27 hmhm, was there any news about n9 during the last two weeks (was in vac) Jul 26 10:27:24 Corsac: no Jul 26 10:27:36 Sicelo: screenshot? Jul 26 10:30:13 just a sec Jul 26 10:33:48 http://www.reghardware.com/2011/07/26/mozilla_plans_open_source_mobile_os/ <-- and finally it's revealed... Jul 26 10:38:48 SpeedEvil: http://ssh.bombshellz.net/~qhubekela/load_applet.jpg Jul 26 10:58:09 ohai Pali Jul 26 10:58:21 hello Jul 26 10:58:35 Pali , how do I get compcache working in power48? Jul 26 10:58:56 I do not know - I did not tested it yet Jul 26 10:59:19 ah Jul 26 10:59:23 I only integrated kernel patch Jul 26 10:59:24