**** BEGIN LOGGING AT Tue Aug 07 02:59:59 2012 Aug 07 03:21:06 ...The Austrian Empire, supported by German and possibly Belgian and Romanian allies, launched a significant fleet of vessels into macrocosmic space, apparently powered by glucose harvested from dust falling onto SCP-1235. Aug 07 03:21:41 * RST38h can't begin to think what kind of controlled substances the author of this shit is on Aug 07 03:22:14 RST38h, makes for great reading, though, right? Aug 07 03:28:18 General: http://scp-wiki.wikidot.com/scp-1235 Aug 07 03:28:27 General: But for Tentacled's sake, EXHALE. Aug 07 03:29:03 ...The americium pellet in Examination Room 3's smoke alarm has also been seized, apparently in the belief that this mass will be fissionable via microcosmic explosives. Aug 07 03:32:04 -.- Aug 07 03:34:22 yawn Aug 07 03:36:00 it seems I ran into this SCP stuff before - probably guided by a dude with nick RST38h, some years ago Aug 07 03:36:48 Probably Aug 07 03:37:53 * DocScrutinizer05 yawns a bit morew Aug 07 03:37:57 -w Aug 07 04:09:42 I have a very dead battery and when it is in the n900 and connected to the charger (either wall or laptop usb) the phone tries and retries to boot and the battery does not charge. How can I stop the cycle of booting attempts and leave te battery to charge "quietly" ? Aug 07 06:10:25 nemo is n900's future? Aug 07 06:11:44 don't summon the troll Aug 07 06:14:19 slonopotamus, what? I don't understand, my english is poor... Aug 07 06:14:26 Do you mena nemo isn't the future? Aug 07 06:17:55 LaoLang_cool: i'm following maemo-related community for quite a long time already and what i saw in all forks/alternatives is constant (and repeatable) rebuild of lower level system packages and then phasing out when things touched gui. that happened to mer1, deblet, meego, tizen, several even less known things. next candidates are mer2/nemo/plasma active Aug 07 06:18:28 oh... Aug 07 06:18:52 well, maybe i'm wrong this time :) Aug 07 06:19:14 slonopotamus, thanks for info, I saw a news that nemo is actively under development these days Aug 07 06:19:21 So I get excited with it Aug 07 06:19:58 LaoLang_cool: and before that, meego was actively under development. and before that, maemo itself was Aug 07 06:21:20 meh, just look at mer homepage. no important news except initial project announcement. Aug 07 06:23:16 slonopotamus, oh... anyhow, I hope this time you're wrong ;p Aug 07 08:24:13 DocScrutinizer: ping Aug 07 08:31:59 DocScrutinizer05: ping Aug 07 08:58:11 kerio: pong Aug 07 09:05:06 DocScrutinizer05: stop mce; start mce caused mce to crash or something Aug 07 09:05:15 it was fun trying to restart it while blind Aug 07 09:05:55 i learned that you can kinda see something, with the backlight off Aug 07 09:07:14 anyway, i suggest not restarting mce at the end of the calibration Aug 07 09:09:29 that's very strange, since quite some situations require restart of mce and I never seen it crash on that (e.g. LED pattern editor) Aug 07 09:09:39 anyway thanks for the report Aug 07 09:10:21 locking the screen via dbus-send sounds like the best option Aug 07 09:28:22 DocScrutinizer05: is it possible to use the n900 without a battery? Aug 07 09:29:09 why? Aug 07 09:29:44 LaoLang_cool: in case my battery dies Aug 07 09:29:52 oh Aug 07 09:30:09 jon_y: use: yes. boot: no. Aug 07 09:30:11 dies as in dead, damaged, burned Aug 07 09:30:15 oh? Aug 07 09:30:43 so if it were on the wall charger it would continue if the battery was removed? Aug 07 09:31:06 if you do it right then yes. I published aq mini-howto on tmo Aug 07 09:31:38 basically you set screen to always on and stop bme Aug 07 09:32:04 ok, its like on life support Aug 07 09:32:16 and probably you want to go tabletmode or airplanemode Aug 07 09:32:37 what's wrong with the gsm? Aug 07 09:32:44 too much power Aug 07 09:32:46 drains more than the charger can provide? Aug 07 09:32:50 hmm ok Aug 07 09:32:51 yep Aug 07 09:33:00 DocScrutinizer05, what do you think about nemo? Aug 07 09:33:31 DocScrutinizer05: I read your article on flimsy usb charger ports, now I'm worried :( Aug 07 09:33:37 I got no dedicated notion about it Aug 07 09:34:30 ~tell jon_y about usbfix Aug 07 09:35:09 I know about those Aug 07 09:35:24 oh Aug 07 09:35:27 just that I'm not too confident on dissembly Aug 07 09:35:38 then just be careful Aug 07 09:36:58 is the usb port accessible after removing the 6 screws behind the cover? Aug 07 09:37:13 or are there more disassembly? Aug 07 09:37:47 e.g. in this particular case it's maybe a good idea to break the iron rule how to remove a plug, and pull out the USB (charger) plug by the *cable*. make sure you pull straight away from N900, not up, down, left or right Aug 07 09:38:30 I will remember that, can hardly find any n900 replacements Aug 07 09:39:12 jon_y: it's a bit more you need to do, particularly unplug that friggin flat plastic cable B2B connector that connects screen half with kbd half Aug 07 09:39:40 I saw that on youtube, doesn't look too simple Aug 07 09:40:19 yep, it's delicate Aug 07 09:41:02 with all the tiny connectors and all, and the impossibly hard to reach screen connector Aug 07 09:42:48 if you don't feel familiar and confident with such stuff, I strongly dicourage you to do the disassembly Aug 07 09:50:51 DocScrutinizer05: yeah, I gathered that much by watching the youtube video on n900 disassembly Aug 07 09:51:04 thanks for all your work though Aug 07 09:52:13 I don't have steady hands that can do BGA solderings with standard sized solders :| Aug 07 09:53:09 well, for soldering down the USB that's not exactly a needed skill Aug 07 09:53:43 unless it's already broken Aug 07 09:54:43 I could handle the soldering, just not the disassembly, my hands are too rough Aug 07 10:04:00 An autistic chimp could solder down the USB socket. Aug 07 10:04:48 vi_: but the FPC B2Bcon is a biatch Aug 07 10:05:10 DocScrutinizer05: the what now? Aug 07 10:05:20 The ribbon that connects the screen to the mobo? Aug 07 10:05:36 tha FlatPlasticCable Board2Board connector Aug 07 10:06:21 DocScrutinizer05: Fuck trying to solder that! Aug 07 10:06:25 that FPC itself also is a bit delicate, but the connector is really silly Aug 07 10:08:14 and the second layer PCB with bridge traces that they soldered on back side of the PCB+connector abomination Aug 07 10:09:35 I mean, soldering two PCB flat on flat is a silly thing in itself, not efen considering the FPC being flexible, and the plug/connector needing mechanical force to mate/unmate Aug 07 10:09:58 s/efe/eve/ Aug 07 10:09:59 DocScrutinizer05 meant: I mean, soldering two PCB flat on flat is a silly thing in itself, not even considering the FPC being flexible, and the plug/connector needing mechanical force to mate/unmate Aug 07 10:54:58 s/bga/surface mount/ Aug 07 11:04:55 i was idly wondering about that for multilayer boards. Aug 07 11:05:00 Solder board on board. Aug 07 11:05:12 Clearly there are mechanical issues. Aug 07 11:10:51 SpeedEvil: I would be more afraid of the parasite capacities Aug 07 11:17:09 There is nothing magical about vias. Aug 07 11:17:34 The capacitance would be if anything, less than a conventionally manufactured board. Aug 07 11:17:58 well, traditional boards are not that different to soldered Aug 07 11:18:02 But, there are severe mechanical issues, the board will be very fragile, as there are huge shear loads when you bend a board. Aug 07 11:18:16 it's a lot of double sided boards pressed together Aug 07 11:18:24 - it would be less due to the lower dielectric constant of air vs pcb Aug 07 11:18:39 And temperature changes will also cause load changes. Aug 07 11:18:45 stress changes Aug 07 11:33:45 DocScrutinizer05: cleven now has a file in /etc/sudoers.d/ Aug 07 11:33:48 ...that doesn't work Aug 07 11:33:59 DocScrutinizer05: LOL! Aug 07 11:34:09 why don't people *test* this stuff? :c Aug 07 11:34:10 well Aug 07 11:34:21 i mean, it's not hard Aug 07 11:34:27 i fixed it in my sudoers file, but still Aug 07 11:34:37 why do people have problems with such a simple stuff like sudoers.d/* Aug 07 11:35:35 user ALL = NOPASSWD: /opt/cleven/clevenHelper.sh *, /opt/cleven/load.sh *, /opt/cleven/unload.sh Aug 07 11:35:42 sudoers.d doesn't work like that! Aug 07 11:35:47 s/\.d// Aug 07 11:35:52 hm Aug 07 11:35:56 infobot: ping Aug 07 11:35:57 ~pong Aug 07 11:35:59 hm Aug 07 11:36:14 care to guess what my fix was? :) Aug 07 11:39:24 kerio: are /opt/cleven/clevenHelper.sh /opt/cleven/load.sh /opt/cleven/unload.sh at least not writable by anybody but root? Aug 07 11:39:50 yeah, 755 root:root Aug 07 11:40:00 as well as /opt/cleven ? Aug 07 11:40:07 yep Aug 07 11:40:11 good Aug 07 11:40:50 the modules are 644 root too Aug 07 11:42:13 on the other hand, passwords are saved in /etc/passwd, hashed with crypt() Aug 07 11:42:26 so yeah :s Aug 07 11:44:55 unrelated: is there a programmatic way to find out what's the size of a partition? Aug 07 11:45:37 kerio: ivstat or something, I can;t remember Aug 07 11:45:52 iostat? that's for statistics Aug 07 11:46:26 no, it gives the size Aug 07 11:46:43 you'll need to divide it by the cluster size or something Aug 07 11:46:51 I can't remember the exact function name Aug 07 11:47:59 sfdisk -s apparently :D Aug 07 11:50:33 kerio: http://linux.die.net/man/2/fstatfs64 Aug 07 11:50:41 I think I used that some time ago Aug 07 11:51:18 a bit overkill if you want to do it from a script Aug 07 11:51:28 yay, sfdisk is busybox Aug 07 11:51:29 so it's installed everywhere! :D Aug 07 11:52:24 hm, iostat is busybox too, but nobody owns that symlink for some reason Aug 07 11:52:28 maybe it's busybox-power? Aug 07 11:52:50 does busybox-power completely replace busybox? Aug 07 11:53:18 yeah Aug 07 11:53:26 but it does so with preinst/postinst scripts, of course Aug 07 11:53:46 if it doesn't break things, sure why not Aug 07 11:54:51 yeah, it's busybox-power Aug 07 12:00:36 hmm, why does reswap use weird commands instead of /proc/partitions and /proc/diskstats? Aug 07 12:03:54 oh balls, how do i add a literal $ in a busybox string? Aug 07 12:20:00 kerio: \$ <-? Aug 07 12:20:10 jon_y: yeah, i figured :) Aug 07 12:20:24 alright, my swapswap script has no dependencies other than busybox Aug 07 12:20:25 i think Aug 07 12:22:40 kerio: heh, nice name Aug 07 12:23:05 :D Aug 07 12:33:55 * DocScrutinizer05 swallows any comment about messybox-bowa Aug 07 12:34:24 except: Aug 07 12:34:28 ~messybox Aug 07 12:34:28 messy... err busybox is meant for lean scripting. Regarding all the missing options and immanent limitations (see su) it's not really the interactive shell of choice. A lot of people hate busybox because a lot of system integrators don't understand the difference between busybox and a decent user interactive shell plus unix utils Aug 07 12:37:17 hell, my fist linux system had 8MB RAM, 50MB HDD, and yet it had a proper bash and complete manpages and all unixtools Aug 07 12:37:27 first* Aug 07 12:37:56 heh Aug 07 12:38:24 at work, I get to deal with csh and all it's idiosyncrasies Aug 07 12:38:44 its ENTERPRISE grade they said Aug 07 12:38:47 (t)csh is a PITA in the long run Aug 07 12:38:50 at least busybox is consistent Aug 07 12:39:07 oh really? Aug 07 12:39:24 "enterprise" is possibly the worst reason to choose a shell Aug 07 12:39:45 kerio: I think its just pure momentum Aug 07 12:40:04 all the scripts are held over from big iron Unix, probably Aug 07 12:40:12 scripting in csh :( Aug 07 12:40:16 and then when Linux came along... Aug 07 12:41:01 also, CTRL+H for backspaces Aug 07 12:41:14 stty? Aug 07 12:41:16 I'm not sure how they got that crazy :) Aug 07 12:41:18 in work I followed instructions for building android, and it been written "tell your admin to install bash as default shell!". I did since my account came with tcsh initially, and answer been "OK, here you are. But don't complain, you're the only one with bash now" Aug 07 12:41:20 yeah Aug 07 12:41:47 lol, I get that too Aug 07 12:41:55 :D Aug 07 12:42:16 DocScrutinizer05: what breaks if you put bash in /bin/sh? Aug 07 12:42:23 tcl is somehow cool to them :| Aug 07 12:42:27 on N900? Aug 07 12:42:38 ancient tcl8.4 Aug 07 12:42:46 DocScrutinizer05: yeah Aug 07 12:43:00 possibly boot of user Aug 07 12:43:28 I just do bash -i as my first command :) Aug 07 12:43:45 not by using bash as default shell, but by replacing the messybox commands by proper fullsized unixtools Aug 07 12:44:00 jon_y: same here Aug 07 12:44:20 I tried that, dpkg got in the way Aug 07 12:44:34 and I fixed the *rc and profile* scripts so only bash gets bin/gnu in $PATH Aug 07 12:45:36 ~jrtools Aug 07 12:45:36 i heard jrtools is http://wiki.maemo.org/User:Joerg_rw/tools Aug 07 12:45:58 meh, bash is both user and root's default shell here Aug 07 12:45:59 and nothing broke Aug 07 12:46:08 i mean, scripts are still launched by busybox Aug 07 12:46:13 "#!/bin/sh" Aug 07 12:47:06 sure, since you neither linked /bin/sh to bash, not included bash to /etc/shells Aug 07 12:47:20 s/not/nor/ Aug 07 12:47:21 DocScrutinizer05 meant: sure, since you neither linked /bin/sh to bash, nor included bash to /etc/shells Aug 07 12:48:20 and I bet your ps is still useless Aug 07 12:48:34 tar is useless Aug 07 12:48:39 that too Aug 07 12:48:40 gtar is better Aug 07 12:48:48 DocScrutinizer05: nope, my ps is the real ps Aug 07 12:48:55 from procps Aug 07 12:49:10 :D Aug 07 12:49:25 gtar == bin/gnu/tar Aug 07 12:49:30 and bash is in /etc/shells Aug 07 12:49:46 or rather Aug 07 12:49:48 full coreutils pleas :) Aug 07 12:49:52 *please Aug 07 12:49:54 g* == bin/gnu/* Aug 07 12:49:56 jon_y: i have that too Aug 07 12:50:14 are those already optified? Aug 07 12:50:19 yep Aug 07 12:50:23 I mean busyboc symlinks are cheap Aug 07 12:50:23 except 4 Aug 07 12:50:28 nice Aug 07 12:50:33 4? Aug 07 12:50:54 rgrep, diff, diff3 and something else Aug 07 12:50:57 what's the idiomatic way of doing if elif else chains in shell scripting? Aug 07 12:51:15 a case ;-P Aug 07 12:51:39 nope Aug 07 12:51:43 i need two different conditions Aug 07 12:52:01 ?? Aug 07 12:52:09 ternary? Aug 07 12:52:35 if $SWAP1 is in /proc/swaps ... else if $SWAP2 is in /proc/swaps ... else Aug 07 12:52:49 ternary only works for numeric in shell Aug 07 12:53:26 I like my perl scripts more than shell scripts because of this :) Aug 07 12:54:10 the thing is, in real programs you're supposed to use the syscalls to get system data Aug 07 12:54:15 not parse the files in /proc Aug 07 12:55:34 yeah, I've seen programmers do that in C Aug 07 12:55:51 oh god why Aug 07 12:55:55 C string manip is horrible Aug 07 12:56:12 because they were script writers before introduced to C Aug 07 12:56:30 csh+tcl everyday will do that to you Aug 07 12:56:38 kerio: case "`cat \proc\swaps`" in "*${SWAP1}*") do-sth;; "*${SWAP2}*") do-sth-else;; *) echo "WTF?!";; esac; Aug 07 12:56:44 also, using C++ just to cat strings Aug 07 12:56:51 DocScrutinizer05: ooh, neat Aug 07 12:58:00 or you use if then elif else Aug 07 12:58:26 kerio: btw, that massive junk to parse text is now rewriten in Java :) Aug 07 12:58:31 jr@halley:~> help if Aug 07 12:58:33 if: if COMMANDS; then COMMANDS; [ elif COMMANDS; then COMMANDS; ]... [ else COMMANDS; ] fi Aug 07 12:58:38 sounds at home on the dailywtf Aug 07 12:59:11 java because regex and they can't handle pointers Aug 07 13:01:50 meh, i can afford two passes over /proc/swaps Aug 07 13:02:23 http://fpaste.org/dHyt/ fuck yeah awk Aug 07 13:03:06 yeah, rarely used tool Aug 07 13:03:15 grep and arithmetics Aug 07 13:04:17 you all make me feel so sad Aug 07 13:04:20 :c Aug 07 13:04:26 DocScrutinizer05: why? Aug 07 13:05:04 just because... and of java Aug 07 13:05:14 and C++ Aug 07 13:05:18 oh, that Aug 07 13:05:43 lol, never underestimate them ever Aug 07 13:06:01 nobody there has heard of diff -u Aug 07 13:06:07 haha Aug 07 13:06:11 they use context diffs Aug 07 13:06:30 also, rcs for version control Aug 07 13:06:58 btw there's absolutely nothing wring with fopen("/proc/foo/bar"), proc-ps does exactly that Aug 07 13:07:06 wrong* Aug 07 13:07:34 well, yeah, but java, reallly, to parse text?? Aug 07 13:07:40 re "supposed to use syscalls" Aug 07 13:07:49 DocScrutinizer05: really? :o Aug 07 13:08:06 guess why it's called procps Aug 07 13:08:33 and yes, even the devels say it's a bit heavy Aug 07 13:09:07 but then otoh, who cares, since you won't call ps in an endless tight loop Aug 07 13:09:40 heh, I've seen ssh called from a GUI thread Aug 07 13:09:53 all fine with me Aug 07 13:10:06 it paused the whole program while the remote command runs Aug 07 13:10:08 wtf is /core? Aug 07 13:10:29 I should get paid more to babysit the programmers :| Aug 07 13:10:39 there are thing I kill with one system() line calling a shell cmd, that otherwise cost me 2 pages of code Aug 07 13:10:52 like e.g. grep Aug 07 13:13:03 and the additional benefit is it's spawning a sub-process that terminates and frees all memory used by sh and grep, as soon as I got my result Aug 07 13:13:36 so tell me what's wrong with it Aug 07 13:13:55 apart from the waste of resources? Aug 07 13:13:57 >:C Aug 07 13:14:26 what friggin waste of resources? YOU are wasting resources reinventing the grep-whell in your code Aug 07 13:14:44 memory resources, and your own resources for coding it Aug 07 13:14:51 grep-whell? Aug 07 13:14:57 oh, wheel Aug 07 13:15:00 wheel Aug 07 13:16:27 it's more mem-heavy for a short while, but frees resources immediately after completing its task, while your code with builtin grep carries this grep code along all the time Aug 07 13:16:44 builtin grep = strcmp? Aug 07 13:16:53 well, strncmp Aug 07 13:16:54 go try that Aug 07 13:17:29 you'll soon find out about boost-regex Aug 07 13:17:40 why would i want to use c++? D: Aug 07 13:18:16 well, you can code your own regex as well Aug 07 13:18:53 or use (3)regex Aug 07 13:19:04 have fun! Aug 07 13:19:56 til you hit "q" on "man 3 regex" I already forgot that system() line I coded Aug 07 13:20:04 hehe Aug 07 13:20:19 plus it just works Aug 07 13:20:38 DocScrutinizer05: Well - yes and no - if the code for your grep is not used, it will be swapped out with enough memory pressure. Aug 07 13:20:47 (well - not swapped out, if it's executable backed) Aug 07 13:20:57 unrelated: can /sys/block/mmcblk*/queue/nr_requests be increased to 4096 or 8192 (as ShadowJK_ said) without too many consequences? Aug 07 13:21:10 SpeedEvil: ack Aug 07 13:21:41 SpeedEvil: probably it's large enough to occupy a whole swap page, or several Aug 07 13:22:04 grep probably is Aug 07 13:22:24 a manual strncmp? not really Aug 07 13:22:42 so what? when I need that grep only once on startup, to e.g. parse initfiles? Aug 07 13:23:13 kerio: and again, grep != strcmp Aug 07 13:23:31 grep == regex + read/write Aug 07 13:24:42 and fun starts when I ned sed instead of grep, and then watch your effort coding that Aug 07 13:24:47 need* Aug 07 13:24:58 * kerio laughs at silly people setting nr_requests to 4 for some weird reason Aug 07 13:25:49 oh, unrelated but inspired by... what's maxterm on fremantle now? Aug 07 13:26:11 iirc on diablo it been 4 or sth iinsane like that Aug 07 13:26:35 maxterm? Aug 07 13:26:50 max number of pty(?) Aug 07 13:28:28 ooh, found a problem - backupmenu can't give me a root console because there's no /bin/bash in backupmenu's jail Aug 07 13:28:34 how fun Aug 07 13:28:50 /proc/sys/kernel/pty/max I guess Aug 07 13:29:18 16, mhm Aug 07 13:29:29 * DocScrutinizer05 wonders if he already fixed that Aug 07 13:30:14 kerio: cat /proc/sys/kernel/pty/max Aug 07 13:30:17 please Aug 07 13:30:30 16 Aug 07 13:30:35 tnx Aug 07 13:30:37 and i haven't fixed it Aug 07 13:31:12 I gathered that Aug 07 13:31:43 since you'd probably have used 10000 when you fix it ;-P Aug 07 13:32:34 how's your reswap going along? Aug 07 13:33:00 DocScrutinizer05: have you "fixed" nr_requests? Aug 07 13:33:08 path? Aug 07 13:33:21 (prolly not, unless swapolube does) Aug 07 13:33:35 /sys/block/mmcblk*/queue/nr_requests Aug 07 13:33:51 4 4 Aug 07 13:33:55 ...wat Aug 07 13:33:56 :-/ Aug 07 13:34:11 cat /sys/block/mmcblk*/queue/nr_requests -> 4 4 Aug 07 13:34:12 ShadowJK_ advised me to use 8192, when doing heavy I/O Aug 07 13:34:42 yeah, I know Aug 07 13:34:50 is 8192 good for normal usage, i wonder? Aug 07 13:35:10 I think I once wrote a wrapper around cp to do that temporarily Aug 07 13:36:16 or not Aug 07 13:36:16 or at least something in between Aug 07 13:36:21 IroN900:~/bin# which cp Aug 07 13:36:23 /usr/bin/gnu/cp Aug 07 13:36:59 I can't recall the downsides it comes with Aug 07 13:37:09 for sure thare are any Aug 07 13:37:35 at very least regarding memory usage or sth Aug 07 13:38:29 waaaait, that's been sooo long ago, I have to check t900 Aug 07 13:39:35 i just don't get why the CPU is used, when copying files Aug 07 13:39:46 i mean, why isn't it I/O bound? Aug 07 13:41:40 DocScrutinizer05: i think that swappolube is the one that put nr_requests to 4 Aug 07 13:42:36 duh Aug 07 13:42:42 * LaoLang_coo_ is back to maemo... Aug 07 13:43:14 meh, 16384 total requests doesn't seem too much to me Aug 07 13:43:20 >:) Aug 07 13:45:50 kerio: PIO mode? Aug 07 13:46:07 copy to cpu cache, copy cache to device Aug 07 13:46:30 meh :c Aug 07 13:47:04 /etc/init.d/rcS: echo 4 >/sys/block/mmcblk0/queue/nr_requests || echo "ohnoes, hope it boots1" Aug 07 13:48:45 dafuq Aug 07 13:49:18 echo 16 > /proc/sys/kernel/pty/max Aug 07 13:49:51 and a *lot* of other nice stuff, inviting to mess with it Aug 07 13:50:27 i don't have anything regarding nr_requests Aug 07 13:50:33 but the other stuff is there Aug 07 13:51:04 i don't get it, why is a small nr_requests a good thing? Aug 07 13:51:17 random IO is really awful, and you gotta use RAM for something Aug 07 13:52:34 echo "/what/me/worry" > /proc/sys/kernel/core_pattern Aug 07 13:52:35 wtf is that Aug 07 13:53:05 oh, no core dumps Aug 07 13:55:05 yeah, git it in my pastebuffer to paste it here ;-) Aug 07 13:59:41 * DocScrutinizer05 idly wonders if there's a pkg bootchart Aug 07 14:01:55 where's friggin bootchart? Aug 07 14:02:46 ok, screw it, i'm setting both nr_requests to 2048 at boot time Aug 07 14:04:18 sounds about sane Aug 07 14:04:32 or maybe 8192? Aug 07 14:04:32 tell me your results Aug 07 14:04:36 ShadowJK_: *poke* Aug 07 14:05:22 i wonder if /etc/init.d/rcS is the right place Aug 07 14:05:27 but hey, lots of parameters are tweaked there Aug 07 14:05:31 obviously Aug 07 14:08:11 ok, put the echo calls right before the exit in /etc/init.d/rcS and they were applied Aug 07 14:08:41 i put them at 8192 because fuck waiting 20 minutes of IO at 200kb/s to copy ONE file Aug 07 14:09:15 I just wonder how /sbin/preinit and /etc/init.d/rcS relate Aug 07 14:09:23 preinit is before init Aug 07 14:09:40 /etc/init.d/rcS is ran by /etc/event.d/rcS, which is ran by upstart (init) Aug 07 14:09:51 mhm Aug 07 14:09:53 thanks Aug 07 14:10:07 at least, that's what i gathered by the comments in /etc/event.d/rcS Aug 07 14:10:28 fsck upstart Aug 07 14:13:56 case "$root" in 3) root_dev=sda1 root_dev_name="USB hard drive" DAFAQ?! Aug 07 14:14:33 too little too late, nokia :c Aug 07 14:14:47 DocScrutinizer06, preinit? Aug 07 14:14:57 :nod: Aug 07 14:15:06 did the n810 have usb otg? Aug 07 14:15:06 yeah, a lot of stuff for booting from somewhere other than onenand was stripped out Aug 07 14:16:00 if [ $serial_console -ne 0 ]; then Aug 07 14:16:01 echo_g "Press any key for interactive startup" Aug 07 14:17:48 default_root=$(cal-tool --get-root-device 2> /dev/null) Aug 07 14:21:15 hope I fsckd my work N900 with cal-tool --set-root-device "" now Aug 07 14:22:33 * DocScrutinizer06 reboots Aug 07 14:25:07 boots. Still I wonder how to recover from cal-tool --set-root-device mmc Aug 07 14:25:19 in case it refuses to boot then Aug 07 14:25:32 reflash won't help a messed up CAL Aug 07 14:26:33 only way to recover from such a thing would be to build something you can use with the "load this and run it" option for the flasher and would fix your CAL Aug 07 14:26:47 yep Aug 07 14:26:58 ~rescueos Aug 07 14:27:00 nin101's rescue-system Aug 07 14:27:07 stupid infobot Aug 07 14:27:22 ~listkeys rescue Aug 07 14:27:24 Factoid search of 'rescue' by key (8): rescuecd ;; htcrescue ;; rescue disks ;; cmd: rescue (.*?) ;; rescue-os ;; superrescue ;; reiserfs-rescue-disk ;; rescue disk. Aug 07 14:28:22 ~rescue kerio Aug 07 14:28:23 * infobot leaps into certain danger to rescue kerio and return kerio to safety! Aug 07 14:28:37 yay Aug 07 14:28:46 !seen robbiethe1st Aug 07 14:28:50 ~seen robbiethe1st Aug 07 14:28:50 ~rescue-os Aug 07 14:28:57