**** BEGIN LOGGING AT Sun Sep 02 02:59:59 2012 Sep 02 04:08:14 freemangordon: hey Sep 02 07:35:05 has any of you a diablo scrachbox toolchain handy ? Sep 02 11:34:25 DocScrutinizer05: you there? Sep 02 11:47:42 DocScrutinizer05: why does your calibration script start, stop and restart bme? Sep 02 11:47:53 can't you just start it once, when it's done? Sep 02 11:49:36 kerio: yep, I'm always here Sep 02 11:50:01 DocScrutinizer05: i know, i'm always here too but sometimes i'm asleep or away from anything ircful Sep 02 11:50:15 well, maybe cargo cult coding. It seemed to have healthy effects whenever I messed around with bme Sep 02 11:50:52 haha Sep 02 11:51:05 often bme recovers from totally insane battery capacity guestimate after second restart Sep 02 11:51:14 anyway, i'm going to rewrite your calibration script to not use bq27200.sh Sep 02 11:51:17 because there's no reason to Sep 02 11:51:30 sure Sep 02 11:51:32 it spams the i2c bus for no reason, considering that you only need vdq and edv1 Sep 02 11:51:53 not that this would matter at all in this context Sep 02 11:52:20 but sure, go ahead, it's sane and feasible Sep 02 11:52:59 I've just been lazy when writing this lil fanzy especially for your convenience, in 10 minutes Sep 02 11:53:25 oh, btw Sep 02 11:53:32 busybox does have a way to do bitwise AND Sep 02 11:53:57 you're sure that's not busybox-power? Sep 02 11:54:37 yep Sep 02 11:54:47 $(( (0x04 & 0x04) / 4 )) Sep 02 11:54:55 (NB bq27200.sh is ShadowJK's work, based on a bashscript called bq27k-detail I wrote) Sep 02 11:54:57 versus $(( (0x04 & 0x03) / 4 )) Sep 02 11:55:28 hm? Sep 02 11:55:37 one is 1, the other is 0 Sep 02 11:55:45 sure Sep 02 11:55:58 in BusyBox v1.10.2 (Debian 3:1.10.2.legal-1osso31+0cssu0) Sep 02 11:56:31 well, bq27k-details is using arrays, and those are missing in messybox Sep 02 11:57:45 I admit I might have copied the bit-dissecting code from shadowjk though Sep 02 11:58:42 so even while I'm aware of bit-ops in bash, the code to dissect flags might still be plain interger arithmetic Sep 02 11:59:01 * ShadowJK assumed sh didn't habe binary arithmetic and used integer arithmetic Sep 02 12:00:00 also, why do you all people use $( ) instead of ` `? Sep 02 12:00:18 simple answer? because you can nest it Sep 02 12:00:23 oic Sep 02 12:00:27 i find it clearer, and also $() is easier to type Sep 02 12:01:10 It's probably not ', and ` and ´ are behind sym keyboard on n900 Sep 02 12:01:10 ambiguous lexems always are prone to fail miserably Sep 02 12:01:20 and I'm not sure which one it is :) Sep 02 12:01:36 that too Sep 02 12:02:09 see PITA with "ticks" Sep 02 12:02:32 and "embeddend'ticks'" Sep 02 12:03:26 luckily in shellscript you can toggle between ' and " usually, so it's parseable Sep 02 12:04:19 how much more convenient is a clear distinction between opening and closing lexem Sep 02 12:05:06 if I wasn't that lazy, I'd even use ${envvar} all the time Sep 02 12:08:00 kerio: bq27k-detail2: >> $(( (0x${dmp[0x0A]} & 2#10000000)?1:0 )) CHGS Charge State flag. A 1 in the CHGS indicates a charge current (VSRP > VSRN). Charging." << Sep 02 12:15:14 DocScrutinizer05: am i a bad person if i make it #!/bin/bash just because i feel safer? Sep 02 12:15:31 no, not at all Sep 02 12:15:42 better than other way around Sep 02 12:16:18 never use #!/bin/sh when you're actually not sure you might have used bashisms Sep 02 12:17:39 actually verifying scripts for basic shell compatibility (no *isms) is a pretty tedious task Sep 02 12:17:55 at least I haven't found any proper tool to do it yet Sep 02 12:24:20 hm, how do i call functions in shell scripts? Sep 02 12:24:43 oh, like binaries Sep 02 12:24:44 k Sep 02 12:26:40 ah goddammit, the semantics of conditions in shell scripting are fucking stupid Sep 02 12:30:33 hmm? Sep 02 12:31:43 (like binaries) actually more like shellscripts (which in turn don't differ much from binaries for the way to invoke them) Sep 02 12:33:14 actually I just started wondering if local env predefines work for function calls intrascript as well (env1=a env2=b myfunction) Sep 02 12:33:47 to be precise I wonder if those envs will get dropped when function returns Sep 02 12:34:09 hm, how does ! work in the context of an if? Sep 02 12:34:17 i mean, what's the distributivity with &&? Sep 02 12:34:30 ! >> && Sep 02 12:35:15 meh, i'll do it inside test Sep 02 12:35:32 how do i output to stderr? Sep 02 12:35:32 if [ !a && !b] <=> if [ ! (a || b) ] Sep 02 12:35:38 yeah, yeah, i figured :) Sep 02 12:35:53 exec >2 Sep 02 12:35:56 ;-P Sep 02 12:35:56 wait, i can do || inside test, too? Sep 02 12:35:59 neat Sep 02 12:36:11 DocScrutinizer05: i actually meant for just a message Sep 02 12:36:20 also it should be >%2, right? Sep 02 12:36:24 aah, then as usual Sep 02 12:36:29 or something Sep 02 12:36:36 i can never remember the syntax for file descriptors :s Sep 02 12:36:45 yeah Sep 02 12:36:53 nope, &2 Sep 02 12:37:25 simple - just remember & means "addr of" Sep 02 12:37:32 ... Sep 02 12:37:42 what you did there, i see it Sep 02 12:40:22 I think busybox also doesn't know base2 consts Sep 02 12:40:32 like 2#10000000 Sep 02 12:41:01 not sure about that either Sep 02 12:43:13 kerio: thanks for making me aware of >> Sep 02 12:43:14 if [ x$BASH_VERSION = x ]; then Sep 02 12:43:16 echo "no bash detected, aborting" Sep 02 12:43:21 << missing a redirect to stderr Sep 02 12:43:53 DocScrutinizer05: what's /usr/sbin/i2cset -y -m 0x80 2 0x6b 0x00 0x80; # timer reset ? Sep 02 12:44:06 err, just that Sep 02 12:44:27 it resets watchdog timer in bq24150 Sep 02 12:44:47 but... you just reset it Sep 02 12:44:52 with /usr/sbin/i2cset -y 2 0x6b 0x01 0xcc Sep 02 12:45:03 or is that just "stop charge"? Sep 02 12:45:08 yep Sep 02 12:45:45 you can't stop watchdog in bq24150. It will engage first time you send *any* I2C command Sep 02 12:45:50 k Sep 02 12:45:57 yeah but you keep resetting it Sep 02 12:46:07 no, wait Sep 02 12:46:07 and on expiry will reset bq24150 to defaults which are "enable emergency charging" Sep 02 12:46:55 that's why I first stop charging, then keep watchdog happy so it stays in off-mode Sep 02 12:47:30 otherwise it would resume emergency charging after watchdog expiry (32s) Sep 02 12:47:34 i see Sep 02 12:47:59 and then when you stop it it immediately restarts because the watchdog has expired? Sep 02 12:48:22 during max 32s, yes Sep 02 12:49:06 might be as little as 15s, due to chip's timebase is rather ??? Sep 02 12:49:24 If software doesn't check in every 32s, every setting made and every command sent is lost and chip resets to defaults: charge @ 100mA to a low level Sep 02 12:49:38 exactly Sep 02 12:49:47 only 32s is upper limit Sep 02 12:50:01 to stay on safe side, we need something like 10s period Sep 02 12:50:15 hm, does sh have a do/while? Sep 02 12:50:20 yep Sep 02 12:50:29 nope Sep 02 12:50:36 it only has a while Sep 02 12:50:37 I lowered the period until resets stopped happening, and ended up with 15s Sep 02 12:50:39 while true; do echo fool:; done Sep 02 12:50:48 that's a while, not a do...while Sep 02 12:50:55 meh Sep 02 12:51:43 it has while and until Sep 02 12:51:56 I hate until Sep 02 12:52:37 dbus-send is standard fremantle, right? Sep 02 12:52:39 since it's a maximum confusing while !cond Sep 02 12:52:47 yep Sep 02 12:52:55 and so are start and stop Sep 02 12:53:05 though any path in any script isn't Sep 02 12:53:19 $PATH that is Sep 02 12:53:44 nope, basically start and stop are convenience stuff for initctl start and initctl stop Sep 02 12:54:05 but they're binary files! Sep 02 12:54:27 ooh, links to initctl Sep 02 12:54:36 so while you may get away with start bme, you actually *ought* use /sbin/initctl start bme Sep 02 12:54:48 can i at least avoid paths for sleep and echo? Sep 02 12:54:51 and test Sep 02 12:55:09 that's a pretty good question Sep 02 12:55:13 :-D Sep 02 12:55:36 probably it's safe to assume "standard" PATH being set Sep 02 12:55:49 which is sth like /bin /sbin Sep 02 12:55:56 meh, i'll choose to assume that root has initctl and dbus-send in the path, and echo and sleep are something Sep 02 12:56:06 yep Sep 02 12:56:13 basically everybody does Sep 02 12:56:17 or should i assume that they're in the place they ought to be, aka /sbin/initctl and /usr/bin/dbus-send ? Sep 02 12:56:38 for those yes you should assume and call them like that Sep 02 12:56:49 (I didn't) Sep 02 12:56:59 shame on me Sep 02 12:57:57 well according to above rationale you may assume initctl is in /sbin and thus "known", while dbus-send obviously isn't Sep 02 12:59:39 most sane way to handle that is to define envs for the "unknown2 stuff, like >> dbus="/usr/bin/dbus-send" Sep 02 12:59:51 then even test for exectuable property Sep 02 12:59:54 nah, /usr/bin/dbus-send is part of dbus Sep 02 12:59:59 it's standard fremantle Sep 02 13:00:09 i2cget/i2cset are in i2c-tools though, and i do test for those like that :) Sep 02 13:00:46 DocScrutinizer05: work in progress: http://fpaste.org/xwxH/ Sep 02 13:01:41 sweet Sep 02 13:01:45 :-) Sep 02 13:02:48 todo: more putting stuff in functions Sep 02 13:02:51 so the logic is clearer Sep 02 13:03:53 # TODO: display TTF here Sep 02 13:04:01 that, too ) Sep 02 13:04:02 :) Sep 02 13:04:03 s/TTF/TTE/ Sep 02 13:04:03 DocScrutinizer05 meant: # TODO: display TTE here Sep 02 13:04:07 ok Sep 02 13:04:09 :c Sep 02 13:04:24 it's DIScharging, so TimeToEmpty Sep 02 13:04:50 i thought it was Time To Finish Sep 02 13:04:56 but alright Sep 02 13:05:06 kerio: kudos, a delight to read such pretty code at a sunday afternoon Sep 02 13:05:11 brb walking the dog Sep 02 13:05:37 also ^__________^ Sep 02 13:16:55 kerio: while true; do kick_bq24k; sleep 10; edv1 && break; done; Sep 02 13:19:17 maybe even Sep 02 13:19:40 while kick_bq24k && sleep 10; do edv1 && break; done; Sep 02 13:20:46 (sometimes simplifies ^C abort Sep 02 13:20:48 ) Sep 02 13:22:07 oh right Sep 02 13:22:07 :c Sep 02 13:22:11 but that makes it uglier! Sep 02 13:22:17 hehe Sep 02 13:22:36 ugly and short, or whining about missing do until Sep 02 13:22:43 ;-) Sep 02 13:22:50 or bloated Sep 02 13:22:51 besides, wouldn't ^C during the sleep exit the loop, causing the rest of the script to go on? Sep 02 13:23:04 depends on settings Sep 02 13:24:11 either a cmd returning with error that isn't cought will abort whole script (incl any while true loop), or it will simply be ignored and the loop happily continues forever Sep 02 13:24:53 iirc it's set +e / -e Sep 02 13:25:14 your scripts inherit that setting from invoking shell Sep 02 13:25:53 all disclaimer "I'm not really awake yet" Sep 02 13:28:17 set +e; while true; do sleep 1; done Sep 02 13:28:22 set -e; while true; do sleep 1; done Sep 02 13:33:13 TTE is in minutes, right? Sep 02 13:34:43 right Sep 02 13:35:19 ah crap Sep 02 13:35:25 Time to empty: 0x01fc minutes Sep 02 13:35:54 >> -e Exit immediately if a pipeline (which may consist of a single simple command), a subshell command enclosed in parentheses, or one of the commands executed as part of a command list enclosed by braces (see SHELL GRAMMAR above) exits with a non-zero status. The shell does not exit if the command that fails is part of the command list immediately following a while or until[...] Sep 02 13:35:56 This option applies to the shell environment and each subshell environment separately (see COMMAND EXECUTION ENVIRONMENT above), and may cause subshells to exit before executing all the commands in the subshell. Sep 02 13:35:57 << Sep 02 13:37:39 whatever, I found sometimes >> while true; do ... sleep 1; done << loops are pretty hard to abort, while you abort them easily when testing for sleep command's return code Sep 02 13:38:06 meh, ^C kills everything Sep 02 13:38:21 maybe i should check for the exit value of sleep and restart bme before closing the script Sep 02 13:38:22 but meh Sep 02 13:38:31 depends (remaining to define on what) Sep 02 13:38:53 maybe you should use a trap to restart bme Sep 02 13:39:12 trap start_bme exit Sep 02 13:39:40 see booston Sep 02 13:39:51 or maybe the user should fuck off and let my script be Sep 02 13:40:22 >:C Sep 02 13:40:30 having a exit trap to cleanup is a pretty nice idea, much underrated Sep 02 13:41:14 and it's so infinitely more convenient than to handle each abort path in a single else branch Sep 02 13:41:15 i should probably display the current flow too Sep 02 13:42:09 http://fpaste.org/b6Ay/ Sep 02 13:42:51 http://maemo.cloud-7.de/maemo5/usr/sbin/booston Sep 02 13:44:02 by the way, what's the correct term? pet the watchdog? tickle the watchdog? kick the watchdog? Sep 02 13:46:03 diablo called it kick Sep 02 13:47:09 I kick dogs, other like to tickle them Sep 02 13:47:37 probably depends on whether you like dogs or not Sep 02 13:48:44 kerio: you'll need root to run that script, and in root you need run-standalone.sh for dbus-send Sep 02 13:49:00 I have recently had to re-install the Maemo 5 scratchbox SDK on a machine running Ubuntu. It installed smoothly, but when I run the x86 desktop withing xephyr, the Hildon widgets don't work properly. Specifically, the drop-down menus don't show any text the first time they are accessed (though they do show text during subsequent accesses) and changes made to the GUI widgets do not show up on the widgets, although the widget events a Sep 02 13:49:00 re seen by the application. Has anyone seen this problem and found a fix? Sep 02 13:53:15 (kick vs tickle) actually both terms are missleading, since the dog is considered state:sleeping unless you forget to do whatever you name the action to keep it sleeping and it wakes up and barks Sep 02 13:53:36 I think neither kicking nor tickling will make a real dog continue sleeping Sep 02 13:54:47 probably the whole term / picture / analogy is a bit... odd Sep 02 13:55:39 since real watchdogs don't usually behave on a timeout policy of any kind Sep 02 14:01:43 sure they do, if you're gone from home too long they eat your shoes and couch Sep 02 14:02:07 hehehehehe Sep 02 15:21:01 WTF of the month: http://gitorious.org/community-ssu/ke-recv/blobs/master/src/events.c#line149 ff Sep 02 15:27:48 DocScrutinizer05: there is goto in C? Sep 02 15:27:56 and it's usable? i mean people use it? Sep 02 15:28:05 args[0]: cool nick Sep 02 15:30:54 ZogG_laptop: *shudder* Sep 02 15:31:36 but that kind of error handeling code is the probably the only sane use for goto Sep 02 15:32:58 Ikarus: i know C a little bit only, never met goto yet. Sep 02 15:33:03 C/C++ Sep 02 15:33:07 both Sep 02 15:33:43 Ikarus: backward goto usually is EVIL Sep 02 15:34:11 a chain of THREE backward goto is MASTER_OF_EVIL Sep 02 15:35:17 forward goto to exit, or jump to errorhandler at end of function - ok'ish Sep 02 15:35:57 but never ever do a backward goto, especially not into a structure element like if{} while{} anything like that Sep 02 15:36:55 hell that looks like COBOL disassembly Sep 02 15:36:56 DocScrutinizer05: hehehe Sep 02 15:58:54 DocScrutinizer05: it's perfectly legal to jump inside a loop in C Sep 02 15:59:25 kerio: legal and any good practice are two differnet things Sep 02 16:00:26 also see duff's device Sep 02 16:08:48 http://pastebin.ca/2200289 Sep 02 16:09:44 ^^^ pretty low effort cleanup, could of course be done smarter and nicer Sep 02 16:10:25 no more backwards goto Sep 02 16:10:58 (yeah I know there's a bit missing around line190 then) Sep 02 16:13:41 actually 188,189 are obsolete then Sep 02 16:14:27 and [190] error3: volume_id_close (vid); needs some love Sep 02 16:15:38 by the way, label: code; on the same line should be ground for execution Sep 02 16:15:45 s/ground/grounds/ Sep 02 16:15:47 kerio meant: by the way, label: code; on the same line should be grounds for execution Sep 02 16:16:42 err, [192] error1: set_localised_label(mmc); needs ome love Sep 02 16:18:45 usually you have only one error: label, and do the >> if (fd < 0) << etc there Sep 02 16:19:37 doesn't matter you already did it before, on >> if (fd < 0) { goto error} Sep 02 16:23:39 kerio: goto jumping into while or for loops usually yields pretty messy unexpected results (if the coder doing that really expected *anything*) Sep 02 16:24:30 jumping into if(){} branches may work Sep 02 16:24:52 but for sure should be avoide3d as much as possible Sep 02 16:45:15 illuminata: please elaborate on details of your boot problem Sep 02 16:45:44 so. i'm hit by the reboot loop bug. and pretty hard apparently because flashing latest nokia firmware didn't fix it Sep 02 16:45:46 illuminata: you (ever) used recent powerkernel? Sep 02 16:45:59 and nope Sep 02 16:46:24 how exactly are the sequence of sympthoms of this reboot loop? Sep 02 16:46:58 does jumping-dots splashscreen show up? Sep 02 16:47:39 home icons either don't draw or disappear. carrier name cycles too. and if by jumping dots you mean the loading screen then yeah Sep 02 16:48:32 and it seems to actually be crashing faster now that i re-flashed Sep 02 16:48:56 probably you got some homescreen settings in your $home dir that cause hildon desktop to segfault. Eventually dsme will reboot device after too many restarts of HD Sep 02 16:49:21 similar was mentioned prior, but shouldn't a firmware flash take care of any of those settings, too? Sep 02 16:49:24 reflashing COMBINED doesn't clear/delete those files Sep 02 16:49:30 aaaaaaaaaah Sep 02 16:49:50 since flashing COMBINED keeps all your "user data" intact Sep 02 16:50:15 you got backupmenu installed? Sep 02 16:50:35 don't think so, unless that's the same as the standard backup app Sep 02 16:50:41 nope Sep 02 16:50:43 ~bm Sep 02 16:50:44 i heard backupmenu is http://talk.maemo.org/showthread.php?t=63975 Sep 02 16:51:11 it would allow you to gather access to a shell in early boot, before HD goes mad Sep 02 16:51:19 dunno if i can use it though considering the phone won't even stay on a minute now Sep 02 16:51:28 in this shell you could manually clean out some files in $HOME Sep 02 16:52:35 maybe i can install it if i work fast enough (but probably not) Sep 02 16:53:05 but assuming i don't care what's on the phone, is there any firmware packages that i can install that'd allow for a 100% clean install? Sep 02 16:53:21 DocScrutinizer05: backupmenu requires a somewhat working system though Sep 02 16:53:32 because it's intended to work when stuff is ok Sep 02 16:53:36 for now there's proabably three possible rescue pathes for you: disable watchdogs via flasher (should stopp reboots, so you *might* gain access to xterm via kbd shortcut). use ~rescue-system. reflash VANILLA (and lose all your user data) Sep 02 16:53:42 ~rescue-system Sep 02 16:53:50 oh infobot Sep 02 16:53:52 ~rescueos Sep 02 16:54:01 ~rescue Sep 02 16:54:06 DocScrutinizer05: you suck Sep 02 16:54:20 ~listkeys rescue Sep 02 16:54:22 Factoid search of 'rescue' by key (8): htcrescue ;; rescuecd ;; rescue-os ;; superrescue ;; rescue disk ;; rescue disks ;; cmd: rescue (.*?) ;; reiserfs-rescue-disk. Sep 02 16:54:26 oh, with a - Sep 02 16:54:28 wtf Sep 02 16:54:30 ~rescue-os Sep 02 16:54:30 rumour has it, rescue-os is http://wiki.maemo.org/User:Nin101/N900_Rescue_OS, or http://nin101.uni.cx/N900/rescueOS/documentation.txt, or http://206.253.166.96/N900/rescueOS/ Sep 02 16:55:17 hmm Sep 02 16:55:18 illuminata: flashing VANILLA is the "easiest" and most reliable way, given you don't care about user data (or have proper backup) Sep 02 16:55:20 infobot: no, rescue-os is http://wiki.maemo.org/User:Nin101/N900_Rescue_OS , or http://nin101.uni.cx/N900/rescueOS/documentation.txt , or http://206.253.166.96/N900/rescueOS/ Sep 02 16:55:20 okay, kerio Sep 02 16:55:25 (added spaces before commas) Sep 02 16:55:42 (not sure if it's just my irc client being stupid, but clicking didn't work properly, it thought , was part of the url) Sep 02 16:55:48 the wiki page is outdated, you might as well remove everything except the URL with the IP Sep 02 16:56:00 kerio: thanks Sep 02 16:56:18 NIN101: huh... why? what's wrong with nin101.uni.cx? Sep 02 16:56:46 i'm getting a malware warning for uni.cx in chrome Sep 02 16:56:58 infobot: no, rescue-os is http://206.253.166.96/N900/rescueOS/ Sep 02 16:56:58 kerio: okay Sep 02 16:57:11 DocScrutinizer05: how do i add redirects for factoids? Sep 02 16:57:18 like, make rescueos an alias for rescue-os Sep 02 16:57:33 ~rescueos is see rescue-os Sep 02 16:57:34 okay, DocScrutinizer05 Sep 02 16:57:43 huh Sep 02 16:57:46 ~rescueos Sep 02 16:57:46 extra, extra, read all about it, rescue-os is http://206.253.166.96/N900/rescueOS/ Sep 02 16:57:49 i see Sep 02 16:58:18 hmm, something bad about thumb2 binaries Sep 02 16:58:18 is there any way i can mass storage mount the phone before it fully boots into maemo? Sep 02 16:58:31 rescueos Sep 02 16:58:40 you can't chroot on a thumby system if you don't have a thumb-aware kernel Sep 02 16:58:47 illuminata: rescueos does that Sep 02 16:58:50 illuminata: probably not, when it bootloops like yours Sep 02 16:58:51 also bootmenu Sep 02 16:58:53 also backupmenu Sep 02 16:59:04 k Sep 02 16:59:11 yeah, except those, of course Sep 02 16:59:20 meh Sep 02 16:59:28 notice that neither one of those will let you access the rootfs Sep 02 16:59:29 just wanted to pull some pics off, then i wouldn't mind flashing vanilla Sep 02 16:59:34 because well, it's not a block device Sep 02 16:59:41 an alternative is usb networking and then ssh Sep 02 16:59:52 illuminata: you *may* disable watchdogs, and then happily mount MSM Sep 02 17:00:06 ~ass-rage Sep 02 17:00:09 no Sep 02 17:00:12 ~ass rage Sep 02 17:00:12 somebody said ass rage was Mass STOrage Sep 02 17:00:24 * kerio giggles Sep 02 17:01:59 illuminata: it's just HD respawning too fast, so dsme reboots device. with watchdogs disabled (lifeguard-wd iirc), it won't reboot and MSM should work Sep 02 17:02:28 with a bit of luck Sep 02 17:03:04 afk Sep 02 17:03:14 cool. i did some steps on a forum while at work a couple days ago, ran cat on some of that lifeguard stuff, and it was saying similarly iirc Sep 02 17:06:16 illuminata: flasher-3.5 --help Sep 02 17:07:11 w0rd. found a string of flags online that should do the trick too. but so far it's stayed alive in mass storage mode without rebooting Sep 02 17:07:25 got images out of fcamera, trying to find where the standard camera images are hidden Sep 02 17:07:34 *standard camera app Sep 02 17:08:46 flasher --set-rd-flags Sep 02 17:08:53 dcim, there we go Sep 02 17:09:09 cool Sep 02 17:10:02 alrighty, thanks for the help everybody. should just be a matter of a vanilla flash now Sep 02 17:10:15 then maybe i'll see about cssu Sep 02 17:10:20 cssu is good! Sep 02 17:10:27 flasher --set-rd-flags=no-lifeguard-reset,no-ext-wd,no-omap-wd Sep 02 17:10:29 stable too? Sep 02 17:10:31 iirc Sep 02 17:10:36 idk Sep 02 17:10:42 i'm on cssu-devel+thumb now Sep 02 17:10:52 but i'm crazy and i don't do anything important in my life anyway Sep 02 17:11:00 lol Sep 02 17:11:03 * kerio starts crying Sep 02 17:11:15 illuminata: yes, CSSU-S is stable (and actually wouldn't have rebooted for you ;-D ) Sep 02 17:11:20 cssu-s is old Sep 02 17:11:53 lol, good to know. maybe i'll give it a shot Sep 02 17:12:01 you should Sep 02 17:12:10 it's recommended for everybody now Sep 02 17:12:16 but is it true that it's hard to roll back from? Sep 02 17:12:30 you don't want to roll back Sep 02 17:12:55 just like you don't want or can rollback from PR1.3 to PR1.2 Sep 02 17:14:41 just think of it like PR1.4 Sep 02 17:15:19 k Sep 02 17:15:26 it's not about new features primarily, it's mere support for all users Sep 02 17:16:06 fixing bugs that Nokia never will touch Sep 02 17:16:25 kerio: yes CSSU stable is old but that's how they become stable. with testing testing and more testing. Sep 02 17:17:11 and next CSSU-S release is at horizon - RSN, only a matter of weeks now Sep 02 17:17:28 that's nice to know Sep 02 17:19:03 illuminata: you most probably won't notice any difference after installing CSSU-S Sep 02 17:19:43 if you do, let us know. There shouldn't be any difference regarding user experience Sep 02 17:20:07 except some annoying bugs missing Sep 02 17:20:10 ;-) Sep 02 17:21:18 all new leete things are strictly optional Sep 02 17:21:29 opt-in Sep 02 17:23:21 on CSSU-Stable that is, CSSU-Testing is a bit more bleeding edge, since we need the new features tested, so in T they are opt-out (or at least that's what we try to achieve, some are not optional [yet]) Sep 02 17:24:35 anyway I couldn't think of any reason why you would wnat to roll back to plain PR1.3.1 from CSSU-S Sep 02 17:24:50 DocScrutinizer05: it's easy to roll back to PR1.3, btw Sep 02 17:25:01 reflash, yes Sep 02 17:25:04 :D Sep 02 17:27:46 lol, gotcha Sep 02 17:34:06 illuminata: one tiny "fix" I'm actually proud of, is that CSSU in your case wouldn't reboot device on HD respawning too fast, but simply give up, leaving you with a device like you could also get with watchdogs disabled Sep 02 17:34:33 that'd be great should similar crop up again Sep 02 17:35:08 press shift-ctrl-x on kbd and you shouls still get a shell Sep 02 17:36:00 worked Sep 02 17:36:12 and for some reason the phone's stayed alive for a bit now Sep 02 17:36:28 I meant on a not-bootlooping fecked CSSU Sep 02 17:36:44 right. just saying i've been lucky thus far on this boot Sep 02 17:36:52 still gonna vanilla it i think though Sep 02 17:36:56 got what i wanted off the phone Sep 02 17:37:35 nice. Well some widget on desktop going apeshit will tear down whole HD process. Sometimes the widget recovers though, or simply waits a bit till it acts up again Sep 02 17:37:55 in your case I guess it's been OMweather or what's the name Sep 02 17:38:00 i tried removing widgets but they'd vanish before i could clear them off from a gui Sep 02 17:38:11 i uninstalled omweather completely, no dice Sep 02 17:39:04 I once had same bootloop problem when I dragged OMweather from one screen to next, while same time $something interrupted the process Sep 02 17:40:01 back in "good ole" Pr1.0 times Sep 02 17:40:45 I wasn't as luckky as you Sep 02 17:40:50 had to reflash Sep 02 17:41:02 vanilla Sep 02 17:41:27 well my luck is only that it hasn't rebooted yet but the icon cycling issues are still being annoying Sep 02 17:42:00 you actually should reflash vanilla too Sep 02 17:42:13