**** BEGIN LOGGING AT Tue Jul 23 02:59:59 2013 Jul 23 04:54:42 hi all Jul 23 04:55:20 I want to format my microsd, what filesystem to use? ext3 or ext2? Jul 23 05:14:20 Don't know, split microSD in half, format one with ext2 and the other with ext3. See which one rubs on you better.. Jul 23 06:10:23 DrCode: ext4 Jul 23 07:14:55 ^^ does mamo support ext4? Jul 23 07:51:14 kernel-power does Jul 23 07:51:26 and the automounter from cssu supports anything you throw at it, with kernel support Jul 23 08:00:31 hmm okay... kerio Jul 23 10:43:30 Hi, i have VERY bad skype video quality... The picture of persion i call is terrible blocked Jul 23 10:43:41 pixelized Jul 23 10:44:05 My video almost hang on the other side Jul 23 10:44:33 But the tiny window on my screen shows good cam performance Jul 23 10:44:38 What's wrong with it ? Jul 23 10:44:49 I have quite good enough connection Jul 23 10:44:53 hi all Jul 23 10:44:58 I can't mount ext4 Jul 23 10:45:01 has loop Jul 23 10:45:20 I got invalid option Jul 23 10:45:20 any idea? Jul 23 10:46:16 DocScrutinizer05: any ideas ? Jul 23 10:49:26 XATRIX: you most probably have either generally flawed poor bandwidth connection to internet, or your carrier (or far end's carrier) do traffic shaping based on SPI, thus violating net neutrality (<-not sure if that's the correct english term) Jul 23 10:50:03 Hm... Jul 23 10:50:14 I call for my partner that is beside me Jul 23 10:50:16 In the same office Jul 23 10:50:29 oh haha Jul 23 10:50:48 :) Jul 23 10:50:53 not that this usually makes any difference, since... it's internet Jul 23 10:51:40 but given the way peer2peer networks like skype work, this particularl situation might introduce special obstacles Jul 23 10:52:07 e.g when you both use same LAN with a router that does no local loopback Jul 23 10:52:35 {v Jul 23 10:52:40 Hmm Jul 23 10:52:41 thus for reaching your far end next to you, you need a relay somewhere outside your LAN Jul 23 10:53:04 Can i skype you? Jul 23 10:53:08 while everybody else you could reach directly with no intermediate hops Jul 23 10:53:15 no skype here Jul 23 10:53:20 :( Jul 23 10:53:34 Anyone can help me to test it ? Jul 23 10:57:15 xatrix make another skype account to your computer and call to that Jul 23 10:57:15 Why I can't mound ext4 loop .img? any idea please Jul 23 10:57:51 japa-fi: that's not helping either Jul 23 10:58:05 since that would still be all local Jul 23 10:58:56 japa-fi: yes, it doesn't matter, i need to make external call Jul 23 10:59:04 XATRIX: you can use wireshark to analyze the VoIP traffic, I guess Jul 23 10:59:20 DocScrutinizer05: it's a bit complicated. Jul 23 10:59:26 dunno if wireshark has dissectors for skype Jul 23 11:00:02 for sure not for the netto traffic since it's encrypted afaik Jul 23 11:00:14 why don't you call the skype test service? Jul 23 11:00:43 In your Contact list, select Skype Test Call. It may also be called Echo / Sound Test Service or Skype Test Call (echo123). Jul 23 11:00:46 I was about to suggest that, but then thought it might not help for video Jul 23 11:00:55 oh video Jul 23 11:03:41 yeah, and "oh, skype" Jul 23 11:03:50 skype is a nightmare Jul 23 11:05:13 and beware! some corporate LANs may have a policy that earns you losing your employment when using skype on the company's LAN Jul 23 11:06:08 skype is pretty hard to impossible to manage on a network admin level, since it acts unpredictably Jul 23 11:06:55 so even when a network admin would try to block it, there are no reliable firewall rules to establish that Jul 23 11:07:57 likewise assigning QoS priority to the skype traffic is basically impossible Jul 23 11:08:45 sometimes skype video sucks without reason Jul 23 11:14:22 btw I wonder what's up with the $internet* Jul 23 11:14:38 many sites are down for maintenance, or simply don't work anymore Jul 23 11:15:22 and then even this: [2013-07-22 21:21:58] [Notice] -christel- [Global Notice] Hi all, some of you may have been experiencing some temporary issues relating to SSL whilst we updated our certificates. Unfortunately this update was necessary for security reasons, you can read a bit more about the issues at http://blog.freenode.net/2013/07/server-hosting-and-trust/ - we apologise for the inconvenience and for the issue arising in the first Jul 23 11:15:24 place. Thank you for using freenode! Jul 23 11:21:04 DocScrutinizer05: just freenode? Jul 23 11:51:42 well, no. It seems to me something particularly nasty is going on - but then that might just be me Jul 23 12:36:48 btw I don't think that the ubuntuphone looks any particularly ugly Jul 23 12:37:49 but again it has no hw keyboard Jul 23 12:38:28 and - duh- a sapphire glass screen? that means capacitive I'd think Jul 23 12:39:08 strange , I have .img file (ext4) , in ubuntu I can mount it but not in maemo Jul 23 12:39:09 Mr Shuttleworth, show me how to use a sketching program on you rphone Jul 23 12:39:31 DrCode: you have KP? Jul 23 12:39:41 what is KP ? Jul 23 12:39:46 :-S Jul 23 12:39:48 ~kp Jul 23 12:39:49 [kp] http://talk.maemo.org/showthread.php?t=85665 Jul 23 12:39:49 kernel Power Jul 23 12:39:54 I am using cssu Jul 23 12:40:01 oh it is then Jul 23 12:40:03 that doesn't mean anything Jul 23 12:40:04 hmm 051r1 Jul 23 12:40:12 yes Jul 23 12:40:18 I hope its ok Jul 23 12:40:23 should Jul 23 12:40:29 lsmod Jul 23 12:40:36 maybe ubuntu has some proprietary ext4 extensions Jul 23 12:40:38 I also load isofs.ko Jul 23 12:40:44 but same results Jul 23 12:41:01 I see Jul 23 12:41:02 ext4 should be autoloaded if you try to mount the image Jul 23 12:41:07 -t ext4? Jul 23 12:41:07 wtf are propietary ext4 extensions Jul 23 12:41:18 invalid option it say Jul 23 12:41:23 Wizzup: ask ubuntu Jul 23 12:41:29 no, I'm asking you Jul 23 12:41:35 iirc there is a mount option to set fs Jul 23 12:41:39 All the ext4 code is gpl, how can it be proprietary Jul 23 12:41:43 well, I can convert it in ubuntu to standart ext4? Jul 23 12:42:04 modinfo ext4 Jul 23 12:42:06 DrCode: I was just kidding about the extensions Jul 23 12:42:11 -t is an option for mount Jul 23 12:42:13 ok Jul 23 12:42:26 Wizzup: gpl or not, violations going to happen Jul 23 12:42:27 I did "-o loop -t ext4" Jul 23 12:42:52 jon_y: you sounds nuts :) Jul 23 12:42:53 -o loop is deprecated iirc Jul 23 12:43:08 modprobe ext4 Jul 23 12:43:11 DocScrutinizer05: Is it? Don't think so -- what is the new syntax for loop devices? Jul 23 12:43:14 seems to be ok modinfo ext4 Jul 23 12:43:19 Wizzup: actually, it happens all the time, just look at android and all the proprietary mods :) Jul 23 12:43:19 I did Jul 23 12:43:23 losetup? Jul 23 12:43:32 kernel blobs Jul 23 12:43:40 jon_y: you still haven't linked me to ubuntu's proprietary ext4 extensions Jul 23 12:44:08 Wizzup: I am trying to say that GPL doesn't mean anything anymore to some developers Jul 23 12:44:09 what is losetup Jul 23 12:44:16 rtfm! Jul 23 12:44:18 just another license to run over Jul 23 12:44:21 man losetup Jul 23 12:44:27 jon_y: Ah, you're working for the FUD company...? Jul 23 12:44:38 nope, I am strongly pro GPL Jul 23 12:44:47 then stop the FUD :) Jul 23 12:44:49 which implies you are anti freedom Jul 23 12:44:52 to the point of being a GPL nazi Jul 23 12:44:53 If people violate the gpl, you can act upon it Jul 23 12:45:00 GPL sucks because it violates my freedom Jul 23 12:45:07 Lava_Croft-- Jul 23 12:45:21 just that I given up trying to fight BSD people that like freedom in chains Jul 23 12:45:35 freedom is chains is exactly the GPL Jul 23 12:45:35 it say invalid argument Jul 23 12:45:44 jon_y: sure, but the thing is, there's no need to bash ubuntu -- they aren't violating the gpl here Jul 23 12:45:57 Wizzup: I know, I was kidding at first Jul 23 12:45:58 mount -o loop -t ext4 my.img /mnt Jul 23 12:46:02 DrCode: modinfo ext4 !!! Jul 23 12:46:13 and sure, people violate the gpl, just like they violate other licenses, and then you can go after them Jul 23 12:46:16 if you wish Jul 23 12:46:20 DrCode: rtfm ! losetup Jul 23 12:46:28 it show 2.6.28.10-power51 Jul 23 12:46:46 sure, I cared once, but now I only look after software that I have stake in Jul 23 12:46:56 some are GPL Jul 23 12:47:05 others I just contribute Jul 23 12:47:28 DocScrutinizer05, It seems to be ok Jul 23 12:47:48 then check proper syntax for losetup Jul 23 12:48:02 Wizzup: also, bashing ubuntu is kind of the in-thing now Jul 23 12:48:06 I didn't use losetup Jul 23 12:48:11 sticking it to the man Jul 23 12:48:13 why do I need losetup? Jul 23 12:48:21 jon_y: ya, I don't use it either, but still Jul 23 12:48:25 do you even read what I write here? Jul 23 12:48:45 mount -o loop is deprectaed and obsolete Jul 23 12:49:00 it might still work Jul 23 12:49:04 I dunno Jul 23 12:49:25 ok Jul 23 12:49:28 I miss it Jul 23 12:49:33 hmm, obsolete how? it just implicitly losetups on mount and deletes on unmount? Jul 23 12:49:37 sorry Jul 23 12:50:00 losetup do I need to use -o ? Jul 23 12:50:18 or just losetup dev/loop0 my.img? Jul 23 12:53:43 >> Jul 23 12:53:44 This type of mount knows about three options, namely loop, offset and encryption, that are really options to losetup(8). (These options can be used in addition to Jul 23 12:53:46 those specific to the filesystem type.) Jul 23 12:53:47 If no explicit loop device is mentioned (but just an option ‘-o loop’ is given), then mount will try to find some unused loop device and use that. If you are not Jul 23 12:53:49 so unwise as to make /etc/mtab a symbolic link to /proc/mounts then any loop device allocated by mount will be freed by umount. You can also free a loop device by Jul 23 12:53:50 hand, using ‘losetup -d’, see losetup(8). Jul 23 12:53:52 << Jul 23 12:54:40 ok Jul 23 12:54:41 thankyou Jul 23 12:55:07 basically you first mount a image file to a *device* (the loop device) - then mount the loop device like a usual partition to a mountpoint Jul 23 12:56:06 ooh offsets, interesting Jul 23 12:56:23 I can mount partitions in whole disk images Jul 23 12:56:37 yes, you might need correct offset to mount an image file Jul 23 12:57:33 you frequently need to deal with that stuff when loopmounting a xen diskimage Jul 23 12:57:47 63 sectors to first partition or something, I can't remember Jul 23 12:57:49 to access the VM disk content without starting the VM Jul 23 12:58:14 need dd to make sure Jul 23 12:58:20 DocScrutinizer05: oh, to access a disk image with a mbr? Jul 23 12:58:29 i always wondered how to do that, actually Jul 23 12:58:29 guys, get out, enjoy the sun! ;) Jul 23 12:58:33 kerio: yes Jul 23 12:58:53 offset might largely depend on your image file - was it a raw disk image incl all MBR and stuff? and was the disk partitioned? Jul 23 12:58:54 win7mac: sun got bought by oracle Jul 23 12:59:10 DocScrutinizer05: likely everything including mbr Jul 23 12:59:22 whatever qemu raw uses Jul 23 13:02:02 from our blade-b history: Jul 23 13:02:04 151 20130723 - 15:00:05 dd if=/dev/zero of=www.img bs=1 count=1 seek=30G Jul 23 13:02:06 152 20130723 - 15:00:05 losetup -f www.img Jul 23 13:02:07 153 20130723 - 15:00:05 losetup Jul 23 13:02:09 154 20130723 - 15:00:05 losetup -a Jul 23 13:02:10 155 20130723 - 15:00:05 mkfs.ext4 /dev/loop1 Jul 23 13:02:12 156 20130723 - 15:00:05 mount /dev/loop1 /mnt Jul 23 13:04:31 seek=30G nice Jul 23 13:05:02 now if I need to find out if my workstation at work is recent enough to support it Jul 23 13:05:31 Linux 2.6.8 is the cutting edge system there Jul 23 13:05:52 so cutting edge that software migration is still on going after 2 years Jul 23 13:06:23 1350 20130314 - 19:17:12 losetup -f www.img Jul 23 13:06:25 1351 20130314 - 19:17:15 losetup -a Jul 23 13:06:26 1352 20130314 - 19:17:28 fsck -yv /dev/loop5 Jul 23 13:06:28 1353 20130314 - 19:17:38 mount /dev/loop5 /mnt Jul 23 13:06:29 1354 20130314 - 19:17:40 cd /mnt Jul 23 13:07:07 does loop actually have much overhead? Jul 23 13:07:15 or is it negligible? Jul 23 13:07:16 1361 20130314 - 19:18:12 umount /dev/loop5 Jul 23 13:07:17 1362 20130314 - 19:18:17 losetup -d /dev/loop5 Jul 23 13:10:45 483 20130722 - 19:43:47 losetup -f monitor.img Jul 23 13:10:46 484 20130722 - 19:43:47 losetup -a Jul 23 13:10:47 485 20130722 - 19:43:47 kpartx -a /dev/loop3 Jul 23 13:28:22 WARNING: Somebody pushed new version of libx11-data *system* package to extras-devel Jul 23 13:28:56 URL? Jul 23 13:29:22 I gonna nuke that package immediately Jul 23 13:29:30 version at downloads.maemo.nokia.com is 2:1.1.99.6+0m5 and pushed version to extras-devel is 2:1.1.99.6+0m5+withxcb Jul 23 13:29:41 http://maemo.org/packages/package_instance/view/fremantle_extras-devel_free_armel/libx11-data/1.1.99.6+0m5+withxcb/ Jul 23 13:29:58 Pali: is nuking it the right thing to do? Jul 23 13:30:38 DocScrutinizer05: I do not know, but system packages should not be pushed to extras(devel/testing) Jul 23 13:30:45 ye Jul 23 13:30:48 yes Jul 23 13:31:03 DocScrutinizer05: it is possible to remove it from extras-devel to prevent some problems? Jul 23 13:31:08 so hurry, I'm about to leave for an appointment i'm already late Jul 23 13:32:09 can be autobuilder configured to reject system packages? Jul 23 13:32:11 yes, I have "remove this package. package will be removed from repository automatically. ONLy FOR AFMINS. USE WITH CARE!" Jul 23 13:33:01 i think he meant automagically Jul 23 13:33:08 like, every system package Jul 23 13:33:37 c'mon "DELETE? yes no" Jul 23 13:34:07 I'm one click from deleteing it and one second from leaving Jul 23 13:35:15 now I tried to do apt-get upgrade on n900 Jul 23 13:35:25 and: Failed to fetch http://repository.maemo.org/extras-devel/pool/fremantle/free/libx/libx11/libx11-data_1.1.99.6+0m5+withxcb_all.deb Size mismatch Jul 23 13:35:52 meh! Jul 23 13:35:59 close enough Jul 23 13:36:35 Package has been removed from the repository. Go back to the package instance page. Jul 23 13:36:47 o/ Jul 23 13:36:49 so delete it is also broken in apt repository Jul 23 13:37:31 btw, on this page http://maemo.org/packages/package_instance/view/fremantle_extras-devel_free_armel/libx11-data/1.1.99.6+0m5+withxcb/ is missing who uploaded that package Jul 23 13:42:09 when will be package removed from apt repository? Jul 23 13:42:18 it is still in extras-devel apt repo Jul 23 13:47:41 DocScrutinizer05: is this normal? http://maemo.org/packages/view/modrana/ Jul 23 13:47:56 try to open above page Jul 23 14:17:45 Hello all, I have a defected N900, can anyone help me solve my issue? Here is a demonstration: http://www.youtube.com/watch?v=yOMQR6FZuek on 8:30 I managed to make it boot, even for some seconds. Jul 23 14:42:17 Unfortunately, it's a decidedly non-trivial fix Jul 23 14:42:42 You likely - if it turns out not to be a connector issue - need to re-work some joint on the motherboard. Jul 23 14:43:13 This will involve at best a high chance of damaging it - even if you're skilled in SMT rework. Jul 23 14:45:27 package is still in repository: http://repository.maemo.org/extras-devel/pool/fremantle/free/libx/libx11/libx11-data_1.1.99.6+0m5+withxcb_all.deb Jul 23 14:45:37 DocScrutinizer05: ^^^ Jul 23 14:45:51 apt-get still show me new version Jul 23 14:48:16 Pali: sorry, on the road Jul 23 14:48:42 merlin1991: ping Jul 23 14:48:54 package promotion to repo said flawed/broken, so I guess is deleting Jul 23 14:48:58 ask merlin1991 Jul 23 15:19:52 needed some help for using scratchbox2 with qemu. but having hardtime understanding the difference in the commands while initializing the sb2-init(scratchbox2) target. Jul 23 15:20:00 what is the difference between sb2-init armel $HOME/path-to-arm-linaro-cross-toolchain/bin/arm-gnueabi-gcc Jul 23 15:20:06 and sb2-init -c qemu-arm ARM /opt/arm-2011.03/bin/arm-none-linux-gnueabi-gcc Jul 23 15:20:19 ignoring the toolchain path Jul 23 15:25:25 anyone? Jul 23 15:28:34 hi Jul 23 15:28:39 regarding that python issue Jul 23 15:29:06 today, earlier when booting up the phone, I got the issue with slow mounts again Jul 23 15:29:18 however, this time, after waiting a few minutes - nothing happened Jul 23 15:29:44 Just to make sure it wasn't just delayed even further, I just played a game meanwhile Jul 23 15:29:53 but after 30 minutes, it had yet to mount Jul 23 15:29:58 So I ended up mounting it manually... Jul 23 15:30:09 Just fyi Jul 23 15:30:15 And I still don't know what the issue is caused by Jul 23 15:30:32 Also looked into /etc/fstab before mounting it manually Jul 23 15:30:37 it wasn't even listed there Jul 23 15:30:48 so when it was auto-generated, they were basically completely missing Jul 23 15:31:24 I think this is an issue with dbus, just a hunch. I was recently experimenting on writing upstart scripts and for some reason my script didn't start even though I simply copied and pasted a working upstart script and renamed a few things here and there. Jul 23 15:32:16 well, my syslog did tell you when you looked into it that there was some potential issues with dbus Jul 23 15:32:27 don't know *what* issues, but yeah Jul 23 15:33:24 dbus was lagging according to your old syslog output, it was waiting on something but I'm not too familiar with upstart and how it works. Jul 23 15:33:37 neither am I :p Jul 23 15:33:54 and I don't know if this is a common issue either Jul 23 15:34:05 i.e. you have the same lagging issues, and that it's normal in N900 for it to do that Jul 23 15:34:12 (you and everyone else I mean) Jul 23 15:36:25 My syslog isn't working (again).. I don't think its normal for dbus to do that, it could be a missing upstart script or something that is causing dbus to be lagging. Jul 23 15:40:19 In theory a properly functional upstart would allow one to be able to write a custom script that would simply show a banner to tell you that upstart has finished. In my case somewhere is broken and I can't understand what or where when even syslogd is failing to output the logs to the right place. Jul 23 15:41:49 You probably have a working syslogd, there might be measures available for you to take to try and take. Jul 23 16:22:56 good finds, I'll look into it later Jul 23 16:22:59 bbl Jul 23 17:04:09 syslog depends on having write access to $LOGDIR/syslog Jul 23 17:05:57 if for whatever reason mounting of /home and even mount -o remount,rw / fails resp takes ages, e.g due to fsck then you shouldn't be surprised to see mounts *to* those filesystems fail Jul 23 17:14:36 Yeah that would be from /etc/event.d/rcS-late or wherever the file resides. Re: syslog issue, the issue seems to be way more complicated than that. When I ran sysklogd manually or even (re)starting it via /sbin/st{art,op} syslogd writes to the aforementioned files. Just not at boot, to make matters even worse, syslogd is started well beforehand and it doesn't seem to have issues from upstart. I Jul 23 17:14:55 I'm turning debugging mode on and reading from the output, hoping to get some sense out of it. Jul 23 17:17:00 psycho_oreos: are you talking to me? Jul 23 17:17:54 FIQ, not for what I wrote above. Did you try and do some experimentations to see where the issue lies? DocScrutinizer05 and I sort of gave hints on where the issue may lie. Jul 23 17:18:23 psycho_oreos: if these hints were given after I posted what I did before - I didn't see those Jul 23 17:18:42 Actually partially, the first sentence that I wrote was about mounting issue where /home may fail (along others iinm). Jul 23 17:19:45 https://mg.pov.lt/maemo-irclog/latest.log.html#t2013-07-23T18:40:19 Jul 23 17:20:26 http://privatepaste.com/f1a96cc863 ... bah DocScrutinizer05 beat me to it lol. Jul 23 17:22:31 ty DocScrutinizer05 Jul 23 17:22:55 /topic is your friend ;-) Jul 23 17:23:18 and of course mgedmin Jul 23 17:24:06 right Jul 23 17:24:27 mgedmin: Tanks Marius for your very long term and pretty much vital support on chanlogging Jul 23 17:25:14 thanks even :-) Jul 23 17:25:38 DocScrutinizer05: but I've never had issues with mounting /home, and the device boots up (I'm very confident /home needs to be mounted before this) just as fast as it did when I first got the device, i.e. there's no notable slowdown Jul 23 17:25:56 the only slowdown I've had (and even fails) are with those python bindmounts Jul 23 17:26:18 strange Jul 23 17:26:20 and when I mounted them earlier manually, it was done instantly Jul 23 17:26:20 i.e. no delay Jul 23 17:26:41 gnutils mount? Jul 23 17:27:02 Usually if there's issues with mounting, I thought syslogd would have logged that. Jul 23 17:27:08 also created a simple script file with those mounts running as a temporary work-around if the mounts refuses to be done in the future... but the optimal solution would be have it working completely Jul 23 17:27:15 Particularly for /home for instance. Jul 23 17:27:35 psycho_oreos: well, the one I gave you (regardless of the private info) was one where python bindmounts was significantly slower than it should (but that time it didn't fail as this time) Jul 23 17:28:29 IroN900:~# ls -l /bin/mount Jul 23 17:28:30 lrwxrwxrwx 1 root root 7 2012-09-11 15:34 /bin/mount -> busybox Jul 23 17:28:59 DocScrutinizer05: same output here Jul 23 17:29:12 and "which mount" shows it's using /bin/mount Jul 23 17:29:16 IroN900:~# busybox Jul 23 17:29:18 BusyBox v1.10.2 (Debian 3:1.10.2.legal-1osso31+0cssu0) multi-call binary Jul 23 17:29:20 FIQ, yeah, I do recall there was no issues with mounting from that log, At least not for non-binding mounts. I'm still trying to get my syslogd to work properly so that I can see if my device would also show slowdowns from dbus as well. I don't expect that would be a natural behaviour but like I said, my syslogd is refusing to work properly so its hard to say. Jul 23 17:30:17 BusyBox v1.21.0 (Debian 1.21.0power2+thumb1) multi-call binary. Jul 23 17:30:24 LOL Jul 23 17:30:31 DocScrutinizer05: do you assume this can be the issue? Jul 23 17:30:40 pretty much, YES Jul 23 17:31:11 (^ got that earlier due to some features I missed, mostly compression utilities IIRC) Jul 23 17:31:13 at least that's what I always warned about Jul 23 17:31:21 cassandra Jul 23 17:31:28 you did? Jul 23 17:31:38 well I guess I can rollback to the default binary Jul 23 17:31:50 ...hmm not sure if uninstalling it will simply pull back the old one Jul 23 17:32:02 I bitched a uncounted number of times "DO NOT mess with init shell!!" Jul 23 17:32:06 or if I need to install some package which that one conflicts with Jul 23 17:32:33 DocScrutinizer05: I did not see this, if you're referring to me :p Jul 23 17:32:41 just saying Jul 23 17:32:46 not your fault Jul 23 17:32:56 oh well Jul 23 17:33:04 how to get back to the old busybox? Jul 23 17:33:15 and after all it might turn out it's _not_ messybox-power fault Jul 23 17:33:17 I'm not sure if simply uninstalling whatever provides busybox will work Jul 23 17:33:26 as I think that would leave me with *no* busybox Jul 23 17:33:46 I just say it *may* be, and that's what I always warned might happen Jul 23 17:33:50 *uninstalling whatever provides busybox-power Jul 23 17:33:52 Well Jul 23 17:33:57 That's easy enough to figure out Jul 23 17:34:20 fix the default one, do a few boots, and if it doesn't happen again for a week I guess we can conclude that was the error Jul 23 17:34:35 FIQ: there's a -(imulate) call in apt-get, no? Jul 23 17:34:43 -s(imulate) Jul 23 17:34:58 and once it fails again, conclude the opposite Jul 23 17:36:00 just fgure some fool "fixed a bug" in messybox power mount plugin Jul 23 17:36:28 or simply linked it against a lib that's not in rootfs Jul 23 17:36:36 or whatever Jul 23 17:37:13 well mount *does* work, after all it can mount non-python bindmounts (including /opt) fine Jul 23 17:37:15 hmm, mine has the same version as FIQ (re: busybox-power) and I don't have that issue with python bind mounts. *shrugs* can't hurt to try out standard busybox either way. Jul 23 17:37:18 so it can't be the latter Jul 23 17:37:22 i.e. making it not work at all Jul 23 17:37:28 virtually any API change or call sequence change can invoke a hidden race condition Jul 23 17:38:02 DocScrutinizer05: yeah, just said it couldn't be linked to a broken lib as that would make mount fail completely, no? Jul 23 17:38:17 probably, yes Jul 23 17:38:20 Actually correction, mine isn't thumb optimised: BusyBox v1.21.0 (Debian 1.21.0power2) multi-call binary. Jul 23 17:38:21 hmm this is strange Jul 23 17:38:39 did dpkg -L busybox-power to see if that's what provided the changed /bin/busybox Jul 23 17:39:08 and while I have a modified /bin/busybox (given the fact that it gives the above result @version), that package does NOT seem to provide this file Jul 23 17:39:21 dpkg -S will tell you where it came from Jul 23 17:39:30 oh, was going to google how to do that lol Jul 23 17:39:31 ty Jul 23 17:40:27 hm I don't understand this output exactly Jul 23 17:41:10 diversion by busybox-power from: /bin/busybox Jul 23 17:41:23 diversion by busybox-power to: /opt/busybox-power/busybox.distrib Jul 23 17:41:36 busybox: /bin/busybox Jul 23 17:42:57 hmm Jul 23 17:43:26 oh, seems like what busybox-power did was to move the original file to that distrib file, and placed its' own file in /bin/busybox Jul 23 17:43:34 then I guess I can simply uninstall busybox-power Jul 23 17:43:50 and *if* that somehow still gets rid of /bin/busybox, reinstall busybox Jul 23 17:45:01 ask kerio, master of dpgk divert Jul 23 17:45:43 yup, kerio ;) Jul 23 17:47:06 I'm quite sure ~50 other users in this channel can teach you about dpkg divert - I can't Jul 23 17:47:16 wut Jul 23 17:47:21 this doesn't make any sense Jul 23 17:47:42 the file in /opt/busybox-power/busybox.distrib is NOT the vanilla busybox version Jul 23 17:47:53 but rather another version of busybox-power Jul 23 17:47:57 that... doesn't make any sense Jul 23 17:49:20 but rather 1.20.2power4 Jul 23 17:49:57 and the removal of busybox-power just made me end up with an older busybox-power version, allthough busybox-power cleaned up its' files Jul 23 17:50:04 so I'll have to reinstall busybox now anyway... Jul 23 17:50:43 there been incredible mess with messybox-power and cssu updates Jul 23 17:51:14 nfc how that got resolved Jul 23 17:51:19 ok Jul 23 17:51:46 now I have 3:1.10.2.legal-1osso31+0cssu0 Jul 23 17:52:08 don't know if this will solve my issue though, but oh well Jul 23 17:52:22 basically it boilded down to: cssu did try update to busybox, that conflicted when user had busybox-power installed, and rendered user's system fubar Jul 23 17:52:45 According to this: https://garage.maemo.org/plugins/ggit/browse.php/?p=busybox-power;a=blob;f=debian/busybox-power.prerm;h=a0d4ea18e79dea8733927588e208f30a1edd8554 original busybox should be in /bin/busybox.distrib, if that doesn't exist it will show an error. Jul 23 17:53:46 psycho_oreos: I have no such file -- allthough this is most likely due to me uninstalling busybox-power Jul 23 17:54:31 FIQ, its supposedly to be part of prerm script. I'm guessing its what dpkg does during removal stages, actually before removing. Jul 23 17:54:42 hmm Jul 23 17:55:04 Anyway you have original busybox back, so the case is now irrelevant. Jul 23 17:55:31 yeah Jul 23 17:55:31 well, cssu busyboc Jul 23 17:55:31 *busybox Jul 23 17:55:35 I've so far encountered 2 issues since optfs restoration Jul 23 17:55:36 Let's see if there will be any more Jul 23 17:55:57 thinking about it, I could even figure how the fstab patch to cssu-t8 never been tested aganst standard busybox Jul 23 17:57:19 seems like I'm running T8.2 Jul 23 17:57:19 FIQ: I think all that bug hunting on your severely tainted system is not fruitful Jul 23 17:57:21 so whatever the patch is, I have it Jul 23 17:58:48 we easily could hunt this and another dozen bugs for the next 6 months Jul 23 17:59:13 lol Jul 23 17:59:29 never being able to reproduce it on a clean system that hasn't suffered massive damage from opt "restauration" Jul 23 17:59:59 heh Jul 23 18:01:16 Yeah I've pointed out before that the setup you have now (FIQ) is very much a zombified setup. Jul 23 18:01:32 usually the standard suggestion is: do a osso-backup, do a BM backup just in case, then reflash and restore from osso-backup Jul 23 18:02:35 sure, in your case that's a decent chunk of work Jul 23 18:03:09 since you also need to restore PK, CSSU-T and CSSU-thumb (aiui) manually Jul 23 18:11:46 * DocScrutinizer05 wonders if FIQ is still "alive" Jul 23 18:12:17 DocScrutinizer05: True Jul 23 18:12:43 and, how feels booting with stock busybox? Jul 23 18:12:48 Worth pointing out that this issue existed before already (the python one), I just didn't bother then -- not sure about the busybox issue Jul 23 18:13:07 Though, you *are* right in that this restore probably introduces some issues Jul 23 18:13:23 that *might* be possible to solve, but would take very long Jul 23 18:13:36 DocScrutinizer05: no issue this boot Jul 23 18:13:42 allthough it was occasional after all Jul 23 18:14:01 yeah, I'd love to see you do 2manual bootloop" ;-) Jul 23 18:14:11 s/2/"/ Jul 23 18:14:22 ~status Jul 23 18:14:28 :-S Jul 23 18:14:29 DocScrutinizer05 meant: yeah, I'd love to see you do "manual bootloop" ;-) Jul 23 18:14:30 Since Wed Jul 17 05:38:45 2013, there have been 10 modifications, 332 questions, 0 dunnos, 0 morons and 150 commands. I have been awake for 6d 12h 35m 42s this session, and currently reference 119097 factoids. I'm using about 23400 kB of memory. With 0 active forks. Process time user/system 1070.09/21.66 child 0/0 Jul 23 18:14:37 ooh Jul 23 18:14:43 as in hard shutdown, or shutting down the right way? Jul 23 18:14:59 never do hard shutdown on purpose Jul 23 18:15:02 DocScrutinizer05: regarding reflashing - how come you didn't ask that from the beginning though? :P Jul 23 18:15:08 yes I thought so Jul 23 18:15:10 except for very dedicated test case Jul 23 18:15:17 just asked, as you put quotes in manual bootloop lol Jul 23 18:16:01 FIQ: you explicitly said you want to avoid reflashing Jul 23 18:16:11 oh I did? Jul 23 18:16:17 and you had a plan how to avoid it Jul 23 18:16:22 ahh Jul 23 18:16:31 this was back when I had a corrupted file system Jul 23 18:16:34 Well then Jul 23 18:16:42 I'll not reflash, unless further issues arise Jul 23 18:16:54 that's a good plan Jul 23 18:17:20 (unless they're minor and I know the problem/solution right away) Jul 23 18:17:34 after all we now have a decent chance to find out if BB-P introduced that mysterious python bindmount issue Jul 23 18:17:46 mhm Jul 23 18:18:31 we still wouldn't know if BB-P got damaged durng your fs corruption + restore Jul 23 18:18:49 but at least we had a clue where to search, or not to search Jul 23 18:19:03 There is another alternative plan to forgo flashing (only temporary), you can use one of those osso scripts (can't remember which) to generate a list of user installed packages. Check and see for each of those packages are still publically available, if not repack the debs (and hope they won't contain broken symlinks, etc). Reflash and restore the list of installed packages that you had along with reconfiguring. Jul 23 18:19:43 psycho_oreos: I don't see how this is avoiding flashing though :P Jul 23 18:19:47 just making it less cumbersome Jul 23 18:20:03 FIQ, its better than fixing up issues as you go along :p lol. Jul 23 18:20:04 sounds like a terrible amount of work Jul 23 18:20:22 psycho_oreos: depends Jul 23 18:20:25 Either way we've all seemed to learn something in this process. Jul 23 18:20:45 Unless the issues are major - I'd rather fix them on the go as they're not really going to make things worse (I think..?) Jul 23 18:21:22 DocScrutinizer05: another solution to avoid reflashing, *and* fix all the bugs is to simply do a combined BM restore Jul 23 18:21:33 i.e. optfs and rootfs Jul 23 18:21:34 The issue you had with python in that very first instance was quite interesting. It slowly began to not get mounted and eventually stops mounting it properly. Jul 23 18:22:03 psycho_oreos: well as I said - I've had this issue for a long time, not really bothering with it Jul 23 18:22:06 Only if you had a current copy of optfs. Jul 23 18:22:07 It could have failed before, I don't know Jul 23 18:22:11 as my solution was to simply reboot Jul 23 18:22:20 psycho_oreos: my optfs restore was from 7 months ago Jul 23 18:22:29 (not my ~, but the optfs) Jul 23 18:22:59 FIQ, I thought you had a mismatched copy of rootfs vs optfs tarballs. Jul 23 18:23:11 psycho_oreos: yes, this is the case *now* Jul 23 18:23:36 psycho_oreos: I asked before the restore (when fsck messed up completely) if I was going to do optfs+rootfs or only optfs in here Jul 23 18:23:44 either you or doc suggested just optfs Jul 23 18:24:04 psycho_oreos: the *only* issue I have with a combined is that there might be kernel mismatches Jul 23 18:24:21 as some (not all!) of my restores are from vanilla, and some for KP (though possibly an older version) Jul 23 18:24:22 FIQ, doubt it would be me, I usually do complete restore. I recall you had pk with mismatched kernel version. Jul 23 18:24:44 Yes, I have several different kernel version restores Jul 23 18:24:51 s/restores/backups/ Jul 23 18:24:51 FIQ meant: Yes, I have several different kernel version backups Jul 23 18:24:55 FIQ, yeah hence optfs restoration was adequate enough to forgo bootloops. Jul 23 18:24:56 you did a mkfs.ext3 on /home Jul 23 18:25:15 DocScrutinizer05: yes, allthough this was undone by BM Jul 23 18:25:23 you asked whether you need to reflash completely or could just fix /home Jul 23 18:25:24 apparently BM actually reformats /home when it restores Jul 23 18:25:43 no, it usually doesn't Jul 23 18:25:48 afaik Jul 23 18:25:55 well during the restore process, it said "Formatting OptFS" Jul 23 18:26:03 as the first step Jul 23 18:26:03 OOOH! Jul 23 18:26:25 that would mean it erradicates ~user/ Jul 23 18:26:38 it did restore /home/user as well yes Jul 23 18:26:57 Allthough ~ was restored by me afterwards from the backup tar I did earlier Jul 23 18:27:03 before I did fsck Jul 23 18:27:13 well, then I have to open a (virtual) ticket against BM, for fixing the wording in menus Jul 23 18:27:36 Should have tar-ed the whole file system eh... Jul 23 18:27:55 Then I wouldn't have a fs mismatch issue (allthough the MH issue would be there still) Jul 23 18:28:39 DocScrutinizer05: and, if so, my wording was a bit bad Jul 23 18:28:41 More like python issue, hence MH. Jul 23 18:28:48 psycho_oreos yeah Jul 23 18:29:00 .. because what I really asked was "restore rootfs+optfs or just optfs?" Jul 23 18:29:16 @ Jul 23 18:29:19 @ (20:25:17) (DocScrutinizer05) you asked whether you need to reflash completely or could just fix /home Jul 23 18:29:35 oh wait no Jul 23 18:29:49 DocScrutinizer05: are you referring to before the fsck? Jul 23 18:30:05 because then yeah my wording was that Jul 23 18:30:48 DocScrutinizer05: what my point really was was that I would rather avoid reflash aka using flasher-3.5 Jul 23 18:31:12 while I would rather avoid rootfs restore too, I see this as far less of an issue Jul 23 18:32:33 (nothing against the flasher software, but it really doesn't like to reflash unless you try several times, possibly due to driver issues) Jul 23 18:32:55 (20:27:07) (DocScrutinizer05) well, then I have to open a (virtual) ticket against BM, for fixing the wording in menus Jul 23 18:33:33 hm... but I got the impression that it did restore the entire optfs (allthough I didn't know if it did reformat or not before I actually tried) Jul 23 18:33:43 it was you who pointed out that it probably just restored /opt Jul 23 18:33:59 (which was why I created the ~ tarball in first place) Jul 23 18:39:29 FIQ: that's weird, busybox-power tends to do the right thing Jul 23 18:39:32 maybe it was an older version Jul 23 18:39:47 well I had 1.21 and 1.20 Jul 23 18:39:51 binaries Jul 23 18:40:01 the 1.20 one appeared when uninstalling busybox-power Jul 23 18:40:04 so something went wrong Jul 23 18:40:19 perhaps it moved the 1.20 version to /bin/busybox.distrib accidentally? Jul 23 18:40:26 when updating to 1.21 I mean Jul 23 18:48:02 FIQ: BM says "backup /opt" afaik Jul 23 18:48:43 if it actually backups /home, incl /home/opt and /home/user, then this is a severely misleading wording Jul 23 18:49:33 basically there's no such thing like an optfs Jul 23 18:51:05 IroN900:~# LANG=en_EN df -h /opt Jul 23 18:51:07 Filesystem Size Used Avail Use% Mounted on Jul 23 18:51:08 /home/opt 2.0G 819M 1.1G 43% /opt Jul 23 18:51:51 I somehow doubt BM would say "backup /home" when the created tarball is labelled optfs. Jul 23 18:52:35 localization/i18n ofgnu-df is abysmal btw Jul 23 18:52:55 an indigestable gibberish mix of english and german Jul 23 18:53:30 lol Jul 23 18:53:49 afk Jul 23 18:54:02 will look into the wording when back Jul 23 18:54:59 so if we assume optfs has any reasonable meaning as a term, then it would be /home/opt Jul 23 18:55:21 "$T2S -s 2 -H left -y $y -T 0 -t " Backing up Optfs with Tar..."; y=$((y+20))" Jul 23 18:55:56 when BM saves /home and not /home/opt then this is wrong, either in wording or in what it actually does Jul 23 18:57:45 IroN900:~# df -h ~user Jul 23 18:57:46 Dateisystem Size Used Avail Use% Eingehängt auf Jul 23 18:57:48 /dev/mmcblk0p2 2,0G 819M 1,1G 43% /home Jul 23 18:58:51 use LANG=C :-) Jul 23 18:58:55 since /opt is a bindmount, it basically has no directly associates filesystem at all Jul 23 18:59:25 Pali: see above, also note that LANG=C does NOT work on maemo Jul 23 19:00:24 I intentionally left out LANG=en_EN to en_passant demonstrate >> an indigestable gibberish mix of english and german Jul 23 19:01:19 I guess you're capable to deduce what the german strings mean, by comparing to above paste Jul 23 19:04:08 psycho_oreos: what does BM use in tar command? /opt? /home? /home/opt? Jul 23 19:05:06 or even an arbitrary generic mountpoint where /dev/mmcblk0p2 got mounted to? Jul 23 19:06:55 sth like "mount /dev/mmcblk0p2 /mnt/0p2; tar ... /mnt/0p2/opt" ? Jul 23 19:08:14 It mounts /dev/mmcblk0p2, and "gtar p"$backupOptCompress"cf $backupOptLocation ./ --ignore-failed-read --numeric-owner 2> /tmp/tarstatus & pid=$!" Jul 23 19:08:23 pondering how BMrev1 did a dd to do volume images, I guess I can understand how this wording and this way to do the backup evolved Jul 23 19:10:49 Mounts /dev/mmcblk0p2 to /tmp/mnt/optfs Jul 23 19:11:21 that's where naming confusion starts Jul 23 19:11:58 I gather it then does a "cd /tmp/mnt/optfs" Jul 23 19:12:17 Yes and tarballs that directory. Jul 23 19:12:18 before tar-ing up ./ Jul 23 19:12:58 well, /dev/mmcblk0p2 is _not_ optfs Jul 23 19:13:39 even when you mount it to a mountpoint dir called optfs, it stays /home volume (ext3fs) Jul 23 19:14:39 Even grepping the output of the (old) compressed tarballs gives /home/user/* files. Jul 23 19:14:43 (there's a bit of widespread confusion here about filesystems and volumes. While a volume can have only one filesystem...) Jul 23 19:15:39 Rather partitions than filesystem :) Jul 23 19:15:57 meh, LVM defeats the notion of partitions Jul 23 19:16:10 it's "volume" Jul 23 19:17:14 funny enough I'm not aware of a proper german translation of the term, other than filesystem/dateisystem Jul 23 19:18:30 a loopmounted image also hardly is a partition, though it might consist of several virtual partitions Jul 23 19:18:44 Heh, not everyone has LVM setup :D. Jul 23 19:18:55 but everybody has volumes Jul 23 19:19:08 not everybody has partitions Jul 23 19:19:19 at least on this visible level Jul 23 19:20:05 iirc e.g. mysql could even use raw devices without any MBR and partition table Jul 23 19:20:19 a floppy has no partitions, yet it's a volume Jul 23 19:20:49 (usually doesn't have partitions...) Jul 23 19:20:58 * psycho_oreos recalls of that windows naming convention, where volumes are used to refer for each hard disk connected regardless of the partitions the volume has or if the whole setup was even in LVM/RAID/JBOD/etc. Jul 23 19:21:37 that are devices aka drives Jul 23 19:22:53 at least that's how I understand all those terms, and try to use them like that Jul 23 19:23:18 Was it called devices? o.O I somehow recalled it was named volumes. Even some of the complex areas like when dealing with something like recovery console or maybe even registry editor where certain portions contain the reference "Vol0". Jul 23 19:23:36 I dunno about windows Jul 23 19:24:07 Ahh, well yeah they tend to have mixed/dubious term usages. Jul 23 19:24:16 windows starts naming confusion by calling itself an OS Jul 23 19:24:58 which evidently it wasn't until at least win-xp Jul 23 19:25:20 or was it win-nt Jul 23 19:25:22 ? Jul 23 19:25:35 I actually thought you were referring to windows before, heh. I'm guessing when it comes to LVM, each device or devices connected together are called volume regardless of what medium/density/etc they are. Jul 23 19:27:23 *shrugs* they created standards which throw people of other platform users off :) Remember that issues with MB/Mb usage with hard disk manufacturers had to make amends? Jul 23 19:27:45 s/es// Jul 23 19:27:47 psycho_oreos meant: *shrugs* they created standards which throw people of other platform users off :) Remember that issu with MB/Mb usage with hard disk manufacturers had to make amends? Jul 23 19:28:11 lol :x Jul 23 19:28:29 * win7mac points to this beauty of a shy flower, taken with N9+macro-lens: https://dl.dropbox.com/lightbox/home/maemo-stuff/N9-macro and like me here if you dare ;) http://talk.maemo.org/showthread.php?p=1361459#post1361459 Jul 23 19:28:29 pebkac haha.. s/es/e/ Jul 23 19:29:29 *ahem* -> #harmattan ;) j/k Jul 23 19:32:32 Error (404 ) We can't find the page you're looking for. Jul 23 19:58:40 hello everybody... quick question, does the Skype implementation within Maemo core mean there was an agreement between the community and Microsoft? If that's the case is there any document about it? Jul 23 19:59:33 no Jul 23 19:59:55 maemo had skype long before it was acquired by M$ Jul 23 20:00:26 als, skype parts of maemo were not provided by community Jul 23 20:00:56 were they provided by Skype itself? Jul 23 20:01:47 Nokia Jul 23 20:02:20 shit, sorry, correct link (I hope): https://www.dropbox.com/s/fl8au9zjdh1yguo/13070049_3x2_LQ.jpg Jul 23 20:02:55 and you're more likely to get hit by a plane than see any of those docs GogoPogo. Jul 23 20:03:23 err... https://dl.dropbox.com/s/fl8au9zjdh1yguo/13070049_3x2_LQ.jpg Jul 23 20:05:21 Sicelo: I got the idea. But it's still not very clear to me how much of maemo's core comes from the community and how much comes from Nokia.. Believing linux meant free community development I believed that an agreement with Skype or Microsoft would make quite not much sense.. Jul 23 20:05:47 everything comes from Nokia. Jul 23 20:06:00 I'm not very linux-literate so it's better for me to ask the savvy Jul 23 20:06:04 read your N900 manual. Jul 23 20:06:04 or this even: http://i.imgur.com/JbCKKWE.jpg Jul 23 20:06:13 first pages Jul 23 20:06:37 Sicelo: thanks for the advise, hope asking didn't irk you Jul 23 20:06:43 GogoPogo: N900 was not developed by community. Jul 23 20:06:49 nop, not irked :P Jul 23 20:07:08 advice* Jul 23 20:07:27 ~closed Jul 23 20:07:28 [closed] http://wiki.maemo.org/Why_the_closed_packages or https://wiki.maemo.org/Fremantle_closed_packages Jul 23 20:07:36 GogoPogo: ^^^ Jul 23 20:08:02 N900 is just great because it is almost standard enough linux (Debian).. but that's as far as it goes. Jul 23 20:08:16 hence, the link DocScrutinizer05 just pointed you to Jul 23 20:09:24 thanks Doc, thanks Sicelo.. I guess the maemo.org wikis should be dumped for Evopedia Jul 23 20:09:37 ? Jul 23 20:09:39 huh? Jul 23 20:10:10 Evopedia, you know.. the application enabling you to read Wikipedia dumps Jul 23 20:10:19 ooh Jul 23 20:10:32 yeah, there been such plans several times Jul 23 20:10:46 I donno if anybody actually compöeted on them Jul 23 20:10:53 completed Jul 23 20:11:02 german keyboard layout-owned Jul 23 20:11:24 Lava_Croft: he's more likely to have pressed Alt+0246 Jul 23 20:11:25 :D Jul 23 20:11:37 ;) Jul 23 20:12:02 anyway, DocScrutinizer05 is there a way to find out? Jul 23 20:12:14 ~pkg Jul 23 20:12:15 i heard pkg is http://maemo.org/packages/ Jul 23 20:12:59 we now can provide whatever raw and cooked dumps of wiki you'd need for such project, thanks to community now managing the infra Jul 23 20:13:35 that's wonderful Jul 23 20:13:46 I'd love to have Maemo wiki with my while I'm on the go... Jul 23 20:13:54 with me* Jul 23 20:14:03 sure, a very sane rationale Jul 23 20:15:13 you *might* look into openmoko wikireader project, which did sth aiui similar to evopedia which I dun't know yet Jul 23 20:15:52 zecke did a huge job compressing complete wikipedia Jul 23 20:16:18 and re-rendering images to 4bit grayscale ;-) Jul 23 20:16:31 including images :O Jul 23 20:17:22 I completely lost track of this project, but it's FOSS as well Jul 23 20:17:40 but how can Openmoko be compatible with maemo? Jul 23 20:17:56 not at all Jul 23 20:18:02 heh Jul 23 20:18:13 but the code is pretty plain common c Jul 23 20:18:34 the core is platform agnostic I'd say Jul 23 20:18:43 I'm more interested in the dumps than in the app.. I already have Evopedia which works flawlessly within maemo's browser Jul 23 20:18:53 ooh Jul 23 20:19:02 well, then it should be simple Jul 23 20:19:27 So if you have dumps for wiki.maemo.org we can download them.. Jul 23 20:19:54 tell me what sort of dump you need, and how it#s supposed to get prepared Jul 23 20:20:11 ok, let me check inside my n900 Jul 23 20:20:34 there's probably a howto somewhere on evopedia pages, eh? Jul 23 20:21:04 wiki.m.o is mediawiki Jul 23 20:21:38 you might want to ask ashley about details and help to get the dump Jul 23 20:22:00 ashley is our wiki maintainer Jul 23 20:22:37 I can facilitate Jul 23 20:23:54 I'm checking evopedia's website and it seems to be able to read wikipedia and wiktionary-like dumps only, Jul 23 20:24:09 oooh, I got a not-too-recent "dump" of wiki on my box: http://maemo.cloud-7.de/wiki/ Jul 23 20:24:33 hi Jul 23 20:24:47 4th power cycle, no issue @bindmounts so far Jul 23 20:24:53 (20:53:56) (FIQ) will look into the wording when back Jul 23 20:25:00 looked into backupmenu Jul 23 20:25:39 it never mentions /opt specifically, it calls the fat partition "MyDocs" (where it's mentioned, i.e. in fsck), and the /home partition "OptFS" Jul 23 20:25:46 that's a 706MB .xml, I guess there may be others way to open it if Evopedia failed.. Jul 23 20:25:49 and the root one "rootfs" Jul 23 20:27:00 yeah Jul 23 20:27:06 yeah Jul 23 20:27:12 and o/ Jul 23 20:30:25 yes, I know, I've been goofing off a bit >.> I should come up with an upgrade plan, publish it, get folks to comment on it and go ahead with it by...the end of August or September or something like that Jul 23 20:33:47 ashley: for now your expertise regarding mediawiki and evopedia would already help a lot Jul 23 20:33:56 agreed Jul 23 20:34:25 nfc over here Jul 23 20:34:27 it would be really helpful Jul 23 20:34:31 alright, what's up? Jul 23 20:35:26 GogoPogo has a thing called evopedia running on his N900, and we all think it's a great plan to have wiki.m.o locally on device Jul 23 20:35:37 it's just Evopedia can read wikipedia and wiktionary-like dumps, is there a way to have mediawiki dumped the same way? Jul 23 20:37:24 sure; there a script called dumpBackup.php in the wiki's maintenance/ directory (http://www.mediawiki.org/wiki/Manual:DumpBackup.php) that can be used to create a(n) .xml dump of the wiki (admittedly it's been a long while since I've had to use it) Jul 23 20:38:23 ashley: I see from the dumps provided by Evopedia that the dumps formats should be .idx and .dat Jul 23 20:38:40 the dumps files onto your device* Jul 23 20:39:12 huh? no idea how such dumps are created, I'm aware only of MediaWiki's standard .xml dumps (generated with the script mentioned above) as well as SQL dumps Jul 23 20:40:10 http://maemo.cloud-7.de/wiki/ Jul 23 20:40:20 wikimedia dumps are like this one right? Jul 23 20:41:55 Wikimedia Foundation (the instance hosting Wikipedia etc.) offers various database dumps (http://dumps.wikimedia.org/), but the standard MW dump format is .xml; I have no idea what could be used to create .idx/.dat files and those definitely are not MediaWiki standard DB dump formats (also, MediaWiki != Wikimedia, MW is the software and Wikimedia is the instance hosting Wikipedia) Jul 23 20:43:44 I think a .xml dump file is just fine too, there must be more than one way to open and browse a .xml file on maemo.. I just believe Evopedia is a very popular application for maemo users and it would make things easier Jul 23 20:44:36 I bet as soon as you get more info about data structure of .idx and .dat, it can get converted from xml to those formats Jul 23 20:45:07 obviously .idx is an index file and .dat is the database file Jul 23 20:45:35 yes, and there's a metadata.txt file indexing the .idx files too Jul 23 20:46:15 find the howto page on evopedia homesite Jul 23 20:46:22 ok Jul 23 20:46:29 DocScrutinizer05: also, it's probably worth noting that the supermajority of "files" at http://maemo.cloud-7.de/wiki/files/ are broken -- the files with a filesize of 8.6K are actually HTML pages (MediaWiki's file description pages instead of the actual images) Jul 23 20:46:38 there jas to be some instruction how to create those .idx and .dat Jul 23 20:47:00 ashley: DAMN! Jul 23 20:47:11 I didn't do that download Jul 23 20:47:24 thanks for letting me know Jul 23 20:47:26 http://maemo.cloud-7.de/wiki/files/Wm-niels%20breet.png for example is a HTML dump of Special:FilePath and I guess so are the others with the same size Jul 23 20:47:59 but for example http://maemo.cloud-7.de/wiki/files/Wazapp-group-create.png seems to be OK Jul 23 20:48:14 no idea what method's been used to dump the files, but obviously it wasn't the ideal one :-/ Jul 23 20:48:33 sth went wrong with the job that downloaded that stuff. it been an incredible manual work Jul 23 20:49:13 some guys were creating a bashscript from special:allpages and special:files or twhatever Jul 23 20:50:01 it's from a time where we had no access to the wiki on an administrative level Jul 23 20:51:41 so any sed command or shellscript command in the created bashscript could have blown chunks on whatever particular property of the filename or file to download Jul 23 20:52:23 whatever, we now have direct access to wiki db Jul 23 20:53:16 so just you guys tell techstaff what you want to get done, and we do it for you - just don't come and ask for pink pony or "I want a .dat file, nfc how you create that" Jul 23 20:55:20 GogoPogo: do you think you can spot the documentation/howto on evopedia? Jul 23 20:56:00 there's none, only downloaded the client to contribute to dumps and see what I can make out of it Jul 23 20:56:09 but I found other useful info elsewhere Jul 23 20:57:17 wikitravel.org has an offline reader option and on its explaination article Evopedia is listed as one of the target formats. Here's what they say about it: Jul 23 20:57:25 Evopedia http://evopedia.info/ Free open source offline dump reader and browser for MediaWiki wikis using devices default web browser to display results. Target devices include Linux tablets and smartphones using Nokia Maemo OS, OpenMoko SHR, and devices and computers running Ubuntu and Arch Linux distros. Evopedia requires processed dumps available at their website or custom databases which... Jul 23 20:57:27 ...can be made with free tools Jul 23 20:57:52 so it says MediaWiki is supported by Evopedia but they're the only ones to say so Jul 23 20:58:12 so I went and check here http://meta.wikimedia.org/wiki/Data_dumps Jul 23 20:58:21 https://mg.pov.lt/maemo-irclog/%23maemo.2010-10-18.log.html#t2010-10-18T19:34:40 Jul 23 20:59:20 Linux tablets and smartphones using Nokia Maemo OS, OpenMoko SHR -- \o/ :-D Jul 23 21:00:30 none seems to fret about the formats so I suppose it's an easy conversion to be done from .xml do .dat Jul 23 21:01:09 http://meta.wikimedia.org/wiki/Data_dumps/Dump_format Jul 23 21:05:22 btw I'm currently downloading the .xml dump you Doc linked to me before and see if it works Jul 23 21:05:44 I keep searching the web for a solution in the meantime Jul 23 21:06:22 o7ywsw but the wiki dump isn't zip files 19:34 Jul 23 21:06:24 o7ywsw it is .idx and .dat files....... 19:35 Jul 23 21:06:25 o7ywsw what's the deal Jul 23 21:06:35 I read it Jul 23 21:06:58 I can't find any reference to .idx and .dat on http://meta.wikimedia.org/wiki/Data_dumps/Dump_format Jul 23 21:07:24 nor did I, that's why I suppose it's a problem that can be solved easily, or at least I hope Jul 23 21:10:52 o7ywsw claimed the downloads consisted of .idx and .dat files in 2010 Jul 23 21:11:07 I don't get it Jul 23 21:11:48 I can't understand Evopedia's choice Jul 23 21:12:12 https://github.com/evopedia/evopedia_python/blob/master/dumpscripts/dumplib.sh Jul 23 21:12:18 everything I come across regarding wikis dumps talks about xml Jul 23 21:13:41 zip -1 wikipedia_"$LANG"_"$dumpdate".zip *.idx *.dat metadata.txt Jul 23 21:15:05 I can't work out anything useful from the code.. do you? Jul 23 21:21:31 wikimedia dumps come in static html format too, one file for each article and no .xml, only and index.html file Jul 23 21:23:42 I guess they would be good too, Evopedia start the browser on local host to the port 8080 and grabs the data from the databases. Plain html should be no hard to open but I fear it will mess up the wikis index inside the application Jul 23 21:24:14 are there any HW-restraints against Linux-compatible Laptops, like graphics or smth.? Jul 23 21:24:43 ...eSATA, FW etc? Jul 23 21:25:48 yes, they are downloading the xml.gz and sql.gz files (line 12 .. 18), then at some place import stuff to a mysql db, and then there's also that magical grep -v '' | "$SCRIPTDIR"/xml2sql -o "$IMPORTTEMPDIR" -v Jul 23 21:30:15 $LIBDIR/datafile_storage.py --convert "$DESTDUMPTEMPDIR/$LANG/" "$dumpdate" "$LANG" "http://$LANG.wikipedia.org/wiki/" Jul 23 21:30:34 yeah I saw that but how can it help? Jul 23 21:31:08 I want to set up an affordable (used) dualboot Win7/Linux Laptop with decent HW. I like the Lenovos, but they're too expensive new Jul 23 21:31:11 it seems this datafile_storage.py creates the .idx and .dat files? Jul 23 21:31:21 yes I'm looking for it Jul 23 21:32:10 hi Jul 23 21:32:12 win7mac: this is not a desktop computer channel, sorry.. Jul 23 21:32:17 just got python issue again Jul 23 21:32:23 so it wasn't bp-related Jul 23 21:32:34 so I just took a look into pymaemo-optify Jul 23 21:32:41 and rcS-late respectively Jul 23 21:33:12 DocScrutinizer05: https://github.com/evopedia/evopedia_python/blob/master/evopedia/evopedia/datafile_storage.py Jul 23 21:33:28 it seems like pymaemo-optify creates 3 relevant files (and some documents in /usr/share/doc) Jul 23 21:33:42 https://github.com/evopedia/evopedia_python Jul 23 21:34:02 one in /etc/default containing configuration (the directores being bindmounted and where the mount actually is) Jul 23 21:34:25 hi all, not specifically a maemo question, but has anyone here used this: http://elektranox.org/n900/index.html ? Jul 23 21:34:36 HRH_H_Crab: i have Jul 23 21:34:41 one in /etc/event.d which starts a daemon (py-optifying is done by a daemon script..?) Jul 23 21:34:46 Sicelo: did you manage to get wifi working? Jul 23 21:34:51 yes Jul 23 21:34:59 everything, even gprs :P Jul 23 21:35:03 i am using wpa-ssid Jul 23 21:35:07 wpa-psk Jul 23 21:35:19 but the little blighter just doesnt seem to want to associate Jul 23 21:35:24 not sure what i might be missing Jul 23 21:35:34 anyway, that's deprecated now.. you're better off with a newer kernel, eg Nemo kernel Jul 23 21:36:11 hmmm Jul 23 21:36:25 you have the firmware in the right place? Jul 23 21:36:38 yep Jul 23 21:36:44 the device is there Jul 23 21:36:54 i have wlan0 phonet0 all that jazz Jul 23 21:36:58 so let's see here Jul 23 21:37:08 i just cant seem to write an /etc/network/interfaces that works Jul 23 21:37:23 wpasupplicant? Jul 23 21:37:29 ive not tried that Jul 23 21:38:02 you can't use wpa-ssid, wpa-psk if you don't have supplicant Jul 23 21:38:26 it works fine on my zaurus'es Jul 23 21:38:54 well, i'm no expert.. at least that's how i understood it. Jul 23 21:39:22 you absolutely sure your Zauruses don't have supplicant? maybe was bundled .. Jul 23 21:39:51 GogoPogo: https://github.com/evopedia/evopedia_python/tree/master/mediawiki-1.14.0-evopediamod sounds good, eh? Jul 23 21:40:35 Sicelo: yeah Jul 23 21:40:40 grep -q " $d " /proc/mounts || mount --bind $OPT_DIR/$d $d - this is how the mounting works, i.e. it first checks if they're already mounted and if grep exits with errors (i.e. it isn't), mounts it Jul 23 21:40:43 i installed debian on them using debootstrap Jul 23 21:40:49 there doesn't seem to be any issues with this Jul 23 21:40:56 ive tried using wpasupplicant before Jul 23 21:40:57 DocScrutinizer05: it does... can you run it? Jul 23 21:41:12 haven't tried yet Jul 23 21:41:12 and ive always found it to be one of the most horrific programs ive ever used Jul 23 21:41:13 unless /home is mounted *after* this script run Jul 23 21:41:13 s Jul 23 21:41:16 :P Jul 23 21:41:33 but maybe this is a good opportunity to try and get to grips with it Jul 23 21:41:38 when i feel i can stomach it Jul 23 21:41:51 is it what you used when you set up wifi on your n900 on deb? Jul 23 21:42:00 GogoPogo: but for all I can tell it seems to me evopedia actively tries to dwonload whatever it needs to duplicate the whole wiki db in a local mysql Jul 23 21:42:08 this could potentially be the issue Jul 23 21:42:29 HRH_H_Crab: i believe so.. i followed all those steps to the letter. Jul 23 21:42:30 it doesn't need you to download any huge files manually Jul 23 21:42:48 DocScrutinizer05: I thought of altering that line in python mount scripts slightly Jul 23 21:44:01 I don't get it, what's the advantage with it? Jul 23 21:44:14 grep -q " $d " /proc/mounts || mount --bind $OPT_DIR/$d $d 2>/home/user/pymounts_failed Jul 23 21:44:25 DocScrutinizer05: would this do what I want? Jul 23 21:44:44 i.e. send any eventual errors with the mount to pymounts_failed as text Jul 23 21:44:59 HRH_H_Crab: it was over a year ago.. so i don't quite remember everything. why don't you try a newer debian with a newer kernel..? or even Arch? Skry made some really great progress there Jul 23 21:45:11 or would that script get overwritten in some way? Jul 23 21:45:35 or worse, error out and make the issue worse Jul 23 21:45:38 FIQ|n900: not exactly, use /var/log/ Jul 23 21:45:51 not /home/* Jul 23 21:46:09 oh Jul 23 21:46:11 but other than that? Jul 23 21:46:14 DocScrutinizer05: also Jul 23 21:46:19 looked into rcS-late Jul 23 21:46:20 looks good to me Jul 23 21:46:37 and btw HRH_H_Crab, i've never really use supplicant directly.. i just had it because my understanding was that it's needed for wpa* stuff to work in /etc/network/interfaces.. guess i was wrong Jul 23 21:46:42 yeah Jul 23 21:46:49 no you may be right Jul 23 21:47:09 and mounting seems to be done by that script unless part of rootfs and already mounted Jul 23 21:47:20 also use mount -v(erbose) Jul 23 21:47:27 so perhaps it's possible that rcS-late is running too late? Jul 23 21:47:53 mhm Jul 23 21:48:18 HRH_H_Crab: did you also create the udev rule as per http://elektranox.org/n900/installation/configuration.html? Jul 23 21:48:28 you'll learn from /var/log/pymounts_failed Jul 23 21:48:28 ok why is my n900 trying to access port 7275 (SUPL?) on 66-152-109-100.tvc-ip.com ... Jul 23 21:48:47 grep -q " $d " /proc/mounts || mount -v --bind $OPT_DIR/$d $d >/var/log/pyoptify_state Jul 23 21:48:55 supl.nokia.com? Jul 23 21:48:57 was thinking on settling with this then Jul 23 21:49:23 add a line echo "mounting $OPT_DIR/$d" >>>/var/log/pyoptify_state Jul 23 21:49:31 ecc3g: what supl server do you have configured? and is gps enabled atm? Jul 23 21:49:33 and you don't want to use >, use >> Jul 23 21:49:43 ah Jul 23 21:49:44 add a line echo "mounting $OPT_DIR/$d" >>/var/log/pyoptify_state Jul 23 21:49:47 dang Jul 23 21:50:48 I was trying different SUPL servers, supl.google.com supl.nokia.com and also tried my supl-proxy Jul 23 21:50:57 set -vx Jul 23 21:51:15 HRH_H_Crab: a more recent debian + regualr maemo kernel tutorial is at http://wiki.debian.org/MaemoAndSqueeze Jul 23 21:51:20 exec 2>>/var/log/pyoptify_state 1>&2 Jul 23 21:51:29 DocScrutinizer05: should we ask ashley to check out the program? Jul 23 21:51:31 at start of rcs-late Jul 23 21:51:33 I was trying to get AGPS working but wonderring why I could not get it to reach my proxy... Jul 23 21:51:51 DocScrutinizer05: exec 2 does what? Jul 23 21:51:57 DocScrutinizer05: but ok, I'll do that Jul 23 21:52:06 ecc3g: and where's the proxy? Jul 23 21:52:08 GogoPogo: I don't know if he's willing to do that. ask him anyway Jul 23 21:52:22 it's not that host :) Jul 23 21:52:34 DocScrutinizer05: pvt message you mean? Jul 23 21:52:56 temporarily running the proxy on my server... Jul 23 21:53:08 FIQ|n900: exec without any parameter is used to redirect any filehandles of the currently running shell interpreter Jul 23 21:53:25 ah Jul 23 21:53:43 fyi, supl.google & supl.nokia are broken for N900 :( Jul 23 21:53:46 set -vx echoes every line as it is executed Jul 23 21:53:49 I'm wondering if this is the SIM-card supplied SUPL server (which, incidentally, that machine is dead) Jul 23 21:54:02 DocScrutinizer05: mhm Jul 23 21:54:07 exec 1>somefile redirects output to somefile Jul 23 21:54:19 2 stderr I guess Jul 23 21:54:22 yep Jul 23 21:54:41 echo "mounting $OPT_DIR/$d" >>>/var/log/pyoptify_state is 3 > here a typo? Jul 23 21:54:48 you wrote it before Jul 23 21:55:06 yes Jul 23 21:55:11 oh makes sense Jul 23 21:55:13 DocScrutinizer05: do you think we could aim it at wiki.maemo.org and see what it comes up with? Jul 23 21:55:33 you don't need that anymore, when you redirect output of whole rcS-late Jul 23 21:55:33 > overwrites, >> appends, so I assumed it was a typo yes Jul 23 21:56:11 DocScrutinizer05: I've long since last programmed anything and even though I liked Python a lot I'm not able to help out anymore Jul 23 21:56:17 but maybe you just want to do that before the interesting part of rcS-late Jul 23 21:56:42 GogoPogo: sure we can try that Jul 23 21:57:13 if there's anyway I can help just tell me Jul 23 21:57:17 any way* Jul 23 21:57:58 ecc3g: i don't think N900 gets supl server from sim even if it's there.. Jul 23 21:58:23 FIQ|n900: probably you even rather go with your solution Jul 23 21:58:32 Hmm... it's just suspicious, I can't seem to contact my supl proxy at all... Jul 23 21:58:35 DocScrutinizer05: I may e-mail Evopedia.info and ask for their thought, is it ok? Jul 23 21:58:36 wiresharking it... Jul 23 21:58:41 it's not as bloated as what I suggested (exec) Jul 23 21:59:22 ask ashley first Jul 23 21:59:27 ecc3g: no typo maybe? Jul 23 21:59:33 GogoPogo: ask ashley first Jul 23 21:59:33 http://pastebin.com/v1MqcLLc rcS-late, http://pastebin.com/ix0ga7AG pymaemo-optify Jul 23 21:59:35 looks good Jul 23 21:59:42 wireshark indicates it is a typo... ha. Jul 23 21:59:55 DocScrutinizer05: hmm but letting rcS-late output may also hint if it's running too early/late Jul 23 21:59:58 DocScrutinizer05: ok, he's not there anyway Jul 23 22:00:03 lol ecc3g Jul 23 22:00:13 but... Jul 23 22:00:27 it doesn't look like a typo in the settings field Jul 23 22:00:29 which would explain the issues I've had with the bindmounts Jul 23 22:00:51 ashley: not here? Jul 23 22:01:13 DocScrutinizer05: unless you see any obvious typos, I'll reboot now.. worst case scenario is I messed up the syntax and maemo doesn't boot, though fixing that would be trivial with bm's root console Jul 23 22:01:35 hope so, since otherwise bootloop Jul 23 22:01:57 hmm.. now that's strange ecc3g. Jul 23 22:02:04 I haven't seen obvious typos but I haven't checked thoroughly either Jul 23 22:02:05 even if it works fine (the bindmounts), I could still check that script to see whether my modifications worked Jul 23 22:02:09 anyone happen to know what file the SUPL server setting is stored in? Jul 23 22:02:16 gconf Jul 23 22:02:22 and if they did, then I know where to look next time it fails Jul 23 22:02:42 /system/nokia/location/supl <-- Jul 23 22:02:46 ecc3g: ^^ Jul 23 22:03:08 ecc3g: gconftool -R /|less "+/supl" -j 5 Jul 23 22:03:16 [00:01:35] hope so, since otherwise bootloop Jul 23 22:03:23 bm - root console - remove my edits Jul 23 22:03:28 and reboot loop solved Jul 23 22:03:33 not too hard really :P Jul 23 22:03:44 brb reboot Jul 23 22:04:16 ecc3g: then hit n Jul 23 22:04:33 gconf looks right Jul 23 22:04:38 hmm. Jul 23 22:04:48 reboot I guess Jul 23 22:05:10 ecc3g: alternatively : gconftool -R /system/nokia/location Jul 23 22:05:58 ecc3g: which process is trying to contact that address? netstat... Jul 23 22:06:05 no clue why wireshark indicates something is trying to look up a mangled hostname... Jul 23 22:06:52 fuuu i need a non-messybox netstat :p Jul 23 22:07:06 lol, for sure Jul 23 22:07:14 ~messybox Jul 23 22:07:29 o.O Jul 23 22:07:36 reboot. Jul 23 22:07:51 ~ping Jul 23 22:07:52 extra, extra, read all about it, messybox is messy... err busybox is meant for lean scripting. Regarding all the missing options and immanent limitations (see su, passwd) 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 Jul 23 22:07:54 ~pong Jul 23 22:08:01 wtf? Jul 23 22:08:35 DocScrutinizer05: is it normal for mount -v not to output anything on a successful mount? Jul 23 22:08:44 dunno Jul 23 22:08:51 why couldn't he just install Netstat? there is a standalone pkg Jul 23 22:08:52 maybe to stadout Jul 23 22:08:56 stdout Jul 23 22:09:05 1>&2 Jul 23 22:09:09 I redirected all output Jul 23 22:09:13 mhm Jul 23 22:09:19 then I dunno Jul 23 22:09:21 for the mount command Jul 23 22:09:22 I was trying to use supl:7276 notation to try to get it to use another port, but I think that may have fubared something. Jul 23 22:09:36 ii netstat 1.60-23maemo1 Non busybox netstat in a small easy to insta Jul 23 22:09:37 also, my edits seems to have worked fine Jul 23 22:09:45 also don't forget that's messybox mount Jul 23 22:09:53 Sicelo: yep i created the udev rule Jul 23 22:09:57 checking that doc now Jul 23 22:10:07 except possibly the edit to the mounting as I dunno if it's supposed to echo anything with -v Jul 23 22:10:13 but i may as well just chroot from maemo and get wpasupplicant installed and see what gives... Jul 23 22:10:22 DocScrutinizer05: hm just tried mount -v w/o any parameters Jul 23 22:10:36 it didn't output anything besides what's given on just "mount" Jul 23 22:10:50 so no output might be normal then Jul 23 22:11:01 Sicelo: will conflict with messybox Jul 23 22:11:15 It did output my date which I mostly added to reveal if it was ran before or after rcS-late Jul 23 22:11:25 what conflicts wit messybox DocScrutinizer05? Jul 23 22:11:29 dpkg divert netstat is the solution I guess Jul 23 22:11:45 (and it appeared at the very bottom - so rcS seems to have been ran first) Jul 23 22:11:47 hmm Jul 23 22:11:59 So, stuff works, great Jul 23 22:12:06 now I at least know what to do next time it fails Jul 23 22:12:16 which may finally tell what the error is Jul 23 22:13:07 add that echo $PYDIR/$d line before mount Jul 23 22:14:03 DocScrutinizer05: also, is rendurancy in pymaemo-optify intended? Jul 23 22:14:03 even echo "mounting py: `date` pydir: $PYDIR/$d ." Jul 23 22:14:13 looks like you can't 'divert' busybox netsta, because it's not installed on its own DocScrutinizer05. Jul 23 22:14:29 Sicelo: possibly Jul 23 22:15:07 it sets a variable containing which directory to bindmount to at start of init.d Jul 23 22:15:11 cya folks, me needs a beer out in the big black bluebox now Jul 23 22:15:28 and then proceed to source /etc/default/pymaemo-optify (which does this as well) lol Jul 23 22:15:32 it's just a symlonk to busybox.. so when installing netstat standalone, it deletes the link and installs a 'real' netstat. if you ever uninstall it, the 'real' binary is deleted, and link to busybox restored Jul 23 22:16:22 Sicelo: that would be awesome, if only busybox wouldn't PROVIDE:netstat Jul 23 22:16:37 so with apt you will run into troubles for sure Jul 23 22:17:03 gtg, will get back to you tomorrow Jul 23 22:17:06 i don't see how. Jul 23 22:17:07 with dpkg I dunno, but definitely will cause inconsistencies next time you run apt Jul 23 22:18:20 Sicelo: that's why all that stuff is installed to bin/gnu/* and symlinked from /bin/g* -> /bin/gnu/* Jul 23 22:19:21 * FredrIQ|n900 altered his PATH to prefer /usr/bin/gnu Jul 23 22:19:26 Sicelo: yep you were right it does need wpasupplicant Jul 23 22:19:29 thanks! Jul 23 22:19:37 Sicelo: actually I guess it's a pkg like busybox-alternatives or whatever that PROVIDEs netstat Jul 23 22:19:51 FredrIQ|n900: me too :-D Jul 23 22:20:06 i have this netstat.. and not any other version of busybox. Jul 23 22:20:16 anyway, i understand your point Jul 23 22:20:26 busybox-symlinks-net-tools provides netstat Jul 23 22:20:48 :nod: Jul 23 22:21:38 now if that's not forced by mp-fremantle-pr then you can uninstall it and install proper netstat. otherwise it will uninstall whole mp-fremantle-pr Jul 23 22:22:06 that's what I actually did to get procps Jul 23 22:22:19 since messybox ps is unbearable either Jul 23 22:22:45 can't you just do it the same way as with coreutils-gnu? Jul 23 22:22:48 since then no more updates, and I'm actually happy about it Jul 23 22:23:18 DocScrutinizer05: sorry, something came up; I'm here now; checked out GogoPogo's link, doesn't seem to be anything new or unseen, just the dumpHTML extension (http://www.mediawiki.org/wiki/Extension:DumpHTML) with some custom patches + a bunch of original shell scripts; the problem with dumpHTML is is that it's been broken (AFAIK) for a while and nobody's cared enough to do any serious work on it and there really isn't an alternative way to create s Jul 23 22:23:18 tatic representations of MediaWiki pages Jul 23 22:23:28 allthough if a lot of packages is going to provide alternatives to busybox binaries, /usr/bin may end up like Program Files@windows after a while lol Jul 23 22:23:38 FredrIQ|n900: sure you could package procps in a way it doesn't conflict with messybox, like *-gnu does Jul 23 22:24:01 anyway, my netstat doesn't come from the pkg FredrIQ|n900 mentioned either Jul 23 22:24:52 ashley: thanks for the info :-) Jul 23 22:25:09 ~$ dpkg -S $(which netstat) Jul 23 22:25:11 netstat: /usr/bin/netstat Jul 23 22:25:39 Sicelo: I dunno Jul 23 22:26:04 IroN900:~# netstat -V Jul 23 22:26:06 net-tools 1.60 Jul 23 22:26:21 Sicelo: the one I mentioned was /bin/netstat though - there's nothing in /usr/bin/ Jul 23 22:26:28 maybe messybox-basic has no netstat at all Jul 23 22:26:34 database dumps are somewhat of a pain point...generating SQL or XML dumps is easy, but that's about it; obviously SQL dumps -- at least full dumps -- aren't suitable for public distribution, which leaves us with the XML ones...but those require a working MediaWiki instance to "read" the contents of such a dump, which isn't exactly ideal either Jul 23 22:27:19 lrwxrwxrwx 1 root root 7 Dec 9 2012 /bin/netstat -> busybox Jul 23 22:27:32 seems like the busybox version maemo has do have netstat Jul 23 22:27:38 i know the one FIQ|n900 mentions. anyway.. back to ecc3g .. any info for us.. which process if trying to access that site? Jul 23 22:27:51 ashley: aiui evopedia uses exactly mysql to get a full mediawiki-lookalike Jul 23 22:28:31 at least those python scripts do Jul 23 22:28:38 aiui Jul 23 22:29:01 https://github.com/evopedia/evopedia_python/tree/master/mediawiki-1.14.0-evopediamod Jul 23 22:29:13 why juse evopedia over mediawiki? Jul 23 22:29:14 *use Jul 23 22:29:53 I have nfc what *is* evopedia Jul 23 22:30:00 except for above link Jul 23 22:30:57 https://github.com/evopedia/ <- if that's evopedia then it uses a mysql db and stores wiki data there to somehow display it on browser locally later Jul 23 22:31:14 it has even a function to generate random-page Jul 23 22:31:42 evopedia is an offline wikipedia etc reader Jul 23 22:31:43 maybe somebody needs that nonsense on their local wiki Jul 23 22:32:07 nox-: exactly Jul 23 22:32:11 * nox- at least finds it useful (only the dumps are kinda old) Jul 23 22:32:19 nox-: and we want to feed it with wiki.m.o data Jul 23 22:32:27 ah Jul 23 22:33:15 and aiui it has all the functions to build his own db from dumps it downloads from http://wikipedia.org via wget Jul 23 22:33:28 s/his/its/ Jul 23 22:33:29 DocScrutinizer05 meant: and aiui it has all the functions to build its own db from dumps it downloads from http://wikipedia.org via wget Jul 23 22:33:31 http://dumpathome.evopedia.info/dumps/finished/en Jul 23 22:34:21 that's some really big size differences on every dumps Jul 23 22:34:49 yeah wp is growing fast... Jul 23 22:35:09 http://dumpathome.evopedia.info/client Jul 23 22:35:12 note how the newest is 1/20 of a 1 year older dump Jul 23 22:35:32 patch that to point to wiki.m.o Jul 23 22:35:34 done Jul 23 22:35:36 cheers Jul 23 22:35:39 * DocScrutinizer05 heads out Jul 23 22:35:44 FIQ|n900, Wikivoyage? thats not wp :) Jul 23 22:36:11 oh misread lol Jul 23 22:37:01 if somebody is willing to patch it, I'm willing to run it on maemo.org locally and provide the dump Jul 23 22:37:58 so no need for you to rape wiki.m.o and your DSL line and local PC Jul 23 22:40:25 doesn't w.m.o run mediawiki? Jul 23 22:42:51 indeed it does Jul 23 22:43:09 http://wiki.maemo.org/Special:Version Jul 23 22:43:28 shall not be hard to patch that client, ah? Jul 23 22:44:05 depends on how it gets the downloads Jul 23 22:44:21 I haven't looked into it Jul 23 22:44:28 neither have I Jul 23 22:44:36 but it *should* be trivial to patch Jul 23 22:44:49 if patching is needed at all Jul 23 22:45:05 URL needs patching for sure Jul 23 22:45:12 unless that's a parameter Jul 23 22:45:24 if it supports importing a local dump, I guess you could simply have a tmo dump instead Jul 23 22:45:39 and I don't see how that would fail Jul 23 22:46:32 unless it *really tries to* avoid "unofficial" wikipedia variants Jul 23 22:47:01 which would be silly Jul 23 22:56:34 Source Code Jul 23 22:56:35 github: http://github.com/evopedia/evopedia_qt (from version 0.4 on) Jul 23 22:56:37 Older versions: http://github.com/evopedia/evopedia_python Jul 23 22:57:26 FIQ|n900: the client, however bloated, has a mediwiki subdir Jul 23 22:59:22 given the tiny size of w.m.o (compared to wikipedia) I don't see any problem to run this client on your local PC and download w.m.o and ""create Wikipedia database dumps for use in the offline Wikipedia reader evopedia."" Jul 23 23:00:45 http://wiki.maemo.org/Evopedia Jul 23 23:04:43 4. start the dump process Jul 23 23:04:45 call createdump.sh Jul 23 23:04:46 (Don't be too optimistic though. After part of the day the database is filled with several gigabytes of downloaded wikipedia content, but something is broken and the final dump gives no results.) Jul 23 23:05:15 great, isn't it? Jul 23 23:05:29 sorry, have to run now Jul 23 23:08:00 sounds funny when they try to download a several terrabytes dump to a VFAT fs like Mydocs Jul 23 23:11:15 why do you want to do this anyway? Jul 23 23:11:28 @ making maemo wiki available offline Jul 23 23:12:01 because it's a sane thing to have the docs at hand when your internet acccess breaks Jul 23 23:12:19 hmm point taken Jul 23 23:12:34 but isn't there an offline maemo wiki reader available already? Jul 23 23:12:34 since, where else than in wiki (or tmo) would you look for the solution to your problem? Jul 23 23:13:00 I dunno, never heard of any Jul 23 23:13:29 o/ o/ o/ o/ o/ o/ o/ o/ Jul 23 23:33:25 huh Jul 23 23:33:39 just went and was going to look for missing symlinks Jul 23 23:34:01 found this: lrwxrwxrwx 1 root root 31 2013-07-21 15:22 python-support.pth -> ../../pymodules/python2.5/.path - why is this a relative link? Jul 23 23:34:19 oh well Jul 23 23:34:32 perhaps it's supposed to be like that lol Jul 23 23:35:00 (that wouldn't explain the issue I had at all as it's irrelevant, just wondered why it was like that) Jul 23 23:36:14 it's only broken in the sense that following it directly (from /opt) would fail, allthough it's not really necessary to make it relative, is it? Jul 24 00:22:43 well there we go, now *that* was fixed Jul 24 00:53:20 http://privatepaste.com/695a4e8d8b Jul 24 00:54:06 HMMM älter als 8 Tage (Das Alter ist 204,3 Tage Jul 24 00:54:17 * DocScrutinizer05 starts updatedb Jul 24 00:56:05 hm... Jul 24 00:57:21 ran into delayed (not failed) bindmounts, but not sure why it stalls Jul 24 00:57:25 hehe quite interesting to see how updatedb made my drive acticity indicator LED shine red Jul 24 00:57:45 # Redirect stderr into the log file (don't forget to restore when exiting!) Jul 24 00:57:46 exec 2 >>"$protocol" Jul 24 00:58:10 then my own log file didn't show anything for a while Jul 24 00:58:21 checked the log file in question that it mentions Jul 24 00:58:23 wtf are you green now, while [2013-07-24 02:22:43] well ther... was red? Jul 24 00:59:45 DocScrutinizer05: possibly because I had the nick FredrIQ|n900 before Jul 24 00:59:54 http://wstaw.org/m/2013/07/24/plasma-desktopph3743.png Jul 24 00:59:59 and xchat didn't refresh until rejoin Jul 24 01:00:41 oh yeah had FredrIQ|n900 at first during that session Jul 24 01:01:08 IroN900:~# time updatedb Jul 24 01:01:10 real 1m22.686s Jul 24 01:01:11 user 0m48.125s Jul 24 01:01:13 sys 0m15.320s Jul 24 01:02:21 I don't see any difference between (red) and (green) Jul 24 01:03:56 "# Check if /home/pot is mount bound to /opt" pot :D Jul 24 01:03:59 DocScrutinizer05: well Jul 24 01:04:23 what Jul 24 01:04:33 color do I have now? Jul 24 01:04:35 * DocScrutinizer05 waits for checkered FIQ|n900 Jul 24 01:04:42 RED Jul 24 01:04:51 that was my point :P Jul 24 01:04:58 lol checkered Jul 24 01:05:24 nfc how konversation picks nick colors Jul 24 01:05:36 based on nick Jul 24 01:05:46 keeps same color on renick Jul 24 01:05:48 I'd have used md5sum%8 or sth Jul 24 01:05:51 before user rejoins Jul 24 01:05:54 on nick Jul 24 01:06:00 yeah Jul 24 01:06:05 that's what it does Jul 24 01:06:14 well not necessarily md5sum Jul 24 01:06:22 but some algoritm based on the nick Jul 24 01:06:38 however it keeps the same color if the user changes nick Jul 24 01:06:45 oooh I seee Jul 24 01:07:04 you say it depends om what nick you *joined* the channel Jul 24 01:07:12 (before the user leaves and come back with this new nick) Jul 24 01:07:13 yeah Jul 24 01:07:35 yes, it seems it does Jul 24 01:07:39 :-D Jul 24 01:08:29 anyway, that logfile didn't really help explaining why bindmounts was delayed Jul 24 01:08:29 so (RED) been < Jul 24 01:08:36 however - if it fails again Jul 24 01:08:42 like it did earlier todat Jul 24 01:08:45 *today Jul 24 01:08:51 perhaps I can figure it out Jul 24 01:09:05 yes Jul 24 01:09:33 make that "time mount bla brubble" Jul 24 01:10:10 also you might want to do "cat /proc/mounts" before Jul 24 01:10:40 if nuttin else helps: strace mount Jul 24 01:10:44 perhaps I should place a date before it switch logfile Jul 24 01:10:48 DocScrutinizer05: well Jul 24 01:11:02 the issue was not with the mount Jul 24 01:11:02 though I'm afraid messybox will neuter that Jul 24 01:11:14 lol Jul 24 01:11:16 but anyway Jul 24 01:11:45 don't forget the timestamp parameter for strace! Jul 24 01:11:59 basically, rcS-late switch to maemo-optify-boottime.log logfile (for exec 2) at a certain part in the script Jul 24 01:12:09 this phase took 2 minutes Jul 24 01:13:02 I'll place a date indicator at this part, and a real seperator between sessions (there's none making it hard to figure out where the current session is) in that other logfile Jul 24 01:13:11 yeah optify-boottime may take some while Jul 24 01:13:38 date +%x is convenient Jul 24 01:14:30 and "trap 'echo ###############' exit" Jul 24 01:14:49 though nfc if that will work in friggin upstart scripts Jul 24 01:15:31 seems rcS-late is in fact a upstart script, according to your pastebin Jul 24 01:16:06 uh huh? Jul 24 01:16:10 that was weird Jul 24 01:16:49 [2013-07-23 23:59:33] http://pastebin.com/v1MqcLLc rcS-late Jul 24 01:16:51 was going to edit rcS-late and search for exec twice, and the thing segfaulted lol Jul 24 01:17:06 (nano) Jul 24 01:17:22 description "late rcS actions" Jul 24 01:17:24 Jul 24 01:17:25 start on started hal Jul 24 01:17:27 stop on stopping hal Jul 24 01:17:28 Jul 24 01:17:30 console none Jul 24 01:17:31 Jul 24 01:17:33 script Jul 24 01:17:34 thats upstart fuckup Jul 24 01:18:32 WOW I managed to post empty lines Jul 24 01:18:47 can't recall to have done this ever before Jul 24 01:18:52 there's 1 space in them Jul 24 01:18:57 Good morning I guess? :) Jul 24 01:19:10 irc doesn't allow empty lines to be posted Jul 24 01:19:12 goo (UGT) morning Jul 24 01:20:11 good* even Jul 24 01:20:18 ~ugt Jul 24 01:20:18 ugt is, like, Universal Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. http://www.total-knowledge.com/~ilya/mips/ugt.html Jul 24 01:21:30 FFS maemo will be my epithaph Jul 24 01:22:17 maybe NCIS on TV Jul 24 01:22:20 cya Jul 24 01:25:17 okay,, Good (UGT) morning Doc...:) Jul 24 01:28:08 hm apt still have localization issues in newest cssu Jul 24 01:30:13 orly? Jul 24 01:30:53 maemo and I18n/l10n seems to not to go together well Jul 24 01:32:12 http://privatepaste.com/842c222548 Jul 24 01:32:22 any questions, Watson? Jul 24 01:34:17 compared to that apt-get http://privatepaste.com/166b40b011 looks awesome Jul 24 01:36:02 https://bugs.maemo.org/show_bug.cgi?id=8291 Jul 24 01:36:04 04Bug 8291: Enable CONFIG_LOCALE_SUPPORT in busybox Jul 24 01:36:13 DocScrutinizer05: not as in a mess between swe/en Jul 24 01:36:21 but rather bogus confirmation requests Jul 24 01:36:31 asking Y/n but in fact only allows j/n Jul 24 01:36:55 I'm searching for the $LANG where "y" means "no" Jul 24 01:36:59 mentioned this on several occasions in here and also ended up posting it in cssu with lack for better place Jul 24 01:37:09 y as in yes Jul 24 01:37:43 but anything else than yes (i.e. anything else than j in swedish locale) aborts Jul 24 01:37:52 so y aborts Jul 24 01:38:03 I'm a fan of hardcore: either use LANG=C, or for maemo use LANG=en_EN Jul 24 01:38:09 even though it asks specifically for y/n Jul 24 01:38:19 *this* is the issue Jul 24 01:38:39 asking for j/n and answering y would abort, is something I would be ok with (annoying but acceptable) Jul 24 01:38:52 asking for y/n and answering y doing an abort is not Jul 24 01:38:59 who's not capable to read english manpages is doomed anyway in IT Jul 24 01:39:17 also just ran into this lovely bug in nano https://bugs.maemo.org/show_bug.cgi?id=8285 Jul 24 01:39:18 04Bug 8285: Always crashes when using search in "nano" (ctrl+w) on device Jul 24 01:39:34 nano wtf Jul 24 01:39:42 nano is for sissies Jul 24 01:39:43 nano ftw Jul 24 01:39:51 kinda annoying but at least there's a straight-forward solution Jul 24 01:40:25 DocScrutinizer05: full emacs isn't available on phone Jul 24 01:40:38 either you use vi or emacs, or you're a quicje-eater like me and use mcedit Jul 24 01:40:51 quiche Jul 24 01:41:06 ~real-programmer Jul 24 01:41:17 ~useless Jul 24 01:41:18 there's qemacs but didn't really like that clone Jul 24 01:41:18 * infobot starts crying and hides from docscrutinizer05 in the darkest corner of the room. :( Jul 24 01:41:32 real programmers use butterflies, np Jul 24 01:42:26 ~real-programmer is http://www-personal.edfac.usyd.edu.au/staff/souters/Humour/Real.Programmer.Stories.html Jul 24 01:42:26 okay, DocScrutinizer05 Jul 24 01:43:33 uhhhhhhh Jul 24 01:43:37 ~real programmer is see real-programmer Jul 24 01:43:38 okay, DocScrutinizer05 Jul 24 01:43:49 whoever made nano-2.0.7 as a solution didn't make the package conflict with nano ... Jul 24 01:44:02 lol Jul 24 01:44:06 w/e Jul 24 01:44:20 fine, since nano prolly is messybox Jul 24 01:44:30 which is worse, nano or messybox-vi :D Jul 24 01:44:38 it's not Jul 24 01:44:44 it's the same binary anyway Jul 24 01:45:44 btw how come some packages' install say that they can't be certified? Jul 24 01:45:47 cosidering there's nano and even pico, I ponder coding an editor called peta Jul 24 01:46:19 lol Jul 24 01:46:33 or micro Jul 24 01:46:41 as a nano successor Jul 24 01:46:45 femto? Jul 24 01:46:46 pico was the propriatory version that came with pine package. nano was just a free clone based off pico. Jul 24 01:46:53 atto Jul 24 01:47:01 zepto Jul 24 01:47:27 Planck-quantum Jul 24 01:47:31 hooray Jul 24 01:47:37 now I can actually search Jul 24 01:48:35 I'll not use editors smaller than milli Jul 24 01:49:09 t-EXA Jul 24 01:49:22 awesome name Jul 24 01:49:52 I started *ahem* BASH scripting with pico (there was no such thing known as nano back then). I saw another person using vi which had noticeably cooler UI including syntax highlighting. I'm certain nano still wouldn't even have syntax highlighting. Jul 24 01:49:55 eight megs and constantly swaps. Jul 24 01:50:27 nano is a hoax Jul 24 01:50:36 80 Jul 24 01:50:53 800 Jul 24 01:51:15 I rather use sed than nano Jul 24 01:51:28 vi/m would easily point out if there was a syntax error (if it knows the language you are {typ,writ,edit}ing in), nano wouldn't even do things like that. Jul 24 01:51:28 HELL, even ed Jul 24 01:51:58 * DocScrutinizer05 idly wonders if messybox has a ed plugin Jul 24 01:52:34 ed: applet not found Jul 24 01:52:36 LOL Jul 24 01:52:41 thought as much Jul 24 01:54:18 vi is the only editor I know of that's *certified* to work on lines of mega-chars length without a single CR Jul 24 01:55:19 everybody else took the cheating easy shortcut to define a max-line-length Jul 24 01:55:40 often 1024 chars Jul 24 01:56:25 if you exceed that, $editor will either segfault, blow chunks, or insert linebreaks on its own Jul 24 01:56:50 ok, accidentally got a loose battery and device did a hard shutdown Jul 24 01:57:08 looked like that ;-P Jul 24 01:57:18 and ext3 partition went corrupt... gg Jul 24 01:57:27 FFS Jul 24 01:58:08 Why isn't the battery cover even installed? Jul 24 01:58:15 http://pastebin.com/q2Tb44RZ seems kind of minor, but still Jul 24 01:58:25 psycho_oreos: of course it is Jul 24 01:59:02 but my battery set isn't exactly stable - merely opening the cover and holding n900 upside down will make it fall out instantly Jul 24 01:59:22 FredrIQ|n900, >_> and the battery managed to slip out of its contacts and somehow landed in the battery case? Jul 24 01:59:54 so occasionally (rarely) if exposed to sudden moves, battery shorts temporary Jul 24 02:00:02 WTF are "fast symbolic links"? Jul 24 02:00:03 Ideally you should have tightened those battery "prongs". Jul 24 02:00:20 psycho_oreos: no idea how to do this Jul 24 02:00:41 it has been like this for a while now, wasn't even aware it was easily fixable Jul 24 02:00:46 DocScrutinizer05: no idea Jul 24 02:00:59 FredrIQ|n900: bend out ypur spring levers! Jul 24 02:01:22 NAO! Jul 24 02:01:31 FredrIQ|n900, you basically need something like a needle and you push towards in each other for each part. Jul 24 02:01:57 bend them out so far yu hardly can insert battery Jul 24 02:02:22 DocScrutinizer05: ok will look into this a more proper time (4am atm) Jul 24 02:02:33 same TZ Jul 24 02:02:47 except I got my private TZ Jul 24 02:02:58 which is fuzzy Jul 24 02:03:02 hmm Jul 24 02:03:12 will try to fsck via BM Jul 24 02:03:26 this seems as a very minor issue so it should be done w/o issues Jul 24 02:03:32 ...hopefully Jul 24 02:03:34 brb Jul 24 02:03:43 yeah. one block (un)claimed seems fixable Jul 24 02:04:25 Should be thankful that its a journaled FS, if it wasn't.. lol. Jul 24 02:04:37 it's NOT Jul 24 02:05:00 journaling on a storage that writes 4MB pages is... brainfuck Jul 24 02:05:38 o.O by default (according to wikipedia that is) its noted as a journal fs. Jul 24 02:06:20 JFFS is not journaling, ext3 is and better be not Jul 24 02:07:07 nfc what fs had that missing block for FredrIQ|n900 Jul 24 02:07:46 fsck on NAND is brainfuck either Jul 24 02:08:10 s/nand/jffs/ Jul 24 02:08:35 ~botsnack Jul 24 02:08:35 DocScrutinizer05: :) Jul 24 02:08:40 hm Jul 24 02:09:01 it says within BM's root console that the fs is clean Jul 24 02:09:16 perhaps those errors show up due to the fact that the fs is mounted? Jul 24 02:09:24 (fsck was done with -nv) Jul 24 02:09:31 Ahh yes my bad, the rootfs is ubifs. I was thinking of /home which is ext3. Jul 24 02:09:57 ubi or jffs, no diff Jul 24 02:10:17 Yup. Jul 24 02:10:22 DocScrutinizer05: should I ignore the issue it shows given that BM fsck show no issues? Jul 24 02:10:56 except friggin ubi can NOT handle "small" volumes like "initrd" mtd Jul 24 02:11:28 ran it several times though, always same output (allthough once it displayed more issues, probably due to some temporary hiccup due to mounted fs) Jul 24 02:11:39 FIQ|n900: I'd assume when it doesn't show it got fixed Jul 24 02:12:21 DocScrutinizer05: yeah, though it still show up within maemo with that command :p Jul 24 02:12:31 and fsck never did anything within BM in first place Jul 24 02:12:35 that's a bug then Jul 24 02:12:36 it just said it was clean Jul 24 02:13:18 fsck has a plethora of cmdline options Jul 24 02:13:26 try verbose+++ Jul 24 02:14:04 what's the suspect volume? Jul 24 02:14:18 / ? Jul 24 02:14:20 /home Jul 24 02:14:24 mhm Jul 24 02:14:53 speaking of / Jul 24 02:14:59 can you even fsck that? Jul 24 02:15:12 ubifs Jul 24 02:15:48 well, run a fsck against that volume in BM, with all the firearms loaded, live and unlocked Jul 24 02:15:57 no Jul 24 02:16:18 fsck of ubi/jffs is a brainfart Jul 24 02:16:23 what if it gets corrupt? Jul 24 02:16:48 it either auto-recovers or you're fux0red Jul 24 02:16:57 mkay Jul 24 02:17:03 see man ubifs Jul 24 02:17:21 no manpage for it apparently lol Jul 24 02:17:39 or jffs for that matter Jul 24 02:18:42 DocScrutinizer05: btw Jul 24 02:18:46 http://en.wikipedia.org/wiki/UBIFS Jul 24 02:18:48 if battery dies Jul 24 02:18:55 is that a clean shutdown? Jul 24 02:19:04 not a excellent starting point, but... Jul 24 02:19:12 yes Jul 24 02:19:38 forced by os yes, but still clean Jul 24 02:19:40 ok Jul 24 02:19:50 since battery can provide power down to 2.5V, but shutdown happens way earlier (if bme is running) Jul 24 02:20:20 noes, no possibility of using all capacity! Jul 24 02:21:59 an unclean/hard shutdown would be one that's triggered by hw/pmu for low voltage, without even kernel taking care of the signals bouncing off every wall Jul 24 02:22:00 DocScrutinizer05: just for curiousity, do "system upgrades" override this decision? given the potential damage if it shuts down I mean Jul 24 02:22:27 hm? Jul 24 02:22:33 well Jul 24 02:22:37 OTA updates Jul 24 02:22:43 cssu and such Jul 24 02:22:48 pr1.x Jul 24 02:23:07 you mean if system acts different regarding shutdown on ow batvolt while upgrading/flashing? Jul 24 02:23:21 low* Jul 24 02:23:26 yeah Jul 24 02:23:41 I'm not sure, but probably yes Jul 24 02:24:11 I dont think a running flashing process gets aborted on lowbat Jul 24 02:24:36 perhaps it puts it in a state that makes resuming the update possible once booted again (allthough this is possibly done automatically) Jul 24 02:24:51 once it started, system should try to finish it by all means Jul 24 02:25:02 yeah Jul 24 02:25:21 like 911 calls Jul 24 02:25:26 shouldn't the system try to put failsafes in place anyway though? Jul 24 02:25:48 the only calls allowed on overtemperature and low battery or too cold battery Jul 24 02:25:53 like a backup binary/script running on boot resuming it immediately on power fail Jul 24 02:25:58 it does Jul 24 02:26:20 you can't resume Jul 24 02:26:23 Anyone familiar with compiling for thumb? Jul 24 02:26:45 you just can refuse to even start flashing when there's no safety headroom Jul 24 02:27:06 that's of course true :p Jul 24 02:27:07 freemangordon is familar with thumb stuff Jul 24 02:27:38 he's not in the channel at the moment I don't think Jul 24 02:28:22 I'm wondering how you can tell if an executable is compiled for thumb or not. the file command gives the same output generally Jul 24 02:28:47 OK, my "nanny" ordered I shall ask for a job. So: does any of you have a job for me? preferably related to maemo maintenance, community moderation, hw-review&QA? Jul 24 02:30:02 shawnjefferson: you can't simply tell for sure, the only way to tell apart thiumb binaries from ARM binaries is by looking at impair jump addr Jul 24 02:31:15 so teh cmdline I came up with had sth like "grep [o-9abcdef]*[13579bdf]" Jul 24 02:31:43 you can mow my lawn for $1/h if you like Jul 24 02:32:13 (jk) Jul 24 02:32:24 tell me your garage account so I can find the lawn Jul 24 02:32:34 lol Jul 24 02:34:15 I'll eventually stop managing/maintainig maemo.org infra when I don't have any source of income that allows me to do that, either by no income at all or by no time at all Jul 24 02:35:28 otoh when sufficient donations or a contract with salary gives me opportunity to continue to dedicate 24/7 of my time to maemo, I for sure will prefer that option Jul 24 02:36:32 3 weeks ago I though what an easy life that would be when every maemo.org user would pay 5ct / month for me to make a living Jul 24 02:37:01 :P Jul 24 02:37:45 make a f2p game, instant rich allthough people will hate you Jul 24 02:39:00 people will hate me when I stop keeping maemo.org et al alive, though only for 12h until they forget about me Jul 24 02:39:24 hahah :P Jul 24 02:39:30 @ the latter Jul 24 02:41:46 sure, you say "why don't you work in a job at McDonalds for 10h, and dedicate the rest of your time to maemo.org, like everybody else?" - simple answer: I can't Jul 24 02:42:34 McD doesn't like me being late due to fixing some urgent stuff in maemo Jul 24 02:43:19 so, as hard as it may sound, this is becoming a "all or nothing" question for me Jul 24 02:44:15 as long as I can afford maemo.org by any means, I'm 100% dedicated to it. But that will have to stop pretty soon Jul 24 02:48:16 Current economy is tough, the only way is to look at alternatives, like spreading the workload out and/or use advertising services. I know there will be lots of anger and hatred but it may keep the service alive. Jul 24 02:49:57 I already pondered a "premium service" where everybody has to register and pay a 8$/year to access *.m.o on a high-availability level Jul 24 02:51:58 if you don't, you have to face no support, no high nadwidth acces, and a number of other convenience options missing Jul 24 02:52:11 bandwidth* Jul 24 02:52:46 as long as you don't make using the repositories harder, do whatever you want Jul 24 02:52:57 @ premium utilities Jul 24 02:53:03 or you volunteer to contribute to maintenance Jul 24 02:53:33 sharing the load is probably the best option... although it's tough to get people involved I suppose (and those with some tech skills too) Jul 24 02:54:05 * FIQ|n900 could definitely see himself donating to keep maemo.org alive Jul 24 02:54:18 yes, it is a game with negative balance for the first few weeks or months Jul 24 02:54:20 not sure if I could help in other ways though Jul 24 02:54:22 * psycho_oreos thinks as soon as there's premium services becomes an option, there would be people willing to try and make a mirror out of *.m.o beforehand. Jul 24 02:55:09 DocScrutinizer05: regarding donations, wouldn't you be able to request donations from fsf? Jul 24 02:55:20 dunno how that works though Jul 24 02:55:21 I dunno Jul 24 02:56:14 for now I manage donations that for 49/50 I may not touch to buy me a pizza Jul 24 02:56:47 but given that you're a not-for-profit organization towards something they support at a reasonable size, they could probably help Jul 24 02:57:34 and I'd prefer to not manage them at all - hope I get rid of this task as soon as new HiFo BoD sorts issues with official paypal account Jul 24 02:58:24 They're still waiting from that one person to sort things out? Jul 24 02:58:36 yep, basically Jul 24 02:59:48 ~xyawn Jul 24 02:59:48 good coffee Jul 24 02:59:53 mhm **** ENDING LOGGING AT Wed Jul 24 02:59:59 2013