**** BEGIN LOGGING AT Sun Jul 21 02:59:58 2013 Jul 21 05:55:32 DocScrutinizer05: i have ext4, yeah Jul 21 05:55:49 first, repartition and align to 4mb Jul 21 05:56:02 then add the raid striping options to ext4 Jul 21 05:56:06 then enjoy ext4 Jul 21 07:37:02 hi all Jul 21 07:37:35 I understand that mail for exchange can sync only 1 account, Emminnig can sync calander Jul 21 07:37:45 is there way to sync my contact ? Jul 21 09:07:58 good (UGT) morning Jul 21 09:57:13 g'morning Jul 21 10:09:49 wtf tempratures Jul 21 10:27:03 wtf Jul 21 10:27:13 why the hell does user not have write access to MyDocs Jul 21 10:27:18 and the SD? Jul 21 10:33:28 you have got to be kidding me Jul 21 10:33:40 i was wondering why i can't save crap from the device to the vfat Jul 21 10:33:51 and i find out that for whatever reason user doesn't hve write permissions after updating to thumb Jul 21 10:34:28 Macer: check perms, check genfstab.awk Jul 21 10:40:55 i know the permissions are wrong Jul 21 10:41:01 it is mounting with root:root as its owner Jul 21 10:41:08 i just changed the fstab manually for MyDocs Jul 21 10:41:11 and rebooted.. Jul 21 10:41:15 but seriously.. wtf? Jul 21 10:41:45 where is genfstab.awk? /etc? Jul 21 10:43:33 this is crazy Jul 21 10:43:49 especially considering MyDocs is vfat Jul 21 10:43:50 lol Jul 21 10:43:59 i don't see a genfstab.awk Jul 21 10:44:13 and you can't change perms on a vfat fs Jul 21 10:45:59 there is no genfstab.awk on my friggin n900 Jul 21 10:47:41 wow. this is amazing. i'm sick of bothering with using old devices. time to just suck it up ebay these and get a GS4 Jul 21 10:47:56 an update shouldn't ruin your ability to write to your SD and MyDocs Jul 21 10:51:53 even attempting to remount with mount with the uid flags set doesn't change it Jul 21 10:53:36 fk this. i'll just swap the sim out move my stuff off of it and ebay it Jul 21 10:53:57 and make someone else happy with a valued n900! \o/ Jul 21 10:54:08 if happy is the word you want to use Jul 21 10:54:38 i like the n900 but when you have shit like this happen where you are trying to figure out why you can't save an img from an mms Jul 21 10:54:41 it's time to go Jul 21 10:54:54 n900 doesnt even support mms out of the box Jul 21 10:55:03 that should tell you a lot right there Jul 21 10:55:06 but even so Jul 21 10:55:11 mms is deprecated Jul 21 10:55:12 it isn't the mms that is the problem Jul 21 10:55:32 lol. if you say so Jul 21 10:55:47 there are still millions and millions of people who use mms to send pics Jul 21 10:55:57 but like i said Jul 21 10:56:01 that isn't the issue right now Jul 21 10:56:07 right now i updated and i can't write to the vfat partitions Jul 21 10:56:21 i run thumb, didnt have such a problem Jul 21 10:56:33 nor can i find an easy fix to do so.. so now i have to dive into it and try to figure out what went wrong Jul 21 10:56:44 and why i can't do something as simple as write to MyDocs and the SD? Jul 21 10:57:06 i hate to say it but these are the sorts of things that should just work. it is no wonder nokia killed maemo off :) Jul 21 10:57:09 same reason why you cannot do other 'simple' taks on other mobiles Jul 21 10:57:14 tasks* Jul 21 10:57:18 er Jul 21 10:57:35 like..... use PAN, MMS, and write to an SD? :) Jul 21 10:57:47 blame nokia for cssu thumb update? Jul 21 10:57:57 why dont you blame MS too, since thats fashionable Jul 21 10:58:00 * Lava_Croft huckles Jul 21 10:58:04 nope. but i can understand why they dropped it Jul 21 10:58:17 you cant Jul 21 10:58:19 you are just enraged Jul 21 10:58:39 Macer: please stop ranting! Jul 21 10:59:00 yeah, ranting is the sole right of DocScrutinizer05 Jul 21 10:59:05 * Lava_Croft chuckles some more Jul 21 10:59:52 yes, since *I* know how to rant, while Macer just makes a fool of himself, by bitvhing about MyDocs not mounted rw Jul 21 11:00:24 opinions probably differ on the first part of your sentence Jul 21 11:00:31 but the last part is correct Jul 21 11:00:48 if that's a bug then report a bug, don't act like a defiant child that assumes everything done just to annoy him Jul 21 11:01:33 and there i thought freemangordon lived off of the misery experienced by people like Macer Jul 21 11:01:37 boy was i wrong! Jul 21 11:01:41 lol Jul 21 11:02:29 and particularly staements like "I'm not going to do $X to get rid of this maemo device, and get a $X instead" is not interesting anybody in this channel Jul 21 11:02:53 well... if anybody wants to buy this thing let me know. i'm going to ebay it in a bit. i'll give it up for $150 Jul 21 11:03:00 s/not/now/ Jul 21 11:03:00 DocScrutinizer05 meant: and particularly staements like "I'm now going to do $X to get rid of this maemo device, and get a $X instead" is not interesting anybody in this channel Jul 21 11:04:11 I suppose you try ebay right away, since all your bitching wasn't a good PR for selling your device at that pricetag in here Jul 21 11:04:21 k Jul 21 11:04:26 just figured i'd ask Jul 21 11:04:35 150? nah Jul 21 11:06:29 I'm also no fan of latest CSSU-T fstab hack. It's basically worth nothing and high risk, just implemented to avoid the problems it's actually introducing. But honestly, that's a CSSU-T issue, not a N900 or maemo issue Jul 21 11:09:00 i would have to agree with you there Jul 21 11:09:22 i still have my e7 tho. so there is that at least Jul 21 11:09:47 but hey, all my fighting for a conservative maintenance of CSSU finally succumbed to the fanboys of bleeding edge and "we want monthly updates, no matter what for" Jul 21 11:10:01 lol Jul 21 11:10:18 yeah. i just figured it would provide a benefit Jul 21 11:10:38 seems not but to each their own. im sure devs still love n900s Jul 21 11:11:06 nowadays i just want something that has a qwerty and works lol Jul 21 11:12:07 s4 has the attachable bt qwerty so that should do the trick Jul 21 11:12:43 and a msd slot. time to give up on the could've beens. Jul 21 11:13:43 ah well. it has been fun. good luck with everything. Jul 21 11:14:17 and it has android Jul 21 11:14:19 no gnu userland Jul 21 11:14:21 samsung hw Jul 21 11:14:24 glass screen Jul 21 11:14:32 have fun with it! Jul 21 11:15:34 meh. i will put a debian chroot on it. the touchscreen is better with multi finger support... can install cm on it to avoid ad-droid based stuff. and it can multitask just fine Jul 21 11:16:25 and i am bold enough to say the samsung hw is probably a bit stronger than what the n900 is working with. ;) Jul 21 11:16:41 CM is still android Jul 21 11:16:47 stronger as in...? Jul 21 11:16:50 it still has no GNU userland Jul 21 11:16:58 and you need a quadcore CPU to run android without lag Jul 21 11:17:02 sure CPU has higher clock Jul 21 11:17:06 well. not like there are many alternatives Jul 21 11:17:16 samsung captiva glide Jul 21 11:17:23 at least that has a HW keyboard Jul 21 11:17:24 maybe ubuntu devices will show promise Jul 21 11:17:37 oh? Jul 21 11:18:04 DocScrutinizer05: http://www.phonespot.nl/images/productimages/big/Samsung%20Captivate%20Glide%20i927%20002.jpg Jul 21 11:18:07 i927 Jul 21 11:18:08 thanks Jul 21 11:18:15 friend of mine got it to replace his n900 Jul 21 11:18:20 apart from android, he is rather happy with it Jul 21 11:18:24 oh, and it lacks a CTRL key:) Jul 21 11:18:36 but there is a CM release for it, so Jul 21 11:18:40 wow, that looks nice Jul 21 11:18:44 its not maemo, but its not that bad Jul 21 11:18:58 wow, uhh, that's a pretty crappy android phone. Jul 21 11:19:03 i think only in the US/Asia tho Jul 21 11:19:07 not EU Jul 21 11:19:46 http://upload.wikimedia.org/wikipedia/commons/thumb/2/20/Samsung_Captivate_Glide_-_SGH-I927_-_013.JPG/1280px-Samsung_Captivate_Glide_-_SGH-I927_-_013.JPG Jul 21 11:19:50 nicer pic Jul 21 11:20:01 meh, get a phone with a large enough keyboard, and you won't be complaining about hwkb unless you use terminal /a lot/ to edit files in vi. Jul 21 11:20:20 hw keyb > * Jul 21 11:20:25 no onscreen keyboard can replace it Jul 21 11:20:36 eeeeek! that looks like one crappy kbd, almost as crappy as N950 or eten M800 Jul 21 11:21:04 Lava_Croft: N9's touch keyboard is fine by me. Jul 21 11:21:11 its no replacement Jul 21 11:21:12 http://www.amazon.com/gp/aw/d/B00D2ZUN1Q/ref=mp_s_a_1_3?qid=1374405622&sr=8-3&pi=SL75 Jul 21 11:21:22 was thinking more of that Jul 21 11:22:13 that will need another case to protect the screen Jul 21 11:22:22 since its samsung, the shell of the phone is made of rubber-like substance Jul 21 11:22:30 making the phone bendable as shit Jul 21 11:22:42 thats why samsung invests so heavily into flexible screens Jul 21 11:22:49 thats easier than learning how to make good HW Jul 21 11:23:09 like falling off usb ports? ;) Jul 21 11:23:44 never happened to me or anyone i know with n900 Jul 21 11:23:46 http://https://play.google.com/store/apps/details?id=org.dyndns.sven_ola.debian_kit&hl=en Jul 21 11:23:49 i only know it from TMO Jul 21 11:24:03 and the AWESOME keys it got! "www." ".com" `speech input´ "smiley" "OK" Jul 21 11:24:18 haha Jul 21 11:24:19 welcome to android Jul 21 11:24:25 can I please get ".de" instead of ".com" ? Jul 21 11:24:30 well... that was the 1st one i saw Jul 21 11:24:31 .com key is normal in android, with onscreen keyb too Jul 21 11:24:35 there are others Jul 21 11:24:52 but they have them to get a qwerty onto an s4 Jul 21 11:25:23 "www." and for sure are my alltime favorites Jul 21 11:25:47 im sure they are easily remapped Jul 21 11:25:59 to tab and pipe Jul 21 11:26:04 heh Jul 21 11:26:23 I hope the printing on key is also scrap-off otherwise too embarrassing Jul 21 11:26:36 to whom? Jul 21 11:26:43 to the bearer Jul 21 11:26:48 ;-) Jul 21 11:27:04 i doubt anybody would notice or care. maybe a cs major Jul 21 11:27:27 I'd be ashamed to carry a device with a smiley hw key Jul 21 11:27:29 and even then theyd probably just like it because it is a qwerty on an s4 Jul 21 11:28:05 cant really find too many high end phones with qwerty Jul 21 11:28:08 if any at all Jul 21 11:28:14 evem on plane-3 Jul 21 11:28:22 probably because you have to sell them internationally Jul 21 11:28:41 wacky germans with z and y swapped Jul 21 11:28:48 heh Jul 21 11:29:10 The whole trend is pretty much what Apple had started, phones/devices lacking a proper hardware keyboard. Jul 21 11:29:29 psycho_oreos: it made sense tho Jul 21 11:29:42 Every other damn manufacturer just followed suit. Jul 21 11:29:47 why make 50 different keyboards? Jul 21 11:29:54 when a vkb is software Jul 21 11:30:16 it probably saved billions so far heh Jul 21 11:30:37 Macer, yeah and why not give end users no freedom to even do simple commands like copy and paste. Afterall, programmers are free to write their own vkb instead. Jul 21 11:31:03 psycho_oreos: that's just because idiotic industry rather copies than trusting in own R&D and design department Jul 21 11:31:22 well i am sure it is all bottom line Jul 21 11:31:33 like using cheap solder points on usb ports heh Jul 21 11:31:53 "iPhone does it this way. iPhone is market leader. So we do it same way, it must be right thing to do" Jul 21 11:32:02 iphone is not market leader anywhere except the US Jul 21 11:32:05 DocScrutinizer05, for various reasons on various scales, its inevitable that hardware keyboard phones/devices faces their immediate death after PDAs. Jul 21 11:32:21 Netherlands is about 70 to 80% android Jul 21 11:32:50 Lava_Croft: but most high end ad-droid phones dont have a kb Jul 21 11:33:14 There was a time when in 2009 some of the top end android devices still featured hardware keyboard. These days hardware keyboard phones/devices are commonly found on dumb/cheap phones. Jul 21 11:33:20 that's since android mimics iPhone Jul 21 11:33:31 but i agree on their reasoning. im sure it also helps with 3rd party international sales Jul 21 11:33:32 thats because people dont WANT hw keyboards anymore Jul 21 11:33:36 not because imitating iphone Jul 21 11:33:39 iPhone been first and always been "leader" Jul 21 11:33:47 since for the most part they can be universal Jul 21 11:34:02 no matter about recent market share Jul 21 11:34:25 Lava_Croft: i think they can just live without them Jul 21 11:34:33 if they had them they would use them Jul 21 11:34:47 they dont want it Jul 21 11:34:53 HW keyboard equals bigger phone Jul 21 11:34:55 but copycats like google simply ignore that even Apple might do decisions based on other rationale than mere product usability considerations Jul 21 11:34:56 but a hw kb adds bulk(weight) Jul 21 11:35:47 DocScrutinizer05: cattle eat hay Jul 21 11:36:19 if they starved to death... theyd eat something else Jul 21 11:36:21 There was a time where innovative features like 9xxx series communicator and E7 had ways to hide the hardware keyboard. Now it seems like hardware keyboard is an unecessary junk/bulk. Jul 21 11:36:39 see "bigger phone" for example. everybody thinks the trend is to ever smaller devices, while every usability expert can elaborate why that's a deadend path. And recent devices show that trend goes to 5" or even 7" *phones* Jul 21 11:36:56 heh Jul 21 11:37:12 i wouldnt say my e7 hides it well Jul 21 11:37:17 where do you get such general statements from like "everybody thinks the trend is towards blabla" Jul 21 11:37:22 i guess it isnt that big Jul 21 11:37:27 im using it now Jul 21 11:38:08 Lava_Croft: that's a simplified statement to summarize trend of whole industry during last 6 years Jul 21 11:38:27 maybe last 8 except last 18 months Jul 21 11:38:35 N900 isn't all that particularly big with hardware keyboard. It was probably because of the design limitations prohibit N900 to be thin, if you looked at N950 which was virtually based off E7 its noticeably thinner. Jul 21 11:38:48 the trend has been towards largest phones for quite some years Jul 21 11:38:57 with even apple copying the androids and getting a bigger screened phone Jul 21 11:39:14 (after stating you would never need a bigger screen, of course) Jul 21 11:39:22 Largest and thinnest phones.. Jul 21 11:40:03 psycho_oreos: yep Jul 21 11:40:52 N900 thickness is basically determined by cam module Jul 21 11:41:34 The wars between trends has always been a key factor, just yesterday I was watching N900 demo/teaser on cutetube and reading comments to see people flaming each other about N900 vs Jul 21 11:42:25 the N900 cam module seems rather poor quality regarding size, in my book Jul 21 11:43:13 In N8, they've improved (probably) that cam module but made it purely symbian. Jul 21 11:43:29 plus it makes for that ugly annoying bump on battery cover Jul 21 11:43:46 I would've shot the ID designers for that Jul 21 11:46:44 but compared to iPhone cam module it's still phantastic, hw wise (my own unsupported notion). Just that apple did a hell of a job on sw side to compensate for the shorcomings of their crap cam Jul 21 11:48:12 which *usually* makes for nice snapshots on iPhone, unless you go to take photos under certain conditions (moved objects) where the limitations show clearly, by generating nasty artifacts Jul 21 11:49:52 but then, Apple is a sw developer that knows his busines sw side and eventually decided to do phone hw, while Nokia been a hw developer always and also generally did a lot of outsourcing for the sw side Jul 21 11:50:19 so the differences in design paradigms are clearly based on compny's history Jul 21 11:52:07 and not even Nokia is rid of limitations the FOSS idea imposes on hw design Jul 21 11:54:01 ask dm8tbr about how boneheaded companies like e.g ST-E can act when it comes to proper licensing of simple stupid .h files for e.g. GFX Jul 21 11:55:56 iirc the MALI400 grafics core can't get used in FOSS just because ST-E refuses to GPL a already disclosed 200lines .h that's "(C) ST-Ericson" Jul 21 11:56:17 Rewrite the header, use different names Jul 21 11:56:23 rearrange them in the header Jul 21 11:56:41 I think if it were that simple, they would have contemplated with seriousness. Jul 21 11:56:54 psycho: no Jul 21 11:57:21 no, since both parties are somewhat boneheaded Jul 21 11:57:31 N900 had lots of cool hardware in general, people just don't realise they could unlock the hardware's inner potential and to look at the obvious shortcomings. Jul 21 11:58:14 well, the fcam guys did Jul 21 11:58:25 and the blessN900 guys as well Jul 21 11:58:55 It was only then had they got these other people with asking about it and wishing their manufacturer would do the same. Jul 21 11:59:03 hmmm Jul 21 11:59:06 mali can't be used by foss? Jul 21 11:59:06 but still they can't compete with what a 500 heads developer team for just the cam drivers can do at Apple Jul 21 11:59:07 this is new to me Jul 21 11:59:16 I mean, there's no finished foss driver Jul 21 11:59:20 There was other things like fmtxd, the wl1251 module, the trio LED notification. Jul 21 11:59:25 Wizzup_: that might have changed meanwhile Jul 21 11:59:27 life with the n900 is generally like this: iphone user looks at n900 and smirks about size, screen, stylus Jul 21 11:59:31 I have it running on my gentoo install Jul 21 11:59:33 but once anything has to be done, n900 has to do it Jul 21 11:59:35 Wizzup_: that info is like 2 years old Jul 21 11:59:39 maybe 3 Jul 21 11:59:40 ok Jul 21 11:59:41 because iphone cant/wont/whatever Jul 21 11:59:49 DocScrutinizer05: libv and others are working on lima Jul 21 11:59:55 I'm hoping it finishes soon Jul 21 11:59:58 at least, comes around Jul 21 12:00:30 Wizzup_: I jusz quoted that instance since I been working inside ST-E at that time, and tried to move sth about the issue with that .h file Jul 21 12:00:40 after dm8tbr told me about it Jul 21 12:01:39 alas Ericsson worked on the LTE modem, not the AP/GFX halve of NovaThor Jul 21 12:01:50 so I couldn't do much Jul 21 12:02:02 hi Jul 21 12:02:24 there's one thing I've become tired of regarding the application "mobilehotspot" Jul 21 12:02:41 every now and then, seemingly completely random, it simply doesn't start Jul 21 12:03:04 the exact error being "ImportError: No module named os" Jul 21 12:03:22 the solution is simply rebooting and hope it doesn't happen again Jul 21 12:03:43 and if that fails, do it again (though that happens rarely) Jul 21 12:03:51 any idea *what* causes this thing? Jul 21 12:03:56 no Jul 21 12:04:01 share syslog Jul 21 12:04:18 sorry? Jul 21 12:04:35 and start mobilehotspot in terminal or with stdout/stderr redirection to a logfile Jul 21 12:04:58 He basically meant share a log where useful information can probably be extracted. Usually from the likes of dmesg/syslog. Jul 21 12:05:11 yep Jul 21 12:05:16 tnx psycho_oreos Jul 21 12:05:32 When X happens instead of Y, its always best to run things in terminal, look in all the possible logs/avenues, etc. Jul 21 12:05:44 mhm Jul 21 12:06:02 You don't go, oh its annoying me *reboot and hope it goes away*. Jul 21 12:06:20 yeah just wasn't sure what he meant with syslog Jul 21 12:06:24 regularly programs spit some diagnostics to stderr or syslog when something inexpected happens Jul 21 12:06:30 ^ Jul 21 12:06:39 psycho_oreos: indeed, but this exact issue *do* get solved by rebooting Jul 21 12:06:47 which I found out accidentally once Jul 21 12:06:56 and this doesn't make any sense to me Jul 21 12:07:01 That's where its useful to debug rather than doing nothing until it got annoying. Jul 21 12:07:24 thus: share syslog Jul 21 12:07:33 ~pkg Jul 21 12:07:34 extra, extra, read all about it, pkg is http://maemo.org/packages/ Jul 21 12:07:36 syslog Jul 21 12:07:37 FIQ|n900, I don't know, that error almost sounds like some issue with python. Though I'm no programmer. Jul 21 12:07:42 sysklog? Jul 21 12:07:56 +d ;) Jul 21 12:08:00 /var/log/syslog Jul 21 12:08:16 only if sysklogd is installed Jul 21 12:08:29 otherwise dmesg, though it's sub-par Jul 21 12:09:14 Speaking of dmesg, I found there was a nice perl script to translate those ktimer messages into meaningful syslog lookalike output. Jul 21 12:09:46 and particularly "run-standalone.sh mobilehotspot 2>~/MyDocs/MH.log 1>&2 Jul 21 12:09:57 " Jul 21 12:10:18 DocScrutinizer05: also I tried to import os manually within a python interpreter Jul 21 12:10:30 and it fails to find it as well Jul 21 12:10:35 so something weird is going on @python Jul 21 12:11:06 psycho_oreos: cooool Jul 21 12:11:10 When it did work, where was the "import os" usually sourced from? Jul 21 12:11:22 DocScrutinizer05, I was going to paste the link, hang on.. I'm going to find it again. Jul 21 12:12:04 DocScrutinizer05, http://linuxaria.com/article/how-to-make-dmesg-timestamp-human-readable?lang=en <-- only the perl script works on N900, python variant failed. Jul 21 12:12:22 mhm Jul 21 12:12:40 anyway, I'd probably rather need it on my PC. Thanks a lot Jul 21 12:13:03 Hehe, oh well, wherever you may find it handy :D Jul 21 12:13:09 (on N900 I got syslog since ~ beginning of times ;-d ) Jul 21 12:13:10 no worries. Jul 21 12:13:16 [14:11:09] When it did work, where was the "import os" usually sourced from? Jul 21 12:13:52 I don't know this one atm, so I guess I'll have to reboot into a working MH to find that out? unless you have another solution Jul 21 12:13:59 * psycho_oreos has had issues with sysklogd failing to start after what was supposedly an issue trying to optify syslog file and wanting to compress it (to save space). Jul 21 12:14:30 DocScrutinizer05: also honestly I fail to understand exactly how dmesg is helpful in this case Jul 21 12:14:30 lol Jul 21 12:14:31 FIQ|n900, I'm no programmer, if it did worked at the time during reboots yeah its probably a way I suppose. Jul 21 12:14:53 ohh hooray, that pastebinit script requires os too lol Jul 21 12:14:55 err I meant after a reboot. Jul 21 12:14:56 FIQ|n900: since MH needs kernel support, it *may* be relevant Jul 21 12:15:32 Hmm I think I have pastebinit also installed.. lemme try something as well from my end. Jul 21 12:15:43 DocScrutinizer05: but when I earlier (before knowing it required KP) tried to use mobilehotspot in vanilla kernel, the application itself started fine Jul 21 12:15:56 it just failed to create a wlan interface, i.e. not same issue as now Jul 21 12:16:26 FIQ|n900: I'm not going to discuss the possible error patterns in any app on this level. It's futile Jul 21 12:16:57 a swift look at syslog is way more ffective, if only to rule out sth odd is going on there Jul 21 12:18:19 to diagnose *any* arbitrary problem with a process, you generally need: stdout.log, stderr.log, syslog Jul 21 12:18:44 and strace if the aforementioned don't help Jul 21 12:19:13 there's hardly any bug in a program that can't get tracked down following this pattern Jul 21 12:19:49 particularly when it's a bug that causes abort or failure to startup Jul 21 12:19:49 as I didn't had syslog installed - http://sprunge.us/MPFe dmesg output Jul 21 12:19:55 Some programs contains hidden extra parameters that can be shown via strings if the program in question is a binary. Jul 21 12:20:35 mobilehotspot is a python script though Jul 21 12:21:30 yeah, dmesg is sub-par, I said it. [ 80.442291] wl1251: down is the only slightly interesting thing in dmesg Jul 21 12:21:35 DocScrutinizer05: in addition I installed sysklogd for future issues for easier debugging Jul 21 12:21:56 http://sprunge.us/KcVH mobilehotspot's stderr Jul 21 12:22:21 make sure to every now and then clean out your /var/log/* from obsolete old logs or too large current logs Jul 21 12:22:34 I do that every 3 months or so Jul 21 12:22:52 http://sprunge.us/PbBW and the actual mobilehotspot python script (it's small) in case it gives any clues Jul 21 12:22:59 http://sprunge.us/jjgN <-- my "dpkg -l| grep python| pastebinit -b http://sprunge.us" output (if that maybe of certain use), goes to show my python version works with pastebinit. Jul 21 12:23:15 FIQ|n900: stderr already gives a good clue, no? Jul 21 12:23:27 DocScrutinizer05: yes it does Jul 21 12:23:37 but in case mobilehotspot is doing something funny Jul 21 12:23:43 doesn't look like it though Jul 21 12:25:27 Also googling for the exact error phrase may also seem to help as well. Jul 21 12:26:42 psycho_oreos: yes I did this before coming here Jul 21 12:26:59 but it didn't turn up anything interesting/relevant Jul 21 12:28:37 err guys, I'm no python crack at all, but I seem to recall a python shebang alsways looks like "#!/bin/env pythonBLAH" Jul 21 12:29:20 /usr/bin/env yeah Jul 21 12:29:31 just installed fcam, rebooted, and I have no /dev/video0 anymore Jul 21 12:29:40 interesting ;-) Jul 21 12:29:41 not "#!/usr/bin/run-standalone.sh python2.5" Jul 21 12:29:55 isn't env used to get the correct path to the python binary? Jul 21 12:30:26 yep, afaik sth like that Jul 21 12:30:40 disclaimer, see above Jul 21 12:30:47 nfc here Jul 21 12:31:52 FIQ|n900, the first link from google was from stackoverflow which one suggested about using $PYTHONPATH. Now my "find / -name "os.pyc"" outputs these: /opt/pymaemo/usr/lib/python2.5/os.pyc | /usr/lib/python2.5/os.pyc | /home/opt/pymaemo/usr/lib/python2.5/os.pyc Jul 21 12:32:50 psycho_oreos: what did you google on? I basically googled the error I pasted above and got no such link Jul 21 12:33:03 but yes I can try looking for os.pyc Jul 21 12:33:26 FIQ|n900, "ImportError: No module named os" and the link: http://stackoverflow.com/questions/12320143/python-importerror-no-module-named-os Jul 21 12:34:14 Wizzup_, somewhere something got messed up horribly, most likely a bad kernel module. Jul 21 12:34:26 psycho_oreos: probably the /usr/lib/python2.5/os.pyc is a symlink to optified /opt/pymaemo/usr/lib/python2.5/os.pyc which is a bindmount to /home/opt/ Jul 21 12:35:15 DocScrutinizer05, yup :) the real python libs would never sit in rootfs anyway considering how limited the space of rootfs is. Jul 21 12:35:29 ""Consider setting $PYTHONHOME to [:]"" Jul 21 12:36:07 seems to me like $PYTHONHOME is unset or incorrect Jul 21 12:36:22 nfc why Jul 21 12:36:36 or why it's needed anyway Jul 21 12:37:07 maybe $PATH or $(pwd) is incorrect Jul 21 12:37:18 I don't have $PYTHONHOME set as well (no outputs from my N900), though one guy from that link said to put this in their .bashrc: PYTHONPATH=/usr/local/lib/python2.6 . Obviously that's the wrong Python version but it seems like its one hackish attempts to rectify the issue. Jul 21 12:37:41 psycho_oreos: it would appear so, but I have no clue why Jul 21 12:37:53 just installed ``FCamera'' from extras-testing, and I am on power kernel 50 Jul 21 12:38:14 yeah, it's probably needed when sth is non-standard thus non-default on your pathnames and mounts and whatnot Jul 21 12:38:41 MIGHT again be a fstab issue ;-P Jul 21 12:38:46 Wizzup_, usually a mismatched kernel with kernel modules loaded can cause issues. Other times it would output invalid module format -1 or something. syslog/dmesg output maybe useful here. I'm using kp52 with fcam modules (latest) and no issues from my end though. Jul 21 12:38:48 check for `mount` Jul 21 12:39:00 if the bindmounts for python are there Jul 21 12:39:08 Yeah that too. Jul 21 12:40:07 psycho_oreos: manual modprobe seems to fail with invalid module format Jul 21 12:40:08 for the randomness of the issue, I might figure the new fstab/fsck hack introduced to CSSU-T may fail to bindmount the python mounts when and only when any of the fs needs fsck Jul 21 12:40:13 that is odd Jul 21 12:41:07 hmm interesting, I have os.py and os.pyo but not os.pyc in that directory.. and I tried to set PYTHONPATH manually (it was not set before - just tried) and it successfully imported that module, however now it errors on another module... will try to see where it is in case I need something else in that path as well Jul 21 12:41:08 http://sprunge.us/IdCF <-- my mount output, if you'll excuse my excessive use of ext3 here *ahem* :) Jul 21 12:41:10 Wizzup_: zhere's a --ignore-vermagic OWTTE option to modprobe Jul 21 12:41:19 Wizzup_, That's the reason why :) Jul 21 12:41:54 s/- just tried/- as seen on echo $PYTHONPATH/ Jul 21 12:41:54 FIQ|n900 meant: hmm interesting, I have os.py and os.pyo but not os.pyc in that directory.. and I tried to set PYTHONPATH manually (it was not set before - as seen on echo $PYTHONPATH) and it successfully imported that module, however now it errors on another module... w... Jul 21 12:42:02 thx,will be back soon Jul 21 12:43:20 FIQ|n900, I'd also start checking on mount as DocScrutinizer05 pointed out. Jul 21 12:43:30 yes which I just did Jul 21 12:43:38 in addition, the mounts @python seems to have no issues, see http://sprunge.us/IYWL Jul 21 12:45:12 strange Jul 21 12:45:36 you start to qualify for a strace session ;-) Jul 21 12:45:41 * psycho_oreos ponders if the issue might also be to do with older versions of python libraries (hence lacking the necessary facility within their respective postinst) to work properly. Jul 21 12:46:03 (non-localized http://sprunge.us/eeRh ) Jul 21 12:46:39 FIQ|n900: not needed ;-D Jul 21 12:46:40 DocScrutinizer05: also I tried to reinstall all python packages in case something was messed up Jul 21 12:46:56 and interestingly it did not go all well Jul 21 12:47:00 Ahh the only difference is what was written on top, not needed :) I know what the content says :D. Jul 21 12:47:12 yep Jul 21 12:47:14 [14:46:39] DocScrutinizer05: also I tried to reinstall all python packages in case something was messed up - like that optfs restoration yesterday Jul 21 12:47:16 FIQ|n900, that's your answer. Jul 21 12:47:41 and interestingly it did not go all well <--- That could be the source of the issues. Jul 21 12:47:50 yes Jul 21 12:47:55 will paste apt results sec Jul 21 12:50:13 I somewhat would also like to think the issues with these are happening is also because you nuked /opt. It sort of makes programs walk on "broken tiles". Jul 21 12:50:22 * DocScrutinizer05 suddenly feels sick, considering that some python programs might now (after fstab hack) run *before* possibly deferred bindmount of *python 'volumes' Jul 21 12:51:44 * psycho_oreos thinks in either case FIQ|n900 is on the way to creating a zombified version of maemo. :) Jul 21 12:52:05 Lots of hits and misses. I'm guessing python won't be the only solitary issue though. Jul 21 12:53:37 psycho_oreos: though this issue did exist before optfs restoration as well - but then the reinstall might have worked a little better Jul 21 12:55:01 FIQ|n900, outstanding issue? o.O Jul 21 12:55:07 and at least I don't think I can make it *worse*, if anything, better Jul 21 12:55:29 psycho_oreos: the optfs restoration was due to a corrupt file system Jul 21 12:56:05 itself caused by me not realizing that hard shutdowns might be a bad idea... lol Jul 21 12:56:05 FIQ|n900, Yeah I know but you had the issue prior to nuking /opt. So it was virtually an issue that you didn't bother to resolve beforehand. Jul 21 12:56:35 yeah Jul 21 12:57:23 http://sprunge.us/cdAE apt log Jul 21 12:58:19 * psycho_oreos recalls having mount warnings with his microSD about unclean mounts. It was only after.. what, 30+ fsck.ext3 that finally yielded a clean result with the majority of the nokia maps file being corrupted. Jul 21 12:58:58 if microSD is ext3, maemo gives an error on mount Jul 21 12:59:15 unless you're running a fairly recent version of cssu Jul 21 12:59:48 FIQ|n900: delete *all* *.pyc Jul 21 12:59:50 At that time I think I was (somewhat), though I had extra mount params like errors=mount-ro. Jul 21 13:00:18 uh huh now that's strange Jul 21 13:00:22 Hmm the warnings from apt log. Doesn't make much sense when /opt was nuked anyway. Jul 21 13:00:25 looking through that log Jul 21 13:00:37 reinstall gave lot less errors this time Jul 21 13:01:17 >>Some bytecompiled files may be left behind.<< is the key Jul 21 13:01:20 I bet Jul 21 13:01:54 * psycho_oreos doesn't think there would really be any bytecompiled files be left on a nuked /opt. Though if you restored from an older opt.tar that could be a potential issue. Jul 21 13:01:57 DocScrutinizer05: last install it blamed on various modules not being found during the installation Jul 21 13:02:16 Yeah that's usually the result of you nuking /opt :) Jul 21 13:02:22 if you got an obsolete *.pyc in some place prior in $(PYTHON)PATH than the new *.py, you inevitably run into trouble Jul 21 13:02:33 had to reinstall *again* to give a meaningful paste Jul 21 13:02:44 in case anyone wonders why I'm implying I did it twice Jul 21 13:02:55 including the above mentioned os, and some others as well Jul 21 13:02:57 Probably a good idea at the time would be to forcibly remove dead packages and reinstall them. That way apt won't complain (well hopefully). Jul 21 13:03:30 er, last line was @ [15:01:56] DocScrutinizer05: last install it blamed on various modules not being found during the installation Jul 21 13:03:35 psycho_oreos: hmm maybe Jul 21 13:04:26 that's interesting Jul 21 13:04:27 Well in either case I'd say go with DocScrutinizer05's suggestion. Clean out old python stuff (after removing python) and then reinstall python again. Jul 21 13:04:43 the first reinstall didn't solve anything - the second reinstall seems to have solved everything Jul 21 13:04:50 That should get rid of the leftover junk and start fresh... well hopefully. Jul 21 13:04:51 because now stuff seem to work Jul 21 13:05:21 Partially, remember that you had to use $PYTHONHOME because it could not find os.pyc. Jul 21 13:05:38 yes true Jul 21 13:05:39 which is odd Jul 21 13:05:49 All that leftover ancient files could cause the confusion. Jul 21 13:06:02 *.pyc, yes Jul 21 13:06:24 but DocScrutinizer05 Jul 21 13:06:31 don't know if you saw me mention it earlier Jul 21 13:06:46 but I didn't have *any* .pyc files in that directory before Jul 21 13:06:56 No guarantees/recommendations here because you do realise you are trying to do something that probably hardly any users have encountered :) Jul 21 13:06:59 that's the whole point Jul 21 13:07:00 (though I did have .py and .pyo files) Jul 21 13:07:10 Its the leftover junk/cruft. Jul 21 13:07:12 you possibly got some in some *other* dir Jul 21 13:07:41 It also could be the issue with restoring old opt.tar. I'm sure it works but to a lesser extent it does. Jul 21 13:07:46 (though now I have some as I reinstalled) Jul 21 13:07:53 so basically Jul 21 13:08:01 and I have nfc what's .pyo, probably something similar to .pyc Jul 21 13:08:12 I should remove all .pyc files, including the newly created files Jul 21 13:08:22 and reinstall all python packages again? Jul 21 13:08:22 iinm, .pyo is compiled (binary blob) of .pyc. Jul 21 13:09:01 Probably uninstall python first (if you could and to make it not uninstall other packages that depends on python like for instance mobilehotspot). Jul 21 13:09:07 afaik .pyc is compiled with symbolic names, while .pyo is compiled w/o symbols Jul 21 13:09:38 anyway both are dynamically created by python interpreter afaik Jul 21 13:09:39 Ahh I tend to think .pyc is more human readable than .pyo. .pyo is all literally machine code (hence compiled). Jul 21 13:10:47 Remove python packages via apt (hopefully.. or use dpkg if you can somehow make it not uninstall MH), then clean out remaining python junk left in /opt/python. Then reinstall python through apt. Jul 21 13:10:57 That might be the best idea imho. Jul 21 13:11:14 :nod: Jul 21 13:12:25 Don't remove stuff in /opt/python then remove via dpkg/apt.. you'll make dpkg/apt even more messed up about python stuff, let them know that you don't want python junk so when it comes to reinstalling it will know from its own database that you have previously removed it. Jul 21 13:13:22 iirc there's a --purge parameter to some of that apt/dpkg stuff Jul 21 13:13:39 remove or purge? Jul 21 13:13:43 @ dpkg Jul 21 13:13:44 prolly useful here Jul 21 13:13:47 *nods* purge is probably more cleaner. Jul 21 13:14:13 I'd also make sure neither of them complain about removing things like MH which depends on python that is to be removed. Jul 21 13:14:31 Whichever you choose, purge would be cleaner iinm. Jul 21 13:14:48 hello, any qt guru here? I need output backtrace when some signal is emitted. Any idea if qt has some object/slot for that? Jul 21 13:14:52 Or even remove, then purge, purge just gets rid of leftover conf files for instance. Jul 21 13:17:17 Pali, not sure but you can try #qt. According to alis, there's 505 users/bots in that channel. Jul 21 13:17:47 ok Jul 21 13:18:32 I know only about GNU/linux backtrace() function in execinfo.h (man 3 backtrace) Jul 21 13:18:37 * psycho_oreos hopes they can fix the issue rather than pointing you back into channels like these :) where its all maemo/meego centric and there's no active Qt gurus active (or at least it seems). Jul 21 13:19:30 * psycho_oreos has very vague information about all of what you said :/ lol. Not a developer so most of this is really flying above my head. Jul 21 13:19:47 o.O Jul 21 13:21:24 ok, everything was removed fine apart from some directories (the bindmounted ones) Jul 21 13:21:32 w/o manual intervention required Jul 21 13:21:37 now installing again Jul 21 13:22:10 FIQ|n900, don't forget to purge and manually remove any leftover cruft/junk. Jul 21 13:23:00 psycho_oreos: everything was removed as I said, checked all python directories Jul 21 13:23:38 psycho_oreos: o.O indeed. Pali has strange habits Jul 21 13:24:01 (the directories themselves was left intact due to bindmounts and I left them like that as I don't want reinstalling them to fill my rootfs Jul 21 13:24:14 ) Jul 21 13:25:19 FIQ|n900, ok that's good. Just don't want to hear apt/dpkg complaining again all of the sudden. Its only wild guesses from here on in, into fixing python. Jul 21 13:25:35 DocScrutinizer05, heh do tell me :D Jul 21 13:26:12 python installation gave no errors this time Jul 21 13:26:18 FIQ|n900, I have for some wild guess those bind mounts may get recreated once you installed base python stuff. Jul 21 13:26:47 psycho_oreos: well dpkg -P attempted to remove the directories - failing due to bindmounts Jul 21 13:26:56 FIQ|n900: for bindmounts see pymaemo-optify - PyMaemo optification package Jul 21 13:27:01 so my assumption is that apt would not redo the bindmounts Jul 21 13:27:03 FIQ|n900, ahh Jul 21 13:27:25 Seems like DocScrutinizer05 has pointed out a vital clue here as well. Jul 21 13:28:14 was some real fun when that pkg got propagated Jul 21 13:28:54 must've been back in time when PR1.2 came out Jul 21 13:29:12 hmm Jul 21 13:29:37 I wonder if there's a quick straightforward way to ask apt about missing files in every single package Jul 21 13:29:55 so I would be able to clean up after that optfs thing Jul 21 13:29:56 err, there is some tool doing that Jul 21 13:30:16 nfc if it's apt or some supplementary tool though, or even dpkg-* Jul 21 13:30:19 simply looking for runaway symlinks isn't enough as some packages install themselves to /opt directly Jul 21 13:30:44 :nod: Jul 21 13:30:50 a really annoying habit Jul 21 13:31:19 I had a real dispute with iirc shapeshifter about alarmed NOT installing a symlink Jul 21 13:31:50 I consider that a flaw for a cmdline tool to not be available anywhere in usual $PATH Jul 21 13:32:33 IroN900:/etc# ll `which alarmed` Jul 21 13:32:34 lrwxrwxrwx 1 root root 23 2011-06-20 15:53 /usr/bin/alarmed -> /opt/alarmed/alarmed.py Jul 21 13:32:41 had to create that myself Jul 21 13:32:51 also sorry for being clueless occasionally with some stuff regarding things like this Jul 21 13:33:06 no reason for feeling sorry Jul 21 13:33:09 DocScrutinizer05: heh Jul 21 13:33:30 we all started learning about that stuff Jul 21 13:33:37 some time Jul 21 13:34:53 N900 was my first serious step into slightly more complex parts of linux, i.e. before I got a n900 I barely knew what "cd" did lol Jul 21 13:34:56 oh well, afk Jul 21 13:35:04 FIQ|n900: you're clearly willing to learn. so it's a please to help you Jul 21 13:35:16 a pleasure* Jul 21 13:35:50 I'm sure whilst you are experimenting with what we have suggested, we too are in the process of learning as well. Jul 21 13:36:00 sure Jul 21 13:36:15 which is why we're all in here doing what we do, isn't it? Jul 21 13:36:17 Its not exactly a "one way street" where we knew what the answer was and could give you a direct answer. Jul 21 13:36:24 yup :) Jul 21 13:36:50 ~question Jul 21 13:36:51 methinks question is If you have a question and want people to give useful answers, make sure you have read this first: http://www.catb.org/~esr/faqs/smart-questions.html Jul 21 13:37:06 has some nice chapters about that Jul 21 13:37:31 it's a recommended read even for experienced hackers Jul 21 13:37:53 dpkg -S I found tells me what files belongs to which package. Though I haven't found a single command that could do it easily. If you could get "find" to output broken symlinks then you maybe able to get dpkg -S to show which broken symlink belonged to which deb. Jul 21 13:37:56 I try to read it at least twice a year Jul 21 13:38:26 hehe I've read that page numerous times, very interesting stuff even though not everyone follows it closely. Jul 21 13:39:05 it's hard to follow it closely, but it gives you a perspective each time you read it Jul 21 13:39:40 I get a little annoyed when people just say "X is broken" "please help", and then you have to start digging clues out of them to get a basic/general understanding of their setup. Jul 21 13:40:01 people who says X is broken are correct Jul 21 13:40:04 Also I guess its also a good reminder to ask questions the "right" way as well. Jul 21 13:40:06 and it cant be helped Jul 21 13:40:37 Its not the program can't be helped, its more like an issue with pebkac ;) Jul 21 13:40:41 http://www.catb.org/~esr/faqs/smart-questions.html#idp5285344 Jul 21 13:42:03 Yeah I've sometimes thought of that as well, but it can get a little irritating when you're also dealing with a situation on hand from your end as well and there's nobody else responding. Jul 21 13:43:07 At times certain commands "collide" when you're helping others on X distro when you're on Y distro. Jul 21 13:44:14 Having to think what the command "equivalent" was to match their case. Jul 21 13:45:48 * psycho_oreos thinks right now there might be a hackish way to output dpkg -L| echo $i| find $i when say for instance under archlinux, pacman -Ql by default lists the name of the package along with the files owned by that package. Jul 21 13:47:16 yep, you can craft some cmdline wizardry Jul 21 13:47:43 but you also can talk the person you try to help thru using all the tools actually available Jul 21 13:48:16 That and the sort of "hackers" way to fix things ;) hackish attempts Jul 21 13:49:12 example of a useful cmdline wizardry: man md|less +"/If an array is created" Jul 21 13:50:33 The power of pipe combined with BASH scripting ;) Jul 21 13:51:35 you wouldn't want to talk the helpee thru less(1) basics Jul 21 13:52:14 or tell them "look at line 255 of `man md´" Jul 21 13:53:05 In that case I guess grep maybe useful here ;) Jul 21 13:53:54 nah, not really unless you want to "quote" a very strictly defined limited section out of some text Jul 21 13:55:12 man md|less +"/If an array is created" -j 9 Jul 21 13:55:21 Yeah that would be a little difficult. Jul 21 13:56:29 I just searched, there's dpkg -C but doubt it's usefulness. Jul 21 14:00:19 man bash|less +"/^[^a-z]*SIGNALS" Jul 21 14:17:10 psycho_oreos: so I guess you have kernel-power from extras-devel? Jul 21 14:18:05 Wizzup_, yup Jul 21 14:18:18 as well as fcam Jul 21 14:18:25 okay, I'll try that, does blessn900 work with that fcam driver? Jul 21 14:19:01 afaik yes. Jul 21 14:19:43 (I really just want blessn900, but I guess I can upgrade the power kernel) Jul 21 14:20:20 You might be able to just get away with newer fcam-drivers insntead of new power kernel. Though I'm not sure. Jul 21 14:20:21 tbh I don't understand why the current ``stable'' fcam driver don't work with my power kernel Jul 21 14:20:30 doesn't* Jul 21 14:21:44 Which version do you have just curiously. Jul 21 14:22:15 2.6.28.10-power50 Jul 21 14:22:47 Err my bad not kp version, fcam-drivers version. Jul 21 14:23:33 0.7.1 I think, let me check Jul 21 14:24:27 http://maemo.org/packages/view/fcam-drivers/ shows why there's a -2 Jul 21 14:24:33 just not sure for what powerkernel that is Jul 21 14:24:39 ok folks. my n800 suddenly stoped working. it turns on and seems to boot but the screen just cycles though some weird vertical stripes then a flash on the whole screen then black Jul 21 14:24:41 dpkg -l fcam-drivers| awk '{ print $3 }' Jul 21 14:25:02 the only thing i did was charge it several times last night. ideas? Jul 21 14:25:50 What did you do before this issue started happening? Jul 21 14:25:51 psycho_oreos: 1.0.7-1 Jul 21 14:26:33 downloaded some maps last nite in maemo mapper. that's it. Jul 21 14:26:34 Wizzup_, that should work with kp50 afaik. It works with kp52 but its been awhile since I looked into what could be the issue. Jul 21 14:26:46 I'll upgrade kp Jul 21 14:27:11 * psycho_oreos doesn't know if that'll necessarily fix the problems though. Jul 21 14:27:11 ah, I should move on with my uinput-mapper package soon... Jul 21 14:27:15 psycho_oreos: we'll see Jul 21 14:28:04 help! Jul 21 14:28:05 Some files within kp (at least 52) has kernel version specific fcam provided kernel drivers. I'm not sure have they been merged. Jul 21 14:29:09 mancha, I misread :/ I thought it was N900 you are talking about not N800. I was going to suggest it could be issue with bootlooping and low battery. I don't have N800 so I don't know I'm afraid. Jul 21 14:29:39 could i have fried my battery? Jul 21 14:30:06 Unlikely unless you charged your battery with some other device. Jul 21 14:30:09 mancha: possible, but i'm afraid your problem is more severe than that Jul 21 14:30:21 DocScrutinizer05: how can the thing just go bad all of a sudden? Jul 21 14:30:33 like all things do Jul 21 14:31:20 is there some kind of safe boot? Jul 21 14:31:37 err, not that I know of Jul 21 14:31:47 I don't own N800 though Jul 21 14:31:52 psycho_oreos: hm, extras-devel is suggesting 51r1 Jul 21 14:32:19 what battery does N800 use? Jul 21 14:32:34 Wizzup_, *grr at whoever is playing with the repository* you can probably install kp52 from Pali's own repository or manually. Jul 21 14:32:35 bp-5l Jul 21 14:33:03 what when you plug it into charger? Jul 21 14:33:20 psycho_oreos: my ``Distribution'' is set to fremantle-1.3, I hope that makes sense Jul 21 14:33:21 nothing Jul 21 14:33:28 :-/ Jul 21 14:33:29 that's the only think I can imagine that may be different Jul 21 14:35:04 Wizzup_, I don't think that would be relevant. Its the issue with the repository still showing 51r1. *looks at his fapman repository configuration*. Jul 21 14:35:22 aha Jul 21 14:35:39 I wouldn't recommend using fapman either :) Jul 21 14:35:47 no, people told me this a few years ago Jul 21 14:35:48 ;-) Jul 21 14:36:28 DocScrutinizer05: it's the batter. Jul 21 14:36:30 * psycho_oreos issues an update and checks via apt-cache on kp52. Jul 21 14:36:44 mancha: glad to hear that Jul 21 14:37:07 mancha, so a replacement battery worked? Jul 21 14:38:17 Wizzup_: that may make a friggin lot of difference, since during migration#1 the fremantle vs fremantle-1.3 duality got messed up Jul 21 14:38:19 psycho_oreos: getting fcam-drivers from devel worked Jul 21 14:38:27 no, but aside from the screen, the bootup sound is faint. Jul 21 14:38:45 Wizzup_, 1.0.7-2?? Jul 21 14:38:47 DocScrutinizer05: yeah, I recall setting it to -1.3 was the right thing to do, I think I did it properly Jul 21 14:38:50 psycho_oreos: yes Jul 21 14:38:50 it seems like it's not getting the right charge...i bet i fried the battery by leaving it plugged in all night Jul 21 14:39:01 psycho_oreos: which is also kind of what his change message says Jul 21 14:39:13 Wizzup_: the sad truth is: right now there is no "right" setting Jul 21 14:39:13 at http://maemo.org/packages/view/fcam-drivers/ Jul 21 14:39:18 DocScrutinizer05: nice ;-) Jul 21 14:39:22 indeed Jul 21 14:39:28 Wizzup_, *nods* I thought that was just some custom one I built (when I was tinkering with abc). I never knew that was in extras-devel. Jul 21 14:40:07 fremantle-1.3 used to be a symlink to fremantle Jul 21 14:40:48 during migration#1 that got messed up, just doing a mirror or whatever of the repo, and that created TWO repos: fremantle and fremantle-1.3 Jul 21 14:41:02 do0es a dead battery sound plausible? Jul 21 14:41:05 heh Jul 21 14:41:15 mancha, I guess it goes back to what DocScrutinizer05 said earlier, sounds like its more severe. I guess there's probably multiple issues, though one step at a time helps I guess. Jul 21 14:41:18 up to anybody's guess which action adresses which repo, also in autobuilder and repo-management Jul 21 14:42:19 mancha, its hard to say :) two of us has admitted we don't own N800, so its hard to know what the problem is if we don't know much about it. Jul 21 14:42:56 psycho_oreos: why do you both assume something more seirous? Jul 21 14:43:48 mancha, to me personally the colour stripes showing on the screen before going black. Prior to that you installed a maps software, it might also be screen playing up but hard to know. Jul 21 14:43:52 mancha: because bad battery usually doesn't cause atrifacts on screen Jul 21 14:44:59 What happens if you tried plugging the charger in whilst having the "dead/bad battery" in place? Does N800 still boot? Jul 21 14:45:27 afaik N8x0 should kinda work with a battery of any quality and charge, as long as plugged into charger Jul 21 14:46:05 Yeah I thought of the same thing as well. At least on N900 you can see the nokia logo before it dies and tries to go back into emergency charge mode. Jul 21 14:46:11 it *might* shut down with battery-low or battery-defect signal Jul 21 14:46:40 but I never heard of artifacts on screen Jul 21 14:47:08 Or the battery is simply too faulty that it fails to even allow charging when power is plugged in. Jul 21 14:47:31 as usual, a comlete reflash might help to distinguish hw from any sw problems Jul 21 14:47:37 Either way, its weird but its beyond my knowledge... I'm virtually stabbing in the dark here. Jul 21 14:48:11 or you just get a replacement battery Jul 21 14:48:18 and check if things change Jul 21 14:48:20 well you guys are a very pessimistic bunch. you're real downers. Jul 21 14:48:23 One step at a time :) Jul 21 14:48:39 mancha, do you have a better suggestion as to what it might be? :) Jul 21 14:48:43 no, we aren't Jul 21 14:48:48 "i don't own an n800 so i have no fucking clue what's up but i'll just pull out my ass it's super serious" :) Jul 21 14:49:09 ~joerg Jul 21 14:49:09 joerg is, like, a natural born EE, ex HW-developer and engineer of Openmoko, now mainly involved in maintenance of maemo and N900. Usually known as DocScrutinizer Jul 21 14:49:25 We've given you a likely scenario on N900's case. Jul 21 14:50:20 mancha: I'm just stating that I don't have N800 to make clear my statements are based on N810 and N900 and general knowledge of embedded hardware Jul 21 14:50:23 The least you could do is try to solve the problem one at a time, first plug it into wall charger with battery in place. If it still doesn't boot, get a replacement battery, if it doesn't boot its probably the screen at fault (or others). Jul 21 14:50:36 ok, if the battery got fried, it's old. and the cells leaked, can this not be providing less volage/amperage than is needed to power the lcd display? Jul 21 14:50:41 question for the EE :) Jul 21 14:51:00 yes Jul 21 14:51:19 still no chance to create artifacts, on all embedded devices I know Jul 21 14:51:26 why not? Jul 21 14:51:50 Probably because of the way its designed I gather. Jul 21 14:51:53 because hw doesn't work that way usually, unless my colleagues messed up N800 fubar Jul 21 14:52:44 and I've seen similar artifacts caused by other hw defects Jul 21 14:52:52 but never by battery Jul 21 14:53:26 actually your error description is very similar to what you get on N900 when flex connector is broken Jul 21 14:53:48 now the N800 doesn't have a flex connector for all I know Jul 21 14:54:17 since it doesn't have a slide-out kbd Jul 21 14:54:59 anyway N8x0 battery charging is notorious fubar, and I suggest you get a spare battery to check Jul 21 14:55:16 The very least you could do is try the suggestion I mention beforehand. I'm not an EE, I don't own a N800 but I'm giving you advice from what a N900 user would try otherwise. Jul 21 14:55:51 it boots. Jul 21 14:56:03 but i can't see anything cause the screen is fucked. Jul 21 14:56:06 If it boots with power in place, then its probably a faulty battery. Jul 21 14:56:17 cmon dude. Jul 21 14:56:22 Yeah sounds like possible multiple issues. Jul 21 14:56:26 no, when LCD still fux0red then it's a hw defect Jul 21 14:56:27 no way Jul 21 14:56:46 mancha, what do you think it is? that it is our fault? Jul 21 14:56:48 look, multiple issues don't just happen out of the blue at the same time. sorry that dog don't hunt. Jul 21 14:57:12 if the multiple issues are related somehow then i'll bite, otherwise it's just nonsense. Jul 21 14:57:12 I hate to say this but N800 is an aging device. Jul 21 14:57:14 he meant multiple alternative causes Jul 21 14:57:31 oh i see. Jul 21 14:57:32 N900 is facing the same path as well. Jul 21 14:57:43 that's irrelevant whether it's an aging device. Jul 21 14:57:58 Electronics generally don't *last* forever. Jul 21 14:58:08 i've not used it much and more important are usage cycles. Jul 21 14:58:31 well Jul 21 14:58:38 get a damn spare battery! Jul 21 14:59:01 lol Jul 21 14:59:03 I know old PSU outputs less power gradually over time before the machine (should the motherboard be intelligent enough) to prevent the machine from booting if PSU is simply not outputting enough power. That's one use case with aging electronics from what I know. Jul 21 14:59:23 DocScrutinizer05: trying to find a store now. Jul 21 14:59:28 the battery is prolly 10 years old Jul 21 14:59:47 reasonably considered dead Jul 21 14:59:54 Batteries also doesn't last forever as well.. the case is very evident when it comes to laptops. Jul 21 15:00:24 You don't see a 10 year old laptop with 10 year old battery working like brand spanking new from the shops. Jul 21 15:01:10 but honestly when your device boots up and acts in reasonable ways to input, just doesn't display on LCD anymore, then you do _not_ have battery issue Jul 21 15:01:47 it sorta boots, sometimes faint boot sound othertimes nothing Jul 21 15:01:48 when your device shuts down any time, by whatever nasty procedure, then you *might* have a battery problem Jul 21 15:01:53 I'm sure almost every PC technician goes by the general rule of thumb, and that is to try one thing at a time.. if this doesn't work try that, if that doesn't work try another. Jul 21 15:01:54 i think it's on last legs Jul 21 15:02:14 There's no secret about debugging hardware issues. Jul 21 15:02:22 DocScrutinizer05: does the n800 have the ability to run batteryless (plugged in) Jul 21 15:02:31 generally not Jul 21 15:02:34 ok Jul 21 15:02:47 so can't test that way... Jul 21 15:02:52 nope Jul 21 15:02:55 best buy has no bp-5l's Jul 21 15:02:57 the fuck Jul 21 15:03:04 otherwise we had suggested to you already Jul 21 15:05:01 To be frank I would leave DocScrutinizer05 to answer that one, I personally don't know of anyone that owns N800 out where I live. I've only seen very few N900 owners. Jul 21 15:05:21 N810 (BP-4) works with BL-5J as well, when you press it to the contacts Jul 21 15:05:32 (BP-4L) Jul 21 15:06:22 ok folks, you've been very helpful and i've been crabby (pissed at my "broken" hw) Jul 21 15:06:32 i'll get a replacement battery and start there. Jul 21 15:06:35 I feel with you Jul 21 15:06:53 if that doesn't work, then it's game over right? nokia won't service this dinosaur, yes? Jul 21 15:07:02 see if you maybe borrow a battery from another device that has identical battery contacts Jul 21 15:07:16 We all will probably go into denial at first with aging stuff. Its inevitable as we try to clamber on the remains. Jul 21 15:07:20 i tried from my phone and the contacts don't line up :/ Jul 21 15:07:42 they won't service it anymore I guess. But it can't hurt to ask - if worst case hits you Jul 21 15:08:03 You might be able to source N800 screen and go from there. The only question from here on in lies in how much money will you want to fork in on the device alone to resurrect it. Jul 21 15:09:58 mancha: I'd almost bet BP-4L and BP-5L have identical battery contacts, so BL-5J should work for you as well. It can be found in quite a number of devices from Nokia Jul 21 15:11:43 mancha: just *carefully* compare writing on your battery and on spare other model battery for same "signals" on same pins (GND or (-), (+), and some "BSI" or whatever on middle contact probably) Jul 21 15:11:56 oops nope!!! Jul 21 15:12:02 * psycho_oreos imagines a few years ahead in time one will probably see N900 dying/withering away. Jul 21 15:12:09 on middle is GND/- Jul 21 15:12:39 the "BSI" is opposite of (+) Jul 21 15:13:02 (+) is on corner of battery, BSI to the center Jul 21 15:13:43 you also can tell GND/- contact in device - it's a tiny bit longer than other two Jul 21 15:14:37 mmmpf, on N810 both + and - are longer, and BSI is the only shorter one Jul 21 15:15:09 probably same on N900 Jul 21 15:15:21 anyway, good luck Jul 21 15:17:05 it's not the battery Jul 21 15:18:01 So now the case continues on possibly the screen being the issue.. either the screen or otherwise. Jul 21 15:18:24 Err the screen, its connectors or otherwise. Jul 21 15:18:41 i have a bootloader on it. wonder if i should reflash with factory Jul 21 15:19:10 The idea of reflashing was suggested earlier on by DocScrutinizer05. Jul 21 15:19:17 when it's only the screen then you probably can reflash Jul 21 15:19:32 if it's only software then you can reflash Jul 21 15:20:14 if it's some problem with main hw components like CPU/SoC then you will face severe problems no matter what you try to do Jul 21 15:20:54 mancha: how did you make sure it's not the battery? Jul 21 15:23:35 i connected a gps battery which is also 3.7 volts and if i held it in place the contacts touched. Jul 21 15:23:43 the thing boots the same way...with artifacts. Jul 21 15:24:07 i will try reflashing now. Jul 21 15:24:20 Good Luck. Jul 21 15:24:22 yes, that rules out battery Jul 21 15:26:14 ~flashing Jul 21 15:26:15 i heard maemo-flashing is http://wiki.maemo.org/Updating_the_tablet_firmware Jul 21 15:26:29 for you: http://wiki.maemo.org/index.php?title=Updating_the_firmware&oldid=17120 Jul 21 15:27:19 you can use flasher-3.5 nevertheless Jul 21 15:31:46 reflash was successful and problem is unfixed. Jul 21 15:33:07 the thing's bricked Jul 21 15:34:35 B-( Jul 21 15:35:21 Bricked? o.O Jul 21 15:35:37 well i didn't brick it in the flash Jul 21 15:35:45 i mean the flash was successful and the problem remains. Jul 21 15:35:53 well, LCD or controller od SoC's video engine defect Jul 21 15:36:24 and that can just go bad? Jul 21 15:36:34 out of the fuckin' blue? Jul 21 15:36:40 or sth as simple as connectors Jul 21 15:37:04 I believe someone mentioned with electronics, its either 1 or 0, Alive or dead, there's no in-between. Jul 21 15:37:14 yes, chips can go defect in an instant Jul 21 15:37:43 Sad. Jul 21 15:38:17 Inevitably electronics are what humans made. Inperfect as they are just as we are inperfect ourselves. Jul 21 15:40:29 s/inp/imp/ Jul 21 15:40:30 psycho_oreos meant: Inevitably electronics are what humans made. Inperfect as they are just as we are imperfect ourselves. Jul 21 15:40:46 bah, case sensitive. Jul 21 15:41:05 and you forgot /g anyway Jul 21 15:41:23 abcabcabcabc Jul 21 15:41:32 s/b/B/ Jul 21 15:41:33 DocScrutinizer05 meant: aBcabcabcabc Jul 21 15:41:37 s/b/B/g Jul 21 15:41:37 DocScrutinizer05 meant: aBcaBcaBcaBc Jul 21 15:41:49 well. thanks all. this sucks. i'll go sulk now. Jul 21 15:42:00 be well. Jul 21 15:42:02 aBcabcAbc Jul 21 15:42:23 s/a/A/g Jul 21 15:42:24 psycho_oreos meant: ABcAbcAbc Jul 21 15:42:36 *shrugs* Jul 21 15:43:16 mancha, there's still hope. you could try with another LCD to see if its LCD that's faulty or if its really the SoC video engine dead. Jul 21 15:43:26 bah. Jul 21 15:58:44 hi how to change font in nokia n900?? Jul 21 15:59:44 anyone here? Jul 21 16:04:12 changing font on nokia n900?? Jul 21 16:07:16 ?? anyone here/ Jul 21 16:07:35 Nope Jul 21 16:07:42 Yes, but no Idea how to change the Font. Jul 21 16:07:50 It's normal X11, so it should not be _too_ difficult. Jul 21 16:08:07 I think the Terminal App has Functionality to change the Font it uses. Jul 21 16:08:55 theme customizer changes the font. but i need to change the font of the textbox in vkb Jul 21 16:09:30 rest, as far as i can see, font has been changed xcpt for the textbox of vkb Jul 21 16:12:06 Have you tried searching/posting thread on tmo as well? Jul 21 18:32:25 DAMN!! was it THAT qwerty123? Jul 21 18:32:43 prolly not Jul 21 18:33:03 I also guess he been called qwerty12 Jul 21 18:34:32 I somehow doubt it'll be him. (qwerty12) Jul 21 18:34:40 yeah Jul 21 18:38:43 dat freenode Jul 21 18:46:21 uh? Jul 21 19:27:12 hi Jul 21 19:27:25 well this os issue I had earlier Jul 21 19:27:37 was apparently caused *solely* by the optfs restoration Jul 21 19:27:51 because, now, the "normal" issue of that just returned Jul 21 19:28:01 i.e. the one where reboots solved it Jul 21 19:28:26 set PYTHONHOME manually Jul 21 19:28:30 nothing Jul 21 19:28:33 I checked the lib directory in case it was missing Jul 21 19:28:47 and found something interesting - there was *nothing* there Jul 21 19:28:51 checked mountpoints Jul 21 19:29:20 and for some reason python failed to be mounted correctly, because all python mountpoints isn't there (none is) Jul 21 19:29:34 i.e. the bindmounts to /opt/pymaemo Jul 21 19:30:08 issuing "mount" fixes this issue Jul 21 19:30:20 but what causes it in first place? Jul 21 19:33:11 I also noticed that this was the first time (since then) where the clock was reset to jan 1 Jul 21 19:33:16 perhaps that's related in some way? Jul 21 19:44:00 hi again Jul 21 19:44:27 http://sprunge.us/fJJc perhaps that can give any clue to why python stuff isn't mounted occasionally but I don't know Jul 21 19:46:32 (also, if you know a better method to get a single session than cat -n syslog | grep syslog, see the relevant line number, and do cat syslog | tail -n +4890 | pastebin, that would be nice to know of lol) Jul 21 19:51:17 the last reboot was fully clean (battery died, I swapped battery, and started up again) Jul 21 19:59:10 uh huh? Jul 21 20:17:08 freenode has issues it seems Jul 21 20:31:42 so.. any ideas regarding my issue? Jul 21 20:31:54 dunno wtf happened with freenode lol Jul 21 20:34:20 stupid kids Jul 21 20:34:23 would be my guess Jul 21 22:26:15 DocScrutinizer51, hey, the extras-devel repository has some md5/sha1sum checksum error , see https://pastee.org/9u47z Jul 21 22:30:26 thanks Jul 21 22:30:41 merlin1991: ^^^ Jul 21 22:34:48 hmm Jul 21 22:34:57 I've used a N900 for umm Jul 21 22:35:04 when was it released? Jul 21 22:35:30 end of 2010 or something like that, wasn't it? Jul 21 22:37:13 oh wait Jul 21 22:37:17 it was 2009 Jul 21 22:37:19 omg Jul 21 22:37:31 I've used this thing for over three years Jul 21 22:37:37 that's crazy lol Jul 22 00:26:51 you know what's crazy? that you can't get repair for those devices after actually only 2 years Jul 22 00:34:02 you couldnt get repairs before those 2 years Jul 22 00:34:04 only replacements Jul 22 00:54:43 FIQ|n900, ping Jul 22 00:58:02 psycho_oreos: pong Jul 22 00:59:31 tilt Jul 22 00:59:42 FIQ|n900, I can't quite see any reasons why /opt/python would fail. Though that last link you posted contains fair bit of private info :) Jul 22 01:01:57 psycho_oreos: mhm? Jul 22 01:02:11 looked at it a little before pasting Jul 22 01:02:27 didn't see anything critical though perhaps I missed something Jul 22 01:03:13 Somehow I don't think failed mounts (particularly at bind mounts) would be shown via syslog. Jul 22 01:03:54 psycho_oreos: and DocScrutinizer05 asked about a syslog dump earlier and I assumed it didn't contain anything critical by that Jul 22 01:04:35 private info how exactly? Jul 22 01:05:22 FIQ|n900, no doubt he did though to this point I believe it was to tell you that we basically needed a lot more information when MH was only partially working. I can't see any information about python issues in syslog personally. Jul 22 01:05:48 ok Jul 22 01:06:38 Your IMEI is visible. There's a few other interesting but not exactly private info, such as your bootup video, your (failed) desktop wallpaper preferences (the images probably failed), and your WAP info. Jul 22 01:06:42 DocScrutinizer05 mentioned something about a potential bug where python bindmounts would fail if a fsck was ran on the partition Jul 22 01:06:59 psycho_oreos: ok Jul 22 01:07:12 my desktop wallpaper is fully intended to not show up Jul 22 01:07:23 I don't care about boot video information Jul 22 01:07:33 IMEI is a bit more critical though :p Jul 22 01:07:37 good to know... Jul 22 01:07:57 FIQ|n900: that been wild speculations Jul 22 01:08:09 DocScrutinizer05: ok Jul 22 01:08:26 Yeah but I think the underlying fault would be to do with /home /opt being the main premise. If that were to occur, /home and /opt would not be mounted imo. Jul 22 01:08:36 "your WAP info" i.e. apn info? Jul 22 01:09:17 isn't this just telling what my operator is? because that's not something I would consider private Jul 22 01:09:49 I said in theory the patch to fstab in CSSU-T might have changed the timing of bindmounts, so in ever farther fetched theory, a python command could run and fail during that time where now python bindmounts are not done while formerly they were Jul 22 01:09:56 allthough perhaps it was silly by me to paste the entire syslog, I dunno Jul 22 01:10:27 FIQ|n900, no not your wallpaper.. I'll give you a bit of info: hildon-desktop[1149]: GLIB WARNING ** default - Error loading cached background image /home/user/.backgrounds/background-6.***. Misslyckades med att öppna filen "/home/user/.backgrounds/background-6.***": No such file or directory Jul 22 01:11:17 Yeah APN info, and your IP address, etc was also visible. Also your SMS centre number. Jul 22 01:11:41 should be intended @ background failing Jul 22 01:15:13 psycho_oreos: regarding desktop background - since I installed cssu, I've walked into desktop settings and removed all desktop views except for number #5 which gave a completely black background as the image was only adjusted for 4 Jul 22 01:15:26 not really what the thing was intended for, but meh :p Jul 22 01:15:32 perhaps that's something else though Jul 22 01:15:51 * psycho_oreos is still quite stumped at the issue. So nothing exists in those binded mounts? and those mounts for python doesn't exist meaning that maybe they're not in fstab or something? Jul 22 01:16:13 but they are Jul 22 01:16:19 issuing "mount" fixes the issue Jul 22 01:16:45 psycho_oreos: well obviously I didn't realize that the info was quite critical but oh well, can't undo it now :P Jul 22 01:16:58 and it wasn't too bad, could have been worse lol Jul 22 01:17:02 FredrIQ|n900, ahh though it goes from 6-9 with that error so yeah. Again not really private info but was somewhat interesting to see (I mean from the perspective of another user with 9 home screens) :D Jul 22 01:17:35 :P Jul 22 01:17:52 I don't use the desktop at all Jul 22 01:17:57 FredrIQ|n900, *nods* oh well you learnt something ;) not that I care about all the private info anyway. Jul 22 01:18:02 that's why I wanted no (or black) bg Jul 22 01:18:46 issuing "mount" fixes issue.. wait.. by mount as in just "mount" with no args? by default "mount" with no args just lists the mounted fs. Jul 22 01:19:13 psycho_oreos: yes, I found this quite strange that it did fix the issue... Jul 22 01:19:46 perhaps it was a coincidence but, mounting being that delayed (startup, attempting mh and failing, firing up terminal) doesn't make any sense... Jul 22 01:20:18 sec Jul 22 01:20:30 k Jul 22 01:20:55 FredrIQ|n900, I somehow don't think its actually coincidential. If you have something like cpumem-applet installed you might be able to see if N900 was still busy loading up or not. Jul 22 01:20:57 (tried to fire up MH to see if the error was apparent right now to be able to debug more, but MH worked fine) Jul 22 01:21:26 psycho_oreos: hmm ok Jul 22 01:21:41 well I'll try to look into it a little next time it fails Jul 22 01:22:07 FredrIQ|n900, also install useful applet like the one I mentioned :) its priceless for power-users imho. Jul 22 01:26:49 There were some notable warnings from dbus within that syslog and then later on issues with ISI_SMS. Whether or not do they play a part in making the device seem sluggish and/or to make "mount" command itself coincidentally fixes up issues with python is a bit beyond me. Though it is probably notable that I'm guessing there are commands/things that the device has to do, the "mount" command itself can't quite possibly automagically mount all the pyt Jul 22 01:26:50 hon binding mounts. It simply just doesn't add up imo Jul 22 01:27:36 I weird idea: umount those python bindmounts and *then* check what's inside the mountpoint dirs Jul 22 01:27:50 psycho_oreos: indeed it didn't make any sense Jul 22 01:28:06 DocScrutinizer05: well I did that indirectly before, didn't I? Jul 22 01:28:10 they was empty Jul 22 01:28:29 didn't follow the details Jul 22 01:28:32 I guess I can check after umounting manually though Jul 22 01:29:54 FredrIQ|n900, its not like the optification somehow automagically hacked the "mount" binary to mount any unmounted stuff needed for python. Another possibility is I/O lag, lag from within the data throughput that is causing the python mounts to not come up instantaneously. Jul 22 01:31:45 psycho_oreos: indeed - it doesn't make any sense Jul 22 01:31:57 so perhaps it just was a coincidence Jul 22 01:32:30 I'm still willing to place my money it isn't a random artifact, there has to be a meaningful explanation. Jul 22 01:33:15 also, checked directories after umounting - empty Jul 22 01:33:37 also tried to issue "mount" - and it did not do anything funny (as expected) Jul 22 01:33:50 so it seems to have been a coincidence then Jul 22 01:34:36 When for the first time you issued that "mount" command, was there any sort of lag? even if it was just a mere 1 second lag? Jul 22 01:35:08 not that I recall Jul 22 01:35:33 Hmm :/ Jul 22 01:36:13 but it's possible that, after my check of directories finding them empty, that I timed the mount command right after they were mounted Jul 22 01:36:27 which would produce such behaviour Jul 22 01:36:33 btw what's isi_sms? Jul 22 01:37:46 Was it at the time when you checked it was empty, the device was still booting up? Jul 22 01:38:06 Not particularly sure, its something to do with text messages lol. Jul 22 01:38:21 Or rather I guess its something to do with text messages :x Jul 22 01:40:15 sounds as a legit guess given "(...)sms" Jul 22 01:41:06 I know there was some guy playing with things like that before, he was the one that got the cell info to work with cssu. I can't recall his handle anymore :/ Jul 22 01:41:19 psycho_oreos: I booted up, entered pin, (tried to) start MH and failing, immediately fire up the terminal and checked if the error was the usual, and then tried to do similar debugging steps as done last time when talking in here Jul 22 01:41:36 Apparently the stuff he was playing also started with ISI, so he might have a clue with ISI might stand for. Jul 22 01:42:17 including setting PYTHONHOME (which didn't help at all), checking directories (which were empty) and then checking mounts with "mount" Jul 22 01:42:19 and you know the rest Jul 22 01:42:42 FIQ|n900, from between after entering the pin correctly to starting MH, how many seconds did you wait? Jul 22 01:43:27 the device lags slightly at first when booted so I usually wait half a minute before doing anything Jul 22 01:43:35 which was the case here too Jul 22 01:44:32 Yeah, I was thinking you didn't give the device anytime to finish loading up properly hence that could be the reason why its still loading hence failed. Then again without cpumem-applet its hard to rule that out. Jul 22 01:45:00 well Jul 22 01:45:05 I have that thing now Jul 22 01:45:12 and if this happens again Jul 22 01:45:19 (and given previous experience - it will) Jul 22 01:45:32 I might try simply waiting Jul 22 01:45:37 before doing stuff Jul 22 01:46:00 Hopefully it won't be just a coincidence with a simple "mount" command :) I'm just really not willing to rest it on pure random "luck" factor. Jul 22 01:46:05 but my assumption was that mounts was taken care of way before hildon even starts up... Jul 22 01:46:45 I'm not sure, the proper (normal) fs mounts are done well beforehand.. notably the binding mounts were done afterwards and even with python. Jul 22 01:47:08 ok Jul 22 01:50:19 I just found the upstart script, it starts after loading rcS-late. Jul 22 01:51:13 it says in /etc/fstab that it's generated by rc--late Jul 22 01:51:14 fstab that is Jul 22 01:51:25 *rcS-late Jul 22 01:52:36 /etc/event.d/rcS-late starts after hal is fully started. The contents shows that it mounts home dir. Jul 22 01:53:18 Correction, home, MyDocs and swap. Jul 22 01:54:13 In other words the script pymaemo-optify should have started quite early. Jul 22 02:01:20 Hmm, I think I recall the guy who tinkered with cell tower info. Jul 22 02:01:38 s/ll/lled/ Jul 22 02:01:40 psycho_oreos meant: Hmm, I think I recalled the guy who tinkered with cell tower info. **** ENDING LOGGING AT Mon Jul 22 02:59:58 2013