**** BEGIN LOGGING AT Wed Jul 22 02:59:59 2015 Jul 22 03:26:58 What does the nokia flashing tool do? Jul 22 03:27:01 Anyone reversed it? Jul 22 04:46:45 ~0xffff Jul 22 04:46:45 0xffff is probably https://github.com/pali/0xFFFF, or http://talk.maemo.org/showthread.php?t=87996 Jul 22 04:46:48 R0b0t1: ^ Jul 22 04:53:54 yas Jul 22 07:42:33 Yesterday my device bricked. Its battery discharged completely, while I was uninstalling custom kernel, no matter charger was plugged. Jul 22 07:43:53 I got a new battery, but device cannot start any more. Tried to reflash it completely, without success. Jul 22 07:45:16 I only see nokia logo on screen, sometimes with led backlit on, sometimes off. Jul 22 07:46:15 as long as usb works and you didnt overwrite nolo (bootloader) you should be fine Jul 22 07:46:32 before trying to turn it on, recharge it for a while, disconnect charger and then turn it on Jul 22 07:46:54 I have reflashed it before, with the same image on the same pc with flasher 3.5, and I'm pretty sure i do things right. Jul 22 07:48:12 Already tried to recharge it the whole night, before flashing. The battery seems ok. Jul 22 07:48:29 is the led solid green ? Jul 22 07:48:54 It sometimes flashes white, sometimes solid yellow Jul 22 07:49:30 solid yellow means the chip is charging it slowly with no software support Jul 22 07:49:55 which usually means it coulnt boot the linux kernel, either because battery is too low, or because it just cant Jul 22 07:51:28 Battery stays about 3,8V loaded with a 300mA load Jul 22 07:52:24 if your kernel is borked I guess it really cant boot it ... could explain that Jul 22 07:53:03 But why cannot enter the flash mode. Jul 22 07:54:52 Is there a way to flash it anyway. I can even unsolder the flash chip, I think it was toshiba, and program it from the outside Jul 22 07:55:14 no, you should be able to flash over usb Jul 22 07:55:18 its pretty resilient Jul 22 07:55:21 Eg with gallep5 programmer Jul 22 07:55:58 1/ turn off device and disconnect usb, 2/ run flasher and get 'waiting', 3/ connect device without turning it on Jul 22 07:57:06 3rd step while holding u button Jul 22 07:57:17 doesnt matter Jul 22 07:58:52 I have tried it, with u key. Flasher gives info about hw version and then continue waiting for suitable device. Jul 22 08:02:51 using linux or windows? Jul 22 08:02:54 Im pretty sure i do things ok. I have done the same last week, everything was ok. Jul 22 08:03:02 Winxp 32 bit Jul 22 08:03:05 try linux Jul 22 08:04:17 I tried first last week, but I got error for ab APE algorithm, then get the job done on my old laptop. Jul 22 08:04:31 I use debian testing 64bit Jul 22 08:06:09 And flasher 3.5 Jul 22 08:06:38 From the debian.org repos Jul 22 08:09:21 as i said, try linux flashing Jul 22 08:10:09 ERROR: APE algorithm has to be provided to flash all the subimages Jul 22 08:10:30 But how to resolve this. Jul 22 08:11:23 must I download flasher deb package from somewhere else. Jul 22 08:11:27 check if your image is ok Jul 22 08:11:31 ~flasher Jul 22 08:11:32 i heard flasher is at http://www.jedge.com/n810/flasher/maemo_flasher-3.5_2.5.2.2.tar.gz (also .exe!), or http://www.chakra-project.org/ccr/packages.php?ID=5027 or generally http://www.google.com/search?q=maemo_flasher-3.5_2.5.2.2.tar.gz. HARMattan(N9): https://aur.archlinux.org/packages/fl/flasher-harmattan/flasher-harmattan.tar.gz, or -- list of filenames/md5sums: http://pastebin.com/sYKdNJSH, or http://galif.eu/nokia/ Jul 22 08:14:05 My image should be ok. The same I used on my laptop. Jul 22 08:14:21 http://talk.maemo.org/showthread.php?t=75000 ? Jul 22 08:14:43 are you sure it's flash 3.5 ? Jul 22 08:14:46 flasher* Jul 22 08:15:10 I will try with the jedge version, and tell you later Jul 22 08:15:38 just grab the one from infobot Jul 22 08:15:39 Yep, I ve read this already. Jul 22 08:18:35 the APE error is most likely not related to your device Jul 22 08:22:03 As far as Ive read in the net, its some incompability of 32 and 64 linuxes and their libs, but nobody gave solution. Jul 22 08:22:24 run livecd 32bit ? Jul 22 08:24:48 or try installing 32b libs (multilib install), or run it in a chroot Jul 22 08:26:10 /32h/32 Jul 22 08:26:53 yeah, if you use flasher designed for harmattan (ie >3.5) it will fail Jul 22 08:35:02 I have ubuntu 12.04 32 somewhere, if jedge version fails, I will try it. Jul 22 08:43:51 DocScrutinizer05: hey may Jul 22 08:43:56 s/may/man/ Jul 22 08:43:56 teotwaki meant: DocScrutinizer05: hey man Jul 22 08:44:14 may flower Jul 22 08:48:32 or james Jul 22 08:49:40 Nokia doesn't even provide the flasher for n810/n900 anymore? Jul 22 08:49:56 nope, nor images Jul 22 08:50:02 ~flashing Jul 22 08:50:02 from memory, maemo-flashing is http://wiki.maemo.org/Updating_the_tablet_firmware, or - on linux PC - download&extract http://maemo.cloud-7.de/maemo5/patches_n_tools/maemo-my-private-workdir.tgz, cd into it, do sudo ./flash-it-all.sh Jul 22 08:50:12 who owns maemo.org? Jul 22 08:50:14 (the domain) Jul 22 08:52:54 according to the whoisdb juiceme is the admin contact Jul 22 08:53:01 and the registrant Jul 22 08:54:31 interesting Jul 22 08:54:44 Well, thanks Microsoft. You've helped us all. Jul 22 08:55:27 nope, it was pre-microsoft Jul 22 08:57:19 KotCzarny: pre-Elop? Jul 22 08:58:10 yeah Jul 22 09:02:24 ok, my bad MS, then. Jul 22 09:02:33 Thanks Nokia board, you've helped us all. Jul 22 09:20:06 i mean, it was migrated in 2013, but i guess it wasnt on elop's 'do or dont list' Jul 22 09:25:02 Found device RX-51, hardware revision 2101 Jul 22 09:25:02 NOLO version 1.4.14 Jul 22 09:25:02 Version of 'sw-release': Jul 22 09:25:02 Booting device into flash mode. Jul 22 09:25:02 Suitable USB device not found, waiting. Jul 22 09:25:41 turn it off, connect usb Jul 22 09:25:59 turn it off, remove battery, connect usb, hold U, put battery in Jul 22 09:27:39 Linux debian 3.16.0-4-amd64 #1 SMP Debian 3.16.7-ckt11-1 (2015-05-24) x86_64 GNU/Linux Jul 22 09:27:52 Using flasher from the first link Jul 22 09:28:17 it doesnt get error message about APE algoritm Jul 22 09:28:36 I have tried it many times Jul 22 09:29:55 with or without pressing U Jul 22 09:30:07 try linux32 ? Jul 22 09:31:53 and I don't see such device with lsusb, only my camera, kb ... Jul 22 09:34:06 try without turning it on? Jul 22 09:36:43 yes, I already tried Jul 22 09:37:53 I may have forgotten to mention, if I leave the phone, it's in a reboot loop, untill it has battery or is on chanrger Jul 22 09:39:46 I can see Nokia logo, but without backlit, in the next moment there is only backlit with black sreen Jul 22 09:42:42 Found device RX-51, hardware revision 2101 - I suppose the USB itself is ok by this Jul 22 09:43:52 can you boot rescueos? Jul 22 09:45:34 I'm not aware of it. Jul 22 09:45:44 Just downloaded the image Jul 22 09:45:59 What to do next Jul 22 09:47:13 flasher-3.5 -k rescueOS_n900_kernel_1.2.zImage -n rescueOS-1.2.img -l -b"rootdelay root=/dev/ram0" Jul 22 09:47:32 it will try to load kernel without flashing (note the -k option) Jul 22 09:48:47 just tried with partedmagic 3.5.3 32bit, no luck Jul 22 09:48:56 partedmagic? Jul 22 09:49:22 this I found easiest to boot to Jul 22 09:49:27 32bit Jul 22 09:49:34 ahm, some livecd Jul 22 09:51:23 so, let me rehash the scenario, 1/ disconnect and poweroff the device, 2/ run flasher line and wait for 'waiting for' 3/ connect usb without powering on Jul 22 09:51:27 this fails, right? Jul 22 09:53:32 yes Jul 22 09:54:42 yes Jul 22 09:55:01 and it seems my battery drains quite fast Jul 22 09:55:32 i think i must recharge it, now phone doesnt start and it led lits solid yellow Jul 22 09:56:47 i try all with su privilegs, is this ok Jul 22 09:58:12 yeah Jul 22 10:02:55 alphadog, you can try script from Jul 22 10:02:57 ~flashing Jul 22 10:02:57 rumour has it, maemo-flashing is http://wiki.maemo.org/Updating_the_tablet_firmware, or - on linux PC - download&extract http://maemo.cloud-7.de/maemo5/patches_n_tools/maemo-my-private-workdir.tgz, cd into it, do sudo ./flash-it-all.sh Jul 22 10:03:04 the one names flash-it-all Jul 22 10:03:20 *named Jul 22 10:04:19 also, you might try another computer Jul 22 10:04:23 or cable Jul 22 10:04:31 maybe there is something wrong hardware wise Jul 22 10:27:37 hi KotCzarny Jul 22 10:28:19 oink zgrr Jul 22 10:57:15 KotCzarny: i might have a time soon to have a look at your brilliant media player. Jul 22 10:57:28 KotCzarny: it does work on n810, doesn't it? Jul 22 11:04:06 I will try flashitall later at home, 10x Jul 22 11:07:03 Hi everyone :-) Jul 22 11:08:06 I'm looking for a package, cannot find the appropriate repository: syncevolution: Depends: libpcrecpp0 (>= 6.7-1osso1+r1) but it is not installable (virtual package, no candidate) Jul 22 11:08:25 Any suggestions? extras, -testing and -devel are enabled. Jul 22 11:10:08 zgrr: yes, latest is in extras-devel Jul 22 11:10:31 fantastisch Jul 22 11:34:50 goldkatze: http://repository.maemo.org/pool/fremantle/free/p/pcre3/ Jul 22 11:35:44 uh, libpcrecpp0 stored in pcre3 folder?! okay. thank you so much! Jul 22 11:37:10 vectis__: Is this part of the SDK repo? Where do I find those for sources.list? I'm rather confused as to why it's not being found by aptitude already as it is right now. Jul 22 11:37:20 debian, ehehe Jul 22 11:37:23 ~maemo-repos Jul 22 11:37:24 it has been said that maemo-repos is http://wiki.maemo.org/Repository#List_of_Maemo_repositories Jul 22 11:37:38 and yes, its for scratchbox (sdk) Jul 22 12:11:11 http://maemo.org/packages/view/libpcre3/ Jul 22 12:11:17 it should be in SSU as well Jul 22 12:11:55 (ie the nokia repositories) Jul 22 12:15:30 bencoh: Is muarf still the mirror to use for those? Since nokia is gone? Jul 22 12:17:33 well, it should work Jul 22 12:18:01 there might be other repositories, and there is no "official" one Jul 22 12:20:48 ok, thanks :-) Jul 22 14:36:06 ~flashing Jul 22 14:36:06 well, maemo-flashing is http://wiki.maemo.org/Updating_the_tablet_firmware, or - on linux PC - download&extract http://maemo.cloud-7.de/maemo5/patches_n_tools/maemo-my-private-workdir.tgz, cd into it, do sudo ./flash-it-all.sh Jul 22 14:44:19 KotCzarny, I need help :) Jul 22 14:44:26 lol Jul 22 14:44:29 what kind? Jul 22 14:44:42 i managed to boot Rescue OS Jul 22 14:44:45 :) Jul 22 14:44:53 now what should I do Jul 22 14:44:58 good, that was just to check if your flasher/device works Jul 22 14:45:15 now close the rescueos and flash the image from the save env Jul 22 14:45:16 :) Jul 22 14:45:39 with vanilla image i suppose Jul 22 14:45:47 ~flashing Jul 22 14:45:48 well, maemo-flashing is http://wiki.maemo.org/Updating_the_tablet_firmware, or - on linux PC - download&extract http://maemo.cloud-7.de/maemo5/patches_n_tools/maemo-my-private-workdir.tgz, cd into it, do sudo ./flash-it-all.sh Jul 22 14:51:23 s/save/same/ Jul 22 14:59:05 It works :) Jul 22 14:59:11 woo-hoo? Jul 22 14:59:22 what was the working combination? Jul 22 14:59:56 the flasher from the link the bot gave Jul 22 15:00:12 :) Jul 22 15:00:20 the script flash it all Jul 22 15:00:54 and a bl-4j battery from my friend phone Jul 22 15:01:14 http://wiki.maemo.org/Updating_the_tablet_firmware Jul 22 15:01:28 so, it was bad battery that prevented flashing interface showing up? Jul 22 15:01:34 from this link I understood that Jul 22 15:02:04 on empty battery its impossible to recharge it from this phone Jul 22 15:02:14 :) Jul 22 15:02:18 yep :) Jul 22 15:02:58 one nice thing nokia did right is that you can hack and reflash till the flash chip dies Jul 22 15:05:20 yes, I found in all g results that hard reset is made with flasher, no *# code or menu option in settings Jul 22 15:05:40 because it's just a linux box Jul 22 15:05:58 no # code to reset, just reinstall Jul 22 15:07:40 yes, but it's not a big deal, as far as I know flash chip can stand up at least 10000 write cycles. Jul 22 15:08:58 And the chip I saw was Toshiba, so this number should be 100% guaranteed Jul 22 15:10:23 the only bad thing is that if your battery is completely dead, you are in a deadlock - you should use 5800 xpress music to recharge it :) Jul 22 15:10:32 yes Jul 22 15:10:54 but its not a big problem usually Jul 22 15:10:59 or have a second battery Jul 22 15:11:06 or a power supply Jul 22 15:11:16 or external charger Jul 22 15:11:41 But the worst is that I didn't know about this, and I would never think about this Jul 22 15:11:52 :) Jul 22 15:12:08 we all thought your battery is charged enough Jul 22 15:12:40 I thought so, I tried it with a potentiometer and voltmetter Jul 22 15:13:09 i will get an polarcell one :) Jul 22 15:16:49 also, rtfm Jul 22 15:16:50 :) Jul 22 15:25:52 now my polarcell is on the way Jul 22 15:30:57 KotCzarny, I owe you a beer, call me if you go to Bulgaria :) Jul 22 15:31:44 :) Jul 22 15:32:52 good that it worked Jul 22 15:34:06 I also found a brand new touchscreen here in Bulgaria Jul 22 15:34:55 for about 15 euro, but I doubt it's OEM, no matter site says "Original Nokia" Jul 22 15:46:30 ~flatbatrecover Jul 22 15:46:31 Remove battery for 1 minute. Insert battery. Plug powered ***NOKIA WALLCHARGER*** to device. Watch steady amber. Let sit and charge. Do NOT try to boot. After 30 min, you got either a) a booted up N900, b) flashing amber which means you can boot, c) steady amber shut off -> start over again with ~flatbatrecover while already searching for a new battery. CAVEAT! Only works when ~rootfs OK (no ~bootloop) Jul 22 15:46:42 ~bootloop Jul 22 15:46:42 i heard bootloop is when your device has broken rootfilesystem, so during reboot it fails on some service startup or kernel module load and thus reboots. This *drains* battery! And you can't reflash to stop bootloop when battery is drained. Recharge your battery by other means before reflashing. E.g. using ~rescueOS. Or external charger or BL-5J compatible other device. Jul 22 15:54:40 "CAVEAT! Only works when ~rootfs OK (no ~bootloop)" thhhat's te important part :) Jul 22 15:56:48 which wasnt, in this particular example Jul 22 16:00:37 ~rootfs Jul 22 16:00:38 rootfs: the content of the storage volume mounted under /, on NAND mtd5 partition. Also called "the system", since it's all the files forming the linux core system (except the kernel itself). Reflash rootfs from COMBINED (see ~flashing). Also see ~bootloop Jul 22 16:01:37 >>Also see ~bootloop<< is the relevant part Jul 22 16:02:02 is /home/user/.profile rootfs? Jul 22 16:02:11 because it also can cause bootloop Jul 22 16:02:12 :P Jul 22 16:02:27 that's right Jul 22 16:02:54 but afaik only when booting up, not in ACT_DEAD Jul 22 16:03:05 I might be wrong Jul 22 16:04:02 yes Jul 22 16:05:06 so you could get out of deadlock by simply letting device charge (which is actually what ACT_DEAD is all about) Jul 22 16:08:09 IOW a broken profile doesn't cause a real bootloop as mentioned in ~flatbatrecover Jul 22 16:09:05 it simply causes reboot when trying to boot device to full operational level, and reflashing is the cure to that Jul 22 16:38:43 hmm, *.m.o down? Jul 22 16:39:19 works here Jul 22 16:39:22 (?) Jul 22 16:41:36 +1 Jul 22 16:41:40 t.m.o works for me Jul 22 16:43:30 hmm not here Jul 22 16:43:43 check if dns resolves Jul 22 16:47:59 isup.me shows down too Jul 22 16:48:27 10. talk.maemo.org 0.0% 5 18.0 18.2 18.0 19.1 0.0 Jul 22 16:48:31 works fine Jul 22 16:49:17 from both my home and friends isps ranges Jul 22 17:08:59 wfm Jul 22 17:10:13 CUBE Jul 22 17:11:12 sixwheeledbeast: mtr ? Jul 22 17:11:18 hi xes, long time no see Jul 22 17:11:19 (the traceroute tool) Jul 22 17:12:53 hi DocScrutinizer05! RL is cruel leaving me without free time ;) Jul 22 17:13:20 same here, sort of Jul 22 17:21:12 I'll investigate further later, off out now. Seems odd only m.o is affected. Jul 22 17:21:53 m.o also works Jul 22 17:41:28 https? Jul 22 17:42:09 yes Jul 22 17:42:23 SSL cipher: 128-bit TLSv1/SSLv3 ECDHE-RSA-AES128-GCM-SHA256 Jul 22 17:47:50 lol, for https://talk,maemo.org i get connection refused Jul 22 17:48:37 wiki and main work Jul 22 17:48:53 talk works without https Jul 22 18:19:01 yes, even for https://talk.maemo.org Jul 22 18:24:57 i still get conn refused for https://tmo Jul 22 18:25:15 talk about chaos Jul 22 18:25:57 could be that this ip range got blocked for spam tho Jul 22 18:26:39 though for my home connection its also conn ref Jul 22 18:28:19 yes, I still get conn refused even for https://talk.maemo.org Jul 22 18:29:17 there is no https on tmo Jul 22 18:29:46 hmm, so passwords go plain? Jul 22 18:30:13 yes, just plain unencrypted passwords Jul 22 18:30:23 which can be easily sniffed Jul 22 18:30:40 fun (since they are used for whole *.maemo.org?) Jul 22 18:31:04 no TMO account is diffetent Jul 22 18:31:12 not SSOed with maemo.org Jul 22 18:31:15 wasnt there a merge? Jul 22 18:31:23 no Jul 22 18:31:25 never Jul 22 18:33:28 ~flashing Jul 22 18:33:29 methinks maemo-flashing is http://wiki.maemo.org/Updating_the_tablet_firmware, or - on linux PC - download&extract http://maemo.cloud-7.de/maemo5/patches_n_tools/maemo-my-private-workdir.tgz, cd into it, do sudo ./flash-it-all.sh Jul 22 19:46:31 what are the symptoms of broken Thumb? Jul 22 19:46:39 could it be responsible for GCC hanging? Jul 22 19:53:36 yes, thumb siErr could cause basically all sorts of erratic behavior Jul 22 19:54:03 though segv is most likely Jul 22 19:55:56 *most lucky Jul 22 19:57:55 SIGILL actually ;) Jul 22 19:58:09 sigbus? Jul 22 19:58:15 Luke-Jr: highly unlikely Jul 22 19:58:20 KotCzarny: no, SIGILL Jul 22 19:58:29 hmm Jul 22 19:58:37 as istructions are executed with wrong T bit value Jul 22 19:58:44 it's a repeatable hang of the process, that I can cleanly kill Jul 22 19:59:01 it is ARM-compied, ain't? Jul 22 19:59:04 yes Jul 22 19:59:13 all: there is a little update maintenance in progress, if you receive some kind of error accessing x.maemo.org please be patient a minute Jul 22 19:59:16 isn't there a second sierr which causes cpu not to invalidate cache when it needs to ? Jul 22 19:59:43 I'm tempted to blame musl and go back to glibc 2.19 until things sanitise Jul 22 19:59:47 bencoh: thumb-related? Jul 22 19:59:57 ARM errata 430973 Jul 22 20:00:00 l29ah had problems with musl Jul 22 20:00:05 and gcc Jul 22 20:00:07 and building Jul 22 20:00:15 bencoh: this is the "main" thumb sierr Jul 22 20:00:22 * Luke-Jr stabs glibc for dropping <2.6.32 support Jul 22 20:00:34 freemangordon: the one triggering the SIGILL should be the other one (?) Jul 22 20:00:39 yep Jul 22 20:01:11 there is another one, something 6xxxxx, which was related to branches Jul 22 20:01:31 Branch Target Buffer ? that's 430973 :) Jul 22 20:01:41 657417 is https://sourceware.org/ml/binutils/2009-05/msg00343.html Jul 22 20:01:58 exactly Jul 22 20:02:43 but both result in SIGILL, because of the wrong context Jul 22 20:02:46 they're both funny as an external spectator ... I guess they're awful to debug :] Jul 22 20:02:57 oh, ok Jul 22 20:03:19 that's even more vicious :] Jul 22 20:05:21 vicious delicious Jul 22 20:24:50 ~flashing Jul 22 20:24:54 maemo-flashing is, like, http://wiki.maemo.org/Updating_the_tablet_firmware, or - on linux PC - download&extract http://maemo.cloud-7.de/maemo5/patches_n_tools/maemo-my-private-workdir.tgz, cd into it, do sudo ./flash-it-all.sh Jul 22 20:25:12 ^^^ slightly updated (option to flash rootfs only) Jul 22 21:11:58 There is a bug with optifying claws-mail-pgp-plugins on the N900. It ends up moving the file to /opt, creating a link to it, and then at some point the original file gets replaced with a circular link to itself. Jul 22 21:50:45 ouch Jul 22 21:51:29 excellent spotting Jul 22 21:53:59 It looks like it installs the files to /opt properly, and then in post install replaces them with symbolic links pointing to where it just installed them. It will probably work fine if the post-install optify is removed. Jul 22 21:56:24 yep, optifying stuff in /opt is contraproductive Jul 22 21:57:28 probably they ran optify on the build before packaging it, incl the optify hook Jul 22 21:58:28 In the post-install it says before the optify function "# Added by maemo-optify, with apologies.", so it looks like it was added after. Jul 22 21:59:18 If it was auto-added to a bunch of packages, it should probably include a check to make sure it isn't creating symbolic links for a file which is already getting installed to /opt Jul 22 22:00:32 This might not be the only package that is broken this way. Jul 22 22:00:48 I dunno if it's auto-added or that's an option you hand to the builder Jul 22 22:01:07 I rather guess the latter Jul 22 22:01:59 Your guess is better than mine; I don't really know much about the package install process. Jul 22 22:04:20 you prepare some file that has a number of info about author, package name, package icon, description, and prolly also whether to run optify on it or not Jul 22 22:06:24 Well, I dissected the package and didn't find anything pertaining to it other than a defined optify function in the "postinst" file. Jul 22 22:07:08 Should I contact the maintainer, or does someone here have the ability to fix the package? Jul 22 22:09:13 /debian/control Jul 22 22:09:22 prolly is that file Jul 22 22:10:08 only maintainer may fix a package Jul 22 22:10:32 I don't see anything in it that mentions /opt or optify, unless you count "Priority: optional". Jul 22 22:10:39 Okay. I will send the maintainer an email. Jul 22 22:11:37 ugh /debian/rules Jul 22 22:12:51 No /debian/rules file in the package. Only ones are the installed files, and control, md5sums, and postinst. Jul 22 22:14:00 sure, in the package are not instructions how to build the üackage Jul 22 22:15:54 but he wonders how it can be possible that it mangles on install Jul 22 22:17:11 starhash-enabler-for assistent-upload.tar.bz2/utar://starhash-enabler_0.1.tar.gz/utar://starhash-enabler/debian/rules Jul 22 22:19:26 KotCzarny: because it's optified twice, once pre-build and once at install-time Jul 22 22:21:00 the install-time optification seems to be auto-added by autobuilder to the postinst file, according to instructions given to the autobuilder in one of the files I mentioned above Jul 22 22:23:18 actually install-time optification is preferable over build-time optification since it allows easy deoptification on devices that have a different partitioning scheme Jul 22 22:24:29 though I guess the main reason been to allow optification of already installed packages from pre-optification (PR1.1) times Jul 22 22:28:41 Though, it will make install fail if you don't have much space on /. And you end up writing to the NAND unnecessarily. Jul 22 22:30:53 Email sent to maintainer. Jul 22 22:33:05 for a temporary "fix" rename the on-device optify function that's called in postinst and instead have an empty (only shebang) shellscript named like the optify function Jul 22 22:33:15 before installing Jul 22 22:33:51 I'm doing same with docpurge permanently Jul 22 22:34:44 Since there are only a few files, I will probably just move the files to their proper location manually rather than repackaging the deb file. Jul 22 22:35:07 who said repackaging? Jul 22 22:35:45 I mean editing it as an archive. Jul 22 22:36:02 nobody suggested that either Jul 22 22:37:33 I don't understand your suggestion then. Why would I edit postinst if I am not going to use it as a deb file? Jul 22 22:37:41 or is optify literally a function() in postinst, rather than a executable called by postinst? Jul 22 22:37:54 It is literally a function(). Jul 22 22:38:02 duh, crap Jul 22 22:38:59 It is also the only thing in postinst, so I don't really need it. Jul 22 22:39:14 echo auto >> debian/optify Jul 22 22:39:18 I thought it calls sth like /usr/sbin/maemo-optify-lib.sh Jul 22 22:39:19 (in the source package) Jul 22 22:40:15 Nope, it doesn't call anything except standard POSIX stuff. Jul 22 22:43:42 gregoa: it double-optifies, rendering the /opt/*/* files as links pointing to themselves Jul 22 22:45:11 gregoa: so rather rm debian/optify, but prior to upload to autobuilder Jul 22 22:45:30 I guess... Jul 22 22:46:35 or remove the line "echo auto >> debian/optify" from some file Jul 22 22:47:09 meh, no idea, I never looked into all that Jul 22 22:49:45 ZetaR: anyway then the easiest fix is prolly to extract the files missing in /opt (resp that got replaced by auto referential links) from claws.deb with mc and copy them to where they belong to Jul 22 22:51:06 Right, that is what I am doing now. Jul 22 22:56:52 (cd "/" && tar cf - "$f") | (cd "/opt/maemo/" && tar xf -); ln -s "/opt/maemo/$f" "/$f" Jul 22 22:56:55 hmmmm Jul 22 22:57:44 where $f = $1 Jul 22 22:58:17 in postinst. That's pretty weird, I don't think postinst gets called for each file of a package Jul 22 22:59:02 It is a function with no call to it in postinst. Jul 22 22:59:14 So it only defines optify() Jul 22 22:59:16 oooh even weirder Jul 22 22:59:45 so wtf? Jul 22 22:59:51 I have no idea. Jul 22 23:00:55 gregoa: http://repository.maemo.org/extras-devel/pool/fremantle/free/c/claws-mail/claws-mail-pgp-plugins_3.7.3-1maemo4_armel.deb Jul 22 23:01:28 gregoa: any idea? Jul 22 23:04:11 it's already "optified" in package: http://wstaw.org/m/2015/07/23/plasma-desktopal1947.png Jul 22 23:12:28 who the fuck calls a function optify () {...} in postinstall? Jul 22 23:13:23 clueless people? Jul 22 23:13:31 no Jul 22 23:13:49 people can't call functions, programs call functions Jul 22 23:13:58 indirectly Jul 22 23:14:01 Lol. Jul 22 23:14:46 also, who asks for person Jul 22 23:14:56 I mean, dpkg calls postinst I'd guess. But how would dpkg have an idea of a function named optify() ? Jul 22 23:15:51 I thought dpkg calls postinst as a shellscript and that's it Jul 22 23:16:07 "The maemo-optify tool helps developers to prepare Debian packages that use /opt. This tool moves selected files inside the package to locations under /opt, and will symbolically link from the original location to the new place of the file. It is available on gitorious as well as in Extras-devel" From https://wiki.maemo.org/Documentation/Maemo_5_Developer_Guide/Packaging,_Deploying_and_Distributing/Installing_under_opt_and_MyDocs Jul 22 23:16:29 yeah Jul 22 23:16:55 It says in postinst that optify() : "# Added by maemo-optify, with apologies." Jul 22 23:16:59 never thought they patched dpkg to add optify Jul 22 23:29:30 Seems to be working after moving files around. Jul 22 23:29:39 \o/ Jul 22 23:30:11 I haven't tested it yet, but I will work on that next. Jul 22 23:30:16 I'm even more puzzled since the dang function starts with a "if [ -h $f ]" Jul 22 23:30:33 You missed a "!". Jul 22 23:30:42 yeah, meant that Jul 22 23:30:54 it shpouldn't touch the symlinks anyway Jul 22 23:30:58 So it only moves it around if it exists and is not a symbolic link. Jul 22 23:31:16 http://wstaw.org/m/2015/07/23/plasma-desktopfM1947.png Jul 22 23:32:10 so who the heck is ruining that package? Jul 22 23:32:17 after installation Jul 22 23:32:28 and how? Jul 22 23:32:50 does apt have logging function? Jul 22 23:35:54 There is this weird file: "/var/log/maemo-optify-boottime.log" Jul 22 23:40:18 the function supposed to work on dirs? Jul 22 23:40:38 if [ -d "/$f" ] Jul 22 23:41:04 in package the files are symlinked, not the complete dir Jul 22 23:41:49 http://wstaw.org/m/2015/07/23/plasma-desktopal1947.png Jul 22 23:48:20 this is the root cause of the problem, the devel symlinked the files, then optify rushing over it and copying&symlinking the complete dir Jul 22 23:49:19