**** BEGIN LOGGING AT Wed Jan 25 03:00:02 2017 Jan 25 08:29:46 hmm curious ... https://www.arduino.cc/download_handler.php?f=/arduino-1.8.1-linuxarm.tar.xz Jan 25 08:29:54 wonder if that'll work on n900 :) Jan 25 08:30:06 guess i should install easydebian or something? Jan 25 08:31:37 what's that ? Jan 25 08:33:16 is that 'easy-deb-chroot' ? Jan 25 08:33:28 bencoh: its the Arduino IDE, they now have a arm release Jan 25 08:33:58 why would you want that on n900? Jan 25 08:34:07 for infield reprogramming Jan 25 08:34:13 with less hardware ;) Jan 25 08:34:44 The environment is written in Java and based on Processing and other open-source software. Jan 25 08:35:00 yeah , gcc , avrdude Jan 25 08:35:14 specific recompiles of gcc though Jan 25 08:35:18 and patched avrdude Jan 25 08:35:20 why would you want to run such a bloat on n900? :/ Jan 25 08:35:28 :) Jan 25 08:35:37 java isnt bloat specifically Jan 25 08:36:03 I'd suggest gcc/avrdude and your favourite text/code editor on n900 Jan 25 08:36:08 * buZz install easy-deb-chroot Jan 25 08:40:37 i'll download the debian image at work :) Jan 25 08:41:01 speaking of which, does anyone here use vim a lot, and if so, does it tend to crash xterm in Maemo for you? Jan 25 08:41:30 I always have to use vim really slowly in osso-xterm, otherwise it crashes. Jan 25 08:42:02 it ocassionally crashes when doing things in irssi too, but nowhere near as often as in vim if I try to do things at a natural speed. Jan 25 08:43:00 writing things in insert mode is fine, but command mode usually crashes if I try entering too many things at once. Jan 25 08:43:45 i havent been able to crash osso-xterm with irssi yet Jan 25 08:44:04 it's fairly rare with irssi. Jan 25 08:44:11 or anything else :P Jan 25 08:44:17 but i only own a n900 since 2017 Jan 25 08:44:49 The irssi thing might have to do with tmux, since I'm pretty much always using irssi in tmux in ssh. Jan 25 08:45:16 i just use screen, cause i'm ancient Jan 25 08:45:28 but vim seems to crash consistently through tmux/ssh or just when running it from maemo or a debian chroot on the device. Jan 25 08:45:44 btw, root exploit found in screen in the last few days. Jan 25 08:45:58 still, if vim crashes, it shouldnt pull xterm along Jan 25 08:46:22 It's not vim crashing, just xterm. Jan 25 08:46:31 local root if you use the logfiles in screen, nice one Jan 25 08:46:35 (i dont use that) Jan 25 08:47:22 It's not a matter of whether you use them. Jan 25 08:47:40 it just depends on where the +s bit is. Jan 25 08:48:05 if it's setuid root, then any user can create empty/log files as root. Jan 25 08:48:18 on multiuser systems yeh Jan 25 08:49:12 well, on systems where you already have programs running under supposedly non-trusted users. Jan 25 08:49:46 there's a reason there are usually 30 or so users in /etc/passwd, even on a "single-user" system. Jan 25 08:50:17 yeah, if you give them passwords to log in, you're not so sane ;) Jan 25 08:50:32 or even a shell Jan 25 08:55:10 Maxdamantus: I use vim on n900 Jan 25 08:55:21 but I ditched osso-xterm a long time ago Jan 25 08:55:29 that's actually one of the very first things I did Jan 25 08:56:36 I use (plain old) xterm with some fine-tuned Xdefaults, and a heavily patched xorg keymap Jan 25 08:57:32 (so that I can type pretty much any character I need) Jan 25 08:58:15 some day I'll ditch maemo input method and replace it with ibus, too ... Jan 25 09:07:30 Yeah, I could mostly get by with rxvt-unicode. The main things I was missing when trying to use it as my main terminal were page up/down and something to interact with X selections. Jan 25 09:07:45 so for the latter I just used `xclip` a lot more than usualy. Jan 25 09:07:53 s/y.$/./ Jan 25 09:09:08 Dunno what to do about page up/down .. my layout is already fairly full. Jan 25 09:09:37 Have pretty much just fit every printable ASCII character on it. Jan 25 09:10:04 with one double modifier. Jan 25 09:48:28 Anybody here? Jan 25 11:55:56 Maxdamantus: for page up/down home/end I use http://paste.debian.net/910499/ in /usr/share/X11/xkb/symbols/nokia_vndr/rx-51 file Jan 25 12:13:28 Mm, I already have up/down/left/right under the modifier. Jan 25 12:13:53 Guess I could just make it shift+level2 Jan 25 12:13:58 for pgup/pgdown Jan 25 12:18:43 lol wtf, easy-deb-chroot's image downloader only has debian 5 or 6? :D Jan 25 12:18:46 kinda very ancient Jan 25 12:20:26 at this point I'd suggest to get alpine chroots instead Jan 25 12:22:31 hm, 'alpine linux' the distro? Jan 25 12:23:04 if i search 'alpine' in app manager, it finds some licensing program? Jan 25 12:23:09 is that what you ment? :P Jan 25 12:23:29 ah Jan 25 12:23:30 Alpine - an Alternatively Licensed Program for Internet. Jan 25 12:23:30 News & Email - is a terminal-based tool for reading, sending, and managing electronic messages. Jan 25 12:23:38 ok, that cant be what you ment :P Jan 25 12:26:31 also xbindkeys is nice, by every shift + mod5 + ctrl + any key combination we can start "evkey -d -t -1 /dev/input/event1" and then any commands Jan 25 12:30:07 buZz: I thin jessie is the newest, so it's debian 8 Jan 25 12:30:36 *think Jan 25 12:32:18 from here http://qole.org/files/ Jan 25 12:48:46 ooo cool! i can use that with easy-deb-chroot ? Jan 25 12:50:35 yes Jan 25 12:50:39 thanks :D Jan 25 12:51:43 more gratitude for sulu :) Jan 25 12:52:16 thanks sulu :D Jan 25 12:52:23 he compiled 2.6 kernel & build distro Jan 25 12:54:14 it work with kernel power, don't know with stock kernel Jan 25 12:55:47 i think i run power kernel indeed Jan 25 12:55:53 i guess most CSSU runners do Jan 25 12:58:16 you may right, ker pow here and it works fine Jan 25 14:55:16 * DocScrutinizer51 idly wonders if it's about time for CSSU to roll a real full system update to fremantle PR2.0 Jan 25 14:57:58 of course with keeping compatibility for *all* apps that do not have a tightly mainatined and thoroughly tested update in CSSU Jan 25 15:00:00 even if that might need LD_PRELOAD or chroot for some apps Jan 25 15:02:23 can't be too hard in theory at least to write a generic backward compatibility wrapper that takes care about all apps that are listed in a config file containing all apps of unknown compatibility to PR2.0 Jan 25 15:07:43 PR2.0 could finally sanitize optification (aka nuke it), introduce new kernel and glibc, and possibly even switch to thumb if we could manage to provide genuine thumbified fiasco image etc Jan 25 15:17:03 for now is it possible to unpack/mount combine img, modify it, pack/build the img and flash device with it? Jan 25 15:20:55 which tool can be used to build combine & fiasco image? Jan 25 15:50:50 use fiascogen Jan 25 15:52:26 i thought that that pali's tool can create images now? Jan 25 15:52:43 that xf000 or something like that Jan 25 15:54:51 fiascogen Jan 25 15:55:16 0xffff maybe too Jan 25 15:55:59 -.- see, that's why i usually shut up. Not even close Jan 25 15:56:56 if fiascogen can generate combine too? Jan 25 16:00:25 that's its purpose. 0xffff maybe only can extract, or somesuch limitation Jan 25 16:00:35 0xFFFF my favourite number :) Jan 25 16:01:18 :) Jan 25 16:01:22 also 0xffff is a flasher, not an image builder Jan 25 16:01:30 fiascogen from maemo sdk Jan 25 16:03:48 too much effort for me with sdk, anyway is there any faster method to access to clipboard than python gtk? Jan 25 16:04:58 python is quite slow Jan 25 16:05:18 effort with sdk? Jan 25 16:06:22 I don't have access to PC now Jan 25 16:06:50 so what do you need a fiasco builder for? Jan 25 16:08:39 modify image, build and use it as fast backup Jan 25 16:09:28 >>for now is it possible to unpack/mount combine img, modify it, pack/build the img and flash device with it?<< yes, with a linux PC and fiascogen Jan 25 16:09:39 also see Jan 25 16:09:46 ~bm Jan 25 16:09:46 backupmenu is, like, http://talk.maemo.org/showthread.php?t=63975, or one-click install handling dependency issues: http://maemo.cloud-7.de/maemo5/et_al/HAM-catalogs/BM.install Jan 25 16:11:07 I *guess* you could also run fiascogen on target (N900) Jan 25 16:11:30 BM doesn't Jan 25 16:11:56 is it possible to have rootfs with latest CSSU packages to use it inside chroot/LXC on another device? Jan 25 16:12:03 but there's an explanation how ro build a fiasco from BM backup Jan 25 16:12:47 NeKit: not really Jan 25 16:16:21 so backupmenu is the best for now I think Jan 25 16:20:49 is it possible to run osRescue.img in RAM and reboot that boot osRescue on the same device? :D Jan 25 16:51:24 DocScrutinizer51: whats' the story with the kernel, why is it stuck on 2.6.28 and not at least 2.6.32 f.example ooi ? Jan 25 16:52:16 it's about the closed blobs mainly PVR afaik Jan 25 16:52:32 maybe also API changes Jan 25 16:53:23 ff_: no Jan 25 16:54:19 because of resetting RAM during reboot? Jan 25 16:56:04 I know that's silly, but I'm just curious Jan 25 16:56:57 ff_: wait, what do you want to do? Jan 25 16:58:10 oh, I think I get it Jan 25 16:58:24 well, you can actually do something quite close to it, using kexec Jan 25 16:58:39 on n900, put osRescue.img into RAM, reboot device, but do something once in boot region that osRescue will boot from RAM Jan 25 16:58:42 but I wouldn't assume kexec for arm worked well with kernel 2.6.28 Jan 25 16:59:44 kexec is actually meant for that - put kernel/initrd in some place in ram, run some command instead of the usual reboot, and kernel jumps to new kernel after doing some cleaning Jan 25 17:02:24 actually motorola droid4 (nice hw kbd / slider, omap4 with separate modem) had a locked bootloader but cyanogen guys were able to load custom kernels on it using kexec Jan 25 17:02:41 but I suspect android kernel there was a bit more recent Jan 25 17:03:25 (if it wasn't for the locked bootloader I'd prolly have bought a droid4 already) Jan 25 17:03:52 yeh, something like that, thanks Jan 25 17:04:39 well ... https://talk.maemo.org/showthread.php?t=70083 Jan 25 17:04:45 doesn't look good Jan 25 17:05:54 it might actually be because of the wdog actually Jan 25 17:06:26 let's say that new kernel doesn't have time to reconfigure wdog ... then it will hang Jan 25 17:06:55 heh, I'm currently trying my luck to get Droid 4 Jan 25 17:07:14 I was thinking of buying droid4, but you mentioned about locked bootloader... Jan 25 17:07:27 :) Jan 25 17:07:43 NeKit: well, if you succeed in doing anything useful with it Jan 25 17:08:10 I mean ... apart from kexecing Jan 25 17:08:35 if it doesn't get confiscated by Russian customs first, but what's wrong with kexec? Jan 25 17:09:06 nothing per-see, it just means that you have to 1. trust their bootloader 2. run an android kernel + some userland first Jan 25 17:09:16 then you can jump to Jan 25 17:10:03 to my knowledge nobody reversed the RIL used there, btw (since cyanogen reuses blobs_ Jan 25 17:10:18 so you'd have to reverse it your self if you wanted to use phone Jan 25 17:10:25 it can be reused through Mer ofono plugin Jan 25 17:10:37 on keep android and slap libhybris/ofono on top of it. nemo/sfos-style Jan 25 17:10:42 s/on/or/ Jan 25 17:10:42 bencoh meant: or keep android and slap libhybris/ofono on top of it. nemo/sfos-style Jan 25 17:11:44 but that means moar blobs Jan 25 17:12:10 it's possible though, that's how sfos works on all those androiphones Jan 25 17:12:42 the good thing that it's relatively hardware-agnostic then Jan 25 17:14:29 this might be a way to get more people to work on the hildon/maemo-like port Jan 25 17:15:29 apart from the fact that maemo-like usually means Xorg, and that wont really work on top of android/hybris afaict Jan 25 17:17:02 xwayland on wayland on android, maybe :> Jan 25 17:17:18 I could start hildon on Redmi Note 2 with libhybris graphics, but it was hacked to render directly to SurfaceFlinger instead of returning picture to Xorg Jan 25 17:17:45 NeKit: hildon-gtk2 or hildon-gtk3? Jan 25 17:18:01 hildon-gtk2, but it shouldn't matter Jan 25 17:18:12 NeKit: maemo apps still expect gtk2, which means xorg Jan 25 17:18:17 unless you hacked gtk2 as well? Jan 25 17:18:38 Xorg is running, just not displaying anything Jan 25 17:18:57 sure but how does gtk2 render? Jan 25 17:20:32 it renders to pixmap as usual, then Hildon loads it to texture and uses OpenGL to render everything Jan 25 17:21:39 well I guess I'm missing something / should dive into the gtk/hildon relationship Jan 25 19:47:57 NeKit: and how's that supposed to work in non-compositing mode? Jan 25 19:48:26 it won't :) Jan 25 19:48:48 :) Jan 25 19:49:41 yeah, my question was rhetoric. Jan 25 19:51:12 I looked at Raspberry Pi OpenGL + Xorg example (their driver doesn't have Xorg support). It's possible to do rendering to offscreen buffer, then copy it to pixmap and to Xorg Jan 25 19:51:41 not sure how much doing multiply copies would hurt perfomance, but it's better solution probably Jan 25 19:51:54 NeKit: it seems you know the stuff, why don;t you join the porting party? lets have it first running on upstream linux, then we'll find a way to run it on wayland/libhybris Jan 25 19:52:26 NeKit: we can use clutter for that I guess Jan 25 19:52:33 (offscreen rendering) Jan 25 19:53:16 but, I guess the performance will be terrible, pipeline flush on every frame Jan 25 19:57:38 well, I'm not sure how I can be helpful to the porting party right now, since I don't have GTK+/Maemo experience Jan 25 19:57:56 what I wanted to try is integrating libhybris builds to make it more or less run on recent devices Jan 25 20:02:10 NeKit: well, porting hildon is just a part of the job that needs to be done, there is stuff like moving to modern init system (be it upstart or systemd), packaging, etc. The point is - if there is something usable, more people will be attracted. At least this is what I hope for. Jan 25 20:26:40 to start apps without source code, like Maemo system apps, would chroot be needed? Jan 25 20:38:04 in a classic Xorg environment, I'd expect both libs to coexist nicely in the same rootfs Jan 25 20:38:19 since different ABI means different lib version number Jan 25 20:39:51 we might have issues with share (/usr/share and the likes) files though Jan 25 20:40:29 if they become incompatible for any reason we'd have to rebuild old libs with a different share prefix Jan 25 20:40:47 (or new ones for that matters) Jan 25 21:44:33 or use mount namespaces Jan 25 21:45:00 per process or per user Jan 25 21:45:28 pretty simple Jan 25 21:45:46 a 5liner wrapper script Jan 25 21:51:47 I'm quite sceptic that stuff like 'modern init system' is anything needed to improve attractiviry of maemo. Rather support for existing apps incl their old init system is a key issue. Seen too many OS fail just because no initial stock of working apps been available so they died from no chicken no egg Jan 25 21:53:47 linux is attractive because of basic compatibility to 1990 apps Jan 25 21:55:56 stuff like wayland and systemd and their proponents are denying that and think new ==leete==better Jan 25 21:56:02 DocScrutinizer51: the problem is that maemo init system is a mixture of upstart and sysv and that doesn't fit in anything upstream Jan 25 21:56:17 so? Jan 25 21:56:54 we already discussed that - the less forks to support, the more time will be used for more sane tasks Jan 25 21:57:34 my init system on my PC also doesn't fit into upstream. Maemo is maemo's upstream *particularly* for sich distro specific stuff like init system Jan 25 21:57:49 give me 5 fulltime job devs and I will make it like that Jan 25 21:58:35 but until then, it is better to stick to the wide-spread distros Jan 25 21:59:16 oh, and the waaay more demanding task to MAKE A WHOLE NEW (to maemo) INIT system is feasible part time? Jan 25 21:59:53 it is not whole new, scripts just need to be ported and this is trivial Jan 25 22:00:14 maemo is using upstart already Jan 25 22:00:34 AND sys4init Jan 25 22:00:40 *sysvinit Jan 25 22:01:36 so? Jan 25 22:01:41 but, those are used in such a hacky way... - or example, lots of things are started by both upstart and sysvinit and they check if they were already started Jan 25 22:02:34 so, the task of moving from /etc/event.d to /etc/init is not a whole new init system Jan 25 22:02:36 when porting init scripts is trivial then why do you need a new init system instead porting NEW apps to the plain sysv init that maemo is to all apps? Jan 25 22:02:50 no, maemo is not plain sysv Jan 25 22:03:12 it evidently is, from a app perspective Jan 25 22:03:17 maemo is more upstart than sysv Jan 25 22:03:53 just take a look in /etc/event.d Jan 25 22:04:12 even my fucking systemd PC still has 80 percent plain sysv init scripts Jan 25 22:04:38 toldya, check what is in /etc/event.d Jan 25 22:04:46 I guess I had more looks to /etc than you, on maemo Jan 25 22:05:00 what is your point then? Jan 25 22:05:09 I *know* what's in there Jan 25 22:05:32 cool, then don;t say maemo init is sysv Jan 25 22:05:38 because it is not Jan 25 22:06:25 and upstart as well as systemd have a sysv compatibility layer, and I don't see a single reason why maemo NEEDS a new init system Jan 25 22:07:06 I didn;t say maemo needs new init system, what I said is that maemo startup scripts have to be ported to some of the exeisting upstream init systems Jan 25 22:07:14 dafaq, read me!!! I *NEVER* said 'maemo is sysv init" Jan 25 22:07:54 "...to the plain sysv init that maemo is..." Jan 25 22:08:15 unless I misunderstand what you said Jan 25 22:08:25 partial quating == abyssmally lame Jan 25 22:08:46 comeon, it is 10 lines ^^^. anyway. Jan 25 22:08:58 "... I misunderstand..." Jan 25 22:09:13 yes you do Jan 25 22:09:17 could be Jan 25 22:09:18 always Jan 25 22:09:50 you're quoting me partial to change meaning of my words Jan 25 22:12:31 ls /etc/init.d/* Jan 25 22:18:53 this works on maemo, old age systems, even ubuntu and opensuse Jan 25 22:19:34 so why do you think maemo acripts need porting? Jan 25 22:23:28 or maemo needs a "new init system' except for obscure upstream compatibility for NEW apps that are ported from upstream Jan 25 22:24:14 or WHY maemo needs... Jan 25 22:28:49 maemo sysvinit scripts are missing LSB headers, upstart scripts need startup dependencies changed Jan 25 22:39:46 couple of things in this chan have left me surprised .. DocScrutinizer51 and Pali have systemd on their machines? :P Jan 25 22:43:03 Well it's a bit difficult not to, hence the arguments over init freedom. You could also have systemd on the machine but not use it for init. Jan 25 22:46:34 understandable Jan 25 22:50:17 Devuan is still in beta too Jan 25 22:53:38 Sicelo: (systemd) just because I hate switching from suse after almost 20 years even more than systemd crap Jan 25 22:55:04 is it your init system? or you setup something else? Jan 25 22:55:39 i have systemd too since 2 or 3 months ago, when i decided to used Debian Sid Jan 25 23:05:28 Sicelo: it's init system atm, and I *hate* it Jan 25 23:06:03 at least nobody can blame me for hating without knowing Jan 25 23:07:00 Sicelo: move to devuan?! Jan 25 23:07:55 ~devuan Jan 25 23:07:57 well, devuan is the most awesome distro, or https://devuan.org Jan 25 23:12:53 sixwheeledbeast: it's still beta but reported rock solid for servers at least. Desktop stuff is too systemd infested and alternatives just emerging. xfce seems to mostly work Jan 25 23:16:59 ~*uptime Jan 25 23:17:19 ~+uptime Jan 25 23:17:20 - Uptime for infobot - Jan 25 23:17:20 Now: 2d 5h 9m 9s running infobot 1.5.4 (SVN) -- linux Jan 25 23:17:21 1: 59d 8h 41m 19s running infobot 1.5.4 (SVN) -- linux, ended Sun Nov 14 18:39:57 2010 Jan 25 23:17:21 2: 57d 3h 9m 23s running infobot 1.5.4 (SVN) -- linux, ended Fri Jun 26 20:39:27 2009 Jan 25 23:17:21 3: 36d 20h 47m 14s running infobot 1.5.4 (SVN) -- linux, ended Tue Aug 4 17:38:59 2009 Jan 25 23:46:42 hmm, looks like infobot lost connection instad Jan 25 23:46:53 it was down, right? Jan 25 23:47:47 interesting distro Jan 25 23:48:13 ~flatbattrecover Jan 25 23:50:48 Which distro would you guys recommend from the BSD family? Jan 25 23:59:53 ibot reboots every now and then, but not since 2d 5h Jan 26 00:01:15 it loses connection on every single netasplit **** ENDING LOGGING AT Thu Jan 26 03:00:01 2017