Sicelo | btw, lol, as nice as ke-recv is, now i have too much stuff being indexed from my ext3 SD card with Nemo. i could edit tracker.cfg, but that would affect other μSD | 00:33 |
---|---|---|
Sicelo | i'll be umounting for now | 00:33 |
*** _rd has quit IRC | 00:40 | |
*** NIN101 has quit IRC | 01:12 | |
kerio | DocScrutinizer05: have you found a way to fix qt? | 01:55 |
DocScrutinizer05 | mickname: I think I recently read somewhere in a changelog or commit comment: "stop unnecessary rotation when switching tasks bla bla bla" | 01:58 |
DocScrutinizer05 | mickname: *maybe* it's a bugfix in most recent microB which freemangordon and others are working on | 01:59 |
DocScrutinizer05 | or was it fireforx? or the rendering engine? | 01:59 |
DocScrutinizer05 | Sicelo: nice point, indeed we should think about how to solve that. Are you aware ke-recv mounting script has a hook to selectively avoid mounting of particular partitions depending on whatever you can probe via cmdline? | 02:02 |
DocScrutinizer05 | probably the proper fix would be to augment trackerd.cfg semantics to exclude not only by path globs but maybe also by calling arbitrary extrernal binaries/scripts with parameter $pathname and not indexing that $pathname when that calle dbinary/script returns non-zero | 02:04 |
kerio | DocScrutinizer05: how do we differentiate a partition with Nemo on vs a data partition with music on? | 02:04 |
DocScrutinizer05 | e.g by uSD/parrtition UUID= | 02:05 |
DocScrutinizer05 | ? | 02:05 |
kerio | no, i meant a new uSD | 02:06 |
DocScrutinizer05 | eh? | 02:06 |
DocScrutinizer05 | not at all | 02:06 |
kerio | what should we default to then? | 02:06 |
DocScrutinizer05 | or you might want to chack if a nemo-specific file is on the partition | 02:06 |
DocScrutinizer05 | check, wtf?? | 02:06 |
DocScrutinizer05 | my typing sucks even more than usually | 02:07 |
kerio | i'm not talking about just nemo | 02:07 |
kerio | si | 02:07 |
kerio | whoops, wrong window | 02:07 |
kerio | DocScrutinizer05: there's no good rationale, apart from asking the user, for indexing or not indexing a partition | 02:08 |
DocScrutinizer05 | well, nothing a script couldn't do, don't you think so? | 02:08 |
DocScrutinizer05 | [2012-09-09 01:04:18] <DocScrutinizer05> probably the proper fix would be to augment trackerd.cfg semantics to exclude not only by path globs but maybe also by calling arbitrary extrernal binaries/scripts with parameter $pathname and not indexing that $pathname when that calle dbinary/script returns non-zero | 02:09 |
nox- | sounds slow | 02:17 |
DocScrutinizer05 | not as slow as trackerd ;-P | 02:22 |
nox- | :) | 02:22 |
DocScrutinizer05 | kidding aside, it's only called once per trackerd index run | 02:22 |
nox- | ah just for the top dir it wants to index? | 02:23 |
DocScrutinizer05 | for whatever shell pattern matches the condition | 02:23 |
nox- | ok but wont the script be called for every subdir it wants to index then at least? | 02:26 |
DocScrutinizer05 | if we'd call it for every file/dir that'd be obviously nonsense, so we only call it for special patterns in exclude statement of trackerd.cfg. This exclude statement has a list of... patterns, which either get excluded unconditionally (that's what we already have) or will call a binary when a special syntax for the pattern is used | 02:26 |
nox- | ah ok | 02:26 |
DocScrutinizer05 | yeah, I'm a slow coder but a fast and usually exact architect | 02:28 |
nox- | :) | 02:28 |
* nox- just misunderstood... | 02:28 | |
*** LaoLang_cool has joined #maemo-ssu | 03:23 | |
*** M4rtinK has quit IRC | 03:40 | |
*** arcean has quit IRC | 04:10 | |
*** nox- has quit IRC | 04:57 | |
*** nox- has joined #maemo-ssu | 04:58 | |
*** ekze has quit IRC | 05:20 | |
*** ekze has joined #maemo-ssu | 05:21 | |
*** amiconn has quit IRC | 05:43 | |
*** amiconn_ has joined #maemo-ssu | 05:43 | |
*** amiconn_ is now known as amiconn | 05:43 | |
*** nox- has quit IRC | 05:55 | |
*** freemangordon has quit IRC | 06:14 | |
*** DocScrutinizer05 has quit IRC | 06:33 | |
*** DocScrutinizer05 has joined #maemo-ssu | 06:33 | |
*** illuminata has left #maemo-ssu | 08:49 | |
*** _rd has joined #maemo-ssu | 09:19 | |
*** krayon has joined #maemo-ssu | 09:31 | |
*** _rd has quit IRC | 09:40 | |
*** LaoLang_cool has quit IRC | 09:52 | |
*** _rd has joined #maemo-ssu | 10:58 | |
*** _rd has quit IRC | 11:04 | |
*** NIN101 has joined #maemo-ssu | 11:21 | |
*** LaoLang_cool has joined #maemo-ssu | 11:38 | |
*** timo^ has joined #maemo-ssu | 11:42 | |
*** _rd has joined #maemo-ssu | 11:46 | |
*** M4rtinK has joined #maemo-ssu | 11:54 | |
*** LaoLang_cool has joined #maemo-ssu | 12:01 | |
*** M4rtinK has quit IRC | 12:11 | |
*** M4rtinK has joined #maemo-ssu | 12:15 | |
*** arcean has joined #maemo-ssu | 12:46 | |
*** LaoLang_cool has quit IRC | 13:09 | |
*** LaoLang_cool has joined #maemo-ssu | 13:10 | |
kerio | do you guys think we could ever add proper /etc/shadow support? | 13:48 |
*** LaoLang_cool has quit IRC | 14:31 | |
*** timo^ has quit IRC | 14:42 | |
*** _rd has quit IRC | 14:43 | |
*** toxaris has joined #maemo-ssu | 14:51 | |
*** _rd has joined #maemo-ssu | 15:56 | |
*** _rd has quit IRC | 16:02 | |
*** timo^ has joined #maemo-ssu | 16:32 | |
DocScrutinizer05 | kerio: please first check if support maybe already is in place. after all it's just about format of /etc/passwd entries and a few system calls that set or compare passwords | 18:01 |
kerio | it's not a syscall | 18:03 |
kerio | how i know there's no /etc/shadow support: | 18:04 |
kerio | /etc/shadow: No such file or directory | 18:04 |
DocScrutinizer05 | meh | 18:10 |
DocScrutinizer05 | touch /etc/shadow | 18:10 |
DocScrutinizer05 | and yes, it seems password support is a tad scarce, anyway there's (3)getpwent et al | 18:11 |
DocScrutinizer05 | >> The getpwent() function returns a pointer to a structure containing the broken-out fields of a record from the password database (e.g., the local password file /etc/passwd, NIS, and LDAP). | 18:11 |
DocScrutinizer05 | I'd hope this includes support for /etc/shadow as well | 18:12 |
DocScrutinizer05 | probably a compile time option | 18:13 |
DocScrutinizer05 | ...whether those calls support shadow, LDAP, NIS/YP | 18:13 |
DocScrutinizer05 | (sorry for my noobish comments, I never looked into all that until now) | 18:14 |
kerio | ooh, /usr/bin/passwd is suid | 18:23 |
kerio | and not busybox | 18:23 |
kerio | for obvious reasons | 18:23 |
*** toxaris has quit IRC | 18:23 | |
kerio | ooooooh | 18:24 |
kerio | "/etc/shadow" is a string in /usr/bin/passwd | 18:24 |
kerio | i don't know how to test this though | 18:25 |
*** toxaris has joined #maemo-ssu | 18:25 | |
DocScrutinizer05 | just create a proper entry in /etc/passwd and corresponding entry in /etc/shadow | 18:28 |
DocScrutinizer05 | [2012-09-09 17:25:00] <kerio> "/etc/shadow" is a string in /usr/bin/passwd ### that's what I meant | 18:29 |
DocScrutinizer05 | in sudo you might find getpwent used and thus no reference to /etc/shadow | 18:30 |
*** toxaris has quit IRC | 18:30 | |
kerio | DocScrutinizer05: how the hell do i check if i can login as the new user | 18:31 |
DocScrutinizer05 | anyway you need to check (1)passwd how it determines whether to use shadow or not. Maybe it's already sufficient to touch /etc/shadow | 18:31 |
DocScrutinizer05 | su - newuser ? | 18:31 |
kerio | DocScrutinizer05: oh you silly | 18:31 |
kerio | do i have to remind you that su is actually busybox | 18:31 |
DocScrutinizer05 | MEEE? ;-P | 18:31 |
kerio | which is not suid root? | 18:31 |
kerio | and thus doesn't actually do *anything* | 18:32 |
*** toxaris has joined #maemo-ssu | 18:32 | |
kerio | hmm | 18:32 |
kerio | ok, passwd actually seems to work | 18:32 |
kerio | but there's no cool hashing support in crypt(3) | 18:33 |
kerio | so we're still stuck with 8-character passwords | 18:33 |
DocScrutinizer05 | >> | 18:34 |
DocScrutinizer05 | Glibc Notes | 18:34 |
DocScrutinizer05 | The glibc2 version of this function supports additional encryption algorithms. | 18:34 |
kerio | DocScrutinizer05: well your man page is full of lies | 18:34 |
DocScrutinizer05 | lol, your system is full of lies | 18:34 |
DocScrutinizer05 | btw use friggin sudo if you got no proper su on your system ;-D | 18:36 |
kerio | /usr/bin/ld: Warning: /usr/lib/gcc/arm-linux-gnueabi/4.2.1/libgcc_s.so: Unknown EABI object attribute 34 | 18:37 |
kerio | ah balls | 18:37 |
kerio | wtf is this shit | 18:37 |
kerio | libc6-dev is already the newest version. | 18:37 |
DocScrutinizer05 | maybe ld isn't, though | 18:39 |
DocScrutinizer05 | which would pretty much explain it | 18:39 |
kerio | well what's the ld i'm supposed to use? | 18:39 |
DocScrutinizer05 | :shrug: | 18:39 |
kerio | (this is on the n900, mind you) | 18:39 |
kerio | meh, fuck this, i'm removing the dev stuff | 18:40 |
kerio | huh, why do i have gcc-4.2-base installed? | 18:42 |
kerio | installed manually, i mean | 18:43 |
DocScrutinizer05 | btw (maybe overly) simple test if sudo reads from /etc/shadow: mkfifo /etc/shadow | 18:46 |
DocScrutinizer05 | cat /etc/shadow will block on read | 18:46 |
DocScrutinizer05 | so I hope would sudo | 18:46 |
kerio | DocScrutinizer05: even better, echo -n >/etc/shadow | 18:47 |
DocScrutinizer05 | so what? | 18:47 |
DocScrutinizer05 | how will you know if sudo reads shadow file then? | 18:47 |
kerio | so sudo reads an empty file, and doesn't actually get stuck | 18:47 |
kerio | because the write succeeds :) | 18:47 |
kerio | i meant with /etc/shadow as a fifo | 18:48 |
DocScrutinizer05 | sudo does writes to shadow? | 18:48 |
kerio | no, sudo reads from shadow | 18:48 |
kerio | if you read from a pipe, you get stuck | 18:48 |
kerio | unless there's something writing on the other side | 18:48 |
DocScrutinizer05 | that's the idea | 18:48 |
DocScrutinizer05 | but of course you can use strace sudo to find out whether it reads from shadow file or not | 18:52 |
kerio | DocScrutinizer05: which debian is fremantle supposed to be similar to? | 19:25 |
DocScrutinizer05 | err | 19:25 |
kerio | was it etch? | 19:26 |
NIN101 | there not too many similiarities besides the package format and a few other things... | 19:30 |
kerio | hm, i should really repartition this :s | 19:31 |
kerio | meh, let's do this | 19:31 |
kerio | leeroooooooooooooooy jeeeeeeeeeeenkinsss | 19:31 |
*** _rd has joined #maemo-ssu | 20:01 | |
kerio | cdebootstrap_0.5.9_armel.tar.gz is pretty neat :D | 20:06 |
*** arcean has quit IRC | 20:10 | |
Sicelo | what is it? | 20:15 |
* Sicelo checks to see what difference exists between that one and debootstrap | 20:17 | |
*** arcean has joined #maemo-ssu | 20:20 | |
kerio | Sicelo: cdebootstrap is a binary | 20:22 |
kerio | and it has a distro-independent, statically-compiled tarball | 20:22 |
kerio | that should work pretty much everywhere | 20:22 |
kerio | DocScrutinizer05: i have no idea how to chroot \_o_/ | 20:39 |
kerio | i mean, what should be exposed inside of the jail? | 20:39 |
kerio | man, this stuff is hard | 20:41 |
kerio | for instance, what to do with /dev/pts? | 20:46 |
kerio | i already have a fake /etc/mtab | 20:46 |
kerio | should i try to make newinstance for devpts work? | 20:46 |
kerio | i don't even know if it's in 2.6.28 or if it was added later | 20:46 |
kerio | man, no wonder the easydebian guys made such a mess | 20:48 |
*** timo^ has quit IRC | 21:12 | |
Sicelo | :P | 21:15 |
Sicelo | what do you want out of it kerio? | 21:15 |
kerio | Sicelo: i don't know! D: | 21:16 |
kerio | i just wanted a working gcc | 21:17 |
DocScrutinizer05 | that's been the point where I stopped posting wise statements last time we discussed the topic | 21:17 |
DocScrutinizer05 | ;-P | 21:17 |
kerio | ._. | 21:18 |
DocScrutinizer05 | kerio: for a gcc you probably need not much more than a working shell | 21:20 |
DocScrutinizer05 | IOW you don't care about much of the original infra getting "re-used" in your chroot | 21:21 |
kerio | eh, i suppose you're right | 21:21 |
kerio | i'm not even entirely sure chrooting like this is a sensible thing to do, considering that debian uses different libraries anyway | 21:21 |
DocScrutinizer05 | :nod: | 21:21 |
DocScrutinizer05 | given the rationale that gcc shouldn't run a s root generally, you'll most likely get away with symlinks for a few of the fat library pathes from / to /opt | 21:22 |
*** toxaris has quit IRC | 21:24 | |
DocScrutinizer05 | I E much like you would do on any arbitrary other linux system: chreate new user, place user's home in a convenient place, get some symlinks form /usr/src to /opt/usr/src etc | 21:24 |
DocScrutinizer05 | su - developeruser | 21:25 |
DocScrutinizer05 | profit | 21:25 |
*** toxaris has joined #maemo-ssu | 21:26 | |
*** toxaris has quit IRC | 21:28 | |
*** toxaris has joined #maemo-ssu | 21:30 | |
*** _rd has quit IRC | 21:35 | |
*** toxaris has quit IRC | 21:35 | |
*** toxaris has joined #maemo-ssu | 21:37 | |
*** toxaris has quit IRC | 21:40 | |
* kerio ponders about optifying /usr/lib/gcc/arm-linux-gnueabi | 21:41 | |
DocScrutinizer05 | sounds as sane as it gets | 21:42 |
*** toxaris has joined #maemo-ssu | 21:42 | |
DocScrutinizer05 | nah | 21:42 |
DocScrutinizer05 | optify /usr/lib/gcc ! | 21:43 |
kerio | k | 21:43 |
*** toxaris has quit IRC | 21:43 | |
DocScrutinizer05 | # (*) only optify the lowest base folder possible | 21:44 |
DocScrutinizer05 | # ie /foo/bar | 21:44 |
DocScrutinizer05 | # instead of /foo/bar/one | 21:44 |
DocScrutinizer05 | (from /etc/maemo-optify-auto-opt.conf) | 21:44 |
*** toxaris has joined #maemo-ssu | 21:46 | |
DocScrutinizer05 | merlin1991: aiui you're involved in migration task from Nokia/Akamai servers to community infra? | 21:46 |
kerio | how do i manually trigger the boot/postinst optification? | 21:46 |
DocScrutinizer05 | merlin1991: I just wonder *how much* of the Nokia closed blobs repos (and src?) Nokia will allow to move? | 21:46 |
DocScrutinizer05 | maemo-optify-auto-opt.sh maemo-optify-firstboot.sh maemo-optify-lib.sh maemo-optify-make-mountbind.sh | 21:47 |
DocScrutinizer05 | pick sth | 21:47 |
kerio | (by the way, isn't -firstboot horrendously dangerous?) | 21:48 |
*** toxaris has quit IRC | 21:48 | |
DocScrutinizer05 | (it's the thing that actually does the whole magic of *creating* /opt ;-D) | 21:48 |
kerio | (i know, but if you launch it again it fucks up your /opt) | 21:48 |
DocScrutinizer05 | yeah, probably XP | 21:48 |
kerio | meh, i'm rebooting | 21:49 |
DocScrutinizer05 | # minimal script to cleanup /home/opt under very specific circumstances. | 21:49 |
DocScrutinizer05 | # will copy contents of rootfs /opt to /home-opt to allow image to be optified. | 21:49 |
DocScrutinizer05 | # will only run if do-not-clean-home-opt does not exist. | 21:49 |
DocScrutinizer05 | # postinst will create this file if used within scratchbox. | 21:50 |
*** timo^ has joined #maemo-ssu | 21:50 | |
*** toxaris has joined #maemo-ssu | 21:50 | |
DocScrutinizer05 | # IMPORTANT! this is called from upstart scripts, so make sure there are no | 21:51 |
DocScrutinizer05 | # error codes returned | 21:51 |
kerio | oh lol | 21:51 |
DocScrutinizer05 | seems it's called each boot | 21:51 |
DocScrutinizer05 | and knows what to (not) do | 21:51 |
DocScrutinizer05 | #only clean the /home/opt if needed | 21:52 |
DocScrutinizer05 | [ -e $optifyflagfile ] && return 0 | 21:52 |
kerio | well this is dumb | 21:52 |
kerio | it doesn't want to work | 21:52 |
DocScrutinizer05 | # if /home is not mounted r/w for some reason -- bail out | 21:53 |
DocScrutinizer05 | # if the /opt->/home/opt mountbind exists -- bail out | 21:53 |
kerio | i see | 21:53 |
DocScrutinizer05 | # we have come back into this function AFTER the binding has occured | 21:54 |
*** toxaris has quit IRC | 21:54 | |
DocScrutinizer05 | log "opt-boottime :: mount /home/opt->/opt exists, already optified, quitting." | 21:54 |
DocScrutinizer05 | # move /opt contents to /home/opt only if they exist | 21:54 |
DocScrutinizer05 | # ensure we update the donotclean flag now so we do not run again. | 21:54 |
DocScrutinizer05 | . | 21:55 |
DocScrutinizer05 | # use a variable to store do-not-clean file location | 21:55 |
DocScrutinizer05 | optifyflagfile="/var/lib/maemo-optify-firstboot-do-not-clean-home-opt" | 21:55 |
*** toxaris has joined #maemo-ssu | 21:55 | |
DocScrutinizer05 | ~ # ls -l /var/lib/maemo-optify-firstboot-do-not-clean-home-opt | 21:56 |
DocScrutinizer05 | -rw-r--r-- 1 root root 0 Mar 6 2011 /var/lib/maemo-optify-firstboot-do-not-clean-home-opt | 21:56 |
kerio | I SEE | 21:56 |
DocScrutinizer05 | ;-P | 21:56 |
kerio | anyway, i did the optification manually | 21:57 |
kerio | i have no idea how maemo-optify-auto-opt.sh is supposed to work | 21:57 |
DocScrutinizer05 | shouldn't be too hard, given there's stuff like cp -l | 21:57 |
kerio | ew | 21:58 |
kerio | i just mv'd | 21:58 |
kerio | well, cp-a'd and then rm'd | 21:58 |
kerio | and then ln-s'd | 21:58 |
kerio | verbing is my new favourite thing | 21:59 |
*** toxaris has quit IRC | 21:59 | |
DocScrutinizer05 | kerio: | 22:00 |
DocScrutinizer05 | # get_optified_path: check if entry or any of its parent dirs was already optified | 22:00 |
DocScrutinizer05 | # conditions: the entry is a symlink and points to equivalent path under /opt | 22:00 |
kerio | it's /usr/lib/gcc | 22:01 |
DocScrutinizer05 | I gather the supposed operation mode is: echo /usr/lib/gcc >>/etc/maemo-optify-auto-opt.conf | 22:01 |
*** toxaris has joined #maemo-ssu | 22:01 | |
kerio | i did | 22:01 |
DocScrutinizer05 | then on next boot it will get optified aiui | 22:02 |
DocScrutinizer05 | or run maemo-optify-auto-opt.sh manually | 22:02 |
kerio | i did both, but /usr/lib/gcc wasn't optified | 22:03 |
DocScrutinizer05 | should do the same | 22:03 |
kerio | i did that manually, no biggie | 22:03 |
*** toxaris has quit IRC | 22:03 | |
DocScrutinizer05 | still puzzling, I'll eventually look into it | 22:03 |
*** toxaris has joined #maemo-ssu | 22:05 | |
*** _rd has joined #maemo-ssu | 22:09 | |
*** toxaris has quit IRC | 22:16 | |
*** toxaris has joined #maemo-ssu | 22:18 | |
*** Sc0rpius has joined #maemo-ssu | 22:21 | |
*** toxaris has quit IRC | 22:30 | |
*** toxaris has joined #maemo-ssu | 22:33 | |
*** toxaris has quit IRC | 22:34 | |
*** toxaris has joined #maemo-ssu | 22:35 | |
*** nox- has joined #maemo-ssu | 22:38 | |
*** BCMM has joined #maemo-ssu | 22:47 | |
*** DocScrutinizer51 is now known as DocPyrotizer | 22:58 | |
amiconn | fireforks? o.O | 23:22 |
*** NIN101 has quit IRC | 23:23 | |
*** arcean has quit IRC | 23:25 | |
*** arcean has joined #maemo-ssu | 23:41 | |
*** toxaris has quit IRC | 23:41 |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!