IRC Logs for #crux-devel Thursday, 2012-01-05

mavrick61Hi Romster: In general, knowledge and understanding of how computers work is low. That the hard disk is part of a computer are not so many users which understand, and that they at any time can cease to function. I work mostly with enterprise systems, but also there must motivate company management so that they invest in backup systems. But When disater strikes it's good to Have Linux to fix the problem.00:43
juejaeger: yeah, the config worked flawless02:04
juejaeger: will fix the libtirpc issue later02:04
juejaeger: installation was without problems, I've build some progs and, most important, didn't see any sign that I'm using a multilib toolchain ;)02:10
Romstermavrick61, testdisk ddrescue photorec got some other tools too all handy but nothing beats a backup, i would expect businesses to have more sense than a home consumer though with backups, i inform customers on every purchase and repair job. about all i can do.02:17
Romsterjue, are oyu trying out multilib and got it in such away that your not using any 32bit ports?02:18
jueRomster: exactly02:44
Romsterand damn it doing a complete recompile of all ports has broken a few things02:45
juewhy do you recompile everything?02:45
Romsterbecause i added -flto02:45
Romsterlibz sudo xorg-libxxf86vm so far since i found out my xine-ui isn't loading atm. looking into what went wrong.02:46
jueyou cannot withstand to do things other than the default?02:46
Romsteri sort of expected something to break but i also expected stuff that wont recompile wont be affected. as all the symbols should all remain the same.02:47
Romsterprobably not but what's wrong with using a few new features.02:48
juewell, you are a official maintainer now, so using the default toolchain should be a matter of course02:49
juethat's at least my opinion ;)02:49
Romsterif ports build in a default chroot and on my system is that good enough?02:50
Romstereveryong in opt/contrib is built in a chroot.02:50
Romsterbut i also build it on my system to install it too.02:51
juewell, it's your decision, you know my opinion now02:51
Romsteri do.02:51
Romsterit's not a major change though, only better optimizing at -O2 over multiple object files.02:52
Romsterbtw new iptables is out02:53
jueI know what LTO is for, but, honestly, is it worth the effort?02:53
Romsterthat's what i'm finding out <_<02:54
Romsterwont know if oyu don't test it02:54
Romsteri'm not using unsafe compiler flags02:55
Romsterdid that in my younger years and learned the hard way.02:55
juethat was a rhetorical question ;)02:55
juehmm, iptables is latest or do you see another release?02:56
Romsteri'vebeen putting off the gold linker due to possible breakage.02:56
Romsterno that's it on freecode02:57
juelooks like your ports-tree is a bit outdated02:58
Romsteroh thats been done already i didn't see the commit off cruxbot in #crux02:58
Romsternah i haven't ports -u yet. will do that after i fix up this breakage issue02:58
Romstercruxbot must off been offline when that iptables got committed.02:59
Romsterthat bot hasn't been reliable lately03:01
Romsterwhat the hell it's advertised on 4th jan 3am on freecode03:06
Romsterbeen done a few days ago i didn't look that far back03:06
Romster4th 3am*03:06
Romstergrr i mean 5th 3am not the 4th03:07
*** sepen has joined #crux-devel03:07
Romsterlate announcement on freecode03:07
Romsterok simple linking issues forced rebuild fixes those.03:10
Romsteri didn't expect LTO to require other libs to be force rebuilt.03:11
Romsterlive and learn.03:11
Romsterjue your lucky you don't have any demented brain issues that i have they drive me crazy sometimes.03:26
*** mavrick61 has quit IRC03:49
*** mavrick61 has joined #crux-devel03:50
*** mavrick61 has quit IRC03:57
*** mavrick61 has joined #crux-devel03:59
*** mavrick61 has left #crux-devel04:05
*** mavrick61 has joined #crux-devel04:19
Romsterhmm that should be bumped asap 6 flaws O_O04:52
teK_wtf is DTLS04:53
Romsteri dunno04:57
frinnstyeah, but nothing really "dangerous", atleast for me04:57
frinnsti think :)04:57
Romsteri probably fall in the same boat but i just bumped it on my firewall05:00
teK_the double-free is for 0.9 releaese only05:02
teK_ssl 3.0.. who uses that? :>05:02
jueno panic ;)05:02
teK_5 days in January and cve lready lists 27 vulns (openssl CVE-2012-0027)05:03
teK_oh yeah05:03
teK_btw.. I gave the java rhino exploit a try with metasploit.. this stuff works shockinlgy well and easy05:03
Romsterhmm i haven't came across that one.05:08
Romsterah that's probably why i've been getting so many virus infected computers at work from drive by sites.05:12
Romsteri always update java and flash and the browser out of habit on all machines.05:12
frinnstpedja: you're more than welcome to maintain virtualbox. but please update the arch maintainer line if you want to take over a port05:44
teK_Romster: you better do so.. that was: visit innocent looking url, java is run and bam.. remote code execution i.e. meterpreter shell with: recording of your mic, taking screenshots, disabling mouse/keyboard, dumping password hashes, yada yada05:47
Romsterinstaling rootkits to the MBR05:53
Romsterlog all key strkes, spam everyone in your address book, change DNS servers to redirect though some dodgy site05:54
Romsteroh and spyware that disables taskmaanager regedit and hides all your icons05:54
Romsteri've seen it all.05:55
Romstereven submitted some virus samples05:56
teK_tbh.. the interesting thing is to get in.. the rest is just boring :)05:57
Romsteronly a few i found xdcc file shares seems the majority want to steal details and prevent you from using your computer.05:58
pedjafrinnst: I somehow forgot to ask you first before I updated the port, sorry.I'll take it, if you want, I use Virtualbox fairly regullarly.06:19
Romsteroh btw teK_ opt-x86_64.git still has issues error: unable to create temporary sha1 filename ./objects/b8: File exists06:21
pedjaxfce4-appfinder update error.06:22
sepenyep I know, new xfce-utils will remove xfce4-run's stuff06:53
sepenyesterday I comment that on #xfce06:54
sepenpedja: I'll be back later from home06:56
sepensee ya'06:56
*** sepen has quit IRC06:56
jaegerjue: glad to hear your test goes well, thanks for the feedback :)08:28
jaegerjue: the kernel .config on that multilib ISO is also quite different from the default CRUX config, I started from scratch to try to fix some boot issues (mostly in virtualbox)08:32
jaegerI haven't re-added all the network drivers that the default config has back in the mix yet but so far it seems to have solved some of the issues08:33
juejaeger: the only module I've to add was ata_piix, without that my CDROM didn't work even if I use ich9 in VBox08:44
jaegeryou had to add it to the iso and rebuild or just load it?08:45
jueno, to your virtualbox kernel-config :)08:46
jaegerohh, gotcha08:46
jaegeryeah, I made a new .config last night as well that should support MOST of the virtualbox hardware08:46
jaegerThat first one was my typical vbox setup, this one is more generic08:46
jaegerIt's uploaded as virtualbox-3.1.5.config-2 if you want to take a look, but if you're already set, don't worry about it :)08:48
juejaeger: thanks08:49
*** mike_k has joined #crux-devel08:50
jaegerGonna try to bootstrap another ISO today, the bootstrap hung up last night :( no idea why08:56
jaegerat least it'll include the new openssl port now08:56
rmullSome concerns from a new user: glib is shipped in core and the only thing that uses it is pkg-config. coreutils is compiled to use shared libraries.08:59
rmullAnything worth looking at here?09:00
juepkg-config is the reason why we moved glib to core09:01
jueand the possibility to add glib stuff to udev, if we ever had to09:02
juewrt coreutils, you mean
jues/had to/have to/09:04
juermull: ?09:13
rmulljue: Presumably09:16 isn't a 'normal' shared lib, it's exclusively used by stduf to change the std-buffers of called programs09:18
juesee stdbuf(1)09:18
juethat's the related mailing ->
rmulljue: Hmm, thanks, I'll forward that along09:25
rmulljue: Actually, the concern was that something like ls is linked against libacl and libattr, so if one of those is deleted, ls won't work09:30
juesure, without that extented acl/attr support won't work, he can always recompile coreutils with --disable-{acl,xattr,libcap} or something like that09:37
rmullIs that an adequate solution?09:43
rmullSo disclaimer - these are the people in #suckless that have their panties in a bunch09:43
rmullAre you familiar with them?09:44
rmullTheir goal is to link everything statically.09:44
rmullYou can see why they might be worried about coreutils being dynamically linked.09:45
*** mike_k has quit IRC09:45
jaegerIs there some reason we need to be concerned about their weirdness for CRUX?09:46
rmullYou don't /need/ to be concerned, of course09:47
rmullI hang around them because their (strong) opinions about software things give me food for thought and they often have good points about why things are bad09:48
jaegerfair enough09:48
rmullAs to whether coreutils should be statically linked, I assume that is so that you can have a very broken system and still have some level of recovery functionality09:49
jaegerthat would at least make sense... *everything* static seems overkill09:50
rmullYeah, that's a different story altogether09:50
rmullThat's one of their subprojects though.09:50
juermull: one point of not removing static libraries (*.a) in our ports is to give others the possibility to link things static09:50
jueso it should be possible to recompile coreutils all static09:51
rmullYes, I would agree that it is possible, and even trivial for someone who would know why they want it09:51
rmullBut the question is whether it should be a default or not09:51
jueIMO not09:51
jueif you want a mostly static system you should go with something like fefe's dietlibc selected programs09:54
juedid that in the past just for fun, works quite nice for many things09:56
rmullBecause the bloat would be too great?09:59
jueyeah, with glibc the minimal program-size is something like 1M IIRC10:01
juenearly, just tested, /bin/cut is around 650k now10:03
juethe other point is that static linking isn't offical supported by glibc10:05
rmullOkay, I see10:07
rmullThanks for the patience10:07
juethat's at least for all stuff the uses passwords etc. a valid point because the whole nss stuff cannot build static10:07
juenss = name service switch10:08
Romsterjaeger, i know i've been hacking a little with LTO lately seems i've nailed most the problems but i did spot glibc-32 won't compile unless i add -m32 to CC there has to be some bug in it's build system, only some lines in gcc haven't got CFLAGS set the rest do.11:00
Romstermost of my LTO issues are either unsupported in gcc or the port brakes on libtool script or something on a handful of ports the rest are due to missing -fPIC11:01
jaegeryeah, I've already fixed that locally, going to push it after this bootstrap succeeds (if it does this time)11:01
Romsterhmp another note /etc/rc.modules needs to be added to pkgmk.conf to not update unless rejmerge is ran.11:05
Romstererr pkgadd.conf sorry but i'm on jue's rc port.11:05
Romsterthere went my virtualbox modprobe modules -_-11:05
Romsteroh and i also noticed /etc/lilo.conf isn't in pkgadd.conf either.11:06
jueRomster: by default /etc is excluded11:07
Romsterah ok so something else must of nuked my lilo.conf last time i did a rebuild.11:08
Romsterwasn't this time but when i went to multilib it happend to overwrite it.11:08
Romstermight of been my mistake who knows.11:08
juesee here ->
Romsteridentical to mine.11:12
Romsterthis makes no sense then. i'll keep an eye on it because my rc.modules got clobbered.11:13
Romsterand i didn't do a rejmerge on it.11:13
Romsteranyway i'm off to nap again seems i've fixed nearly all the issues i had. system seems more responsive too.11:14
jueRomster: sorry, forgot that you're on rc-ng where rc.modules is a symlink to rc.hooks/sysinit/late and that is matched by the ^etc/rc.*$ rule12:06
jaegerbootstrap finished successfully this time13:48
jaegerRomster: does this work for glibc-32 with your ccache setup?: CC="${CC:-gcc} -m32" ../glibc-$version/configure14:15
*** y3llow_ has joined #crux-devel14:31
*** y3llow has quit IRC14:33
*** y3llow_ has quit IRC14:33
*** y3llow has joined #crux-devel14:34
*** y3llow has quit IRC14:36
*** y3llow has joined #crux-devel14:36
*** y3llow has quit IRC14:38
*** y3llow has joined #crux-devel14:38
*** acrux|G4 has joined #crux-devel15:32
*** acrux|G4 has quit IRC15:35
Romsterjaeger, i haven't tested but that looks right.19:01
Romsteruse $CC if set else use gcc19:02
jaegeryeah, that's the idea19:17
jaegerI tested it by setting CC=omgwtf and it failed as expected, so yeah19:17
jaegerman, I love Virtualization, so handy for some of my build use cases19:29
jaegerperfect example with the gnome ports19:29
jaegerI install crux multilib, take a snapshot19:29
jaegerinstall vbox guest tools, take a snapshot19:29
jaegerinstall gnome ports, test things19:29
jaegerdecide I want to start the gnome install from scratch with tweaked deps and port list, etc.19:30
jaegerrevert to snapshot #2 and go19:30
rmullThis is why I'm looking forward to btrfs19:45
jaegerI use both ZFS and VMware snapshots at work a lot, so nice :)19:45
*** mavrick61 has quit IRC21:23
*** mavrick61 has joined #crux-devel21:24
Romsternew nvidia 295.0921:30
Romsterhaven't tested it yet21:30
Romsterjaeger, something like this to remove duplicates
Romsterstill working on it it only echos so far.23:31
Romsterrun that as part of ports -u after it's done fetching files then preferhigher yes be ok to use.23:32
*** Romster has quit IRC23:33
*** Romster has joined #crux-devel23:34
*** Romster has joined #crux-devel23:34
Romsterreckon that be ok?23:34
Romstergotta go out be back in like 45minutes23:35
Romsterthough there is a down side you can't look at the Pkgfiles to see if the other one in the repo is newer version etc. unless i copy them to the branch above as Pkgfile.repo name.23:58
Romsterthough should use another tool to check the versions off the crux site perhaps i dunno.23:59
Romstero jsut off the git tree be better.23:59
Romsteror just*23:59

Generated by 2.11.0 by Marius Gedminas - find it at!