IRC Logs for #crux-devel Thursday, 2011-12-08

*** clb has joined #crux-devel00:23
Romsterjaeger, no but i would like it to be.00:49
Romsterhow far have you got on multilib tool chain?00:49
jueteK_: it's not possible to create a new file that has the same md5sum of an existing one, but it's possible to create two files with different content but identical md5sum01:15
jaegerRomster: still in progress, been pretty busy with work01:16
jaegernow off to bed, 1:16 here01:16
juegood night jaeger01:16
juermull: don't think so ;)01:18
Romsterg'night jaeger01:35
Romsteryeah pretty busy here too got some time now to mess with stuff. got work tomorrow then i got the weekend off.01:35
teK_jue: so it's *really* theoretical02:56
*** tilman has quit IRC03:08
*** tilman has joined #crux-devel03:08
RomsterEduardo Diaz has described a scheme by which two programs could be packed into two archives with identical MD5 hash. A special "extractor" program turn one archive into a "good" program and the other into an "evil" one.  http://www.codeproject.com/KB/security/HackingMd5.aspx03:13
Romsterhttp://www.mscs.dal.ca/~selinger/md5collision/03:13
Romsterand sha1 is broken too03:17
Romsterhttp://www.schneier.com/blog/archives/2005/02/sha1_broken.html03:17
Romsterthat's why predatorfreak and i've been suggesting sha256 since 2008 or so. but let it rest, i'm tired of "flogging a dead horse". http://en.wikipedia.org/wiki/Flogging_a_dead_horse03:19
frinnst:>03:23
frinnstiirc, the argument for remaining with md5 was that it did the job of verifying that the download was complete and not corrupt "good enough"03:37
Romstershe'll be right... in aussie slang.03:55
Romstermeaning at some point it will fail spectacularly.03:56
Romsternow back to code stuff.03:56
jueRomster: as I said, it's possible for a malicious packager to create two different files with the same md5sum, but not create one new file that has the same md5sum as an already existing file04:16
juebut I wrote that in FS#223 already ...04:17
teK_that's afine but important difference to what I thought earlier ;)04:17
jueyeah, indeed04:17
teK_I did not follow the agrument too close thinking we could switch to shaX without too much hassle .. so why not :-)04:20
teK_*argument04:21
frinnst+104:21
*** mike_k has joined #crux-devel04:21
juesure, I'm not against it, but the arguments of FS#223 are not really a reason for it04:22
juefor our usage at least04:23
Romsterjeu, yes i read it, i don't mind you know more detains than i do about it. plus you do one heck of a job maintaining crux04:24
Romsterjue*04:24
Romsters/detains/details04:24
Romsteri've noticed it's mostly jue that's looking after core04:25
Romsterare we that low on active devs now?04:25
teK_as long as jue/tilman  and the users won't complain :-)04:26
juewell, we got a new opt maintainer lately ;)04:27
teK_:>04:27
juebtw, if we are talking on new features04:28
teK_not listed at http://crux.nu/Main/About04:28
jueoops, could you fix that, please?04:28
teK_sure04:28
teK_+ I did not forget the pmwiki update..04:29
jueanyone brave enough to test Michal's rc-ng?04:29
teK_maybe next week04:30
jueI've prepared a rc port in my privat repo, so the update should be pretty painless04:30
jueone thing is very important though:04:30
juerun rejmerge after the update04:30
jue!04:30
teK_I will try to remember that one ;)04:32
juewould be very helpfull if we have a test from someone who is using raid/lvm stuff04:33
* jue looks at Romster04:33
teK_done +  moved the profile page to http://crux.nu/Profiles/DannyRawlins04:34
mike_kwas there an agreement to repost rc-stuff summary message to the general ML?04:35
Romsterindeed i need to do more in opt too.04:35
Romsteri'm currently looking over my contrib ports an fixing a few things.04:36
juemike_k: yep, there was04:36
juebut looks like tilman forgot to do it04:36
teK_Han Boetes is not listed at http://crux.nu/Main/RetiredMembers :D04:37
Romsteri'm using raid and lvm but i got runit i'd had to do major surgery to try that rc-ng04:37
jueteK_: he was never a CRUX member04:37
Romsteri could however install on a fresh root disk and access my raid/lvm stuff that way.04:37
teK_oh ok04:37
Romsteri do have quote a bit of lvm and some mdadm knowledge.04:38
jueRomster: maybe time to switch to more standard CRUX ;)04:38
Romsteri'm also quite good at ddrescue/testdisk and recovering data off faulty clients disks.04:38
Romsterbut i like runit's speed :/04:39
teK_I never felt /etc/rc to be slow04:39
Romsterthough i have been considering the effort involved in what i've been doing the past few years.04:39
Romsteri did redo my firewall with standard crux this time however.04:39
jueon improvement of rc-ng is that you can start services parallel :)04:40
Romsterload a few more services then it is slow.04:40
Romsterhmm and watchdog restarting?04:40
jueno04:40
Romsterthat and parallel starting of services might get me away from runit.04:40
teK_jue: I did that by appending '&' :P04:41
teK_Romster: for this laptop there are 11 services to start04:41
Romsteri do notice that my system does some head thrashing at startup i'm getting a SSD later that will migrate that.04:42
Romsterthough i don't reboot that much either.04:42
Romsterhonestly why do i want to pursue runit... is it that too complicated for crux to consider it's use too?04:43
Romsteri should look at rc-ng at least too see what the fuss is about it.04:43
Romsterjue, on another note any chance of this making it into crux 2.8 http://romster.dyndns.org/linux/ports/hvlinux/core/binutils/enable-relro-new-dtags-and-optimise-by-default.patch i've been using that for a couple of years now with no issues. other than it requires everything to be rebuilt.04:52
jueRomster: sure, we can consider that, best would be to have a feature request in our FS04:55
Romsteri'll add one. it's a bit quicker and i've had zero issues from it.04:55
Romsterdone't know if it's ok on 64bit though04:56
Romsterlooking over rc off crux as it is and the one in jues repo now04:57
jueI've just made FS#233 a CRUX 2.8 task, so we should finally decide what to do05:01
Romster223 or 233 i can't see a 23305:05
juesorry, was 22305:10
jueRomster: wrt watchdog, I'm using contrib/monit on my server, it's quite nice and simple to use05:16
Romsterah05:19
Romsteri guess that would do the job though what happens if monit itself crashes. where in my case it's the init system.05:22
Romsterthough pretty rare case there.05:23
Romsterit's more tempting ot move back to crux bar having that binutils patch and gcc 4.6.205:24
Romsterhmm i need to check over anything else i did modify and submit FS feature requests if it's in crux's interest to consider them.05:24
jueit's explained in the README how to setup monit within inittab05:24
Romsterk i might look into that on the weekend.05:25
Romsternot sure what i'm doing yet.05:25
*** jaeger has quit IRC08:21
*** clb has joined #crux-devel08:23
*** clb has joined #crux-devel08:35
*** clb has joined #crux-devel08:45
*** jaeger has joined #crux-devel08:46
*** clb has joined #crux-devel08:55
*** clb has joined #crux-devel09:07

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!