IRC Logs for #crux-devel Friday, 2013-01-25

*** horrorSt1uck has joined #crux-devel00:50
*** horrorStruck has quit IRC00:53
jue.oO new xorg-server 1.13.201:44
frinnstyeah: "No new commits since rc1."02:38
frinnstuseful changelog in the announce mail02:38
*** joe9 has quit IRC04:48
*** heroux has quit IRC05:06
*** heroux has joined #crux-devel05:13
jaegerBuilt a new server with a motherboard that has a type A USB port on the board, inside the case... I'm going to use CRUX to build a little rescue image for it, could be handy in the future :)05:21
*** joe9 has joined #crux-devel05:48
*** horrorSt1uck has joined #crux-devel06:13
frinnstjaeger: yeah more and more servers have that06:29
frinnstperfect for esxi, just install it on an usb-stick06:29
frinnstwe have esxi installed on sd cards on our newish bladeservers06:30
jaegeryep, it's the same motherboard I used in my ESXi server at home, love it06:32
jaegerhave a Lexar Echo ZX usb drive in it, it's tiny06:32
jaegerWill get another one of those as it will fit into a 1U case without any trouble06:33
jaegerour ESXi servers here at work are a few years old and boot from CF but the new ones I'm ordering this year will boot from SD, should be a decent speed increase over the CF06:33
frinnstour previous gen blades booted from disk06:34
frinnstpointless and expensive06:35
jaegerwe have a mix... I want a couple VM hosts with local disk just in case our NAS is down but for the majority they boot from flash06:41
frinnstIf our san would go down, i'd flee the country :D06:51
frinnsteveything would still be around, but dat rebuild06:51
frinnstthat06:52
jaegerhehe06:52
frinnsthopefully a new netapp so we can mirror it properly in the future06:53
*** felixos has joined #crux-devel07:06
jaegerhow odd... blkid from busybox doesn't print anything08:50
jaegerwell, that's inconvenient08:52
frinnstand pointless :)09:06
jaegeryeah09:06
jaegerI want to use it for my little USB rescue thing with rr=UUID on the command line09:06
jaegersince it doesn't work, though, I'll have to fall back to brute force checking everything in /proc/partitions09:07
jaegerit would have been helpful for the ISO as well09:07
jaegeralso odd is that extlinux can't seem to boot memtest86+ for me10:45
jaegerjust get a repeating "0104"10:45
jaegerah, I bet that's because it has a .bin extension and extlinux thinks it's an ISO or something10:47
jaegerwill try removing the extension10:47
jaegerok, that fixes memtest10:56
*** joe9_ has joined #crux-devel11:40
jaegerugh... I hate comparing vendor quotes for servers11:41
teK_don't you have a supplier that does that for you? :11:42
teK_+ hi folks11:42
jaegerthe most expensive one has the weakest CPUs but it includes a SAS RAID controller I specifically asked them to remove11:42
teK_nice11:42
jaegerIt depends... we go straight to dell (via a rep) for dell stuff, we go through CDW for HP stuff, we go directly to oracle (via a rep) for oracle stuff11:42
teK_oh ok11:43
jaegerThe HP ProLiant DL360e Gen8 is looking most appealing right now but I need oracle to give ma revised quote without the RAID controller and drives11:43
jaegerI don't want any drives in these at all11:43
*** joe9 has quit IRC11:43
jaegerThey're going to boot ESXi from an SD card and talk to a couple NASes11:44
*** felixos has quit IRC11:46
jue.oO something strange happend with our opt repo12:30
teK_I looked at the log and it seemed fine12:32
teK_no?12:32
jueno12:32
teK_fuck12:33
juelooks like everything we push to 3.0 exclusively has been merged into 2.812:33
jues/push/pushed/12:33
teK_yeah there was a  bogus assumption on my side12:34
juee.g. libnl12:34
teK_so http://stackoverflow.com/a/1895095 would be the route to take?12:35
juedid you merged from 3.0 to 2.8?12:35
teK_yes12:35
teK_I confused it with 2.8 -> 3.012:35
teK_...12:35
juef***12:35
teK_it's been a long day *gasp*12:35
jueFTR here the wrong merges:12:35
juejdk jre libnl nspr nvidia pango pangox-compat python valgrind wine wpa_supplicant12:37
juenot sure about valgrind12:37
jueteK_: are you going to revert this?12:39
teK_still figuring out what to do tbh12:39
juewell, checkout 2.8 and revert the last commit to jdk, jre etc.12:40
jueI don't see a possibility to do this in one commit12:41
juebut I'm a git dummy, ask Tilman probably12:41
teK_how did you identify the affected ports?12:43
juef***, I forgot some merges in the above list:12:43
jueat-spi2-atk at-spi2-core epdfview firefox-java-plugin flash-player-plugin git grub2 gkt3 harfbuzz12:44
jueI pulled to opt and saw the mess ;)12:45
jue.oOo...12:46
teK_I cannot even find the commit that did the merge12:48
teK_just like the others12:48
juethe date of the commit is on both branches the same12:50
juepuuuuh, probably easier to restore a backup ...12:52
juejust looked at libnl, I did several commits to 3.0 and everything is merged into 2.8 now, of course12:53
rmullSo it's the 2.8 branch that got wacky? Full of stuff that should have been 3.0 only?12:53
juelooks like i have to revert 5 commits for it12:53
teK_jue: we cannot to backups atm because NDC won't let me ssh to frinnst's server12:54
juermull: yeah12:54
juehmm12:54
teK_you don't have to do the reverts12:55
teK_I'm just a bit lost atm, tbh12:55
teK_so you say commit 1c4dc22b27cd72db32f12c69c73f02460cc5cbf4 has to be reverted, for example?12:57
juejust a moment12:58
jueyeah that works, I've done it for libnl, let me push it12:59
jueheh, nice :)13:00
teK_how do I know if reverting only the last commit for all ports from above will suffice13:00
jueteK_: you don't13:04
jueit's hairy, look at nspr for example13:07
teK_do I have to add that I'm sorry? -.-13:07
juedo a: git checkout 2.8 ; tig nspr13:08
teK_ok13:08
teK_ah13:08
juethe problem here is the commit 'Initial consolidation of opt ports for CRUX 3.0'13:09
teK_so in 2.8 it is supposed to have 99ef3ad60c81e30821dc13f23b7d2750d06b3a03 as the most recent commit?13:09
teK_i.e. bevore the 'branching'?13:09
teK_s/v/f/13:09
juebecause it adds --enable-64bit and is the first commit where 2.8 and 3.0 differs13:09
juewell, almost13:10
teK_but 60279f1fb3de6a32f6dc341973122c060294caa2 also is part of the diverged graphical branch13:10
*** mike_k has quit IRC13:11
juethe commit 'updated to 4.9.4' is correct13:11
jueso you have to revert only the one commit13:11
jueit's a pain but we have to do it ;)13:13
teK_I have to do it13:13
jaegerMan, I leave for a few minutes and teK_ blows up git!13:53
teK_always a pleasure, sire13:53
jaeger:D13:54
teK_DEAREST of all sires13:55
jaegerI have no offspring!13:55
jaegeralso my inbox exploded13:55
teK_excuse me I have to clean up the mess I left13:56
juejaeger: are there any diff between 2.8 and 3.0 for grub2?14:02
jaegerI don't think so14:04
jueok, thanks14:04
teK_jaeger: is pangox-compat in 2.8?14:07
jueno, it's 3.0 only14:12
teK_ok14:12
juemachst du dann noch bitte: firefox-java-plugin flash-player-plugin14:15
jue?14:15
teK_klar14:15
teK_jaeger: did you pull in during the last 60-80 minutes?14:16
teK_more precise: last 120 minutes14:16
teK_if not I would higly appreciate seeing a copy of opt/wine14:17
teK_:O14:17
jaegersorry, was driving home. I have not pulled recently14:41
jaegerI'll reboot and get wine if you still need it14:42
teK_I think jue and I already figured it out but you can show my the Pkgfile if you want14:42
teK_just to be sure14:42
frinnstgeez, whats happened?14:54
frinnst:)14:54
jue.oO don't ask :)14:54
teK_you got mail14:56
teK_except frinnst, haha14:57
jueour history in the 2.8 branch is now a bit, let's say strange, but besides that everything should work again14:57
jaegersorry, had a phone call for a while there, back now15:09
teK_no problem15:09
jaegerhttp://pastebin.com/bFy0bJjj15:09
jaegerugh, someone should really take that apostrophe out :D15:10
teK_hrhr15:10
teK_we have a Term for that at least in german15:11
teK_it's the deppen-Apostroph15:11
teK_you can see it all the time ;)15:11
jaegerapostrophitis15:11
teK_ok so this the 3.0 branch I think15:11
jaegeroh, did you want a 2.8 wine?15:11
jaegersorry if I missed that15:12
teK_please :)15:12
teK_SIRE15:12
jaegermultilib or no?15:12
jaegerI guess if you're fixing things in git, non-multilib15:13
jaegerhttp://pastebin.com/jbyNUexN15:13
teK_ok thank you15:15
teK_could you upload the .footprint and .md5sum, too?15:16
jaegersure, sec15:17
teK_thank you15:18
jaegerhttp://jaeger.morpheus.net/linux/crux/tmp/wine.tar.xz15:18
teK_yay15:19
jaegerout of curiosity what made this stuff hard to fix? was it not possible to just go back to a previous hash?15:20
jaeger(I was pretty busy at work so didn't follow closely)15:20
*** joe9_ has quit IRC15:20
teK_well.. which hash15:20
teK_I could not find it15:20
jaegerah15:21
jaegerwell, when you do a merge like 2.8 -> 3.0 it adds X commits to the 3.0 tree, right? so you could, theoretically, go back X commits using git log or gitweb or something?15:21
jaegerexcept in the 2.8 tree instead of 3.0, just used that as an example15:22
teK_I was not quite sure about that, tig just looked messed up15:22
teK_I hope we fixed everything15:22
teK_while wine disturbs these hopes a bit15:22
teK_D:15:22
jaegerdoh15:22
teK_it took me about three hours so at least I was punished already *g*15:23
jaegeroof, that sucks15:23
teK_on a somewhat related sidenote: watching Flight now15:24
teK_;)15:24
jaeger:)15:25
jaegerman, I love remote KVM access15:31
teK_yes it's nice15:33
teK_fujitsu requires you to enter a license code if you purchase one their servers with a builtin kvm device15:34
jaegerI think HP and dell do some of that, too15:36
jaegeroracle doesn't, supermicro doesn't15:36
teK_it's just stupid15:36
teK_now they are flying upside down15:40
teK_Denzel is a DEVIL15:41
jaegerheh15:41
teK_nooo Denzel don't drink that liquor while flying15:50
teK_dammit boy15:50
rmullIs it safe to run a ports -u opt on 2.8 now?15:59
teK_when did you run it the last time?16:04
teK_rmull16:04
teK_if you pulled today or yesterday and not during the last four hours  you really should send me a tarball of your opt repo16:08
rmullLet me see if I can figure out when that was16:16
rmullpretty sure I sysupped either this morning or last night16:17
teK_yay16:18
teK_thanks ;)16:18
rmullSeems like the list thing I built may have been ghostscript?16:18
rmullLet me prtsweep it, then I'll send it16:19
rmullActually, maybe you don't want it prtsweeped?16:20
rmullNah, I think it'll be okay16:21
teK_if it won;'t contain packages send it16:21
rmullteK_: http://lambda.rmull.com/opt.tar.xz16:33
rmullLet me know when you've got it so I can take it down16:33
teK_connection refused16:35
teK_works16:38
rmullMust have been a problem on your end <_<16:39
rmullpesky router16:39
teK_sure16:40
teK_250KB/s are a nice upload16:40
teK_I have that too <316:40
teK_done16:46
teK_thank16:46
teK_s16:46
rmullcheers, hope you find what you're looking for16:47
rmullagain, not sure if it was taken inside the window that you needed, so tread carefully16:47
rmullIt's probably right though16:48
teK_kk16:51
frinnstnot many changes between xorg-server 1.13.1 -> 1.13.217:11
frinnstdid anybody push an update yet?17:11
frinnstdont look like it - i'll do it17:12
Romstergit merge can be dangerous, use with caution. But I guess teK_ now knows that lol.17:36
Romsterhttp://stackoverflow.com/questions/1895059/revert-to-a-commit-by-sha-hash/1895095#1895095 <- isn't that rewriting the history, that's even more dangerous than git merge.17:37
teK_reverting is not rewriting I think17:40
jaegerI suppose an easy but messy way would be to check out the old hash and commit it as a new commit17:41
jaegermessy in that it would look weird but it wouldn't mess up the history17:41
teK_yes that was my idead, too17:41
Romsterreverting isn't but hard resetting is.17:46
Romsteri would never do a hard reset or even a soft reset after it's been pushed.17:46
Romstercompat-32 is maintained by what group now? i keep seeing the maintainer line and want to edit them all.17:52
Romsterwould it be CRUX compat Team, compat-ports at crux dot nu17:53
Romsteror something.17:53
teK_there's no such alias (yet) I think17:54
teK_and I won't touch anything sensitive tonight17:54
teK_not even frinnst17:54
Romsterand we could do a i686 overlay for those still stuck on 32bit systems. and keep that as a group project for anyone that's trusted and on 32bit.17:54
frinnsthaha17:55
Romsterlol17:55
*** joe9 has joined #crux-devel18:02
*** joe9 has quit IRC19:04
*** joe9 has joined #crux-devel19:06
*** __mavrick61 has quit IRC19:40
*** __mavrick61 has joined #crux-devel19:42
*** joe9 has quit IRC20:21
Romsternew pkg-config is out 0.2822:26
Romstersure hope that don't break anything.22:26
*** mike_k has joined #crux-devel23:37

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