IRC logs for #baserock for Monday, 2015-01-26

*** zoli__ [] has quit [Remote host closed the connection]00:31
De|ta_ is now known as De|ta08:02
*** mike [] has joined #baserock08:07
mike is now known as Guest4560308:08
*** zoli__ [] has joined #baserock08:23
*** mariaderidder [] has joined #baserock08:42
*** CTtpollard [] has joined #baserock09:00
*** rdale [] has joined #baserock09:04
*** jonathanmaw [] has joined #baserock09:17
*** tiagogomes_ [] has joined #baserock09:38
*** grahamfinney [] has joined #baserock09:51
richard_mawsorry straycat, no inline chunks yet09:53
* pedroalvarez_ uses virt-manager to spin up a VM in baserock 09:58
Kinnisonpedroalvarez_: woot09:59
pedroalvarez_only thing needed was openbsd-netcat09:59
persiaNo changes to morph?09:59
pedroalvarez is now known as pedroalvarez10:00
*** ssam2 [] has joined #baserock10:00
Mode #baserock +v ssam2 by ChanServ10:00
pedroalvarezso I have a PC with the virtualization.morph stratum, and I connected to it with virt-manager to setup a VM10:01
pedroalvarezno, no changes to morph nor to definitions, since I installed `nc` manually just for tetsting10:01
Kinnisonpedroalvarez: I take it the issue is that busybox's netcat isn't openbsd cmdline compatible?10:03
persiaAh, so baserock delivering a tested-working libvirt-based hypervisor, rather than baserock deploying to a libvirt-based target.  Apologies for my confusion.10:03
pedroalvarezKinnison: virt-manager needs a netcat version with the '-U' option available10:04
pedroalvarez-U: Use UNIX domain socket10:04
* Kinnison prefers openbsd-netcat anyway, so perhaps we could make that default10:05
pedroalvarezKinnison: I found difficult to find the source, and I just installed a fork of it. Would you help me to find the right one?10:06
KinnisonI'd start from whatever Debian has10:06
pedroalvarezDebian has a fork of the openbsd one10:06
pedroalvarezwait, this was more hecky than I normally do. I Copied the `nc` binary from my Debian OS. :)10:09
persiadebinject was writted specifically to avoid the need to do that10:09
pedroalvarezI think this is the one:
*** Krin [] has joined #baserock10:13
KinnisonYou might find of use10:13
pedroalvarez\o/.git! Thanks10:15
KinnisonTTOTD, apt-cache showsrc blahblah | grep Vcs10:16
Kinnisona handy way to find stuff10:16
KinnisonNot always immediately obvious how to get thence to upstream, but it's a good starting point10:16
pedroalvarezI see.10:17
pedroalvarezI only know how to do those things in baserock :) 10:17
*** lachlanmackenzie [] has joined #baserock10:18
persiadebian/copyright in that repo claims as upstream10:21
KinnisonMmm, but the patches in the debian repo look useful10:21
persiaYes, but they aren't commits, so need a fork repo to be consumed by morph10:21
Kinnisonaye, tis awkward10:22
*** sambishop [] has quit [Ping timeout: 252 seconds]10:31
*** sambishop [] has joined #baserock10:31
*** grahamfinney [] has quit [Ping timeout: 246 seconds]10:32
*** franred [] has joined #baserock10:53
*** grahamfinney [] has joined #baserock11:32
grahamfinneyHi, doing the initial Baserock quick start I hit a barrier. I have detailed it at Any ideas?11:48
paulsherwoodwhere did the <grahamfinney> problem with: morph init workspace -> /use/bin/morph: line 3: permission denied11:50
paulsherwoodtypo come from?11:50
persiagrahamfinney: Is this with use-latest-morph?11:50
tiagogomes_`chmod +x /use/bin/morph` ?11:51
persiaThe problem is "use", should be "usr", not an actual permission issue.11:52
persiaI think a symlink or path is broken somewhere11:52
persiagrahamfinney: Just to check, what does `which morph` return/11:55
grahamfinneyI uses /usr/bin/morph11:57
radiofreepaulsherwood: what's the output of 'cat /usr/bin/morph'11:57
persiaHow about `ls -l /usr/bin/morph`?11:57
radiofreepersia: i think /usr/bin/morph is a script rather than a direct symlink?11:58
radiofreeunless the instructions have changed11:58
radiofreesorry, grahamfinney, what's the output of 'cat /usr/bin/morph'11:58
grahamfinneypersia: latest morph? I think so, I just pulled it down11:59
straycatit's a script that appends /src/morph to the $PYTHONPATH and runs /src/morph/morph11:59
persiagrahamfinney: BY "pulled down", do you mean "downloaded the build system"?12:00
radiofreeit sounds like there's a problem inside the /usr/bin/morph script12:00
persiagrahamfinney: I wanted to confirm you were not doing
SotKgrahamfinney: what happens if you do `export PYTHONPATH="$PYTHONPATH:/src/morph"; /src/morph/morph init workspace`?12:00
grahamfinneypersia: cat /usr/bin/morph -> PYTHONPATH="$morphpath" "$morph" "$@"12:02
persiagrahamfinney: Check the values of those variables in your environment12:03
radiofreegrahamfinney: that looks wrong12:03
grahamfinneyNo I downloaded the binary12:03
persiaWhat binary?12:03
persiaExcept don't do that unless you are hacking morph: for most folk, it's better to just upgrade morph in the development system12:04
radiofreegrahamfinney: if that's all that's in your /usr/bin/morph file, you're not setting "$morphpath"12:04
grahamfinneyI doanloaded the binary pointed to by the quick start tut12:05
radiofreeedit /usr/bin/morph and add "morphpath=/src/morph" above the PYTHONPATH=... line12:05
grahamfinneyI'll give that a go radiofree12:05
persiagrahamfinney: ?12:06
grahamfinneyI can't keep up with the flood......12:06
grahamfinneythat's alde12:07
grahamfinneythat's already in there radiofree12:08
radiofreeare you sure you git cloned morph into /src/morph?12:09
grahamfinneyi.e. morphpath is set12:09
* persia continues to believe that putting morph in /src/morph is generally to be avoided12:09
grahamfinneynot sure at all radiofree. I'll look over it ...12:10
radiofreepersia: why?12:11
persiaradiofree: Because 1) there are better ways to install/upgrade software in devel systems, and 2) it increases the potential to cause issues working only in /src to run software from /src12:12
persiaObviously, if folk are hacking morph, it is critical to be able to use /src/morph, but I think most end users should not be doing that.12:12
*** grahamfinney_ [] has joined #baserock12:17
*** grahamfinney [] has quit [Ping timeout: 246 seconds]12:18
straycattbf, the wiki does say "If you want to use 'bleeding edge morph'" I was hoping the term "bleeding edge" might suggest most users don't want this12:22
*** grahamfinney_ [] has quit [Ping timeout: 246 seconds]12:23
persiastraycat: I think it should.  The bits that worry me are 1) the presence of "use" instead of "usr" somewhere: I'd like to understand where that comes from, and 2) telling folk to set /src/morph for morphpath before we understand the current values.12:25
*** grahamfinney [] has joined #baserock12:31
straycatpersia, i'll update the wiki to make the clearer then12:32
persiaIf you think it is necessary.  I think that new readers of the old text will not try it without prompting, and that folk who have been using /src/morph for a long time will continue to recommend it if not reminded (but are unlikely to read the relevant page).12:33
tiagogomes_I didn't find any "use/" on the documentation. I don't think that the instructions for using the last morph should be on "Using Baserock", they could be moved to Guides perhaps12:34
grahamfinneyTurns out the problem is that PYTHONPATH wasn't quite right. it was set to "$morphpath" "$morph" "$@", but "$morph" should have been "$morphpath/morph". Thanks ripsum.12:38
persiatiagogomes_: Moving it sounds sane.12:38
persiaI think the "use" is a typo somewhere, but I don't know where.12:38
persiaFrom the investigation so far, I suspect the environment variables that morph is consuming.12:38
persiaBut it worries me that "/src/morph" was already in /usr/bin/morph12:39
tiagogomes_persia, I am pretty sure that it doesn't come configured to look at /src/morph from factory12:40
persiatiagogomes_: That's why I'm worried.  I think grahamfinney's environment differs from default in odd ways, and want to understand how so it can be repaired.12:41
* SotK assumed that "use/" was a typo in the original error description on friday since it wasn't in the pastebin that grahamfinney sent today, but could be wrong.12:46
persiaThat would be a sensible explanation.  grahamfinney ^^ ?12:47
grahamfinneyI appeared made an error in copying over the commands (from "Using the latest version of morph")13:23
persiaI'd start again, as it probably will be easier than untangling that.13:24
persiaAnd not go through "Using the latest version of morph" next time, unless you really need to hack on morph.13:24
persiaIf you do need to hack on morph, then troubleshooting your current installation will probably be educational in a way that assists your hacking on morph.13:25
grahamfinneyI don't recall if I did that, but I remmeber having a loook at it, so I may have,13:26
*** tiagogomes_ [] has quit [Quit: Leaving]15:00
*** tiagogomes [~tiagogome@] has joined #baserock15:03
*** mdunford [] has quit [Ping timeout: 256 seconds]15:30
*** grahamfinney [] has quit [Ping timeout: 245 seconds]15:58
jjardonHi, anyone knows how big is the minimal baserock system?16:05
ssam2a few hundred MB16:05
rjekISTR ~150MB, but this number is out of date16:05
pedroalvarezjjardon: is a bit big right now, but there was a patch to strip more the gcc chunk16:06
pedroalvarezmaking it smaller than 30M (I believe)16:06
jjardonoh, thats nice, thanks!16:07
flatmush1 is now known as flatmush16:26
mauricemoss_I can't make libstdc++ available to stage2-gcc when cross-bootstrapping for mips64 does anyone have an idea how to solve this?16:38
*** flatmush [] has quit [Quit: Leaving.]16:43
*** flatmush [] has joined #baserock16:47
*** zoli__ [] has quit [Remote host closed the connection]16:50
*** zoli__ [~zoli_@] has joined #baserock16:50
tiagogomesmauricemoss_  where that came from? /src/moph.log? I never saw exceptions on build failures17:06
mauricemoss_tiagogomes, yep that's from morph.log17:07
tiagogomesmauricemoss_, where is libstdc++ installed?17:07
mauricemoss_I found it in ../tools/lib32/libstdc++.a17:10
mauricemoss_actually that makes me wondering, what gcc excpects for mips64. probably not /lib3217:11
tiagogomesargh, it should go to /tools/lib17:12
mauricemoss_let me check..17:13
tiagogomeswhat `./tools/bin/mips-baserock-linux-gnu-ld --verbose | grep SEARCH` shows, if executed from the failed staging area17:13
tiagogomesI don't expect that the executable name is named `./tools/bin/mips-baserock-linux-gnu-ld` exactly17:13
mauricemoss_SEARCH_DIR("=/tools/mips64-bootstrap-linux-gnu/lib32"); SEARCH_DIR("=/tools/lib"); SEARCH_DIR("=/tools/lib64"); SEARCH_DIR("=/tools/mips64-bootstrap-linux-gnu/lib");17:15
*** zoli__ [~zoli_@] has quit [Remote host closed the connection]17:16
*** zoli__ [] has joined #baserock17:16
mauricemoss_oh I forget to mention I'm not using your v2 branch yet, because I figured out there was no change in this area.17:16
tiagogomes mauricemoss_ there you go, /tools/lib32 is not on the ld path17:16
tiagogomesmauricemoss_, did you pass --with-lib=/tools/lib to gcc configure's script?17:17
*** zoli__ [] has quit [Remote host closed the connection]17:18
mauricemoss_I didn't, will add it :)17:20
tiagogomesmauricemoss_, unfortunately, I don't think that is going to do any difference, gcc seems to ignore that flag. 17:21
* ssam2 starts a build of a system based off latest master and watches freshly built artifacts get downloaded from cache.baserock.org17:21
ssam2the system works!17:22
tiagogomesssam2, I just realized that I forgot to update the unpetrify-refs, I'm cooking a patch17:22
ssam2doh, it's got to stage2-gcc and started building17:25
ssam2I guess my laptop caught up with :)17:26
mauricemoss_tiagogomes, if gcc is ignoring the flag, how can I make the libstdc++ know to gcc?17:27
ssam2it might help to look at the configure script to see why gcc is ignoring the flag17:30
tiagogomesmauricemoss_, did you disabled multilib as well?17:30
mauricemoss_I did.17:30
ssam2running 'grep '/tools/lib32' *' in a failed staging area might give you some useful info17:30
*** jonathanmaw [] has quit [Quit: Leaving]17:31
*** mariaderidder [] has quit [Quit: Ex-Chat]17:33
*** zoli__ [] has joined #baserock17:39
*** ssam2 [] has quit [Remote host closed the connection]17:43
*** Krin [] has quit [Remote host closed the connection]17:46
*** mdunford [] has joined #baserock17:50
*** grahamfinney [] has joined #baserock17:57
*** tiagogomes [~tiagogome@] has quit [Quit: Leaving]18:09
straycatahh, i think some of the requirement processing must have changed between the setuptools version we're using in baserock and current, cause i get different dependencies for python-glanceclient18:16
straycat12.0.5 was released pretty recently, we could probably move to that and our delta against setuptools would reduce to one line18:18
persiaUpgrades that reduce delta are good.18:22
*** Guest45603 [] has quit [Quit: Leaving]18:23
straycatyeah sadly that's not actually the problem here >.>18:27
pedroalvarezI wonder if cross-bootstrap is broken in general or only for the new architecture that mauricemoss_ is trying to use 18:35
persiapedroalvarez: Do you still have ppc64 hardware?  Maybe run a test there?18:52
pedroalvarezI'll make a note to do it. 18:53
*** radiofree [radiofree@unaffiliated/radiofree] has quit [Ping timeout: 244 seconds]19:25
*** lachlanmackenzie [] has quit [Ping timeout: 252 seconds]19:30
*** rdale_ [] has joined #baserock19:30
*** rdale [] has quit [Ping timeout: 276 seconds]19:33
*** rdale [] has joined #baserock19:37
*** rdale_ [] has quit [Ping timeout: 240 seconds]19:38
*** radiofree [radiofree@unaffiliated/radiofree] has joined #baserock20:33
*** radiofree [radiofree@unaffiliated/radiofree] has quit [Remote host closed the connection]20:33
*** radiofree [radiofree@unaffiliated/radiofree] has joined #baserock20:33
*** jjardon [sid723@gateway/web/] has quit [Remote host closed the connection]21:12
*** jjardon [sid723@gateway/web/] has joined #baserock21:24
*** grahamfinney [] has quit [Ping timeout: 245 seconds]22:36
*** franred [] has quit [Quit: Leaving]23:04
*** zoli__ [] has quit [Remote host closed the connection]23:56

Generated by 2.15.3 by Marius Gedminas - find it at!