IRC logs for #baserock for Monday, 2014-10-20

*** ridgerunner [~robjones@access.ducie-dc1.codethink.co.uk] has quit [Ping timeout: 250 seconds]03:21
*** ridgerunner [~robjones@access.ducie-dc1.codethink.co.uk] has joined #baserock03:21
*** dutch [~william@82.70.136.246] has joined #baserock07:16
*** fay_ [~fay@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock07:23
pedroalvarezgood morning!07:32
*** tiagogomes [~tiagogome@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock08:02
*** mSher [~mike@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock08:08
pedroalvarezmeh, adding stage2-bash before glibc "solves" the problem but causes another08:12
pedroalvarezglibc replaces a placeholder in some files  (@BASH@ ?) pointing to wherever bash is installed08:13
pedroalvarezand adding a stage2-bash causes glibc to think that it's on /tools/bin/bash08:14
*** jonathanmaw [~jonathanm@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock08:18
KinnisonSo glibc's configuration needs teaching where bash is *going* to be :-)08:19
KinnisonDo you know which subtree does that?08:20
pedroalvarezKinnison: not yet08:22
KinnisonI think I can see one instance (ldd)08:22
KinnisonAnd it'll be a pain08:22
Kinnisonbecause it substitutes what the bash shell itself thinks it is08:22
Kinnisonwhen you run bash, it sets $BASH to the path to itself08:22
Kinnisonand this doesn't seem to be overrideable08:22
KinnisonWe might actually have to build bash specially08:24
Kinnisonand I wonder if that'll cause "issues" for the bootstrap08:24
Kinnison:-(08:24
pedroalvarezindeed, after merging any of these changes I have to test cross-bootstrap08:25
pedroalvarezbefore***08:25
pedroalvarezmaybe I can try to fix the root cause:08:28
pedroalvarezchecking for bash... /tools/bin/bash08:28
Kinnisonthat's not the core root cause :-(08:28
Kinnisonbash itself knows it was built for /tools08:29
Kinnisonso when it runs, $BASH == /tools/bin/bash08:29
Kinnisonand that doesn't seem to be controllable from the external environment08:29
Kinnisonplease stop pulling that face08:29
pedroalvarezcan we hack bash?08:29
pedroalvarezso it thinks that it's on /bin/bash?08:30
Kinnisonpossibly, but who knows what might break08:30
KinnisonI'm looking now08:30
ridgerunnerThere was talk last week of a meeting this afternoon to discuss Baserock. Is that actually on? If so, when?08:30
petefothridgerunner: yes. checkout http://doodle.com/9azzxpn9zvh8g8tq. The decision is to meet at 15:00 today in 30208:32
petefothand to cover (the first four points of) the agenda at http://piratepad.net/8TnPqU6CIj08:33
Kinnisonpedroalvarez: http://git.baserock.org/cgi-bin/cgit.cgi/delta/bash.git/tree/variables.c#n70408:33
Kinnisonpedroalvarez: I think it's that function which derives the value of $BASH08:33
petefothridgerunner: It's all in an email from jjardon :)08:33
*** De|ta [~arc@195.242.156.171] has joined #baserock08:36
*** violeta_ [~violeta@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock08:39
pedroalvarezKinnison: thanks! but looks complicated to me :(08:39
juergbipedroalvarez: i simply symlink /bin/bash (and /bin/sh) to the stage2 bash08:39
juergbiwith the right PATH, this should work. at least it does for me (not baserock, though)08:40
Kinnison/bin/bash symlinked might work08:40
KinnisonThat function does trawl $PATH so providing /bin is before /tools/bin in the PATH it should work08:40
* Kinnison finds reading GNU-style C very tiring, so I might have gotten confused though08:41
Kinnisondepressingly it might find it from elsewhere too08:42
Kinnisonbleurgh08:42
pedroalvarezI need more understanding about how we do stage1 and stage2 08:42
*** CTtpollard [~tom@82.70.136.246] has joined #baserock08:48
*** rdale [~quassel@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock08:48
*** tlsa [wjJZNA6ndJ@gateway/shell/pepperfish/x-kjfwgfawtfpsxwsh] has joined #baserock08:53
*** rdale [~quassel@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Ping timeout: 272 seconds]08:53
*** rdale [~quassel@82.70.136.246] has joined #baserock08:53
ridgerunnerpetefoth: thanks08:54
petefothridgerunner: shhh!08:54
ridgerunner:-)08:55
*** ssam2 [~ssam2@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock09:06
pedroalvarezjuergbi: I was confused before because I thought that we were already doing that.09:36
juergbipedroalvarez: if so, maybe /tools/bin is in the path before /bin?09:37
pedroalvarezthe problem is that we are also setting /tools/bin in the PATH for stage309:37
pedroalvarezexactly09:37
juergbii would say that /tools/bin should be after /bin in the PATH after stage209:37
juergbihowever, i don't really know baserock-specific aspects of bootstrapping09:38
pedroalvarezthat fix would be easier than checking when to set /tools/bin and when not09:38
pedroalvarezKinnison, juergbi: thanks for your help :)09:38
ssam2the logic to add /tools/bin to the PATH is based on the 'prefix' setting of the dependent chunks10:34
ssam2basically any chunk that sets 'prefix' gets that prefix prepended to the PATH of all future builds10:34
ssam2hopefully we can stop doing that because it's a bit dumb10:35
ssam2it's dumb because /tools/bin is added to the PATH for every chunk, even though it doesn't exist for most of the build process10:35
pedroalvarezi'm not 100% sure it's dumb, but I think I'm sure that adding it before /bin and before /usr/bin is a bad idea10:37
*** sambishop [~sambishop@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock11:31
pedroalvarezhm... I wonder now what happens when cross-bootstrap a system. I believe that in that case the bin --> tools/bin symlink won't be there when building the stage3 chunks. 11:57
pedroalvarezI may be wrong11:57
persiaDoes cross-bootstrap do something especially different than regular bootstrap?  If so, can it be replaced with some nifty trick to build a foreign system that when booted builds a normal system?12:04
richard_mawthe important difference when cross-bootstrapping is that it doesn't use linux-user-chroot for build isolation, and it installs with DESTDIR=/12:10
pedroalvarezright12:11
pedroalvarezI will test my glibc patch tomorrow with cross-bootstrap12:11
persiaAh, yes, that is an essential difference, which I expect changes a few things :)12:12
*** rdale_ [~quassel@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock12:16
*** rdale [~quassel@82.70.136.246] has quit [Read error: Connection reset by peer]12:16
*** fay__ [~fay@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock13:30
*** fay_ [~fay@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Read error: Connection reset by peer]13:30
*** dutch [~william@82.70.136.246] has quit [Ping timeout: 244 seconds]14:01
*** genii [~quassel@ubuntu/member/genii] has joined #baserock14:10
*** CTtpollard [~tom@82.70.136.246] has quit [Ping timeout: 244 seconds]14:14
*** dutch [~william@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock14:23
*** dutch [~william@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Ping timeout: 272 seconds]15:04
*** CTtpollard [~tom@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock15:08
*** dutch [~william@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock15:09
* straycat was quite surprised gentoo folks didn't turn up here after that post on lwn15:53
straycatWe seem to be solving the problem they want to solve.15:54
pedroalvarezstraycat: caon you point me to that post?15:56
straycathttps://lwn.net/Articles/612975/15:57
*** dutch [~william@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Quit]16:00
pedroalvarezSotK: nice patch16:13
wikicatWiki change: Add notes of meeting held at Codethink http://source.baserock.branchable.com/?p=source.git;a=commitdiff;h=82ef0c116:22
radiofreestraycat: might it be useful to have the bot also link to the wiki page?16:26
*** jonathanmaw [~jonathanm@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving]16:26
*** ssam2 [~ssam2@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving]16:32
*** tiagogomes [~tiagogome@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving]16:34
straycatmaybe16:34
pedroalvarezhm.. but if the commit changes more than one page...16:38
pedroalvarezI like how wikica.t works right now16:38
* persia also: in addition to showing the quick change, it encourages folk to use git to deal with the wiki, leading to better logs as the practice is socialised, etc.16:45
*** logger_ [~ubuntu@85.199.252.110] has joined #baserock16:52
pedroalvarezhttp://85.199.252.110/17:01
pedroalvarezha!17:01
pedroalvarezlet's see if it's alive or not17:01
persiapedroalvarez: Nice!17:02
* persia was thinking to look at that this evening, and has been entirely outclassed :)17:02
pedroalvarezirssi + irclog2html + 20 miniutes17:02
pedroalvarezand a cron job17:02
radiofreepedroalvarez: you should update the channel topic to inform people this channel is logged, and the location of the logs17:03
persiaWe really need to get API access to the baserock.org DNS: there's more and more cool services that are hard to discover.17:03
* persia agrees with radiofree17:03
* pedroalvarez has to leave right now17:03
radiofreealso, i believe it is (was?) part of the freenode philosophy to provide a way to have what you say not logged17:03
jmacsradiofree: I think you're right17:03
persiaMost of the channels I use on freenode are logged.17:04
radiofreei don't think there's any issue with logging, as long as you tell people you're logging17:05
*** mSher [~mike@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving]17:05
radiofree" Be sure to provide a way for users to make comments without logging "17:05
radiofreeit's not a hard rule though17:05
jmacsMost channels I use on freenode are logged by people privately, but it's considered rude to publish those logs17:05
jmacsI don't think I could find any written statement to back this up last time I looked though.17:06
radiofreeyes i'm talking about published logs17:06
radiofreehttps://freenode.net/channel_guidelines.shtml17:06
radiofreeat the bottom of the page17:06
jmacsAh, cool17:09
pedroalvarezWhat do I have to do then? 17:11
pedroalvarezSeems like the cronjob is not working anyway 17:12
persiapedroalvarez: Be sure that there is a logs announcement in the /topic17:12
persiaI'm less certain about the way to provide users with a way to make comments without logging: that's something that wasn't done at all in other communities where IRC was a prevalent topic (except perhaps via /query)17:14
radiofreeyeah we can probably pass on that one17:15
straycatmake another channel that's not logged17:15
persia#baserock-chatter ?17:15
straycatI guess, though given the size of this channel at present, it probably doesn't matter.17:16
persiaAgreed.  I mostly wanted logs because sometimes my bot hiccoughs.17:17
*** rdale_ [~quassel@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Ping timeout: 250 seconds]17:47
*** genii [~quassel@ubuntu/member/genii] has quit [Read error: Connection reset by peer]22:56

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!