| *** VVIBIN [~VVIBIN@202.0.77.198] has joined #baserock | 04:35 | |
| VVIBIN | @ssam2/Paul - my mistake,selected a wrong network adapater in VM for bridge.Corrected and baserock eth0 is properly assigned the IP. | 04:38 |
|---|---|---|
| *** VVIBIN [~VVIBIN@202.0.77.198] has quit [Quit: Leaving] | 04:41 | |
| *** ajay-pai [~pajai@202.0.77.198] has joined #baserock | 05:06 | |
| *** ajay-pai [~pajai@202.0.77.198] has quit [Client Quit] | 05:09 | |
| *** aja-pai [~pajai@202.0.77.198] has joined #baserock | 05:09 | |
| aja-pai is now known as ajay-pai | 05:10 | |
| *** hari_ [uid49867@gateway/web/irccloud.com/x-frypqlabymmvoexm] has joined #baserock | 05:14 | |
| *** ajay-pai [~pajai@202.0.77.198] has quit [Remote host closed the connection] | 05:18 | |
| madhu | in the middle of the build, it repeatedly asking for root@localhost password: | 05:46 |
| madhu | sed -i "s|^- morph: .*$|- morph: $1|" "$2" | 05:46 |
| madhu | morph deploy --upgrade "$2" self.HOSTNAME=$(hostname) self.VERSION_LABEL=TEST | 05:46 |
| madhu | root@localhost's password: | 05:46 |
| madhu | root@localhost's password: | 05:46 |
| madhu | root@localhost's password: | 05:46 |
| madhu | 2014-10-28 05:41:18 [systems/devel-system-x86_64-generic.morph][self]Unpacking system for configuration | 05:46 |
| madhu | 2014-10-28 05:44:06 [systems/devel-system-x86_64-generic.morph][self]System unpacked at /src/tmp/deployments/tmpU9kGHj/tmpf8a9FZ | 05:46 |
| madhu | 2014-10-28 05:44:06 [systems/devel-system-x86_64-generic.morph][self]Writing deployment metadata file | 05:46 |
| madhu | 2014-10-28 05:44:07 [systems/devel-system-x86_64-generic.morph][self]Configure system | 05:46 |
| madhu | 2014-10-28 05:44:09 [systems/devel-system-x86_64-generic.morph][self]Processing NETWORK_CONFIG=lo:loopback;eth0:dhcp,hostname=$(hostname) | 05:46 |
| madhu | 2014-10-28 05:44:09 [systems/devel-system-x86_64-generic.morph][self]Writing to device | 05:47 |
| madhu | 2014-10-28 05:44:10 [systems/devel-system-x86_64-generic.morph][self]Finding device that contains "/" | 05:47 |
| madhu | root@localhost's password: | 05:47 |
| madhu | root@localhost's password: | 05:47 |
| madhu | 2014-10-28 05:44:32 [systems/devel-system-x86_64-generic.morph][self]Creating remote mount point | 05:47 |
| madhu | root@localhost's password: | 05:47 |
| madhu | 2014-10-28 05:44:35 [systems/devel-system-x86_64-generic.morph][self]Mounting root disk | 05:47 |
| madhu | root@localhost's password: | 05:47 |
| madhu | 2014-10-28 05:44:36 [systems/devel-system-x86_64-generic.morph][self]Creating /tmp/tmp.maTzyi/systems/TEST | 05:47 |
| madhu | root@localhost's password: | 05:47 |
| madhu | 2014-10-28 05:44:38 [systems/devel-system-x86_64-generic.morph][self]Creating "orig" subvolume | 05:47 |
| madhu | root@localhost's password: | 05:47 |
| madhu | 2014-10-28 05:44:40 [systems/devel-system-x86_64-generic.morph][self]Populating "orig" subvolume | 05:47 |
| madhu | root@localhost's password: | 05:47 |
| *** VVIBIN [~VVIBIN@202.0.77.198] has joined #baserock | 05:59 | |
| *** VVIBIN [~VVIBIN@202.0.77.198] has quit [Quit: Leaving] | 06:05 | |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 06:20 | |
| paulsherwood | madhu: that's in the middle of the deploy | 06:48 |
| paulsherwood | looks like the ssh-copy-id has not happened properly? | 06:49 |
| paulsherwood | madhu: assuming you entered the password (lots of times :-) ) did it finish ok? | 06:50 |
| madhu | paulsherwood, yes, it had completed. | 06:51 |
| paulsherwood | woot! :) | 06:51 |
| paulsherwood | madhu: so let's fix the password thing | 06:52 |
| paulsherwood | in your ssh terminal onto the guest, type ssh-copy-id root@127.0.0.1 | 06:53 |
| paulsherwood | now if you re-run the script/cycle.sh command, the build should skip because it's already cached, and hopefully the deploy will work without all those prompts | 06:54 |
| madhu | paulsherwood, thanks, it worked, I also run the ssh-keygen before ssh-copy-id | 07:09 |
| *** mSher [~mike@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 07:41 | |
| *** hari__ [~hari@202.0.77.198] has joined #baserock | 07:43 | |
| paulsherwood | aha! | 07:45 |
| *** aananth [~caananth@122.165.96.181] has joined #baserock | 07:51 | |
| * petefoth now has plenty of material for the 'Common Errors' page which will exist soon :) | 07:52 | |
| aananth | The step "scripts/cycle.sh systems/devel-system-x86_64-generic.morph clusters/upgrade-devel.morph" takes too long. My machine is currently at "[Build 89/166] [linux-x86-64-generic] Cloning upstream:linux" | 07:52 |
| aananth | We have a training session for day 2 in another 30 mins. | 07:53 |
| aananth | Is there a short cut to copy /src contents from some of my colleagues? | 07:53 |
| aananth | Btw, I am following steps in page http://wiki.baserock.org/guides/no-frills/ | 07:55 |
| *** ssam2 [~ssam2@cpc7-asht7-2-0-cust170.10-1.cable.virginm.net] has joined #baserock | 08:05 | |
| Mode #baserock +v ssam2 by ChanServ | 08:05 | |
| *** franred [~franred@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 08:12 | |
| paulsherwood | aananth: it's ok. we'll discuss that, and shortcuts :) | 08:15 |
| aananth | Ok *Paul* | 08:17 |
| *** dutch [~william@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 08:23 | |
| paulsherwood | aananth: is there some meaning to the asterisks? :-) | 08:29 |
| *** VVIBIN [~VVIBIN@202.0.77.198] has joined #baserock | 08:33 | |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection] | 08:39 | |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 08:43 | |
| pedroalvarez | good morning | 08:45 |
| Kinnison | petefoth: for wiki.baserock.org it's the site admins (me, sam and richard_maw effectively) | 09:11 |
| Kinnison | petefoth: feel free to put together an editors-help page which might cover things like how to make nicely labelled links etc | 09:11 |
| *** jonathanmaw [~jonathanm@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 09:15 | |
| petefoth | Kinnison: I shall do so (and have created a Trello task to remind myself) | 09:17 |
| *** tiagogomes [~tiagogome@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 09:26 | |
| *** LeProvokateur [~LeProvoka@dslb-088-075-010-149.088.075.pools.vodafone-ip.de] has joined #baserock | 09:33 | |
| *** LeProvokateur [~LeProvoka@dslb-088-075-010-149.088.075.pools.vodafone-ip.de] has quit [] | 09:34 | |
| * straycat meows | 09:47 | |
| straycat | 19:22 straycat$ so http://sprunge.us/ABfY probably needs to go in as well | 09:47 |
| straycat | can i do that? | 09:47 |
| straycat | on top of the other thing | 09:47 |
| straycat | it's almost trivial enough not to bother asking | 09:47 |
| persia | To make sure I understand the syntax correctly, that prints the value of original_ref if defined, and otherwise the value of ref? If I've read it correctly, I concur it to be trivial and necessary. | 09:50 |
| straycat | yes | 09:50 |
| persia | Then yes :) | 09:53 |
| ssam2 | might make more sense to print the sha1, so the user knows exactly what commit is being built | 09:55 |
| ssam2 | thus always printing 'ref' | 09:56 |
| straycat | yeah but as you pointed out those shas are temporary | 09:56 |
| straycat | it's not going to be the sha at the repo's head | 09:56 |
| ssam2 | oh, that's true if a temporary build branch was created | 09:56 |
| *** hari__ [~hari@202.0.77.198] has quit [Quit: Ex-Chat] | 09:57 | |
| straycat | we could also always print the sha, would possibly cause some confusion for newcomers though | 10:09 |
| persia | How so? | 10:10 |
| * persia doesn't think that we can meaningfully hide SHAs from users | 10:11 | |
| Kinnison | Several of our users want us to | 10:11 |
| Kinnison | Or at least, paulsherwood does | 10:11 |
| Kinnison | Well, he dislikes full shas anyway | 10:11 |
| Kinnison | He might accept a 7char abbrev-sha | 10:11 |
| persia | It's like git. One can get a short way while pretending everything is short and pretty, and then as soon as one wants to do something interesting, one needs to know the sha. | 10:12 |
| straycat | Actually it wouldn't be confusing at all really, "huh? why isn't morph build HEAD. oh, because i have uncommitted changes" | 10:12 |
| persia | I'm not opposed to abbreviated shortshas, with digit width being the minimum required for unique identity, but that's essentially still SHA. | 10:13 |
| persia | straycat: That would be my interpretation, yes. | 10:13 |
| ssam2 | yeah, I guess once it's clearer about how to get Morph to build the thing you have checked out locally, showing the ref name instead of its sha1 shouldn't be a problem | 10:19 |
| paulsherwood | i don't want to hide shas from uses. but if tags are available, users will want to use them | 10:20 |
| paulsherwood | and short shas are easier to grok than long ones | 10:20 |
| franred | git users would be used to use long and short shas, with long shas you have all the info when short ones doesn't so I don't know why we want to use the short version when on the screen/log are going to use few characters more | 10:27 |
| straycat | It's also worth pointing out that the build ref for a temporary build branch is much longer than a sha, and that's what we print currently | 10:30 |
| straycat | (if we're building from a temporary branch) | 10:31 |
| paulsherwood | yes - this information is of no help to a user, afaict? | 10:31 |
| * paulsherwood would like temporary build branches scrapped :) | 10:32 | |
| straycat | I don't know really, I've never paid much attention to it before now. It does tell them the temporary branch which might be more use than a sha. | 10:33 |
| paulsherwood | if my build works, i don't need that info. if my build doesn't work, i need to get to the failed staging area... in which case i still don't need that info afaict | 10:34 |
| straycat | paulsherwood, Me too really, but I have to concede that they are useful both for distbuild and for building from your current working tree. | 10:34 |
| * paulsherwood nods, grumpily :) | 10:36 | |
| straycat | I'll just merge the first patch for now, we can change the output later if we're not happy with it | 10:37 |
| persia | The problem with the temporary build branches is that is causes all sorts of messiness in the parentage and git history, causing git visualisation tools to give confusing output. | 10:38 |
| ssam2 | i'm gonna head to the office, see you in 20 mins | 10:40 |
| *** ssam2 [~ssam2@cpc7-asht7-2-0-cust170.10-1.cable.virginm.net] has quit [Quit: Leaving] | 10:40 | |
| Kinnison | the temporary branches are unlikely to show up in history, surely? | 10:41 |
| petefoth | So is the Baserock Trello is set up to automatically archive items once they're moved to the 'Done' lane? | 10:42 |
| Kinnison | I've never known trello to do that, can it? | 10:43 |
| paulsherwood | Baserock Trello? pardon? | 10:43 |
| Kinnison | Pedro set up a trello for tracking his upstream work | 10:44 |
| Kinnison | others have chosen to join him on it | 10:44 |
| petefoth | paulsherwood: https://trello.com/b/rShl98Nc/baserock-things which some of us are using in the absence of other task trackers | 10:44 |
| petefoth | for people who value the Kanban process :) | 10:45 |
| paulsherwood | over open source software :) | 10:45 |
| robtaylor | maybe using a free software kanban implmenation would be good? ;) | 10:45 |
| persia | Kinnison: The temporary branches always show up in history, sadly, unless you restrict your history view to that of a single branch. They exist as branches at commits with timestamps, so wider views show them (confusingly), and some tooling that tries to determine true ancestry through blob analysis gets truly confused. | 10:46 |
| Kinnison | they're supposed to be removed by the tooling when builds are completed | 10:46 |
| paulsherwood | robtaylor: i believe we even developed one :) | 10:47 |
| persia | That has not been my experience, although I haven't tried to use git visualisation tools on definitions in some time (as the output was less useful than had been my experience with other projects). | 10:47 |
| robtaylor | paulsherwood: indeed | 10:47 |
| * persia continues to believe that kanban-style systems are inapplicable in environments with elastic resources | 10:47 | |
| robtaylor | paulsherwood: well you and rdale did ;) | 10:47 |
| petefoth | paulsherwood: robtaylor: we've been talking about gettign a public facing task-tracker in place, but we got distracted by trying to make storyboard work *in baserock*. In the meantime, aomething is better than nothing. I'm more than happy to use an open source tool if / when there is one in place | 10:47 |
| richard_maw | Kinnison: they're removed from the git server, they still exist locally | 10:48 |
| robtaylor | https://taiga.io/ also looks cool | 10:48 |
| Kinnison | richard_maw: aah | 10:48 |
| persia | petefoth: You should chime in on ssam2's thread about the potential for using non-baserock-hosted tooling as baserock infrastructure. Deploying something isn't the hard part, it's reaching consensus on the amount of dogfood to eat. | 10:48 |
| petefoth | persia: I'll do that. From my perspective the problem is that the baserock community spend a long time in dicussions ("We could use tool foo", "bar looks really interesting", "wouldn't it be good to get qux working in Baserock") but no *decisions* get made. I would rather spend my time getting on with doing stuff (that I hope will be useful). Contributing to discussions gets old quite quickly when nothing seems to happen as a result. | 10:52 |
| rdale | the lorried version of the qt3d repo seems to be behind the upstream one and is missing a lot of branches | 10:53 |
| persia | petefoth: My understanding of the decision process is that of loose consensus, and a two-affirmation requirement. If ssam2 proposes it, and you and I agree, then we should be good unless someone disagrees, which forces discussion. | 10:54 |
| petefoth | The consensus seems to be to use StoryBoard, but it isn'tin place yet because the prosepective hosting company aren't yet ready. I'll happily migrate to StroyBoard once the public instance is up and running. In the nmean time Trello works for me (and for pedroalvarez and richard_maw ) | 10:54 |
| pedroalvarez | just to be clear, this trello board is not part of the baserock infra. | 10:55 |
| persia | I thought the issue was whether it was hosted on Baserock alone, regardless of anything else. | 10:55 |
| petefoth | persia: see above. I think we have consensus to use StoryBoard but it isn;t in place yet (because reasons) | 10:55 |
| persia | petefoth: So, let's fix that. Give a +1 to the idea that we can use non-baserock-hosted stuff, and we can do that. | 10:56 |
| persia | And when we have the ability to self-host, we can migrate. | 10:57 |
| petefoth | persia: done. Let we know when it's working :) | 10:58 |
| persia | petefoth: Thanks for helping the decision process. | 10:58 |
| franred | rdale, are you sure that you lorried the repo that you wanted? | 10:59 |
| franred | rdale, I've cloned the repo and it shows the same tag and branches as g.b.o trove has | 11:00 |
| franred | s/tag/tags/ | 11:00 |
| madhu | how do I include only selected chunks from a stratum | 11:01 |
| Kinnison | madhu: You will need to look at the rules for stratum splitting. richard_maw may be able to give you hints, otherwise look at the minimal system and work backwards from there. | 11:02 |
| rdale | franred: ok i'll double check qt3d again | 11:03 |
| richard_maw | madhu: in the strata list in the system definition (systems/minimal-system-x86_64-generic.morph is an example) you can list the artifacts to include | 11:03 |
| richard_maw | by default every stratum has two variants: -runtime and -devel, -devel includes things like the headers, while runtime has the executables, libraries and default config files | 11:04 |
| richard_maw | you can define other variants for a stratum by adding a field called products to the stratum morphology (see strata/build-essential.morph for an example) | 11:05 |
| richard_maw | you can specify regular expressions to specify patterns for which parts of the chunks to include | 11:06 |
| richard_maw | or, in the list of chunks, you can add an artifacts key (see `name: eglibc` for an example) | 11:06 |
| richard_maw | which lets you override the regular expression match per chunk, and say that these specific splits go into a different stratum variant | 11:07 |
| richard_maw | the default rules for splitting chunks up produce -bins -libs -misc -devel -doc and -locale, but you can define your own | 11:08 |
| rdale | franred: i've cloned the upstreamed qt3d again and now it matches the lorried one. I'm not sure was going on - the old upstream copy of qt3d suddenly became an empty directory | 11:08 |
| richard_maw | you can look at strata/build-essential/eglibc.morph to see how we've overridden the default ways to split a chunk's build up, and we've added an -nss split | 11:09 |
| *** ssam2 [~ssam2@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 11:09 | |
| Mode #baserock +v ssam2 by ChanServ | 11:09 | |
| madhu | richard_maw, Kinnison : thank you, I will look at them | 11:10 |
| franred | rdale, no idea what could happen, if everything is ok now is good :) | 11:11 |
| Kinnison | ssam2: Are the issues with your trove fixed now? | 11:12 |
| ssam2 | Kinnison: my primary goal was to set up a mirror of git.baserock.org | 11:17 |
| ssam2 | so I solved that my having my mirror clone from git.baserock.org rather than baserock-clone | 11:17 |
| Kinnison | ssam2: :-) | 11:17 |
| Kinnison | ssam2: is baserock-clone fixed? | 11:17 |
| ssam2 | not yet | 11:17 |
| Kinnison | okay | 11:18 |
| ssam2 | i'll look at it later | 11:18 |
| Kinnison | let me know if you need more help | 11:18 |
| * pedroalvarez builds and deploys a base system with glibc :) | 11:30 | |
| pedroalvarez | and it works | 11:31 |
| straycat | cool :*) | 11:31 |
| ssam2 | I am incredibly impressed ;) | 11:31 |
| ssam2 | speaking as someone who got part way through updating just to eglibc 2.16, and gave up due to lack of time | 11:32 |
| robtaylor | nice! | 11:32 |
| pedroalvarez | to be honest, this is being easy | 11:32 |
| *** zoli__ [~zoli_@0x5e91887a.adsl.cybercity.dk] has joined #baserock | 11:36 | |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Ping timeout: 272 seconds] | 11:37 | |
| pedroalvarez | let's build a devel system now | 11:39 |
| aananth | sam: how to check if the cycle.sh is really running. In my machine it stays at "[Build 89/166] [linux-x86-64-generic] Cloning upstream:linux" for the last 3 hours. | 11:42 |
| aananth | I stopped and re-ran, here are the messages 2014-10-28 09:59:28 Deciding on task order | 11:43 |
| aananth | 2014-10-28 10:03:40 Caching git repository upstream:gperf | 11:43 |
| aananth | 2014-10-28 10:04:59 Caching git repository upstream:linux | 11:43 |
| aananth | No updates for the last 1 hr | 11:43 |
| ssam2 | aananth: ok, seems that it's taking a /really/ long time to download linux.git | 11:44 |
| ssam2 | it doesn't give progress info for the download, sadly (it'd be better if Morph did that) | 11:45 |
| aananth | ~ 40 KiB/s is my average speed. | 11:45 |
| aananth | Then, It is good to know that it is working! | 11:45 |
| aananth | Let me connect to Visteon network and re-try. | 11:46 |
| ssam2 | if you want to double check, you can find the process ID of 'morph' with: 'pgrep -fl morph' | 11:47 |
| ssam2 | (do that in a separate session inside the same VM) | 11:48 |
| madhu | aananth, you need to configure proxy if you connect to Visteon | 11:48 |
| rdale | i'm getting an 'internal compiler error' when i build mesa in a kvm with '-m 2048' - would increasing the memory help? | 11:49 |
| ssam2 | rdale: is there any C++ in Mesa? if so, perhaps yes | 11:49 |
| rdale | i don't know - i'll try it with some more memory - i assume it isn't an actual bug in the compiler | 11:50 |
| ssam2 | Mesa has built successfully many times, so I doubt it | 11:54 |
| madhu | does baserock generate packages, like rpms, dkgs which we build and install later. | 11:55 |
| ssam2 | madhu: Baserock only generates whole systems | 11:56 |
| aananth | madhu: I hope the environment variable "http_proxy" and "https_proxy" setting should be fine, right? | 11:56 |
| ssam2 | madhu: the idea of that is that allowing systems to be customised after they're build introduces a lot of complexity | 11:57 |
| ssam2 | it makes sense for a developer desktop environment | 11:57 |
| ssam2 | but it doesn't make as much sense for embedded or cloud systems | 11:57 |
| madhu | yes, we also setup corkscrew as we do in Ubuntu, | 11:57 |
| madhu | aananth, we need corkscrew for accessing git, there was another option with "nc | 11:58 |
| madhu | but I didnt know how to configure nc with username/password, | 11:59 |
| pedroalvarez | is g.b.o down? :/ | 12:00 |
| madhu | corkscrew is not default in the baserock, so I copied from my machine | 12:00 |
| ssam2 | pedroalvarez: I can reach cgit on git.baserock.org | 12:00 |
| ssam2 | and git:// works | 12:01 |
| ssam2 | so I think it might be you :) | 12:01 |
| pedroalvarez | then is my connection | 12:01 |
| *** aananth [~caananth@122.165.96.181] has quit [Ping timeout: 256 seconds] | 12:01 | |
| pedroalvarez | odd | 12:02 |
| madhu | ssam2, incase if I want to stack trace gstreamer for a crash, I can install the gstreamer dbg package to the system and proceed, I dont need to deploy entire system. | 12:10 |
| ssam2 | madhu: the way we do this right now, is that we don't strip any binaries by default | 12:11 |
| ssam2 | so you always have the debug info available, unless you explicitly strip things (as you'd probably want to for a final release :) | 12:11 |
| ssam2 | there are ways to improve this though, I think separating out the debug info would be quite cool | 12:12 |
| ssam2 | and would reduce the download size of artifacts | 12:12 |
| madhu | when we share the system image locally here, we give only production image which will not contain debug symbol | 12:13 |
| radiofree | madhu: submit a patch to add corkscrew to baserock :) | 12:13 |
| radiofree | maybe in foundation | 12:13 |
| ssam2 | yes, that'd be useful, although maybe it should go in 'tools' | 12:14 |
| madhu | radiofree, thanks for the suggestion, :-), does baserock maintain corkscrew sources | 12:17 |
| madhu | brb | 12:17 |
| ssam2 | madhu: not right now. there's kind of a two step process for adding new stuff to the Baserock reference systems: 1st you send a patch to add a 'lorry' file to import the sources, 2nd you send a patch to actually built the code | 12:18 |
| *** aananth [~caananth@74.112.167.117] has joined #baserock | 12:19 | |
| ssam2 | you can start step 2 without waiting for step 1 if there's already a Git mirror you can use somewhere | 12:19 |
| ssam2 | but http://www.agroman.net/corkscrew/ seems to only link to a tarball | 12:20 |
| *** paulk-aldrin [~paulk@armstrong.paulk.fr] has joined #baserock | 12:22 | |
| franred | hi paulk-aldrin | 12:24 |
| paulsherwood | ssam2: +1 to lorry that? | 12:24 |
| ssam2 | I was wondering if madhu would want to have a go at writing a .lorry file | 12:25 |
| paulsherwood | :) | 12:25 |
| paulk-aldrin | hi | 12:25 |
| ssam2 | maybe it's easier if we add a lorry so he can try adding to his system | 12:25 |
| paulsherwood | ssam2: alternatively i can put it on github | 12:25 |
| ssam2 | I don't know what's best. Corkscrew seems small, and useful | 12:26 |
| paulsherwood | madhu: do you want to submit a lorry file patch? | 12:26 |
| paulsherwood | or we can do it | 12:26 |
| paulsherwood | normally i just copy one from http://git.baserock.org/cgi-bin/cgit.cgi/baserock/local-config/lorries.git/tree/open-source-lorries | 12:27 |
| franred | paulk-aldrin, we have built systems for the jetson tegra TK1, I think paulsherwood and radiofree can give you more information about it | 12:27 |
| ssam2 | http://git.baserock.org/cgi-bin/cgit.cgi/baserock/local-config/lorries.git/tree/open-source-lorries/itzam-tarball.lorry is a tarball lorry that could be adapted for corkscrew | 12:28 |
| paulk-aldrin | nice | 12:34 |
| paulk-aldrin | well, I'm a bit interested in the graphics part | 12:34 |
| paulk-aldrin | I see you have wayland and mesa running | 12:34 |
| paulk-aldrin | that's nice | 12:34 |
| paulsherwood | it's a work in progress :) | 12:35 |
| paulk-aldrin | how hard would it be to have X11 accelerated with glamor? | 12:35 |
| paulk-aldrin | and also, what's the display interface on mainline? DRM/KMS, fb? | 12:35 |
| paulsherwood | that's a question for radiofree | 12:35 |
| paulsherwood | and possibly jjardon | 12:36 |
| paulk-aldrin | ok | 12:36 |
| ssam2 | I know kms works, I've seen kmscube running :) | 12:38 |
| ssam2 | I believe we've done Weston using the fbdev backend, that might have been unaccelerated though | 12:39 |
| ssam2 | I think we've also got accelerated Weston running with DRM, although it took a bit of special patching | 12:40 |
| ssam2 | mainly because the Tegra exposes 2 devices for rendering, and nothing really expects that at the moment | 12:40 |
| jjardon | paulk-aldrin: not difficult, we need to upgrade the xserver we currently build though | 12:45 |
| paulk-aldrin | alright | 12:49 |
| madhu | paulk-aldrin, ssam2: I am ok if you add corkscrew to baserock, | 12:56 |
| petefoth | ssam2: Kinnison: richard_maw: will one of you have time to enable videos on http://wiki.baserock.org/video-archive/ (please)? | 12:56 |
| ssam2 | madhu: if I get it into git.baserock.org, could you try adding it to your Baserock system? | 12:57 |
| ssam2 | petefoth: I'll have a look now | 12:57 |
| petefoth | ssam2: ta! | 12:57 |
| madhu | sure | 12:58 |
| jjardon | paulk-aldrin: I actually though about doing it but Ive been more focused in wayland stuff lately. It could be a fun Sunday project though ;) | 12:58 |
| paulk-aldrin | nice | 12:59 |
| paulk-aldrin | well I sure will look into it when I get a board | 12:59 |
| pedroalvarez | I want to update the lsof tarball lorry, and I found 2 different urls: | 13:01 |
| pedroalvarez | http://ftp.u-strasbg.fr/securite/lsof/lsof_4.88.tar.gz | 13:01 |
| pedroalvarez | ftp://sunsite.ualberta.ca/pub/Mirror/lsof/lsof_4.88.tar.gz | 13:01 |
| ssam2 | petefoth: I disabled 'htmlscrubber' on 'video-archive' and it seems to embed the videos now | 13:01 |
| ssam2 | what do we currently use ? | 13:02 |
| petefoth | ssam2: thanks | 13:02 |
| ssam2 | pedroalvarez: seems we curently use an ftp mirror from purdue.edu | 13:03 |
| ssam2 | which is out of date | 13:03 |
| madhu | I would like to know how do you generate SDK in baserock, we provide SDK to suppliers for building their binary deliveries , | 13:04 |
| pedroalvarez | ssam2: yeah, they don't have the latest one | 13:04 |
| ssam2 | pedroalvarez: the fedora spec file point sto http://people.freebsd.org/~abe/ but not sure if that's up to date | 13:04 |
| ssam2 | madhu: we do have a way of doing that | 13:04 |
| madhu | can you share me the link | 13:05 |
| ssam2 | there's an example of using the mechanism here: http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/definitions.git/tree/clusters/sdk-example-cluster.morph | 13:05 |
| ssam2 | I can guide you through how it works, if you like | 13:06 |
| ssam2 | the key is that you have two components: (1) the system for your device, and (2) the cross-compile SDK | 13:08 |
| ssam2 | the deployment instructions I linked to show you how to create (2), and make it contain (1) so you have libraries and headers available | 13:09 |
| ssam2 | in this example, 'devel-system-armv7lhf-highbank.morph' is (1) | 13:10 |
| ssam2 | wait, worry | 13:10 |
| ssam2 | 'devel-system-armv7lhf-highbank.morph' is (2) :) | 13:10 |
| ssam2 | no, 'devel-system-armv7lhf-highbank.morph' is (1). I'm confusing myself by using numbers ! | 13:11 |
| ssam2 | sorry | 13:11 |
| petefoth | I'vce just stumbkled across the 'Baserock Team' page at http://wiki.baserock.org/team/. Is this something we want to keep and maintain? | 13:11 |
| ssam2 | I see no harm in it, and maybe some value. Perhaps those who aren't listed can add themselves | 13:12 |
| pedroalvarez | after updating lsof to latest version, the devel system builds with glibc :) | 13:17 |
| ssam2 | yay | 13:18 |
| pedroalvarez | time to deploy it and check that lsof works :P | 13:19 |
| *** VVIBIN [~VVIBIN@202.0.77.198] has quit [Ping timeout: 260 seconds] | 13:19 | |
| ssam2 | madhu: if you're reading baserock-dev@baserock.org, you'll see that I've just sent a patch to import 'corkscrew' into git.baserock.org | 13:20 |
| pedroalvarez | what is corkscrew for? | 13:21 |
| ssam2 | it allows tunneling SSH through HTTP connections. Handy if you happen to, for example, be behind a mean corporate proxy | 13:21 |
| pedroalvarez | aha | 13:23 |
| *** Krin [~mikesmith@host86-190-220-106.range86-190.btcentralplus.com] has joined #baserock | 13:24 | |
| *** hari__ [~hari@202.0.77.198] has joined #baserock | 13:24 | |
| madhu | ssam2, looks good to me, | 13:25 |
| pedroalvarez | ssam2: also to me: +1 | 13:29 |
| pedroalvarez | great, lsof work in the devel system after upgrading it to the latest :) | 13:32 |
| madhu | ssam2, can 't I build a sdk for devel-system-x86_64-generic.morph right now | 13:33 |
| Krin | hey all, mind if i ask for an addition to the lorry system? needing the ant build environment, one of the requirements for zookeeper that i'm still working on http://pastebin.com/DgPfG2b2 | 13:35 |
| ssam2 | madu: yes, you could do that | 13:35 |
| Krin | i'd try myself but i'm scared on using my "this thing breaks now" magic >.< | 13:35 |
| Kinnison | Krin: what's the reasoning for calling it ant_build_environment rather than just ant ? | 13:35 |
| ssam2 | madhu: it'll take some changes though, because the current SDK example is designed for cross-compiling to ARM | 13:36 |
| Krin | i have come across a few things called ant, some in windows, some in linux, i just wanted to define it incase anyone was looking for the other things | 13:36 |
| ssam2 | cross-compiling from x86_64 to x86_64 is not something we needed to do yet :) | 13:36 |
| Krin | one example is ant is also a populat interface building tool for games. | 13:37 |
| ssam2 | Krin: looks good, I wonder if it'd be better to call it 'java/ant' or some such | 13:37 |
| Krin | would the slash cause issues? | 13:37 |
| ssam2 | Krin: we already have ruby-gems/* and python-packages/*, so it'd kinda follow a pattern | 13:37 |
| ssam2 | no, slashes are cool! | 13:37 |
| * Kinnison thinks java/ant would probably be better | 13:37 | |
| Kinnison | +1 for whoever merges it, with the rename to java/ant | 13:38 |
| * Krin agrees now that he knows the above info | 13:38 | |
| ssam2 | ok, I'll merge it and rename it | 13:38 |
| * Krin cheers | 13:39 | |
| ssam2 | Krin: done, it should appear in this list in due course: http://git.baserock.org/lc-status.html | 13:41 |
| Kinnison | ssam2: I responded to your corkscrew posting btw, not sure if you've seen | 13:42 |
| ssam2 | I did, thanks | 13:42 |
| ssam2 | I didn't realise those fields weren't needed | 13:43 |
| ssam2 | I shall merge once it gets another +1 | 13:43 |
| Kinnison | cool | 13:43 |
| * pedroalvarez sends the +1 via email this time | 13:45 | |
| *** hari__ [~hari@202.0.77.198] has quit [Quit: Ex-Chat] | 13:48 | |
| ssam2 | oh, I missed that one before. thanks pedro! | 13:49 |
| ssam2 | oh, seems paul merged it already | 13:49 |
| ssam2 | and fixed it. thanks! | 13:50 |
| ssam2 | I am going to eat some lunch, will be back in half an hour or so | 13:50 |
| paulsherwood | ssam2: i thought my merge had failed | 13:53 |
| paulsherwood | i'll reply to the list | 13:53 |
| paulsherwood | http://git.baserock.org/cgi-bin/cgit.cgi/delta/corkscrew.git/ | 13:54 |
| *** genii [~quassel@ubuntu/member/genii] has joined #baserock | 14:00 | |
| madhu | ssam2, not just for cross compiling, but to be sure you do not link to different version of the library when you build outside. | 14:12 |
| *** tiagogomes_ [~tiagogome@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 14:13 | |
| *** tiagogomes [~tiagogome@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Read error: Connection reset by peer] | 14:13 | |
| *** thecorconian [~thecorcon@eccvpn1.ford.com] has joined #baserock | 14:21 | |
| *** thecorconian [~thecorcon@eccvpn1.ford.com] has quit [] | 14:21 | |
| ssam2 | madhu: with Morph everything is built in an isolated chroot (except for the bootstrap part) | 14:26 |
| ssam2 | so anyone building stuff with Morph doesn't have to worry about linking to wrong versions of libraries, it's just not possible | 14:27 |
| ssam2 | but if you have developers who aren't using Morph then yeah, I can see why a 'fake' cross toolchain might be useful | 14:27 |
| ssam2 | and an SDK definitely | 14:27 |
| ssam2 | I guess we can demo it for you later in the week | 14:27 |
| *** CTtpollard [~tom@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 14:55 | |
| CTtpollard | Would anyone have a problem with Baserock videos being hosted on Youtube, instead of or as well as Vimeo? | 14:56 |
| ssam2 | I think it'd be cool having them on both, as long as they were kept in sync | 14:57 |
| ssam2 | as in, there wasn't a random selection of different versions of videos on both sites, but not all of them on either site :) | 14:58 |
| pedroalvarez | I've been uploading videos to youtube in the past | 14:58 |
| petefoth | I really think we should pick one, so that we don't have to maintain two sets of links | 14:58 |
| rdale | i watched the two latest, and i thought the sound was fine and didn't notice paul's cold | 14:59 |
| petefoth | I don't mind whether it's vimeo, youtube or anywhere else | 14:59 |
| CTtpollard | I thought about this when the email went around the mailing list for Baserock in the community | 14:59 |
| petefoth | Are there any video storring/streaming site that ae particularly FOSS friendly? | 15:00 |
| CTtpollard | I'm not bringing this up in relation to which is the better option technically, more in relation to reaching a larger viewerbase | 15:00 |
| paulsherwood | iirc i posted the first video, and my weapon of choice at the time was vimeo | 15:01 |
| petefoth | CTtpollard: I suspec that most video viewers will come to the videos via the wiki. I'm not sure any will stumble on baserock in either vimeo or youtube | 15:01 |
| paulsherwood | when the possibility of moving came up, it struck me that we'd lose the 'proof' that the original trebuchet demo happened in 2011 | 15:01 |
| CTtpollard | I tend to agree petefoth, but I have been amazed in the past what I have stumbled on via recommended videos / channels across various parts of Youtube | 15:02 |
| petefoth | re: the 'error: Nonexistent host networking interface, name '' (VERR_INTERNAL_ERROR) reported by one of the trainees: can anyone tell me when it was encountered? (my backscroll buffer isn't big enough) | 15:02 |
| paulsherwood | after creating the vm, on attempting to boot it | 15:02 |
| CTtpollard | ssam2 is crucial though I think, in synce or not all between 2 platforms to avoid confusion | 15:04 |
| petefoth | paulsherwood: ta! | 15:04 |
| petefoth | CTtpollard: +1. And as, if they are on twp systems they *will* get out of sync, we shoulkd only host them on one system (IMHO) | 15:06 |
| paulsherwood | well, if someone wants to garden this, i'm ok to go along with it | 15:07 |
| paulsherwood | i'd recommend youtube since maybe autocaption will work | 15:07 |
| * Kinnison will pass off the Baserock channel control to someone the project nominates | 15:08 | |
| Kinnison | Actually, Richard Maw has access to it | 15:16 |
| straycat | ssam2, In your distbuild fix, I take it the original_ref field is optional purely to save us having to redeploy existing distbuild nodes? | 15:18 |
| ssam2 | yes | 15:19 |
| * richard_maw checks if he remembers the password | 15:26 | |
| petefoth | I have made quite a lot of edits in the wiki pages in tha last couple of days. I have prrof-read them, and will do do again in the morning, but I *will* have missed some errors. So if anyone fancies doing some checking, that would be great. The pages are videos, video-archive, common-errors, index, guides/no-frills, guides/chroot, guides/vm-setup | 15:27 |
| richard_maw | Kinnison: nope, forgotten | 15:27 |
| * richard_maw checks mail | 15:27 | |
| Kinnison | richard_maw: see | 15:28 |
| Kinnison | Subject: Baserock YouTube channel | 15:28 |
| Kinnison | Message-ID: <20130725164542.GX30070@somnambulist.local> | 15:28 |
| pedroalvarez | ssam2: | 15:30 |
| pedroalvarez | oops, sorry | 15:30 |
| petefoth | if some does move the videos from vimeo to youtube, please keep me informaed so I can updated the links in the wiki (unless you feel moved to update the wike yourself)? | 15:35 |
| petefoth | s/some/someone/ | 15:35 |
| aananth | paul | sam: I have flashed Baserock Development Image into Jetson. As per the page, I need to follow "Quick start". But the instructions sounds like "they are for VirtualBox". I hope I can proceed? | 15:43 |
| paulsherwood | aananth: which page have you followed so far? | 15:44 |
| *** zoli__ [~zoli_@0x5e91887a.adsl.cybercity.dk] has quit [Remote host closed the connection] | 15:44 | |
| aananth | paul: I followed "http://wiki.baserock.org/guides/baserock-jetson/" for Jetson | 15:45 |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 15:45 | |
| paulsherwood | ok, so now i think you have a baserock machine on jetson | 15:45 |
| aananth | Yes, I am able to log in and change password for root. | 15:46 |
| aananth | Messages from Jetson: Welcome to Baserock! | 15:46 |
| aananth | Baserock is a system and toolset for developing Linux systems. | 15:46 |
| aananth | For more information on working with Baserock, see: | 15:46 |
| aananth | * http://wiki.baserock.org/ | 15:46 |
| aananth | * http://wiki.baserock.org/quick-start/ | 15:46 |
| aananth | * http://wiki.baserock.org/devel-with/ | 15:46 |
| aananth | Thank you for trying Baserock. | 15:46 |
| aananth | baserock login: root | 15:46 |
| aananth | Password: | 15:46 |
| aananth | No mail. | 15:46 |
| paulsherwood | ok, so on quikstart it says vm, but the same instructions should now work on the machine you are in | 15:46 |
| paulsherwood | even though it's not a vm | 15:46 |
| aananth | How about hard disk? Instructions differ? | 15:46 |
| paulsherwood | we should think how to fix this to make it clearer | 15:47 |
| paulsherwood | good question | 15:47 |
| paulsherwood | i don't think they do... | 15:47 |
| paulsherwood | please hold | 15:47 |
| paulsherwood | can you confirm do you have dev/sdb ? | 15:48 |
| paulsherwood | aananth: ^^ | 15:49 |
| aananth | paul: I have not yet bought SATA hard disk | 15:49 |
| aananth | :) | 15:49 |
| paulsherwood | ah :) | 15:49 |
| paulsherwood | then you don't | 15:49 |
| aananth | I hope I will do it tomorrow. | 15:49 |
| paulsherwood | ok | 15:49 |
| paulsherwood | you could maybe use an ssd card? | 15:49 |
| paulsherwood | radiofree: would the above be possible/ | 15:50 |
| aananth | paul: Can I proceed without hard disk? | 15:50 |
| paulsherwood | not really... you don't have enough storage to do morph build. | 15:50 |
| paulsherwood | df -h | 15:50 |
| paulsherwood | will show you how much you have available | 15:50 |
| aananth | Ok. Thats clear. | 15:51 |
| aananth | ~ # df -h | 15:51 |
| aananth | Filesystem Size Used Available Use% Mounted on | 15:51 |
| aananth | /dev/mmcblk0p1 4.0G 1.1G 2.6G 30% / | 15:51 |
| aananth | devtmpfs 849.4M 0 849.4M 0% /dev | 15:51 |
| aananth | tmpfs 873.1M 0 873.1M 0% /dev/shm | 15:51 |
| aananth | tmpfs 873.1M 268.0K 872.9M 0% /run | 15:51 |
| aananth | tmpfs 873.1M 0 873.1M 0% /sys/fs/cgroup | 15:51 |
| aananth | tmpfs 873.1M 0 873.1M 0% /tmp | 15:51 |
| aananth | /dev/mmcblk0p1 4.0G 1.1G 2.6G 30% /home | 15:51 |
| paulsherwood | could maybe use an ssd to act as /src | 15:51 |
| aananth | /dev/mmcblk0p1 4.0G 1.1G 2.6G 30% /root | 15:51 |
| aananth | /dev/mmcblk0p1 4.0G 1.1G 2.6G 30% /srv | 15:51 |
| aananth | /dev/mmcblk0p1 4.0G 1.1G 2.6G 30% /opt | 15:51 |
| aananth | /dev/mmcblk0p1 4.0G 1.1G 2.6G 30% /var | 15:51 |
| aananth | ~ # | 15:51 |
| paulsherwood | if you have one that's 30GB or more | 15:51 |
| aananth | Ok, I will do that tomorrow. | 15:51 |
| paulsherwood | or SD card | 15:52 |
| * paulsherwood keeps getting confused between the different storage things | 15:52 | |
| aananth | Ok, I will check for SD card also. | 15:52 |
| ssam2 | aananth: I'm glad you're getting somewhere, at least! | 15:52 |
| ssam2 | when you're pasting more than a couple of lines of text, could you use a pastebin service like http://pastebin.com in future ? | 15:53 |
| ssam2 | otherwise the IRC window gets very full of text! | 15:53 |
| aananth | sam: Yeah, but the VirtualBox is still on stage 106/166 | 15:53 |
| aananth | Ok | 15:53 |
| ssam2 | aananth: oh, probably slow download speeds still ... | 15:53 |
| paulsherwood | even after switching to 14.40.1 ? | 15:53 |
| aananth | Sorry for incorrect response. I wll check pastebin. | 15:54 |
| ssam2 | is it possible to leave it going overnight ? | 15:54 |
| ssam2 | hopefully it'll be done in the morning! | 15:54 |
| aananth | Yeah, but it is better than the earlier network, which claims 8 Mbps. | 15:54 |
| aananth | Yeah, I will do it tonight | 15:54 |
| ssam2 | so 1 MB/s ... it'd certainly be finished by now if it was really 1 MB/sec ! | 15:55 |
| ssam2 | about the Jetson, external hard disk will definitely be best if you can find one | 15:55 |
| ssam2 | SD card will be quite slow | 15:55 |
| aananth | sam: Ok, thanks. | 15:55 |
| *** aananth [~caananth@74.112.167.117] has quit ["Leaving"] | 15:58 | |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection] | 16:03 | |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 16:04 | |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection] | 16:08 | |
| *** zoli_ [~zoli_@0x5e91887a.adsl.cybercity.dk] has joined #baserock | 16:09 | |
| *** zoli_ [~zoli_@0x5e91887a.adsl.cybercity.dk] has quit [Changing host] | 16:09 | |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 16:09 | |
| jjardon | Hi, Im using "morph edit gtk+" but baserock is picking the gtk+ from the gtk2 stratum instead gtk3 one. How can I change this? | 16:23 |
| richard_maw | rename the chunks in the gtk2 and gtk3 strata to be gtk2+ and gtk3+ | 16:24 |
| jjardon | oh nevermind, I should use morph edit gtk3 instead | 16:24 |
| jjardon | is there any tool in baserock to check if any chuck is defined in 2 different places? | 16:26 |
| straycat | git grep 'name: foo' | 16:27 |
| jjardon | straycat: that would work if you know "foo" | 16:37 |
| paulsherwood | jjardon: 'git grep '^name: ' | sort | uniq -c' | 16:38 |
| jjardon | paulsherwood: thanks! but I think -d is what we want here? | 16:40 |
| * paulsherwood no entiendo | 16:41 | |
| jjardon | oh -c shows the number of occurrences | 16:42 |
| Krin | hey all, is there a way that i can make sure that a command for building is executed in the right spot while making a new baserock build? i'v noticed other cunks have src && <command> in them, but i'm not getting any joy with that one | 16:42 |
| richard_maw | Krin: that's how it usually works, can you pastebin the morphology you're using? | 16:42 |
| paulsherwood | jjardon: yup:) | 16:42 |
| jjardon | paulsherwood: btw in this case it would be "paulsherwood no entiende" ;) | 16:43 |
| * paulsherwood was futzing with unix command line while jjardon was still in daipers | 16:43 | |
| jjardon | still, I think morph should warn about duplicate chunks in your system | 16:44 |
| paulsherwood | agreed | 16:44 |
| straycat | Is there any reason we lorry setuptools from http://svn.python.org/projects/sandbox/ ? Current upstream happens in a git repo on bitbucket. | 16:45 |
| straycat | jjardon, that would be nice I think | 16:46 |
| rdale | if i get a 'Error reading config file' from u-boot is there a way of finding out what is wrong with the extlinux.conf file? | 16:46 |
| franred | straycat there are a bunch of python packages which changes the place from where we should track like pyyaml or libyaml... feel free to add a new lorry file for yours | 16:46 |
| franred | although we may want to label some repos as old at some point | 16:48 |
| straycat | franred, I'm not sure I follow, you just mean they lorry from old mirrors? | 16:48 |
| franred | straycat, yes | 16:48 |
| Krin | richard_maw, http://pastebin.com/vacHmFEE i'v had to write that one in manualy, cant copy out of the VM today for some reason >.< | 16:49 |
| richard_maw | ok, 2 things | 16:49 |
| Krin | ok? | 16:50 |
| richard_maw | 1. exports don't persist between different commands, you'd need to put it in the same command, I'll give you an example | 16:50 |
| robtaylor | overleyfs has been merged! | 16:50 |
| robtaylor | http://lwn.net/Articles/618141/ | 16:50 |
| richard_maw | 2. I assume the `src::` is a transcription error, and you actually meant `cd src &&`? | 16:50 |
| Krin | richard_maw, your right on 2, transcript error | 16:51 |
| richard_maw | robtaylor: about time we had a unionfs in the kernel | 16:51 |
| robtaylor | yep :D | 16:51 |
| Krin | well, no i dont have the cd in there, the others i'v looked at didnt have cd i dont think *looks again incase his brain is playing tricks again* | 16:51 |
| richard_maw | Krin: you can write it like http://pastebin.com/WceexWhj | 16:52 |
| Kinnison | robtaylor: wootle! | 16:52 |
| pedroalvarez | I wonder if this script is still valid now that we support hard float platforms: http://git.baserock.org/cgi-bin/cgit.cgi/delta/eglibc2.git/tree/morph-arch-config?h=baserock/2.15-build-essential | 16:52 |
| Kinnison | Do we use it? | 16:52 |
| richard_maw | Krin: we tend not to have autocd enabled, if you do find one that doesn't use the cd command, it needs to be squashed with a hammer | 16:53 |
| pedroalvarez | Kinnison: if that question was for me: yes, we use it when building eglibc | 16:53 |
| Krin | leave it to me to pick the one thats not quite to standard from a list of a few hundred XD | 16:53 |
| Kinnison | pedroalvarez: hmm | 16:53 |
| Kinnison | pedroalvarez: I think we no longer build (or need to build) soft-float | 16:54 |
| Krin | nope, my brain was playing tricks richard_maw, there is a CD there >.< | 16:54 |
| straycat | I have a pretty basic lorry question but I just want to check, I want to lorry from a new place, I take it all the git objects that were lorried in from the previous place will stay around, so people can still build their old systems (assuming their definitions are not floating) | 16:54 |
| Kinnison | straycat: assuming their anchored their commits in a baserock/ branch, it should be fine | 16:55 |
| pedroalvarez | Kinnison: yeah, but we are also using the --without-fp flag in hard-float | 16:56 |
| straycat | Even if they didn't, why would the existing objects be removed? | 16:57 |
| Kinnison | pedroalvarez: Interesting | 16:57 |
| Kinnison | pedroalvarez: along with the vfpv3d16 arguments? | 16:57 |
| pedroalvarez | Kinnison: is the first time I see that one | 16:58 |
| * Kinnison may be confused | 16:58 | |
| Kinnison | it has been a long time since he looked at build-essential | 16:58 |
| robtaylor | oh, sadness. overlayfs only suports 2 layers. | 16:58 |
| Kinnison | can you overlay onto an overlay? | 16:59 |
| pedroalvarez | Kinnison: no worries, I will investigate | 16:59 |
| robtaylor | Kinnison: i think so, but its possible you hit issues. mitgh not be an issue for the staging area build case though | 17:00 |
| Kinnison | Mmm :-( | 17:01 |
| * Kinnison is still thinking that a fuse FS might be best for the staging area | 17:01 | |
| robtaylor | unionfs-fuse? | 17:01 |
| richard_maw | or we could knock something together quickly with pyfilesystem, as it lets you export your logical filesystem as a FUSE one | 17:02 |
| robtaylor | http://developerblog.redhat.com/2014/09/30/overview-storage-scalability-docker/ | 17:02 |
| robtaylor | what does that gain you? | 17:03 |
| robtaylor | the main problem with fuse is that it will have a pretty constant impact on the build speed itself, not sure to what degree, but i do recall the ubuntu folks dropped the unionfs-fuse approach for live cds as it was way too slow | 17:05 |
| robtaylor | but i'm interested what you;d do with a more programmatic interface for the staging fs? | 17:06 |
| jjardon | pedroalvarez: kill it! | 17:06 |
| jjardon | pedroalvarez: still working in the glibc port? ;) | 17:07 |
| pedroalvarez | jjardon: i'm getting there | 17:07 |
| jjardon | pedroalvarez: \o/ | 17:08 |
| pedroalvarez | jjardon: I wanted to send some patches today but not sure I'll have the time | 17:08 |
| straycat | pedroalvarez, I know how you feel, yesterday I didn't have time for Sarah & Duck | 17:08 |
| *** ssam2 [~ssam2@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving] | 17:09 | |
| pedroalvarez | jjardon: do you have plans for glibc? i remember that was you who requested the glibc lorry | 17:09 |
| pedroalvarez | straycat: Sarah & Duck? | 17:09 |
| jjardon | pedroalvarez: nope, I wanted to fix the coreutils stuff first, then maybe systemd | 17:09 |
| * straycat nods | 17:09 | |
| * pedroalvarez listens to straycat | 17:10 | |
| *** dutch [~william@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Quit] | 17:10 | |
| straycat | pedroalvarez, I can pm you some links if you like | 17:10 |
| *** Krin [~mikesmith@host86-190-220-106.range86-190.btcentralplus.com] has quit [Remote host closed the connection] | 17:11 | |
| pedroalvarez | jjardon: re: kill it! | 17:11 |
| pedroalvarez | jjardon: are you sure? | 17:11 |
| robtaylor | straycat: sarah and duck rocks | 17:12 |
| straycat | this must be safe though, we surely don't need to lorry to a new place everytime upstream moves repo | 17:12 |
| jjardon | jjardon: we do not build no-hard float systems. And I doubt we are going to do it in the future | 17:12 |
| straycat | robtaylor, :) | 17:13 |
| jjardon | pedroalvarez: ^^ | 17:15 |
| * jjardon should stop talking with himself | 17:15 | |
| pedroalvarez | jjardon: I'm confused | 17:16 |
| robtaylor | richard_maw: Kinnison: is it so we could mount the tars directly, rather than extract them? | 17:16 |
| pedroalvarez | jjardon: ah, double negation. ok | 17:16 |
| *** Krin [~mikesmith@host86-190-220-106.range86-190.btcentralplus.com] has joined #baserock | 17:16 | |
| *** zoli__ [~zoli_@0x5e91887a.adsl.cybercity.dk] has joined #baserock | 17:18 | |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Read error: Connection reset by peer] | 17:18 | |
| * Kinnison was more thinking from the PoV of rapid construction without the hardlinking, and tracking of changes for system artifact construction | 17:19 | |
| Kinnison | And you don't mount tars if you want efficiency | 17:19 |
| Kinnison | tars are awful for that | 17:19 |
| Kinnison | they lack a central index | 17:19 |
| robtaylor | yeh, thats what i was thinking | 17:20 |
| robtaylor | Kinnison: but you can just use overlayfs stacked to do the rapid conctruation | 17:21 |
| Kinnison | I worry about having 900+ stacked mounts | 17:21 |
| robtaylor | Kinnison: i'm trying to understand what the benefit would be to using a fuse fs | 17:21 |
| * Kinnison is imagining that you prepre the staging area, make the system integration changes, unmount, and get a tbdiff patch out the other side | 17:21 | |
| robtaylor | Kinnison: itd need testing yea. I know aufs does it happily, but.. | 17:21 |
| robtaylor | Kinnison: i'm curious what docker does here, going to go read up | 17:23 |
| Kinnison | AIUI docker has pluggable backends, but I'm not familiar with it | 17:23 |
| robtaylor | Kinnison: yeah, worth having a read of the link i posted a little earlier | 17:23 |
| * richard_maw didn't look at the aufs stuff, but the VFS backend will use `cp -a --reflink`, and the btrfs backend uses snapshots | 17:24 | |
| *** thecorconian [~thecorcon@136.1.1.104] has joined #baserock | 17:24 | |
| robtaylor | unrelatedley, has everyone intereted in CI seen drone? https://github.com/drone/drone | 17:24 |
| *** thecorconian [~thecorcon@136.1.1.104] has quit [] | 17:24 | |
| * Kinnison adds the link to his to-read folder | 17:24 | |
| robtaylor | https://github.com/docker/docker/pull/7619 | 17:24 |
| Kinnison | I saw drone, I saw drone.io at least was tied to gihub, I stopped looking | 17:24 |
| robtaylor | Kinnison: apparently also does bitbucket and gitlab | 17:25 |
| robtaylor | Kinnison: i presume i has to have something that supports pull requests by its nature | 17:26 |
| Kinnison | I believe the Baserock project is looking toward gerrit, as such Zuul is the obvious coordinator/gatekeeper and something based around gearman for the build workers for CI/CD | 17:27 |
| Kinnison | richard_maw: did you find the email I referenced earlier? | 17:30 |
| richard_maw | only just finished writing the mail I was composing | 17:31 |
| richard_maw | found it | 17:31 |
| Kinnison | :-) | 17:32 |
| richard_maw | it's asked for verification now, because it hasn't been logged into by my browser | 17:33 |
| richard_maw | fingerprint | 17:33 |
| Krin | hmm, still confused, it seems like the build.sh is not being found, though that should be on the top level of ant... confused krin is confused | 17:34 |
| richard_maw | oh, and it's throwing the same verification hissy fit if I try to log into the gmail | 17:34 |
| Kinnison | richard_maw: I'll come see if I can help with the verification in a sec | 17:35 |
| Kinnison | Krin: Could you pastebin your morphology and the error? | 17:36 |
| Kinnison | Krin: If for whatever reason copy/paste isn't working from your VM <-> Host, then scp the morphology file out first :-) | 17:36 |
| richard_maw | Krin: build.sh is in the top level, so why do you have to chdir into the src directory? | 17:36 |
| *** franred [~franred@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving] | 17:36 | |
| Krin | richard_maw, it was something i was trying, as just executing the command alone didnt work, and i was thinking that the cd src && was a step i didnt know about. | 17:37 |
| richard_maw | http://ant.apache.org/manual/install.html#buildingant doesn't say anything about needing to chdir into src | 17:37 |
| Krin | it was something i was seeing in other morphs | 17:38 |
| Krin | hmm, ok, some other change i made seems to have fixed the command not being executed, now it's telling me that i havent exported JAVA_HOME corectly... but it's progress! PROGRESS! | 17:43 |
| * Krin goes for a cup of PROGRESS tea | 17:44 | |
| jjardon | mmm, so if I morph edit a chunk, go to the directory and commit the changes, what its getting build by morph? seems morph builds master instead the generated branch here | 17:49 |
| straycat | Can you rephrase? I couldn't follow that | 17:50 |
| paulsherwood | jjardon: what was your original morph command? branch or checkout? | 17:53 |
| jjardon | straycat: what ref is built when I morph edit and then make a change in the chunk? | 17:54 |
| paulsherwood | jjardon: if you did morph branch foo, it builds whatever it thinks is the foo branch in the repo you created in morph edit | 17:55 |
| paulsherwood | if you did morph checkout master, it maybe trying to build master. bug | 17:55 |
| persia | Krin: Re: JAVA_HOME: annoyingly that's something that you'll need to reset over and over and over again. For building, you can probably hardcode it, but to use it, you may either want wrapper scripts to launch things, or to do interesting things to the user environment. | 17:55 |
| * paulsherwood abandoned use of morph checkout a long time ago, when he couldn't work out how to get around this issue | 17:55 | |
| jjardon | ah, ok: I see, morph modify definitions to point to the generated branch in the chunk | 17:56 |
| * paulsherwood abandoned morph edit more recently, when persia showed him it was unnecessary | 17:56 | |
| straycat | paulsherwood, so you just clone from the trove and edit the ref manually? | 17:56 |
| paulsherwood | jjardon: see https://vimeo.com/110065776 for examples of hacking on morph files directly | 17:57 |
| paulsherwood | maybe not even clone :) | 17:57 |
| paulsherwood | straycat: yup | 17:57 |
| jjardon | paulsherwood: thanks, will do | 17:57 |
| persia | For the record, morph edit provides a benefit that not using it doesn't: that being that morph will automatically commit the current uncommitted changes in the working directory for the edited chunk. At least for me, this is a misfeature, but there are people who want to avoid running "git commit" themselves, and prefer morph to do it for them. | 17:57 |
| *** mSher [~mike@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving] | 17:59 | |
| straycat | It's not that it's edited, it's that it's a repo in the workspace with uncommitted changes in it, unless I've misunderstood. | 18:00 |
| * paulsherwood wants an end to workspaces, too :) | 18:00 | |
| rdale | it is certainly difficult to learn this stuff because 'morph edit' seems to be depracted, but a replacement doesn't seem to be in place yet | 18:01 |
| paulsherwood | rdale: pls watch the video, let me know if it's still muddy afterwards | 18:02 |
| straycat | it's not deprecated as far as i'm aware | 18:02 |
| rdale | i've watched the two new videos, and they were very good | 18:02 |
| paulsherwood | but didn't clear this up? | 18:02 |
| paulsherwood | straycat: it is for me :-) | 18:03 |
| rdale | but the videos didn't show 'morph edit' - you changed tags/branches that pointed to the trove, but that isn't the same as making local changes to checked out stuff from the trove and testing whether they build etc | 18:04 |
| paulsherwood | rdale: ah, ok. | 18:04 |
| paulsherwood | rdale: same process. git clone, then edit repo: to be file:///the/local/directory/you/cloned/to | 18:04 |
| paulsherwood | and of you go. | 18:05 |
| paulsherwood | off | 18:05 |
| paulsherwood | but i feel another video coming on :) | 18:05 |
| paulsherwood | i'll get that done soonish | 18:05 |
| rdale | yes, i can do that sort of thing with 'file://..', but it feels very hand crafted | 18:05 |
| paulsherwood | :) | 18:06 |
| persia | rdale: For which feature of morph edit do you seek a replacement? | 18:07 |
| rdale | well i wasn't an expert in using 'morph edit' in the first place | 18:07 |
| *** jonathanmaw [~jonathanm@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving] | 18:08 | |
| rdale | there should be a way of editing the ref that a chunk points to in the lorries, where you could checkout the lorried repo locally and test. then commit back both the change to the stratum you are working on at the same time as commiting back any changes to the lorried repo that they point to | 18:10 |
| *** Krin [~mikesmith@host86-190-220-106.range86-190.btcentralplus.com] has quit [Remote host closed the connection] | 18:13 | |
| paulsherwood | there is. most of it is git :) | 18:15 |
| paulsherwood | sorry, it's been a long day... :-) | 18:15 |
| * paulsherwood drops off | 18:16 | |
| *** CTtpollard [~tom@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Ex-Chat] | 18:29 | |
| *** tiagogomes_ [~tiagogome@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving] | 18:50 | |
| *** tiagogomes [~tiagogome@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 19:01 | |
| *** tiagogomes [~tiagogome@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving] | 19:09 | |
| *** paulk-aldrin [~paulk@armstrong.paulk.fr] has quit [Ping timeout: 250 seconds] | 19:14 | |
| *** VVIBIN [~VVIBIN@202.0.77.198] has joined #baserock | 19:16 | |
| *** VVIBIN [~VVIBIN@202.0.77.198] has quit [Ping timeout: 258 seconds] | 19:20 | |
| *** DavePageCT [~davepage@valueofnothing.org.uk] has joined #baserock | 19:21 | |
| * DavePageCT waves | 19:22 | |
| *** paulk-aldrin [~paulk@armstrong.paulk.fr] has joined #baserock | 19:23 | |
| *** paulk-aldrin [~paulk@armstrong.paulk.fr] has quit [Ping timeout: 245 seconds] | 19:40 | |
| *** zoli__ [~zoli_@0x5e91887a.adsl.cybercity.dk] has quit [Remote host closed the connection] | 20:19 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has quit [Quit: Bye] | 20:37 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has joined #baserock | 20:38 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has quit [Client Quit] | 20:38 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has joined #baserock | 20:38 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has quit [Client Quit] | 20:40 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has joined #baserock | 20:40 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has quit [Client Quit] | 20:40 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has joined #baserock | 20:41 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has quit [Client Quit] | 20:42 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has joined #baserock | 20:42 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has quit [Ping timeout: 256 seconds] | 20:56 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has joined #baserock | 20:56 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has quit [Ping timeout: 256 seconds] | 21:17 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has joined #baserock | 21:17 | |
| * paulsherwood waves back, and wonders what's happened to ct irc ... | 21:17 | |
| paulsherwood | DavePageCT: ^^ | 21:18 |
| paulsherwood | probably offtopic for here, though... | 21:18 |
| *** genii [~quassel@ubuntu/member/genii] has quit [Read error: Connection reset by peer] | 21:24 | |
| * robtaylor too | 21:40 | |
| *** zoli_ [~zoli_@0x5e91887a.adsl.cybercity.dk] has joined #baserock | 21:46 | |
| *** zoli_ [~zoli_@0x5e91887a.adsl.cybercity.dk] has quit [Changing host] | 21:46 | |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 21:46 | |
| *** mSher [~mike@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 22:26 | |
| *** mSher [~mike@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving] | 22:32 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has quit [Ping timeout: 256 seconds] | 22:38 | |
| *** radiofree [radiofree@unaffiliated/radiofree] has joined #baserock | 22:38 | |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection] | 22:41 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!