| *** petefoth [~petefoth@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 07:23 | |
| *** rdale [~quassel@92.Red-88-20-2.staticIP.rima-tde.net] has quit [Ping timeout: 272 seconds] | 08:25 | |
| *** wdutch [~william@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 08:35 | |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 08:37 | |
| *** rdale [~quassel@92.Red-88-20-2.staticIP.rima-tde.net] has joined #baserock | 08:49 | |
| *** tiagogomes [~tiagogome@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 08:53 | |
| *** mariaderidder [~maria@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 08:58 | |
| *** mariaderidder [~maria@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Client Quit] | 08:58 | |
| *** De|ta [~arc@195.242.156.171] has quit [Quit: reboot] | 09:07 | |
| *** bashrc [~motters@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 09:13 | |
| *** ssam2 [~ssam2@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 09:20 | |
| Mode #baserock +v ssam2 by ChanServ | 09:20 | |
| *** Krin [~mikesmith@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 09:59 | |
| *** jonathanmaw [~jonathanm@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 10:02 | |
| *** ssam2 [~ssam2@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Ping timeout: 265 seconds] | 10:24 | |
| franred | where is the "manual page" for lorry? in README nad in lorry wiki it says: "see the manual page for instructions on using" but I can not find it | 10:26 |
|---|---|---|
| petefoth | franred: I think teh information that will/should be in the man page is just the command line help | 10:27 |
| paulsher1ood | franred: http://wiki.baserock.org/Trove/reference/ | 10:27 |
| paulsher1ood | petefoth: it took me some time to find the trove reference earlier, shouldn't it be in guides? and perhaps linked in other places too? | 10:28 |
| franred | petefoth, I disagree, lorry is fairly complex to not give any example and more when you can mess up g.b.o | 10:28 |
| paulsher1ood | +1 | 10:29 |
| paulsher1ood | franred: to be more precise http://wiki.baserock.org/Trove/reference/#index4h2 | 10:29 |
| straycat | mess up gbo? why do you need to use lorry on gbo? | 10:30 |
| paulsher1ood | straycat: i think he meant mess up a trove | 10:30 |
| straycat | oh okay | 10:31 |
| franred | straycat, http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/lorry.git/tree/README | 10:31 |
| franred | if you have push access to g.b.o you could mess it up testing lorry | 10:32 |
| petefoth | franred: paulsher1ood: there is currently a ‘Story’ to ‘improve’ that page. In the meantime it might be worth adding paulsher1ood’s link to the README | 10:32 |
| franred | Im about testing a patch to upgrade git, I want to test that we can lorry without errors with the new version so rather than deploy a trove Im going to try to use lorry. | 10:33 |
| * straycat nods | 10:34 | |
| straycat | I'm not sure you can do damage with the defaults anyhow, since it's ssh://gitano@roadtrain.codethink.co.uk/delta rather than gbo, but I may be wrong | 10:36 |
| straycat | In any case you can obviously use --pull-only | 10:37 |
| Kinnison | roadtrain?! | 10:37 |
| Kinnison | meep | 10:37 |
| Kinnison | that is a long long time ago | 10:37 |
| *** ssam2 [~ssam2@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 10:37 | |
| Mode #baserock +v ssam2 by ChanServ | 10:38 | |
| straycat | yeah i have no idea what that is | 10:38 |
| Kinnison | it's a server from before we even had the concept of a trove | 10:38 |
| Kinnison | the roadtrain had a lot of lorries on it | 10:38 |
| Kinnison | *baddumtish* | 10:38 |
| straycat | bahahahah. | 10:38 |
| straycat | >.> | 10:38 |
| straycat | so maybe we should patch lorry so that it doesn't have default push urls and so can't do any damage | 10:38 |
| straycat | (not that I'm saying it can do damage as it is) | 10:39 |
| Kinnison | well default push urls to things which don't exist are what it currently has | 10:39 |
| Kinnison | that seems fairly safe | 10:39 |
| straycat | okay | 10:40 |
| franred | straycat, I would like to use also push, so Im looking if I can push into the same machine Im pulling the repo | 10:40 |
| pedroalvarez | franred: hmm... then you have to run a git server in the machine as well | 10:41 |
| straycat | you can push straight to git repos | 10:41 |
| straycat | locally | 10:41 |
| pedroalvarez | oh yeah | 10:41 |
| franred | cool, I can do that | 10:42 |
| franred | thanks straycat :) | 10:42 |
| richard_maw | also, there's a git server that comes with git | 10:43 |
| * richard_maw uses it in the test suite | 10:43 | |
| *** petefoth [~petefoth@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [] | 10:43 | |
| bashrc | I can add some hashes to the baserock images, but folks will need to keep them up to date when new images are released | 10:43 |
| *** petefoth [~petefoth@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 10:44 | |
| ssam2 | bashrc: that'd be good. we need to have this done as part of the release process. I'm not sure what state the release process is in at the moment, we did a lot of work on automating it a while back | 10:45 |
| bashrc | yes really it should just be part of a release script | 10:45 |
| pedroalvarez | sigh... I didn't spend time working out how these scripts work since I always think we are stopping of doing releases | 10:47 |
| bashrc | never release anything. I guess that's one way of dealing with the problem | 10:50 |
| jmacs | doffm_: Are you around? | 10:51 |
| paulsher1ood | petefoth: /win 30 | 10:51 |
| paulsher1ood | bah!!!!! | 10:51 |
| paulsher1ood | now folks know i talk to petefoth, and have 30 windows :-) | 10:51 |
| bashrc | so is the plan to always build from source and not have any images? | 10:51 |
| paulsher1ood | yes | 10:52 |
| * straycat has 31 windows | 10:52 | |
| paulsher1ood | but images are created, of course :-) | 10:52 |
| straycat | (in case anyone was wondering) | 10:52 |
| paulsher1ood | straycat: i have more than that :-) | 10:52 |
| rjek | paulsher1ood: Talking to petefoth is indeed a huge stigma that you will now have to live with ;-) | 10:52 |
| petefoth | It’s OK - petefoth doesn’t listen, and only has 9 windows | 10:52 |
| * rjek grins. | 10:53 | |
| Kinnison | straycat: I have around 18 channels and somewhere approaching 60 queries currently | 10:53 |
| straycat | I'm glad I know all this :) | 10:53 |
| *** sambishop [~sambishop@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Ping timeout: 250 seconds] | 11:00 | |
| *** petefoth [~petefoth@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Ping timeout: 245 seconds] | 11:01 | |
| *** sambishop [~sambishop@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 11:12 | |
| *** petefoth [~petefoth@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 11:37 | |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection] | 11:46 | |
| *** aananth [~caananth@74.112.167.117] has joined #baserock | 11:47 | |
| richard_maw | I've gotten two +1s to my runtime-depends patch, are we in agreement that we want it now? | 11:58 |
| pedroalvarez | I guess it would have -1's otherwise | 12:00 |
| richard_maw | it was a contentious issue at one point, so I thought I'd best check first | 12:01 |
| franred | richard_maw, I though was paulsher1ood the one who was concern about it | 12:08 |
| *** De|ta [~arc@195.242.156.171] has joined #baserock | 12:09 | |
| franred | maybe it is worth to check with him if that concerns has been disappeared | 12:09 |
| *** zoli_ [~zoli_@cpe.ge-0-2-0-366.abnqu1.dk.customer.tdc.net] has joined #baserock | 12:20 | |
| *** zoli_ [~zoli_@cpe.ge-0-2-0-366.abnqu1.dk.customer.tdc.net] has quit [Changing host] | 12:20 | |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 12:20 | |
| paulsher1ood | i still don't see the point, to be honest.... richard_maw - is there an urgent need for this? | 12:27 |
| paulsher1ood | if there isn't, i would like to throw my ybd experiment onto the list, and see if the srtucture i'm proposing addresses the need, or not? | 12:28 |
| paulsher1ood | s/structure/definitions structure/ | 12:28 |
| straycat | the need is that we end up including build dependencies in the resultant system afaics | 12:29 |
| richard_maw | no urgent need yet, except that there's a trend towards splitting strata up (see lorry as the latest); I was just a little annoyed that it's sitting in review and hasn't been merged | 12:29 |
| straycat | richard_maw, sorry for not being able to get around to it :( | 12:30 |
| paulsher1ood | richard_maw: ok, if i may, i'll send an RFC on the ybd stuff today, and see whether you refute it, or see merit in adopting some of it | 12:31 |
| richard_maw | paulsher1ood: and stratum runtime-depends would allow us to handle the split up strata nicely now, but there'd be more implementation time required to have your ybd format definitions | 12:31 |
| *** petefoth [~petefoth@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Ping timeout: 255 seconds] | 12:49 | |
| *** petefoth [~petefoth@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 12:49 | |
| jjardon | richard_maw: an example system using the capabilities of your branch would be helpful | 13:14 |
| *** petefoth1 [~petefothe@access.ducie-dc1.codethink.co.uk] has quit [Ping timeout: 250 seconds] | 13:18 | |
| *** petefoth1 [~petefothe@access.ducie-dc1.codethink.co.uk] has joined #baserock | 13:24 | |
| *** tiagogomes [~tiagogome@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving] | 13:39 | |
| *** tiagogomes [~tiagogome@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 13:39 | |
| *** jmacs [~jimmacart@access.ducie-dc1.codethink.co.uk] has quit [Ping timeout: 250 seconds] | 13:43 | |
| *** DavePage [~dpage@access.ducie-dc1.codethink.co.uk] has quit [Ping timeout: 245 seconds] | 13:43 | |
| *** ridgerun1er [~robjones@access.ducie-dc1.codethink.co.uk] has quit [Ping timeout: 272 seconds] | 13:43 | |
| *** petefoth1 [~petefothe@access.ducie-dc1.codethink.co.uk] has quit [Ping timeout: 265 seconds] | 13:44 | |
| *** jmacs [~jimmacart@access.ducie-dc1.codethink.co.uk] has joined #baserock | 13:52 | |
| *** petefoth1 [~petefothe@access.ducie-dc1.codethink.co.uk] has joined #baserock | 13:54 | |
| *** ridgerunner [~robjones@access.ducie-dc1.codethink.co.uk] has joined #baserock | 13:56 | |
| paulsher1ood | does the current jetson flash script mentioned on http://wiki.baserock.org/guides/baserock-jetson/ include max sizing? | 14:03 |
| *** DavePage [~dpage@access.ducie-dc1.codethink.co.uk] has joined #baserock | 14:03 | |
| *** mariaderidder [~maria@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 14:07 | |
| paulsher1ood | hmmm... the instructions at http://wiki.baserock.org/guides/baserock-jetson/ no longer work... | 14:09 |
| pedroalvarez | paulsher1ood: at what point do they fail? | 14:11 |
| paulsher1ood | pedroalvarez: sadly, at a point last edited by you :-) | 14:11 |
| pedroalvarez | yeah, I know that | 14:11 |
| pedroalvarez | :) | 14:11 |
| pedroalvarez | that's why I'm asking | 14:11 |
| paulsher1ood | cd baserock... do stuff.. then ./flash... but flash is in the directory above | 14:12 |
| pedroalvarez | ah! ok | 14:12 |
| paulsher1ood | cd .. doesn't work... | 14:12 |
| paulsher1ood | Error: missing system image | 14:12 |
| pedroalvarez | oh yeah, I know what it is | 14:12 |
| paulsher1ood | should the flash script be in the baserock directory? | 14:13 |
| pedroalvarez | The problem is that it should be "./baserock-jetson-flash.sh /full/path/to/Linux_for_Tegra/baserock/baserock-current-build-system-armv7lhf-jetson.img" | 14:13 |
| paulsher1ood | after cd .. | 14:14 |
| paulsher1ood | ? | 14:14 |
| *** franred [~franred@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving] | 14:14 | |
| pedroalvarez | yes | 14:14 |
| paulsher1ood | ok i'll try it, and fix? | 14:14 |
| *** franred [~franred@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 14:14 | |
| pedroalvarez | paulsher1ood: could you fix it if that works? | 14:15 |
| *** franred [~franred@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Remote host closed the connection] | 14:16 | |
| *** franred [~franred@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 14:17 | |
| *** franred [~franred@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Client Quit] | 14:17 | |
| *** franred [~franred@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 14:18 | |
| pedroalvarez | maybe we can do something clever with `curl` instead of `cd`ing to directories, etc | 14:20 |
| doffm_ | jmacs: I'm here. | 14:27 |
| *** fay_ [~fay@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving] | 14:32 | |
| *** fay [~fay@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 14:32 | |
| fay | ck fay_ | 14:32 |
| fay is now known as fay_ | 14:33 | |
| *** jmacs_ [~jimmacart@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 14:35 | |
| jmacs_ | doffm_: Hi. We're having some IRC problems here | 14:35 |
| jmacs_ | doffm_: I was just wondering if you can remember the work you did with ceph a year ago. I found your ceph.configure file but wasn't sure what state it was in, whether you had an example of it working, for example | 14:36 |
| *** robtaylor [~robtaylor@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Remote host closed the connection] | 14:42 | |
| doffm_ | I can vuagely remember. | 14:49 |
| doffm_ | I | 14:49 |
| doffm_ | jmacs_: I'll check if I have an example cluster file. | 14:49 |
| jmacs | doffm_: That would be useful. | 14:52 |
| *** aananth [~caananth@74.112.167.117] has quit [Quit: Leaving] | 14:53 | |
| paulsher1ood | pedroalvarez: i've fixed it, i think | 15:10 |
| doffm_ | jmacs: I can't seem to find an example. Sorry. Not sure it would work a year later anyway, given updates to ceph in that time. | 15:17 |
| doffm_ | The configuration script seems moderately well documented. | 15:18 |
| jmacs | Yeah, the configuration script is good | 15:18 |
| pedroalvarez | paulsher1ood: it shouldn't be "baserock-14.40.1-devel-system-armv7lhf-jetson.img" | 15:18 |
| jmacs | There must be some manual configuration work needed to set up the cluster after it's deployed though? | 15:18 |
| pedroalvarez | paulsher1ood: (when calling the flash script) | 15:19 |
| doffm_ | jmacs: I believe there was ONE step that had to be performed on the running instances. For the life of me I can't remember what. I'll try and find the e-mail. | 15:19 |
| paulsher1ood | pedroalvarez: i didn't touch that bit. | 15:20 |
| paulsher1ood | i guess so. i'll retry with the latest image, then? :-) | 15:20 |
| jmacs | I can set up ceph clusters, that's not a problem - it's just that my way needs several more steps to set up, and you seem to be doing it in a simpler way which I don't quite understand | 15:20 |
| doffm_ | I'll send you the e-mail that I sent with the patch. Me of 1 year ago probably understood this better. | 15:21 |
| jmacs | Cheers | 15:22 |
| doffm_ | jmacs: In general though the ceph-deploy script uses ssh to run commands on the nodes. The ceph configure script runs the same commands (Or similar) but at configure time rather than on the running instance. | 15:27 |
| doffm_ | Thats a bit hacky, but lots of the commands just create data directories or install auth keys, so for many it wasn't a problem. | 15:28 |
| doffm_ | The magic is in the ceph init scripts. The init script starts daemons based on the existance of data directories. | 15:29 |
| *** robtaylor [~robtaylor@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 15:29 | |
| doffm_ | So the script just creates them and lets the ceph init script bring up the correct daemons. | 15:29 |
| doffm_ | Most of the script is just moving configuration files, auth files to the right places and creating the data directories. | 15:30 |
| doffm_ | jmacs: Unfortunately ceph-deploy went through a BIG change this last year. The init scripts / data dir locations might have changed also. I hope not, but you never know. | 15:32 |
| jmacs | Righto. Well, ceph still works. It's just the automation of setup that we need to finish. | 15:33 |
| jmacs | Thanks for the email. | 15:34 |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection] | 15:42 | |
| bashrc | is baserock 14.40 recent ? | 15:51 |
| *** tiagogomes [~tiagogome@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving] | 15:52 | |
| pedroalvarez | bashrc: it is from the week 40 of the year 2014 | 15:52 |
| *** tiagogomes [~tiagogome@82-70-136-246.dsl.in-addr.zen.co.uk] has joined #baserock | 15:52 | |
| bashrc | I'll take that as a yes | 15:52 |
| pedroalvarez | is from 6 weeks ago | 15:52 |
| pedroalvarez | we released 14.46 last week | 15:53 |
| bashrc | requires 40G, and yet the install guide creates a 30G system #seemsinconsistent | 15:54 |
| pedroalvarez | why 40G? | 15:55 |
| pedroalvarez | I used to use a 30G disk, but I had to clean up very often | 15:55 |
| bashrc | "/tmp/morph_tmp requires 4000000000 bytes free, has 1048940544" | 15:55 |
| pedroalvarez | that is 4G | 15:56 |
| bashrc | oh | 15:56 |
| paulsher1ood | bashrc: did you remember ln -sv /src/morph.conf /etc/morph.conf | 15:58 |
| paulsher1ood | ? | 15:58 |
| paulsher1ood | oh, ignore me | 15:58 |
| paulsher1ood | /ignore off | 15:59 |
| paulsher1ood | bashrc: use cycle.sh - it will clean up for you :) | 15:59 |
| ssam2 | bashrc: paul was right first time, in fact | 16:02 |
| bashrc | strange. /dev/sda is only 6G. I thought I'd fallocated 30 | 16:02 |
| ssam2 | bashrc: morph unpacks whole systems into its tmpdir, so using /src/tmp (so that it isn't an in-memory tmpfs) is recommended | 16:02 |
| paulsher1ood | bashrc: we normally have a separate device for source? | 16:03 |
| ssam2 | bashrc: the default /etc/morph.conf found on http://wiki.baserock.org/quick-start/#index4h2 takes care of setting the correct tmpdir location, make sure you followed that (and the instructions in the previous section for setting up a separate /src partition, as paul points out) | 16:04 |
| richard_maw | paulsher1ood: I thought you were only going to show us what your proposed changed definition syntax would be | 16:06 |
| bashrc | so, the instructions on http://wiki.baserock.org/devel-with fail | 16:08 |
| straycat | bashrc, can you be more specific? | 16:08 |
| paulsher1ood | richard_maw: i did that previously, i thought? | 16:09 |
| paulsher1ood | then i was guided to produce poc code | 16:09 |
| paulsher1ood | maybe i misunderstood | 16:10 |
| bashrc | entering the commands up to morph --verbose build systems/base-system-x86_64-generic.morph doesn't build | 16:10 |
| richard_maw | my understanding of my own comments at least was that I'd look forward to seeing your poc | 16:10 |
| paulsher1ood | bashrc: please paste output | 16:10 |
| paulsher1ood | richard_maw: well now you can :-) | 16:10 |
| richard_maw | the bit to convert definitions to your proposed new format is the bit I'm after | 16:11 |
| bashrc | ERROR: Insufficient space on disk: | 16:11 |
| bashrc | /tmp/morph_tmp requires 4000000000 bytes free, has 1048940544 | 16:11 |
| bashrc | Please run `morph gc`. If the problem persists increase the disk size, manually clean up some space or reduce the disk space required by the tempdir-min-space and cachedir-min-space configuration options. | 16:11 |
| paulsher1ood | richard_maw: https://github.com/devcurmudgeon/ybd/blob/master/morph-converter.sh | 16:12 |
| bashrc | also the 30G drive appears to be missing | 16:12 |
| pedroalvarez | bashrc: have you followed http://wiki.baserock.org/quick-start first? | 16:12 |
| bashrc | do I need to mount it? | 16:12 |
| richard_maw | paulsher1ood: was already trying that out before you posted the link | 16:12 |
| paulsher1ood | bashrc: please check have you followed the instructions at http://wiki.baserock.org/guides/no-frills/ ? | 16:12 |
| bashrc | ok it looks like I didn't set up the second disk | 16:13 |
| paulsher1ood | richard_maw: note i haven't done anything fancy like moving directories around yet | 16:13 |
| paulsher1ood | so 'systems' are still in systems directory, etc. but they are no longer different from 'strata' or 'chunks' in this scheme | 16:14 |
| paulsher1ood | bashrc: i'm interested to know how that happened? we've seen an example already this week where something on the wiki was wrong... did you follow a page that didn't cover it? | 16:14 |
| richard_maw | paulsher1ood: I feel this is useful for the renaming discussion, but in its current state not so useful for debating whether we need runtime depends, which I understood to be the main point for posting the code | 16:18 |
| paulsher1ood | it's the main point for posting the code *now*, but not in general | 16:19 |
| paulsher1ood | i'm suggesting contents are in at runtime, build-depends are not. | 16:19 |
| paulsher1ood | i still don't understand what runtime-depends is except that they need to be present at runtime? | 16:20 |
| * paulsher1ood admits, he may be being very obtuse here... having failed to grasp the concept properly after several discussions and emails | 16:21 | |
| richard_maw | I've tried to communicate this point before in my responses to the e-mails. In my model we have a separation between runtime-depends and chunks because how they are referred to is different. In your model runtime depends and chunks are put together in the contents list. | 16:21 |
| richard_maw | So we're both agreeing on the need for runtime-depends hopefully | 16:22 |
| paulsher1ood | yes. is there some weakness in my model from your pv? | 16:22 |
| richard_maw | I don't agree that build-dependencies listed in the definition should be applied to runtime-depends | 16:22 |
| paulsher1ood | 'applied to'? | 16:23 |
| richard_maw | my understanding of your model, is that if you have build-depends in a definition then all the entries in your contents list have the build-depends that you've listed added to their list of build-depends | 16:23 |
| paulsher1ood | correct | 16:24 |
| paulsher1ood | which is what currently happens in morph, i think? | 16:24 |
| richard_maw | yes, but that's because we only allow chunks in there | 16:25 |
| paulsher1ood | in which part? a stratum's build-depends include strata? | 16:25 |
| richard_maw | we only have chunks in the "chunks" field which will become "contents" | 16:26 |
| paulsher1ood | right | 16:26 |
| paulsher1ood | in my model any build depend could be a single component, or a set | 16:27 |
| richard_maw | that's fine, and I want to be able to do that | 16:27 |
| richard_maw | I'll try to put together an example where I think this will cause problems | 16:27 |
| paulsher1ood | ok. if i'm being obtuse, sorry for the confusion | 16:28 |
| richard_maw | at least you're giving me lots of opportunity to get used to my new keyboard | 16:28 |
| paulsher1ood | :-) | 16:29 |
| radiofree | ok, so i've got the flashing script working to create a ext4 boot partition, and then flash the btrfs image to mmcblk0p2, and it boots | 16:35 |
| bashrc | building... | 16:35 |
| radiofree | i'm thinking of adding a BOOTFS_PATH="/dev/mmcblk0p1" to morph/system-version-manager | 16:36 |
| radiofree | since this could be used in a generic way (i.e any u-boot system that doesn't have btrfs) | 16:36 |
| paulsher1ood | would that be ok for non-u-boot systems too? | 16:37 |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 16:37 | |
| paulsher1ood | should/could the name make it clear it's u-boot only? | 16:37 |
| radiofree | system-version-manager is very extlinux.conf focused, as long as the bootloader can look for /extlinux.conf in whatever the boot partition is, then yes, i don't see it being a problem | 16:37 |
| radiofree | paulsher1ood: it could be used for other things in the future | 16:38 |
| radiofree | richard_maw: are you happy with BOOTFS_PATH before i start down that road/ | 16:39 |
| pedroalvarez | indeed is very extlinux.conf focused, I'd be ok with it supporting u-boot as well instad of modifying u-boot to read an extlinux.conf file | 16:39 |
| paulsher1ood | cool then :) | 16:39 |
| richard_maw | AAAAAAAAaaaaaa too many things happening at once! | 16:39 |
| richard_maw | what does BOOTFS_PATH *do*? | 16:39 |
| radiofree | pedroalvarez: we don't need to modify u-boot, u-boot already supports that | 16:40 |
| paulsher1ood | richard_maw: welcome to my world :-) | 16:40 |
| pedroalvarez | radiofree: does it? I thought we patched it. | 16:40 |
| radiofree | richard_maw: sorry, BOOT_DEVICE, not BOOTFS_PATH | 16:41 |
| radiofree | it's the cousin of ROOT_DEVICE | 16:41 |
| richard_maw | what does BOOT_DEVICE *do*? | 16:41 |
| radiofree | i.e system-version-manager will check for BOOT_DEVICE and use that for updating the extlinux.conf | 16:41 |
| radiofree | else, use whatever is mounted at / | 16:41 |
| radiofree | (current behaviour) | 16:41 |
| radiofree | pedroalvarez: we patched u-boot to support reading extlinux.conf from a btrfs partition | 16:42 |
| pedroalvarez | ahh | 16:42 |
| radiofree | if a wandboard has u-boot with the ums command on it (probably?) you could use BOOT_DEVICE to allow a wandboard to be upgraded as well | 16:42 |
| radiofree | ums command is not essential, as long as you can partition it with a BOOT_DEVICE and ROOT_DEVICE it would work | 16:43 |
| richard_maw | so when it goes to update /extlinux.conf and copy the kernel somewhere the bootloader can read it, rather than mounting env.get(ROOT_DEVICE, get_device_of(/)), it mounts BOOT_DEVICE to a tempdir and updates the kernels and extlinux.conf relative to that path, rather than the mounted ROOT_DEVICE? | 16:43 |
| radiofree | same goes for any arm board | 16:43 |
| radiofree | richard_maw: exactly | 16:43 |
| radiofree | i'll use the same file structure as well, since that's quite nice | 16:43 |
| richard_maw | in the absence of being able to use the same partition for /boot and /, due to btrfs in u-boot reliability issues, it looks like a workable solution | 16:45 |
| richard_maw | I'm just a little concerned about the interactions of where it gets used at write time | 16:45 |
| richard_maw | since ideally you could use BOOT_DEVICE, ROOT_DEVICE and a field describing the partition structure to handle this without requiring you to say which dev-board | 16:46 |
| radiofree | you could probably do that by just setting the BOOT_DEVICE and ROOT_DEVICE? BOOT_DEVICE="/dev/whateverthebootpartitionisonaboard" ROOT_DEVICE="/dev/myrootfspartition" | 16:48 |
| richard_maw | you'd have to work backwards from the device names to know which partition numbers they are | 16:48 |
| richard_maw | and you're not describing how large to make the partitions | 16:48 |
| richard_maw | nor what fs to make them | 16:49 |
| richard_maw | though we can assume BOOT=ext2 and ROOT=btrfs for nwo | 16:49 |
| *** genii [~quassel@ubuntu/member/genii] has joined #baserock | 16:53 | |
| richard_maw | paulsher1ood: I attempted to write up my thoughts in http://paste.baserock.org/idupejawex.md | 16:58 |
| *** jmacs_ [~jimmacart@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: leaving] | 16:58 | |
| * paulsher1ood goes to sit in a corner and properly engage cpu | 17:00 | |
| richard_maw | it's a bit straw-manny because I've been unable to get a full comprehension of what your proposal is | 17:00 |
| paulsher1ood | richard_maw: why does *anything* have build-essential as contents, unless you want the build tools in the target? | 17:01 |
| richard_maw | because it contains essential system libraries | 17:01 |
| richard_maw | libgcc.so comes from gcc | 17:02 |
| richard_maw | there's also zlib in there because parts of build-essential require it | 17:02 |
| paulsher1ood | taking second first, zlib could be in build-essential (for building) and in core, as contents if required? | 17:03 |
| paulsher1ood | regarding libgcc - tht is a subset of gcc 'chunk' at the moment? | 17:04 |
| richard_maw | paulsher1ood: I think we'd need to drastically split up build-essential in ways morph can't currently handle for that; it may be possible but it's too much of a shift for me to contemplate at the moment | 17:05 |
| paulsher1ood | i think in my model gcc is in build essential, and a subset (one or more products of gcc as they are called currently?) is in core or foundation or whatever | 17:05 |
| richard_maw | paulsher1ood: how exactly, your current proposal doesn't handle artifact splitting | 17:06 |
| paulsher1ood | only because i haven't implemented it yet... it's not that hard, i think? | 17:07 |
| richard_maw | well, supposing you split up definitions sufficiently that you could include libgcc as contents and have your definition include gcc-compiler as a build-depend | 17:08 |
| richard_maw | in my understanding of this, you'd end up saying that to build gcc, you need to build gcc | 17:09 |
| paulsher1ood | don't you? :-) | 17:09 |
| richard_maw | also, many things sound easy until you start to implement them, then they require a complete redesign of the model | 17:09 |
| paulsher1ood | you do, i think. unless you've already built it. | 17:09 |
| paulsher1ood | i | 17:09 |
| paulsher1ood | i think i've already implemented the redesign of the model | 17:10 |
| richard_maw | my point is you may need to do it again to handle artifact splitting | 17:10 |
| paulsher1ood | it's the hard deep linux foo that beats me :-) | 17:10 |
| paulsher1ood | richard_maw: no i don't think so. | 17:10 |
| richard_maw | you won't know that until you start to implement it | 17:10 |
| paulsher1ood | agreed. but i'm quietly confident :-) | 17:11 |
| * paulsher1ood has been wrong before, of course | 17:11 | |
| richard_maw | that's fine, but claiming that a model is the solution to all our problems, when it hasn't taken into account some of the problems our current model is currently solving is… frustrating to deal with | 17:12 |
| paulsher1ood | heh | 17:12 |
| paulsher1ood | i'm not saying i have a solution to all problems. i'm saying i have solutions to some problems. | 17:12 |
| richard_maw | which is fine, it's just causing difficulty when attempting to find merge points in the models | 17:14 |
| paulsher1ood | to try to simplify what i've said already.. i'm raising this because you are proposing to *add* complexity to a model which imo is already too complex. if others agree with you, fair enough | 17:16 |
| * Kinnison notes that often the first step in simplification is to add a required complication | 17:16 | |
| paulsher1ood | not in my worldview :-) | 17:17 |
| Kinnison | premature optimisation is an evil not a good. | 17:17 |
| paulsher1ood | that's a fair generalization, not sure how it applies here, though. | 17:18 |
| ssam2 | my experience is that artifact splitting made a lot of things a bit more ugly | 17:18 |
| ssam2 | but we do need it | 17:18 |
| ssam2 | I don't think the way we expose it to the user is very good, yet | 17:19 |
| paulsher1ood | where do we use it exactly? | 17:19 |
| ssam2 | in the minimal system | 17:19 |
| ssam2 | we don't use it nearly enough, or our devel systems wouldn't be 1.5GB | 17:19 |
| paulsher1ood | so, frankly, we don't use it. | 17:19 |
| ssam2 | I seem to recall a time when you were very pleased that Richard Maw had made a Baserock system that ran and was only 28MB | 17:20 |
| ssam2 | that was done with artifact splitting | 17:20 |
| paulsher1ood | we've proved it can work... but we don't use it | 17:20 |
| paulsher1ood | i'm not saying we don't need it. just that in practice it is not used | 17:20 |
| ssam2 | ok, but i'm not sure that's a helpful comment since what matters is whether we need it :) | 17:21 |
| richard_maw | that's because the project it was going to be required for ended abruptly | 17:21 |
| paulsher1ood | anyway, i should getback to digesting richard_maw's example | 17:21 |
| pedroalvarez | artifact splitting is being used in the initramfs system, and we are using it | 17:21 |
| ssam2 | personally i'm very sad that our systems are 1.5GB, a good portion of which is debug info that could be stripped and supplied separately | 17:22 |
| pedroalvarez | so am I | 17:22 |
| paulsher1ood | ok, i was not aware of that. it's not in any of the systems i am using (devel, build, base, genivi) afaik | 17:22 |
| richard_maw | I suppose my main point re: runtime dependencies, is that both our models allow it, it's just that we're expressing it differently, and that I feel that paulsher1ood is conflating complexity of model with complexity of implementation. | 17:22 |
| paulsher1ood | richard_maw: you may be right | 17:23 |
| paulsher1ood | richard_maw: could you email your example to the list, so it doesn't get lost in traffic? | 17:24 |
| richard_maw | ok | 17:24 |
| paulsher1ood | tvm | 17:24 |
| richard_maw | though I think we got a bit sidetracked on that, rather than focussing on the idea of whether runtime-depends need to be in the model | 17:24 |
| paulsher1ood | well, if you send your example as reply to my email, it's part of the RFC i am seeking | 17:26 |
| bashrc | what's the preferred way of getting the kernel source? | 17:26 |
| richard_maw | for which purpose | 17:26 |
| paulsher1ood | richard_maw: you and others are welcome to proceed with runtime-depends anyway ... i think you have +2 | 17:27 |
| paulsher1ood | bashrc: assuming you're in definitions, 'morph edit linux-x86-64-generic' for example would work i think | 17:27 |
| richard_maw | if it's just to look at a `git clone` would work | 17:29 |
| paulsher1ood | that too :-) | 17:29 |
| richard_maw | if you then want to build it, then either `morph edit`, or clone locally and find where the system you're building has its definition for linux, and change the (repo, ref) to (file://path/to/local/clone, HEAD) | 17:30 |
| * paulsher1ood gets email from Amazon 'Can you answer this question from a fellow customer about Acer CB5... 'can you play sims 3/4 on this?' | 17:31 | |
| * richard_maw snorts | 17:32 | |
| richard_maw | perhaps the Android versions that aren't reall the sims 3 or 4, they're just branded the same | 17:32 |
| * rjek giggles. | 17:32 | |
| * paulsher1ood wants to tell them to answer their own support questions | 17:32 | |
| paulsher1ood | (expletives deleted) | 17:32 |
| rjek | heh | 17:33 |
| pedroalvarez | I wonder when should I do `mount -o loop` and when just `mount` | 17:33 |
| richard_maw | you could probably do it with wine, but you'd need to buy your own windows licence, for a device that can't run windows, for that to be legal AIUI | 17:33 |
| Kinnison | paulsher1ood: welcome to the electric monks | 17:33 |
| pedroalvarez | is the former to mount files and the latter to mount devices? | 17:33 |
| richard_maw | yep | 17:33 |
| Kinnison | pedroalvarez: in theory you need -o loop whenever you need to mount a file as a device | 17:33 |
| pedroalvarez | right | 17:33 |
| pedroalvarez | because both work in both cases | 17:34 |
| Kinnison | pedroalvarez: the 'loop' refers to /dev/loopN which is the kernel's "loopback" block device | 17:34 |
| richard_maw | though I think util-linux mount is smart about it | 17:34 |
| Kinnison | It is | 17:34 |
| pedroalvarez | it is | 17:34 |
| pedroalvarez | :) | 17:34 |
| Kinnison | pedroalvarez: please may I have my braincell back? | 17:34 |
| pedroalvarez | Kinnison: only 5 minutes more please | 17:34 |
| juergbi | richard_maw: why would you need a windows license for wine? or am i misunderstanding what you wrote? | 17:35 |
| pedroalvarez | then I guess I should support both cases even if we are using util-linux now | 17:35 |
| *** wdutch [~william@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Quit] | 17:36 | |
| richard_maw | juergbi: you understood what I wrote correctly, I don't know the details, but that's what I've been told. | 17:36 |
| Kinnison | pedroalvarez: So long as this time you brush the fluff and hair off before you return it | 17:36 |
| pedroalvarez | Kinnison: braincells with hair? | 17:37 |
| Kinnison | I have no idea what you did with it last time | 17:37 |
| pedroalvarez | heh | 17:38 |
| * pedroalvarez continues mounting things | 17:38 | |
| juergbi | richard_maw: afaik, it's perfectly legal to use wine without windows license - assuming you don't replace wine libraries with microsoft dlls from a windows installation | 17:38 |
| richard_maw | I think redistributable binaries provided by various applications are restricted to distribution for Windows. | 17:38 |
| juergbi | such restrictions may exist but i don't think that's a very common issue | 17:40 |
| juergbi | also, distribution license is not an issue if you install an application that has been distributed by someone else | 17:41 |
| richard_maw | heh, so wine's existance means that you could force EA to violate a redistributable license by not checking whether a game is being installed in wine | 17:41 |
| juergbi | it would probably be sufficient if they include the same restriction in their license agreement | 17:45 |
| juergbi | but i would expect it to be an installation restriction, not distribution restriction | 17:45 |
| *** ssam2 [~ssam2@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving] | 17:45 | |
| juergbi | anyway, there is not much i need wine for anyway | 17:46 |
| * richard_maw uses it occasionally for games when the Linux native ones aren't satisfying, and an application for configuring a switch | 17:47 | |
| *** Krin [~mikesmith@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Remote host closed the connection] | 17:56 | |
| *** bashrc [~motters@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Lost terminal] | 18:02 | |
| *** mariaderidder [~maria@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Ex-Chat] | 18:03 | |
| straycat | pedroalvarez, do you know anything about 6f46308dcc457b258b1f7f8f7b6eb2873d598de5 ? | 18:27 |
| straycat | (in definitions) | 18:27 |
| straycat | I rebased at some point and now I have it in my history, yet I don't see it in master | 18:28 |
| pedroalvarez | I can't access to that commit | 18:28 |
| pedroalvarez | what is the log message for it? | 18:28 |
| straycat | http://sprunge.us/KVih | 18:29 |
| jmacs | No, it's not there | 18:29 |
| pedroalvarez | ah, that might be in the latest release branch | 18:29 |
| straycat | so master and the release diverged? | 18:30 |
| straycat | *have | 18:30 |
| pedroalvarez | yes, the release normally need some fixes that can't be merged on time | 18:30 |
| *** tiagogomes [~tiagogome@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving] | 18:30 | |
| straycat | okay, should the change not be in master as well? | 18:31 |
| pedroalvarez | I think it is now in master | 18:31 |
| straycat | it isn't, unless you've just put it there | 18:31 |
| *** jonathanmaw [~jonathanm@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Quit: Leaving] | 18:32 | |
| straycat | i based off the release so i could build more quickly | 18:32 |
| pedroalvarez | straycat: http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/definitions.git/commit/?id=d3d894d1f06d64005086b4b0c15ba2e8350ffd33 | 18:32 |
| pedroalvarez | is not the same change, but equivalent | 18:32 |
| straycat | oh, fair enough | 18:32 |
| pedroalvarez | you can base your work off master now that we have mason populating the cache of 85.199.252.93 | 18:33 |
| *** locallycompact [~lc@82-70-136-246.dsl.in-addr.zen.co.uk] has quit [Ping timeout: 264 seconds] | 18:33 | |
| straycat | i wonder if it would be better to merge the release branch next time | 18:33 |
| straycat | pedroalvarez, yeah i didn't think of that | 18:33 |
| straycat | anyhow i can safely drop this commit now | 18:34 |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection] | 18:54 | |
| radiofree | hmm | 19:04 |
| radiofree | subprocess.check_call(['btrfs', 'subvolume', 'snapshot', orig_dir, run_dir]) is throwing an exception on my system :\ | 19:04 |
| radiofree | i ran the command manually and it worked | 19:04 |
| radiofree | ("'list' object is not callable",) oO | 19:10 |
| *** zoli_ [~zoli_@0x5e91887a.adsl.cybercity.dk] has joined #baserock | 19:11 | |
| *** zoli_ [~zoli_@0x5e91887a.adsl.cybercity.dk] has quit [Changing host] | 19:11 | |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 19:11 | |
| jmacs | Looks OK to me | 19:11 |
| radiofree | i didn't touch that line either, i *have* made some modifications to system-version-manager | 19:11 |
| radiofree | but... i don't see why it's failing there | 19:12 |
| radiofree | hmm maybe it's not there.. | 19:16 |
| radiofree | it's probably my fault, i'll try again tomorrow | 19:21 |
| radiofree | i bet this has something to do with this being a weird genivi image... | 19:23 |
| radiofree | i'll try with a devel image tomorrow | 19:23 |
| *** franred_ [~franred@host-80-41-248-121.as13285.net] has joined #baserock | 19:27 | |
| *** franred_ [~franred@host-80-41-248-121.as13285.net] has quit [Client Quit] | 19:27 | |
| *** cosm [~Unknown@host-78-150-56-250.as13285.net] has quit [Ping timeout: 240 seconds] | 20:17 | |
| *** cosm [~Unknown@cspc154.cs.man.ac.uk] has joined #baserock | 20:42 | |
| paulsher1ood | radiofree: could orig_dir or run_dir be something unexpected? | 21:38 |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection] | 21:38 | |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 21:52 | |
| radiofree | paulsher1ood: no i had them printed out | 22:06 |
| radiofree | it's almost certainly a change i made though | 22:06 |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection] | 22:07 | |
| paulsher1ood | same here :-) | 22:08 |
| radiofree | paulsher1ood: i was testing having it upgrade when /boot is on a different partition, not too much to go i think, should be able to send patches tomorrow for testing | 22:09 |
| paulsher1ood | cool | 22:10 |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 22:23 | |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection] | 22:29 | |
| *** zoli_ [~zoli_@linaro/zoli] has joined #baserock | 22:36 | |
| *** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection] | 22:36 | |
| radiofree | paulsher1ood: would requiring an sd card to to the initial flash be ok? | 22:39 |
| radiofree | that way we could use the entirely open-source tools available | 22:39 |
| * straycat looks at ybd | 22:55 | |
| straycat | (in case anyone was wondering) | 22:55 |
| * straycat is suddenly reminded of http://sackheads.org/~bnaylor/spew/away_msgs.html | 22:56 | |
| radiofree | young british designers? | 22:57 |
| straycat | no it's a prototype for a new morph that paulsher1ood has been working on | 22:58 |
| straycat | (yaml build deploy) | 22:58 |
| *** genii [~quassel@ubuntu/member/genii] has quit [Read error: Connection reset by peer] | 23:06 | |
| straycat | paulsher1ood, Do you think you could include some example uses on how its meant to be used? Or base it on cliapp so we can easily find help for the commands? | 23:25 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!