| *** edcragg has quit IRC | 00:06 | |
| paulsherwood | configure: error: Python module libxml2 is needed to run this package | 01:10 |
|---|---|---|
| paulsherwood | 3 15-10-19 00:00:43 [3/48/771] [itstool] ERROR: log file is at /src/artifacts/itstool.2f1d064084c15674eb5f68db9519a0d4ad2205cb173a098eba06808a7a46e841.build-log.3 | 01:10 |
| paulsherwood | (from master ci.morph using ybd) | 01:11 |
| *** gtristan has joined #baserock | 01:12 | |
| paulsherwood | gtristan: how near are you to lotte hotel world? | 01:13 |
| gtristan | paulsherwood, :) | 01:14 |
| gtristan | not near... about 1 hour away | 01:14 |
| paulsherwood | k. maybe we could do drinks/food later? | 01:14 |
| paulsherwood | any ideas on the itstool thing? | 01:15 |
| gtristan | I'm not sure which you're talking about... are you missing adapters ? | 01:16 |
| paulsherwood | i'm building ci.morph from definitions mastrer using ybd.. i wonder if something has changed recently to break the definition of itstool | 01:18 |
| paulsherwood | (last week this was working fine) | 01:19 |
| gtristan | hmmm | 01:19 |
| gtristan | no idea, I have a *ton* of unsubmitted patches, itstool is pretty low-level | 01:19 |
| gtristan | just recently hyphen/enchant & webp was added | 01:20 |
| gtristan | but that shouldnt have any effect | 01:20 |
| paulsherwood | np i can track this down | 01:20 |
| gtristan | paulsherwood, so around what time are you thinking, afternoon / evening ? | 01:33 |
| * gtristan is about to launch another attempt at WebKit... configures correctly but actually rendered my laptop completely unresponsive while building | 01:34 | |
| gtristan | for hours, and only got to 53% of the build :-/ | 01:34 |
| paulsherwood | gtristan: use AWS | 01:35 |
| paulsherwood | or post your branch and i can build it for you | 01:35 |
| paulsherwood | :) | 01:35 |
| gtristan | heh | 01:35 |
| * gtristan thinks the ram is the bottleneck, only 8GB on this dinasaur | 01:36 | |
| gtristan | and swapping like mad | 01:36 |
| paulsherwood | hmmm... that doesn't sound right to me | 01:36 |
| paulsherwood | i've successfully built webkit in < 1 hour with 4GB | 01:36 |
| gtristan | I recall having done so in the past too | 01:37 |
| gtristan | actually I recall because I had to patch gnu make (now its cmake) | 01:37 |
| gtristan | there was some hardwired memory limit | 01:37 |
| gtristan | well, I'll try to launch it again, maybe something else was to blame for that - I was asking for a lot | 01:38 |
| paulsherwood | we could meet later in the week... pedroalvarez will be here too. i expect he'll be way too tired today | 01:39 |
| gtristan | oh I don't mind the trek out there | 01:41 |
| paulsherwood | :) | 01:43 |
| gtristan | well, I would like to stop by there anyway, pedroalvarez will be exhausted and will need to sleep surely | 01:43 |
| gtristan | surely, I guess he's landing this evening | 01:43 |
| gtristan | ok, let me give this webkit build another shot | 01:44 |
| gtristan | finish my morning coffee injection :) | 01:44 |
| gtristan | and I'll swing by the area in the afternoon | 01:45 |
| paulsherwood | sounds good | 01:46 |
| gtristan | paulsherwood, where are you staying exactly again ? directly *at* lotte world near jasmil metro ? | 01:47 |
| paulsherwood | at lotte, it's rather luxurious :) | 01:49 |
| gtristan | I'll be underdressed I'm sure :) | 01:50 |
| paulsherwood | 893259d724707190f763a980bc055ad9c87f7535 broke itstool, according to git bisect | 01:51 |
| paulsherwood | gtristan: i doubt it | 01:51 |
| paulsherwood | http://paste.baserock.org/buxodenico | 02:25 |
| paulsherwood | so libxml2 is being installed in the sandbox, but i'm guessing this is something to do with python2 + python3 coexistence | 02:26 |
| gtristan | Odd, perhaps... I noticed that libxml2 is a part of core.morph, but XML-Parser is there too - I have no idea why the build failure though | 02:52 |
| * gtristan has not yet rebased the python upgrade | 02:52 | |
| * gtristan thinks python2 and python3 are intended to be parallel installable, though | 02:52 | |
| paulsherwood | ack | 02:52 |
| gtristan | I *guess* you probably need a different XML-Parser for python3 | 02:53 |
| paulsherwood | erk :) | 02:53 |
| gtristan | nod | 02:58 |
| gtristan | Imported from /srv/lorry/lorry-area/XML-Parser/XML-Parser-2.41.tar.gz. | 02:58 |
| gtristan | or that is a perl parser, eh... | 03:01 |
| gtristan | paulsherwood, do you have 1b486bd7e7e79583331648f322a52424fe27d297 ? | 03:03 |
| gtristan | paulsherwood, looks like jjardon fixed that issue | 03:03 |
| paulsherwood | aha! :) | 03:05 |
| paulsherwood | wow, that's a big rebuild | 03:06 |
| * gtristan is at 61% of WebKit and trucking along :-S | 03:08 | |
| gtristan | no lockups at least ! | 03:09 |
| paulsherwood | :) | 03:10 |
| paulsherwood | gtristan: strata/qt5-tools/qtwebkit.morph:max-jobs: 1 | 03:12 |
| paulsherwood | maybe we should fix that, forever | 03:13 |
| * paulsherwood wonders how many emgineer-weeks have been wasted by that setting | 03:14 | |
| * gtristan is building webkitgtk with 8 threads :) | 03:15 | |
| paulsherwood | how many cores in your machine? | 03:16 |
| gtristan | quad i7 | 03:16 |
| paulsherwood | do you think 8 is the right setting, then? | 03:16 |
| * paulsherwood doesn't understand cores vs threads here | 03:17 | |
| gtristan | i7 does 2 tasks per core | 03:17 |
| gtristan | with 4 cores I do 8 things at the same time | 03:17 |
| gtristan | so yeah, but I dont think it needs a setting, ybd conveniently detects that at the beginning of the build | 03:18 |
| paulsherwood | ok, so have you specified max-jobs: 8 in ybd.conf | 03:18 |
| gtristan | I did not specify anything | 03:18 |
| gtristan | probably cmake is doing what gnu make normally does with 'make -j' ... just take the amount of cores available | 03:19 |
| paulsherwood | interesting .... in python, what does multiprocessing.cpucount() tell you? | 03:19 |
| gtristan | I would need an explicit setting if I wanted to reserve jobs for doing something else | 03:19 |
| gtristan | I dont know... /me greps ybd sources | 03:20 |
| * gtristan suspects something like cat /proc/cpuinfo | something | 03:20 | |
| paulsherwood | (ybd has config['max-jobs'] = cpu_count() / config.get('instances', 1)}) | 03:20 |
| paulsherwood | i guess it' says 8 | 03:21 |
| gtristan | instances is basically, how many ybd's you have running at the same time I guess ? | 03:22 |
| paulsherwood | yup. | 03:22 |
| gtristan | indeed | 03:23 |
| paulsherwood | you might find that instances: 2 and max-jobs: 8 gives faster builds overall | 03:23 |
| gtristan | my build logs always say: [SETUP] Max-jobs is set to 8 | 03:23 |
| gtristan | hmmm, so wait... if I set that up... then ybd will parallelize builds of modules which dont depend on eachother ? | 03:24 |
| gtristan | all by itself ? | 03:24 |
| gtristan | that would definitely make faster builds in general | 03:24 |
| gtristan | many modules dont parallelize builds in their build scripts for one, and for another builds tend to burst in cpu, even if they do parallelize | 03:25 |
| gtristan | 74% :) | 03:25 |
| paulsherwood | gtristan: yup | 03:26 |
| gtristan | on the other hand, when reaching your target and there are no separate instances to run... you dont want to limit your remaining interdependent builds to be limited to half of your available jobs | 03:27 |
| gtristan | so probably some tweaking can be done there :) | 03:27 |
| paulsherwood | that's how i can get one machine to build all of ci.morph, from scratch with no artifacts, in 85 minutes | 03:27 |
| *** brlogger has joined #baserock | 03:48 | |
| gtristan | 99% :) | 03:55 |
| gtristan | 2hrs of webkit, eesh | 03:59 |
| gtristan | this machine would not build everything from scratch in 85min - getting clunky | 03:59 |
| gtristan | paulsherwood, so are you all set for connectors and gadgets and such ? | 04:00 |
| gtristan | I'm thinking to make a stop at yeongsan elecrtronics market on the way down | 04:00 |
| gtristan | anyway | 04:00 |
| gtristan | but I guess you've sorted out all your adaptors and such ? | 04:01 |
| *** gtristan has quit IRC | 04:25 | |
| *** toscalix__ has joined #baserock | 04:42 | |
| paulsherwood | i think we should be ok, thanks | 04:54 |
| *** toscalix__ has quit IRC | 05:08 | |
| *** zoli_- has left #baserock | 05:36 | |
| *** zoli_ has joined #baserock | 05:39 | |
| *** gtristan has joined #baserock | 06:40 | |
| jjardon | very soon we will build 3 different web engines: qtwebkit, webkitgtk and qtwebengine (based on blink, chromium fork of webkit); only a note for the challenges the ciat team need to solve | 06:57 |
| paulsherwood | :) | 06:58 |
| paulsherwood | jjardon: parallel instances of ybd | 06:58 |
| gtristan | jjardon, speaking of which... there is a lorry challenge I'm not quite sure how to handle from reading: http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/lorry.git/tree/README | 06:59 |
| gtristan | or anyone haha :) | 06:59 |
| gtristan | basically, the WebKit repository, where the WebKitGtk port lives, is here: https://svn.webkit.org/repository/webkit/ | 06:59 |
| gtristan | and what we want is the webkit-2.10 branch | 07:00 |
| gtristan | I was able to test the build by cloning the upstream git mirror | 07:00 |
| gtristan | and following these instructions: http://trac.webkit.org/wiki/WebKitGTK/2.10.x#Howtoaddawebkit-2.10branchtoexistinggit-svnclone | 07:00 |
| jjardon | gtristan: why only that branch? I guess we will need another branch in the next GNOME release | 07:01 |
| gtristan | which essentially fetch a specific version and then git svn rebase... I presume the lorry controller is using git-svn | 07:01 |
| gtristan | jjardon, good point | 07:01 |
| gtristan | the part that I dont understand, is how the lorry file should be written to handle the scheme in upstream webkit | 07:02 |
| jjardon | gtristan: if you are going to not lorry everything, I would be ok to lorry tarballs for now, if that makes life easier for you | 07:02 |
| gtristan | which is not only tags/ branches/ trunk/ ... but also has releases/ | 07:02 |
| gtristan | well, I could prepare a tarball lorry which we could fix later | 07:03 |
| jjardon | only remember to name the lorry repo <something>-tarball , so we can identify it easily | 07:03 |
| paulsherwood | :/ | 07:03 |
| jjardon | as IMHO is a bug to work with tarballs | 07:03 |
| * paulsherwood doesn't like that we keep settling for tarballs | 07:04 | |
| * jjardon neither | 07:04 | |
| gtristan | right, if we do settle for the tarball, we should name it as such, because then switching to a proper repo will not break... but are we capable of handling the svn ? | 07:05 |
| jjardon | paulsherwood: btw, 안녕하세요! :) | 07:05 |
| paulsherwood | jjardon: i'll take your word for it :) | 07:05 |
| rjek | ISTR there are two reasons for using tarballs: 1) there is no publically-facing VCS for a project, 2) GNU-style change logs murder git | 07:05 |
| jjardon | paulsherwood: https://storyboard.baserock.org/#!/story/10 | 07:05 |
| rjek | 2 seems surmountable with effort and engineering perhaps, 1 is unavoidable really | 07:05 |
| paulsherwood | it's 2 i'm not ok with | 07:06 |
| jjardon | rjek: 3. lorry cant manage to clone the repo (I think this is the gcc case) | 07:06 |
| paulsherwood | we could do shallow for that | 07:07 |
| paulsherwood | s/could/should/ | 07:07 |
| * paulsherwood is running out of battery | 07:07 | |
| * gtristan has a queue of 9 patches which basically separate some modules outside of strata/gnome/, and add strata/WebKitGtk.morph which depends on those common deps on those | 07:07 | |
| jjardon | paulsherwood: send patches :) | 07:07 |
| jjardon | gtristan: can you put those in a branch so I can take a look before you submit them? | 07:08 |
| rjek | jjardon: That's reason 2, IIRC | 07:08 |
| gtristan | well, I can, but they *also* depend on the unresolved lorry thing | 07:08 |
| gtristan | jjardon, also, what do you mean by "a branch", the regular git review topic branch ? | 07:09 |
| jjardon | gtristan: yes | 07:09 |
| jjardon | so we do not have to battle in gerrit too much :) | 07:09 |
| jjardon | gtristan: btw, I wanted to proposse you a challenge | 07:10 |
| gtristan | they should not be merged before the libwebp/enchant and hyphen definitions morphs land either, they depend on that | 07:10 |
| gtristan | jjardon, shoot :) | 07:10 |
| jjardon | GNOME 3.18.2 is going to be released the 11th of November, what about having a complete GNOME system by them so we can offer an baserock image to the community to try it? | 07:11 |
| gtristan | jjardon, we can probably make that date, I have a branch-scheme / CI proposal pending that I want to send / discuss on the ML | 07:12 |
| jjardon | to give you an idea, these are the modules part of the GNOME 3.18.1, as released by the GNOME RT: https://www.mirrorservice.org/sites/ftp.gnome.org/pub/GNOME/teams/releng/3.18.1/versions | 07:12 |
| gtristan | basically I'd like to see CI done both on master, _and_ on gnome-3.18 branches if at all possible | 07:13 |
| jjardon | gtristan: ok, lets try it; if not we can try with 3.19.1 | 07:13 |
| gtristan | getting the OS to build with everything should not be a problem | 07:13 |
| jjardon | gtristan: that would be awesome | 07:13 |
| gtristan | but sorting out what is stable and what is master right now needs thought | 07:13 |
| * gtristan has a writeup anyway, but we'll see how it works with what you guys have in place already | 07:14 | |
| *** ssam2 has joined #baserock | 07:15 | |
| *** ChanServ sets mode: +v ssam2 | 07:15 | |
| jjardon | ssam2: 08:11 <jjardon> GNOME 3.18.2 is going to be released the 11th of November, what about having a complete GNOME system by them so we can offer an baserock image to the community to try it? :) | 07:16 |
| jjardon | ssam2: I still do not see tracker in the GNOME stratum :) | 07:17 |
| ssam2 | that would be cool | 07:41 |
| ssam2 | it would be nice if the import tool could import from jhbuild so we could do that for every gnome release! | 07:43 |
| jjardon | ssam2: jhbuild tracks branches | 07:43 |
| jjardon | mostly | 07:43 |
| jjardon | we need a tool that takes the last tag for a given repo, or the last tag for a given repo and from a specific branch | 07:44 |
| jjardon | ssam2: does the import tool supports jhbuild yet? | 07:45 |
| ssam2 | no | 07:46 |
| ssam2 | finding the last tag for a given branch is really easy | 07:46 |
| ssam2 | `git describe --abbrev=0` (not that easy I guess, I had to look it up :) | 07:50 |
| *** 18WAAX2PD has joined #baserock | 07:51 | |
| jjardon | ssam2: yeah, but to be useful with our definitions maybe we have to use branches un the unpetrify-ref: field? Update would be very easy in that case | 07:52 |
| *** paulwaters_ has joined #baserock | 07:54 | |
| *** edcragg has joined #baserock | 07:55 | |
| ssam2 | jjardon: would be better to add a new 'tracking-branch' field or something like that | 07:56 |
| ssam2 | which we could actually document, etc :) | 07:56 |
| jjardon | ssam2: lets do it. is V7 already merged? | 07:56 |
| ssam2 | yeah, just waiting for release | 07:56 |
| jjardon | ah no, we need to do a release | 07:57 |
| *** 18WAAX2PD has quit IRC | 07:57 | |
| ssam2 | no reason we can't propose more changes though | 07:57 |
| jjardon | ssam2: what is needed to do a release? | 07:57 |
| jjardon | is the process documented somewhere? | 07:57 |
| ssam2 | http://wiki.baserock.org/guides/release-process/ | 07:57 |
| ssam2 | basically, new versions of the images at http://wiki.baserock.org/download need to be uploaded | 07:58 |
| ssam2 | usually the process involves lots of yak shaving though | 07:58 |
| ssam2 | e.g. http://mason-x86-32.baserock.org seems to be broken at the moment, so there's no x86_32 build, so that needs to be fixed 1st | 07:58 |
| ssam2 | and getting an arm build is still a pain | 07:59 |
| ssam2 | i would do one if I had a couple of days free to do it | 07:59 |
| paulsherwood | ssam2: i have lots of arm machines. how can i help? :) | 08:00 |
| ssam2 | paulsherwood: turn one into a Mason | 08:00 |
| jjardon | why not use the images mason is building a put them somewhere, then tag definitions and we are done? | 08:00 |
| jjardon | is there a mason for ARM? | 08:00 |
| ssam2 | jjardon: that's exactly what we want to do :) | 08:00 |
| paulsherwood | ssam2: why not ciat? | 08:01 |
| ssam2 | paulsherwood: i've never used it. does it work? | 08:01 |
| *** franred has joined #baserock | 08:01 | |
| ssam2 | it just needs to deploy the images | 08:01 |
| paulsherwood | it seems to, yes. wdutch is your man | 08:02 |
| ssam2 | and I guess produce morph-compatible artifacts... but we could punt on that if it produces working .img files | 08:02 |
| ssam2 | a ciat instance that can provide working builds of build-system-armv7lhf-jetson.img would be really helpful | 08:02 |
| paulsherwood | how do mason artifacts get to downloads.baserock.org? | 08:02 |
| ssam2 | by artifacts, do you mean the .img, .tar.gz files, or the actual chunk tar files? | 08:03 |
| ssam2 | the former are uploaded with scripts/release-upload | 08:03 |
| ssam2 | the latter don't go there, they appear on cache.baserock.org automatically, and get manually copied to git.baserock.org for releases | 08:03 |
| paulsherwood | why bother with the copy? | 08:04 |
| ssam2 | we've discussed before what's needed to do away with this silly dual-artifact-cache setup we have -- a way of cleaning up unwanted stuff from cache.baserock.org without removing old release artifacts | 08:04 |
| ssam2 | cache.baserock.org was a quick hack and still is | 08:04 |
| paulsherwood | ssam2: do you think we *need* to keep all chunks for releases, as opposed to just current release? | 08:05 |
| ssam2 | not sure | 08:05 |
| ssam2 | problem is still the same even if we only care about 1 release though | 08:05 |
| paulsherwood | ack | 08:05 |
| ssam2 | the files all look the same from `rm`'s point of view | 08:05 |
| * paulsherwood has been ponderign on this for ybd lately | 08:06 | |
| * wdutch has a look at build-system-armv7lhf-jetson | 08:06 | |
| paulsherwood | wdutch: could we move ciat to a cheaper box, while you're on? :) | 08:06 |
| ssam2 | jjardon: anyway, I like the idea of tracking-ref or something. It might overlap with what firehose is doing, though | 08:06 |
| ssam2 | jjardon: but probably better to have the info in definitions than in firehose's config files, if possible | 08:07 |
| paulsherwood | ssam2: ciat uses latst firehose, iiuc | 08:07 |
| *** bashrc has joined #baserock | 08:07 | |
| wdutch | paulsherwood: yes, I need to figure out how to redirect ciat.baserock.org to a new server after I move it | 08:07 |
| paulsherwood | btw what's the minimum solt to make a server program keep running after i logout of a box (something more elegant/reliable than screen) | 08:09 |
| paulsherwood | solution | 08:09 |
| rjek | Add code to daemonise | 08:10 |
| rjek | (detach from terminal and such) | 08:10 |
| ssam2 | systemd units are fairly easy to write | 08:10 |
| *** tiagogomes_ has joined #baserock | 08:10 | |
| ssam2 | e.g. http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/definitions.git/tree/extensions/mason/mason.service | 08:10 |
| paulsherwood | tvm | 08:11 |
| jjardon | who do I have to bribe to add the GNOME system to ciat? :) | 08:12 |
| * wdutch is open to bribery | 08:15 | |
| paulsherwood | jjardon: there's your answer | 08:15 |
| * jjardon opens the bribery chest | 08:18 | |
| jjardon | but lets fix the artifacts stuff first, so making a release is not as painful as its now! | 08:19 |
| jjardon | wdutch: how difficult would be that? any idea? | 08:19 |
| wdutch | jjardon: just for x86? | 08:20 |
| gtristan | jjardon, I'm pushing a topic branch called adding-webkitgtk | 08:21 |
| ssam2 | we currently release the BR reference systems for x86_32, x86_64 and armv7lhf (jetson) | 08:21 |
| jjardon | wdutch: that would be the first step, yes | 08:21 |
| gtristan | as mentioned, it wont build without other patches applied, and the tip, the WebKitGtk.morph points to a local repo, it needs the lorry first | 08:21 |
| jjardon | gtristan: cheers | 08:22 |
| gtristan | jjardon, so it's _only for review_, no merging ! | 08:22 |
| gtristan | hehe | 08:22 |
| jjardon | gtristan: ack :) | 08:22 |
| jjardon | gtristan: I was actually thinking on pushing a branch to the repo, not to gerrit :) dont worry, I will make my comments there | 08:23 |
| jjardon | ssam2: do you know if someone is working on activate cgit on gerrit? | 08:24 |
| gtristan | jjardon, right I dont have access for now | 08:24 |
| paulsherwood | l;et's take care of that? | 08:24 |
| gtristan | yay ! | 08:25 |
| * gtristan jumps in circles | 08:25 | |
| gtristan | cartwheels | 08:25 |
| * paulsherwood can confirm that this is exactly what is happening in real life | 08:25 | |
| paulsherwood | it's causing quite a stir in this posh hotel | 08:25 |
| ssam2 | jjardon: don't think anybody is actively working on that | 08:26 |
| *** mariaderidder has joined #baserock | 08:27 | |
| paulsherwood | i think we're trying to move away from gerrit if at all possible, but i may be wrong | 08:27 |
| ssam2 | first I've heard if so ... | 08:27 |
| jjardon | paulsherwood: are we? | 08:27 |
| jjardon | I though we are moving all the baserock components to gerrit, so we need to activate cgit in gerrit to easily see branches and stuff | 08:28 |
| jjardon | and keep git.baserock.org for deltas | 08:28 |
| * Kinnison doesn't understand the desire to migrate the projects off g.b.o | 08:29 | |
| radiofree | please don't :( | 08:30 |
| paulsherwood | i thought that straycat's git candidate stuff can be usedto work *withh* gbo | 08:30 |
| * paulsherwood crosses his fingers | 08:30 | |
| jjardon | gtristan: you are creating too many strata | 08:31 |
| jjardon | maybe we need to find a way to group them? | 08:33 |
| *** persia has quit IRC | 08:33 | |
| jjardon | Kinnison: contributors can not commit to g.b.o; and we tell them to clone the repos from gerrit, not g.b.o | 08:34 |
| *** persia has joined #baserock | 08:34 | |
| jjardon | Kinnison: if we activate cgit in gerrit we do not loss any functionality, I think | 08:35 |
| Kinnison | I'd rather just have the httpd on the gerrit server redirect to g.b.o | 08:35 |
| jjardon | Kinnison: contributor still can not see its branches | 08:36 |
| *** jonathanmaw has joined #baserock | 08:37 | |
| Kinnison | Do you mean the review refs? | 08:37 |
| jjardon | no, I mean normal branches | 08:38 |
| Kinnison | Why would we want arbitrary contributors pushing normal branches to gerrit? that sounds like a recipe for doom | 08:38 |
| jjardon | Kinnison: not arbitrary, gtristan for example | 08:39 |
| jjardon | Kinnison: we tell contributors to clone from gerrit but then we push to another different repo in g.b.o? It looks confusing to me | 08:40 |
| paulsherwood | 'nobody in the community is saying "i *love* how the strata are organised" | 08:59 |
| Kinnison | jjardon: mostly I think of the gerrit as a convenience function for handling candidates. Gerrit is one of the least-good git servers I know of so I wouldn't want to make it 'default' | 09:00 |
| Kinnison | jjardon: either way, g.b.o will need to have updates from the gerrit | 09:00 |
| Kinnison | jjardon: because other troves mirror from g.b.o | 09:00 |
| gtristan | ok, the requiring icu-common from graphics-common is probably safe to apply ;-) | 09:06 |
| *** franred has quit IRC | 09:07 | |
| *** CTtpollard has joined #baserock | 09:08 | |
| CTtpollard | I guess my raspi is in korea? | 09:08 |
| CTtpollard | oops wrong channel | 09:08 |
| paulsherwood | CTtpollard: this channel is fine... we have korean representatives here :) | 09:09 |
| paulsherwood | CTtpollard: do you want us to ook for it? | 09:09 |
| paulsherwood | look | 09:09 |
| gtristan | jjardon, lets get it to just build: tell me how you want the strata arranged and I can prepare a new branch based on that | 09:09 |
| CTtpollard | paulsherwood: I don't know if it will have any distinguishing features, no worries | 09:10 |
| paulsherwood | :) | 09:11 |
| *** nowster has joined #baserock | 09:11 | |
| gtristan | guys, can we at least get the ICU refactor out of the way ? https://gerrit.baserock.org/#/q/topic:share-icu-common | 09:12 |
| paulsherwood | merged | 09:15 |
| *** nowster_ has joined #baserock | 09:15 | |
| * tiagogomes_ will start reviewing patches soon | 09:18 | |
| *** nowster has quit IRC | 09:18 | |
| paulsherwood | i seem to have triggered some merge issues in gerrit... eg https://gerrit.baserock.org/#/c/1224/ | 09:26 |
| paulsherwood | can anyone help? | 09:26 |
| tiagogomes_ | paulsherwood have you tried to press the rebase button? | 09:29 |
| *** nowster__ has joined #baserock | 09:29 | |
| *** persia has quit IRC | 09:32 | |
| *** nowster_ has quit IRC | 09:32 | |
| *** persia has joined #baserock | 09:33 | |
| *** persia has quit IRC | 09:33 | |
| *** persia has joined #baserock | 09:33 | |
| ssam2 | you have to rebase, +2 again, then submit again | 09:37 |
| ssam2 | it's dumb | 09:37 |
| tiagogomes_ | yeah, that did the job | 09:38 |
| * gtristan looked for a "rebase button" | 09:39 | |
| *** nowster_ has joined #baserock | 09:39 | |
| tiagogomes_ | gtristan I think the rebase button to appear, you had to be signed in | 09:39 |
| * gtristan is signed in | 09:40 | |
| *** nowster__ has quit IRC | 09:42 | |
| gtristan | tiagogomes_, not allowed to rebase your own patch ! | 09:43 |
| gtristan | haha | 09:43 |
| gtristan | I have "Abandon" instead | 09:43 |
| paulsherwood | merged, for real this time | 09:43 |
| ssam2 | maybe rebase is restricted to people in Mergers group? | 09:44 |
| ssam2 | which doesn't really make sense | 09:44 |
| ssam2 | gerrit access controls are quite complex, it takes time to get them right | 09:44 |
| wdutch | can I not clone from ssh://git@git.baserock.org without permission? | 09:44 |
| jonathanmaw | paulsherwood: are you still interested in misc fixes to the GDP? I've been looking at the browser PoC, and while taking tpollard's patch to fix cursor trails is easy, catching up with master is trickier, since our baserock branch has changes which I had hoped to get merged into browser-poc, but weren't | 09:45 |
| paulsherwood | wdutch: ssh -A ? | 09:45 |
| ssam2 | wdutch: you need a gitano account on git.baserock.org, yes | 09:46 |
| ssam2 | try `ssh git@git.baserock.org whoami` | 09:46 |
| paulsherwood | jonathanmaw: i'm in your hands on this | 09:46 |
| tiagogomes_ | wdutch you can, but not using the SSH protocol | 09:46 |
| wdutch | CIAT clones from ssh urls everywhere | 09:47 |
| tiagogomes_ | wdutch, git clone git://git.baserock.org/baserock/baserock/morph should work anonymously | 09:47 |
| wdutch | I feel liek there is a reason it always uses ssh but I can't remember why | 09:47 |
| * tiagogomes_ can't comment on CIAT | 09:48 | |
| SotK | pushing candidate branches maybe (idk if it actually does)? | 09:48 |
| jonathanmaw | paulsherwood: I'll see if the only problem is the unmerged build-system patches, and see if I can chase the maintainer. | 09:48 |
| paulsherwood | jonathanmaw: one thing i'm interested in is how to make gdp containg agl :) | 09:48 |
| SotK | wdutch: I guess CIAT could be given a gitano account? | 09:49 |
| jonathanmaw | paulsherwood: that's an interesting question which requires defining "agl", hopefully in a way that doesn't conflict with the GENIVI specification | 09:49 |
| gtristan | jjardon, ummm: https://gerrit.baserock.org/#/c/1258/ | 09:54 |
| paulsherwood | jonathanmaw: i agree | 09:54 |
| gtristan | jjardon, how come we are building libxml2 twice ? | 09:54 |
| gtristan | does the python thing not depend on core.morph and naturally pull in libxml2 ? | 09:55 |
| gtristan | if not, isn't the sensible thing to split out libxml2 from core.morph and share it ? | 09:55 |
| *** mariaderidder has quit IRC | 09:56 | |
| *** nowster__ has joined #baserock | 09:56 | |
| paulsherwood | +1 | 09:58 |
| *** nowster_ has quit IRC | 09:59 | |
| *** nowster_ has joined #baserock | 10:00 | |
| *** nowster__ has quit IRC | 10:04 | |
| wdutch | I've put the keys from the old ciat into the new one and it's still saying permission denied | 10:08 |
| wdutch | curious... | 10:08 |
| wdutch | oh because we were using a different trove | 10:08 |
| wdutch | derp | 10:08 |
| *** nowster__ has joined #baserock | 10:10 | |
| wdutch | who can add a ciat key for me? | 10:11 |
| wdutch | if I ask nicely O:) | 10:11 |
| paulsherwood | gary_perkins: ? | 10:11 |
| jjardon | gtristan: not sure we can do that | 10:11 |
| jjardon | components in core depends on libxml2, so you will have to move those as well | 10:12 |
| paulsherwood | jjardon: only libxslt iiuc | 10:12 |
| *** nowster_ has quit IRC | 10:12 | |
| *** mariaderidder has joined #baserock | 10:13 | |
| jjardon | paulsherwood: yep, but then you have to check why those components are there to not break other statra: it can be done but it will take time, but happy if gtristan wants to spend time on it | 10:15 |
| gtristan | jjardon, python2-core *already* depends on core.morph | 10:17 |
| gtristan | it seems that libxml2 is entirely unneeded there | 10:17 |
| gtristan | aha ! | 10:18 |
| gtristan | jjardon, that needs thought indeed | 10:19 |
| gtristan | jjardon, I wonder if libxml2 builds something different when python2 is present, as compared to python3 | 10:20 |
| gtristan | or if it builds something for each of them | 10:20 |
| gtristan | perhaps we just need libxml2 to sit on top of both pythons for everything to work out | 10:21 |
| gtristan | anyway, we're shutting down and will see if the others arrive | 10:22 |
| * gtristan shuts down | 10:22 | |
| *** gtristan has quit IRC | 10:23 | |
| ssam2 | wdutch: are you talking about access to git.baserock.org accounts? | 10:25 |
| ssam2 | wdutch: if you have the old key, you can use it to add the new key | 10:25 |
| ssam2 | wdutch: cat new_key.pub | ssh -i old_key.pub git@git.baserock.org sshkey add new_key | 10:25 |
| ssam2 | (not tested) | 10:25 |
| wdutch | I shall try that | 10:26 |
| *** nowster_ has joined #baserock | 10:26 | |
| wdutch | but that will add it as me, maybe as SotK said CIAT should have an account, does this matter at all? | 10:26 |
| *** franred has joined #baserock | 10:27 | |
| *** nowster__ has quit IRC | 10:29 | |
| *** nowster_ has quit IRC | 10:30 | |
| *** mariaderidder has quit IRC | 10:31 | |
| *** franred has quit IRC | 10:47 | |
| jjardon | gtristan, we build againt pythyn2 to build the python2 bindings needed by itstool | 10:54 |
| jjardon | and we build libxml in core because probably something in foundations needs libxml2 | 10:54 |
| *** franred has joined #baserock | 10:54 | |
| *** nowster has joined #baserock | 10:58 | |
| ssam2 | wdutch: if you do it with the old ciat key, instead of yours, then it should do it as the ciat account | 11:11 |
| ssam2 | you could do: `ssh -i old_ciat_key.pub git@git.baserock.org whoami` first to check what account it thinks you are | 11:12 |
| wdutch | ssam2: there is no ciat account afaik, we used a different trove in the past | 11:12 |
| ssam2 | ah, ok. so what you need is an account | 11:12 |
| ssam2 | i can do that | 11:12 |
| ssam2 | pm me the public key | 11:12 |
| ssam2 | sorry, missed that it didn't have an existing account on git.baserock.org | 11:13 |
| *** nowster has quit IRC | 11:32 | |
| *** franred has quit IRC | 12:42 | |
| *** franred has joined #baserock | 12:58 | |
| * pedroalvarez waves | 13:05 | |
| *** nowster has joined #baserock | 13:15 | |
| *** franred has quit IRC | 13:32 | |
| *** brlogger has joined #baserock | 13:57 | |
| *** locallycompact has joined #baserock | 13:59 | |
| jonathanmaw | hrm, cherry-picking only the buildsystem patch did not result in a working browser-poc, rebasing did, however. Any objections to me pushing a branch "baserock/jonathanmaw/genivi-demo-platform-2" and pushing a modified branch of definitions that uses it? | 15:09 |
| *** locallycompact has quit IRC | 15:17 | |
| *** locallycompact_ has joined #baserock | 15:17 | |
| *** franred has joined #baserock | 15:25 | |
| * jonathanmaw is not sure whether tacit approval applies :/ | 15:27 | |
| * radiofree has never sought approval when pushing personal branches to gbo | 15:29 | |
| ssam2 | i have no objection to you pushing a branch, if that helps | 15:30 |
| radiofree | i think it's ok (at least no one has commented on it) | 15:30 |
| jonathanmaw | done! there's a branch "baserock/jonathanmaw/k1.0-fixes" in definitions, which has a few more fixes to browser-poc. The most notable is that you don't see cursor trails before the browser finishes loading a page | 15:35 |
| CTtpollard | :) | 15:36 |
| jjardon | jonathanmaw: definitions in g.b.o or in gerrit? | 16:08 |
| jonathanmaw | jjardon: g.b.o | 16:10 |
| *** rjek has quit IRC | 16:29 | |
| *** tlsa has quit IRC | 16:29 | |
| *** sebh has quit IRC | 16:29 | |
| *** tlsa has joined #baserock | 16:30 | |
| *** rjek has joined #baserock | 16:30 | |
| *** paulwaters_ has quit IRC | 16:31 | |
| *** nowster has quit IRC | 16:33 | |
| *** sebh has joined #baserock | 16:34 | |
| *** jonathanmaw has quit IRC | 16:36 | |
| *** bashrc has quit IRC | 17:04 | |
| *** brlogger has joined #baserock | 17:29 | |
| *** edcragg has quit IRC | 17:30 | |
| *** franred has quit IRC | 17:36 | |
| *** locallycompact_ has quit IRC | 17:57 | |
| *** paulwaters_ has quit IRC | 18:05 | |
| *** ssam2 has quit IRC | 18:06 | |
| *** toscalix__ has joined #baserock | 20:51 | |
| *** edcragg has joined #baserock | 21:08 | |
| *** toscalix__ has quit IRC | 21:41 | |
| *** toscalix__ has joined #baserock | 21:55 | |
| *** edcragg has quit IRC | 22:14 | |
| *** toscalix__ has quit IRC | 22:28 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!