*** jlrmagnus has quit IRC | 00:07 | |
*** mdurnev has joined #automotive | 00:09 | |
*** chbae has quit IRC | 00:20 | |
*** AlisonChaiken has quit IRC | 00:41 | |
*** tjamison has left #automotive | 00:58 | |
*** jackjack has joined #automotive | 04:12 | |
*** jackjack has left #automotive | 04:29 | |
*** Tarnyko2 has quit IRC | 05:10 | |
*** jobol has joined #automotive | 05:23 | |
*** DSVenky has quit IRC | 05:23 | |
*** mdp has quit IRC | 05:25 | |
*** leon-anavi-cloud has quit IRC | 05:25 | |
*** Tartarus has quit IRC | 05:25 | |
*** albanc has quit IRC | 05:25 | |
*** jjardon has quit IRC | 05:25 | |
*** FelixH has quit IRC | 05:25 | |
*** leon-anavi-cloud has joined #automotive | 05:26 | |
*** FelixH has joined #automotive | 05:26 | |
*** DSVenky has joined #automotive | 05:28 | |
*** mdp has joined #automotive | 05:31 | |
*** albanc has joined #automotive | 05:32 | |
*** jjardon has joined #automotive | 05:33 | |
*** Tartarus has joined #automotive | 05:34 | |
*** leon-anavi has joined #automotive | 06:37 | |
leon-anavi | morning | 06:38 |
---|---|---|
*** sanjeev has quit IRC | 06:42 | |
*** rajm has joined #automotive | 07:15 | |
CTtpollard | morning | 07:22 |
*** CTtpollard has quit IRC | 07:24 | |
*** fredcadete has joined #automotive | 07:28 | |
*** CTtpollard has joined #automotive | 07:28 | |
*** toscalix has joined #automotive | 07:53 | |
*** sanjeev has joined #automotive | 08:06 | |
*** jani_ is now known as jani | 08:16 | |
*** sanjeev has quit IRC | 08:18 | |
jani | good morning.. | 08:18 |
jani | hey, is that qt ivi project announchement stufff you guys have been working on or do they have their own project ? | 08:20 |
*** sanjeev has joined #automotive | 08:26 | |
JEEB | some people here develop their solutions with Qt, but I don't think we've seen too many Qt people around here | 08:29 |
jani | yeah, i just saw agl mentioned and mistake or not, thought it was associated with most of you guys .. and the fact that demo ivi was qt .. | 08:31 |
*** Tarnyko has joined #automotive | 08:31 | |
leon-anavi | jani, as far as I know for the moment Qt is involved in both GDP and AGL. | 08:32 |
jani | check. | 08:39 |
jani | dunno if i ever mentioned this but company i work for is doing similar stuff as qt for automotive field. | 08:40 |
leon-anavi | jani, which is this company? | 08:43 |
*** jonathanmaw has joined #automotive | 08:46 | |
CTtpollard | does anyone know how AGL (AGL Demo) CI handles testing patchsets/changes that requires changed to both agl-demo and meta-agl or the agl 'repo' at the same time? | 08:47 |
CTtpollard | It's a problem I want to tackle with meta-genivi-dev and genivi-dev-platform | 08:49 |
CTtpollard | leon-anavi: when you added sota-client to AGL this would be one of the situations | 08:52 |
*** fredcadete has quit IRC | 09:29 | |
*** fredcadete has joined #automotive | 09:37 | |
leon-anavi | hi CTtpollard, when a patch is submitted to AGL through Gerrit, the AGL Jenkins in Gerrit users gives it Code-Review-1. If the build is successful AGL Jenkins gives Code-Review+1 | 09:40 |
leon-anavi | so a jenkins build is started for each new patch | 09:40 |
leon-anavi | you can monitor the build statuses here: https://build.automotivelinux.org/ | 09:41 |
CTtpollard | leon-anavi: but how does it handle patch to two different repositories that need to be built together? so in your case the commit to add sota recipes, and then the commit to the 'repo' to add meta-rust? | 09:41 |
leon-anavi | I am not sure if the problem that you are tacking is solved in AGL | 09:41 |
leon-anavi | You better ask Stephane or someone else from iot.bzh regarding this | 09:42 |
CTtpollard | oki doke | 09:42 |
CTtpollard | :) | 09:42 |
leon-anavi | btw Tarnyko, can you share some details related to AGL jenkins setup? | 09:42 |
CTtpollard | we could not test PR's to meta-genivi-dev, and only test PR's to genivi-dev-platform. But if a problem was found at that stage it could mean having to revert the commits | 09:43 |
leon-anavi | right | 09:48 |
leon-anavi | btw the testing should have another step if the package is not included by default in a certain image. | 09:48 |
leon-anavi | the build system should add it to local.conf as IMAGE_INSTALL_append | 09:49 |
CTtpollard | indeed | 09:51 |
CTtpollard | currently we don't carry an optional packages as such, but we did for sota/rvi core | 09:52 |
fredcadete | CTtpollard: from my Jenkins experience, the only way I imagine that to work is to setup a build that "follows" meta-genivi-dev and just happens to build it in the environment of genivi-dev-platform | 09:55 |
fredcadete | it's quite possible to make a SCM setup where meta-genivi-dev is taken as one Git SCM, and genivi-dev-platform is taken as another with the sparse-checkout option to *not* checkout meta-genivi-dev | 09:56 |
leon-anavi | CTtpollard, btw one of the things that I really enjoy in AGL is that there are several different images: agl-image-minimal, agl-image-ivi, agl-demo-platform. | 09:56 |
leon-anavi | I think they just inherit each other therefore the maintenance effort should not be that big. | 09:57 |
leon-anavi | It is very similar to what we have in Poky and in my opinion it is quite convenient for development | 09:57 |
CTtpollard | fredcadete: We do that now, all PR's for meta-genivi-dev are taken as a separate checkout against master of genivi-dev-platform, for each target | 09:58 |
leon-anavi | I mentioned this approach as a proposition for GDP during the open call several weeks ago but I understood that the majority of other people involved in the project don't like it much :) | 09:58 |
CTtpollard | leon-anavi: we have the -ivi image included from meta-ivi, but no differential for a -minimum yet | 09:59 |
leon-anavi | yes | 09:59 |
CTtpollard | I agree it would be nice to have, but I don't think GDP is big enough yet to justify splitting | 10:00 |
leon-anavi | ok | 10:00 |
*** rajm has quit IRC | 10:00 | |
CTtpollard | but, that's just my opinion | 10:00 |
CTtpollard | fredcadete: I need to create a pipeline/system in which it can somehow link two github PR's across two repo's together, and build them in parallel | 10:01 |
CTtpollard | and instance of this is the current PR's to adapt meta-ivi10 | 10:02 |
fredcadete | CTtpollard: that would be tricky. why is that necessary? | 10:02 |
CTtpollard | it requires changes to recipes in meta-genivi-dev to support ivi10/yocto2.0, and changes to genivi-dev-platform to update the submodules/layers to jethro branches. To be built together | 10:02 |
CTtpollard | if either at tested separately, they fail due to the dependency on each other | 10:03 |
fredcadete | oh, I see, this one? https://github.com/GENIVI/genivi-dev-platform/pull/14 | 10:05 |
CTtpollard | so, we could just merge the meta-genivi-dev changes without testing, and then only test the PR which tells genivi-dev-platform to use this sha, as well as the other changes I listed before | 10:05 |
CTtpollard | but, it's not nice that way | 10:05 |
fredcadete | this is certainly not an automotive-specific problem | 10:06 |
fredcadete | but I don't know of a solution | 10:06 |
CTtpollard | fredcadete: yep that PR | 10:06 |
CTtpollard | another instance of this situation can be seen here: https://github.com/GENIVI/genivi-dev-platform/pull/6 | 10:07 |
CTtpollard | an update of meta-raspberrpi had to be built with an update of meta-genivi-dev at the sametime | 10:07 |
CTtpollard | The only real solution I've been able to see aiding this situation is merged both repo's together | 10:08 |
CTtpollard | so meta-genivi-dev is not an external submodule | 10:08 |
CTtpollard | that would be my eventual proposal for the GDP | 10:11 |
fredcadete | that would solve the issue when the interdependency is meta-genivi-dev<->other, not between other<->other. But that's maybe enough for GDP | 10:13 |
CTtpollard | if it's other-other, those are would be existing sha's that we can point to | 10:14 |
*** rajm has joined #automotive | 10:14 | |
CTtpollard | I think it's enough for what GDP can realistically aim for | 10:16 |
fredcadete | that would probably be much easier than trying to "link" PRs to different repos | 10:16 |
CTtpollard | :) | 10:18 |
toscalix | leon-anavi: size is one thing, the fact that we have Baseline is another one | 10:32 |
*** rajm has quit IRC | 10:33 | |
*** rajm has joined #automotive | 10:33 | |
*** clopez has quit IRC | 10:33 | |
toscalix | we are concentrating first in moving fast towards shipping talest software and providing a major release with several targets | 10:34 |
toscalix | when we achieve both goals, we can increase the offering | 10:34 |
leon-anavi | ok | 10:34 |
leon-anavi | sounds as a plan | 10:34 |
toscalix | Did you read the document I sent to the project mailing list about GDP viewed as portfolio? | 10:34 |
leon-anavi | toscalix, sorry, I haven't read it yet. | 10:35 |
toscalix | that is an attempt to describe target -> deliverable | 10:35 |
leon-anavi | I busy with some other tasks but I will read it by the end of the week. | 10:35 |
toscalix | ok, please let me know what you think about it | 10:35 |
toscalix | I thought about the idea of "flavors" | 10:35 |
toscalix | but based on my experience.... I think we need to prioritise some other things first | 10:36 |
toscalix | having Baseline in place | 10:36 |
toscalix | what I like about AGL is that all those flavours come from the same "trunk" | 10:37 |
toscalix | that is the most efficient way of dacing the challenge involved in supporting them | 10:37 |
*** clopez has joined #automotive | 10:37 | |
toscalix | s/dacing/facing | 10:37 |
toscalix | my talk at 14th AMM about GDP portfolio was in that direction | 10:38 |
leon-anavi | yes | 10:38 |
leon-anavi | I like this in AGL too | 10:39 |
CTtpollard | toscalix: the structure of AGL is virtually mirrored in GDP | 10:39 |
CTtpollard | meta-agl = meta-ivi, agl-demo = gdp | 10:39 |
toscalix | as output, yes | 10:39 |
CTtpollard | building on top of poky releases, using OE | 10:39 |
toscalix | we deffer in how we deliver it | 10:39 |
toscalix | differ | 10:40 |
Tarnyko | leon-anavi: I do not really knox the Jenkins setup, sorry for that | 10:59 |
leon-anavi | Tarnyko, ok. CTtpollard was asking about it? Could you point him in the right direction to someone with experience in it? | 11:00 |
Tarnyko | this was the initial setup notice : https://lists.linuxfoundation.org/pipermail/automotive-discussions/2015-September/000822.html . I guess Jan-Simon is the good person to ask about | 11:05 |
CTtpollard | thanks all | 11:07 |
*** gunnarx has joined #automotive | 11:51 | |
*** gunnarx has joined #automotive | 11:51 | |
gunnarx | CTtpollard, read your email now | 11:53 |
gunnarx | ^^ I have read | 11:53 |
gunnarx | English is an ambiguous language in writing :) | 11:54 |
gunnarx | CTtpollard, can you explain more about "It also can't handle the parallel PR dependency, such as can be seen with the 'GDP10' situation"? | 11:54 |
*** praneeth has quit IRC | 12:06 | |
*** praneeth has joined #automotive | 12:06 | |
CTtpollard | gunnarx: been discussing the in here today, but can quote an example | 12:14 |
CTtpollard | '<CTtpollard> and instance of this is the current PR's to adapt meta-ivi10 | 12:14 |
CTtpollard | <fredcadete> CTtpollard: that would be tricky. why is that necessary? | 12:14 |
CTtpollard | <CTtpollard> it requires changes to recipes in meta-genivi-dev to support ivi10/yocto2.0, and changes to genivi-dev-platform to update the submodules/layers to jethro branches. To be built together | 12:14 |
CTtpollard | <CTtpollard> if either at tested separately, they fail due to the dependency on each other ' | 12:14 |
CTtpollard | this situation has also occurred for example in the upgrade of the raspi kernel recently, and the addition of rvi sota client | 12:15 |
gunnarx | sorry, I can go back read the logs. been off IRC for a little while | 12:15 |
CTtpollard | np | 12:15 |
CTtpollard | I believe agl and potentially other yocto system using submodules/google-repo will have this problem, but I've yet to find out how they handle it | 12:16 |
gunnarx | ah yes, now I understand. Yes changes that need to be done in m-g-d and g-d-p simultaneously? | 12:16 |
CTtpollard | yup | 12:17 |
CTtpollard | so, you can seen on the genivi.go that the pipelines we've introduced to do m-g-d(PR) against master g-d-p fail, because they require the PR in g-d-p to bring in dependencies | 12:18 |
gunnarx | after thinking about it 30 seconds only... yeah maybe the PR can't be built automatically. | 12:18 |
gunnarx | I mean in fact you can't create a g-d-p PR until you have a commit hash in m-g-p to refer to | 12:18 |
CTtpollard | yes unless the github api/go.cd plugin has a smart way to link PR's across repos | 12:19 |
CTtpollard | which I highly doubt | 12:19 |
gunnarx | so there's some working around there anyhow. at first you need your submodule change accepted, (alternatively provide a fork) then you can propose a submodule change to point to that and that can be built (automatically) | 12:20 |
gunnarx | A special incoming branch on m-g-d that anyone can commit to. Then you havea commit hash and you can point g-d-p PR to that. It will be built. The PR can be merged on g-d-p, then m-g-d updated. It will only work if the merge is a fast forward so the hash doesn't change.... | 12:22 |
gunnarx | m-g-d updated means merging the commit over to master which as I said cannot be an actual merge as such. Fun times for maintainers :) | 12:23 |
gunnarx | OK, so regarding PR integration. https://github.com/GENIVI/meta-genivi-dev/pull/7 seems to be built here: https://go.genivi.org/go/tab/build/detail/GDP-Yocto-RPI2-meta-genivi-dev-GITHUB-PR/1/build/1/init_and_bitbake agreed? | 12:25 |
gunnarx | I think reporting the result back to GitHub is the failing step. | 12:26 |
gunnarx | CTtpollard, are you ignoring me or busy trying to make sense of my rambling :) | 12:28 |
gunnarx | I am not offended either way :) | 12:29 |
CTtpollard | gunnarx: no ignoring, just currently on lunch :) | 12:29 |
gunnarx | lunch is important | 12:30 |
CTtpollard | very cheap supermarket sushi, but it's not all that bad | 12:30 |
gunnarx | If I want the PR build to fail as fast as possible, do you have a suggestion of a file in meta-genivi-dev to screw up? :) | 12:36 |
gunnarx | I'm not sure what is read first from the layer... | 12:37 |
CTtpollard | gunnarx: I usually just add a dummy file to test pr's | 12:37 |
CTtpollard | going back to parse your comments now | 12:38 |
gunnarx | yes but I want the build to fail also. I'll put something in layer.conf, I suppose that should work | 12:38 |
*** Saint_Isidore has quit IRC | 12:39 | |
CTtpollard | ah ok then, just bork a file path in layer/bblayers | 12:39 |
*** Saint_Isidore has joined #automotive | 12:40 | |
gunnarx | hmm, was it meta-genivi-dev that was deleted before? | 12:41 |
CTtpollard | gunnarx: the way works, but as you point out it requires the shas to match, meaning m-g-d has to stay stagnant during the PR process. This happens a lot, and I usually just amend, or ask the submitter to rebase the PR and amend | 12:43 |
CTtpollard | again, and example of that: https://github.com/GENIVI/genivi-dev-platform/pull/13 | 12:43 |
gunnarx | the forks that were created in GitHub are no longer connected to GENIVI/m-g-d because of its deletion. They are not part of the GENIVI network. My fork suddenly derives from toscalix/m-g-d :-) | 12:45 |
CTtpollard | once the meta-genivi-dev side was merged, the sha in g-d-p PR did not match due to commits in the meantime to m-g-d | 12:45 |
toscalix | wow, I am upstream :-) | 12:45 |
gunnarx | Yes I see, so it's a common thing then. I guess the process kind of works then if wanted to set it up that way. | 12:45 |
gunnarx | Everyone knows toscalix is the definitive version of GDP | 12:45 |
toscalix | :-) | 12:46 |
gunnarx | The downside is I can't seem to create PRs back to the GENIVI copy. GitHub doesn't know any connection to that repo apparently. | 12:47 |
CTtpollard | ..... | 12:47 |
CTtpollard | Nicr | 12:47 |
CTtpollard | *e | 12:47 |
pedroalvarez | so, problem found | 12:48 |
pedroalvarez | (that's good) | 12:48 |
gunnarx | Question or statement pedroalvarez ? | 12:48 |
pedroalvarez | celebration | 12:48 |
gunnarx | :? | 12:48 |
pedroalvarez | I mean, we know what the problem is, that's good | 12:48 |
gunnarx | We are 6 people in a private network on meta-genivi-dev | 12:48 |
gunnarx | Which problem? | 12:48 |
gunnarx | You think this affects the PR builds you mean? | 12:49 |
pedroalvarez | oh, I thought you were thinking that | 12:49 |
CTtpollard | gunnarx: all that I can deem, taking AGL for example (all be it that's using google-repo,jenkins,gerrit but same concept) is that changes to agl-demo don't get tested, it only gets tested at the point when 'submodule' sha is updated in the 'repo'(g-d-p in our instance) side | 12:49 |
gunnarx | don't know if there's a connection to that yet | 12:49 |
gunnarx | CTtpollard, makes sense | 12:49 |
CTtpollard | but that can lead to having to revert changes in 'demo' if they don't work once the testing is carried out at the repo/gdp stage | 12:50 |
pedroalvarez | gunnarx: sorry, totally misunderstood the conversation | 12:51 |
gunnarx | pedroalvarez, FYI you are part of the "disconnected" network :) Doesn't matter now but I will re-fork the repo from GENIVI later on. | 12:51 |
gunnarx | np pedroalvarez, we will get to the PR discussion also, I'm juggling a few different things | 12:51 |
CTtpollard | So if people are happy with that model of CI then I'm happy to trial it, although I still believe combining the repo's is the ultimate solution | 12:51 |
gunnarx | Hmm, I'm quite surprised I can't even by typing in the name force GitHub to consider a pull request from my copy... Can PRs only be made within the known "network" of forks...? | 12:53 |
gunnarx | meta-genivi-dev has not changed the hash of its base commit or anything weird right? It's still the same content? | 12:54 |
CTtpollard | should be | 12:57 |
jani | leon-anavi: rightware | 12:59 |
jonathanmaw | Does anyone here have experience with booting the renesas Silk? I've got it loading the kernel and dtb, but nothing after "Starting kernel ...", which seems to be a problem with the kernel's tty interface. | 13:06 |
gunnarx | FYI I temporarily replaced ct_test resource with yocto_build resource in the PR pipeline, so another agent can be used | 13:08 |
CTtpollard | gunnarx: ok, we also added the renesas binaries to the codethink agent so it should be part of the pool for those pipelines | 13:12 |
gunnarx | Pending check! https://github.com/GENIVI/meta-genivi-dev/pull/10 | 13:13 |
CTtpollard | woop | 13:15 |
gunnarx | plot thickens | 13:15 |
gunnarx | bugs galore... | 13:15 |
gunnarx | A paused pipeline is not paused for PRs. PRs will trigger it anyhow? wth. | 13:15 |
pedroalvarez | so all of them have been triggered, but only the non-paused one has actually comunicated "check pending" to github | 13:18 |
gunnarx | yes | 13:20 |
gunnarx | might be a coincidence though. maybe first one wins on the reporting | 13:20 |
gunnarx | Might be some weirdness going on. It's a little hard with the pipelines that have multiple material - they will trigger on either repo changing. These 4 seem to be PR triggered though AFAICT | 13:21 |
gunnarx | Anyhow, I think we're back on track. I just wanted to confirm first. The gocd user seems to need write access to the repo to report on checks. The pipelines are triggered with only read access, but no indication inside GitHub without write. | 13:22 |
gunnarx | Go itself only has a OAUTH token though, so there are some limits on what it can actually do | 13:23 |
CTtpollard | if it only needs write to report, then it should be granted? Unless it has capabilities to take successful PR's as triggers to auto-merge like some tools then no | 13:24 |
gunnarx | no parse | 13:25 |
gunnarx | The username is a collaborator with Write access now (it was Read). | 13:26 |
gunnarx | Go server logs in with username and token, so the scope of the token itself also limits | 13:27 |
CTtpollard | If it just needs write access to report on success/failure then fine, I'm just aware other 'gatekeeper' type workers also can take success==merge | 13:27 |
CTtpollard | ok then | 13:27 |
gunnarx | OK I see what you mean. We don't have any auto merge functionality, no. | 13:29 |
gunnarx | Hmm, this thing single PR triggering multiple pipelines might still need some work. | 13:31 |
*** gunnarx has quit IRC | 13:39 | |
*** gunnarx has joined #automotive | 13:39 | |
pedroalvarez | can we un=pause them now? or will you sort that out, gunnarx? | 13:47 |
gunnarx | just a sec | 13:48 |
gunnarx | OK, yeah I guess it built, failed as expected, and reported the failure. In other words successful PR check. | 13:49 |
gunnarx | OK, I unpaused them again | 13:50 |
CTtpollard | Failure==Success \o/ | 13:50 |
gunnarx | exactly | 13:50 |
gunnarx | still not sure where we stand with multiple pipelines triggered by the same PR | 13:50 |
gunnarx | looks like only one is being handled and reported. I guess we keep testing some more | 13:51 |
CTtpollard | yeh, ultimately we'd want it to report all targs | 13:51 |
pedroalvarez | I was hoping that is what happens | 13:51 |
gunnarx | No wait, look https://github.com/GENIVI/meta-genivi-dev/pull/10 | 13:51 |
gunnarx | 4 checks! | 13:52 |
CTtpollard | :D | 13:52 |
pedroalvarez | purrrfect | 13:52 |
gunnarx | I guess we have been accustomed to being too negative. It works as intended apparently | 13:52 |
gunnarx | that's one win for today at least :) | 13:53 |
pedroalvarez | this is the closest thing we have ever had to a proper CI, so YAY! | 13:53 |
CTtpollard | +1 | 13:54 |
gunnarx | +1 | 13:54 |
radiofree_ | is SOTA/Software Manager being used in the CI pipeline? | 13:55 |
CTtpollard | not directly | 13:56 |
CTtpollard | sota client is integrated into gdp, but there's no interaction taking place as of yet | 13:57 |
*** steve_l has joined #automotive | 14:01 | |
CTtpollard | gunnarx: will try and be more eloquent with explaining the 'parallel PR' situation next time, now there's some existing PR's that show it it's easier to describe/visualise though | 14:02 |
toscalix | rajm: I added you to the group GDP maintainers | 14:03 |
gunnarx | disk full on ct agent :( | 14:03 |
gunnarx | https://go.genivi.org/go/tab/build/detail/GDP-Yocto-MINNOWBOARD-GITHUB-PR/1/build/1/init_and_bitbake | 14:03 |
rajm | thanks toscalix! | 14:04 |
gunnarx | We have a new GDP maintainer? | 14:04 |
gunnarx | Ah yes | 14:04 |
toscalix | rajm: Robert Marshall | 14:04 |
gunnarx | never mind :) | 14:04 |
gunnarx | Good to see you in action rajm | 14:05 |
* steve_l hears hdd is cheap but agents are harder to come by :( | 14:06 | |
* steve_l although Spook agents are ten-a-penny seemingly | 14:07 | |
*** mdurnev has quit IRC | 14:13 | |
rajm | Thanks gunnarx looking forward to the work! | 14:13 |
*** fredcadete has quit IRC | 14:14 | |
*** rajm has quit IRC | 14:21 | |
*** rajm has joined #automotive | 14:35 | |
*** fredcadete has joined #automotive | 14:41 | |
*** fredcadete has quit IRC | 14:46 | |
*** fredcadete has joined #automotive | 15:00 | |
*** fredcadete has quit IRC | 15:00 | |
*** fredcadete has joined #automotive | 15:01 | |
*** rajm has quit IRC | 15:09 | |
*** rajm has joined #automotive | 15:09 | |
*** toscalix has quit IRC | 15:10 | |
*** jlrmagnus has joined #automotive | 15:10 | |
*** toscalix has joined #automotive | 15:11 | |
gunnarx | CTtpollard, pedroalvarez still there? | 15:14 |
pedroalvarez | yup | 15:14 |
CTtpollard | o/ | 15:15 |
*** toscalix_ has joined #automotive | 15:15 | |
gunnarx | Do you guys have the graphics binaries, it looks like the recipes you set up for porter are doing copy_gfx_ etc. | 15:15 |
*** toscalix has quit IRC | 15:15 | |
gunnarx | Do you have them on your agent I mean | 15:15 |
CTtpollard | they should be pedroalvarez? | 15:15 |
CTtpollard | ls /var/go/sgx_bin/ shows them | 15:16 |
gunnarx | Previously I've set up the porter builds to only build on a machine that was allowed to have the binaries | 15:16 |
gunnarx | And I used a different path | 15:16 |
gunnarx | Just basically looking to see that we agree on paths | 15:16 |
CTtpollard | I took the path from the existing renesas template iirc | 15:17 |
gunnarx | Oh. Didn't know there was a defined path. In the long run recipes should be buildable on multiple agents. currently PR recipes are tied to CT agent. | 15:17 |
gunnarx | And the template for the PR build, it seems like we could use the same template everywhere really. | 15:18 |
gunnarx | I'm happy to adapt to yours, just saying we seem to be developing the same in parallel but different | 15:18 |
gunnarx | so let's sync up. | 15:18 |
pedroalvarez | the goal of our pipeline is to try to save some diskspace. | 15:18 |
pedroalvarez | our agent is full of pipelines that take up to 60G | 15:19 |
gunnarx | great, why don't we discuss some strategies for that. | 15:19 |
gunnarx | yes, you are using a shared DL_DIR, OK | 15:19 |
pedroalvarez | my current one is to have sstate and downloads in a common place for all of them | 15:19 |
pedroalvarez | exactly | 15:20 |
gunnarx | I've been requesting that we agree on paths for that | 15:20 |
gunnarx | Although I was more leaning towards PREMIRRORS, but that's because d/l isn't that big after alll | 15:20 |
gunnarx | it's a few gigs I think | 15:20 |
gunnarx | Most of the 60G is stuff generated during the build right? | 15:20 |
pedroalvarez | ah, I also remove tmp/ when the build finishes | 15:21 |
gunnarx | and before I noticed :) | 15:21 |
gunnarx | sorry got interrupted. afk | 15:21 |
pedroalvarez | oh, just before, not after | 15:22 |
CTtpollard | instead of removing /tmp it might be cleaner to inherit="rm_work" | 15:22 |
steve_l | 60G! rm_work really is your friend. http://git.projects.genivi.org/?p=genivi-demo-platform.git;a=blob;f=gdp-src-build/conf/local.conf;h=d3b433fb42007670fce222f2fe724448350c7892;hb=refs/heads/porter#l6 | 15:23 |
CTtpollard | I'm not sure it that will actually save space constantly though | 15:23 |
CTtpollard | snap steve_l | 15:23 |
steve_l | Indeed. Though note my comment below that line about navit. | 15:24 |
CTtpollard | yeh I kept that when switching to the templates https://github.com/GENIVI/genivi-dev-platform/blob/master/gdp-src-build/conf/templates/porter.local.conf | 15:25 |
steve_l | CTtpollard: Did Jonathan solve his silk kernel issue do you know? He's unresponsive on irc :) | 15:25 |
CTtpollard | however, now that some form of FSA functionality has been merged to meta-genivi, I might send a PR to remove it | 15:25 |
CTtpollard | steve_l: I believe he's managed to boot into a shell now, if I stare across the room long enough he might notice | 15:26 |
CTtpollard | jonathanmaw ^ | 15:27 |
jonathanmaw | steve_l: yep, the problem was that I was specifying ttySC6, not ttySC10 | 15:27 |
jonathanmaw | I got it booting into a shell. from there I can figure out why I'm not getting anything out to the touchscreen. | 15:28 |
* steve_l he's <insert that hipster word for being in the zone> | 15:28 | |
steve_l | ^ ahh good :) | 15:28 |
gunnarx | OK I'm back but in another meeting and have to go soon | 15:32 |
* paulsherwood is amazed to see automotive folks actually using irc productively now... who'dathunk? :) | 15:37 | |
* jbocklage2 not | 15:37 | |
paulsherwood | not amaze, not productive, or wouldn't have considered it possible? :) | 15:37 |
CTtpollard | paulsherwood: I can ask about my intermittent clutch slipping if you want | 15:38 |
jbocklage2 | hehe :) No. 3 | 15:38 |
paulsherwood | CTtpollard: i'm a software guy, that sounds like hardware to me | 15:38 |
CTtpollard | gunnarx: I was referencing http://go.genivi.org/go/admin/templates/Yocto_GDP_Build_Renesas/stages/build/job/init_and_bitbake/tasks for the sgx path | 15:39 |
gunnarx | Right CTtpollard. And I wrote that :) | 15:40 |
steve_l | CTtpollard: That's because your using an old generation single core clutch. If you were using a modern multi-core clutch you could integrate all the other h/w into just the clutch. | 15:40 |
gunnarx | facepalm | 15:40 |
*** tjamison has joined #automotive | 15:40 | |
*** rajm has quit IRC | 15:40 | |
CTtpollard | gunnarx: I know, so it made sense to do the same :) | 15:41 |
CTtpollard | steve_l: have you been on that IPA you mentioned already today? | 15:41 |
* gunnarx logs off irc in shame | 15:41 | |
gunnarx | seriously I have to go. let's sync up on aligning the templates later on. if you want you can use resource yocto_build to open up for more agents running the builds | 15:42 |
gunnarx | bye | 15:42 |
*** gunnarx has quit IRC | 15:42 | |
CTtpollard | o/ | 15:42 |
steve_l | CTtpollard: sadly no but in the current humidity it is a mightly attractive idea | 15:43 |
*** mpaolino has joined #automotive | 15:44 | |
steve_l | pedroalvarez: There is some detris in the folder tree as well but for illustration my GDP build with rm_work on Porter is 15GB. download directory outside of the tree | 15:45 |
*** ashwasimha_ has joined #automotive | 15:50 | |
*** jlrmagnus has quit IRC | 15:51 | |
*** clopez has quit IRC | 15:59 | |
*** clopez_ has joined #automotive | 16:02 | |
*** jlrmagnus has joined #automotive | 16:03 | |
jonathanmaw | well, that's the HDMI display problem fixed. The cable was loose >.< | 16:08 |
*** fredcadete has quit IRC | 16:08 | |
*** chbae has joined #automotive | 16:15 | |
chbae | hello | 16:19 |
*** chbae has quit IRC | 16:26 | |
*** jonathanmaw has quit IRC | 16:31 | |
*** toscalix_ is now known as toscalix | 16:32 | |
*** Tarnyko has quit IRC | 16:37 | |
*** jlrmagnus has quit IRC | 16:42 | |
*** |\n has joined #automotive | 16:47 | |
|\n | hello guys, i'm trying to find a replacement for bosch blower motor or at least maybe there is some catalog where i could see dimensions | 16:48 |
|\n | the point of interest is 0130111040 | 16:49 |
*** jobol has quit IRC | 17:05 | |
toscalix | |\n: check the topic of the channel, at the top part of your IRC client | 17:10 |
toscalix | this channel is for automotive....software discussions | 17:11 |
toscalix | you do not want software engineers to provide you support on mechanical problems....trust me | 17:11 |
|\n | oh, my fault, toscalix, is there anything similar to can bus triple project btw? | 17:12 |
|\n | if regarding software | 17:12 |
leon-anavi | toscalix, well said :) | 17:17 |
*** Tarnyko has joined #automotive | 17:18 | |
toscalix | :-) | 17:20 |
*** mpaolino has quit IRC | 17:20 | |
*** leon-anavi has quit IRC | 17:38 | |
*** steve_l has quit IRC | 17:41 | |
*** toscalix has quit IRC | 17:42 | |
*** kooltux_ has quit IRC | 17:58 | |
*** kooltux__ has joined #automotive | 17:58 | |
*** JEEB has quit IRC | 17:58 | |
*** JEEB has joined #automotive | 17:59 | |
*** jjardon has quit IRC | 18:12 | |
*** jjardon has joined #automotive | 18:17 | |
*** jjardon has quit IRC | 18:22 | |
*** jjardon has joined #automotive | 18:22 | |
*** IOBN has quit IRC | 18:30 | |
*** IOBN has joined #automotive | 18:31 | |
*** ashwasimha_ has quit IRC | 19:01 | |
*** |\n has left #automotive | 19:12 | |
*** jlrmagnus has joined #automotive | 20:20 | |
*** Saint_Isidore has quit IRC | 21:21 | |
*** kooltux_ has joined #automotive | 21:51 | |
*** kooltux_ has quit IRC | 21:58 | |
*** Saint_Isidore has joined #automotive | 22:04 | |
*** tjamison has quit IRC | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!