*** jlrmagnus has quit IRC | 01:00 | |
*** mdurnev has joined #automotive | 01:38 | |
*** AlisonChaiken has quit IRC | 01:58 | |
*** AlisonChaiken has joined #automotive | 02:01 | |
*** fury has quit IRC | 05:04 | |
*** fury has joined #automotive | 05:17 | |
*** fury has quit IRC | 05:23 | |
*** weston has joined #automotive | 07:31 | |
weston | hi all | 07:31 |
---|---|---|
*** weston has quit IRC | 07:39 | |
*** weston has joined #automotive | 07:41 | |
*** weston has quit IRC | 07:49 | |
*** leon-anavi has joined #automotive | 08:01 | |
*** Sisco has quit IRC | 08:16 | |
*** ZrZ is now known as RzR | 08:16 | |
*** toscalix has joined #automotive | 08:35 | |
*** CTtpollard has joined #automotive | 08:35 | |
*** ctbruce has joined #automotive | 08:39 | |
*** Tarnyko has joined #automotive | 09:10 | |
Tarnyko | morning | 09:10 |
CTtpollard | hello | 09:12 |
*** gunnarx_ has joined #automotive | 09:16 | |
*** jbocklage_ has joined #automotive | 09:27 | |
*** jbocklage has quit IRC | 09:29 | |
*** jobol has joined #automotive | 09:36 | |
*** jonathanmaw has joined #automotive | 09:50 | |
*** fury has joined #automotive | 10:03 | |
*** Sisco has joined #automotive | 10:07 | |
*** gunnarx_ has quit IRC | 10:38 | |
*** gmacario has quit IRC | 11:18 | |
*** gunnarx_ has joined #automotive | 11:40 | |
*** gunnarx_ has quit IRC | 11:54 | |
*** fredcadete has joined #automotive | 12:39 | |
*** phongtran has joined #automotive | 12:45 | |
*** gunnarx has joined #automotive | 13:03 | |
*** gunnarx has joined #automotive | 13:03 | |
*** andreas_ has joined #automotive | 13:05 | |
*** andreas_ is now known as Guest12131 | 13:06 | |
Guest12131 | ping / test | 13:06 |
jbocklage_ | pong | 13:09 |
Guest12131 | thanks ;-) this was my first ICQ post in the last 10 years ;-) | 13:11 |
JEEB | and the first IRC message in how many years? ;) | 13:14 |
Guest12131 | \nick awarnke | 13:14 |
Guest12131 | Yes - i just saw the typo-error | 13:15 |
CTtpollard | :) | 13:15 |
*** Guest12131 is now known as awarnke | 13:15 | |
gunnarx | Hi awarnke, welcome back | 13:33 |
*** alex_h has quit IRC | 13:41 | |
*** waltminer has joined #automotive | 13:49 | |
*** alex_h has joined #automotive | 13:52 | |
*** phongtran has quit IRC | 13:54 | |
*** steve_l has joined #automotive | 13:56 | |
awarnke | Hi gunnar - good to hear you! | 13:58 |
toscalix | GENIVI tools meeting through iRC today? | 13:59 |
paulsherwood | yup | 14:00 |
paulsherwood | == GENIVI Tools Team Meeting Starts == | 14:00 |
*** philrob has joined #automotive | 14:00 | |
paulsherwood | sorry folks, was on a call... am 100% here now :) | 14:00 |
awarnke | Question: Is this regular meeting at 14:00 or at 15:00h? | 14:00 |
paulsherwood | it is regularly at this time | 14:00 |
awarnke | ok. thx. | 14:01 |
paulsherwood | but there is a suggestion to move it to wednesday morning, to follow the BIT | 14:01 |
paulsherwood | we don't do a rollcall normally, but who's here? :) | 14:01 |
gunnarx | me | 14:02 |
paulsherwood | hi gunnarx | 14:02 |
toscalix | me | 14:02 |
*** alex_h has quit IRC | 14:02 | |
*** mdurnev has quit IRC | 14:02 | |
awarnke | me, Andreas | 14:02 |
paulsherwood | well i take this as a quorum.. unfortunately klaus birken can not be here today | 14:02 |
gunnarx | :( | 14:03 |
paulsherwood | before i go with the usual agenda, dnoes anyone have new items to propose? | 14:03 |
steve_l | here | 14:03 |
paulsherwood | hi steve_l | 14:03 |
* paulsherwood proposes the mirroring topic for AOB | 14:03 | |
pedroalvarez | patchtracking is already a topic? | 14:04 |
* paulsherwood proposes xmas break for AOB | 14:04 | |
paulsherwood | pedroalvarez: actually no... we can add it | 14:04 |
paulsherwood | any more topics? | 14:04 |
gunnarx | good one | 14:05 |
paulsherwood | last week's transcript is at https://at.projects.genivi.org/wiki/display/TOOL/TT-minutes-20151130 | 14:05 |
paulsherwood | if anyone thinks of things, feel free to pipe up :) | 14:05 |
paulsherwood | - - Build Tools - | 14:05 |
gunnarx | Is that the go topic? keep forgetting. | 14:06 |
gunnarx | :) | 14:06 |
paulsherwood | yup... but previous discussions on this were heavily with klausbirken i think | 14:06 |
paulsherwood | oh, the 'go' topic... no | 14:07 |
paulsherwood | i think that has fallen into automated testing previously... | 14:07 |
gunnarx | Yep. It's easy to misread. I take it as no, it's later in the agenda, not as "oh no, the go topic again" | 14:07 |
paulsherwood | do you want to update on 'go' now? | 14:07 |
gunnarx | sure, it has to come some time | 14:08 |
paulsherwood | please proceed :) | 14:08 |
gunnarx | So I realized there was a problem with user account creation but I hope I'm fixing it now as we speak | 14:08 |
gunnarx | Wanted to have that ready. Otherwise I expect later today we will finally roll it out to users | 14:08 |
paulsherwood | is there anything visible today? | 14:09 |
gunnarx | yeah ^^ later today :) | 14:09 |
paulsherwood | ok... i was hoping to mention a url here | 14:09 |
paulsherwood | but it can wait | 14:09 |
paulsherwood | pedroalvarez: i believe you've been loking into this too... any comments? | 14:10 |
toscalix | Codethink is working to use go.cd in production. | 14:10 |
gunnarx | I'll be expecting people's SSH pub keys if you want to create a personal account, so maybe prepare for that (or send to me now) | 14:10 |
toscalix | the coming weeks | 14:10 |
gunnarx | toscalix, good for you. should be some synergies possible | 14:10 |
toscalix | so no real news for now but soon | 14:10 |
gunnarx | I'm considering a read-only guest account. | 14:11 |
pedroalvarez | well, yes, some comments, but nothing to be discussed here | 14:11 |
paulsherwood | toscalix: by production you mean genivi, or something else? | 14:11 |
gunnarx | is that a good idea? | 14:11 |
toscalix | gdp | 14:11 |
paulsherwood | ok | 14:11 |
paulsherwood | pedroalvarez: why not here? | 14:11 |
pedroalvarez | gunnarx: yeah, we will need that | 14:11 |
gunnarx | Ah, ok, so the one and same instance then toscalix? | 14:11 |
toscalix | yes | 14:11 |
pedroalvarez | paulsherwood: during this meeting I think | 14:11 |
pedroalvarez | they are just ideas of next steps, and such | 14:12 |
gunnarx | thought you meant an additional one internally in CT. OK, that's what I expected of course that we set up a bunch of GDP stuff | 14:12 |
paulsherwood | i think that ouwld be of interest here, pedroalvarez ? | 14:12 |
paulsherwood | unless gunnarx prefers the discussion to be elsewhere? | 14:12 |
toscalix | pedroalvarez: go ahead | 14:12 |
gunnarx | yeah go ahead | 14:13 |
pedroalvarez | not now please :) I didn't have my brain ready for this | 14:14 |
toscalix | the ideas is... | 14:14 |
gunnarx | ok, no problem. I'm available for support | 14:14 |
paulsherwood | hah. ok we can take it at another time | 14:14 |
toscalix | to use go.cd for building and basic acceptance criteria | 14:14 |
gunnarx | the idea is... to build all GDP variants, to track upstream layers with continuous integration seeing where they break. | 14:14 |
toscalix | so we can put some basi processes in place that allow us to | 14:14 |
paulsherwood | toscalix: acceptance by whom, of what? | 14:14 |
gunnarx | the idea is... to take artifacts from the build and upload to a server for download | 14:14 |
gunnarx | and so on. don't think we need to go over this, it's pretty obvious and sure we've done it before | 14:14 |
toscalix | move forward to gdp based on baseline v9 using go.cd | 14:14 |
paulsherwood | ok | 14:14 |
toscalix | basic acceptance criteria like...the image boots? ping, ssh? | 14:15 |
paulsherwood | sounds exciting... maybe it would be good to get an 'official' slot in the agenda for this at future meetings? | 14:15 |
toscalix | so start simple with current gdp to use it in 2016Q1 for gdp v9 | 14:15 |
*** Sisco has quit IRC | 14:15 | |
paulsherwood | ack | 14:15 |
gunnarx | FYI The reason you use ssh to create a personal account is that you will get a frontend that allows you to choose user name and password. There's no good web front end so I created one over SSH. And that means the password is secure for your eyes only. It's stored in hashed form on server. | 14:16 |
paulsherwood | cool | 14:16 |
paulsherwood | gunnarx: shall we have an official item for go in future? | 14:16 |
*** alex_h has joined #automotive | 14:16 | |
gunnarx | not terribly cool, the login script is restricted shell that talks to you. and taunts you sometimes, lol | 14:16 |
pedroalvarez | I thought it was "Automated Testing" | 14:17 |
paulsherwood | and do we have any actions from this discussion? | 14:17 |
gunnarx | Well Go is build, test, deploy, ... hard to define | 14:17 |
paulsherwood | gunnarx: true. but it's (becoming) an actual project, so maybe worth an explicit slot? | 14:17 |
gunnarx | One action is for those who want accounts to send or pastebin me SSH public key. You don't have to be a GENIVI member but if I don't know you at all, please include a little introduction. | 14:18 |
* pedroalvarez wonders if this is better on an email | 14:18 | |
gunnarx | We can have a slot sure, at least to drive forward the creation of pipelines / recipes | 14:18 |
gunnarx | There will be an intro email from me. Just giving you a head start :) | 14:19 |
paulsherwood | gunnarx: ack. i've notionally assigned this to you https://at.projects.genivi.org/jira/browse/TOOL-10 | 14:19 |
toscalix | I would like to point out that, the usage of the tool implies that we are hosting images | 14:20 |
paulsherwood | (i've been told to improve on our actions get tracked) | 14:20 |
paulsherwood | toscalix: are those images going to be published? | 14:20 |
toscalix | the tool is public, our work is public..... | 14:20 |
toscalix | I assume so | 14:21 |
gunnarx | toscalix, that's one planned usage yes. with necessary sources of course, as mailing list has beaten to death. | 14:21 |
paulsherwood | :) | 14:21 |
toscalix | gunnarx: I read the threads..... and I hope we can move forward on this. | 14:21 |
gunnarx | there's no image hosting unless we put it in, and then we do it right. do you have any concern remaining about this? | 14:22 |
toscalix | I do not have any. It is a common practice in other projects | 14:22 |
gunnarx | ok | 14:22 |
paulsherwood | toscalix: https://at.projects.genivi.org/jira/browse/TOOL-11 just so we don't forget :) | 14:22 |
* paulsherwood is in action-slinging mode | 14:23 | |
paulsherwood | shall we move forward? | 14:23 |
*** jobol has quit IRC | 14:23 | |
gunnarx | if there are no other questions we can move on to other things, and I can fix the user login | 14:23 |
*** jobol has joined #automotive | 14:23 | |
paulsherwood | k | 14:24 |
paulsherwood | - UML Modeling - | 14:24 |
paulsherwood | any updates on this? | 14:24 |
paulsherwood | maybe this waits til after the F2F? | 14:25 |
paulsherwood | when is that, btw? | 14:25 |
gunnarx | Exactly, I hope to have a discussion on the SAT F2F about UML strategy | 14:25 |
paulsherwood | cool | 14:25 |
*** ToxicGumbo1 has quit IRC | 14:25 | |
gunnarx | next week, during 14-18 December, roughly | 14:26 |
paulsherwood | gunnarx: would someone report bakck here on that? will the discussion be private or public? | 14:26 |
gunnarx | SAT meetings are documented inside of GENIVI atm | 14:26 |
paulsherwood | ack | 14:27 |
gunnarx | Not sure what people want to know. Someone cares about UML, participate now... ask questions... | 14:27 |
paulsherwood | well, in my dreamworld it would be possible to trace from UML to code to image | 14:28 |
paulsherwood | is that possible? | 14:28 |
gunnarx | in theory, yes. | 14:29 |
paulsherwood | also... would GENIVI consider using plantuml instead of EA? | 14:29 |
paulsherwood | s/would/could/ | 14:29 |
gunnarx | Requirements traceability we use of course. If you mean full out model based development then few companies succeed, which I would say makes it doubly difficult for an alliance to have such a shared model. Sharing code seems difficult enough :) | 14:29 |
paulsherwood | i don't mean model based development, i think that's unfeasible, especially since much of what we care about has been previously developed, without models? :) | 14:30 |
gunnarx | I want to ask the SAT about that. Previously differing opinions whether introducing a lightweight format for alternative diagrams is good or bad, overlaps the "real model". | 14:30 |
paulsherwood | is there some limitation to plantuml that makes folks think it can't do the 'real model'? | 14:31 |
paulsherwood | anyways... those were my questions for the F2F... any more on this? | 14:31 |
awarnke | PlantUML is really cool. You can easily add a small diagram within the doxygen comments of some function documentation | 14:31 |
paulsherwood | +1 | 14:32 |
awarnke | but for big models - i do not see it | 14:32 |
gunnarx | I like it too, but it seems like a complement | 14:32 |
gunnarx | +1 | 14:32 |
paulsherwood | maybe somone could *try* to port a big model into it? | 14:32 |
rjek | PlantUML is excellent as long as you are willing to let go of your irrational desire to have pixel-level control of the layout of your datagram :) | 14:32 |
paulsherwood | :-) | 14:32 |
*** ToxicGumbo has joined #automotive | 14:33 | |
awarnke | A big model has the advantage of being consistent: You have a meta-model and drag the elements into your diagrams. | 14:33 |
awarnke | Plant-UML just has diagrams | 14:33 |
gunnarx | "someone" is welcome to try. :-) I'm not sure I've heard the reason why though. | 14:33 |
paulsherwood | ok | 14:33 |
gunnarx | thanks awarnke | 14:33 |
paulsherwood | any actions from this? | 14:33 |
awarnke | you're free to contradict ;-) | 14:33 |
paulsherwood | gunnarx: would you be happy to report back from the F2F on this topic, or should that fall to someone else? | 14:34 |
paulsherwood | (to the level that public reporting is possible) | 14:34 |
* paulsherwood proposes to move on... | 14:35 | |
gunnarx | IMHO, the TT topic needs guidance from what SAT/GENIVI want to achieve, sure. But I want to see the TT being creative and free, try out tools, evaluate... so plantuml was not a bad example to throw in the mix, even if it doesn't replace EA | 14:35 |
paulsherwood | ack | 14:36 |
paulsherwood | gunnarx: can i action you with feeding back to us, then? | 14:37 |
gunnarx | you can action what you like, I will still do what I want :) | 14:37 |
paulsherwood | heh | 14:37 |
gunnarx | No problem, might not be much to say, we wil; see | 14:37 |
paulsherwood | gunnarx: thanks .... https://at.projects.genivi.org/jira/browse/TOOL-13 | 14:38 |
paulsherwood | - Debugging and Analysis - | 14:38 |
paulsherwood | i failed to even raise the action on this last time... | 14:39 |
paulsherwood | https://at.projects.genivi.org/jira/browse/TOOL-12 | 14:39 |
*** dzen has quit IRC | 14:39 | |
paulsherwood | so i have not done it | 14:39 |
paulsherwood | any new comments on this topic? | 14:39 |
*** dzen has joined #automotive | 14:40 | |
paulsherwood | ok, moving on, then | 14:40 |
paulsherwood | - Automated Testing - | 14:40 |
paulsherwood | we already talked about the go project earlier | 14:41 |
paulsherwood | any other comments wrt automated testing? | 14:41 |
toscalix | not from my side | 14:41 |
gunnarx | sorry, no progress from me. As stated before I'd like to see investigate the engine part of JTA integrated with Go instead of Jenkins. Keeps the important parts equal. | 14:42 |
paulsherwood | i have one... | 14:42 |
gunnarx | to be more specific, I would like an evaluation of JTA engine to see if it's any good | 14:42 |
pedroalvarez | I still don't know what JTA is | 14:42 |
paulsherwood | would GENIVI be interested in establishing a physical board farm (or multiple farms) where community members could run tests? | 14:42 |
gunnarx | Jenkins Test Thingy | 14:43 |
gunnarx | Thingy spelled with an A | 14:43 |
pedroalvarez | hah :) | 14:43 |
paulsherwood | gunnarx: maybe easiest on that is to get feedback from AGL? they are using it i believe | 14:43 |
gunnarx | sure. the meeting is open for input, the topic has been up before | 14:43 |
paulsherwood | gunnarx: i think we'd have to ask explicitly, this timing is not great for some of the AGL folks | 14:44 |
steve_l | or ask LTSI list | 14:44 |
gunnarx | waltminer, do you know about AGL experience with JTA so far? | 14:44 |
*** awarnke has left #automotive | 14:45 | |
waltminer | I would say not great so far | 14:45 |
gunnarx | pedroalvarez, URL again for reference https://bitbucket.org/cogentembedded/jta-public/ | 14:45 |
waltminer | dl9pf - Jan-Simon can chime in as well | 14:45 |
paulsherwood | waltminer: can you provide any more detail? | 14:45 |
pedroalvarez | gunnarx: thanks | 14:45 |
* paulsherwood wonders if JTA is a cogentembedded-only project? | 14:46 | |
steve_l | no | 14:46 |
dl9pf | reading ... | 14:46 |
dl9pf | ok, so far as I can tell it is a jenkins with extra plugins and a set of scripts to fetch an image, *build* the ltp/other tests using a toolchain, putting it into a dir for tftp/nfs boot and starting up the board with tftp - | 14:48 |
* paulsherwood is cloning | 14:48 | |
paulsherwood | dl9pf: so it's a whole jenkins setup? | 14:49 |
*** alex_h has quit IRC | 14:49 | |
dl9pf | right now yes | 14:49 |
gunnarx | there is a bash-coded engine part and an integration with jenkins. the engine looks independent, and not too bad, if you can stomach a bash based program. | 14:49 |
* paulsherwood kills the cloning, muttering 'wtf' :) | 14:49 | |
dl9pf | well, the scripts are shell scripts, so that could be used anywhere | 14:49 |
waltminer | it uses a separate tool chain IIRC from our regular build | 14:50 |
dl9pf | in a way it is a reimplentation of the lava idea | 14:50 |
gunnarx | I agree | 14:50 |
waltminer | +1 | 14:50 |
paulsherwood | i like bash-based programs... but why would someone include some in a whole fork of jenkins? | 14:50 |
gunnarx | paulsherwood, I know! | 14:50 |
dl9pf | personal comment - I think the building of the tests should be done at imaging time and not by JTA. It should just run it | 14:51 |
gunnarx | but the point is engine looks independent | 14:51 |
* paulsherwood does know why, actually, but should not rant in public | 14:51 | |
gunnarx | so you could influence upstream or fork... point is figure out if it's a good approach | 14:51 |
paulsherwood | is anyone already on that at AGL? or have you already decided? | 14:52 |
dl9pf | paulsherwood: it is no jenkins, you have to setup a jenkins and add these scripts | 14:52 |
dl9pf | qiaonuohan is setting up a POC for porter | 14:53 |
paulsherwood | ack | 14:53 |
waltminer | Jan-Simon can you explain the whole need for these huge binaries that need to get uploaded | 14:53 |
dl9pf | JTA hosts all their tests in the same git repo as they do for the shell scripts | 14:53 |
dl9pf | the repo is huge due to that | 14:53 |
paulsherwood | erk | 14:54 |
dl9pf | during the test process it will build the test binaries from those sources and place them into the filesystem | 14:54 |
dl9pf | (see my note above about when do build stuff) | 14:54 |
* toscalix needs to drop in 2 min. I have another meeting. | 14:54 | |
paulsherwood | for GENIVI, i propose we wait a bit longer to see how AGL gets on with this :) | 14:54 |
paulsherwood | dl9pf, waltminer - thanks for the input | 14:55 |
waltminer | any time | 14:55 |
paulsherwood | shall we move on to AOB? | 14:55 |
gunnarx | +1 open for discussion if anyone gets the time to dig into the bash parts a bit | 14:55 |
gunnarx | I sense a backronym / rename coming after we make it Jenkins independent... Jolly Test Automation? Jovial? Juvenile? | 14:55 |
gunnarx | :) | 14:56 |
paulsherwood | :) | 14:56 |
paulsherwood | - AOB - | 14:56 |
gunnarx | time for patch tracking? | 14:56 |
paulsherwood | yup... please go ahead | 14:56 |
* paulsherwood thinks AGL folks would say gerrit | 14:57 | |
pedroalvarez | so I have been wondering if we are going to move somewhere else soon | 14:57 |
gunnarx | else? | 14:57 |
pedroalvarez | for goCD makes sense to have some of the supported patch trackers | 14:57 |
paulsherwood | ie gerrit or? | 14:58 |
pedroalvarez | github or stash | 14:58 |
pedroalvarez | we can also use something else, and implement the plugins needed ourselves | 14:58 |
paulsherwood | cool! let's have gerrit and github for starters :) | 14:58 |
pedroalvarez | but I think we are not here to do that kind of things | 14:58 |
gunnarx | suppose I agree, but what does the support consist of? triggering builds? | 14:58 |
pedroalvarez | triggering builds per patch, and giving feedback | 14:58 |
gunnarx | ack | 14:59 |
paulsherwood | gunnarx: any reason we shouldn't just start with gerrit? | 14:59 |
pedroalvarez | github for example, doesn't require any infra | 14:59 |
* paulsherwood assumes genivi doesn't want to move git.projects.genivi.org/* to github | 15:00 | |
pedroalvarez | yeah, I guess so | 15:01 |
gunnarx | no I don't think that's the plan, but the question is bigger isn't it? | 15:01 |
paulsherwood | howso? | 15:01 |
gunnarx | You should know... I mean bringing full git mirrors into the picture, if that still wanted, then it's unlikely to mirror 300 packages on github for the purpose of a yocto build | 15:01 |
gunnarx | Owned gerrit or equivalent is then the way | 15:02 |
pedroalvarez | hm... I kind of assumed that the mirrors didin't have to be on the same place as genivi git's | 15:03 |
*** philrob has quit IRC | 15:03 | |
paulsherwood | gunnarx: actually, it wouldn't be hard to go the github road for that, i think? but i'm not recommending it | 15:03 |
paulsherwood | in any case the repos for code review don't have to be the same phsical repos as for the build | 15:04 |
gunnarx | pedroalvarez, if you wanted to do patch tracking on such projects before they go upstream. I've been over this with paulsherwood so many times I am wrongly assuming everyone is on the same page | 15:04 |
paulsherwood | can have clones (of clones) if necessary | 15:04 |
gunnarx | same questions rehashed :) But why not. First do people like the pull request workflow? That then points to Gitlab. | 15:05 |
gunnarx | instead of Gerrit | 15:05 |
paulsherwood | i think 'people' includes a range of preferences | 15:05 |
* paulsherwood needs to drop off this soon... | 15:05 | |
pedroalvarez | but gitlab I think is not supported by go.. | 15:06 |
gunnarx | and also not Gerrit as far as you have seen? | 15:06 |
pedroalvarez | gerrit, github and stash | 15:06 |
pedroalvarez | there might be more non-official plugins for others, i can research more | 15:07 |
paulsherwood | action on this? | 15:07 |
pedroalvarez | nothing else to say _o_ | 15:07 |
gunnarx | OK good. I thought I had made a mistake and just assumed Gerrit is a well known integration. Anything is possible of course with the APIs but good to have it out of the box. | 15:08 |
pedroalvarez | not sure about what action | 15:09 |
pedroalvarez | ask the people? or start using something for testing? | 15:09 |
paulsherwood | gunnahttps://at.projects.genivi.org/jira/browse/TOOL-14 | 15:09 |
paulsherwood | oops... https://at.projects.genivi.org/jira/browse/TOOL-14 | 15:09 |
gunnarx | A Gerrit setup seems pretty logical. I've looked around and it's amazingly well packaged nowadays, docker files and stuff. So it's low hanging | 15:09 |
paulsherwood | next up... mirrorinr | 15:09 |
paulsherwood | mirroring. there's been so many emails on this i don't know what to do | 15:10 |
paulsherwood | gunnarx: what do you suggest? | 15:11 |
gunnarx | I tried to summarize in one of the emails. Items 1. 2. 3. Me and Jan-Simon described the same things. | 15:12 |
gunnarx | Mirroring for license compliance, no more discussion - it will be done as we put binaries up. | 15:12 |
paulsherwood | yup. | 15:12 |
dl9pf | so build stage and development stage left | 15:12 |
* steve_l dropping.. meeting calls | 15:13 | |
gunnarx | yes | 15:13 |
*** steve_l has quit IRC | 15:13 | |
dl9pf | development is either in your patch tracker and for the rest a mirror can be considered for availability | 15:13 |
*** Sisco has joined #automotive | 15:14 | |
gunnarx | Build stage I hadn't run into the weird behavior I would have continued testing the git-annex setup for myself. Might've been something people adopt. | 15:14 |
gunnarx | and yes, what he said :) development is still open. | 15:14 |
dl9pf | pristine-tar might be interesting btw. | 15:15 |
gunnarx | *If I hadn't | 15:15 |
gunnarx | that's a debian thing? | 15:15 |
paulsherwood | dl9pf: not sure what you mean. developers would normally have their own clone? the question is where should they publish to? | 15:15 |
dl9pf | yes, but tizen with obs used it internally if i'm not mistaken | 15:15 |
dl9pf | paulsherwood: you're right, the push location is one thing to consider | 15:16 |
paulsherwood | dl9pf: so one of the things that occurs when there is not a clear destination for engineers to publish to, is they publish in random places.... | 15:17 |
paulsherwood | or worse, stick their patches in with layers :) | 15:17 |
dl9pf | well, upstream first? | 15:17 |
dl9pf | paulsherwood: yes, that is worse | 15:17 |
gunnarx | paulsherwood, we haven't agreed on all the details but you've been influential | 15:17 |
paulsherwood | jrandom developer does not by default have ability to push upstream | 15:17 |
gunnarx | Gerrit with local/private namespaces is one preference I've come up with based on our discussions. | 15:18 |
paulsherwood | so jrandom developer by default puhes to his/her own github, or corporate server, or nowhere? | 15:18 |
gunnarx | j random developer creates a repo jrandom/systemd and pushes to it and then asks other people to look at it | 15:18 |
paulsherwood | gunnarx: can gerrit handle all the cases where developer wants just to put their wip stuff, not get it reviewed? | 15:19 |
gunnarx | see above for my proposal. | 15:19 |
* paulsherwood does not know if gerrit does namespaces | 15:19 | |
gunnarx | It's just naming of repos I presume | 15:20 |
gunnarx | You can do subproject IIRC. dl9pf can you confirm ? | 15:20 |
dl9pf | iirc you can define a container like "external" and "internal" and everything "external/*" can inherit the settings from "external" | 15:20 |
paulsherwood | ok | 15:21 |
dl9pf | its more of a toplevel project which holds your default settings | 15:21 |
dl9pf | path based basically | 15:21 |
gunnarx | thanks, as I thought | 15:21 |
gunnarx | Just like now publishing under GENIVI namespace = controlled process. Publishing under your own namespace = not much controlled. Creating git repos easily is a useful service to provide. | 15:22 |
gunnarx | It's github but we won't trigger your OCD with stuff being all over the place ;) | 15:22 |
paulsherwood | :) | 15:22 |
* paulsherwood has to go very soon | 15:23 | |
gunnarx | OK. Solved then? :) | 15:23 |
paulsherwood | so the proposal is gerrit for mirroring, too? | 15:23 |
gunnarx | well.... for development | 15:24 |
gunnarx | or patch review rather, we need to flesh out the details | 15:25 |
paulsherwood | ok... so the remaining question is jusrt build? | 15:25 |
gunnarx | in a way, yes. I | 15:25 |
gunnarx | I'm not sure what needs are seen by individuals and companies here. Companies doing serious dev have their own mirror already | 15:26 |
paulsherwood | true | 15:26 |
paulsherwood | but they may not be able to publish from there | 15:27 |
gunnarx | Others... would they rsync/clone a published directory? Are they interested in getting a guaranteed package of sources, not sure. (my git-annex setup could robustly guarantee that you have everything before the build starts) | 15:27 |
paulsherwood | i'm out of time.... | 15:28 |
* paulsherwood proposes as final AOB that next week should be final TT meeting til new year | 15:29 | |
paulsherwood | i'm dropping off... could somone else close the meeting please? | 15:29 |
gunnarx | I think you just did :) | 15:30 |
paulsherwood | == GENIVI Tools Team Meeting Ends == | 15:30 |
paulsherwood | thanks, folks... | 15:30 |
paulsherwood | i'll write up outstanding questions, promise | 15:31 |
*** waltminer has quit IRC | 15:53 | |
*** RzR has quit IRC | 15:59 | |
*** RzR has joined #automotive | 15:59 | |
*** jlrmagnus has joined #automotive | 16:01 | |
*** waltminer has joined #automotive | 16:02 | |
*** mvick has joined #automotive | 16:06 | |
*** jlrmagnus has quit IRC | 16:08 | |
*** Sisco has quit IRC | 16:18 | |
*** Sisco has joined #automotive | 16:19 | |
*** Sisco has quit IRC | 16:36 | |
*** Tarnyko has quit IRC | 16:46 | |
*** egy has joined #automotive | 17:11 | |
*** MalditoDev has quit IRC | 17:15 | |
*** ctbruce has quit IRC | 17:17 | |
*** aeiche has joined #automotive | 17:20 | |
*** jobol has quit IRC | 17:37 | |
*** waltminer has quit IRC | 17:41 | |
*** fredcadete has quit IRC | 17:49 | |
*** waltminer has joined #automotive | 17:49 | |
*** jonathanmaw has quit IRC | 18:13 | |
*** Sisco has joined #automotive | 18:25 | |
*** toscalix has quit IRC | 18:30 | |
*** gmacario has joined #automotive | 18:45 | |
*** waltminer has quit IRC | 18:46 | |
*** waltminer has joined #automotive | 19:05 | |
*** gmacario has quit IRC | 19:19 | |
*** egy has quit IRC | 19:27 | |
*** leon-anavi has quit IRC | 20:13 | |
*** waltminer has quit IRC | 20:45 | |
*** youreeck has joined #automotive | 20:58 | |
*** egy has joined #automotive | 21:04 | |
*** waltminer has joined #automotive | 21:06 | |
*** gmacario has joined #automotive | 21:29 | |
*** youreeck has quit IRC | 21:33 | |
*** Sisco has quit IRC | 22:22 | |
*** Sisco has joined #automotive | 22:23 | |
*** waltminer has quit IRC | 22:44 | |
*** waltminer has joined #automotive | 22:45 | |
*** egy has quit IRC | 23:39 | |
*** mdurnev has joined #automotive | 23:54 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!