| *** bbranch has quit IRC | 00:00 | |
| *** waltminer has joined #automotive | 00:08 | |
| *** waltminer has quit IRC | 01:18 | |
| *** FabienH has joined #automotive | 05:25 | |
| *** gunnarx has quit IRC | 06:43 | |
| *** fredcadete has joined #automotive | 06:59 | |
| *** txenoo has joined #automotive | 07:23 | |
| *** CTtpollard has joined #automotive | 07:37 | |
| *** gunnarx has joined #automotive | 07:38 | |
| jeremiah | Connected a screen, but no luck. | 07:56 |
|---|---|---|
| jeremiah | I'll try Steve's advice and connect a mouse and keyboard. | 07:56 |
| CTtpollard | jeremiah: what is your problem? | 07:57 |
| jeremiah | CTtpollard: can't start weston / HMI | 07:58 |
| jeremiah | But a bunch of other things fail along the way, like correct config of sshd (likely due to hostnamectl failing?) | 07:58 |
| CTtpollard | jeremiah: is this the gdp on porter? sorry I had a very intermittent connection yesterday afternoon | 07:58 |
| jeremiah | GDP on porter. :-) | 07:58 |
| *** gunnarx has quit IRC | 07:59 | |
| *** gunnarx has joined #automotive | 08:23 | |
| *** jacobo has joined #automotive | 08:43 | |
| *** jonathanmaw has joined #automotive | 08:45 | |
| *** gunnarx has quit IRC | 08:57 | |
| *** gunnarx has joined #automotive | 08:57 | |
| *** FabienH has quit IRC | 09:00 | |
| jeremiah | CTtpollard: Does this look familiar? --> Couldn't stat device /dev/pts/ptmx | 09:07 |
| *** gunnarx has quit IRC | 09:07 | |
| *** wschaller has joined #automotive | 09:07 | |
| jeremiah | This is preventing hostnamectl from running (I think) | 09:07 |
| *** gunnarx has joined #automotive | 09:07 | |
| CTtpollard | jeremiah: I have not seen it, I'll aim to boot a porter build today to take a look | 09:09 |
| *** gunnarx has quit IRC | 09:12 | |
| *** gunnarx has joined #automotive | 09:12 | |
| jeremiah | CTtpollard: Cool. It looks to be pseudo terminal stuff connected to namespaces. | 09:12 |
| jeremiah | No rush, this is not a blocker BTW | 09:12 |
| *** FabienH has joined #automotive | 09:13 | |
| *** gunnarx has quit IRC | 09:18 | |
| *** gunnarx has joined #automotive | 09:19 | |
| *** jacobo has quit IRC | 09:24 | |
| *** gunnarx has quit IRC | 09:34 | |
| *** gunnarx has joined #automotive | 09:41 | |
| CTtpollard | I've put forward this question before in other channels, personally I think the irc logger should not record joins / quits. I think this is especially true for ease of reading of meetings | 09:44 |
| CTtpollard | or if we want to log joins etc, then something like https://wiki.debian.org/MeetBot should be used for meetings | 09:45 |
| jeremiah | CTtpollard: +1 | 09:52 |
| *** gunnarx has quit IRC | 09:54 | |
| *** gunnarx has joined #automotive | 09:55 | |
| CTtpollard | I've been in openstack meetings that use meetbot and have found it highly useful | 10:07 |
| paulsherwood | noted | 10:19 |
| *** gunnarx has quit IRC | 10:28 | |
| *** apinheiro has joined #automotive | 10:34 | |
| *** FabienH has quit IRC | 10:37 | |
| jeremiah | meetbot works great | 10:55 |
| jeremiah | We used to use it in MeeGo | 10:55 |
| CTtpollard | ok booting the porter now | 10:58 |
| CTtpollard | the systemd unit for the hostname service is failing for me at boot | 10:59 |
| jeremiah | yeah, it does that. | 11:17 |
| jeremiah | And it doesn't like to come up. | 11:17 |
| CTtpollard | ok, capture the systemctl status of it | 11:17 |
| CTtpollard | *captured | 11:17 |
| jeremiah | I looked at that too, but there's not much info there. | 11:17 |
| jeremiah | nor in the logs | 11:17 |
| jeremiah | The logging set up is remarkably sparse | 11:18 |
| jeremiah | neard is failing on boot too | 11:20 |
| CTtpollard | systemd-hostname doesn't as far as I can see have a .service file like I would have expected | 11:23 |
| jeremiah | I think its called systemd-hostnamed | 11:25 |
| jeremiah | Which is, umm, counter-intuitive. | 11:25 |
| jeremiah | I think it may be the private network setting that's at issue, at least so says Google. | 11:25 |
| CTtpollard | jeremiah: yes, but that is a binary :) | 11:26 |
| jeremiah | Well, not the service file. :-) | 11:28 |
| jeremiah | i.e. /lib/systemd/system/systemd-hostnamed.service | 11:28 |
| CTtpollard | the service file is what I was expecting to find | 11:28 |
| jeremiah | Change #PrivateNetwork from yes to no | 11:29 |
| CTtpollard | my build does not have the service file | 11:29 |
| jeremiah | ? | 11:29 |
| jeremiah | Weeeeeird | 11:29 |
| CTtpollard | and find isn't pulling the location either | 11:30 |
| CTtpollard | sorry I'm on the Koelsch now not the porter | 11:33 |
| CTtpollard | although if you have a service file and I don't that is boggling | 11:33 |
| CTtpollard | jonathanmaw: could you see if you have the service file as jeremiah does? | 11:36 |
| CTtpollard | I'll reboot in the meantime | 11:37 |
| *** FabienH has joined #automotive | 11:37 | |
| jonathanmaw | my systemd-hostnamed has PrivateNetwork=yes | 11:37 |
| jeremiah | jonathanmaw: Do you get hostnamectl failuers? | 11:39 |
| jeremiah | Failures even | 11:39 |
| jonathanmaw | jeremiah: yep, it seems I have failures as well | 11:41 |
| jonathanmaw | status=225/NETWORK | 11:41 |
| jeremiah | Yeah, config out the PrivateNetwork and that goes away, fwiw | 11:42 |
| *** txenoo has quit IRC | 11:46 | |
| *** txenoo has joined #automotive | 11:47 | |
| *** waltminer has joined #automotive | 12:06 | |
| *** txenoo has quit IRC | 12:06 | |
| *** waltminer has quit IRC | 12:06 | |
| *** waltminer has joined #automotive | 12:08 | |
| *** gunnarx has joined #automotive | 12:08 | |
| gunnarx | ping CTtpollard | 12:13 |
| gunnarx | Just wondering if you have looked at https://github.com/johnnytk/meta-genivi-demo/commit/2afcb22a569 | 12:19 |
| gunnarx | Except for the BSP it's the only diff from koelsch to minnowboard. It's based on the Wiki instructions for minnowboard. | 12:19 |
| gunnarx | As of now I've moved back gdp-submodules completely to the standard URLs for the layers and also minnowboard is now set to the same meta-genivi-demo commit as koelsch/porter, meaning the diff I point out here is _not_ included. | 12:20 |
| gunnarx | weston_ini.patch (previously used by minnowboard build) includes some useful things, like the correct names of the graphics files that are stored in the dir. Koelsch/porter build must have these wrong, afaict. | 12:22 |
| gunnarx | The minnowboard build patches weston.ini whereas the koelsch build installs its own weston.ini. The latter seems a cleaner approach, i.e. keep an exact copy of the weston.ini we want among the layer files and just install it. | 12:23 |
| gunnarx | I'd propose to: Include the koelsch installation of weston.ini also for minnowboard (this is what should happen now since they use the same layer definition). Step 2. Put the right graphics names from the weston_ini.patch into the weston.ini that is checked in. Step 3. Consider if there are different graphics settings and other differences required in weston_ini for the two boards (I have not looked into that). If differences re | 12:27 |
| gunnarx | quired, either use yocto magic to make sure the right file is installed depending on the machine you build for, or create two branches in meta-genivi-demo where weston.ini differs. (minnowboard/porter (or intel/renesas)) | 12:27 |
| gunnarx | "Include the koelsch installation" means "Include the installation of weston.ivi, just like it is done right now on the koelsch branch" | 12:28 |
| fredcadete | is the same weston.ini expected to work independently of the screen used by each developer? | 12:30 |
| radiofree | yes | 12:31 |
| fredcadete | great | 12:31 |
| radiofree | though the weston.ini in this meta-genivi-demo sets the keyboard layout to german | 12:32 |
| jeremiah | I think the porter and koelsch use the same graphics drivers so dropping in a correct weston.ini should "just work"? | 12:32 |
| radiofree | jeremiah: yes, you change the graphics backend when launching weston | 12:32 |
| radiofree | (though you can also do it in the weston.ini) | 12:33 |
| jeremiah | So I ought't just monkeypatch my running weston.ini on the porter since I'm not getting any graphic output. | 12:33 |
| radiofree | jeremiah: run weston with --log=/tmp/weston.log and pastebin the output | 12:34 |
| radiofree | seems renesas have implemented their own rendered as well - https://github.com/renesas-devel/weston/blob/RCAR-GEN2/1.5.0/vsp1-dev/src/vsp-renderer.c | 12:35 |
| radiofree | so i guess there's going to be some work to forward port that to a newer weston | 12:35 |
| radiofree | which will need to happen at some point, considering master of wayland-ivi-extension doesn't work with any weston < 1.8.0 | 12:35 |
| jeremiah | radiofree: http://pastebin.com/eq66vga9 | 12:36 |
| jeremiah | radiofree: Yeah, there's been talk of upgrading but I think the hope it to make it through October with older versions | 12:37 |
| rjek | Afternoon, automotivites. | 12:38 |
| paulsherwood | radiofree: would it be *easier* to move forwards? | 12:38 |
| radiofree | paulsherwood: whats the deadline? | 12:38 |
| CTtpollard | sorry gunnarx I was afk | 12:39 |
| radiofree | jeremiah: are you manually running weston there? | 12:39 |
| jeremiah | yeah | 12:39 |
| radiofree | jeremiah: and there's a monitor attached | 12:39 |
| jeremiah | rjek: Hey there! | 12:39 |
| jeremiah | radiofree: yep | 12:39 |
| radiofree | are you running that command via serial/ssh? | 12:40 |
| radiofree | the weston one sorry | 12:40 |
| jeremiah | yeah, via ssh | 12:40 |
| jeremiah | yeah | 12:40 |
| jeremiah | Although weston-launch runs on boot, successfully according to systemctl | 12:40 |
| radiofree | systemclt stop weston | 12:41 |
| radiofree | (or whatever the service name is) | 12:41 |
| radiofree | then weston --tty=1 ........ --log=/tmp/weston.log | 12:41 |
| jeremiah | yep, its weston and stopped. | 12:41 |
| jeremiah | Yep, I'll give it a tty | 12:41 |
| jonathanmaw | my investigation as to why the "back" button in gdp-hmi-panel isn't working has eventually led me to the conclusion that weston's systemd unit finishes too soon (hence gdp-hmi-launcher tries to start, and fails; hence gdp-hmi-panel gets started and stores the old PID; then gdp-hmi-launcher gets re-launched with a different PID; hence gdp-hmi-panel's signalling doesn't work) | 12:42 |
| radiofree | jonathanmaw: how are they notifying systemd that it's ready? | 12:43 |
| radiofree | sd_notify? | 12:43 |
| jeremiah | that gets me a kernel module: loaded module: libgbm_kms.so | 12:43 |
| jonathanmaw | radiofree: Type=simple, so straight away, I think | 12:43 |
| radiofree | eek | 12:43 |
| jonathanmaw | I am currently looking into whether I can change that to sd_notify successfully. | 12:43 |
| radiofree | it's a relatively quick fix to have the apps use sd_notify(READY) and change the type to "notify" | 12:43 |
| radiofree | of course you'll have to decide at what point the app is "ready" | 12:44 |
| radiofree | ah, already thinking that | 12:44 |
| radiofree | jeremiah: ok that's a good sign | 12:45 |
| radiofree | what does the weston log say? | 12:45 |
| jeremiah | :-) | 12:45 |
| radiofree | success with ivi-controller is, unhelpfully here, a black screen :) | 12:45 |
| jeremiah | Let me get a paste . . . | 12:46 |
| jeremiah | lots of output | 12:46 |
| jeremiah | http://pastebin.com/1pTHds2d | 12:47 |
| jeremiah | Now it only really seems to complain about the backlight. | 12:47 |
| CTtpollard | gunnarx: on first read-through of your suggestion, my initial feeling is we don't want board specific branches of genivi demo. However it would fit nicer into the submodule approach | 12:47 |
| radiofree | jeremiah: assuming you killed it, then it worked! | 12:48 |
| radiofree | looks like you'll need to modify the weston startup service to do the same as your command line | 12:48 |
| radiofree | with weston-launch you do it | 12:48 |
| jeremiah | radiofree: I did kill it. | 12:48 |
| radiofree | weston-launch - --log=/tmp/weston.log ........ | 12:48 |
| jeremiah | Okay, I'll hack on the service file | 12:49 |
| jeremiah | radiofree: Thanks for your help. | 12:49 |
| *** txenoo has joined #automotive | 12:50 | |
| jeremiah | this isn't good: Output LVDS1 | 12:55 |
| jeremiah | The output should be HDMI | 12:55 |
| radiofree | i can't for the life of me remember how (or if you can) specify the output to use via the command line | 12:57 |
| radiofree | you can do it via weston.ini | 12:57 |
| jeremiah | I'll just hack it via weston.ini | 12:57 |
| radiofree | assuming the gdp-* services are robust enough, you should be able to test it with `systemctl reload weston` | 12:58 |
| radiofree | if reload seems to work, but you still see a blank screen, i can give you some commands to test output | 12:59 |
| paulsherwood | jonathanmaw: if you are able to reply to tanikawa-san's email maybe worth adding genivi-projects in cc? | 12:59 |
| radiofree | providing `LayerManagerControl get scene` worked | 12:59 |
| *** waltminer has quit IRC | 13:00 | |
| jonathanmaw | paulsherwood: the multimedia requirements one? | 13:00 |
| gunnarx | CTtpollard: Yes it works with submodules, but I agree with you. As I said there is an alternative to simply put the different file versions in the layer, I was just asking a bit in #yocto. | 13:00 |
| radiofree | jeremiah: if LayerManagerControl get scene worked | 13:00 |
| radiofree | http://fpaste.org/258973/14405076/ | 13:00 |
| *** FabienH has quit IRC | 13:00 | |
| radiofree | should show you some mock navigation app | 13:00 |
| gunnarx | Since the difference is so small, might as well put it in there directly. | 13:00 |
| paulsherwood | jonathanmaw: his latest, re audio on porter | 13:01 |
| radiofree | gunnarx: difference between gdp and renesas gdp? | 13:01 |
| gunnarx | radiofree: difference between minnowboard and renesas builds, yes. | 13:01 |
| radiofree | renesas patch weston | 13:02 |
| gunnarx | It's all in the link/info I provide above if you can see that far. | 13:02 |
| radiofree | they use their own renderer https://github.com/renesas-devel/weston/blob/RCAR-GEN2/1.5.0/vsp1-dev/src/vsp-renderer.c | 13:02 |
| radiofree | i think meta-board is fine, as long as it only contains the things needed to apply over the top of the base gdp | 13:02 |
| radiofree | e.g for renesas weston, binary blobs.... | 13:03 |
| gunnarx | radiofree: I don't see this patch in meta-genivi-demo. Or where is it? We should be able to see it then - can you look? | 13:03 |
| radiofree | i think it's only in meta-renesas at the moment? | 13:04 |
| gunnarx | radiofree: Ah right. I forgot the "layered approach" for a moment, lol | 13:04 |
| gunnarx | but that's as it should be, if it is bsp specific | 13:04 |
| gunnarx | I'm right now discussing (with CTtpollard) to unify the build recipes within meta-genivi-demo for the different boards. | 13:05 |
| jeremiah | gdp-hmi-launcher is segfaulting. *sigh* | 13:05 |
| jeremiah | Time for PMO | 13:05 |
| gunnarx | time for gdb more like | 13:05 |
| gunnarx | :) | 13:05 |
| radiofree | CMU Sphinx has the best piece of advice for debugging "Troubleshooting is not rocket science. For all issues you may blame yourself. You are most likely the reason of failure." | 13:07 |
| CTtpollard | gunnarx: the patch file is very minimal, do we have any record to the reason this was implemented? including the reason for changes that are required to weston.ini? | 13:07 |
| gunnarx | weston_ini.patch? | 13:09 |
| *** FabienH has joined #automotive | 13:11 | |
| CTtpollard | yes, I don't know the background to why it was added | 13:11 |
| CTtpollard | and the changes to weston.ini for minnow | 13:11 |
| gunnarx | No I have no background but it is applied in minnowboard instructions. First of all it seems to use the _correct_ names for the .png files. So that needs to be moved over to Koelsch I would assume. | 13:11 |
| CTtpollard | yeh, and a different filepath for weston.ini | 13:12 |
| jeremiah | heh | 13:12 |
| jonathanmaw | paulsherwood: having read Tanikawa's E-mail, I think the best I can do is make sure meta-ivi and meta-renesas' maintainers are aware of this | 13:12 |
| gunnarx | It also adds some clients audiomanager and browser looks like | 13:12 |
| gunnarx | No I guess it's just different startup behavior for weston. We would have to run/compare what clients are shown and available. I don't know more but I have the address to the intel guy who wrote the wiki page. | 13:13 |
| gunnarx | CTtpollard: The final q? for me is the browser_poc_hack.patch, is it renesas unique (as the build defines it today) or can it be applied to both builds... | 13:16 |
| CTtpollard | gunnarx: I think contacting the source of the change is the place we should start, and if you have found yocto magic that would let there be to weston.ini and it would pick the correct one at bitbake time then that seems sound | 13:17 |
| CTtpollard | s/to/two | 13:18 |
| gunnarx | My understanding is that the yocto magic is there already. The koelsch subdirectory is there and it provides files that override the other ones if the MACHINE is koelsch | 13:19 |
| gunnarx | So what I'm told, you just do the same for minnowboard, but that's assuming you need any difference. I'm not convinced that we do. | 13:19 |
| gunnarx | for weston.ini that is | 13:20 |
| gunnarx | browser_poc_hack remains as the main uncertainty | 13:20 |
| gunnarx | but what's needed is really to build, test and compare | 13:21 |
| CTtpollard | looking at the build log of the weston append, the patch was was switched to the complete weston.ini file on Feb 5th 2015 | 13:22 |
| CTtpollard | if the wiki instructions were made before that, that would explain it | 13:22 |
| CTtpollard | the instructions don't make it clear if everything in the block is to be added to weston.ini | 13:24 |
| CTtpollard | i.e it's not a diff | 13:24 |
| CTtpollard | I would suspect if you were to the append at the time it was written, it would have already had the patch file inclusion there | 13:25 |
| gunnarx | That's true but I suspect only koelsch was being worked on at that time. The proposed minnowboard changes were documented March 26 it looks like | 13:26 |
| gunnarx | In any case I think a reasonable guess is that the differences are not needed. | 13:27 |
| *** wschaller has quit IRC | 13:27 | |
| CTtpollard | the .patch file was commited to the repo on Feb 5th also | 13:27 |
| gunnarx | Weird. I'm not sure there's much knowledge to be acquired here. I'm mostly concerned with what to bring into the shared version, like the correct png file names for example. I would just merge all of it together, and it needs to be tested. | 13:28 |
| CTtpollard | but the append was updated to remove the .patch 7 mins later | 13:28 |
| CTtpollard | so I want to say the patch is redundant | 13:29 |
| gunnarx | :) If you're drawing some conclusions let me know, because I stopped listening. | 13:29 |
| jeremiah | or reading, as the case may be. | 13:29 |
| gunnarx | but why does it define graphics, icons, and stuff | 13:29 |
| gunnarx | is that stuff not useful stuff? | 13:29 |
| gunnarx | I'm just trying to avoid leaving behind anything useful is all | 13:30 |
| CTtpollard | gunnarx: I think it's pretty safe to say that the correct file names are included anyway, as it has the correct background.png on koelsh without requesting the patch | 13:32 |
| fredcadete | the icons stuff are used by the desktop shell or by the example hmi-controller.so that comes from ivi-shell | 13:33 |
| gunnarx | CTtpollard: It does? I did not see this in the source. | 13:35 |
| gunnarx | "GDP_Background.png"? I see it in the SRC_URI of the bb, but how does it get into weston.ini without the patch? | 13:36 |
| *** waltminer has joined #automotive | 13:37 | |
| CTtpollard | well as fredcadete says, if ivi-shell is providing them there is no need to provide them via the weston patch | 13:42 |
| *** gunnarx has quit IRC | 13:44 | |
| *** gunnarx has joined #automotive | 13:45 | |
| CTtpollard | I'm not sure if you managed to see what I last commented gunnarx | 13:46 |
| *** wschaller has joined #automotive | 13:47 | |
| CTtpollard | the weston recipe in meta-ivi contains 31.patch files.... | 13:51 |
| *** gunnarx has quit IRC | 13:53 | |
| *** gunnarx has joined #automotive | 13:54 | |
| jonathanmaw | CTtpollard: yep, I think poky defines weston as a tarball of 1.5.0, plus a handful of patches. Then meta-ivi adds 31+ patches, then meta-genivi-demo adds a couple more | 13:58 |
| jonathanmaw | so it takes a bit of work to reconstruct what the source tree looks like (though yocto probably has a clever tool to do it for them, somewhere) | 13:58 |
| *** yoshi_ito has joined #automotive | 14:01 | |
| *** gunnarx has quit IRC | 14:01 | |
| *** gunnarx has joined #automotive | 14:02 | |
| *** FabienH has quit IRC | 14:02 | |
| CTtpollard | I don't know what the best course of action is without knowing the reason a past maintainer / wiki editor did it in the first place | 14:05 |
| waltminer | *** AGL Layer Desiign Call started for those interested in joining *** | 14:06 |
| waltminer | Topic - How is gerrit process working? | 14:07 |
| *** Tanikawa has joined #automotive | 14:07 | |
| *** gunnarx has quit IRC | 14:09 | |
| *** gunnarx has joined #automotive | 14:09 | |
| waltminer | Tanikawa said we need tag naming rules and branch name rules | 14:15 |
| *** bbranch has joined #automotive | 14:17 | |
| waltminer | Jan-Simon would be helpful to get automated build running so we can use that as a check in the review process | 14:17 |
| * CTtpollard notices that the weston patch was removed from the append before the 0001-configure patch was added, so is lost to why the instructions list both | 14:17 | |
| CTtpollard | and systemd | 14:18 |
| waltminer | Paul discussing Jenkins vs other CI systems | 14:18 |
| waltminer | Buildbot from Yocto | 14:18 |
| *** gunnarx has quit IRC | 14:19 | |
| *** gunnarx has joined #automotive | 14:19 | |
| waltminer | Tanikawa can you add your concern about committer ID | 14:20 |
| Tanikawa | Curretly, I used commit ID to sync multiple repositories depend each others and decribes it into README.md | 14:21 |
| waltminer | Jan-Simon suggests that we get a jenkins instance up and running for the gerrit workflow | 14:22 |
| waltminer | JSM can configure the server on LF server | 14:22 |
| Tanikawa | But when push multiple patches into multiple repositories, it is a bit troublesome. So I think we should use tag to point and sync multiple repositories. | 14:22 |
| paulsherwood | +1, sincie it will provide build validation for all reviews | 14:22 |
| waltminer | Walt to add Jira issue for JSM to perform the work | 14:22 |
| paulsherwood | s/for/to support/ | 14:22 |
| *** gunnarx has quit IRC | 14:24 | |
| *** gunnarx has joined #automotive | 14:24 | |
| *** txenoo has left #automotive | 14:25 | |
| waltminer | Tanikawa suggest we have two branches - master and development | 14:25 |
| waltminer | possibly a third for CI | 14:26 |
| waltminer | Ci / release | 14:26 |
| * paulsherwood notes that linux kernel and other projects don't do it that way | 14:26 | |
| waltminer | This is a more clearcase type of thinking | 14:27 |
| gunnarx | CTtpollard: sorry, in a meeting, missed the last discussion. I saw "ivi-shell" providing them. So is our understanding that the png files are installed by the weston recipe but used from ivi-shell? | 14:27 |
| CTtpollard | gunnarx: i'm trying to work out currently why the minnowguide seems to have cherrypicked what patch of the append it is running on | 14:32 |
| *** gunnarx has quit IRC | 14:33 | |
| CTtpollard | the history of the append shows it move to using weston.ini, and then moving to use systemd, | 14:33 |
| *** gunnarx has joined #automotive | 14:33 | |
| CTtpollard | the append for the minnow had the systemd stuff, but not the weston.ini change that came before it | 14:33 |
| yoshi_ito | How about the risk of "build error" on master branch when we use just 2 branch to manage repository. | 14:34 |
| yoshi_ito | May we use another branch than master to prepend build error on master branch? | 14:34 |
| paulsherwood | yoshi_ito: the point would be that gerrit can accept a +1 from the CI, so it should have been built successfully before merge | 14:34 |
| paulsherwood | for example in review.openstack.org i believe all master commits are gated *only* on the ci | 14:35 |
| paulsherwood | but i understand Tanikawa's concerns about meta-data being different from code | 14:35 |
| yoshi_ito | However as Tanikawa-san mentioned in voice conf., sometimes meta files affects large scale build error. | 14:36 |
| gunnarx | Sorry about my bad connection. CTtpollard, I guess people have been working on different versions. We have not had a really shared project. | 14:36 |
| CTtpollard | gunnarx: no worries, I had constant disconnections yesterday afternoon from freenode | 14:37 |
| gunnarx | Speaking of which, let's discuss the setup of "official" submodule based project, since I think you support that. | 14:39 |
| CTtpollard | I can only suggest we try and build the minnow image from the default append koelsh and porter use, (and qemu) and see if it works | 14:40 |
| CTtpollard | gunarx: sure, I've beem a supporter for quite a while | 14:41 |
| CTtpollard | *been | 14:41 |
| gunnarx | OK so I think it's pretty much ready. Everything is now based on upstream layers. The readme can be improved. | 14:44 |
| gunnarx | And I think we should have a quick mailing list poll about the name maybe? | 14:44 |
| gunnarx | A nicer name, basically just "genivi demo platform" would make sense to me | 14:44 |
| gunnarx | jeremiah can set up the git once we have a name. I might retain a github fork for some experiments but that's normal | 14:46 |
| CTtpollard | gunnarx: how do you feel about keeping the manual instructions along side instructions for the submodule method? | 14:46 |
| waltminer | Ito-san said that gstreamer 1.2.3 should be used to enable hardware codecs | 14:47 |
| waltminer | Tanikawa-san said there are no instructions on how to enable the hw codecs | 14:47 |
| gunnarx | it's a matter of maintenance | 14:49 |
| gunnarx | and to some extent the instructions are just duplicating the same info, so I'd say they should probably be rewritten and minimized at least. | 14:49 |
| gunnarx | Some instructions are of course necessary but checking out exact hashes I doubt people will miss that much, but we should ask our community I guess (genivi-projects) | 14:49 |
| gunnarx | and various patches / manual modifications are now being committed instead... | 14:49 |
| *** mdunford has quit IRC | 14:50 | |
| waltminer | tanikawa-san has some things to try based on informaiton from Ito-san | 14:50 |
| CTtpollard | gunnarx: +1 on reduction of duplicates | 14:50 |
| waltminer | next up after gstreamer is qt to finish phase 1 | 14:52 |
| CTtpollard | especially with porter and koelsch | 14:52 |
| gunnarx | OK, let's present our plans on mail and ask for feedback. I think this is a proposal from you as maintainer, not from me? ;) | 14:53 |
| gunnarx | I make enough noise... | 14:53 |
| CTtpollard | constructive noise | 14:53 |
| *** wschaller_ has joined #automotive | 14:54 | |
| *** wschaller has quit IRC | 14:58 | |
| *** yoshi_ito has left #automotive | 15:00 | |
| *** mdunford has joined #automotive | 15:05 | |
| *** bbranch has quit IRC | 15:07 | |
| gunnarx | +license information ought to be added | 15:08 |
| *** txenoo has joined #automotive | 15:09 | |
| *** mdunford has quit IRC | 15:10 | |
| gunnarx | CTtpollard: Something to consider about workflow, I removed the master branch to clarify it should not be relied upon. However, for changes that are useful for all variants I find it quite nice to have a branch to put them on, and then merge into each of the specific branches | 15:10 |
| CTtpollard | gunnarx: the steps to place the propertairy renesas drivers in the correct dirs still needs to be included | 15:10 |
| CTtpollard | gunnarx: master would fit the purpose for global changes I think | 15:11 |
| CTtpollard | what do you think? | 15:11 |
| gunnarx | But to do that you need to have a common parent from which all the build-specific changes have been applied after. Basically when you merge from master, only the new diff should come along. If I were to merge from koelsch to minnowboard for example, the minnowboard branch would be messed up with koelsch stuff | 15:13 |
| gunnarx | that no longer exists :) | 15:13 |
| gunnarx | re: renesas, yes I agree | 15:13 |
| gunnarx | how about adding a directory where such blobs live (but not committing them) | 15:13 |
| gunnarx | binary-tmp according to the description. but it could be called firmware / extras ... | 15:13 |
| gunnarx | resources ? | 15:13 |
| gunnarx | then you could commit the copying scripts. The only thing the user needs to do is drop the binaries into the directory. | 15:13 |
| gunnarx | so to recap, recreating a base branch "master" from which all other changes stem might need a little trickery now. | 15:14 |
| gunnarx | naming: I'd propose calling that one develop or such. People have different workflows but many assume that something called "master" is buildable, which would not be guaranteed here, IMO | 15:14 |
| *** jlrmagnus has joined #automotive | 15:15 | |
| jlrmagnus | Morning | 15:15 |
| gunnarx | morning, Mr CAN frames :) | 15:16 |
| CTtpollard | gunnarx: the copying scripts already come with meta-renesas so there is probably no point commiting it in two locations | 15:16 |
| jlrmagnus | Tsk. You should know that I know UDP frames as well. | 15:17 |
| jlrmagnus | Jag har många strängar på min lyra. | 15:17 |
| CTtpollard | gunnarx: develop / core or suites me | 15:19 |
| *** gunnarx has quit IRC | 15:21 | |
| *** gunnarx has joined #automotive | 15:21 | |
| *** mdunford has joined #automotive | 15:23 | |
| CTtpollard | gunnarx: what changes do you see multiple boards both requiring, as I see it it would be shared changes to local.conf? | 15:30 |
| gunnarx | just in general. | 15:31 |
| *** fredcadete has quit IRC | 15:35 | |
| *** gunnarx has quit IRC | 15:35 | |
| *** gunnarx has joined #automotive | 15:36 | |
| gunnarx | all I'm saying is that if we wanted to create a devel branch it needs to branch off early in the tree like that, if I'm not totally confused about things, which happens | 15:36 |
| gunnarx | you could rebase on top of master also each time to avoid the merge commits, but the history would be quite misleading. | 15:37 |
| gunnarx | s/master/devel/ | 15:44 |
| CTtpollard | yeh it's an interesting approach to think about | 15:45 |
| gunnarx | like I said that is just a result of the original master having been thrown away really, you would need to find the right place to recreate it | 15:48 |
| gunnarx | otherwise forget that, make common changes on any of the branches and cherryppick it over to the others. It generally works too but does not feel as clean somehow. | 15:48 |
| CTtpollard | takes away the git magic | 15:53 |
| *** waltminer has quit IRC | 15:57 | |
| CTtpollard | gunnarx: I think we've hashed out enough to put into an email, it'll be the first job on my list for the morning | 15:59 |
| gunnarx | good man | 15:59 |
| gunnarx | today is tuesday, therefore a large burger for dinner. good night all. | 16:00 |
| CTtpollard | ah, it's international burger day on thursday so I shall wait for that | 16:00 |
| CTtpollard | night gunnarx | 16:00 |
| *** gunnarx has quit IRC | 16:01 | |
| CTtpollard | ah, seems to be a UK thing | 16:01 |
| myself | not very international! | 16:01 |
| CTtpollard | indeed, 28th May seems to be the international day. I shall celebrate both | 16:02 |
| *** mdunford has quit IRC | 16:02 | |
| *** Tanikawa has quit IRC | 16:16 | |
| *** mdunford has joined #automotive | 16:21 | |
| *** mdunford has quit IRC | 16:25 | |
| *** mdunford has joined #automotive | 16:26 | |
| *** josemic has joined #automotive | 16:26 | |
| *** mdunford has quit IRC | 16:38 | |
| *** jonathanmaw has quit IRC | 16:44 | |
| *** bbranch has joined #automotive | 16:47 | |
| *** mdunford has joined #automotive | 16:52 | |
| *** bbranch has quit IRC | 16:57 | |
| *** Egy has joined #automotive | 17:01 | |
| *** wschaller_ has quit IRC | 17:04 | |
| *** mdunford has quit IRC | 17:19 | |
| *** Egy has quit IRC | 17:28 | |
| *** mdunford has joined #automotive | 17:30 | |
| *** txenoo has left #automotive | 17:52 | |
| *** bbranch has joined #automotive | 17:53 | |
| *** txenoo has joined #automotive | 17:54 | |
| *** bbranch has quit IRC | 17:56 | |
| *** bbranch1 has joined #automotive | 17:59 | |
| *** txenoo has left #automotive | 18:06 | |
| *** waltminer has joined #automotive | 18:09 | |
| *** Egy has joined #automotive | 18:32 | |
| *** mdunford has quit IRC | 18:35 | |
| *** bbranch1 has quit IRC | 18:42 | |
| *** stevel_ has joined #automotive | 19:06 | |
| stevel_ | quick comment before I go home... | 19:07 |
| *** josemic has quit IRC | 19:07 | |
| stevel_ | If the koelsch weston.ini comments are referring to the following in the GDP src /recipes-graphics/wayland/weston/koelsch/weston.ini | 19:07 |
| *** waltminer has quit IRC | 19:08 | |
| *** apinheiro has quit IRC | 19:08 | |
| stevel_ | then that is not a requirement of the BSP. Short story is that it was put there by the then maintainer as a workaround for a perceived issue in Layer Manager. | 19:08 |
| stevel_ | Feeling was that LM was not consistentally enumerating displays out to the same values and this may be part of the early display issues that were seen. | 19:09 |
| stevel_ | As GDP started on Koeslch that weston.ini was created to disable the other displays so the issue would not appear. | 19:09 |
| stevel_ | Personally, although that may or may not be an issue in LM, I don't think that was the cause of the gfx issues and could be safely removed. That appears to be born out by the fact Porter does not have one. | 19:10 |
| stevel_ | Although as Tom says would also make sense to follow up with the originator. | 19:11 |
| *** stevel_ has quit IRC | 19:14 | |
| *** waltminer has joined #automotive | 19:23 | |
| *** bbranch has joined #automotive | 19:57 | |
| *** waltminer has quit IRC | 20:56 | |
| *** bbranch has quit IRC | 21:58 | |
| *** waltminer has joined #automotive | 22:00 | |
| *** bbranch has joined #automotive | 22:03 | |
| *** Egy has quit IRC | 22:54 | |
| *** bbranch has left #automotive | 23:01 | |
| *** waltminer has quit IRC | 23:20 | |
| *** jlrmagnus has quit IRC | 23:23 | |
| *** waltminer has joined #automotive | 23:39 | |
| *** bbranch has joined #automotive | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!