| sicelo | Wizzup: so in arm-sdk, i see a function for building the kernel for a device. and i also see the kernel exists as a standalone package. how does this work? or the function in arm-sdk is not used? | 11:01 |
|---|---|---|
| sicelo | at least n900 in arm-sdk looks like it would build old kernel ... so i guess that's not used? | 11:03 |
| Wizzup | sicelo: we don't use arm-sdk way to build kernel | 11:13 |
| Wizzup | that's a devuan-only thing | 11:13 |
| Wizzup | ew just build a package and install/depend on it during out bootstrap | 11:13 |
| sicelo | ah, ok | 11:16 |
| Wizzup | do you need help with l5 things? | 11:18 |
| halftux | Wizzup: what is the helper command to use with the arm-sdk is it build_image_dist or build_arm_dist? | 11:42 |
| sicelo | build_arm_dist | 11:43 |
| halftux | is the arm64-generic build already fixed? | 11:44 |
| Wizzup | halftux: let me check @ generic | 11:55 |
| Wizzup | https://phoenix.maemo.org/view/Images/job/leste-image-arm64generic/6/ | 11:55 |
| Wizzup | looksl ike it's not set to auto build, heh | 11:55 |
| Wizzup | we can also make a daedalus one if this passes | 11:58 |
| halftux | sure fine with me | 11:58 |
| halftux | so the kernel needs a device specific package. Is there also a package for the device tree files? | 12:00 |
| sicelo | the dts is generally part of the kernel package | 12:03 |
| Wizzup | yes, this is leste-config-xyz | 12:04 |
| Wizzup | oh, the device tree is part of the kernel | 12:04 |
| Wizzup | honestly when it comes to specific kernel stuff I'm hoping we could just merge the support into a single kernel package since they are a pain to maintain, do you have a link to a kernel tree and config somewhere? | 12:04 |
| sicelo | what do you mean by single kernel package? | 12:06 |
| Wizzup | not to have a kernel per device, but a kernel for multiple devices | 12:07 |
| Wizzup | like we have a single one for all our omap devices, ideally we'd also have a single one for arm64 devices | 12:07 |
| Wizzup | and we could extend the omap kernel to other 32 bit devices | 12:07 |
| Wizzup | we modify builddeb scripts in kernel, so it's not a bunch of extra work | 12:08 |
| sicelo | unfortunately that can clash. even within omap, for some reason d4 doesn't want kernel with THUMB, and we would like to have it for N900. as a result, we currently have it disabled for N900 | 12:08 |
| sicelo | anyway, for L5 i prefer this kernel, https://salsa.debian.org/Mobian-team/devices/kernels/librem5-linux, and it's the one I've been testing with so far ... unless halftux prefers something else | 12:09 |
| Wizzup | for things that need their own kernel we will do it of course | 12:10 |
| Wizzup | I don't know what halftux is working on, I assumed (maybe wrong) raspi | 12:10 |
| sicelo | L5 as well, afaik :-) | 12:13 |
| Wizzup | ok | 12:17 |
| halftux | I only wanted to know how everything works in general to add maybe new devices by myself. I have odroid c4 in mind and yes I have L5 and raspi2-5 at home | 12:46 |
| halftux | when I tried to build the raspi 3 and 4 it arm-sdk builds an old kernel from raspi git so I guess I need to remove this custom build and provide a package | 12:48 |
| halftux | for odroid c4 I am looking atm at armbian how they build the kernel, many patches are involved | 12:54 |
| halftux | https://github.com/armbian/build/tree/main/patch/kernel/archive/meson-6.12 | 12:59 |
| Wizzup | halftux: yes, it would be great to get a proper kernel for the raspis | 13:08 |
| Wizzup | ideally just mainline | 13:09 |
| Wizzup | I would suggest to start with raspi over odroid-c4, since it is hopefully easier and more impactful | 13:10 |
| halftux | ok checkout daedalus image-builder and running arm64-generic then trying to get raspi4 mainline kernel running on it | 13:51 |
| Wizzup | cool :) | 13:58 |
| Wizzup | freemangordon: this is not something we'd want to work on now, but nlnet is asking if I know what it would take to move us to gtk3 | 14:33 |
| Wizzup | as far as I know we got quite far, even have the input module working, but we don't have the theming working yet, which I think in turn made us not move forward, we might be missing some fremantle specific widgets too. | 14:34 |
| arno11 | Wizzup: AI could help a lot imo | 14:54 |
| Wizzup | arno11: with what, themes? | 14:56 |
| arno11 | idk exactly for theming but i asked leste assistant, seems able to directly convert gtk2 code to gtk3. | 14:58 |
| arno11 | if you have specific example to provide and test, i can try | 14:59 |
| Wizzup | the challenge isn't really porting gtk2 to gtk3, it's the fact that we have themes, widgets, etc that need to be ported | 14:59 |
| Wizzup | to a different plugin architecture at that | 14:59 |
| Wizzup | some of this work was done already, but it's not like 'port this gui application from gtk2 to gtk3' | 14:59 |
| arno11 | i understand that | 15:00 |
| arno11 | i just mean: gtk2 to gtk3 sounds like a monumental work for just few devs with not a lot of spare time. | 15:08 |
| Wizzup | yup, hence not a now thing, but having an idea of how much work it would be is helpful | 15:09 |
| Wizzup | and nlnet might offer to fund some of the work | 15:10 |
| arno11 | i see. | 15:10 |
| inky | Wizzup, i guess no hope for rafael to do that in close time, i think if you pull the pinephone kernel and send to phoenix, updated one will work, i can test, my pp is on devel. | 15:20 |
| Wizzup | just a test | 15:50 |
| Wizzup | https://git.maemo.org/merlijn/test-bugtracker | 15:50 |
| Wizzup | oh, ha, it didn't copy the issues | 15:50 |
| Wizzup | let me retry xD | 15:50 |
| halftux | error in: strapdir to image no space left on device | 15:54 |
| halftux | /dev/loop0p2 2.7G used 2.7G | 15:55 |
| halftux | when trying to build generic arm64 with daedalus | 15:55 |
| Wizzup | halftux: I have a fix for this, let me see | 15:57 |
| Wizzup | halftux: edit arm-sdk/boards/arm64-generic.sh and change size=3000 | 15:57 |
| halftux | your bugtracker is brewing me a green coffee | 15:57 |
| Wizzup | make it size=4000 | 15:57 |
| halftux | ok | 15:58 |
| Wizzup | halftux: yeah it's still migrating | 15:58 |
| Wizzup | it's supposed to be a copy of the github bugtracker | 15:58 |
| halftux | or better it is a green tea | 15:58 |
| Wizzup | gitea | 15:58 |
| Wizzup | looks like the chimaera arm64 generic 64 build passed | 16:00 |
| arno11 | yes indeed | 16:00 |
| Wizzup | shall I build a daedalus image on ci as well (try to?) | 16:00 |
| Wizzup | https://phoenix.maemo.org/view/Images/job/leste-image-arm64generic-daedalus/ | 16:02 |
| halftux | hmm now I am getting failure resolving maedevu.maemo.org but in another terminal I could ping the server | 16:04 |
| halftux | chroot-script-intern fourthstage | 16:06 |
| Wizzup | could be a temporary issue? | 16:06 |
| Wizzup | oh, there is one part where that does happen, but it should not be fatal | 16:06 |
| Wizzup | that gitea migration is surely taking a while | 16:07 |
| Wizzup | lol, just as I said that it finished | 16:08 |
| Wizzup | https://git.maemo.org/merlijn/test-bugtracker/issues/751 | 16:08 |
| Wizzup | https://git.maemo.org/merlijn/test-bugtracker/issues/766 | 16:09 |
| Wizzup | (some examples) | 16:09 |
| Wizzup | the milestones are there too: https://git.maemo.org/merlijn/test-bugtracker/issues?type=all&state=open&milestone=10&project=0&assignee=0 | 16:10 |
| arno11 | cool, really similar to github | 16:11 |
| arno11 | i mean, the ui | 16:12 |
| Wizzup | yeah | 16:12 |
| Wizzup | I will see if folks can use their github account to log in too, at least for say reporting issues and such | 16:12 |
| arno11 | ok | 16:13 |
| Wizzup | I guess maybe the next step is to migrate some repos en masse | 16:14 |
| Wizzup | just to see what it looks like | 16:14 |
| arno11 | yeah | 16:15 |
| freemangordon | Wizzup: we will have to port libhildon to gtk3 | 16:17 |
| freemangordon | porting hildon-desktop should be trivial, AFAIK | 16:19 |
| freemangordon | however, this port cannot be done as a one-shot task | 16:20 |
| freemangordon | we shall do on phases | 16:20 |
| Wizzup | yes, but what would the phases look like I am wondering | 16:30 |
| freemangordon | libhildon3 is the first one | 16:31 |
| Wizzup | I think we already have that building | 16:31 |
| freemangordon | then themes | 16:31 |
| freemangordon | but no idea how complete it is | 16:31 |
| Wizzup | well | 16:37 |
| Wizzup | that depends on how we do it | 16:37 |
| Wizzup | I don't think we should try to parse gtk2 themes and make them work in gtk3 | 16:37 |
| Wizzup | I think we just need to make themes thegtk3 way | 16:37 |
| Wizzup | which is html/css style, rather than small bit maps | 16:38 |
| Wizzup | in which case we just need to ... make them | 16:38 |
| Wizzup | or we have to write code to parse/understand gtk2 pixmap themes and somehow render them in gtk3 | 16:38 |
| Wizzup | btw, random example repo: https://git.maemo.org/merlijn/alarmd | 16:38 |
| Wizzup | oh it's private,sec | 16:39 |
| Wizzup | fixed | 16:40 |
| arno11 | works fine | 16:41 |
| Wizzup | I think some migrations are not ok | 16:45 |
| Wizzup | https://git.maemo.org/merlijn/hildon-status-menu | 16:45 |
| Wizzup | this seems to have used whatever we forked it from | 16:45 |
| Wizzup | not https://github.com/maemo-leste/hildon-status-menu | 16:46 |
| arno11 | lot of branches are missing | 16:46 |
| arno11 | ah yeah even master is wrong | 16:48 |
| Wizzup | same here | 16:49 |
| Wizzup | https://github.com/maemo-leste/libconic | 16:49 |
| Wizzup | https://git.maemo.org/merlijn/libconic | 16:49 |
| Wizzup | 'original_url': 'https://github.com/fremantle-gtk2/libconic.git' | 16:49 |
| Wizzup | the problem is that our repos seem to be 'forked' from fremantle-gtk2 | 16:49 |
| arno11 | same for h-s-m | 16:50 |
| Wizzup | yeah a bunch I bet | 16:51 |
| Wizzup | I stopped the mass test import for now | 16:51 |
| dsc_ | does it import users too | 16:54 |
| dsc_ | and issues | 16:54 |
| dsc_ | well, mock users | 16:54 |
| dsc_ | else nobody can ever register, unless gitea can merge that | 16:55 |
| dsc_ | btw Gitea has this really annoying 'archiving' feature (that I dont really understand) that loves to fill your HDD until it is full | 16:56 |
| dsc_ | it has a job for that | 16:56 |
| dsc_ | please look into `cron.archive_cleanup` and `cron.delete_repo_archives` config options | 16:57 |
| dsc_ | because it will 100% kill the instance lol | 16:58 |
| dsc_ | (speaking from experience) | 16:58 |
| Wizzup | hold up | 16:58 |
| Wizzup | I already linked some examples and the test-bugtracker clearly shows the comments/issues imported from github | 16:58 |
| Wizzup | but that is not the issue here | 16:59 |
| Wizzup | the issue is that gitea seems to 'follow' github's fork network | 16:59 |
| Wizzup | so it's not cloning say github.com/maemo-leste/libconic | 16:59 |
| Wizzup | but rather github.com/fremantle-gtk2/libconic | 16:59 |
| Wizzup | or even fremantle-gtk3/libconic in this case | 17:00 |
| dsc_ | strange | 17:00 |
| dsc_ | why would it follow | 17:00 |
| Wizzup | no idea, but that's what it seems to do, at least with the script that I used | 17:00 |
| Wizzup | let me try manually for libconic | 17:00 |
| Wizzup | when I do it manually it seems better https://git.maemo.org/merlijn/libconic | 17:01 |
| arno11 | indeed, better | 17:03 |
| arno11 | at least, all branches seem there | 17:03 |
| Wizzup | mabye the github api returns the wrong urls | 17:03 |
| Wizzup | yeah I see it | 17:07 |
| Wizzup | I'll fix it in a bit | 17:07 |
| arno11 | able to see history commits but just from a folder or a file. no way to see full history | 17:10 |
| Wizzup | can you share a link | 17:13 |
| arno11 | *from the website, works from command line | 17:16 |
| arno11 | a link about what ? | 17:19 |
| arno11 | ah you mean showing commits history ? | 17:20 |
| Wizzup | https://git.maemo.org/merlijn/libhildonfm/commits/branch/master/configure.ac | 17:22 |
| Wizzup | https://git.maemo.org/merlijn/libhildonfm/commits/branch/master | 17:22 |
| arno11 | ah yes :) | 17:26 |
| Wizzup | I added a few more repos, they look better now | 17:26 |
| arno11 | cool | 17:27 |
| Wizzup | (ofc we wouldn't have this under /merlijn/, just testing) | 17:27 |
| arno11 | yes ofc | 17:27 |
| Wizzup | trying hildon-desktop now to see if it migrates both open and closed prs | 17:28 |
| arno11 | can't see any prs | 17:31 |
| arno11 | *in /merlijn/hildon-desktop/issues | 17:32 |
| arno11 | scratch that | 17:33 |
| arno11 | /merlijn/hildon-desktop/pulls is ok | 17:33 |
| arno11 | similar to github | 17:34 |
| Wizzup | cool | 17:34 |
| Wizzup | it's impressive, really | 17:34 |
| Wizzup | https://git.maemo.org/merlijn/hildon-desktop/pulls?type=all&state=closed&milestone=0&project=0&assignee=0 | 17:34 |
| mkf | hello. | 17:37 |
| mkf | i've (finally) managed to get that android tablet running once again | 17:37 |
| mkf | it's touch rougly works now, albeit it's still bit off. | 17:37 |
| mkf | is there any way to calibrate it? i'm using firmware from android, but it's flakey. | 17:38 |
| freemangordon | hmm, are you sure you use the correct FW? | 17:40 |
| freemangordon | please check in fex | 17:40 |
| freemangordon | (or whatever that strange format was called) | 17:41 |
| mkf | yes, sys_config.fex apperently uses that fw | 17:42 |
| freemangordon | does it work properly under android? | 17:42 |
| mkf | and for all fws inside that rom, this one behaves works best | 17:42 |
| mkf | yes, this works fine in android. | 17:42 |
| freemangordon | hmm, could be driver issue then | 17:42 |
| mkf | tho, it's not the stock rom this device shipped with, but touch works fine in this rom dump. | 17:43 |
| freemangordon | sorry, can;t help with calib | 17:43 |
| mkf | no problem. would xinput_calib work? i've tried it but not sure if it worked | 17:43 |
| Wizzup | I think xinput calibrator doesn't work if libinput is used (IIRC), but should work for other things | 17:44 |
| Wizzup | arno11: btw, jfyi (since you're watching this), it looks like the mass importer still doesn't actually properly import MRs and such | 17:45 |
| Wizzup | but when I do it through the UI it does :D | 17:45 |
| mkf | i see. do beowulf images use libinput? | 17:45 |
| mkf | (i'm using old sunxi image until i manage to get usb and touch running, it's a bit easier to debug with those.) | 17:46 |
| arno11 | Wizzup: ah ok | 17:47 |
| sicelo | mkf: i saw this some weeks ago, https://github.com/kreijack/xlibinput_calibrator ... worth a try | 17:49 |
| mkf | sicelo: thanks. compiling it atm :D | 17:49 |
| mkf | god bless qemu authors, they surely made lives easier. | 17:49 |
| Wizzup | mkf: no they should not, beowulf is also really old fwiw | 17:52 |
| Wizzup | I know you know, but like, we might not remember the status of things on beowulf | 17:52 |
| mkf | how can i check? also, how can i upgrade? | 17:55 |
| mkf | it says using input driver libinput for silead_ts | 17:56 |
| mkf | (Xorg.0.log) | 17:56 |
| halftux | maybe input_set_abs_params could help for ts problem | 17:58 |
| Wizzup | arno11: ok, found the issue, the script I am using does not pass along my token to gh so it can't pull in the other stuff, will test fix | 17:58 |
| arno11 | ok cool | 18:00 |
| mkf | sicelo: xlibinput_calibrator fixed it. \o/ | 18:03 |
| sicelo | lovely | 18:19 |
| Wizzup | arno11: k, fixed, will continue with this later, let me see if I can figure out a fix to the fstab thing now | 18:55 |
| arno11 | Wizzup: oh great | 19:32 |
| arno11 | i think there is a commit to revert | 19:32 |
| arno11 | commit 621d8ca in arm-sdk, no ? | 19:34 |
| mkf | is gconf problem solved? | 19:37 |
| Wizzup | mkf: in daeadalus, yes | 19:37 |
| Wizzup | and maybe in chimaera | 19:39 |
| Wizzup | definitely not in beowulf | 19:39 |
| Wizzup | (but you can work around it) | 19:39 |
| Wizzup | arno11: hm... maybre | 19:39 |
| mkf | i'm upgrading to chimaera. | 19:42 |
| mkf | did the workaround tho. | 19:42 |
| Wizzup | btw, here's the gtk3 issue we had: https://git.maemo.org/merlijn/test-bugtracker/issues/261 | 19:50 |
| Wizzup | or milestone I guess | 19:50 |
| Wizzup | arno11: I think that is one part of the puzzle | 19:50 |
| Wizzup | arno11: we can certainly remove it and see if things just work | 19:51 |
| Wizzup | revert it | 19:51 |
| arno11 | ok, at least, from jenkins console, that seems to be the last thing which overrides fstab before the end of the build | 19:52 |
| arno11 | but maybe i'm wrong | 19:52 |
| mkf | is gittea a mirror of github or complete replacement? | 20:10 |
| mkf | Err:1 https://maedevu.maemo.org/leste chimaera InRelease | 20:23 |
| mkf | The following signatures were invalid: EXPKEYSIG 545FEC4E0927F6FD spam1@wizzup.org | 20:23 |
| mkf | ^ any ideas? | 20:23 |
| mkf | installing maemo-keyring manually fixed it \o/ | 20:29 |
| arno11 | mkf: ATM gitea stuff is just a test | 20:33 |
| arno11 | *with leste | 20:33 |
| mkf | wow, chimaera update was sure worth it, got battery indicator. :D | 21:30 |
| arno11 | Wizzup: before building a new img, maybe we should add transitions, no ? | 21:54 |
| arno11 | ah cool, daedalus arm64 is ok | 22:12 |
| Wizzup | arno11: well ... ok we don't know, but it built :) | 22:32 |
| Wizzup | arno11: yeah we should fix fstab and add a few more things | 22:32 |
| arno11 | yep | 22:41 |
| arno11 | renice xorg, mce, H-D and systemui could be useful but it works only with sudo so i've been only able to renice mce on boot | 22:47 |
| arno11 | hmm not renice, nice sorry | 22:48 |
| arno11 | so the problem is not sudo | 22:49 |
| arno11 | ah now i remember | 22:50 |
| arno11 | indeed even with nice we need sudo | 22:50 |
| arno11 | to get prio @-8 or -9 | 22:51 |
| arno11 | that's the problem | 22:51 |
| arno11 | so atm it works on boot only with mce | 22:52 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!