| mkfx | neat. got leste running on android. | 06:34 |
|---|---|---|
| mkfx | is it fesiable to have a "android" port or it's against goals of leste? (native and mainline only?) | 06:37 |
| freemangordon | mkfx: we are fine, the issue is that such port has maybe 30% of the functionality | 12:04 |
| Wizzup | mkfx: please document the process | 13:06 |
| Wizzup | and make it replicable | 13:06 |
| arno11 | mkfx: did you use proot or chroot ? or another method ? | 13:34 |
| arno11 | btw under proot, i see let's say 85-90% of things working. even mobile data works through the dummy network | 13:35 |
| sicelo | Wizzup: let me try today (re:maeotp icon) :-) | 13:40 |
| sicelo | but the name ... i like it as is, https://github.com/maemo-leste/bugtracker/issues/736 | 13:41 |
| Wizzup | sure @ name | 13:44 |
| dsc_ | I didnt test QML @ vkb, ill try | 14:43 |
| dsc_ | ugh :P | 15:17 |
| Wizzup | hm? | 15:17 |
| Wizzup | arno11: it would be really nice to get documentation on how to make a proot image | 15:17 |
| Wizzup | so that we can make it in CI | 15:17 |
| arno11 | Wizzup: agree, but idk what exactly diejuse modified in the original old pi3 ascii img | 15:22 |
| dsc_ | Wizzup: some complications with the QML stuff because hildon steals focus | 15:24 |
| dsc_ | ill fix it.. | 15:24 |
| Wizzup | freemangordon: I ported the tp-haze python scripts to py3 | 15:27 |
| Wizzup | should we try to send that upstream? | 15:27 |
| Wizzup | I don't know if they work btw, would be worth checking on chimaera whether the results are the same | 15:28 |
| sicelo | arno11: didn't diejeuse document it somewhat? | 15:30 |
| arno11 | sicelo: somewhat yes, but not the part about img modifications | 15:31 |
| arno11 | but it was a very old img | 15:32 |
| arno11 | maybe it works better with a recent one, with few modifs in the starting script under termux | 15:33 |
| arno11 | i really don't know. i've been able to dist-upgrade to chimaera -devel so i didn't try anything else | 15:34 |
| arno11 | (i mean i didn t try the process with another img) | 15:36 |
| Wizzup | freemangordon: maybe I am confused and looking at the wrong tp haze branch, I thought you already build a newer one with more patches | 15:36 |
| Wizzup | freemangordon: ah, yes, indeed, I used the wrong branch | 15:36 |
| Wizzup | ok, that was a wasted time :D | 15:37 |
| Wizzup | too bad master wasn't in sync | 15:37 |
| arno11 | Wizzup: i still need 1-2 weeks for maemo-gpt and then i'll have a look @proot stuff | 15:46 |
| Wizzup | freemangordon: pulling in sfos patch for tp-ring python2 code to port it to py3 | 15:51 |
| Wizzup | we should rebase on sfos later, but they have deps on their own internal daemons in it now | 15:52 |
| dsc_ | I need to modify HIM to make it work with QML, even then the fix is hacky | 15:53 |
| Wizzup | dsc_: got some details? | 15:53 |
| dsc_ | keyboard raises, steals focus from Qt app, keyboard goes away, focus re-enters, which triggers the keyboard again | 15:54 |
| dsc_ | this does not happen with regular widgets because they do not call showInputPanel() a second time | 15:54 |
| dsc_ | proposed solution is to add a command to HIM "keyboard close" so I can throttle it | 15:55 |
| dsc_ | another possible solution is not to steal focus | 15:55 |
| dsc_ | https://github.com/community-ssu/qt-x11-maemo/blob/b3282763fc0a95b7f6eddc7b8caf7406003975ac/src/gui/inputmethod/qhildoninputcontext_x11.cpp#L492-L495 | 15:56 |
| dsc_ | this work-around is not possible, because Qt5 has a new api | 15:56 |
| dsc_ | the new api = showInputPanel() | 15:56 |
| Wizzup | ok, let's not add a new command to h-i-m just yet and see what we can do here | 15:57 |
| Wizzup | does h-i-m explicitly steal focus, or does it just get focus? | 15:57 |
| dsc_ | most likely it just gets focus | 15:58 |
| dsc_ | from it being a new x11 thing | 15:58 |
| dsc_ | and tbh maybe thats fine | 15:58 |
| dsc_ | but I'm not sure how to prevent the loop from happening | 15:59 |
| kiva | interesting old phone: https://www.gsmdome.com/fujitsu-windows-7symbian-phone-loox-f-07c-fully-detailed-here | 16:02 |
| kiva | although link says "fully detailed" the page does not have even fully specification..it has full phone it...but boots first with Intel Atom..so it might have possibility replace boot code. | 16:04 |
| kiva | sadly quite expensive in second hand markets, so I dont have money for test boot to linux. | 16:07 |
| kiva | anyway it should be easy examine can it make to Maemo Leste device, because it can boot to Win7. | 16:31 |
| Wizzup | freemangordon: I've got a few more pkgs to go but then I think the main blockers for daedalus will be mesa/sgx and tracker | 16:32 |
| Wizzup | there's also some tinymail complication issue but I will try to solve that, same for osso-calculator-engine | 16:32 |
| sicelo | kiva: if you have the device and are able to get linux booting, then yes, sounds doable | 16:36 |
| kiva | sicelo, too expensive for me, but if somebody have money, there is one with dock: https://www.ebay.com/itm/387756831364 | 16:45 |
| Wizzup | I would image the battery life will be bad | 16:46 |
| Wizzup | imagine | 16:46 |
| sicelo | kiva: maybe we need to forget about that device then (at least for the time being) :-) | 16:49 |
| kiva | battery life bad indeed...in win7 use it is 2 hours...but old intel atom can drop CPU speed very slow on the fly to wait phone calls and wake up when call coming. | 16:52 |
| kiva | sicelo, yes that is true..it is not first thing to do, but if somebody finds it for cheap for tests,it would be good to test. | 16:54 |
| sicelo | Wizzup: freemangordon: i'm confused by the N900 camera's dmesg, which reports "could not get clock." | 16:59 |
| freemangordon | maybe DTS is wrong/missing | 17:00 |
| sicelo | driver looks like https://elixir.bootlin.com/linux/v6.13-rc3/source/drivers/media/i2c/et8ek8/et8ek8_driver.c#L1436 ... so we only ask for some default clock, not the one specified in dts ("extclk") | 17:00 |
| freemangordon | Wizzup: what's with mesa? | 17:00 |
| sicelo | now the curious thing is ... this part of driver has been this way since day 1, and same for DTS | 17:00 |
| freemangordon | sicelo: perhaps back then there was a fallback if named clock does not exist | 17:01 |
| sicelo | no idea how this ever worked. also looks like return value for `devm_clk_get` has also been stable over the years when given NULL as id | 17:01 |
| freemangordon | sorry, when default clock does not exist | 17:01 |
| sicelo | anyway, i will test the driver with explicit "extclk" instead of null and see | 17:02 |
| Wizzup | freemangordon: ah, I thought we wanted newer mesa | 17:09 |
| Wizzup | freemangordon: do you want me to import chimaera mesa and sgx for now | 17:09 |
| Wizzup | s/import/build/ | 17:09 |
| Wizzup | well, our chimaera | 17:09 |
| freemangordon | sec | 17:11 |
| freemangordon | what about this https://gitlab.freedesktop.org/StaticRocket/mesa? | 17:11 |
| freemangordon | what is mesa version in daedalus? | 17:12 |
| Wizzup | I think that's why I was asking for your help :D | 17:12 |
| Wizzup | https://packages.debian.org/source/bookworm/mesa | 17:12 |
| Wizzup | 22.3.6 | 17:12 |
| Wizzup | I don't have a good idea of what mesa patches we have, what amber or not menas for us, etc | 17:13 |
| Wizzup | and I don't know about this repo either | 17:13 |
| freemangordon | this is the official TI repo for sgx mesa | 17:13 |
| freemangordon | basicalli it has the same as what I did back then for 'our' mesa | 17:13 |
| Wizzup | last commit 1 year ago, or do we want other branch | 17:13 |
| freemangordon | 23.2.1? | 17:14 |
| freemangordon | I guess we can uplift 22.3.5 to 22.3.6 | 17:14 |
| Wizzup | would be better if we can stay on 22.3.x, then we don't need special libdrm etc | 17:15 |
| freemangordon | ok, lets do it then | 17:15 |
| freemangordon | I will rebase 22.3.5 to 22.3.6 | 17:15 |
| freemangordon | or, maybe can just have sgx patches in /debian | 17:15 |
| freemangordon | hmm? | 17:16 |
| Wizzup | I prefer not to use debian/patches | 17:16 |
| freemangordon | ok | 17:16 |
| freemangordon | do you have link to debian mesa? | 17:16 |
| Wizzup | do we also want newer sgx-ddk-um? | 17:16 |
| freemangordon | yes, sure, but not immediately | 17:16 |
| Wizzup | https://salsa.debian.org/xorg-team/lib/mesa | 17:16 |
| Wizzup | (this is linked from the https://packages.debian.org/source/bookworm/mesa page) | 17:16 |
| freemangordon | ok | 17:17 |
| Wizzup | https://salsa.debian.org/xorg-team/lib/mesa/-/tree/debian-bookworm?ref_type=heads | 17:17 |
| freemangordon | lets see if that allwinner with its 4 cores can do fatser :) | 17:17 |
| Wizzup | probably not great | 17:18 |
| freemangordon | yeah | 17:18 |
| Wizzup | I can maybe set something up | 17:18 |
| Wizzup | btw, we also need to enable gles v1, etc | 17:18 |
| Wizzup | there's a lot of patches we carry, also just to debian/ | 17:18 |
| Wizzup | https://github.com/maemo-leste-upstream-forks/mesa/commits/maemo/chimaera/ | 17:18 |
| freemangordon | yeah | 17:19 |
| Wizzup | I wonder if the kirkstone branch is for us | 17:20 |
| Wizzup | maybe it's not | 17:20 |
| freemangordon | I think it is | 17:21 |
| Wizzup | ok | 17:21 |
| Wizzup | also note that they dropped __DRIimageExtension to 7 from 8, we set it to 8 from 14 | 17:22 |
| freemangordon | could be | 17:22 |
| Wizzup | (yes) :D | 17:22 |
| freemangordon | will see if there are issues, will fix thema st they arise | 17:22 |
| Wizzup | https://gitlab.freedesktop.org/StaticRocket/mesa/-/commit/c9f0919367589b38f5682183846de9d60eec082d | 17:22 |
| Wizzup | I think if you take this branch and rebase for the minor one in debian we will already have the commit, then we just need to pick our changes | 17:23 |
| freemangordon | mhm, something like that | 17:23 |
| freemangordon | but, I need a machine to build on | 17:23 |
| freemangordon | do we already have VM images? | 17:23 |
| Wizzup | for daedalus? no | 17:24 |
| freemangordon | dist-upgrade? | 17:24 |
| Wizzup | I don't know if this will build on amd64 either | 17:24 |
| freemangordon | why not? | 17:24 |
| Wizzup | maybe I should try dist-upgrade first and let you know, there might be a few repos missing | 17:24 |
| Wizzup | maybe it will, don't know if it needs sgx specific libs/headers | 17:24 |
| freemangordon | no, it does dload/dlsym | 17:24 |
| Wizzup | ok | 17:25 |
| freemangordon | IIRC | 17:25 |
| Wizzup | well, you can certainly snapshot and then try dist upgrade | 17:25 |
| freemangordon | ok, lemme clone my chimaera | 17:25 |
| freemangordon | mhm | 17:25 |
| Wizzup | I didn't do this yet but think we're close | 17:25 |
| Wizzup | for updates/status see https://github.com/maemo-leste/bugtracker/issues/751 | 17:27 |
| freemangordon | mafw-tracker will take some time | 17:28 |
| Wizzup | right, it's not essential atm | 17:28 |
| freemangordon | hmm, I wonder what we can do for hildon python bindings | 17:29 |
| freemangordon | do we need gtk3 hildon? | 17:29 |
| Wizzup | right now we will just have nothing with python/gtk2 | 17:30 |
| Wizzup | but yes, later on we need to deal with this | 17:30 |
| freemangordon | no applets? | 17:30 |
| Wizzup | well there's actually very very little that uses gtk2 python | 17:30 |
| Wizzup | there's the binary clock applet | 17:30 |
| Wizzup | and there's syncevolution-frontend | 17:30 |
| Wizzup | that's it I think | 17:31 |
| Wizzup | so yes, eventually we'd want gtk3 hildon, but we can live without it even until trixie | 17:31 |
| Wizzup | the main upcoming issue will be gconf2 no longer existing in trixie, so we need to go to gsettings | 17:31 |
| Wizzup | but for bookwork (daedalus), we're ok | 17:31 |
| Wizzup | bookworm* | 17:31 |
| freemangordon | ok | 17:31 |
| Wizzup | even for chimaera I had to put in a lot of effort to get some py2 stuff going, and I'm happy not to even try again | 17:31 |
| freemangordon | :) | 17:32 |
| freemangordon | lets see if I will be able to cross-compile mesa | 17:32 |
| Wizzup | I can set up a vm for you on our honeycomb build machine | 17:32 |
| freemangordon | hmm, wait, I am on ubuntu 20 | 17:32 |
| freemangordon | or 22 | 17:32 |
| freemangordon | yeah 20 | 17:33 |
| freemangordon | I should not need VM, no? | 17:33 |
| Wizzup | merlijn@honeycomb:~$ ls devuan-vm-aarch64-plain/ | 17:33 |
| Wizzup | devuan.qcow2 efi mini.iso | 17:33 |
| Wizzup | I can set that up if you want | 17:33 |
| freemangordon | 21.2.6-0ubuntu0.1~20.04.2 | 17:33 |
| freemangordon | that's what my native OS is on | 17:34 |
| freemangordon | hmm old | 17:34 |
| Wizzup | I am setting you up with a vm | 17:34 |
| freemangordon | lemme first see if I can do with local VM | 17:34 |
| freemangordon | unless you have very fast internet | 17:34 |
| Wizzup | 1gbit/s | 17:34 |
| freemangordon | hmm... | 17:34 |
| freemangordon | still, lemme try local first | 17:35 |
| freemangordon | as I will need it anyways | 17:35 |
| freemangordon | I am using nfs mount from PC->VM for access to source files | 17:36 |
| freemangordon | and my inet is nto that fast as yours (perhaps 75MBs) | 17:37 |
| Wizzup | ok, in any case I set up the dev vm | 17:39 |
| freemangordon | sure | 17:40 |
| Wizzup | as in, I did it already | 17:40 |
| Wizzup | I'll dist upgrade it | 17:40 |
| freemangordon | Wizzup: hmm, why dist-upgrade wants to remove qt applications? | 17:48 |
| freemangordon | https://pastebin.com/HSmQy06P | 17:50 |
| Wizzup | let me check | 17:52 |
| Wizzup | freemangordon: don't know, I think you can just let it remove them for now and it will get pulled in again later | 17:52 |
| Wizzup | I will try a dist upgrade once I've done a few more repos | 17:53 |
| freemangordon | yep, that's what I did | 17:54 |
| Wizzup | I am dist-upgrading the dev vm to daedalus now btw | 18:00 |
| Wizzup | (armhf vm) | 18:00 |
| freemangordon | me too (amd64) | 18:22 |
| freemangordon | Wizzup: I ran out of time today, will continue tomorrow | 18:23 |
| Wizzup | ok | 18:24 |
| Wizzup | let me know if you need a armhf vm | 18:24 |
| freemangordon | sure | 18:24 |
| arno11 | Wizzup: btw for daedalus, indeed pa-modules-nemo/cmtspeech-n9xx are not needed | 19:16 |
| Wizzup | yeah | 19:17 |
| Wizzup | :) | 19:17 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!