libera/#maemo-leste/ Tuesday, 2022-12-27

WizzupOSError: /usr/lib/arm-linux-gnueabihf/rtcom-eventlogger/plugins/libchat.so: undefined symbol: rtcom_el_plugin_chat_get_group_title02:20
Wizzuphrm02:20
Wizzupwonder what I missed :)02:20
Wizzupwill check tomorrow02:20
* dsc_ stabs cmake03:32
buZzwoa10:09
buZzhttps://patchwork.kernel.org/project/linux-omap/cover/20181221011752.25627-1-sre@kernel.org/10:09
buZzthats about a functional FM radio in droid4 :O10:09
buZz>10:11
buZzThere is also a TX firmware. The Droid 4's chip should support this,10:11
buZzbut I don't know if there is audio routing for this feature.10:11
buZzwtf10:11
sicelothe radio does work in android. i've used it, so it should work in linux. i did try the stuff on that page. iirc it didn't work because serdev10:22
sicelos/radio/receiver/10:23
buZzi'm kinda suprised its a v4l2 device10:24
sicelowhat should it be? :-)10:25
sicelov4l2 is used for most media stuff that has controls (besides mixers, of course)10:26
siceloeven the TX is a v4l2 device10:26
buZzthe FM TX?10:27
siceloyes10:27
buZzoh, FM TX works in android too?10:27
sicelono idea10:28
siceloi meant, fm tx is a v4l2 function. not necessarily in connection with working on d410:29
sicelotbh, i don't know what chip has fm tx on d4, although i've heard it mentioned.10:29
buZzthat thread seems to suggest the same chip10:31
siceloah10:32
sicelothat would need an antenna though10:36
buZzyes, RX aswell, according to that thread10:38
buZzthe antenna = headphone cable10:38
siceloi mean TX ..10:39
siceloRX is usually always the headphone. but i'm not sure the same one will work well for TX10:40
buZzwell, wavelength is the same :P10:40
Wizzupdsc_: cmake, automake, pick your poison :p10:41
buZzbuild.sh10:41
buZz:P10:41
siceloloooks like wifi is broken (at least for N900) on 6.2-rc1 :'(10:42
sicelohttps://paste.debian.net/1265418/10:42
siceloalthough maybe it's not N900 per se10:42
sicelobuZz: at least N900 has a separate, shorter antenna for tx10:44
buZzyeah, thats really nice10:44
Wizzupsicelo: any progress on the modem oops so far?10:49
sicelono. no serial :-)10:50
WizzupI'm not sure how the serial matters, it resets with a known trace right?10:50
Wizzupor do you mean it's harder for you to bisect?10:51
siceloyes, not easy to bisect this if you can't see how/where it fails10:51
sicelobut i'll try10:51
sicelothe problem is that going back, you end up with non-working screen, and for my skill levels, things get confusing quickly at that point. i think serial will help a lot10:52
buZzsicelo: do you need me to make you such a cable?10:53
buZzi still need to make one for myself10:53
sicelowill make it soon ... the stuff that's been keeping me too busy is almost over now. last bits should be out of the way by weekend hopefully10:53
buZzbut would be more motivated if second one is needed :P10:53
sicelobuZz: i'll make it, but very rudimentary10:53
buZzhehe is there any other way?10:53
buZzfyi; the 18650 holder droid4 emptied itself over night, while connected to charger :) will replace cables later today probably10:54
siceloi mean - just putting things together so it works, at least for RX side (this doesn't need level conversion too)10:54
Wizzupsicelo: I think going back to the version that worked, with the revert, adding some prints to see if it uses dma or not, and which paths it takes, is helpful10:55
buZzi think i'd just make it '6 pins' so i can use those cheapass USB-UART things i have around10:55
Wizzupbut I do not have a good understanding of how it works10:55
buZzwhich do 3.3V already (or is it 1.8V?)10:55
Wizzupit's neither iirc10:56
Wizzuphttp://n900.elektranox.org/serial-adapter.html10:56
buZz0.9? O_o10:56
buZz2.7V .. i think that should work with a 3.3V uart without issues?10:56
Wizzupnope10:57
Wizzupyou'll burn the ports10:57
Wizzupread the page ;)10:57
Wizzup>For kernel debugging purposes one wants to have access to a serial port on embedded hardware. The Nokia N900 has such a port available, but it’s under the battery and is not 3.3 Volt tolerant (so no cheap USB-TTL Adapters can be used).10:57
buZzoh, n90010:57
Wizzupyes, d4 cable is trivial10:57
buZzi thought it was about d4 :)10:57
sicelo okay looks like the N900 wifi breakage might just be as simple as making some adjustments in dts10:59
sicelothey did some work on wl125111:00
sicelofor the serial/modem, i think it would have been easier if i knew how to make mtdoops or pstore work11:03
Guest47Hello everyone, do you need to upload a new image to the SD card or is there another way to update?11:05
Wizzuptypically, normal updates are installed with 'apt update' and 'apt upgrade'11:05
Wizzupor 'apt dist-upgrade' even11:06
buZzkinda depends how old the install is :)11:10
buZzbut -usually- just apt update; apt upgrade; works fine11:11
Guest47ok thanks11:11
Guest47Is there a way to install multi-boot on the latest version from the community? (10/10/2015 - 2011/21/38-1Tmaemo11+thumb0)11:11
Guest47Nokia N90011:12
sicelomulti-boot?11:13
Guest47Yes, downloading maemo-leste via a laptop every time is not convenient.11:15
siceloi'm not sure i follow :-)11:19
WizzupGuest47: install leste on the sd card, then you can boot from fremantle/cssu on emmc and leste from sd11:20
siceloWizzup: i'm bisecting now for the wireless regression. i know it was working in 6.1-rc7, and it's broken in 6.2-rc1. should i do a `make mrproper` between builds?11:34
WizzupI honestly don't know, sorry11:35
Wizzupif you have a powerful machine, then probably a good idea11:35
siceloactually,let me add wireless tree, so i don't waste time moving around unrelated points in the history11:41
Wizzupto the bisect you mean?11:41
siceloi mean add a remote for one of the wireless-* repos, then bisect from that, i hope11:44
sicelohttps://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless-next.git/log/ , https://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless.git/log/11:44
WizzupI mean, in general it won't matter too much, but you can try11:44
Wizzupit's binary so you wouldn't be taking that many more steps with a bigger history11:44
Wizzup(it's binary search*)11:44
sicelook. when i bisected the musb regression, it ended up with a drm commit being the first bad commit :-P11:45
Wizzup'bad commit' here means that the first commit where the problem also occurs, right, it doesn't mean this commit is to blame11:46
Wizzupiiuc11:46
siceloyes.11:48
Wizzupfreemangordon: so the sip plugin works for several free accounts I have tried + the one I have at home11:50
Wizzupjust not for linphone11:51
sicelo< buZz> i thought it was about d4 :) ... hehe, there are lots of people working on D412:08
buZz:)12:11
Guest47I want to do this but I can't install12:17
Guest471. Install "U-Boot with kernel 2.6.28-omap1"12:17
siceloyou need to be inside the old maemo. is that one still working on your N900?12:17
Guest47Installed bootmenu added:12:18
Guest47cat > /etc/bootmenu.d/30-maemo-leste.item << "EOF" &&12:18
Guest47ITEM_NAME="Maemo Leste"12:18
Guest47ITEM_KERNEL="uImage"12:18
Guest47ITEM_DEVICE="${EXT_CARD}p1"12:18
Guest47ITEM_FSTYPE="ext2"12:18
Guest47EOF12:18
Guest47but at startup the second boot option from a flash card does not appear12:18
Guest47sicelo yes12:19
sicelodid you run `u-boot-update-bootmenu` after that?12:23
Guest47sicelo /bin/sh: u-boot-update-bootmenu: not found12:24
siceloso fix that :-)12:24
siceloforgot the pkg that contains it... probably u-boot-tools or something like that12:24
siceloyes, u-boot-tools, http://maemo.org/packages/package_instance/view/fremantle_extras-testing_free_armel/u-boot-tools/2013.04-2/12:32
sicelohow did you get u-boot-installed without this ... seems u-boot-tools is a dependency for u-boot itself12:34
Guest47The problem is that there is no U-boot for Community SSU/Thumb 21.2011.38-1Tmaemo1214:03
Wizzupthere should be I think14:04
sicelothere isn't14:09
sicelowhat you do is ... install the standard u-boot, then install the thumb bootimg, and make a bootmenu entry for it14:10
sicelobecause you can't boot thumb kernel with the kernel that's appended in u-boot14:11
freemangordonWizzup: great!14:29
freemangordonWizzup: BTW, antisip work in fremantle as well14:32
freemangordon*works14:32
Wizzupcool14:32
freemangordonat least chat and audio calls14:32
Wizzupso now I guess we need a sphone tp module that can deal with tp-rakia, tp-gabble and tp-ring14:32
freemangordonfor video calls I have to find good android client to install on my GFs android14:32
Wizzup(probably mostly the same code for all)14:33
Wizzupah hehe14:33
freemangordonantisip's own application does not support chat :D14:33
Wizzuptwinkle does14:33
freemangordonok, will install that14:33
freemangordonI tried to do audio call from the VM, but seems RTP cannot pass through to it14:34
Wizzupmight need some NAT on your side14:35
freemangordonyes, but I am not sure I want to do that14:35
Wizzupagreed14:35
freemangordonIIRC RTP ports are dynamic14:36
freemangordonso it will not be simple14:36
freemangordonso I'd rather not play with that14:36
Wizzupright, if you use different qemu network, like bridge or macvtap it might be easier14:36
Wizzupbut I agree14:36
freemangordonmhm14:36
freemangordonI am using virtualbox14:37
freemangordonbut yeah14:37
freemangordondoes sphone support conference calls?14:37
freemangordonand video calls?14:37
freemangordonIOW - does it have a concept of those?14:38
WizzupI don't think so, not at this point14:39
freemangordonhmm...14:39
Wizzupuvos: when you read logs ^^14:39
Wizzupfreemangordon: I think we can just start with one to one calls, that'll be a big improvement already14:39
* freemangordon installs qtcreator in chimaera VM :)14:51
freemangordonwow, chimaera is on qt 5.15!14:51
Wizzupyes14:52
WizzupI did the port already14:52
dsc_im using another IDE but qtcreator is not bad, used it for a while years ago15:12
dsc_just try to use CMake, not qmake :P15:17
freemangordonI am fine with qmake, why cmake?15:24
dsc_nvm15:27
freemangordonWizzup: dsc:   osso_abook_contact_chooser_new() works like a charm in qt application :)16:45
freemangordonthis https://pastebin.com/u2d71HQx16:46
freemangordonvoid MainWindow::on_pushButton_clicked() {     tst(effectiveWinId()); }16:46
dsc_does it block the qt eventloop?16:47
freemangordonumm, why should it?16:47
freemangordonlook at the code16:47
dsc_what should I look at?16:48
freemangordongtk_widget_show_all(chooser); is the last line in tst() function16:48
freemangordonand then it returns to qt event loop16:48
freemangordonlemme just check that I can select a contact16:49
Wizzupcool16:52
* dsc_ was under the impression `abook` would move to Qt :)16:53
dsc_instead we are embedding GTK16:54
freemangordonre-implement aboot in qt?16:54
Wizzupthat seems like a sensible first step, given how large it is and we also need it in gtk16:54
freemangordondsc_: this is not the first thing gtk embeded16:55
dsc_mkay16:55
freemangordonsee https://github.com/maemo-leste/qtmaemo5/blob/master/src/maemo5/qmaemo5datetimepickselector.cpp16:55
dsc_Yeah, I'm not superrrr interested in figuring out how to embed GTK in Qt17:03
dsc_i.e: if some example code could be provided that would be cool :P17:04
freemangordonsec17:07
Wizzupideally we would have libabook-qt which provides a qt pkg that we can be used to call parts of abook17:10
freemangordonhttp://95.43.220.235/leste/abook/17:10
Wizzupthat can be used*17:10
freemangordonWizzup: exactly17:10
freemangordona part of https://github.com/maemo-leste/qtmaemo5/17:10
freemangordonjust LMK what you need first17:11
dsc_freemangordon: thanks :)17:11
freemangordondsc_: this is a working example, but hacky17:12
freemangordonideally osso_abook_contact_chooser_new() (and others) should be dlsym() - ed17:12
freemangordonthat way no extern "C" stuff will be needed17:13
freemangordonI will have some hard times wrapping OssoABookContact, but well... :)17:14
dsc_right17:14
dsc_so17:16
dsc_the .pro includes `abc.c` as source17:16
dsc_yet no file references it17:16
freemangordonummm17:16
dsc_and then `extern` magically exposes it or something?17:16
dsc_:P17:17
freemangordonextern is a declaration17:17
dsc_its like a forward declaration17:17
freemangordonno17:17
dsc_oh17:17
dsc_I see17:17
freemangordonin C all functions are exported unless declared static17:17
freemangordonextern "C" void tst(uintptr_t anid); is for the compiler to know how to call it from mainwindow.cpp17:18
freemangordonWizzup: so, what is the first class we need?17:18
dsc_why not use `#include 'abc.c'`17:18
freemangordonbecause... it is source file17:19
freemangordonnot a header17:19
dsc_well ok, `abc.h` :P17:19
freemangordonbecause this was a quick-n-dirty POC :p17:19
freemangordon"(18,12,04) freemangordon: dsc_: this is a working example, but hacky"17:20
dsc_I do not have the ability to differentiate between what is needed (as in: 'the way to do it') and things that are 'shortcuts'17:20
dsc_i.e: I would have ported your code as-is, under the impression this EXTERN trick is needed17:21
dsc_because of <insert reason>17:21
dsc_thats why I'm asking17:21
freemangordondsc_: no, my question was - what is the first *abook functionality* that we need17:21
freemangordonin qt17:22
siceloapparently sdm845 already has good/reasonable runtime PM17:28
freemangordonWizzup: dist-upgraded my d4, not xorg wouldn't start21:24
freemangordon*now21:24
freemangordonis that a common issue?21:24
buZzi did notice that xorg wont start if storage is full21:40
norayrwhere is the d4's or pp's default xkb map? there should be some file that describes the kbd, but i cannot find it.21:50
norayromg, it's in dts: arch/arm/boot/dts/omap4-droid4-xt894.dts21:52
norayri won't touch it then...21:52
sicelo:-)21:54
sicelowhat did you want changed?21:55
norayrhmmm, the pp kbd behaves differently under pmos.21:55
norayrlet's say the dash is right alt and 721:55
norayrand i cannot type dash symbol without onscreen kbd with pinephone at all.21:55
norayri can create a xkbmap for myself but that won't fix the problem.21:56
sicelo i think it will ... keymaps live in /usr/share/X11/xkb or similar21:57
norayryes, but there are default linux keymaps.21:58
norayrnot pp specific.21:58
norayror d4 specific.21:58
siceloand PP kb might have a keymap already in xkb-data, although I won't be surprised if devuan has older one21:58
sicelonorayr: i'm quite sure there's a pp kb in upstream xkb-data at least21:58
sicelohttps://github.com/freedesktop/xkeyboard-config/commit/8b4ca4168059f80330e792ffc432bf037fec90ff22:02
sicelo and of course, if you come up with a nicer keymap for yourself, you can show it here, and open a PR against leste so it can be added for other pp users. or open it against upstream xkb, whichever you prefer :-)22:05
buZzi think norayr is talking about console keymap?22:07
buZzinstead of xorg22:07
siceloxkb map, perse said ... anyway let's hear which was meant22:08
norayri need an xkb map22:11
norayrlet me see your link22:11
sicelobuZz: at least i can see Alt + 7 => - in that xkb-data commit22:12
norayrdo you think 'setxkbmap -model pinephone' should work? it works now but nothing changes.22:12
norayrah ppkb it is called22:13
norayrwitth current map it didn't work.22:13
norayri'll try to replace the files. the file? let me see.22:13
norayrbtw, uvos, volume up or down doesn't trigger the onscreen kbd on pp.22:14
sicelonorayr: so leste has forked xkb-data, and that fork doesn't have anything for pp22:17
siceloso i suggest that if you want to use it with leste, backport the upstream commit to  https://github.com/maemo-leste-upstream-forks/xkb-data/22:17
siceloshould be straightforward22:17
norayroh so that is in xkb-data package.22:18
norayrthank you.22:18
siceloyw22:21
norayrok i changed those three files like in the commit you pointed out, but only this didn't help. even rebooted to make sure.22:29
norayror maybe setxkbmap -model ppkbd is not what is necessary to type22:29
norayralso setxkbmap -model ppkbd us doesn't help. :/22:29
norayrok i'll think about it if i have time. now it is night and i am tired.22:30
Wizzupfreemangordon: I didn't have that problem, but apart from you and me, others probably should not dist-upgrade yet22:36
Wizzupfreemangordon: elogind installed?22:36

Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!