| Wizzup | arno11: I am a bit confused now, is anything missing or not? did you attach gdb to check the segfault, or? | 00:03 |
|---|---|---|
| arno11 | apparently nothing missing | 00:04 |
| arno11 | currently running gdb before zzz | 00:04 |
| arno11 | segfault seems to come from x11 | 00:07 |
| Wizzup | ok | 00:36 |
| Wizzup | sorry, it's been a busy day at work | 00:36 |
| freemagordon | sicelo: create a new testcase? | 06:42 |
| sicelo | Wizzup: we have a dpkg diversion in wl1251-cal package for N900. i'm working on a PR that will need to undo that diversion. instead of removing/diverting /lib/firmware/ti-connectivity/wl1251-nvs.bin, we will now want to keep it, but the package will change its contents. | 16:40 |
| sicelo | whenever you have time, please help me achieve both. | 16:41 |
| sicelo | rationale: actually that whole udev stuff needs to go. hasn't worked in a long time because udev stopped supported the SUBSYTEM="firmware" mechanism some years ago already | 16:43 |
| sicelo | https://github.com/systemd/systemd/commit/be2ea723b1d023b3d385d3b791ee4607cbfb20ca | 16:43 |
| sicelo | https://github.com/eudev-project/eudev/commit/3b717594600fa717cdf9bcfd0c7c1b703b245482 | 16:44 |
| sicelo | so now the idea is - blacklist wl1251 driver so it doesn't load before we play with wl1251-cal. as early as possible in the boot process, run wl1251-cal, and instead of spitting the nvs via /sys/firmware, just echo it back to /lib/firmware/ti-connectivity/wl1251-nvs.bin. after this is done successfully, probe wl1251 and you're home free | 16:46 |
| sicelo | this does work just fine and i've tested it. just the packaging part now ... the divert to be specific | 16:48 |
| arno11 | Wizzup: @work, np :) | 17:07 |
| arno11 | good news and bad news btw | 17:07 |
| arno11 | tg desktop works (randomly), when it works it works fine | 17:08 |
| arno11 | something seems really wrong with qt5/x11 | 17:08 |
| arno11 | picodrive works fine but only in a small sdl window. no clue what's going on | 17:09 |
| arno11 | drnoksnes still works only from terminal, gui launch returns gtk critical stuff with osso | 17:10 |
| arno11 | and crashes | 17:12 |
| sicelo | Wizzup: https://github.com/sicelo/wl1251-cal/commit/59767d9166a54dc98bae508ab9ef929eed7bfd45 ... i'm still going to build it on N900 and see if the packaging is fine | 18:05 |
| Wizzup | ok | 18:12 |
| Wizzup | how will the module be loaded with it beign blacklisted | 18:13 |
| sicelo | we load it via openrc | 18:16 |
| sicelo | https://github.com/sicelo/wl1251-cal/commit/59767d9166a54dc98bae508ab9ef929eed7bfd45#diff-e6b11caa01583d3f6e39013cd1160e9b3317dc7381888303b2e8f00a0b904d76R13 | 18:17 |
| Wizzup | ok | 18:49 |
| sicelo | arno11: interestingly my leste install never really has those apt/ham slowdowns, even though my card is pretty much low quality. just got offered updates to Core Meta and Meta packages (still on chimaera). | 19:31 |
| sicelo | i was busy building wl1251-cal and wasn't even aware updates were happening in the background | 19:32 |
| sicelo | Wizzup: and yes, my package is reporting ominous dpkg-divert issues when installing. | 19:34 |
| Wizzup | do you have a log of the warn/error? | 20:12 |
| sicelo | made some adjustments now, and the installation succeeds, although this doesn't look too clean, https://paste.debian.net/1349677/ | 20:21 |
| Wizzup | run dpkg again and see if one of the errors goes away? | 20:22 |
| sicelo | this is with https://paste.debian.net/1349678/ in postrm/postint | 20:22 |
| sicelo | without the `--no-rename`, install would fail completely | 20:22 |
| sicelo | no error/warning when running it again | 20:23 |
| Wizzup | I mean if this works, I am not against it, I didn't look in detail though | 20:24 |
| sicelo | the wl1251-nvs.bin.disabled is still there ... which i guess is a bit annoying | 20:27 |
| sicelo | will push my latest stuff | 20:28 |
| Wizzup | I don't think it's weird that those remain tbh | 20:29 |
| Wizzup | that's the point of divert | 20:29 |
| arno11 | sicelo: interesting. maybe your card is good @small blocks | 20:58 |
| arno11 | anyway, apt-worker is mostly a problem on a fresh install | 21:00 |
| arno11 | otherwise, almost no real troubles | 21:01 |
| sicelo | ok @ fresh install. makes sense | 21:08 |
| arno11 | the good news @n900 daedalus is that is usable ootb (with proper swap and maybe custom transitions) | 21:18 |
| arno11 | if someone fix qt5/x11 troubles, we are good | 21:19 |
| arno11 | ram/swap usage is quite low (120/120MB) | 21:24 |
| sicelo | i think the divert stuff is fine now in latest form of https://github.com/sicelo/wl1251-cal/commits/unudev/ | 21:34 |
| sicelo | arno11: if you're feeling adventurous, please give that a spin. should give your N900 the same wlan MAC address as in Fremantle | 21:35 |
| sicelo | Wizzup: so i think the divert stuff is fixed. next problem - there's some 'displace' stuff in leste-config-n900. we need to drop the 00mac file. it's the displace that i'm not sure about atm | 21:36 |
| sicelo | hope i can figure it Out | 21:36 |
| arno11 | sicelo: sorry, not now, currently looking @qt5/X11 | 21:45 |
| sicelo | cool | 21:50 |
| sicelo | what x11 trouble? | 21:50 |
| arno11 | every qt5 apps segfault on closing | 21:50 |
| arno11 | gdb returns segfault @libx11.so | 21:51 |
| arno11 | could be interesting to doublecheck on d4 btw | 21:52 |
| freemagordon | arno11: going to test in my daedalus VM | 21:55 |
| freemagordon | (qt) | 21:55 |
| arno11 | cool thx | 21:56 |
| freemagordon | could you provide backtrace (with symbols installed) | 21:56 |
| arno11 | it depends what exactly you want me to do (no machine around atm, just n900) | 21:58 |
| freemagordon | backtrace of the segfault | 21:58 |
| freemagordon | which application (for example) shall I use? | 21:59 |
| arno11 | ah ok, i tried with hansterfiler | 21:59 |
| arno11 | but bt full doesn't returns lot of details | 22:00 |
| freemagordon | perhaps you need debufg symbols | 22:00 |
| freemagordon | *debug | 22:00 |
| arno11 | i have them | 22:00 |
| freemagordon | may I see that backtrace? | 22:01 |
| arno11 | yep | 22:01 |
| arno11 | #0 0xb5633444 in _XSend () at /usr/lib/arm-linux-gnueabihf/libX11.so.6 | 22:02 |
| arno11 | #1 0xb563360e in () at /usr/lib/arm-linux-gnueabihf/libX11.so.6 | 22:02 |
| arno11 | that's when i close the app | 22:02 |
| freemagordon | you lack debug symbols for libX11.so.6 | 22:03 |
| arno11 | ah ok | 22:03 |
| freemagordon | libx11-6-dbgsym or something | 22:03 |
| arno11 | yep | 22:03 |
| sicelo | freemagordon: Wizzup: https://github.com/maemo-leste/wl1251-cal/pull/1 and https://github.com/maemo-leste/leste-config/pull/56 are up for review. please triple check whatever you can to be sure i didn't especially botch the divert and displace. it all seems to be fine on my end | 22:04 |
| freemagordon | hmm, dummy network seems broken on daedalus | 22:04 |
| arno11 | freemangordon: hmm, can't find them from repo | 22:05 |
| freemagordon | you have to add additionsl repos | 22:06 |
| freemagordon | deb http://deb.debian.org/debian-debug/ bookworm-debug main | 22:06 |
| freemagordon | add ^^^ to sources.list | 22:06 |
| arno11 | ok, but bookworm ones ? | 22:07 |
| freemagordon | sorry, can't parse | 22:07 |
| freemagordon | ah... | 22:07 |
| freemagordon | yes, bookworm ones | 22:07 |
| arno11 | ok let s go | 22:08 |
| freemagordon | daedalus is based on bookworm | 22:08 |
| arno11 | needs few min | 22:09 |
| arno11 | sources.list refuses bookworm line | 22:11 |
| sicelo | so yesterday while running mobian (while working on ofono), i came across something that highlighted the wisdom of using maemo/hildon as a base for an OS. so, i got the asan log, and then needed to copy it so i could pastebin it. lo and behold - there's no way to highlight text in the gnome-console application when using a touchscreen! | 22:11 |
| freemagordon | arno11: what does that mean? | 22:12 |
| arno11 | it says Malformed entry 6 in list file /etc/apt/sources.list | 22:12 |
| arno11 | when i try to update | 22:12 |
| freemagordon | then you did something bad | 22:13 |
| Wizzup | sicelo: I can help fix the displace in leste-config np | 22:13 |
| Wizzup | sicelo: oh you already did it ;) | 22:14 |
| * Wizzup slow today | 22:14 | |
| freemagordon | 'deb http://deb.debian.org/debian-debug/ bookworm-debug main' without the quotes is what I have here | 22:14 |
| sicelo | do verify it :-) | 22:14 |
| Wizzup | freemagordon: your name lacks an 'n' btw | 22:14 |
| Wizzup | now it reads like "free ma' gordon" | 22:14 |
| freemagordon | name? you mean as nick? | 22:14 |
| sicelo | for example, i used a dummy version .1.129, assuming that'll be next version (in the rm_conffile section) | 22:14 |
| freemangordon | oh :D | 22:15 |
| arno11 | freemangordon: ofc but no result atm. let me doublecheck lol | 22:16 |
| freemangordon | Wizzup: does dummy net work for you in daedalus? | 22:17 |
| arno11 | freemangordon: your url was wrong :P should be fine now | 22:18 |
| sicelo | url seems ok to me. what was wrong with it? :-) | 22:19 |
| arno11 | deb http://deb.debian.org/debian-debug bookworm-debug main | 22:20 |
| freemangordon | and how is that different? | 22:20 |
| arno11 | with no '/' after debian-debug | 22:20 |
| freemangordon | ah | 22:20 |
| freemangordon | well, apt here does not have issue | 22:20 |
| arno11 | apt ? | 22:21 |
| freemangordon | yes, apt, what do you use to install debug symbols? | 22:22 |
| arno11 | apt is ok for me too | 22:22 |
| arno11 | with the correct url | 22:22 |
| freemangordon | anyway | 22:22 |
| arno11 | yep anyway, just need few min to provide bactrace | 22:23 |
| arno11 | *backtrace | 22:24 |
| freemangordon | Wizzup: re HAM domains: I made a change (here) so on every update, if there is a package that does not belong to any domain, apt-worker assign the domain of the repo with the highest priority (in HAM terms) that is currently enabled | 22:25 |
| freemangordon | does that sound ok to you> if yes, I'll push it | 22:25 |
| freemangordon | also, shall I split system domains to leste/devuan/debian and if yes, do we want different priorities? | 22:26 |
| arno11 | freemangordon: 0xb5633444 in _XSend (dpy=dpy@entry=0x45f1e0, data=data@entry=0x0, size=size@entry=0) at ../../src/xcb_io.c:548 | 22:28 |
| arno11 | 548../../src/xcb_io.c: No such file or directory. | 22:28 |
| freemangordon | this is not a backtrace :) | 22:28 |
| arno11 | lol let me paste the proper things | 22:29 |
| arno11 | i got the result but i don't have much things install on daedalus | 22:30 |
| arno11 | so let me try to pastebin | 22:30 |
| arno11 | freemangordon: https://paste.debian.net/hidden/433e6ea4/ | 22:40 |
| freemangordon | please install symbols for libxcb.so.1 | 22:44 |
| freemangordon | and all other libs in the backtrace | 22:45 |
| arno11 | okay | 22:45 |
| freemangordon | could do it tomorrow as well | 22:45 |
| freemangordon | anyway I will not look now, zzz time | 22:46 |
| arno11 | anyway, tyvm for your help | 22:47 |
| arno11 | gn | 22:47 |
| Wizzup | freemangordon: can you share the change/diff ? | 22:47 |
| Wizzup | freemangordon: didn't test dummy net I guess | 22:48 |
| arno11 | freemangordon: Wizzup: https://paste.debian.net/hidden/fa306ee7/ | 23:12 |
| sicelo | arno11: btw, someone came along and picked up work on dillo, already with an updated version. maybe will find time to package the new one | 23:22 |
| Wizzup | arno11: ok, so that's in the hildon input module | 23:22 |
| Wizzup | sicelo: that would be great | 23:22 |
| sicelo | for use on N900 | 23:22 |
| arno11 | Wizzup: ok | 23:23 |
| arno11 | sicelo: ah cool @new dillo | 23:24 |
| arno11 | Wizzup: but the question is why | 23:26 |
| sicelo | https://dillo-browser.github.io/25-years/ | 23:29 |
| arno11 | there is a port to atari :) | 23:31 |
| Wizzup | arno11: looks to me like sending something to non-existing x display conn | 23:31 |
| Wizzup | let me work through the other backlog of today first :D | 23:32 |
| arno11 | sure :D | 23:33 |
| Wizzup | sicelo: it looks like dillo is in bookworm and trixie, just as old version | 23:33 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!