| Wizzup | https://www.phoronix.com/scan.php?page=news_item&px=X.Org-Server-21.1-21.0.99.1 maybe now is a time to push on some of the bugs we filed | 00:50 |
|---|---|---|
| stan | with non-stock battery, one unit reports current, the other gives me c=NA | 04:09 |
| stan | still gives me power consumed though | 04:10 |
| Wizzup | uvos: ping, I did the thing with the UCM as you suggested | 08:34 |
| Wizzup | but it just returns less being visible in pa | 08:35 |
| Wizzup | I think the playbackpcm stuff doesn't match or so | 08:35 |
| Wizzup | fixed it | 08:54 |
| Wizzup | (the names had to match specifically HDA-Intel, both dirname and the main config | 08:54 |
| Wizzup | I had to strace to figure it out) | 08:55 |
| Wizzup | unrelated, but lol: https://github.com/mumble-voip/mumble/issues/1069 | 09:02 |
| Wizzup | looks like current PA already has code to cork and paste roles from .desktop files | 09:03 |
| Wizzup | freemangordon: I guess that properties like these are also assigned/changed dynamically: | 09:24 |
| Wizzup | x-maemo.alsa_sink.mixer_tuning = "HP DAC=0:-6000,38:-4000,42:-3800,46:-3600,50:-3400,54:-3200,58:-3000,62:-2800,66:-2600,70:-2400,74:-2200,78:-2000,86:-1600,90:-1400,94:-1200,98:-1000,102:-800,106:-600,110:-400,114:-200,118:0" | 09:24 |
| stan | experimenting with PA daemon.conf -- using the 'trivial' resampler didn't reduce cpu use. | 09:24 |
| uvos | Wizzup: oh jeah the profile needs to be for the card you have ofc. this is checked by the alsa name | 09:34 |
| uvos | looks like cork + stream restore + ucm gets us pretty far | 09:35 |
| Wizzup | not that far I thrink | 09:36 |
| Wizzup | I need a break, was woken up too early | 09:36 |
| uvos | ok | 09:36 |
| Wizzup | but yeah, I have a semi working ucm | 09:36 |
| Wizzup | (it only lists "Make a phone call" and "Off", not the hifi equiv) | 09:36 |
| uvos | hmm | 09:36 |
| Wizzup | do you have a leste vm? and do you use qemu? | 09:37 |
| uvos | yeah | 09:37 |
| Wizzup | ij | 09:37 |
| Wizzup | ok* | 09:38 |
| Wizzup | so I can share some configs eventually | 09:38 |
| uvos | there probubly is no CapturePCM "hw:${CardId},1" | 09:38 |
| uvos | arecord -l? | 09:38 |
| Wizzup | right yeah, I didn't add a capture device | 09:40 |
| Wizzup | oh, that's why HiFI doesn't work? likelyu | 09:41 |
| uvos | yeah the profile just fails if the card isent available | 09:41 |
| uvos | you need to remove all the capture stuff | 09:41 |
| Wizzup | or figure out how to add capture to qemu | 09:42 |
| uvos | sure | 09:42 |
| Wizzup | probably too much work | 09:43 |
| Wizzup | brb | 09:43 |
| freemangordon | Wizzup: not really "changed", rather "provided to volume control based on the current sink" | 10:14 |
| freemangordon | IIUC | 10:14 |
| Wizzup | freemangordon: hm ok | 12:25 |
| Wizzup | uvos: just to confirm, yes, removing the capture stuff from the UCM worked | 12:57 |
| uvos | great | 13:02 |
| Wizzup | I'm going to try to get a sink list similar to fremantle | 13:03 |
| Wizzup | e.g. load the null sink, but also this kind of stuff: | 13:03 |
| Wizzup | .ifexists module-policy-enforcement.so | 13:03 |
| Wizzup | load-module module-policy-enforcement null_sink_name=sink.null | 13:03 |
| Wizzup | .endif | 13:03 |
| Wizzup | my fremantle has 7 sinks, at least last I ran pactl info | 13:04 |
| Wizzup | I wonder if we can get jusa (from sfos/nemo/mer) to discuss it some with us | 13:16 |
| Wizzup | btw https://github.com/nemomobile/pulseaudio-module-cmtspeech-n9xx/commit/0a1db966f67550c3334dd6126fee57ef90ec9425 | 13:26 |
| Wizzup | :-) | 13:26 |
| Wizzup | https://github.com/nemomobile/pulseaudio-settings-n900/blob/master/etc/pulse/arm_nokia_rx_51_board.pa | 13:27 |
| Wizzup | uvos: do you think we should turn this into ucm? https://github.com/nemomobile/pulseaudio-settings-n900/tree/master/usr/share/pulseaudio/alsa-mixer | 13:29 |
| Wizzup | maybe I should just take this as-is for now | 13:30 |
| Wizzup | get it to work, and then look at ucm and how to change the policy enforcement for it | 13:30 |
| Wizzup | freemangordon: looks like volumes are here https://github.com/nemomobile/pulseaudio-settings-n900/tree/master/var/lib/pulse-nokia/algs/mainvolume | 13:30 |
| freemangordon | yes | 13:31 |
| Wizzup | I fear some of this may need work, since it's PA 4 and nemo is now at PA 12 and they didn't update it | 13:31 |
| Wizzup | we'll see.. | 13:31 |
| Wizzup | freemangordon: parazyd: any clue what provides pulsecore? | 13:50 |
| Wizzup | It looks like debian doesn't have a package for it | 13:50 |
| Wizzup | libpulsecore is part of pulseaudio pkg | 13:50 |
| Wizzup | but there is no pulsecore.pc or something | 13:50 |
| Wizzup | probably debian didn't package it or something | 13:53 |
| Wizzup | ... | 13:53 |
| Wizzup | yes every pulse module they list in debian is part of the pulseaudio source package :( | 13:54 |
| parazyd | Wizzup: There's libpulse-dev which should provide libpulse.pc | 14:05 |
| parazyd | Then all the libs will be available | 14:05 |
| parazyd | Just append -lpulsecore | 14:05 |
| Wizzup | that is not pulsecore though | 14:23 |
| Wizzup | what about the include files? | 14:23 |
| Wizzup | there's a lot of pulsecore include files | 14:23 |
| Wizzup | maybe it's a fedora (or w/e sfos uses) vs debian thing | 14:23 |
| Wizzup | ok, looks like the nemo stuff wants pulseaudio 14 | 15:33 |
| Wizzup | bullseye has 14 - wonder if we can get it | 15:33 |
| Wizzup | I could also try to use older nemo code.. | 15:34 |
| Wizzup | actually, hang on, I think I mostly got an older tag to compile | 15:35 |
| Wizzup | yeah, tag 12.2.30 (latest 12.2) tag compiled finally | 15:46 |
| Wizzup | sudo dpkg -i ../pulseaudio-modules-nemo_12.2.30_amd64.deb ../libmeego-common-dev_12.2.30_amd64.deb | 17:10 |
| Wizzup | :) | 17:10 |
| bencoh | libmeego? uh | 17:10 |
| Wizzup | yeah, it's what the pa common stuff is called | 17:11 |
| lel | MerlijnWajer created a repository: https://github.com/maemo-leste/pulseaudio-modules-nemo | 17:18 |
| bencoh | hmm ... do you guys use ${shlibs:Depends} when packaging? the recursive deps thing seems pretty silly to me :( | 18:29 |
| parazyd | Yeah you should always use that | 18:30 |
| parazyd | Also ${misc:Depends} | 18:30 |
| bencoh | seriously? :( | 18:30 |
| parazyd | What's the problem? | 18:30 |
| bencoh | I am not sure whether it really links against all those libs at compile time, or if it just resolves recursively | 18:31 |
| bencoh | if the former, then everything is fine | 18:31 |
| parazyd | It generates package dependencies | 18:31 |
| parazyd | Meaning, the deps will be equivalent to what your binary is linked to | 18:32 |
| bencoh | if the latter, then it means that changing a direct dependency would force us to rebuild the end-of-the-chain (consumer) package just to update its dependencies | 18:32 |
| Wizzup | I think debian should do the right thing here | 18:32 |
| bencoh | Wizzup: which is? | 18:32 |
| Wizzup | I don't think it recurses everything | 18:33 |
| Wizzup | you can also see what it resolves too by opening the .deb file I think | 18:33 |
| bencoh | alright | 18:33 |
| parazyd | Yeah they'll be listed in the .deb, in the control file | 18:33 |
| bencoh | according to readelf, the .so really includes all those deps | 18:34 |
| bencoh | so ... alright then | 18:34 |
| uvos | oh btw can you suppess the unresolved symbol warings that this creates? | 18:34 |
| uvos | mce generates a huge amount of those | 18:34 |
| uvos | because of all the pluginns | 18:34 |
| bencoh | I do get this at build time: http://pastebin.notk.org/pastebin.php?show=m561eaa44 | 18:34 |
| bencoh | apparently I'm not the only one thinking that something is silly with that link | 18:35 |
| uvos | Wizzup: freemangordon: would have some time to talk about the mce pr rn | 18:36 |
| uvos | if you like | 18:36 |
| parazyd | bencoh: -Wl,--as-needed in ldflags | 18:37 |
| bencoh | parazyd: I have the feeling --as-needed is one of the last things I'd like to enforce by myself when building packages | 18:37 |
| bencoh | considering how it changes linker behavior | 18:38 |
| bencoh | (and dependency order requirements) | 18:38 |
| Wizzup | uvos: I don't have time right now, meetings, can we pick another day this week maybe, or later tonight for me | 18:39 |
| uvos | Wizzup: ok | 18:39 |
| uvos | so it gernerates a huge amount of dpkg-shlibdeps: warning: debian/mce/usr/lib/mce/modules/libinactivity-inhibit.so contains an unresolvable reference to symbol dbus_send_message: it's probably a plugin | 18:40 |
| uvos | would be nice if you could specify that "yes this i a plugin" for eatch module somehow | 18:41 |
| Wizzup | https://man7.org/linux/man-pages/man1/dpkg-shlibdeps.1.html | 18:47 |
| Wizzup | /usr/local/etc/dpkg/shlibs.override | 18:47 |
| Wizzup | Per-system overriding shared library dependency | 18:47 |
| Wizzup | information. | 18:47 |
| Wizzup | ? | 18:47 |
| Wizzup | not sure what you can override | 18:47 |
| Wizzup | uvos: the man page says | 18:52 |
| Wizzup | If the binary is really a plugin, then disregard this warning. But there's always the possibility that it's a real library and that programs linking to it are using an RPATH so that the dynamic loader finds it. In that case, the library is broken and needs to be fixed. | 18:52 |
| Wizzup | you could set --warnings to anything but binary | 18:53 |
| Wizzup | although, maybe that's not the name for the warning,weird | 18:53 |
| Wizzup | hm. looks like it cannot silence those | 18:53 |
| parazyd | I wouldn't worry about it | 18:55 |
| uvos | its not a problem | 18:55 |
| uvos | ist just ulgy that the build throws literaly 100s of warnings | 18:56 |
| Wizzup | yup | 18:56 |
| lel | MerlijnWajer edited a repository: https://github.com/maemo-leste/pulseaudio-modules-nemo | 19:00 |
| Wizzup | not sure how I edited it, but ok | 19:00 |
| parazyd | branch changed from master to leste :p | 19:01 |
| bencoh | autotools' distcheck doesn't play nice with plugindir :( | 19:02 |
| Wizzup | parazyd: oh, I see | 19:03 |
| lel | MerlijnWajer created a repository: https://github.com/maemo-leste/pulseaudio-module-cmtspeech-n9xx | 20:54 |
| lel | MerlijnWajer created a repository: https://github.com/maemo-leste/pulseaudio-policy-enforcement | 20:54 |
| tmlind | uvos: sounds like there are both secure (hs) and general purpose (ns) versions of the hardware available for some tablets | 21:01 |
| uvos | tmlind: interesting | 21:28 |
| uvos | those files are in mz615 files | 21:28 |
| uvos | aka euro xyboard 10 inch | 21:28 |
| uvos | so there are effectivly unlocked devices out there | 21:29 |
| uvos | hmm | 21:29 |
| tmlind | yeah so it seems | 21:35 |
| tmlind | uvos: so you suspect the allow-mbmloader-flashing-mbm.bin is actually packed into motoboot.bin? | 21:35 |
| uvos | well can you flash mbmloader.bin on your device with stock mbm? | 21:36 |
| uvos | if no it must be somewhere | 21:36 |
| uvos | because the updates contain updates to mbmloader | 21:36 |
| uvos | its in the sbf files for xoom 2 for sure | 21:36 |
| uvos | there are forum potst about this | 21:37 |
| uvos | some guy extraced the sbf and posted it | 21:37 |
| uvos | but the upload long gone | 21:37 |
| tmlind | uvos: hmm so what is sbf? | 21:37 |
| uvos | and i cant find a copy of that extract or the sbf | 21:37 |
| uvos | tmlind: sbf is a packed image format moto used with the windows flashing tool | 21:38 |
| tmlind | uvos: ok, never used that | 21:38 |
| uvos | it was for internal use only afaik, sbf files got leaked for some devices only | 21:38 |
| tmlind | ok | 21:38 |
| tmlind | so doing strings on motoboot.bin shows mbm and mbmloader | 21:39 |
| uvos | if you load it in a hex editor | 21:39 |
| uvos | you will find lots of partion names | 21:39 |
| uvos | so yeah | 21:39 |
| uvos | idk | 21:39 |
| uvos | what it is | 21:39 |
| tmlind | i guess i'll try to extract it at some point and see what happens | 21:39 |
| uvos | well can you flash mbmloader.bin on your device with stock mbm? <-- so what about this | 21:40 |
| uvos | just reflashing the same mbmloader.bin should be safe to try | 21:40 |
| tmlind | uvos: yeah ok i'll try next time i play with it | 21:41 |
| uvos | ok | 21:41 |
| uvos | if that works the probubly just changed from having a permissive mbm for updates and a normal mbm to just one slightly more permisive mbm | 21:41 |
| uvos | for updates | 21:41 |
| uvos | with the sbf containg the other build | 21:42 |
| uvos | that would be bad for us | 21:42 |
| tmlind | yeah that would suck | 21:43 |
| uvos | Wizzup: btw the tegra 2 soc used on the xoom dose have pcie | 21:47 |
| uvos | Wizzup: so maybe the wirgly 4g lcm 2.0 lte card IS pcie | 21:48 |
| uvos | tmlind: happen to have one of those? ^^^ | 21:48 |
| tmlind | uvos: no, don't have one, and it's not pcie, just pcie connector | 21:50 |
| uvos | ok | 21:50 |
| Wizzup | yeah, I think it does usb | 21:52 |
| uvos | yeah that was the most likely thing. tmlind have any info on it? | 21:52 |
| tmlind | yeah somebody tried plugging one to a laptop, it prevents it from booting, did not fry anything | 21:55 |
| uvos | heh thats brave | 21:55 |
| Wizzup | hm, do you know how it prevented it (assuming it's not some laptop blacklist) | 21:55 |
| Wizzup | the modem I have/had in my x230 was also mini-pcie but just did usb afaict | 21:55 |
| uvos | Wizzup: squinting at this | 21:56 |
| tmlind | no idea, but it's not a pcie card so anything can happen | 21:56 |
| uvos | https://www.ifixit.com/Teardown/Motorola+Xoom+Teardown/4989 | 21:56 |
| uvos | reviels its not pcie | 21:56 |
| uvos | as the blank has no chips | 21:56 |
| uvos | but still is connected to the antenna somehow | 21:56 |
| uvos | so the wifi chip must be driving the ant though the connector or something | 21:56 |
| uvos | hmm | 21:56 |
| tmlind | heh that's a dummy card, the real one was not ready in time or something | 21:57 |
| uvos | yeah i know | 21:57 |
| uvos | or maybe they just dident want the lte antenna leeds to be loose | 21:57 |
| Wizzup | btw, pulseaudio-modules-nemo, pulseaudio-policy-enforcement, both built fine | 21:57 |
| Wizzup | pulseaudio-module-cmtspeech-n9xx required porting to newer PA api but it was trivial and also builds | 21:57 |
| * tmlind unplugs from computer | 21:58 | |
| Wizzup | ttyl | 21:58 |
| sicelo | nice to hear (about porting pa) -- so with some luck, it should be possible to get Nemo-grade voice calls on N900 too | 21:58 |
| Wizzup | or perhaps better :) | 22:00 |
| parazyd | ^_^ | 22:01 |
| lel | MerlijnWajer edited a repository: https://github.com/maemo-leste/pulseaudio-policy-enforcement | 22:16 |
| Wizzup | ok, now we need pulseaudio forked to get a pulsecore pkg | 22:18 |
| parazyd | Wizzup: Could it work by just providing headers and a .pc file in that repo? | 22:20 |
| uvos | thats slightly ugly | 22:21 |
| uvos | we could have a seperate package with just those | 22:22 |
| uvos | that would be more stomacheable | 22:22 |
| Wizzup | parazyd: yes, but it's a hack | 22:23 |
| Wizzup | of course, us forking pulse is also a hack | 22:23 |
| uvos | Wizzup: did you petition debian/devuan to stop being silly with with the headers? | 22:23 |
| Wizzup | no, not yet | 22:23 |
| Wizzup | it's only been like 6 hours :p | 22:23 |
| uvos | im just asking :P | 22:24 |
| Wizzup | it looks like debian, maybe even upstream, doesn't ship pulsecore.pc | 22:24 |
| Wizzup | so maybe it comes from sfos people | 22:24 |
| uvos | oh ok | 22:24 |
| uvos | then its different story | 22:24 |
| Wizzup | but there is also no way to get the headers + libs on debian | 22:24 |
| uvos | let me search arch | 22:24 |
| uvos | what do you need exacly? | 22:25 |
| Wizzup | let me pm | 22:25 |
| uvos | pulsecore.pc and what header | 22:25 |
| Wizzup | https://paste.villavu.com/raw/3rN0qBUYtmovt1Mk2zMbG5R7eGzulX6KkhNCQ8c5/ | 22:25 |
| Wizzup | + made my own pulsecore.pc | 22:26 |
| Wizzup | (so the cp of libpulse.pc is not enough, of course, I modified it) | 22:26 |
| uvos | ok | 22:27 |
| uvos | can you list one header in /usr/include/pulsecore/ (just any) id like to check where these come from | 22:27 |
| uvos | see if its in the repo | 22:28 |
| Wizzup | /usr/include/pulsecore/mutex.h | 22:28 |
| Wizzup | https://gitlab.freedesktop.org/pulseaudio/pulseaudio master doesn't have pulsecore.pc either | 22:29 |
| Wizzup | maybe I should ask jusa | 22:29 |
| uvos | ok arch dosent have that file either | 22:29 |
| uvos | so yeah | 22:29 |
| uvos | its froms somwhere else | 22:29 |
| uvos | *from | 22:29 |
| Wizzup | what file specifically? | 22:30 |
| uvos | any thing that contains "pulsecore/mutex.h" | 22:30 |
| Wizzup | here are the headers https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/tree/master/src/pulsecore | 22:30 |
| uvos | any package that is | 22:30 |
| Wizzup | meson lists them as libpulsecommon_headers | 22:31 |
| Wizzup | https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/blob/master/src/meson.build#L71 | 22:31 |
| Wizzup | sid doesn't package them still: https://packages.debian.org/sid/amd64/libpulse-dev/filelist | 22:32 |
| uvos | since this is something that is across distros | 22:32 |
| uvos | i suspect the pulse build system just dosent install these files | 22:32 |
| Wizzup | https://github.com/sailfishos/pulseaudio/blob/master/rpm/pulseaudio.spec#L270 | 22:33 |
| bencoh | Wizzup: I guess I should push wifi-switcher to a github repo, or is a private (non-github) repo fine? | 22:33 |
| uvos | Wizzup: hmm ok | 22:34 |
| uvos | strange then | 22:34 |
| Wizzup | https://github.com/sailfishos/pulseaudio/commit/0f0e393e8f4aecf81b5bb0473a97cc2f1f28a761 | 22:34 |
| Wizzup | see msg | 22:34 |
| Wizzup | bencoh: you can request one at maemo-leste-extras maybe | 22:34 |
| uvos | ah ok | 22:35 |
| bencoh | Wizzup: hmm, alright | 22:35 |
| uvos | Wizzup: no point petitioning debian then | 22:35 |
| Wizzup | bencoh: https://github.com/maemo-leste-extras/bugtracker | 22:35 |
| bencoh | oh and ... looks like dpkg-buildpackage includes .git in the source package ... | 22:35 |
| Wizzup | uvos: yeah, lame though | 22:35 |
| bencoh | I'd have thought it wouldn't, in 2021 :/ | 22:35 |
| Wizzup | bencoh: that's weird, it doesn't for me | 22:35 |
| Wizzup | I think? | 22:36 |
| bencoh | do you use dpkg-buildpackage or git-buildpackage? | 22:36 |
| bencoh | oh and, maybe I missed something that should go in debian/whatever | 22:36 |
| Wizzup | I use dpkg-buildpackage | 22:37 |
| bencoh | uh ... | 22:43 |
| Wizzup | uvos: parazyd: I suppose I could make a package that uses the pulse source, and just installs the headers and the pc file (from sfos), if we don't want to re-build/pkg pulse | 22:46 |
| Wizzup | annoying how we have to keep track of more and more :-D | 22:46 |
| bencoh | :] | 22:48 |
| parazyd | If you wanna go a bit weird | 22:48 |
| parazyd | You can apt-get source from the build | 22:48 |
| parazyd | So you pull in the actual corresponding source code from Devuan | 22:48 |
| parazyd | just lmk if you decide to do this, as I need to whitelist it for network in Jenkins | 22:49 |
| Wizzup | probably better not | 22:49 |
| bencoh | Wizzup: do you use -i and/or -I ? | 22:49 |
| Wizzup | bencoh: dpkg-buildpackage -b -uc | 22:49 |
| Wizzup | that's what I use | 22:49 |
| bencoh | hmm | 22:49 |
| Wizzup | so I guess I don't usually generate src pkgs | 22:49 |
| bencoh | ah, right | 22:49 |
| Wizzup | but our CI should not create a source pkg with git in it | 22:49 |
| bencoh | alright | 22:49 |
| bencoh | then everything's fine | 22:50 |
| bencoh | (maybe) | 22:50 |
| bencoh | anyway, time to push that to some server and call it a day (somehow) | 22:50 |
| bencoh | or not ... I just realized we have a workflow guidelines ... time to read that | 22:52 |
| Wizzup | it should be real simple :) | 22:52 |
| Wizzup | (and if not, let us know) | 22:52 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!