Wizzup | uvos: re: duplication stuff, do you mean leste-config, image-builder and the like? | 00:50 |
---|---|---|
Wizzup | image-builder has support for the OUYA, which has fastboot, so probably image-builder can deal with it | 00:51 |
Wizzup | I guess we'll have another kernel pkg, right? | 00:54 |
buZz | Wizzup: woa, leste on Outa? :D | 02:32 |
buZz | ouya* | 02:33 |
buZz | that would be legendary, haha | 02:33 |
uvos | Wizzup: sure copy what ever makes the most sense, image builder wise. | 09:42 |
uvos | for the meta package, leste-config, copy pp | 09:43 |
uvos | we will need a new kernel yes, its not really _that_ manny patches, we could maybe integrate it later | 09:43 |
uvos | we also need to package quite a bit for firmware | 09:44 |
uvos | not sure what to do about that, pmos has a script that scrapes fw from the android partitions on boot | 09:44 |
uvos | they dont pacakge the fw either, for the most part | 09:44 |
Wizzup | freemangordon: if device is locked and opened by program, say alarm, then ts does not work | 10:14 |
Wizzup | I had to lock and unlock manually | 10:14 |
Wizzup | uvos: oh, that is a bit odd | 10:14 |
Wizzup | @ scraping | 10:15 |
uvos | fw has unclear licence | 10:17 |
uvos | i dont think we really need to care about that tho | 10:19 |
uvos | los dosent either | 10:19 |
uvos | and they are the mutch bigger and more visible target | 10:19 |
Wizzup | I agree | 10:35 |
Wizzup | we might want to watch for fw being different for the different models | 10:36 |
freemangordon | Wizzup: maybe pester uvos about it, that's mce realm | 11:59 |
freemangordon | i may have a look at it as well though | 11:59 |
freemangordon | Wizzup: or, perhaps TS is inhibited? | 12:02 |
freemangordon | could you check? | 12:02 |
Wizzup | freemangordon: I think it could be the inhibit thing | 12:05 |
Wizzup | freemangordon: yeah ok will check today | 12:05 |
norayr | uvos, btw i checked svkbd from suckless/sxmo projects, they do setxkbmap by code. | 12:11 |
norayr | i checked the source because i added two layouts there, dvorak and armenian. | 12:12 |
norayr | so my understanding is that they also have to use setxkbmap, and it has to be in the code, for this layout you do setxkbmap this, and for that - setxkbmap that. no detecting, just mapping. | 12:13 |
norayr | everybody: other interesting thing is that i like that tapping on the input field doesn't bring the keyboard up. | 12:15 |
norayr | for bringing up the keyboard one needs to specifically use the gesture in sxmo to bring it up. | 12:17 |
norayr | and to bring it down the same. | 12:17 |
norayr | i happen to like it very much. and it behaves just as regular xorg onscreen keyboard. it types in everything. | 12:17 |
norayr | >apping on the input field doesn't bring the keyboard up. | 12:18 |
norayr | maybe i intent to use hw kbd, or i tapped to then do ctrl-a ctrl-c and copy what is there. | 12:18 |
sicelo | that could just be toggled by presence or state of hwkbd, i guess ... like with the stock OSes of the hwkbd phones | 13:09 |
uvos | usbnet dosent apear to work | 14:35 |
uvos | shouldent something bing up the interface? | 14:35 |
uvos | device side | 14:35 |
uvos | i have to do that by hand | 14:35 |
Wizzup | uvos: do you have configfs and such? | 14:38 |
Wizzup | uvos: ke-recv should do this | 14:38 |
uvos | CONFIG_USB_CONFIGFS=y | 14:39 |
uvos | jup | 14:39 |
uvos | dosent seam to work | 14:40 |
uvos | but it dosent seam to work on my d4 eihter | 14:40 |
Wizzup | huh | 14:41 |
uvos | btw xorg crashes on reboot, poweroff | 14:41 |
uvos | beacause logind is stopped before xorg | 14:41 |
uvos | xorg really really dosent like this | 14:41 |
Wizzup | we can add 'before logind' for stopping to the init script I think | 14:41 |
Wizzup | uvos: I get this: | 14:42 |
Wizzup | [2739419.234385] rndis_host 6-1.4:1.0 usb0: register 'rndis_host' at usb-0000:07:00.4-1.4, RNDIS device, 7a:77:d5:b8:3c:af | 14:42 |
Wizzup | maybe you forgot we now use rndis rather than the usual method | 14:43 |
Wizzup | (just a guess) | 14:43 |
Wizzup | (on d4) | 14:43 |
uvos | thats not the problem | 14:44 |
uvos | the problem is that if i configure the interface on the host side, i cant access the phone side untill i configure the interface there too by hand | 14:44 |
uvos | or am i missing how this is supposed to work | 14:45 |
uvos | where is the phone suppsoed to get an ip address from on the usb0 if | 14:45 |
Wizzup | right, ke-recv-extra sets this | 14:46 |
Wizzup | sec | 14:46 |
uvos | then its not working | 14:46 |
uvos | anyhow wlan wroks on xt1602 now | 14:46 |
uvos | so imidate need is reduced | 14:46 |
Wizzup | sorry, ke-rexc* | 14:46 |
Wizzup | recv* | 14:46 |
Wizzup | pcsuite_enable.sh should bring it up and assign it | 14:47 |
Wizzup | if the interface called usb0? | 14:47 |
Wizzup | logger "$0: starting pcsuite" | 14:47 |
Wizzup | # TODO: check error code here | 14:47 |
Wizzup | /usr/sbin/hildon-usb-gadget-network | 14:47 |
Wizzup | ifconfig usb0 up | 14:47 |
Wizzup | ifconfig usb0 192.168.42.2 | 14:47 |
uvos | ifconfig | 14:47 |
uvos | command not found | 14:47 |
uvos | obvious problem | 14:48 |
uvos | that whent away with chimaera | 14:48 |
uvos | unless you have the compat package installed | 14:48 |
Wizzup | uvos: that's one, and then there is that it might not detect it being hotplugged | 14:50 |
Wizzup | if it is extcon it might just work | 14:50 |
Wizzup | uvos: can you check if it works if ifconfig is installed? | 14:50 |
uvos | well missing ifconfig def is the reason why it cant work on my d4 | 14:51 |
Wizzup | mhm | 14:51 |
Wizzup | I meant on the moto g 4 | 14:51 |
uvos | right | 14:51 |
uvos | qtwebbrowser is really fast on xt1602 | 14:51 |
uvos | but the hildon button in the top left corner makes it impossible to use the deivce left handed :| | 14:52 |
uvos | er one handed i mean | 14:52 |
Wizzup | in portrait I guess? | 14:52 |
uvos | i mean both | 14:53 |
uvos | in landscape its maybe ok if your one hand is the left hand | 14:54 |
uvos | but then there is no getting to X | 14:54 |
Wizzup | ok | 14:56 |
uvos | for some reason mce cant handle the device haveing 4095 pannel brightness levels | 14:59 |
Wizzup | building new ke-recv | 15:00 |
Wizzup | uvos: maybe log scaling or something? | 15:00 |
uvos | maybe, should not be to hard to figure out | 15:00 |
Wizzup | uvos: ah, is this what 'turns off' the screen? | 15:00 |
uvos | no theres more to it than that | 15:01 |
uvos | somethings wrong with freedreno/glamor | 15:01 |
uvos | it works fine in sw mode | 15:02 |
Wizzup | ok | 15:03 |
sicelo | Porting the script to use iproute2 instead of ifconfig is trivial | 15:08 |
uvos | pretty sure Wizzup did so just now | 15:08 |
Wizzup | no, I just had it depend on ifconfig | 15:11 |
Wizzup | net-tools rather | 15:11 |
uvos | hm ok | 15:11 |
Wizzup | since we rely on it in other places too | 15:11 |
Wizzup | I bet | 15:11 |
uvos | but those will go away eventually | 15:11 |
Wizzup | probably :P | 15:11 |
Wizzup | sicelo: if you want to find all usage of 'ifconfig' in maemo-leste and submit PRs for each then we can stop depending on it | 15:12 |
Wizzup | for now I just fixed the regression | 15:12 |
sicelo | ok. will try and find time for that | 15:12 |
Wizzup | as far as I am concerned you don't have to btw | 15:15 |
Wizzup | mostly just indicating that I personally don't care too much about this | 15:15 |
Wizzup | whenever I find that ifconfig isn't installed on a machine, I just install it | 15:16 |
Wizzup | :p | 15:16 |
Wizzup | btw, I have invited to speak at https://www.dorscluc.org/ | 15:16 |
Wizzup | and might have a stand too | 15:16 |
Wizzup | (maemo) | 15:16 |
sicelo | oh cool | 15:18 |
Wizzup | it won't be anything new talk wise | 15:18 |
Wizzup | but at least it's publicity :p | 15:19 |
sicelo | it'd be great to get a user who has pinephone pro, so support for it could be done as well | 15:19 |
Wizzup | how much do they cost? | 15:21 |
Wizzup | 400 eur? | 15:22 |
sicelo | mmm, no idea tbh :-) | 15:22 |
sicelo | should be less than that | 15:22 |
Wizzup | 400 dollar then | 15:22 |
sicelo | just checked, yes | 15:23 |
Wizzup | ok, hm | 15:23 |
Wizzup | I think we could just order one or two | 15:23 |
buZz | i'm still hunting for a affordable Pocophone F1 :P | 16:11 |
Wizzup | to do a leste port? | 16:14 |
buZz | well, someone started one already? at least its somewhere between our files | 16:16 |
buZz | supposedly that Xiaomi phone can run mainline | 16:16 |
uvos | yeah the pocophone has most everything in place | 16:34 |
uvos | it just needs a maintaner | 16:34 |
uvos | it probubly bitroted quite a bit by now | 16:34 |
buZz | i guess, i just mostly want to see how well mainline works, and then try whatever i can find :) maybe pp rootfs already gets me most the way | 16:38 |
Wizzup | we have code to build poco images | 16:39 |
Wizzup | but it will take quite a lot of time to make it work well I guess | 16:40 |
freemangordon | uvos: re xorg/logind I think we shall decide what we need stopped actually in off runlevel | 16:46 |
freemangordon | I think we can have xorg/logind running till the end | 16:47 |
freemangordon | this will speed-up poweroff too | 16:47 |
freemangordon | or maybe not, maybe this will prevent unmounting filesystems | 16:48 |
buZz | Wizzup: yeah i have no dreams of making it work well ;) just work would be enough for me | 16:48 |
buZz | and seems like fun to try :) | 16:49 |
Hatman | Hi | 17:14 |
Wizzup | hello | 17:19 |
uvos | freemangordon: i dont think leaving the session/xorg running untill the kernel kills every process right at the end is sane | 17:44 |
uvos | also xorg will still likely exit with error in this case | 17:44 |
freemangordon | uvos: not that important, but it is not the kernel that kills the processes, there is 'sendsigs' init script that does it. | 18:11 |
freemangordon | however, I don;t see what is the benefit of having xorg exit gracefully | 18:11 |
freemangordon | could you elaborate? | 18:12 |
freemangordon | also, dsme can be used to control elogind/autoboot/xorg, so it is not that complicated to fix the order | 18:14 |
freemangordon | well, not elogind, but autologin/xorg | 18:14 |
freemangordon | but, if we go that route, we shall stop all processes running in xorg before stoppong xorg | 18:15 |
freemangordon | I implemented that in dsme, but reverted it, as it takes 10-15 seconds on d4 to terminate all dsme processes gracefully | 18:16 |
Wizzup | Hatman: what's up | 18:32 |
Hatman | Hi. All fine just joined | 18:57 |
Wizzup | cool, let us know if you have any questions | 19:09 |
freemangordon | Wizzup: do you fix ke-recv? | 19:26 |
freemangordon | because there is some bug causing it to segfault in VM | 19:26 |
freemangordon | Wizzup: https://pastebin.com/kJ9s8DwD | 19:29 |
freemangordon | shall I fix it? | 19:30 |
Wizzup | freemangordon: never saw that, pls check if you can | 19:56 |
Wizzup | freemangordon: looks like you already did | 19:59 |
Wizzup | :D | 19:59 |
freemangordon | yep | 20:00 |
arno11 | Wizzup: quick news about n900 calls. i've restarted from the beginning using pulse.c from master branch with very few modif and no hack | 21:12 |
arno11 | results: | 21:12 |
arno11 | i get 29ms latency very stable with good sound quality | 21:13 |
arno11 | in one way there is almost no latency | 21:13 |
arno11 | and in the other way few more | 21:13 |
arno11 | but acceptable i think | 21:14 |
arno11 | when i receive a call there is a weird 3 sec latency at the beginning of the call (PA is setting up latency) | 21:15 |
arno11 | and then everything is fine | 21:15 |
arno11 | at the opposite when i call another phone, no latency at the beginning and everything is fine | 21:16 |
norayr | folks, does droid4 vibrate when getting libnotify message? | 21:16 |
norayr | how can i make it vibrate when dino sends notification? | 21:16 |
Wizzup | arno11: 'no hack'? | 21:24 |
arno11 | Wizzup: i mean no pactl latency command | 21:25 |
arno11 | like we tried with your pulse.c version | 21:26 |
Wizzup | ok, but with my changes? | 21:26 |
arno11 | no i used the master branch version with few other changes | 21:27 |
arno11 | (just 2 or 3 values changed) | 21:27 |
Wizzup | well, before you had a much higher latency, so that's weird | 21:28 |
Wizzup | in any case the code in master is quite clearly wrong (imo) wrt latency handling | 21:28 |
Wizzup | what did you change? | 21:28 |
arno11 | well, the original code is not as wrong as we think | 21:28 |
arno11 | sec | 21:29 |
arno11 | because pa_usec_to_bytes for example cannot work | 21:31 |
arno11 | it is just like a request for pulseaudio | 21:32 |
arno11 | and it is canceled by other PA stuff | 21:32 |
arno11 | and it is in fact unstable if we do several calls | 21:33 |
arno11 | other example | 21:33 |
Wizzup | cancelled how? | 21:33 |
arno11 | sec | 21:33 |
Wizzup | online doc says this is the best way to tell PA to do low latency without hickups and other issues, IIUC | 21:34 |
Wizzup | if it gets slower maybe there is too much resampling going on (?) | 21:34 |
arno11 | yes the best way if tsched is deactivated | 21:36 |
arno11 | no diff with resampling | 21:36 |
arno11 | in every case still high | 21:36 |
arno11 | i even tried to change your settings and PA canceled in fact everything | 21:37 |
arno11 | how i know that ? i tried to test latency with my voice using always the same syllables and different settings in pulse.c and compared with pactl latency results | 21:40 |
arno11 | result is there is almost no diff. excepting stability and quality of sound | 21:40 |
arno11 | i mean, even if pactl says there is 10ms latency, it is not the case in real conditions | 21:42 |
arno11 | i think in every cases we will have a lot of resampling because other stuffs are running at 44100hz or 48000hz (like ringtones) | 21:54 |
Wizzup | (sorry, work mtg) | 22:02 |
Hatman900 | connecting to irc in n900 nice } | 22:03 |
Wizzup | -) | 22:05 |
arno11 | oh yeah no probs man, sorry. Anyway everything is running well :) | 22:07 |
arno11 | Hatman900: yeah irc rocks on n900 :) | 22:07 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!