missMyN900 | wow I put the Pinephone away the day before yesterday in a drawer after having installed Mobian | 00:09 |
---|---|---|
missMyN900 | not sure if it was even fully charged but I forgot it was still on and now I see the battery is still 34%! | 00:09 |
missMyN900 | this is really an insane difference compared to pmOS and ML | 00:10 |
missMyN900 | suspend is working really well on Mobian | 00:10 |
missMyN900 | not going to change my decision to sell it but it does show what is possible with the PP for ML eventually | 00:11 |
missMyN900 | having to rely only on suspend is primitive PM for sure but it does work really well | 00:12 |
uvos | you can suspend in ml | 00:14 |
uvos | nothings stopping you | 00:14 |
missMyN900 | but I thought there was a bug? | 00:14 |
missMyN900 | oh it just dropped to 20% suddenly after using it briefly... | 00:14 |
missMyN900 | someone on here said it would crash sometimes | 00:15 |
uvos | you have to disable longpress action | 00:15 |
uvos | in mce | 00:15 |
uvos | its not a crash | 00:15 |
uvos | its shuting down because you pressed the power button | 00:15 |
missMyN900 | but would that not occur consistently? | 00:15 |
missMyN900 | I am talking about something that would happen randomly | 00:16 |
missMyN900 | but yeah I did try suspend once briefly using ssh | 00:16 |
uvos | yes because mce by default shutsdown when you longpress the power button | 00:16 |
uvos | depending how long you press and the interval between the press and the real resume | 00:16 |
missMyN900 | and it was working but did not test it for an extended period of time to study the power consumption | 00:16 |
uvos | it might end up looking like longpress | 00:16 |
missMyN900 | I thought you were talking about something like that wrt a mapphone yesterday? | 00:17 |
uvos | no that was about the pp | 00:17 |
uvos | the mapphones obv dont have sutch an issue | 00:17 |
uvos | they suspend/resume 30 times a second | 00:17 |
uvos | or more | 00:17 |
missMyN900 | yes TI OMAP is very nice | 00:17 |
missMyN900 | I wonder if the i.MX6/8 are just as good in that aspect | 00:18 |
missMyN900 | (I have an old Wandboard Dual with i.MX6) | 00:18 |
missMyN900 | I have always liked i.MX | 00:18 |
uvos | anyhow | 00:18 |
missMyN900 | sadly there are not that many images for the Wandboard | 00:18 |
missMyN900 | and most require using the serial port which is annoying | 00:18 |
missMyN900 | I never got a serial cable | 00:19 |
uvos | if you have even trival c knowlage you could write a mce module to suspend when the screen is off | 00:19 |
missMyN900 | I see, that is nice | 00:19 |
missMyN900 | I will still sell the PP though because of other things | 00:19 |
missMyN900 | I am hoping I can get a DB eventually | 00:19 |
missMyN900 | but right now it does not work for me | 00:20 |
missMyN900 | too much stuff to do and too many devices already | 00:20 |
missMyN900 | I would like to but I have to be realistic | 00:20 |
uvos | i mean bionics are <30 euro and installing leste on one takes 45min but ok | 00:21 |
missMyN900 | in theory it takes 45 min for an experienced person like you ;) | 00:21 |
missMyN900 | for me it will probably take much more than that | 00:21 |
missMyN900 | and it is not just installation | 00:22 |
missMyN900 | I also need to find the time to actually do something with it | 00:22 |
uvos | sure | 00:22 |
missMyN900 | otherwise it will just be sitting there with an empty battery that is degrading | 00:22 |
missMyN900 | so I will have to see how things go | 00:23 |
missMyN900 | I have still have many things to do and I want to mod qtwebbrowser first anyway | 00:23 |
missMyN900 | if I am going to use the DB once every 6 months, I might as well just run the QEMU VM | 00:24 |
missMyN900 | uvos: by the way, have you seen the new Pocket Reform Mini | 00:27 |
missMyN900 | it is really awesome | 00:27 |
missMyN900 | don't think it has touch screen so not really a potential ML target but still a really cool mobile Linux device | 00:27 |
missMyN900 | it comes with an i.MX8M Plus by default but can also use a Raspberry Pi CM4 or Pine64 RK3566-based SO module as motherboard | 00:28 |
missMyN900 | uvos: do you think you will ever get the Motorola LTE modem to work on the DB and other mapphones? | 00:34 |
missMyN900 | because I recall you saying that it is possible to make it work on other bands than the official ones or maybe that was referring to another modem | 00:34 |
uvos | i dont see the point of those types of devices really | 00:34 |
missMyN900 | DB? | 00:34 |
missMyN900 | or the Motorola modem in the mapphones? | 00:35 |
uvos | prm | 00:35 |
missMyN900 | what is prm? | 00:35 |
missMyN900 | oh | 00:35 |
missMyN900 | the Reform laptop | 00:35 |
missMyN900 | why not? | 00:35 |
missMyN900 | the size? | 00:35 |
missMyN900 | or are you also referring to the original Reform? | 00:35 |
uvos | i dont see the usecase where i could carry that | 00:35 |
uvos | but not a laptop | 00:35 |
missMyN900 | yeah | 00:35 |
uvos | and laptops are plety linux compatabil | 00:36 |
missMyN900 | I can understand that | 00:36 |
missMyN900 | true | 00:36 |
missMyN900 | but there are no proprietary blobs on the i.MX8 AFAIK | 00:36 |
uvos | sure | 00:36 |
missMyN900 | there is no proprietary BIOS | 00:36 |
missMyN900 | and no proprietary firmware | 00:36 |
missMyN900 | *UEFI | 00:36 |
uvos | i could still buy a old thinkpad if that was a priority for me | 00:36 |
uvos | and get better perf | 00:36 |
missMyN900 | sure but this is something that you can buy new | 00:36 |
uvos | (it isent a priority tho) | 00:36 |
missMyN900 | those won't be available forever | 00:37 |
missMyN900 | that is why the PP project was also started, I think | 00:37 |
uvos | then a chromebook fits the bill | 00:37 |
uvos | anyhow | 00:37 |
uvos | wirt the modem | 00:37 |
uvos | no idea | 00:37 |
missMyN900 | of course, the Reform is closer to a Librem 5 than the PP | 00:37 |
missMyN900 | PBP is of course closer to the PP;) | 00:37 |
uvos | its not a priority atm | 00:37 |
uvos | i also have a limesdr | 00:38 |
missMyN900 | LOL | 00:38 |
uvos | that could be a lte basestation | 00:38 |
uvos | to try it | 00:38 |
missMyN900 | are you going to iplemet LTE on it? | 00:38 |
uvos | sure would not be so hard | 00:38 |
uvos | to test the modem | 00:38 |
missMyN900 | I don't think the power consumption is going to be good for mobile... | 00:38 |
missMyN900 | oh | 00:38 |
uvos | no i mean as a basestation | 00:38 |
uvos | with us specs | 00:38 |
uvos | but its not a priority as you can imagine | 00:38 |
missMyN900 | I have an RTL-SDR | 00:38 |
missMyN900 | but I am not doing that advanced stuff | 00:39 |
missMyN900 | I failed to decode any digital signal with it | 00:39 |
missMyN900 | compiled lots of stuff manually | 00:39 |
uvos | well that dosent have close to the specs to do lte anyhow | 00:39 |
missMyN900 | yep | 00:39 |
missMyN900 | but you said 2G and 3G are also shutting down in Europe? | 00:41 |
uvos | only 3g | 00:42 |
uvos | for now (2030 ish) | 00:42 |
missMyN900 | oh | 00:43 |
missMyN900 | then there is still plenty of time to get the Motorola modem working ;) | 00:43 |
missMyN900 | or to start a NeoDroid4 project :D | 00:43 |
missMyN900 | it can have a 5G modem ;) | 00:44 |
missMyN900 | hopefully Mediatek learns something from their Chromebook experience | 00:44 |
norayr | uvos, Wizzup, this atrix 2 device will work with lapdock under linux? | 01:16 |
uvos | norayr: sure | 01:17 |
uvos | norayr: but so dose droid 4 or bionic | 01:17 |
uvos | they all share the same port layout so you can connect them to eatch others lapdocks | 01:18 |
uvos | (except the atrix 1 one its reversed) | 01:18 |
norayr | omg i did not know. | 01:27 |
norayr | and what about droid 3? | 01:27 |
uvos | i think it should work to | 01:27 |
norayr | i don't know actually how they connect. via usb? | 01:27 |
uvos | but it dosent have offical support i think | 01:28 |
uvos | so all motorla devices of this time period have hdmi out and usb otg in the same place | 01:28 |
uvos | the lapdock has a cradle that connects to both | 01:28 |
norayr | understand. | 01:29 |
uvos | for the first lapdock the cradle was specifc to the deivce | 01:29 |
uvos | (ie the atrix 4g) | 01:29 |
uvos | but later its just a slot with the connectors | 01:29 |
uvos | so any phone will fit that has the same port layout | 01:29 |
uvos | and all the phones also had the linux install in parralell with android | 01:30 |
uvos | (ie ubuntu 12.04 based linux/xorg) | 01:30 |
norayr | oh true, i start to remember that. | 01:32 |
norayr | so when connected to lapdock they were running ubuntu. | 01:32 |
norayr | not android. | 01:32 |
uvos | both | 01:32 |
norayr | wow. | 01:32 |
uvos | it spins up a container | 01:32 |
norayr | yes. | 01:32 |
uvos | the d3 cant do it | 01:33 |
uvos | (probubly because of the 512mb ram) | 01:33 |
norayr | i probably won't get a lapdock, but it's interesting. | 01:33 |
norayr | and for d3 i would run only linux anyway. | 01:34 |
norayr | the possibility to reuse old mapphones is amazing. | 01:35 |
norayr | since i only use d4 recently | 01:35 |
norayr | i noticed people don't even realize it is an old device. | 01:35 |
norayr | they think it is something new and shiny. | 01:35 |
norayr | i just had this idea that we already live in a dystopia. | 01:42 |
norayr | it's there when 'war is peace' and when (also in post apocalyptic societies) people search and trade devices that are no more produced. | 01:51 |
norayr | why omap chipsets were well supported in mainline? should we thank ti for that? did they contribute to mainline? | 01:53 |
norayr | and are. | 01:53 |
mglbg[m] | I have been fiddling around with ML for a bit now, and it appears that removing the NOPASSWORD: part from /etc/sudoers.d/maemo breaks the app manager. how do you guys feel about that safety wise, and is this a correct observation or did I do some other stuff as well that broke it? | 09:26 |
uvos | mglbg[m]: that is correct | 10:27 |
Wizzup | mglbg[m]: yeah that is a problem | 10:27 |
uvos | mglbg[m]: you must provide a askpass implmentation | 10:27 |
Wizzup | mglbg[m]: although I think the app manager has it's own specific entry in sudoers AFAIK | 10:27 |
uvos | it dosent work | 10:27 |
uvos | never did | 10:27 |
uvos | i reported this before | 10:27 |
Wizzup | ok, well, we'd need to fix that | 10:28 |
Wizzup | uvos: sorry, didn't remember | 10:28 |
uvos | thers also a bug somewhere | 10:28 |
uvos | in git | 10:28 |
Wizzup | hm | 10:28 |
uvos | i mean a bug report | 10:28 |
uvos | btw did you see the kernel? | 10:28 |
Wizzup | uvos: will fix perms, just woke up | 10:28 |
Wizzup | vi | 10:28 |
uvos | ok i wont be around today | 10:29 |
uvos | maybe just pull from my repo and do a release | 10:29 |
uvos | i hope i dident miss anything | 10:29 |
Wizzup | I will also be with my parents part of the time | 10:29 |
Wizzup | let me give you perms anyway | 10:29 |
uvos | patch wise | 10:29 |
uvos | anyhow i dident test it on n900 yet | 10:29 |
uvos | so we have to do that too | 10:30 |
Wizzup | ok | 10:30 |
Wizzup | we can use -experimental | 10:30 |
Wizzup | nice work on the audio patch, at least starts ofi t | 10:30 |
Wizzup | I was thinking to write to sre/tmlind and the list and see what they think - since sre wrote a lot of the audio stuff (at least per the copyright in the file) | 10:30 |
uvos | he did | 10:31 |
uvos | but the modem audio stuff isent form him | 10:31 |
uvos | (nor is it in mainline) | 10:31 |
uvos | its something tmlind hacked in there | 10:31 |
Wizzup | ok | 10:36 |
Wizzup | mglbg[m]: can you file an issue for this problem? (if it doesn't exist already) | 10:37 |
uvos | https://github.com/maemo-leste/bugtracker/issues/403 | 10:38 |
Wizzup | ty | 10:38 |
Wizzup | I don't think we want to use askpass for this per se if it calls sudo for various things | 10:38 |
Wizzup | right? | 10:38 |
uvos | we def want askpass as a backup | 10:39 |
uvos | also i dont like ham being able to use apt | 10:39 |
uvos | without pw | 10:39 |
uvos | anyhow its not just ham | 10:39 |
uvos | the theme switcher and qutie some other stuff also breaks | 10:39 |
Wizzup | yeah | 10:40 |
uvos | askpass also makes you see all of these | 10:40 |
Wizzup | so which branches do you want me to sync over to droid4-linux, or will you do it? | 10:40 |
uvos | well maemo-5.18 at least | 10:40 |
uvos | if you want he feature branches -516 -518 then those too | 10:41 |
uvos | maemo-5.18 comes to be by merging tmlinds -pending with all of the feature branches | 10:41 |
uvos | (and 5.18.8 stable) | 10:41 |
bencoh | rafael2k: bayer can be anywhere between 8 and 16, it's often 10b or 12b | 11:00 |
bencoh | it really depends on the sensor and how it was configured | 11:01 |
Wizzup | what do you think makes the most sense for our -dbgsym packages and h-a-m | 12:11 |
Wizzup | we can either hide them alltogether, or show them only in 'all', or give them a 'debug' section | 12:11 |
Wizzup | I think we should just hide them, no developer will use h-a-m to install the debug symbols imho | 12:11 |
Wizzup | https://github.com/maemo-leste/bugtracker/issues/482 relevant issue | 12:12 |
Wizzup | https://github.com/maemo-leste/hildon-application-manager/commit/37bf83aacce04c8ce36cd9a83c67fbcf71bee63e | 12:16 |
lel | MerlijnWajer closed an issue: https://github.com/maemo-leste/bugtracker/issues/482 (HAM: Hide (or toggle) dbgsyms) | 12:54 |
bencoh | Wizzup: yeah, I'd just hide them | 13:30 |
rafael2k | bencoh: right, tks | 14:06 |
rafael2k | took just my pinephone to the weekend (without the keyboard), I was pretty happy to read news and listen you youtube music on the LTE connection, using QT WebBrowser, which has a nice vkb | 14:37 |
rafael2k | btw, good news from camera side, it seems libcamera's gui app, qcam, is closer to working | 15:29 |
rafael2k | aaaand, we have /usr/lib/aarch64-linux-gnu/gstreamer-1.0/libgstlibcamera.so | 15:31 |
rafael2k | : ) | 15:31 |
rafael2k | so no urgency for chimaera | 15:32 |
rafael2k | I just have some some colorspace conversions to sort out in order qcam start to work... it seems. At least it opens up already and correctly list the cameras. | 15:44 |
bencoh | neat | 16:15 |
rafael2k | just got it working, with fields looking funny in the screen, dunno why I'm getting an error when I set qcam to use gles rending | 16:17 |
rafael2k | *rendering | 16:17 |
rafael2k | hum QXcb in qcam says neither GLX nor EGL is enabled | 16:22 |
rafael2k | we should have some of those right? | 16:22 |
rafael2k | qcam -s pixelformat=YUYV | 16:23 |
rafael2k | it starts, but colorspace conversion is fuck up, funny images on screen | 16:23 |
bencoh | well, if it receives bayer, that's expected | 16:24 |
bencoh | if you really want to render bayer without converting it, you should choose a grayscale pixelformat | 16:24 |
bencoh | (luminance-only, 8b if working with RAW8) | 16:25 |
bencoh | (and 16b if working with RAW10/RAW12, and expect endianness issues) | 16:25 |
rafael2k | writing to libcamera crew... they seem to be interested in supporting us to have it working, lets see | 16:25 |
bencoh | regarding glx/egl, we should support EGL outofthebox (at least a certain version/subset of it) | 16:26 |
rafael2k | indeed, nobody really made libcamera work properly with the PP cameras setup | 16:26 |
rafael2k | right | 16:26 |
rafael2k | btw, can anyone remind me again how to set a volume button to trigger the vkb? | 16:33 |
rafael2k | this is very useful when not having a working keyboard | 16:34 |
rafael2k | btw, anyone wanting to play with libcamera in ML: https://www.abradig.org.br/maemo-crazyness/libcamera/ | 16:38 |
rafael2k | these are libcamera packages without any change from upstream (I just set a 0.1 version) | 16:38 |
rafael2k | hopefully soon in our repos | 16:40 |
rafael2k | : ) | 16:40 |
rafael2k | I already uploaded my pet repo to https://github.com/maemo-leste-upstream-forks/libcamera | 16:40 |
rafael2k | btw, libcamera was supposed to convert from bayer to yuyv automatically when I chose a different color spacẹ.. which clearly did not happen - more likely an implementation bug in libcamera | 16:42 |
rafael2k | QXcbIntegration: Cannot create platform OpenGL context, neither GLX nor EGL are enabled | 19:50 |
rafael2k | hum | 19:50 |
rafael2k | in mobian it seems to work | 19:50 |
rafael2k | I wonder what it could be | 19:55 |
uvos | rafael2k: try with -platform xcb | 21:08 |
rafael2k | lemme try | 21:13 |
rafael2k | hum, images appear now! but with some crazy swapped fields | 21:18 |
rafael2k | what does it mean this "-platform xcb" | 21:21 |
uvos | uses desktop linux qt | 21:22 |
uvos | instead of maemo qt | 21:22 |
rafael2k | qcam -r gles -spixelformat=SBGGR8,width=1280,height=720 -platform xcb | 21:23 |
rafael2k | fucking shit, it worked!! | 21:23 |
rafael2k | 7 fps for the viewer | 21:23 |
rafael2k | not bad | 21:23 |
rafael2k | : ))) | 21:24 |
rafael2k | now we can take pictures with a gui in ML in PP | 21:25 |
rafael2k | : ))) | 21:25 |
rafael2k | front camera driver still needs some love (image is a but funny still) but the main rear camera is looking fine! | 21:28 |
dsc_ | nice one | 21:30 |
rafael2k | yay! | 21:42 |
rafael2k | tks uvos! | 21:43 |
rafael2k | still some artifacts... https://www.abradig.org.br/maemo-crazyness/back_camera-qcam-2.jpg | 21:58 |
rafael2k | but my daughter seems happy with it already | 21:58 |
rafael2k | bencoh: gimme some hint of this type of artifact ^ | 21:59 |
bencoh | :] | 22:05 |
bencoh | are you sure it's supposed to be BGGR8? | 22:06 |
bencoh | and are you using the right picture size? | 22:06 |
rafael2k | Now I'm just driving a truck đownstream without breaks | 22:07 |
bencoh | uh | 22:07 |
rafael2k | running "qcam" with all autogatic conversions | 22:07 |
bencoh | oh | 22:07 |
rafael2k | *automagic | 22:08 |
rafael2k | seems to work almost ok, tbh | 22:08 |
bencoh | personally I'd use v4l2 without libcamera, convert it manually using makeDNG+ufraw, check I understand what really happens there, compare with what libcamera does (just to make sure you use the same format) | 22:08 |
rafael2k | no need to go that far, with libcamera "cam" app I can get the "original" raw pictures, which look fine indeed | 22:09 |
rafael2k | using any canonical color conversion tool gives me an output which says the capture side is "ok" | 22:10 |
rafael2k | libcamera "raw" capture seems ok... automagic conversion seems a bit fucked up | 22:11 |
bencoh | ah, I see | 22:12 |
rafael2k | even considering trashy fffmpeg conversion | 22:13 |
rafael2k | https://www.abradig.org.br/maemo-crazyness/back_camera.jpg | 22:13 |
rafael2k | I don't see those strange "blocky" patterns I get from qcam | 22:13 |
uvos | pixel format is clearly wrong | 22:14 |
rafael2k | clearly seem here https://www.abradig.org.br/maemo-crazyness/back_camera-qcam-2.jpg | 22:14 |
rafael2k | uhum | 22:14 |
uvos | it has less lines | 22:14 |
uvos | probubly be even factor | 22:14 |
rafael2k | I think developers realized that | 22:14 |
rafael2k | trying to push them to fix... as it is clear the error | 22:14 |
rafael2k | hummmm true | 22:15 |
rafael2k | lemme try another res | 22:15 |
rafael2k | anyway, need to sleep, tomorrow I'll try a bit more, but anyway, the packages are up for anyone with a PP willing to try the luck: https://www.abradig.org.br/maemo-crazyness/libcamera/ | 22:17 |
rafael2k | with the help of uvos, anyone wanting risk yourself: qcam -r gles -spixelformat=SBGGR8,width=1280,height=720 -platform xcb | 22:18 |
rafael2k | is a starting point | 22:18 |
Wizzup | uvos: will do the kernel now | 22:35 |
Wizzup | rafael2k: great news re: kernel | 22:43 |
Wizzup | well, re: camera | 22:44 |
Wizzup | uvos: I need to make the version 5.18.8 instead of 5.18.0 but otherwise it's running atm | 22:56 |
uvos | on what? | 23:06 |
Wizzup | d4 linux for beowulf experimental | 23:21 |
Wizzup | hmm the source build failed | 23:22 |
Wizzup | should I remove the debian/patches ? | 23:22 |
Wizzup | oooh I built the wrong repo | 23:22 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!