kshatriya | this article describes (perhaps) what is done to get usb OTG host mode working, while charging both host and peripherals with external power https://blog.workshoptwelve.com/2019/01/04/otg-usb-charge-hub-wtf/ | 04:10 |
---|---|---|
roalyr | So since we are flashing a microSD card, how does using ML affects its lifespan? | 08:14 |
roalyr | And how large of a SD card would you suggest? | 08:15 |
parazyd | sdcards are pretty good these days | 08:40 |
parazyd | I mostly use 32G, but even 8 is decent for Leste | 08:41 |
roalyr | Well yes, it is not like we have the luxury of using persistence mode like in MX linux with that amount of RAM. | 08:42 |
roalyr | Looks like I am stuck at "fastboot flash mbm allow-mbmloader-flashing-mbm.bin" step. | 08:53 |
roalyr | Terminal output is: "target didn't report max-download-size | 08:53 |
roalyr | And it is staying at sending 'mbm' (512 KB)... | 08:54 |
roalyr | Fastboot returns the devices list though. | 08:55 |
sicelo | kshatriya: looks like complicating things ... I thought d4 can charge and host same time ootb, if I recall what uvos said/tried. At least N900, with its hacky/broken otg dies | 08:56 |
sicelo | *does | 08:56 |
roalyr | Okay, looks like I have to try different USB ports. | 08:57 |
lel | parazyd renamed a repository: https://github.com/maemo-leste/tor-network-applet | 09:00 |
uvos | roalyr needs to use flash:raw | 09:16 |
uvos | as to why thats nessecary on some d4's no idea, might be old mbm version or a interaction with fastboot version | 09:16 |
uvos | roalyr: if you come back and read the logs please fastboot --version as well as the mbm version (the nummber displayed on device under AP Fastboot Flash Mode in fastboot mode) | 09:18 |
roalyr | What is the way to oem unlovk D4? | 09:21 |
uvos | you cant | 09:22 |
uvos | the bootloader is permanently locked | 09:22 |
roalyr | So how do I even flash it? | 09:22 |
uvos | we use a vunerability in the boot process to essentally boot a second bootloader | 09:22 |
uvos | thats unlocked | 09:22 |
roalyr | In my case flashing mbm hangs and doesn't proceed. | 09:23 |
uvos | plase provide the information above | 09:23 |
uvos | please insure the the allow-mbmloader-mbm.bin is free from defect | 09:23 |
uvos | its md5sum should be 8875e9a9860522aa99568b15896fa076 | 09:24 |
roalyr | It is as such. | 09:25 |
uvos | plaase provide the information requested above | 09:26 |
roalyr | Which one? Please re-send the message. | 09:26 |
uvos | [09:16] <uvos> as to why thats nessecary on some d4's no idea, might be old mbm version or a interaction with fastboot version | 09:26 |
uvos | [09:18] <uvos> roalyr: if you come back and read the logs please fastboot --version as well as the mbm version (the nummber displayed on device under AP Fastboot Flash Mode in fastboot mode) | 09:26 |
roalyr | Fastboot (Xubuntu 20.10): 1:8.1.0+r23-8 | 09:27 |
roalyr | AP: 0A.77 | 09:28 |
uvos | thats really old | 09:29 |
uvos | (fastboot, mbm is fine) | 09:29 |
uvos | could you try with https://developer.android.com/studio/releases/platform-tools? | 09:29 |
roalyr | I see, got it from standard repo. | 09:29 |
uvos | i use 30.0 and this works fine | 09:30 |
uvos | if the fastboot version is not it, then i dont know why some d4s react differently | 09:30 |
uvos | in this case you have to use flash:raw | 09:30 |
uvos | but i dont reccomend this | 09:30 |
uvos | as it skips some checks and you are flashing a bootloader | 09:31 |
uvos | with obvious bricking implications | 09:31 |
roalyr | Let me try the new fastboot first. | 09:32 |
roalyr | It worked! | 09:33 |
uvos | yay | 09:33 |
roalyr | Phew. It would have been depressing to get a brick that fast. | 09:33 |
uvos | tmlind: ^^^^^ could you please add a note to kexecboot instructions that at least fastboot v30.0 is needed | 09:34 |
uvos | clearly some old versions have issue with mbm | 09:34 |
roalyr | Yay, all good. | 09:36 |
roalyr | BTW, that means I could also flash some more modern android-based OS too... Brings me nostalgic memories of custom ROMs. | 09:38 |
uvos | the maximum is android 7.0, mostly because no one cares you could run android 11 on the mainline kernel | 09:38 |
uvos | or 12 even | 09:39 |
roalyr | I'd stick for 10. | 09:39 |
uvos | well you would have to build it yourself | 09:39 |
uvos | but in theory it should work | 09:40 |
roalyr | I guess. | 09:40 |
roalyr | It loads ML! | 09:45 |
uvos | great | 09:52 |
roalyr | what browser should I get? | 10:01 |
uvos | firefox-esr | 10:02 |
uvos | and then set some evvars to allow thumb scrolling | 10:02 |
uvos | and customize the ui a bit | 10:02 |
uvos | works really well | 10:02 |
roalyr | Is there home folder? | 10:02 |
uvos | http://uvos.xyz/maserati/firefox-demonstration.mp4 | 10:02 |
uvos | roalyr: ? | 10:02 |
uvos | maemo leste is just debian | 10:02 |
uvos | so yeah | 10:02 |
roalyr | As in for user files. | 10:02 |
uvos | there is /home/user | 10:02 |
parazyd | fwiw: https://leste.maemo.org/Tricks#Touchscreen_support_in_Firefox_.28firefox-esr.29 | 10:05 |
parazyd | btw. one caveat to setting Android useragent is that it blocks you from installing certain plugins | 10:06 |
roalyr | That's fine. | 10:09 |
roalyr | How is battery calibrated? | 10:09 |
roalyr | And how is virtual keypad called? | 10:09 |
roalyr | Oh, figured out thaa it pops up when physical is not deployed. | 10:10 |
roalyr | Okay, the most important question - how to disable auto-rotation to prevent graphics from dying? | 10:13 |
uvos | graphics should not die on rotation | 10:14 |
uvos | on d4 | 10:14 |
uvos | what do you mean "die"? | 10:14 |
roalyr | A few rotation and X server (or whatever server) dies and device reboots. | 10:15 |
uvos | this can happen due to a bug in pvr when the virtual keyboard is open | 10:15 |
uvos | otheriwse it never happens to me | 10:15 |
roalyr | I see. | 10:15 |
roalyr | But I'd still switch the turning off. | 10:16 |
uvos | you can download oritation lock applet | 10:16 |
uvos | and activate that | 10:16 |
uvos | (name aproximate) | 10:16 |
uvos | id also recommend the simple brightness widget | 10:16 |
uvos | go apt search for them | 10:16 |
roalyr | Are maemo package mirrors down? | 10:24 |
uvos | no | 10:25 |
uvos | you must apt update first | 10:25 |
roalyr | Oh, I think wifi settinhs didn't work out. | 10:25 |
roalyr | How is it enabled? | 10:35 |
uvos | roalyr: how is what enabled? | 10:43 |
sicelo | Landscape only users are rare. I'm getting curious | 10:48 |
sicelo | roalyr: should be able to bring up the status menu by clicking the time display, then you'll see "Connect" or similar | 10:50 |
roalyr | I am not seeing neither orientation nor brightness applets. Should be restarting, I guess? | 10:57 |
sicelo | Or restarting the status menu ... I don't use those applets, so I don't know what happens when you install them | 11:10 |
roalyr | Okay, restarting enabled the notification area applets. | 11:46 |
roalyr | But what about droid 4 symbols on physical keyboard/ | 11:47 |
uvos | what about them | 11:48 |
uvos | please be more specific when asking questions | 11:48 |
roalyr | Well, as far as I understand by defauly 'sym' key does not enable the access to the symbols | 11:49 |
uvos | no symb is alt | 11:50 |
uvos | various symbols ar available via ok+ whatever | 11:50 |
uvos | otherwise you can use vkb | 11:50 |
roalyr | there is no alt key. | 11:51 |
uvos | sym==alt | 11:51 |
uvos | caps-lock==shift | 11:51 |
uvos | shift==ctrl | 11:51 |
uvos | on d4 hwkbd | 11:51 |
roalyr | Oh... | 11:51 |
uvos | ok == sym (sorta) | 11:52 |
uvos | ok+p is pipe for instance | 11:52 |
uvos | there are several ok+ for special chars | 11:52 |
roalyr | Well, it seems like there is no straight way to access the symbols then | 11:53 |
uvos | ? | 11:53 |
roalyr | On hw keyboard | 11:53 |
uvos | ? | 11:54 |
uvos | i just told you | 11:54 |
roalyr | Ok dkey does nothing. | 11:54 |
uvos | via ok+ several keys (just try it) you can acess everything | 11:54 |
uvos | ok+o is } for instance | 11:54 |
uvos | and ok+p is | | 11:54 |
uvos | and so on | 11:54 |
roalyr | Doesn't work in my case. | 11:55 |
uvos | run as root | 11:55 |
uvos | udevadm hwdb --update | 11:55 |
uvos | udevadm trigger | 11:55 |
uvos | and then hildon-desktop-rotate-touchscreen.sh as normal user | 11:56 |
roalyr | This did work;:}| | 11:57 |
uvos | parazyd: ^^^^ this still happens | 11:58 |
roalyr | Are the symbols such as plus, underscore, uestion mark mapped | 12:04 |
sicelo | They work as marked. Alt + symbol, I think | 12:21 |
Wizzup | morning | 12:30 |
parazyd | Besides what is currently in leste-config, we could have an apt hook to run this | 12:49 |
uvos | parazyd: the problem is not apt | 12:56 |
uvos | parazyd: its that it dosent run on first boot | 12:56 |
uvos | udev should update the hwdb on first boot | 12:56 |
uvos | but apearntly it dosent | 12:56 |
uvos | maybe this is caused by the timestamaps from jenkins | 12:56 |
uvos | vrs the d4 which will have its time a 1. jan 1970 innitally on first boot | 12:57 |
parazyd | Can it run in the image build or it needs to be on-device? | 12:57 |
uvos | no | 12:57 |
uvos | it must be on device | 12:57 |
parazyd | ok | 12:57 |
uvos | udev creates the hwdb with info from the kernel | 12:57 |
uvos | running kernel that is | 12:57 |
uvos | parazyd: could we use fake-hwclock to ensure that the system time is after the build time of the image | 12:58 |
uvos | (if that is the problem at all) | 12:58 |
uvos | on first boot | 12:58 |
parazyd | We could, but I had plans to finish reproducible image builds, so this'd break it | 12:59 |
uvos | hmm yeah | 12:59 |
uvos | maybe we should just run it by hand on first boot | 12:59 |
parazyd | We already have an initscript that runs on first boot | 12:59 |
uvos | to force udev to update | 12:59 |
uvos | regardless of time | 12:59 |
parazyd | zzz something | 12:59 |
parazyd | Also hildon-base | 13:00 |
kshatriya | roalyr: i added a keyboard layout graphic to the wiki https://leste.maemo.org/Motorola_Droid_4#Notes | 13:18 |
Wizzup | cool | 13:19 |
kshatriya | need to add ` sorry, fixing | 13:20 |
Wizzup | kshatriya: did you see the ones I autogenerated? | 13:22 |
kshatriya | i based it off your post https://github.com/maemo-leste/bugtracker/issues/347 | 13:22 |
Wizzup | https://github.com/maemo-leste/bugtracker/issues/347#issuecomment-640127353 | 13:22 |
Wizzup | ok | 13:22 |
Wizzup | ool | 13:22 |
Wizzup | cool! | 13:22 |
kshatriya | i dont understand the * you have on modifier-a | 13:22 |
kshatriya | for soem reason the 'serial cable' entry is to the right of the keyboard picture | 13:26 |
parazyd | Wizzup: Do you want me to implement a shared lib for generating torrc from gconf so it's also usable in icd? | 13:27 |
parazyd | It's be part of the applet repo/package | 13:28 |
Wizzup | I don't know, I'd say no for now since I don't know if I need it | 13:28 |
uvos | kshatriya: loose the left | 13:28 |
uvos | i think | 13:28 |
Wizzup | how about a binary? | 13:28 |
parazyd | There's a bit more generic code there too, so I could make a nice lib | 13:28 |
parazyd | Wizzup: Could do that as well | 13:28 |
parazyd | Then you can actually: gentorrc | tor -f - | 13:29 |
parazyd | Good idea | 13:29 |
uvos | [[File:Maemo-leste-droid4-kbd-2021-07.png|center|620px|thumb|maemo-leste droid4 keyboard]] | 13:29 |
uvos | kshatriya ^^^ | 13:29 |
kshatriya | thanks, fixed | 13:32 |
uvos | aemo-leste droid4 keyboard | 13:32 |
uvos | is missing a m | 13:32 |
kshatriya | fixed | 13:32 |
parazyd | It could also be a shell script | 13:32 |
uvos | {{center|maemo-leste droid4 keyboard}} | 13:32 |
uvos | in place of just |maemo-leste droid4 keyboard | 13:32 |
uvos | is supperior i thin | 13:32 |
uvos | k | 13:32 |
uvos | but is ok as is | 13:32 |
kshatriya | [[File:Maemo-leste-droid4-kbd-2021-07.png{{center|maemo-leste droid4 keyboard}}]] doesn't show me a picture in preview | 13:33 |
uvos | [[File:Maemo-leste-droid4-kbd-2021-07.png|center|620px|thumb|{{center|maemo-leste droid4 keyboard}}]] | 13:33 |
kshatriya | that puts a link below pic which says "Template:Center" | 13:34 |
uvos | hmm | 13:34 |
uvos | ok | 13:34 |
uvos | nvm then | 13:34 |
uvos | its not important | 13:34 |
kshatriya | i just noticed ~ tilde is also on OK+z | 13:35 |
kshatriya | unless that's just on mine | 13:35 |
uvos | no | 13:35 |
uvos | thats how it is | 13:36 |
uvos | no idea why | 13:36 |
uvos | probubly a mistake | 13:36 |
uvos | dont document it | 13:36 |
kshatriya | ok | 13:36 |
uvos | ok+z == XK_Next | 13:44 |
uvos | wierd | 13:44 |
uvos | Wizzup: ^^^ | 13:44 |
kshatriya | this shows off h265 HEVC playback on droid4 nicely (8MB) Axel Borg on his flying chair - doubt he'll mind sharing the video http://0x0.st/-Wta.mkv | 14:04 |
roalyr | How to properly calibrate the battery? | 14:23 |
kshatriya | use device until battery is low, then before it shuts off, charge it to full without a reboot | 14:24 |
kshatriya | did you see the keyboard layout graphic i made roalyr ? | 14:25 |
kshatriya | the battery status monitor could have an uncalibrated mode of operation - where it alarms on low voltage and prompts user to charge to full | 14:26 |
kshatriya | but someone would need to code that | 14:26 |
roalyr | Sorry, couldn't keep up, on a lunch in between work now, gotta run soon again. I've explored the Ok+ combos, alas couodn't find the +, ?, * symbols. | 14:28 |
kshatriya | plus is shift = | 14:32 |
kshatriya | asterisk is shift 8 | 14:32 |
kshatriya | question mark is shift / | 14:32 |
u0_a482 | By the way, is hardware acceleration possible? Tried to run a few applications and it outputted that omapdrm_dri is not found. I presume that's the case for a missing mesa library? | 15:08 |
u0_a482 | Or is there more to it? | 15:08 |
Wizzup | hardware accel for what? | 15:09 |
Wizzup | opengl? | 15:09 |
u0_a482 | Yes. | 15:09 |
Wizzup | opengl|es accel is there and working | 15:09 |
Wizzup | opengl is not supported by the driver but you can look at stuff like gl4es | 15:09 |
Wizzup | (without opengl|es, hildon-desktop would -not- be smooth) | 15:09 |
parazyd | Yeah we have also gl4es but didn't experiment a lot | 15:10 |
u0_a482 | So I just ignore that error in terminal? | 15:10 |
Wizzup | u0_a482: no, you might need to do some more work, like ld preload gl4es | 15:10 |
Wizzup | brb | 15:10 |
u0_a482 | Okay. Will look into it. | 15:10 |
roalyr | Looks like working things out will get me to learn linux much more deeply. | 15:11 |
parazyd | https://github.com/maemo-leste-upstream-forks/gl4es | 15:14 |
parazyd | There might be some upstream commits missing | 15:14 |
buZz | is gl4es apt-getable now? | 15:31 |
parazyd | Yeah, but risky afaik | 15:32 |
buZz | ok, i'll refrain for now then :) | 15:32 |
parazyd | I mean, you can always delete kt | 15:32 |
parazyd | I'd be good if someone tries it out though | 15:32 |
kshatriya | i think sdl will default to opengl if available, not using the -es path | 15:33 |
kshatriya | oh i'm installing and deinstalling it all the time | 15:33 |
buZz | ah hmhm | 15:33 |
buZz | would it force hildon to work through gl4es then? | 15:33 |
Wizzup | no | 15:33 |
kshatriya | no | 15:33 |
Wizzup | gl4es is LD_PRELOAD only | 15:33 |
buZz | then i guess fine :) | 15:33 |
buZz | ahh | 15:33 |
Wizzup | so it really shouldn't affect system wise | 15:33 |
Wizzup | wide | 15:33 |
uvos | gl4es is system wide | 15:34 |
Wizzup | uvos: wait what | 15:34 |
Wizzup | do we install it system wide? | 15:34 |
kshatriya | does it still symlink /usr/lib/arm-linux-gnueabihf/libGL.so to /usr/lib/gl4es/libGL.so.1 | 15:34 |
uvos | Wizzup: yes | 15:34 |
Wizzup | hm | 15:34 |
uvos | thats why its risky | 15:35 |
uvos | i breaks things | 15:35 |
uvos | because apps that ignore llvmpipe and use es | 15:35 |
uvos | use gl4es now | 15:35 |
parazyd | Yeah the point was to have gl4es to be a replacement | 15:35 |
Wizzup | I guess it makes some sense | 15:35 |
Wizzup | ok | 15:35 |
parazyd | But we need to use it and experiment | 15:35 |
uvos | we should have glvnd | 15:35 |
parazyd | It was just a oneshot | 15:35 |
parazyd | uvos: We have it | 15:35 |
uvos | and switch to gl4es on demand | 15:35 |
uvos | yeah but gl4es dosent use glvnd atm | 15:36 |
uvos | so you cant switch it in or out | 15:36 |
parazyd | mhm | 15:36 |
parazyd | It needs some love | 15:36 |
parazyd | For Leste we have the newest mesa and glvnd | 15:37 |
uvos | right | 15:38 |
parazyd | That's not yet in buster | 15:38 |
uvos | but glvnd is useless | 15:38 |
uvos | because we dont have pvr or gl4es as glvnd backend | 15:38 |
kshatriya | would it make sense to ask the SDL guys for an environment variable to switch preference of backend GL/GL-es ? | 15:38 |
parazyd | Yeah we need a mechanism for this | 15:38 |
uvos | sdl isent really involved in that | 15:39 |
parazyd | Gentoo has eselect with is something like what we want | 15:39 |
uvos | its the app that asks sdl for gl or gles | 15:39 |
parazyd | (Basically a symlink manager) | 15:39 |
kshatriya | thx uvos i was thinking wrong | 15:40 |
lel | clort81 opened an issue: https://github.com/maemo-leste/bugtracker/issues/562 (Droid4 5.11 Fullscreen rendering errors working/nonworking tests.) | 16:34 |
roalyr | How to disable the desktop manager and stay in console? | 18:43 |
uvos | choose the other option in the bootloader | 18:44 |
roalyr | That one just disables the xserver? | 18:46 |
uvos | on bionic yes, on d4 it boots single user mode | 18:46 |
uvos | but you can switch runlevel from there | 18:47 |
roalyr | Thanks. | 18:52 |
roalyr | How many cycles does it take to calibrate the battery? | 18:53 |
kshatriya | this TI driver isn't even reporting GL_VERSION . This breaks mpv's opengl-es detection. | 18:58 |
kshatriya | there is some talk about newer driver for droid4 maybe? with fixes? | 19:00 |
kshatriya | i do get EGL_VERSION = 1.4 build 1.9@2253347 | 19:31 |
kshatriya | /mesa-demos-8.4.0/src/egl/opengl has some tests | 19:32 |
Wizzup | uvos: hm I also have a win7 img around here, well, going to make a xp install anyway | 23:44 |
Wizzup | uvos: internet mentions this https://gsmserver.com/sigmakey/ - looks interesting | 23:59 |
Wizzup | (wrt droid 3 unlock) | 23:59 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!