**** BEGIN LOGGING AT Sat May 12 03:00:04 2018 May 12 05:26:56 sheepbat, Magisk works fine for me May 12 08:00:29 MonkeyofDoom: you can also set the brightness with repowerd (see info added to DebianTP wiki yesterday) May 12 08:03:59 depscribe[m]: yes we need some tweaks to repowerd to pickup the light sensor, I just commented that stuff out for now to get more important things working May 12 08:12:50 updated Android images from Planet along with a tool to pick any size of A/D size split and to pick whether A or D are the default boot (the tool just swaps the boot/linux boot image names - the types are unchanged so obviously not a problem) May 12 08:13:21 see: http://support.planetcom.co.uk/index.php/Gemini_Firmware May 12 08:14:42 actually that stuff I removed might have been more about detecting a face against the device to turn the screen off when your in a call, which we'll never want, don't remember if it did adaptive brightness as well May 12 09:35:01 > "Khalid Kamal Naji Only God can help you. We reached the point of no interest in the product due to sustainable delays. The company needs more focus and cut using backers money to attend conferences and marketing. They need to focus more on operations and quality. Seriously they have an issue." haaaaaahahahahahha May 12 09:35:02 Salty... May 12 09:41:45 > oh please IphoneX the most innovative piece of art of all times lmfao May 12 09:41:46 So innovative that it's killing itself. May 12 09:41:46 https://www.cnbc.com/2018/04/20/apple-iphone-x-discontinued-this-year-analyst-says.html May 12 09:45:34 > but crap. People are flat out dumb when trying to be smart May 12 09:45:35 That because smart people don't act that dumb. I am sure I can tweak exif information as well as location data if needed. Tools are all out there and it is smart to plan ahead ;) May 12 10:06:24 wondering what "sustainable delays" are? May 12 10:08:52 Just flashed to get me a dual boot system! Yey! May 12 10:09:06 Congrats May 12 10:09:22 First job: delete the google crap May 12 10:09:32 jeffers_[m]: I just updated the wiki with some tweaks and re-orderings based on the new updates from planet, so if you happend to have that open do a refresh May 12 10:09:59 that is for the Debian side May 12 10:10:21 adam_b: wait, where? May 12 10:10:55 the new release from planet just under the image files is the link to the DebianTP page which is a wiki that any of us can update, but so far mostly me May 12 10:11:10 ah ok May 12 10:12:17 adam_b quick question though: I have no idea how to delete chrome and the rest for example. Do I need magisk? I'm assuming the dual boot gives me a rooted android install May 12 10:12:49 jeffers_[m]: I've not done much on the android side (never even rooted it), others here will be able to help though May 12 10:12:51 adam_b: yes, I used the latest one May 12 10:13:24 the dual boot won't give you a rooted install by default no May 12 10:13:34 adam_b: smashed the screen on my nexus 4 this morning, so gives me the impetus to get the gemini set up asap May 12 10:13:36 so if you made your own scatter file for the dual boot May 12 10:13:39 oh May 12 10:13:44 then you just need to change the android kernel May 12 10:13:58 oh the img file of course May 12 10:14:22 yea depending on whether you picked primary boot android or linux you need to swap a different file May 12 10:14:51 let me boot back into windows and finish the job then May 12 10:14:54 you don't need to reflash the whole thing either, just pick download and then update a single file (using tick boxes) May 12 10:14:55 cheers! May 12 10:15:08 yes, quick that way May 12 10:15:30 though if you've not changed any data on the device its not so much of a big deal May 12 10:15:52 compare the android and rooted android scatter files for the change to apply to your dual boot one May 12 10:26:37 adam_b: looks like the geminiplanet blog page is correct. I need the boot-verfied file May 12 10:27:27 patched_boot? May 12 10:27:58 patched one comes with the dual boot scatter May 12 10:28:09 my diff Gemini_Android.txt Gemini_Android_Rooted.txt shows < file_name: boot.img > file_name: patched_boot.img May 12 10:28:23 verified is the one in the rooted android May 12 10:28:41 the rooted android is included in the dual boot download May 12 10:28:52 with a different name so you have the choice May 12 10:29:18 if you've also downloaded a blob containing just the rooted android then yes it probably has just one name May 12 10:29:20 I may have it the wrong way round! May 12 10:29:39 you have downloaded this today? May 12 10:29:48 or yesterday evening May 12 10:30:04 geminiplanet blog is probably out of date May 12 10:30:11 last hour! May 12 10:30:37 I'll go with the patched from the rooted one May 12 10:30:38 cool, so in that one you have the option of flashing just android/rootedandroid/dualboot May 12 10:31:47 I'll suggest another tick box on planets partitiontool for whether to have rooted android or not May 12 10:32:49 it gives me May 12 10:32:50 PMT changed for the ROM, it must be downloaded. Please select firmware upgrade scene and try again! May 12 10:33:43 this is with the ticking only one item and selecting download from the dropdown? May 12 10:34:11 its in the new documentation as supposed to work, but I've not tested it and it might be a linux flash tool only thing May 12 10:34:13 yes, just boot selected, download only May 12 10:34:33 if that fails you'll need to do the usual firmware update whole thing :( May 12 10:35:03 I don't feel like mucking it up! May 12 10:35:16 then just do what worked last time May 12 10:36:10 If I just put rooted android on, I won't have the linux side May 12 10:36:50 no the idea is that you edit the Gemini_Dual_Boot.txt file to apply the change May 12 10:36:58 then flash with that May 12 10:38:53 the android only txt files have an entirely different partition structure so that makes sense it would'nt let you flash a single part of that May 12 10:39:09 ah, now I follow you May 12 10:40:08 so the dual boot file needs to say, hey, "flash this rooted one if you don't mind" May 12 10:40:50 yea should be just a matter of a find and replace for boot.img -> patched_boot.img in that dual boot file May 12 10:42:29 ok that's done May 12 10:42:41 it might let you flash just one bit now May 12 10:42:47 as the partition table will match May 12 10:43:23 showing patched_boot in the tool May 12 10:43:24 try just the download only with the boot file alone? May 12 10:43:47 yep, it'll either work or give an error like last time May 12 10:44:08 same as last time May 12 10:44:19 so its a windows/linux flash tool thing then May 12 10:44:35 or its just nonsense that someone thinks should work but doesnt May 12 10:46:25 well, never mind! May 12 10:46:26 I can set it up when that's sorted May 12 10:46:41 cheers for the help though May 12 10:46:41 I want to help with debian when I can May 12 10:47:15 from the debian side you can easily swap whats on the boot partitions with a 'dd' May 12 10:48:12 so your doing a full firmware upgrade I guess May 12 10:48:15 oh yeah! so convenient May 12 10:48:43 well saves going back to the flash tool, but can also get you stuck May 12 10:48:43 yeah, it's booting into android now. next put the magic thing on May 12 10:49:06 I despise android, I was on ubuntu touch for a while May 12 10:49:35 yea I'm still UT on my N4 May 12 10:50:45 I heard the purism phone will have it May 12 10:53:36 People keep talking about purism, I guess its going in the many options like the Gemini May 12 10:55:01 Well, I figure enough people have had enough of shoddy privacy invasions that there's a market now May 12 10:55:41 just the last person to point me in that direction was I'm sure a gnome mobile type direction for it May 12 10:56:05 rather than ubports May 12 10:56:08 well, there is that. But the KDE one looks incredible May 12 10:56:36 as in plasma mobile? May 12 10:56:57 don't forget magisk May 12 10:56:59 yes that one May 12 10:57:10 sorry, wrongwindow May 12 10:57:10 downloading that now May 12 10:57:33 sounds like Zevv is helping someone else out on almost the same path May 12 10:58:51 same tasks! May 12 11:04:31 This is why I like lineage, no google crap by default May 12 11:06:13 well that is also an option May 12 11:06:58 (bottom of the firmware page - 'unoffical' yet on the 'official' firmware page...) May 12 11:07:26 I didn't know it was on the official page May 12 11:23:35 the next job is figure out how to handle language switching May 12 11:50:04 jeffers_: lineage doesn't have phone/messaging working yet. May 12 11:50:46 depscribe: yes, I saw that May 12 12:03:28 I can't quite figure out how to switch between two languages. I'm sure I'm missing something. May 12 12:31:36 hmm.. I did the swap to the rooted boot-verified.img May 12 12:31:52 I'm reasonably sure I did the installation right May 12 12:31:58 but magisk doesn't seem to be working May 12 12:32:01 anyone have any ideas? May 12 12:33:27 with the dual boot scatter? May 12 12:35:10 oh, just doing android for now May 12 12:37:17 you installed the magisk May 12 12:37:34 did you open magisk and do the installation inside? May 12 12:38:17 i opened magisk, but it just hangs at the start screen May 12 12:39:02 oh! May 12 12:41:23 I would try a reboot then and try again May 12 12:45:50 hmm.. looks like it worked May 12 12:45:52 thanks May 12 12:49:38 are the changes being committed by the developers who hang out here being incorporated in the debian tree, or is there a line i should add to sources.list? May 12 12:49:50 happy to help! May 12 13:03:59 do we know if the default android installation does storage encryption? May 12 13:04:21 I'm sure it does but haven't got that far yet May 12 13:04:36 I'm trying and failing to get my pebble to connect at the minute May 12 13:08:32 ooh May 12 13:08:37 which pebble? May 12 13:08:53 2 May 12 13:10:18 SE or HR? May 12 13:10:47 HR May 12 13:10:59 great watch May 12 13:11:07 I'm using one myself May 12 13:11:24 Indeed great May 12 13:11:36 Something's off with the pairing May 12 13:11:42 I'm not intending to use the gemini as my main phone, so it's paired with my oneplus May 12 13:12:33 but a smartwatch seems a necessity with this device May 12 13:12:40 That sounds good May 12 13:12:50 Yeah, that's why I got this pebble May 12 13:12:53 I suppose I could always wear two watches... May 12 13:13:08 you just bought it? May 12 13:14:20 about 2 months ago, not new but barely used May 12 13:15:29 it's so useful when I'm on my bicycle May 12 13:18:53 finally got it set up!! May 12 17:43:50 I am trying to build the official kernel and it does not seem to be including the mt6797 path includes... anyone know what I am missing? May 12 17:54:46 I was thinking perhaps that the paths were just too long but that isnt it May 12 18:26:38 Beiriannydd: intersting question, I'll give it a go. Which compiler are you using, and where did you find the proper .config? May 12 18:27:51 ndk 16b and the only one in arch/arm64/config May 12 18:29:32 hm funny start, the makefile sais 3.18.41, but it's not "Diseased Newt" May 12 18:29:39 so it's some kind of hybrid May 12 18:30:22 yeah I figured that when trying to find where to patch it into the android repo May 12 18:34:56 aeon6797_6m_n_defconfig, right? May 12 18:35:02 yes May 12 18:35:57 which kernel tree? https://github.com/NotKit/kernel-3.18-geminipda? May 12 18:36:13 yes May 12 18:36:37 you have to build out-of-tree (mtk makefiles are broken) and for new compilers may need to patch makefiles that set EXTRA_CFLAGS May 12 18:37:05 the latter being https://github.com/mdgem/kernel-3.18-geminipda/commit/747d63fde06c15fe4ae0a03d0b0dc09315cd29ec May 12 18:38:19 What's the relation between the Notkit and dguidipc trees May 12 18:38:22 they're not forks May 12 18:39:04 I'm not sure May 12 18:49:29 Zevv: I think the dguidipc is just out of date or rather tracks the kernel in the current download, not the latest changes being worked on May 12 18:51:21 it seems yes May 12 18:53:04 I think once someone other that TheKit has something to push it'll get branched into the gemian repo with everything else May 12 19:05:50 Beiriannydd: mt_gpufreq.h: No such file or directory May 12 19:05:51 ? May 12 19:06:04 yeah May 12 19:06:21 adam_b: you seem to know the trick to this? May 12 19:06:39 drivers/misc/mediatek/base//power/mt6797/Makefile does not seem to add the proper includes to fin the mt_gpufreq.h May 12 19:06:44 that is the -I not containing MTK_PLATFORM May 12 19:06:49 exactly May 12 19:10:00 maybe -I. May 12 19:10:22 +ccflags-y += -I$(srctree)/drivers/misc/mediatek/base/power/$(MTK_PLATFORM) May 12 19:10:32 but after that it breaks in other places May 12 19:11:13 adam_b probably knows what's up, he noted you need an out-of-tree build May 12 19:12:33 is that just O=? May 12 19:12:41 yeah May 12 19:15:47 Zevv: sorry not me, I've been tempted to try kernel tinkering but so far always found it easier to bug someone else May 12 19:16:02 hehe May 12 19:16:21 At this time my build breaks with missing includes. I fix them, and run into the next May 12 19:16:45 they're all easily fixed, but there's probably a trick to the build May 12 19:16:57 there was definately chat about using an old compiler making it easier May 12 19:17:45 see the scrollback here though I think it was with MonkeyofDoom so maybe your already working from that tree May 12 19:18:42 gcc 7.3.x seems to get upset by -WError (which I patched around but and didn't produce get a bootable kernel out of), but gcc 6.4.x worked and the android-4.9 toolchain worked May 12 19:19:07 an elementary question: there being no version of an application i want that's made for debian-arm64, how difficult would it be to port it over? here's the application: https://github.com/protonmail-desktop/application May 12 19:19:54 i'm hoping it's as simple as specifying the architecture in the configuration file, but fear it is a lot more complicated than that. May 12 19:19:57 depscribe[m]: chromium itself is fine on arm64, so I'd expect there'd be minimal trouble May 12 19:20:17 since chromium is most of electron and electron is most of protonmail-desktop May 12 19:20:49 may take some compiling but shouldn't need much/anything in the way of actual code written May 12 19:21:24 indeed, if you can build on debian on intel, you should probably be just fine on arm64 May 12 19:21:27 "but and didn't produce" wow my grammar has been better May 12 19:21:51 Yeahh, you should watch your capitals and punctuation as well :) May 12 19:24:13 Beiriannydd: well, I have no clue at this time. I can fix tens of -I pathss, but that's probably not how it works May 12 19:24:23 the kernel is just very mediateky May 12 19:24:24 I am on it May 12 19:24:28 Zevv: did you apply the patch I linked above? May 12 19:25:47 kernel builds have changed since the last time I did this... May 12 19:26:31 MonkeyofDoom: I did, but that does not seem to be the issue May 12 19:26:34 hmm May 12 19:26:50 I am going to come out of this with a dockerfile for building kernels and instructions May 12 19:26:54 and you are building out-of-tree? O=wherever May 12 19:27:16 that is the piece I hadn't done before May 12 19:29:48 it will be a while before I hit breaky bits May 12 19:33:46 I'm in the middle of copying files to my fast machine for builds and then I'll try a few more and write things down ;) May 12 19:36:02 I filed an issue on the NotKit github kernel repo asking for build instructions May 12 19:37:21 I basically want to build nfs support and full cgroup support for docker in for my linux chroot May 12 19:47:48 NotKit kernel Out of Tree build is looking good so far May 12 19:49:26 it is in base/power right now May 12 19:50:49 got through that, I think that means this will work May 12 19:52:26 funny, what's your exact build cmdline? May 12 19:55:38 step1 export PLATFORM=aeon6797_6m_n; export MTK_PLATFORM=mt6797; make O=/usr/src/build V=1 May 12 19:56:21 after doing make O=/usr/srcc May 12 19:56:26 graaarh May 12 19:56:57 after doing make O=/usr/src/build ${PLATFORM}_defconfig May 12 19:57:26 no ARCH=arm64 on the cmdline then May 12 19:57:44 already have that exported May 12 19:57:49 ok May 12 19:58:05 and the cross_compile May 12 19:58:17 darn error May 12 19:58:39 oh the ucCenterChannel one May 12 19:59:06 I fixed that by making a sane default in the other tree May 12 19:59:26 anyone got a recommendation for a good windows irc client? May 12 19:59:59 ssh to a linux machine :) May 12 20:00:14 lol ok for linux then May 12 20:00:34 since my windows is also a linux machine May 12 20:00:41 wsl for the win May 12 20:01:15 Slack has made my irc fu weak May 12 20:08:15 I'm testing build instructions based on https://github.com/bq/aquaris-M10-FHD May 12 20:08:27 if it works, will publish in wiki May 12 20:08:38 http://paste.ubuntu.com/p/GsJTSZwtrn/ is what is running here May 12 20:08:55 ok fixed rlm_domain.c May 12 20:11:25 I have a vmlinux May 12 20:13:22 my gcc is 4.9 btw May 12 20:13:46 4.9.x May 12 20:14:02 my only source change needed was UINT_8 ucCenterChannel = 0; May 12 20:14:13 ooh I would NOT do that May 12 20:14:30 :) May 12 20:14:37 Just to get it through the build May 12 20:14:52 looking at what the code does elsewhere I made it equal ucPrimaryChannel May 12 20:15:15 that seems like a sensible default May 12 20:15:41 sorry PriChannel May 12 20:15:42 It probably gets set somewhere else to a sane value May 12 20:15:48 suuuuure May 12 20:16:04 but still, I'm positively surprised to get this to build May 12 20:16:07 except there is a reason for the may be used uninitialized May 12 20:16:28 mali midgard is building right now May 12 20:17:03 well, I'm out for the night May 12 20:17:04 have fun! May 12 20:17:30 Goodnight from PDT May 12 20:21:57 I suspect that all which is required for an intree build is -I. used judiciously May 12 20:41:10 OK My opinion on irc client is probably weechat instead of irssi just because PERL in irssi May 12 20:50:00 yeah, this will work. May 12 20:50:52 OK, so now that building a kernel is complete, has anyone replaced their android kernel and modules yet? May 12 20:51:25 https://github.com/gemian/gemini-keyboard-apps/wiki/KernelCompilation May 12 20:51:38 but is for Debian May 12 20:51:48 *this is May 12 20:52:02 Ubuntu (WSL) isn't too different May 12 20:52:14 I meant for Linux kernel on Gemini PDA May 12 20:52:18 ohhhh May 12 20:52:43 https://github.com/vianney/bootimgtool is known to work for repacking boot.img on Gemini May 12 20:53:03 gtk May 12 20:53:59 looks good TheKit, very useful May 12 20:56:28 Beiriannydd[w]: I've been replacing the kernel with new ones built by TheKit with the 'dd' method, which works fine for the 'linux_boot' partition but won't work with the new linux as primary boot option as the 'boot' partition is protected May 12 20:57:13 It would probably be useful to be able to configure alternate boot images May 12 20:57:17 'recovery' mode doesn't have the protection for 'boot' partition though May 12 20:58:28 there is some chat about implementing this https://github.com/gemian/gemini-keyboard-apps/wiki/Bootloader#four-combinations-only-three-in-use May 12 20:58:49 I've noticed that MediaTek lk has support for kind of env variables: https://github.com/dguidipc/gemini-lk/blob/master/lk/platform/mt6797/include/platform/env.h May 12 20:58:55 if thats what you mean by 'alternate boot images' May 12 20:59:05 maybe that could be used as config storage by us to change some options from Linux May 12 21:00:38 that would be persistent over a reboot/power off? that would be cool to have an option to say next reboot linux should be Sailfish or Debian etc May 12 21:02:34 I think it uses other mechanism for "reboot recovery/bootloader" in Android May 12 21:16:16 Fixed a mistake with mkbootimg command in wiki page. Interestingly, without dtb appended the kernel is still able to save log to ram console and panic due to failing to find ramdisk May 12 23:15:59 so when trying to build dtbs with O=, I can see it is including ./include (the object directory) and not $srctree/include so it's missing the mt6797-clk.h file May 12 23:40:29 Beiriannydd[w]: yep May 12 23:46:09 ok, fixed May 13 00:03:20 changing the Makefile.lib caused me to have to do a full rebuild, but I think it will be fine. May 13 00:05:51 lots of linking May 13 00:07:30 I'll have a PR for you shortly, TheKit May 13 00:46:59 so I have xt_CAPITALIZED.o but not xt_lowercase.o and it's trying to link the lowercase versions of: dscp, hl, rateest, tcpmss... any ideas? May 13 00:58:50 OK I commented the duplicates out for now and that made it work. May 13 00:58:57 make complete. May 13 01:15:58 OK, so the mkbootimg in the wiki you posted TheKit, it appears to want to build with a dynamic and static link library May 13 01:18:36 taking off the extra -L. and -lmincrypt made it build May 13 01:23:21 and now I have a 9MB linux_boot.img May 13 02:31:50 i'm encountering the strangest bug on android and i have no idea where to report it May 13 02:32:08 it sends f11 every time i close the lid May 13 02:33:11 I am getting the same thing but it hadn't caused me any problems yet May 13 02:43:10 I wondered if F11 was being intercepted under Android but not under Linux... are you seeing the F11 in Android itself? May 13 02:49:08 sounds as if when the keyboard is squashed as the gadget is closed, it's squashing some keys a little too much -- might be a hardware issue. May 13 02:49:49 depscribe[m]: there's no F11 on the keyboard. I think it is more likely that it is an event caused by closing the lid. May 13 02:50:58 meanwhile, on mine under linux, if i'm logged in and leave it closed but running for several hours, it logs me out, which is worrisome. May 13 02:51:09 oh, interesting May 13 02:51:21 Beiriannydd[w]: is there no key combination that produces F11? May 13 02:51:48 If there's a way to get F keys out of the device, I'd love to kno May 13 02:52:47 i'd love to find a right-mouse-button emulation keystroke, too! May 13 02:53:43 well, key combination. May 13 02:54:01 you mean like alt-tap? May 13 02:54:29 (not that alt-tap works, as far as I know) May 13 02:54:29 that'd do. May 13 02:56:52 it doesn't. i just checked. May 13 02:57:23 I don't know if alt on its own sets the modifier map May 13 02:57:38 if it did, then you could do it May 13 02:58:35 i was testing to see how much standby time the gemini has under linux. began at 9 a.m. and checked every few hours -- put it in "standby," such as it is, by closing the lid -- and if i check again fairly soon i'm still logged in, but if i wait a few hours, i'm logged out. which would be bad for unsaved work. May 13 02:59:05 it sounds as though it isn't going to sleep May 13 02:59:33 when i get a little while tomorrow i'll look at dmesg and see if it casts any light on the subject. **** ENDING LOGGING AT Sun May 13 03:00:06 2018