**** BEGIN LOGGING AT Sat Apr 21 03:00:06 2018 Apr 21 05:41:27 I get an error when I try to build initramfs-tools-halium Apr 21 05:41:31 depmod: ERROR: Bad version passed touch-armhf Apr 21 05:41:52 http://paste.ubuntu.com/p/vYbxfrVTmX/ Apr 21 05:43:15 That's not fatal, and this error is normal because we don't have a kernel version specific initramfs (touch-armhf is not a kernel version) Apr 21 05:43:39 It should continue anyway Apr 21 05:44:27 oh yeah, there is smt in out/ Apr 21 06:36:58 the initrd that I've built now contains a shell "init" as expected, but I can't figure out which source file this i built from Apr 21 06:37:26 [Edit] the initrd that I've built now contains a shell script "init" as expected, but I can't figure out which source file this i built from Apr 21 06:37:40 [Edit] the initrd that I've built now contains a shell script "init" as expected, but I can't figure out which source file this is built from Apr 21 06:37:43 It's mostly the default debian initrd with a few additional files from initramfs-tools-halium. Apr 21 06:39:14 ah ok Apr 21 06:58:50 Fwd from Siber GΓΌvenlik: https://twitter.com/e2d3m/status/986242538104410112 … Eyvallah @e2d3m Dostum Apr 21 09:52:48 Hi Dalton, Florian, Marius and team. I wonder whether its be beneficial to ask XDA for official forums for Halium and Uborts for exposure and collaboration with all of those wonderful kernel / LOS minds, and kinda replicate how LOS does it with official and unofficial builds etc. Apologies if this has already been considered. Apr 21 10:33:43 @Garth, @UniversalSuperBox, @Flohack Apr 21 10:33:50 @Garth, I dont know if they are interested. They are all Java fanboys maybe πŸ˜† Apr 21 10:38:07 Am just thinking they have the kernel in common, consumer electronic devices, java via anbox, and maybe they will want to be able to say that we supported and collaborated with halium and ubports before it was mainstream and cool (of course it's already cool!) 😁 Apr 21 10:39:56 @Flohack, There are some get-a-"real"-linux-onto-my-device threads on xda Apr 21 10:42:01 I'm not really that clued up, but I have the idea that treble is going to make this alternative (non-spyung OS) far more accessible going forward. Apr 21 10:43:39 what's a treble? Apr 21 10:46:27 That's not really affecting us Apr 21 10:46:42 The Treble thing Apr 21 10:46:59 and if it is, it can only do good thing Apr 21 10:47:48 It's like halium but by google ;_; … Basically no need for vendor/device specific blobs and you can build and run generic roms across all devices Apr 21 10:47:49 Thank you for entertaining my thoughts; excitement re these projects got me hyped Apr 21 11:04:37 @alamo18, That's not true... Apr 21 11:05:23 It surely relies on blobs, it's just that they have been moved to a separate partition and rely on a stable API Apr 21 11:09:38 @ilyaishere, Yea i said device specific blobs Apr 21 11:09:44 Not blobs alone Apr 21 11:09:58 @alamo18, These are device specific Apr 21 11:10:11 Oo Apr 21 11:10:17 I see, thanks ;_; Apr 21 11:20:00 @UniversalSuperBox, I can't seem to get it to work. I put a line into init, but it doesn't seem to end up in the kernel I build Apr 21 11:20:06 https://paste.ubuntu.com/p/g3QZB8KZWB/ Apr 21 11:52:33 Does anyone know if a kernel config modified using the Ubuntu check-config script does still work with the reference rootfs and hybris-boot? Apr 21 12:04:09 i can still boot pm ok off of hybris-boot after fixing the check-config stuff Apr 21 12:04:49 i skipped like two configs bec otherwise the kernel wouldnt compile Apr 21 14:43:48 LOTNSAI was added by: LOTNSAI Apr 21 16:51:59 Can I bug someone to make sure this doesn't break hybris-boot in their tree? Apr 21 16:52:00 https://github.com/Halium/android_build/pull/14 Apr 21 17:42:33 Greetings. I compiled a kernel but some flags e.g. CONFIG_IKCONFIG_PROC are not set despite me setting them in the `cedric_defconfig` file (`cedric` is my device's codename). (I didn't find this out until I booted and read the diagnostic log over telnet...) Apr 21 17:42:34 Is there any way I can find out what file is overriding my configuration? Apr 21 17:44:48 on most devices the lineage_cedric_defconfig file is used. Does a file with this name exist? Apr 21 17:46:32 No, there is no such file, nor anything which would match these find patterns: `"*_cedric_defconfig"` `"*cedric*_defconfig*"` Apr 21 17:47:10 other than the cedric_defconfig file itself, that is Apr 21 17:50:57 * lambdaperl[m] sent a long message: lambdaperl[m]_2018-04-21_17:50:57.txt Apr 21 17:54:03 you can change the CONFIG_LOCALVERSION="-lambdaperl-001" or smt to make sure you are changing the right file Apr 21 17:54:15 that will show up in dmesg when the kernel starts Apr 21 17:55:02 @ubports_bot, You need to make sure your changes aren't overriden by a config option at the end of the defconfig. Apr 21 17:55:20 Each config option should only be in there once Apr 21 17:56:20 * lambdaperl[m] sent a long message: lambdaperl[m]_2018-04-21_17:56:19.txt Apr 21 17:56:31 also I have a config "output" in out/target/product/deb/obj/KERNEL_OBJ/.config Apr 21 17:56:44 (after running `mka systemimage` and letting it run through the config) Apr 21 17:57:04 * lambdaperl[m] sent a long message: lambdaperl[m]_2018-04-21_17:57:03.txt Apr 21 17:57:09 ok. what JBB said then Apr 21 17:59:29 * lambdaperl[m] sent a long message: lambdaperl[m]_2018-04-21_17:59:28.txt Apr 21 17:59:52 a very good questions Apr 21 17:59:54 [Edit] a very good question Apr 21 18:00:00 When I remove those lines for example, the warning about reassigning those symbols goes away... which makes it seem like a config option even though it looks like a comment Apr 21 18:00:01 I'm not sure! Apr 21 18:00:41 I had the feeling it kinda represent a turning off ... but I'd hope someone around here actually knows!? Apr 21 18:02:10 * lambdaperl[m] sent a long message: lambdaperl[m]_2018-04-21_18:02:09.txt Apr 21 18:02:12 CONFIG_MIND=blown Apr 21 18:02:23 sounds like it's not just a comment Apr 21 18:02:43 hehe Apr 21 18:03:52 well maybe that just might be what has had me on a wild goose chase for a while :-) Apr 21 18:09:40 I'll send a PR in later to add that to the docs, it might save someone a few hours :-) Apr 21 18:09:53 It looks so innocent with even a space after the hash Apr 21 19:52:56 Hi everyone, i'm trying to create a port for my phone a htc desire Apr 21 19:53:35 it's supposed that in lineage os github repo for my device i will find files named cm.dependencies or lineage.dependencies Apr 21 19:53:46 but i don't see these files in the device repo Apr 21 19:54:16 Guest43863: Hello there Apr 21 19:54:56 Can we have a link to the device repo? Apr 21 19:54:59 it's supposed by the porting doc, that in this repo https://github.com/LineageOS/android_device_htc_desirec the files should be present Apr 21 19:56:44 or i may have misunderstood and have to check in another repo Apr 21 19:59:28 looks like it should be the right repo to me (bear in mind I'm very much new to this) Apr 21 20:01:24 do you know if there's some kind of alternative to build the image without these files? Apr 21 20:04:11 I'm not very knowledgeable about this (maybe someone else could tell you) but you could try looking at a similar device that DOES have the __.dependencies file and trying to adapt it Apr 21 20:04:43 yeah trying that... thank you Apr 21 20:06:09 If you know the chipset (MSMxxxx probably looking at another Desire?), you should hopefully find a repo for that with not too much trouble Apr 21 20:06:15 I'm looking at https://github.com/LineageOS/android_device_htc_hiae/blob/lineage-15.1/lineage.dependencies for example (not sure how similar it is to the device you're after) Apr 21 20:06:38 Already I recognise qcom_common and the devicesettings stuff from my device which isn't even a HTC (but does have a Qualcomm MSM8937 chipset) Apr 21 20:07:04 oh great, i can check on that Apr 21 20:07:09 thanks Apr 21 20:08:23 And if you get really stuck, maybe just try and build it and see what errors fall out, that might (probably) tell you if you're missing something Apr 21 20:09:17 do you know what chipset/SoC your device uses? Apr 21 21:43:14 Hi everyone Apr 21 21:43:27 How do I change my kernel settings? Apr 21 21:44:06 (Photo, 1280x719) https://irc.ubports.com/WSkX0Ivo.png Apr 21 21:44:47 Just open the file you just checked using mer-verify-kernel-config with an editor and do the requested changes. Apr 21 21:44:59 I did that with nano Apr 21 21:45:08 All files are set to Y value Apr 21 21:45:18 Isn't it right? Apr 21 21:45:55 I mean, those settings that are required to be set to Y are already set Apr 21 21:46:03 According to the file Apr 21 21:46:33 Many are still completely missing in the file (aka unset) Apr 21 21:46:46 You need to add them at the end of the defconfig file Apr 21 21:47:07 like … ```CONFIG_NET_NS=y``` Apr 21 21:47:17 Ok, thanks Apr 21 21:47:30 But I found some of them there already Apr 21 21:47:38 I'll show you Apr 21 21:47:49 [Edit] I'll show you in a few minutes Apr 21 21:51:08 dohniks, [21.04.18 13:19] … [In reply to Dalton Durst] … I can't seem to get it to work. I put a line into init, but it doesn't seem to end up in the kernel I build … dohniks, [21.04.18 13:20] … https://paste.ubuntu.com/p/g3QZB8KZWB/ Apr 21 21:51:22 @JBBgameich any idea what I'm doing wrong? Apr 21 22:00:59 @Guest43863, is there a Lineage OS 14.1 port for it? Apr 21 22:01:28 I don't see it here https://wiki.lineageos.org/devices/ nor there https://forum.xda-developers.com/htc-desire-c/orig-development#romList Apr 21 22:05:19 Am I correct in thinking that I need to make my `userdata` partition not encrypted in order for hybris-boot/hybris-recovery to be happy with it? Apr 21 22:05:46 Yeah, you need to format it unencrypted before installing the rootfs. Apr 21 22:06:08 do I format as f2fs? Apr 21 22:06:32 ext4 should work Apr 21 22:07:25 Also, is it normal that when I plug my phone in to charge (or turn the phone 'off' when it's on charge), it starts the hybris-boot/recovery? Apr 21 22:08:56 I don't think hybris-boot distinguishes between charging and normal boot. Apr 21 22:09:49 ah, that might explain why I saw the charging animations in the boot logo partition a while ago Apr 21 22:10:04 sounds like the bootloader is in charge of that stuff Apr 21 22:10:16 or boot.img I should say Apr 21 22:10:46 later when your port works you can use halium-boot, which properly handles charging animations etc. Apr 21 22:11:05 ooo Apr 21 22:13:27 @JBBgameich, I think it fixed the error, thank you Apr 21 22:16:29 Anyone nows how to get rid of doze? Apr 21 22:17:07 I'm getting an error when building mkbootimg Apr 21 22:17:18 Can you paste them somewhere? Apr 21 22:17:20 would you happen to know how can I 'unmount'/get rid of `/dev/block/dm-0` so I can format the underlying block device? Apr 21 22:17:30 @JBBgameich, sure Apr 21 22:17:53 @lambdaperl[m], I think TWRP has a GUI for that Apr 21 22:18:35 aha Apr 21 22:18:39 Wipe > Format Data Apr 21 22:19:12 but failed with error 255 Apr 21 22:19:14 @JBBgameich, https://paste.ubuntu.com/p/G2VmTWtdM9/ Apr 21 22:19:26 which is the same error gotten when I try to do it manually Apr 21 22:22:20 @mayltonfnds, SettingsLib is java, you can remove the line that includes it in the device tree. (It's in some of the .mk files) Apr 21 22:23:33 let me try Apr 21 22:27:19 well I rebooted TWRP and formatted data again and it worked... Apr 21 22:30:04 I had to change the `halium-install` script to pack the images on the host, because my device just wasn't letting me mount loop devices as r/w (although it works under the kernel I compiled myself in hybris-boot) Apr 21 22:34:06 I already rewrote the halium-install script to do that btw (https://github.com/JBBgameich/halium-install) Apr 21 22:34:36 the wiki needs updating then or so ;-) Apr 21 22:34:37 did you have a similar issue to me? Apr 21 22:35:06 The busybox tar utility in the TWRP build of my device just didn't work. Apr 21 22:35:32 hahahaha Apr 21 22:35:51 @dohniks, FWIW, it works when I set BOARD_USE_LOCAL_INITRD:=true in halium-boot/Android.mk Apr 21 22:36:03 for me, all loop mounts are always `ro` no matter what I do.. it just silently mounts as ro Apr 21 22:40:22 @dohniks, So it worked? Apr 21 22:40:56 jup, my locally built initrd is on the device now Apr 21 22:41:14 it seems it didn't pick up the variable from the board config Apr 21 22:41:28 Strange that setting it in boardconfig didn't work since that gets sourced when you run breakfast or lunch Apr 21 22:41:38 ohhh Apr 21 22:41:46 do I need to rerun breakfast? Apr 21 22:41:46 "Sourced" Apr 21 22:41:50 Yes Apr 21 22:41:52 hehe Apr 21 22:41:56 welll Apr 21 22:42:09 I didnt know that Apr 21 22:43:19 oh deary me... the `loop` device mounting as `ro` always is back and now it's happening in hybris-recovery too Apr 21 22:44:06 Does the data partition need a fsck maybe? Apr 21 22:44:47 `/data` is fine, it's: Apr 21 22:44:48 `/dev/block/loop0 on /target type ext2 (ro,relatime,errors=continue,user_xattr)` Apr 21 22:45:36 Can you do dmesg so you can see what the mount system is doing to get that result Apr 21 22:45:38 * lambdaperl[m] sent a long message: lambdaperl[m]_2018-04-21_22:45:37.txt Apr 21 22:46:54 should I umount it, clear dmesg, then mount it again? Apr 21 22:47:14 No, it should complain about loop0 somewhere Apr 21 22:47:26 well, I did that, and... empty dmesg, `0` exit code return from `mount` Apr 21 22:47:50 this happened earlier in TWRP, but then was fine in hybris-boot (before I compiled a new kernel) Apr 21 22:49:13 Empty? Apr 21 22:49:20 Like, you have no kernel logs Apr 21 22:49:35 I cleared the dmesg buffer before mounting it again Apr 21 22:49:48 if you want I can reboot and give a full log Apr 21 22:51:14 (Photo, 1280x720) https://irc.ubports.com/TFGKZtE7.png Does it take much time? Apr 21 22:51:46 Unless hybris-boot and hybris-recovery are different enough, then I suspect this is a kernel issue.. Apr 21 22:52:33 @mayltonfnds, Building the kernel for the first time takes a few minutes depending on your hardware. Apr 21 22:56:03 It's finished Apr 21 22:58:10 What about it? Apr 21 22:58:12 https://paste.ubuntu.com/p/Q59vdzKqw9/ Apr 21 22:58:23 i was trying to build system image now Apr 21 23:03:49 I don't have libcurl_intermediates directory Apr 21 23:06:55 can you remove hs20-osu-client from Android.mk Apr 21 23:07:14 I can try Apr 21 23:07:41 Which Android.mk has it btw? Apr 21 23:07:46 main? Apr 21 23:08:20 dunno, I'm just picking out tips out of old irc logs :) Apr 21 23:09:00 Oh, I see Apr 21 23:09:02 maybe grep for it? Apr 21 23:09:03 `grep 'hs20-osu' $(find -iname '*.mk')` Apr 21 23:09:04 I think the same where halium docs say to remove the apks ... Apr 21 23:13:03 @dohniks, I'll try this first Apr 21 23:13:18 I guess it's similar to this http://docs.halium.org/en/latest/porting/common-system-build-errors.html?highlight=apk#common-system-build-errors Apr 21 23:13:32 maybe it's not actually an Android.mk .... Apr 21 23:14:02 grep -R hs20-osu vendor device Apr 21 23:15:25 Yeah Apr 21 23:15:31 I'm reading that Apr 21 23:15:57 it's really difficult to edit files on terminal, specially on google cloud Apr 21 23:16:15 I can't see the cursor, I don't know the line i'm at Apr 21 23:16:36 aha Apr 21 23:16:44 different editor? Apr 21 23:17:04 Well, I'm new at porting and developing stuff Apr 21 23:17:12 I don't know how to use vim Apr 21 23:17:21 I'm using nano Apr 21 23:17:35 download,edit,upload? Apr 21 23:18:00 that sounds dreadful Apr 21 23:18:01 try using `nano` if you want something nice and simple Apr 21 23:18:19 @lambdaperl[m], I'm doing it already Apr 21 23:18:41 what terminal are you using to access nano? It normally shows the position for me o_O Apr 21 23:26:24 I'm using google cloud Apr 21 23:26:31 that's the reason i suppose Apr 21 23:27:17 That's strange, there's no ask or jar file on Android.mk in my vendor folder Apr 21 23:27:28 [Edit] That's strange, there's no apk or jar file on Android.mk in my vendor folder Apr 21 23:27:50 I think I know why Apr 21 23:28:22 does it let you connect via SSH? Apr 21 23:28:49 @lambdaperl[m], Yes Apr 21 23:29:26 I'm using free credit Apr 21 23:29:33 what OS are you running on your own PC? Apr 21 23:29:33 for first year trial Apr 21 23:29:51 macOS Apr 21 23:30:32 ahh Apr 21 23:30:44 macOS might have ssh built in, maybe it's worth a try? Apr 21 23:35:35 Maybe later Apr 21 23:35:57 Well, I couldn't find any apk or jar in Android.mk files Apr 21 23:36:42 No, what I meant is: i think you face a similar problem as with those apk or jar files. Apr 21 23:36:55 Namely, too much is attempted to be build Apr 21 23:37:01 But you don't need all that Apr 21 23:37:07 halium needs less than android Apr 21 23:37:18 in your case the thing that is too much is not an apk or jar Apr 21 23:37:40 but something called hs20-osu Apr 21 23:37:43 I see Apr 21 23:37:57 in order to try to find it Apr 21 23:38:11 grep -R hs20-osu vendor device Apr 21 23:38:19 https://paste.ubuntu.com/p/zxJ9vCJGsT/ Apr 21 23:38:24 @dohniks Apr 21 23:39:02 two hits Apr 21 23:39:27 try removing the line from msm8996.mk Apr 21 23:39:39 ok Apr 21 23:40:03 make sure you completely remove that line, don't leave a blank line Apr 21 23:40:38 Maybe # is fine? Apr 21 23:40:42 no Apr 21 23:40:49 the line ends in \ Apr 21 23:41:09 which means it ... uhm ... how do I say this ... Apr 21 23:41:24 the line after directly continues Apr 21 23:41:36 I'm not expressing myself clearly Apr 21 23:41:58 the backslash means the line "continues" Apr 21 23:42:02 makes sense? Apr 21 23:43:04 I think so Apr 21 23:43:30 the line-ending is ignored so the line carries on to the next line ? ;-) Apr 21 23:43:47 thanks :) Apr 21 23:45:09 I've found out that my mounting loop files as `ro` problem actually doesn't depend on the kernel config changes I've made, but does vary depending on what filesystem the file-to-be-mounted is on (rootfs OK, f2fs NOT OK, ext4 to be tested) Apr 21 23:46:11 I deleted the line Apr 21 23:46:29 Let's try now Apr 21 23:47:10 Now it's building! Apr 21 23:47:22 Well, lets see if works or not πŸ˜‚πŸ˜‚πŸ˜‚ Apr 21 23:48:37 woot Apr 21 23:48:42 @lambdaperl what exactly are you trying to achieve Apr 21 23:48:59 which step are you at? Apr 21 23:49:23 @dohniks when hybris-boot starts up, it mounts the /data/rootfs.img to /target Apr 21 23:49:49 BUT the mount is `ro` (read-only) Apr 21 23:50:30 where do you see that? in telnet? Apr 21 23:50:36 yeah Apr 21 23:50:58 the same thing happened earlier in TWRP when building the rootfs.img, so I adapted the script to build it on the host PC instead.. Apr 21 23:51:19 OOO Apr 21 23:51:21 ah and your rootfs.img is f2fs?! Apr 21 23:51:38 it seems to mount `rw` as long as `/data` is ext4 Apr 21 23:52:24 my rootfs.img is ext2 Apr 21 23:52:30 but before my /data was formatted as f2fs Apr 21 23:53:15 ok. so it works now? Apr 21 23:53:27 because for some strange reason, the filesystem containing rootfs.img (or any loop mount file) silently affects whether it can be mounted read/write or just read-only Apr 21 23:53:46 yes, it looks like that is one problem out of the way for now Apr 21 23:54:09 well, if the image file is on a filesystem where the kernel doesn't have rw support ... Apr 21 23:54:13 not sure what to make of it. Kernel bug maybe? the kernel is 3.18.49 Apr 21 23:54:31 then I guess you can't really change the image either ... make sense Apr 21 23:54:47 the kernel DOES have `rw` support for f2fs Apr 21 23:54:57 ah Apr 21 23:55:28 well, maybe a bug then, but probably not important for now?! Apr 21 23:57:01 yeah, I'm just going to take what I can get for no Apr 21 23:57:04 now Apr 21 23:58:01 at least there's some comfort in that it does work with ext4 as the parent filesystem Apr 22 00:00:43 cheers for your help today and the other day @dohniks but I'm going to go to bed now :-) Apr 22 00:01:34 πŸ‘‹ Apr 22 00:01:52 o/ Apr 22 00:03:27 Dohniks, what am I supposed to do after system image build is finished? Apr 22 00:04:12 I know i have to work on rootfs Apr 22 00:06:15 I must backup my files in order to save them in case something goes wrong, right? Apr 22 00:06:34 You're working on your daily phone? Apr 22 00:07:14 Yes Apr 22 00:07:28 It's the only one I have πŸ˜‚πŸ˜‚ Apr 22 00:07:40 As long as you're okay with wiping everything so often Apr 22 00:07:50 Yes. Back up everything now. Apr 22 00:08:19 okay Apr 22 00:21:37 Yes Apr 22 00:21:55 Guess who was typing in the wrong group? Apr 22 00:22:14 Tofe, I found a little interesting blurb at the bottom of http://www.webos-ports.org/wiki/Halium_Considerations Apr 22 00:22:20 "We are integrating libhybris/libcompat in our middleware. This means a dedicated "flavor" of the Android build is needed, or a way to overlay a base system with specific bits." Apr 22 00:22:41 If you still require this, halium-boot has you covered with the overlay Apr 22 00:23:13 I say "the" overlay, but I mean the one from the system image. Files from there get mounted into the rootfs Apr 22 01:15:47 hΔ±. find d802 fixup-mountpoints Apr 22 01:15:57 [Edit] hΔ±. how find d802 fixup-mountpoints Apr 22 01:16:20 You make them Apr 22 01:19:25 ok. but macine make hard brick Apr 22 01:25:09 @ankaos, LG G2? I had one Apr 22 01:25:48 please help me Apr 22 01:26:15 sorry I cant Apr 22 01:26:37 I've never dev anything for it Apr 22 01:28:07 I need fixup-mountpoit Apr 22 01:28:24 Making that is described in the documentation Apr 22 01:28:33 Maybe you'd like to post what you have? Apr 22 01:30:08 Dalton, Do you know how can I download my files from google cloud? Apr 22 01:30:36 You have an SSH session? Apr 22 01:32:07 yes Apr 22 01:32:14 Use `scp` Apr 22 01:42:12 I found a way to upload my files to google drive from terminal Apr 22 01:42:16 that did the job Apr 22 01:42:20 thank you **** ENDING LOGGING AT Sun Apr 22 03:00:04 2018