**** BEGIN LOGGING AT Mon Mar 12 03:00:01 2018 Mar 12 07:31:54 Been gone a while due to stuff and life, have I missed anything exciting (: Mar 12 08:31:59 Morning! Mar 12 08:32:46 thejsa: Those things happen. Tofe's been looking to integrate Halium a bit more. Not sure to what extend that would have a positive effect on the issues with Athene and the image Mar 12 09:01:58 By the way, if anyone has an idea on how to convert a sparse android image to a minimal-size ext4 image, I'm all ears :) Mar 12 09:02:40 The version with sudo is here https://bpaste.net/show/334ac8f6d233 Mar 12 09:04:22 Tofe: I guess JaMa might have some clues for that when he's around Mar 12 09:06:06 Tofe: Could this be of any help? Seems the AsteroidOS faced a similarish issue: https://github.com/AsteroidOS/asteroid/issues/31 Mar 12 09:09:59 Maybe android-tools-native can be of some help. I'll have a look. Mar 12 09:39:04 Tofe: It seems it should: https://patchwork.openembedded.org/patch/140076/ Mar 12 09:39:39 uh Mar 12 09:39:52 it's super compleicateed way (re: changing sparse image to ext4 image) Mar 12 09:40:51 Tofe: https://github.com/Halium/halium-scripts/blob/master/replace-android-image#L34 Mar 12 10:22:32 bshah: copied directly from halium-install script ;) Mar 12 10:22:52 halium-install script needs copy of replace-android-image Mar 12 10:22:53 :P Mar 12 10:23:05 But thanks for the other method, e2resize is exactly what I need Mar 12 10:23:54 advantage of resize method is, you don't need to worry about varying sizes of system.img per target.. so you can have 100 MB system.img or 500 MB system.img Mar 12 10:24:07 you won't need to manually play with values Mar 12 10:28:01 bshah: it would be good to find a way to avoid getting 32bit binaries in 64bit halium images Mar 12 10:28:19 why is it issue? Mar 12 10:28:23 size, mainly Mar 12 10:28:29 mido is 450MB Mar 12 10:28:40 instead of 260MB really useful Mar 12 10:29:04 well.. you will have issues if "something" like e.g camera blob doesn't really work for 64 bits and then you would want 32 bits Mar 12 10:29:12 so keep it as it is :) Mar 12 10:30:29 Tofe: I agree with bshah there. Not everything exists in 64 bits and we need fallback on 32 in some cases Mar 12 10:31:35 ok Mar 12 11:33:12 bshah: You around? Mar 12 11:33:42 yep Mar 12 11:34:57 I noticed that the repo manifests in android and halium-devices are all over the place in terms of tab vs space... Any preference? Mar 12 11:35:04 Will clean it up a bit then :) Mar 12 12:28:39 Herrie|Laptop: adding that in android-system-image-mido should help https://bpaste.net/show/a5738741921c Mar 12 12:28:49 But it's not at all tested! Mar 12 12:29:13 hmm why is this mido specific? Mar 12 12:46:48 bshah: system.img for mako,hammerhead and tenderloin were not sparse (but maybe that's related to 5.1 vs 7.1) Mar 12 12:51:01 But maybe that stuff would make more sense in our jenkins job we use to build system.img Mar 12 12:53:57 Herrie|Laptop: ^ you have an opinion on this? Mar 12 13:03:04 Tofe: Not sure really... Seems sparse is something qcom specific to a certain extend from what I could see Mar 12 13:03:28 Seeing N4 and N5 are Nexus therefore AOSP they might not be sparse... Mar 12 13:03:41 Tenderloin always has been a bit of a different duck Mar 12 13:03:51 Tofe: Let me test your fix in a bit Mar 12 13:16:00 Tofe: Well it builds :P Mar 12 13:16:08 Anything I can check before flashing? Mar 12 13:29:54 Herrie|Laptop_: not really Mar 12 13:30:46 OK I'll just flash then Mar 12 13:31:03 bshah: Any opinion on the space vs tab on the android and halium-devices repo? Mar 12 13:31:16 I'm preparing some PR's so can just as well do some cleanup too Mar 12 13:31:27 I prefer 4-spaces everywhere :P Mar 12 13:35:48 bshah: Even where there are 2 currently? Mar 12 13:36:07 Herrie|Laptop_: tbh, do whatever is consistent and widely used Mar 12 13:36:24 Well it's all over the place currently :P Mar 12 13:36:36 I am in-different about what is used in code .. ( 4-spaces is more like personal preference and not halium standard) Mar 12 13:37:50 bshah: http://tinypic.com/r/2r219qs/9 Mar 12 13:37:53 To give you an idea :P Mar 12 13:40:48 Herrie|Laptop_: also keep the vertical alignment for example for the "remote" parameters Mar 12 13:41:01 But yes, please, spaces :) Mar 12 13:55:05 bshah: 7.1 one: https://github.com/Halium/android/pull/18 Mar 12 13:55:07 Will do 5.1 in a bit Mar 12 14:21:23 Tofe: Mido boots again ;) Mar 12 14:43:46 bshah: Seems that piggz disabled some services on Android side for mido due to conflicts with Mer, do you know if all these should be disabled for Halium too? I suspect some might cause the issues I'm having on my mido with things not working ;) Mar 12 14:44:02 I guess I could revert back to the LineageOS device repo to try :P Mar 12 14:44:17 which are not enabled in mer but enabled in halium? Mar 12 14:48:15 bshah: Yeah Mar 12 14:48:26 I know you said I could kill pm-service & qseecom stuff Mar 12 14:48:59 Though pm-service seems to be related to wakelocks and might give me other issues Mar 12 14:50:06 But at least for now I don't have logspam ;) Mar 12 15:14:53 Ah, good ! Mar 12 15:15:17 Still have other issues, but those are due to my device repo I reckon Mar 12 15:15:32 I might start from LineageOS instead of Piggz to avoid some of those Mar 12 15:36:18 I'll start a build with that, just to exclude piggz changes being cause of some of the issues I see Mar 12 16:09:49 Tofe: Herrie|Laptop_ is this ready for merge? https://github.com/shr-distribution/meta-smartphone/pull/63 Mar 12 16:21:53 JaMa: Well we were fighting a bit with sparse vs ext4 and looking where to best solve that Mar 12 16:21:57 Otherwise it seems to work fine Mar 12 16:22:05 Unless you have some remarks/suggestions? Mar 12 16:25:36 I didn't really look into it :/ just had some time to do the merge before disappearing again Mar 12 16:41:13 JaMa: OK Mar 12 16:41:20 Well no rush in general Mar 12 16:41:45 When you have some time to review that would be great since it's a change with quite some impact in general Mar 12 16:44:28 ok Mar 12 17:57:29 Herrie: https://github.com/Tofee/meta-smartphone-1/commit/087763dc7e4e522741dd68fb96ffa8e5056c2cd6 Mar 12 19:09:37 Tofe: That looks like a more generic solution, similar to Halium, I like it :) Mar 12 19:09:52 Once JaMa reviewed we're good to go I'd say :) Mar 12 19:10:47 My new build with other device tree seems to behave similar, but I applied LineageOS cleanly with just the jar/apk grep changes. So that saves patchingg Mar 12 19:11:42 I don't have logspam even though I didn't disable the pm-service it seems Mar 12 19:11:59 Need to dig a bit further when kids are asleep Mar 12 19:12:06 Might need to fix fstab still Mar 12 19:12:11 To remove the context Mar 12 19:12:36 Need to double check Halium device repos. Can probably find a nice generic grep solution with regexp too. Mar 12 19:13:39 Also cleaned up the Halium device-manifests for 7.1 and will PR those. Then will add our devices (well Tenderloin, Mido, Athene and Onyx). Just Athene and Onyx need to test new Halium build still Mar 12 19:14:02 Then will have a look at a rework of our Android repo to remove duplication Mar 12 19:14:17 And see if I can contribute some porting guide bits into Halium. Mar 12 19:14:44 Tofe: PS still stuck with the libglibc from Mer Mar 12 19:15:10 I either get the .h files in the image but it can never seem to find the functions even though .so are in image Mar 12 19:15:22 So I'm probably missing something trivial Mar 12 19:30:36 Great :) Mar 12 19:30:59 the libglibc ? what's this ? Mar 12 19:31:33 Tofe: Some library from Mer used for tests in newer oFono which I suspect I need for mido Mar 12 19:33:37 Seems I might need to add FILES_${PN}-libs ? Mar 12 19:34:04 Tofe: This one: https://github.com/webOS-ports/meta-webos-ports/commit/3e3b3cca2c27d1ab2ba28ab57bf48169a2a79f03 Mar 12 19:34:35 ah yes I remember Mar 12 19:35:36 the headers are there in the -dev package ? Mar 12 19:41:48 Tofe: WEll I've been playing around a bit... My current recipe looks like https://bpaste.net/show/c2d1fdf89c0f Mar 12 19:43:39 I guess I should try just with just FILES bits Mar 12 20:32:20 Tofe: GEtting https://bpaste.net/show/1efc865c6ecc Mar 12 20:32:38 I can do INSANE_SKIP_${PN} += "arch" but that seems incorreect to me? Mar 12 20:40:55 This is what I have now as recipe: https://bpaste.net/show/bade0de703ec Mar 12 20:50:16 Herrie: your binaries are build using host's arch, not the target one Mar 12 20:51:41 Herrie: I'll give it a try on my side, taking your recipe as a starting point Mar 12 20:58:30 Tofe: Yeah that's what I figured Mar 12 20:58:35 This should be done by ldflags Mar 12 20:59:17 YOu can put this in your ofono_%.bbappend: https://bpaste.net/show/68ca0d8d06b7 Mar 12 21:27:10 Herrie: you'll need to patch the Makefile file I think Mar 12 21:27:33 In order to get "CC ?= $(CROSS_COMPILE)gcc" instead of "CC = $(CROSS_COMPILE)gcc" Mar 12 21:27:50 Then in do_compile, just run "oe_runmake" Mar 12 21:29:49 You can also drop the FILES_${PN}, OpenEmbedded manages just fine it seems Mar 12 21:30:29 So in the end, it's https://bpaste.net/show/ce5bb3ddb5b9 + the patch for Makefile Mar 12 21:37:26 Tofe: Thnx iwll check in the morning! Mar 12 21:37:51 I didn't check ofono, I'll go to bed too :p Mar 12 21:38:01 Tofe: Well if ofono builds we're good Mar 12 21:39:28 yup; we'll see. At least on my side the libglibc packages look good Mar 12 21:40:01 OK great Mar 12 21:40:05 More to test tomorrow Mar 12 21:42:55 gn8 **** ENDING LOGGING AT Tue Mar 13 03:00:03 2018