**** BEGIN LOGGING AT Fri Jun 29 03:00:00 2018 Jun 29 07:39:43 Morning! Jun 29 08:20:15 morning Jun 29 08:20:56 Morning ! Jun 29 08:55:03 Herrie|Laptop: about the PR discusssion Jun 29 08:55:22 Do you see what I mean by the "hook" idea for the setup ? Jun 29 08:55:23 Tofe: Yeah Jun 29 08:55:30 Not really Jun 29 08:55:40 Ah didn't see your last comment Jun 29 08:55:41 Let me check Jun 29 08:56:50 Tofe: We wouldn't need luneos_targets.xml anymore if we use halium-devices/setup Jun 29 08:57:13 We'd just need to upstream our device manifests to halium-devices (all are already except for Rosy) Jun 29 08:57:44 See https://github.com/Halium/halium-devices/tree/halium-7.1/manifests and https://github.com/Halium/halium-devices/tree/halium-5.1/manifests Jun 29 08:57:50 Already did those ;) Jun 29 08:58:56 But still we need to remove hybris-boot and add the luneos_hal, isn't it? Jun 29 09:00:06 Tofe: luneos_hal we don't need Jun 29 09:00:15 Since setup script uses regexp magic to do that now Jun 29 09:00:36 ah, we only used our patches for that Jun 29 09:00:46 Wallpaper for 5.1 for example: https://github.com/Halium/halium-devices/blob/halium-5.1/setup#L72 Jun 29 09:01:35 And all the apk jar stuff in https://github.com/Halium/halium-devices/blob/halium-5.1/setup#L56 Jun 29 09:02:00 Why the [w][a][l][l][p][a][p][e][r] syntax ? Jun 29 09:02:09 and not just wallpaper ? Jun 29 09:03:49 Just habit I have when writing regexps :P Jun 29 09:03:53 wallpaper could work too Jun 29 09:03:59 In this case Jun 29 09:04:05 Just try to be consistent Jun 29 09:04:09 In some other cases it wouldn't Jun 29 09:04:58 so... we don't need luneos_hal, and hybris_boot maybe we could keep it actually (just ignore what it generates) Jun 29 09:05:45 There's also halium-boot now Jun 29 09:05:56 Tofe: Yeah in general we don't need luneos_hal anymore Jun 29 09:06:01 That was the whole point Jun 29 09:06:18 Is there one scenario where we still need it ? Jun 29 09:06:18 So when vendor binaries update we don't have to rely on the shady patching that's prone for errors Jun 29 09:06:34 Tofe: Well there are some 7.1 targets that might need more regexps Jun 29 09:06:42 I.e. this doze/cmaction/lineageaction bits Jun 29 09:06:46 That's what I'll upstream still Jun 29 09:06:51 So far that's only Onyx it seems Jun 29 09:07:15 For rosy I've patched the vendor repo, for the moment Jun 29 09:08:01 what about these two ? https://github.com/webOS-ports/android_luneos_hal/tree/halium-5.1/patches/hardware/libhardware Jun 29 09:14:42 Rosy vendor should be OK Jun 29 09:14:47 The setup script should take care of this Jun 29 09:14:54 I tested this with a dozen or so devices Jun 29 09:15:15 From different manufacturers in halium-devices Jun 29 09:17:39 Tofe: Well this is actually a UB Ports issue Jun 29 09:17:55 That's why we need the patches Jun 29 09:19:11 For this I would suggest to keep our own fork and do a remove of the halium version and add ours without the last 2 commits from https://github.com/Halium/android_hardware_libhardware/commits/ubp-5.1 Jun 29 09:19:19 That should only be for 5.1 and would only affect TP Jun 29 09:20:20 Since UB Ports assume all targets are AOSP I guess Jun 29 09:32:30 yes Jun 29 09:32:48 but I mean, this means we still need our local manifest Jun 29 09:32:57 however we put Jun 29 09:32:59 it Jun 29 09:33:57 Though I'd prefer not forking that particular repo: we might miss important fixes from Halium even on 7.1 Jun 29 09:33:59 Tofe: We can remove the repo that's added by https://github.com/Halium/android/blob/halium-5.1/default.xml#L140 Jun 29 09:34:14 And add our fork Jun 29 09:34:17 Tofe: Well it's only for 5.1 Jun 29 09:35:06 yes, ok, why not. But it doesn't solve the initial question then :) Jun 29 09:35:35 I'll discuss it with Halium guys a bit Jun 29 09:35:55 Well there are only a small number of 5.1 targets Jun 29 09:36:13 https://github.com/Halium/halium-devices/tree/halium-5.1/manifests Jun 29 09:36:26 For 5.1 it's not a problem, I agree: it's pretty much old stuff now anyway Jun 29 09:37:13 Most of the work today is on 7.1, and the edgy work is on 8.0 Jun 29 09:37:18 Yup Jun 29 09:38:39 So, in the end the question would be how to keep a minimal gap between our 5.1 build script and 7.1 one (which wouldn't even need our local manifest) Jun 29 09:40:34 oh but wait: what about introducing this "remove hardware repo and replace it with a patched one" only in the touchpad manifest Jun 29 09:41:51 Because, if I'm correct, it's really only about TP Jun 29 09:47:23 Tofe: Well it shouldn't matter for the others Jun 29 09:47:31 Since we don't set the flags during the build for those Jun 29 09:47:52 It's just that UB Ports until recently at least had no way to have CAF and non-CAF side by side Jun 29 09:47:53 And we do Jun 29 09:48:03 That's what they faced when they started to play with 7.1 Jun 29 09:53:12 Tofe: We shouldn't have issues, because we build libhybris based on https://github.com/shr-distribution/meta-smartphone/blob/pyro/meta-hp/conf/machine/tenderloin.conf#L30 Jun 29 09:53:16 Which is machine specific Jun 29 09:53:26 However UB Ports didn't have this option so they faced issues Jun 29 09:53:30 And needed AOSP Jun 29 09:53:59 So we should be OK to do the remove & add in default.xml for all 5.1 targets Jun 29 11:19:40 will do another webOS OSE build on jenkins soon, this time it should work without the work arounds for qemux86 Jun 29 11:24:39 JaMa: OK great Jun 29 11:24:45 I saw some new bits pushed a few hrs ago Jun 29 12:16:53 Herrie|Laptop: nothing really interesting for us unfortunately Jun 29 12:17:02 JaMa: Yes saw that but well Jun 29 12:17:18 Seems some A1's will be in the mail shortly, got image ready to start testing Jun 29 12:17:41 Likely to fail, but Tofe got his Rosy behaving quite OK too already Jun 29 12:17:48 did they ask for your address already? Jun 29 12:17:52 So we're making progress Jun 29 12:17:55 Yes last week Jun 29 12:17:59 Or early this week Jun 29 12:18:03 aha, good Jun 29 18:12:04 Herrie: https://github.com/Halium/halium-devices/pull/101 Jun 29 18:29:00 Herrie: could you push the https://github.com/Tofee/android_hardware_libhardware/tree/qcom-bsp-5.1 branch to our webos-ports account ? It's the same as upstream, without the two commits for AOSP headers Jun 29 18:31:29 oooh but I just discovered that hp_tenderloin.xml already has what is needed for halium-devices.. silly me... Jun 29 18:31:39 Tofe: We have cm12.1branch already Jun 29 18:32:05 yes, sorry, I'm lagging behind tonight :) Jun 29 18:33:21 Herrie: but then for 5.1 we can drop luneos_hal completely and start using halium's manifest + device/setup ?... Jun 29 18:34:41 Tofe: Yes :-) Jun 29 18:34:58 ok, I'll change our build script then Jun 29 18:35:07 Small bits I was working on to make things a bit more smooth and easier to maintain Jun 29 18:35:29 Just you might want to force sync because otherwise repo will moan about locally changed files Jun 29 18:35:48 I.e. we might want to pass force-sync to setup script Jun 29 18:36:08 doesn't the setup script do that already ? Jun 29 18:36:13 Repo sync options were added not too long ago Jun 29 18:36:24 Tofe: Not force sync I think Jun 29 18:36:50 I usually would just drop my device & vendor folders in between targets but that's nasty Jun 29 18:37:40 At least doesn't force sync by default but you can pass the command line params for repo now, if I recall correctly Jun 29 18:48:44 ok, I think it's correct now Jun 29 18:49:01 Herrie: it's too soon for doing the same for 7.1, right ? Jun 29 18:49:29 https://github.com/Halium/halium-devices/blob/halium-7.1/README.md#L22 Jun 29 18:49:38 Tofe: Well all work except for Onyx really Jun 29 18:50:41 I.e. I need to add some bits to setup script to deal with https://github.com/webOS-ports/android_luneos_hal/tree/halium-7.1/patches/device/oneplus/onyx Jun 29 18:50:51 And https://github.com/webOS-ports/android_luneos_hal/tree/halium-7.1/patches/device/motorola/athene Jun 29 18:51:06 So the various Doze variants, CMActions, LinegeActions that's about it Jun 29 18:52:40 Doesn't look easy to automatize Jun 29 18:56:27 for athene, there's a halium-7.1 branch which looks quite clean Jun 29 18:57:31 Well got the regexps ready more or less just need find time to wrap up and test for soem targets Jun 29 19:00:14 I have quite a hard time reading the [a][b][c][d] regexps... may I propose a PR for simplifying this ? Jun 29 19:09:04 Tofe: I like them like that, but let the majority decide :P Jun 29 19:17:08 https://github.com/Halium/halium-devices/pull/102 & 103 **** ENDING LOGGING AT Sat Jun 30 03:00:03 2018