| T42 | <loey1199> Does any one had the sailfish os file for oneplus 6T? | 02:16 |
|---|---|---|
| T42 | <fredldotme> Sad but understandable, we also made some packaging efforts without icd, avoiding conflicts with libvulkan1. (re @FakeShell: @fredldotme i gave u...) | 05:17 |
| Mister_Magister | @loey1199 verdanditeam.com | 07:51 |
| T42 | <mettska111> So, I still have the problem that rpm/dhd/helpers/build_packages.sh --mw fails | 11:44 |
| T42 | <mettska111> I searched around and did make modules | grep gralloc (what works, but nothing changed | 11:45 |
| T42 | <mettska111> https://paste.opensuse.org/pastes/93216e414229 | 11:46 |
| T42 | <mettska111> where can I go on with my research? I have no real idea. | 11:46 |
| T42 | <mettska111> here is my libhybris.spec | 11:48 |
| T42 | <mettska111> https://paste.opensuse.org/pastes/6677652a74ec | 11:48 |
| mal | @mettska111 did you apply patches correctly? | 13:14 |
| mal | also making sure you haven't done repo sync or something after applying patches which might remove the applied patches | 13:18 |
| T42 | <mettska111> I made exactly like written in the hadk-faq under hybris-16.0 | 13:30 |
| T42 | <mettska111> it was a time ago, but I think i didnt made a repo sync afterwards | 13:30 |
| T42 | <mettska111> but if you dont have another idea, i will apply patches again and than try again | 13:30 |
| mal | go to folder frameworks/native and run "git log" do you see some patches with commit messages starting with (hybris) | 13:31 |
| mal | wait | 13:32 |
| mal | I'll check one thing | 13:32 |
| mal | the patch should be ok | 13:33 |
| T42 | <mettska111> https://paste.opensuse.org/pastes/a3670be5e024 | 13:34 |
| mal | so the patches have not been applied, I wonder if all patches are missing or just some | 13:35 |
| mal | see what happens when you rerun "hybris-patches/apply-patches.sh --mb" there is a chance it might cause issues if only some repos are missing the patches | 13:36 |
| T42 | <mettska111> gives me that error: | 13:37 |
| T42 | <mettska111> fatal: previous rebase directory .git/rebase-apply still exists but mbox given. | 13:37 |
| T42 | <mettska111> ah, wrong directory | 13:37 |
| T42 | <mettska111> no^^, was the right directory | 13:37 |
| mal | which one fails? | 13:38 |
| T42 | <mettska111> hybris-patches/apply-patches.sh --mb | 13:39 |
| mal | I mean which repo | 13:39 |
| mal | does it print anything? | 13:39 |
| T42 | <mettska111> only this: fatal: previous rebase directory .git/rebase-apply still exists but mbox given. | 13:39 |
| mal | go to folder bionic and run "git status" | 13:39 |
| T42 | <mettska111> Not currently on any branch. | 13:40 |
| T42 | <mettska111> You are in the middle of an am session. | 13:40 |
| mal | git am --abort | 13:40 |
| T42 | <mettska111> no reapply patches? | 13:40 |
| mal | then back to $ANDROID_ROOT and try apply-patches.sh again | 13:40 |
| T42 | <mettska111> *now | 13:40 |
| T42 | <mettska111> https://paste.opensuse.org/pastes/5cba3e82dacb | 13:42 |
| mal | hmm, it seems patches fail for some reason, I'll have a look | 13:44 |
| T42 | <mettska111> thanks | 13:44 |
| mal | how did you even manage to build hybris-hal without patches | 13:44 |
| T42 | <mettska111> thats a pretty good question. But it worked. | 13:45 |
| mal | check that bionic folder again and do "git am --abort" and the check "git log" to see if the patches were applied again | 13:45 |
| mal | *applied already | 13:45 |
| T42 | <mettska111> it looks like it worked, right? | 13:49 |
| T42 | <mettska111> https://paste.opensuse.org/pastes/bb5bd8c9d650 | 13:49 |
| mal | so you seem to have partially applied patches at the moment | 13:50 |
| mal | I'll update my local hybris-22.2 checkout and see if the patches apply correctly or not and will fix the patches if needed | 13:50 |
| mal | I had no trouble applying the patches, not sure why you have only partially applied patches | 13:58 |
| T42 | <mettska111> ok, so I start with git clone again. | 13:58 |
| mal | no | 13:58 |
| mal | @mettska111 before running some commands to fix the issue, did you do changes to kernel or device repos | 13:58 |
| T42 | <mettska111> ah, yes, like hadk-guide says | 13:59 |
| mal | maybe run "repo diff" and store the output so you have backup of any changes | 13:59 |
| mal | another question, when you did the local changes did you use "git commit"? | 14:02 |
| mal | that will affect how to proceed with fixing | 14:02 |
| T42 | <mettska111> yes, I did it exactly like given in the hadk-guide | 14:03 |
| mal | then you need to make sure you backup the modifications to kernel and device repos | 14:05 |
| T42 | <mettska111> ok, I'll do | 14:05 |
| mal | once you have the changes backed up then run "repo sync -l" which should remove any local patches and after that you can apply patches again and make sure it works without issues, and finally apply any changes to kernel and devices you just backed up | 14:08 |
| T42 | <mettska111> ok, thanks. I'll do | 14:11 |
| T42 | <birdzhang> What is debugedit? My droid-hal-$DEVICE failed to build on obs, it was working before 5.1, the url is https://build.sailfishos.org/package/show/home:birdzhang:nemo:devel:hw:xiaomi:ginkgo/droid-hal-ginkgo | 14:19 |
| mal | @birdzhang btw, devices are moving to using short version in repos like 5.1, I think community-adaptation package should be doing that automatically but seems you have a custom community-adaptation package for some reason | 14:24 |
| Mister_Magister | mal: not to nag but any chances for camera2? :P | 14:27 |
| mal | @birdzhang in case you still want to use custom version get this from upstream https://github.com/mer-hybris/community-adaptation/commit/69e98ebadef5ca706bf45940b4c5b7c38c0d86db | 14:27 |
| mal | that has been supported since 4.6 in ssu | 14:28 |
| mal | @birdzhang you had the repos in wrong order in Meta in your obs project | 14:30 |
| mal | @birdzhang I modified that Meta, please check it | 14:45 |
| T42 | <birdzhang> Thanks mal, working now | 15:32 |
| Mister_Magister | UGhhhhhh sending sms does not work | 17:58 |
| Mister_Magister | rinigus: ping about SMS, i'm not sure if your issue on nagara was the same so lemme ask, if i don't register for volte using ext-qti i can send and receive sms no issue, but once i register for volte I can only receive sms and not send, but from what i gathered you had opposite you could send but not receive? | 18:11 |
| Mister_Magister | I do get imsradio0< [00000005] 43 sendImsSms | 18:19 |
| Mister_Magister | but like no confirmation and it never arrives | 18:20 |
| Mister_Magister | i've been looking at qti_ext so long maybe i'll start to understand what it does | 18:21 |
| Mister_Magister | i got turbo brainded idea | 18:30 |
| Mister_Magister | if on 1.2 interface I was getting ids that were +2… what if i send sms with id +2 | 18:31 |
| rinigus | Mister_Magister: it was a while ago. note that nagara has volte over aidl, so we have separate implementation for it | 18:41 |
| Mister_Magister | gotcha i'm just trying to fix my sms sending xd | 18:41 |
| Mister_Magister | any help is welcome | 18:41 |
| rinigus | that's in https://github.com/rinigus/ofono-binder-plugin-ext-qti | 18:42 |
| Mister_Magister | ye but i'm using binder one | 18:43 |
| Mister_Magister | so i take it you have no tips for me? | 18:43 |
| rinigus | note that the "official" one (hidl) has few bugs, as far as I understand. I did open issue on their repo, but that's ignored so far - at least was when I checked last time | 18:43 |
| rinigus | writing tips <begin/> | 18:44 |
| rinigus | you could check aidl implementation and look into what's missing in hidl. probably some interfaces are similar | 18:44 |
| Mister_Magister | aw man | 18:45 |
| rinigus | what helped me was decoding stock android all kind of java libs/apps and then trying to find how sms and other staff was done and communicated to services | 18:45 |
| Mister_Magister | that's only way beyond my capabilities xd | 18:46 |
| rinigus | then, for nagara, we had to have this running - https://github.com/sailfishos-sony-nagara/fake-qcrilmsgtunnel . maybe needed if you are on qcomm. | 18:46 |
| Mister_Magister | that's what I asked you about xd | 18:46 |
| Mister_Magister | if that stuff only fixes receiving or did you have issue with sending | 18:47 |
| rinigus | it fixed everything | 18:47 |
| Mister_Magister | all issues | 18:47 |
| Mister_Magister | heck won't hurt to try ig | 18:48 |
| rinigus | its similar to what one of qcom services is doing already from older times and I have no idea how we managed to work without it. | 18:48 |
| rinigus | sure | 18:48 |
| rinigus | </end tips> | 18:48 |
| rinigus | question is just how universal is it... | 18:48 |
| Mister_Magister | i suspect it won't work at all xd | 18:49 |
| Mister_Magister | as I don't have any process named vendor.qti.hardware.radio.qcrilhook | 18:49 |
| rinigus | then it wouldn't :) | 18:50 |
| Mister_Magister | eugh | 18:50 |
| rinigus | then you have to go over your capacity and decode APKs. I also thought that it was way too much, but when in the end you cannot get SMS, its kind of sucks too much | 18:51 |
| Mister_Magister | just don't use volte and sms work fine just fine xd | 18:51 |
| rinigus | not sure it is long-term solution :( | 18:52 |
| Mister_Magister | expected 'vendor.qti.hardware.radio.ims@1.1::IImsRadio' but read 'vendor.qti.hardware.radio.ims@1.2::IImsRadio' now that's interesting | 18:54 |
| Mister_Magister | every time i send sms | 18:55 |
| Mister_Magister | rinigus: https://paste.opensuse.org/pastes/4457c3c49083 | 19:00 |
| rinigus | Mister_Magister: one more tip - receive confirmation for volte SMS does not work - I think it is not wired to SFOS stack in ofono | 19:05 |
| rinigus | when you start fake daemon, did you end up getting "Sending ATEL ready" message? | 19:06 |
| rinigus | as that's all this daemon is about.. | 19:06 |
| Mister_Magister | yes | 19:07 |
| rinigus | then it should be OK on that side. did it help? other messages are for info, we have resp_id=525306 also. the other one we don't | 19:08 |
| Mister_Magister | my tester takes long time to respond if it worked xd | 19:09 |
| rinigus | then probably no luck | 19:11 |
| Mister_Magister | rinigus: what's interesting is that even if you disable the volte registration it still doesn't work, like, if you start ofono with registration unticked, it will work, if you tick registration and it registers, it doesn't work, if you untick it, it still doesn't work | 19:17 |
| rinigus | those all kind of combinations are probably plenty. it is difficult to get volte running and requires significant amount of time, unfortunately | 19:22 |
| Mister_Magister | no volte works fine just fine | 19:28 |
| T42 | <b100dian> rinigus: for the nagara spinner when 2nd sim is enabled do you have any logs? maybe get recovery boot img and telnet for that if not even wireless works | 22:28 |
| T42 | <b100dian> is there any way to disable the 2nd sim initially through a sparse file 🤔? | 22:45 |
| nightishaman | Would there be a way to tell the hybris build script to use a prebuilt kernel that I would specify, because Pixel kernels have to be built using bazel and don't really support Makefiles anymore. At least I've been hacking together a solution using Makefiles the last days but I really don't like it. I'd rather modify the defconfig and let bazel | 23:21 |
| nightishaman | build it with the kernel modules, then tell hybris to use those | 23:21 |
| mal | I have used bazel to build kernel for a sailfish device, I think you just need to remove the prebuilt kernel, I remember seeing that the build might actually detect whether prebuilt exists or not and might then fall back to building the kernel if no prebuilt it found | 23:37 |
| nightishaman | mal How did you get it to work? Currently it calls the makefiles and so I had to add TARGET_KERNEL_EXT_MODULES | 23:39 |
| mal | well the build just used bazel, I didn't have to do anything really | 23:42 |
| mal | various lineage kernel repos seem to have a script build_with_bazel.py | 23:43 |
| nightishaman | mal hmm, i don't how to get this to work with the repos im using, lineageos-22.2 used prebuilt kernels and it doesn't seem to have a file to build it yourself, id have to try to export the environment variable, but I don't know if it will do it | 23:55 |
| nightishaman | and as far as im aware there is no hybris for android 16 yet | 23:55 |
| mal | so did you have this in local manifests https://github.com/LineageOS/android_kernel_google_gs-6.1_manifest/blob/lineage-23.2/default.xml | 23:57 |
| mal | hmm, but that is only for 23.2 | 23:58 |
| mal | or 23 | 23:58 |
| nightishaman | yeah I tried with that since the kernel imported is for android 14 anyway, so it should also work with android 15 | 23:58 |
| mal | there is hybris-23.2 but excluding working droidmedia (for camera) | 23:59 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!