[01:13:57] *** Quits: Horwitz (~mich1@p200300ec9f2b8400022268fffe64e7c4.dip0.t-ipconnect.de) (Ping timeout: 272 seconds) [01:26:07] *** Joins: Horwitz (~mich1@p200300ec9f3e0800022268fffe64e7c4.dip0.t-ipconnect.de) [01:26:07] *** ChanServ sets mode: +o Horwitz [03:04:44] *** Quits: Cyker (~Cyker@host-92-28-104-61.as13285.net) (Ping timeout: 265 seconds) [03:06:17] *** Joins: Cyker (~Cyker@host-92-28-111-125.as13285.net) [10:55:53] *** Quits: mvglasow (~mvglasow@78-61-158-97.static.zebra.lt) (Ping timeout: 256 seconds) [11:08:39] *** Joins: mvglasow (~mvglasow@78-61-158-97.static.zebra.lt) [12:22:01] hi all [12:35:41] mvglasow: heres my current diff for android and gradle: https://gist.github.com/jkoan/fdd4a7e1af4fc42661b7fb4ef661d646 [12:40:54] jkoan great, have you tested it successfully yet? are you planning to include that in your Android build activities? (So we don't end u developing the same thing at the same time.) [12:41:26] my current problem is to reintegrate the images which where build from cmake (which is working) back into gradle. Basically "make install" into a directory which could be created by gradle to then keep those images back up... [12:43:27] the build itself is working with the CMAKE_SYSROOT, yes. [12:56:50] in navit/CMakeLists.txt around line 17, you change if (${HAVE_GLIB}) to if (HAVE_GLIB), is that needed for the sysroot issue? [12:57:40] no, i just found it, and classified it as wrong, so i fixed it [15:28:09] looks good, at least the F-Droid build works now. Local build and smoke test on a device are still pending. Oddly though, sanity-check fails because NDK is not found (running from my own fork, not the official repo). [15:38:52] doh! found the error, checkstyle chokes on our use of `android.ndkDirectory` in build.gradle, probably because it is running on a different image than the actual build [15:41:54] I meant sanity_check, not checkstyle [15:42:09] jkoan: I see the Android build runs on circleci/android:api-29-ndk whereas sanity_check runs on navit/sanity_check:latest [15:44:47] does sanity_check need anything that is not available in circleci/android:api-29-ndk? if not, it would make sense to just use the Android image [15:47:54] I just tried and it seems to work, running time is slightly longer (39 vs. 33 seconds) [15:49:06] and the APK I built locally seems OK, starts up and I get a menu... if libraries were missing, I wouldn't expect it to get that far [15:49:34] Nice :) [15:55:19] if you want to earn yourself so e reputation on StackOverflow, feel free to add an answer here: https://stackoverflow.com/questions/66339827/how-to-tell-cmake-invoked-through-externalnativebuild-where-to-look-for-librar/66340906#66340906 [15:56:57] Actually I don't have a SO account (yet) [15:57:46] I can post it as well, just didn't want to take credit for your answer without asking first :-) [16:03:21] feel free to do so :-) [16:09:05] once the PR is merged i will create a new release. Its probably lone overdue due to the vehicle_geoclue plugin as well (for example Pinephone) [16:10:41] @aerostitch do you have some spare time today? i would like to create a release and would also like to have the debian package updated as well so that the mobian-project could get up to date integration for vehicle_geoclue on Pinphone/Pinetab/... [16:12:25] mvglasow: any whishes for planing navit 0.7.0 i could fulfill from you? [16:13:41] currently my main wish is that we get an updated version out on F-Droid (possibly an interim one as we haven't had any updates in a while) [16:14:29] TraFF 0.8 support for Android is merged, I might still write a TraFF HTTP plugin, but no guarantee on timeframe [16:15:33] also, someday a list of traffic events in the GUI would be nice, but again, subject to availability of my time [16:16:17] yes, i do wish for the same. Do we need to add a new recipe for f-droid before we can trigger an update? [16:17:05] I'll look into it, I can finally finish the F-Droid CI thing, after that I can tell for sure [16:17:54] from my understanding it should be enought to update: https://download.navit-project.org/api/version.json [16:18:20] for the F-Droid recipe we need to change the path to the output file [16:18:39] ah wait, i also got the fix for that [16:19:06] we'll have to look at NDK, as the `ndk` directive doesn't work on CI, I have instead installed NDK 20 as part of the recipe [16:19:20] but maybe that will work as it did before [16:30:44] i also think the actual build from f-droid are done on vm's where the ndk from the ndk tag is installed [16:35:38] mvglasow: why do we use those sudo steps? I remember that you added them, but we dont need protobuf-c for anything android related, because only maptool needs them [16:38:54] what are you referring to? the F-Droid build recipe? [16:39:02] yep [16:39:24] i think we can skip sudo all together [16:39:44] we can try to strip that down further and see if the build still completes [18:08:02] jkoan after some tests, the F-Droid build completes without protobuf, so that part can go [18:09:44] That sounds nice because it also removes the problem with sudo (makes manual testing easier) [18:32:27] NDK may still be an issue: I removed the manual installation and reinstated the ndk directive, but with that the build fails on CI. That'll be for the F-Droid guys to figure out as the live server behaves differently from CI; I'll file a merge request with the information they need and post the link here [19:09:09] here it is: https://gitlab.com/fdroid/fdroiddata/-/merge_requests/8472 [21:00:01] Failed: mvglasow's build (#22934; push) in navit-gps/navit (fdroid-ci) -- https://circleci.com/gh/navit-gps/navit/22934?utm_campaign=chatroom-integration&utm_medium=referral&utm_source=irc [21:00:25] Failed: mvglasow's build (#22935; push) in navit-gps/navit (fdroid-ci) -- https://circleci.com/gh/navit-gps/navit/22935?utm_campaign=chatroom-integration&utm_medium=referral&utm_source=irc