[00:12:41] *** Quits: ilovekiruna (~ilovekiru@2a02:8108:d00:110:c148:52f5:d957:d25) (Remote host closed the connection)
[00:16:03] <mvglasow> jkoan CI on F-Droid not matching the build environment is an open issue there, at least as far as the NDK part goes
[00:17:56] <mvglasow> besides that, I cloned their fdroiddata repo and enabled CI on my fork (with F-Droid's config), which works for me
[00:18:48] <mvglasow> and if something builds on CI but not on the build server (or vice versa), that would be an issue, but so far I've managed to reproduce it
[00:20:34] <mvglasow> btw, the getcwd issue only happens with later NDK versions, and I haven't seen it on F-Droid yet (though the other error might just have occurred earlier)
[01:35:05] *** Quits: Horwitz (~mich1@p200300ec9f294500022268fffe64e7c4.dip0.t-ipconnect.de) (Ping timeout: 272 seconds)
[01:47:44] *** Joins: Horwitz (~mich1@p200300ec9f3a6100022268fffe64e7c4.dip0.t-ipconnect.de)
[01:47:44] *** ChanServ sets mode: +o Horwitz
[10:19:16] *** Quits: mvglasow (~mvglasow@78-61-158-97.static.zebra.lt) (Ping timeout: 240 seconds)
[10:32:20] *** Joins: mvglasow (~mvglasow@78-61-158-97.static.zebra.lt)
[11:19:42] *** Joins: ilovekiruna (~ilovekiru@2a02:8108:d00:110:c148:52f5:d957:d25)
[11:19:42] *** ChanServ sets mode: +v ilovekiruna
[12:07:00] <jkoan> hi mvglasow, i will try to look into the issues now (hopefully i don't need to go to move snow every couple of minutes)
[12:14:01] <ilovekiruna> hi jkoan
[12:14:06] <jkoan> hi ilovekiruna 
[12:14:39] <ilovekiruna> the talk I saw just now about onboarding wasnt so great
[12:15:24] <jkoan> good to know :d i have just seen your messages from yesterday
[12:17:07] <ilovekiruna> the susi.ai people are not as communicative as I hoped :-(
[12:17:33] <ilovekiruna> one guy told me, one could test german and japanese
[12:17:44] <ilovekiruna> right now I have no time unfortunately
[12:18:11] <ilovekiruna> but would be nice to present a speech interface and the finally setup device farm next year
[12:18:20] <ilovekiruna> would like to set that as goal for me
[12:19:18] <ilovekiruna> how is the snow in Peine?
[12:23:28] <jkoan> Its like Dust, but the amount is growing in such rates that we could move it away immediately after we are finished on the other end. And than the snow moving vehicles are comming right after we finished the task just to put the snow from the street back to the sidewalk. And we already had moved the snow from the street in addition to the sidewalk into our dooryard...
[12:35:00] <ilovekiruna> doesnt seem so bad here at least in front of the house
[13:12:08] <ilovekiruna> jkoan: do you have a few minutes to look on some code of mine?
[13:12:33] <jkoan> i am just back from moving the next batch of snow, so yes
[13:13:45] <ilovekiruna> do you know why I get a segfault in https://dpaste.org/mhix
[13:19:41] <jkoan> Just used a debugger to see where the segfault is: its in Line 22 *pointDispl_ is NULL
[13:21:03] <jkoan> I just use Eclipse and its Debugger as it automatically brakes when a segfault is detected, the last state can than be debugged
[13:21:55] <ilovekiruna> my question is rather in understanding why it happens
[13:25:09] <ilovekiruna> btw: this talk could be interesting for all of us a bit: https://fosdem.org/2021/schedule/event/community_devroom_mental_health_free_software/
[13:28:59] <jkoan> I think its because the definition has ** instead of * and the new double then takes the place of *x instead of **x. Also shouldnt it be new double(1) instead of []? But i am also not really sure
[13:51:29] <mvglasow> jkoan great! Up here in Lithuania we got our dose of snow two weeks earlier, I've been through the same... today the weather's great for cross-country skiing :-)
[13:52:18] <jkoan> hi mvglasow :-)
[13:52:55] <jkoan> for android-builddir you probably also need to adjust the path inside navit/android/build.gradle
[13:53:12] <jkoan> but i am still testing
[13:54:06] <jkoan> do you know if there is a prebuild docker/vm image from f-droid with there build environment?
[13:55:18] <mvglasow> there is a CI image, not 100% identical to the build environment but somewhat close (they might have a different NDK version on the live machine)
[13:55:31] <mvglasow> to use it:
[13:55:48] <mvglasow> clone gitlab.com/fdroid/fdroiddata/
[13:56:26] <mvglasow> then ensure CI is enabled on your clone as described in https://stackoverflow.com/questions/66070323/no-ci-cd-on-forked-gitlab-project/66070324#66070324
[13:56:58] <mvglasow> create a branch in your fdroiddata fork
[13:57:24] <mvglasow> now every time you change org.navitproject.navit.yml and push the changes, CI will start an F-Droid build
[13:58:07] <mvglasow> the one thing that doesn't work on CI is specifying an ndk version in the build recipe (see ticket)
[13:59:35] <jkoan> ah, but i can test localy by just using the things from .gitlab-ci.yml directly inside docker, thx
[13:59:41] <mvglasow> you can see my attempts at getting a working build recipe at https://gitlab.com/mvglasow/fdroiddata/-/commits/navit-055-fix
[14:00:44] <mvglasow> if you want to go the hard way and set up your own F-Droid build server, sure, that is an option as well
[14:01:31] <mvglasow> I skipped that part as I'm not that savvy about all that container stuff :-)
[14:02:21] <jkoan> i have a suspission that this doesn't match up: /builds/mvglasow/fdroiddata/... and /usr/include/glib-2.0 the android build should use the glib from the ndk not from the host
[14:03:27] <mvglasow> then I wonder why that works on our end but not on F-Droid
[14:03:31] <jkoan> doing it locally is just a matter of running this: https://gitlab.com/fdroid/fdroiddata/-/blob/master/.gitlab-ci.yml#L149-216 inside the containter defined in Line 150
[14:03:41] <jkoan> yes, me to
[14:03:54] <jkoan> will try to figure this out
[14:04:34] <jkoan> probably because fdroid will run gradle again after our build_android.sh is done? (just a guess for now)
[14:05:53] <mvglasow> it doesn't even get that far, at least from what I can tell... though it's hard to figure out, as clang spews all kinds of warnings I thought we had disabled (#1086)
[14:06:27] <jkoan> i have seen the issue, but i want to try to fix the build first
[14:27:07] <jkoan> mvglasow: https://gist.github.com/jkoan/0d0f786c4fec42c0b94ddb7e5d7fe407 this should be the recipe to create a Docker image which has everything needed to build navit (as its converted from there ci file). Currently its running, so i will see what happens. Probably i will also remove the last two lines to make it interactive to test
[14:30:36] <mvglasow> are you planning to add an F-Droid build to our CI chain? Nice...
[14:31:24] <jkoan> not really, but would be an idea, but than i would do that on github actions instead of circleci
[14:34:46] <jkoan> afk some time, hope the image does build in the meantime
[14:36:43] <mvglasow> actually, running an F-Droid build as part of CI sounds like a cool idea, might want to consider this for my own projects (which are on GitLab and already use GitLab CI)
[15:03:37] <ilovekiruna> @jkoan, i think i solved my issue
[15:04:41] <Navit> <jkoan> What was the problem?
[15:08:37] <ilovekiruna> changing*pointDispl_[i] = my_float  to pointDispl_[i] = &my_float;
[16:08:12] <ilovekiruna> so something is apparently wrong with my memory allocation
[17:06:34] <mvglasow> jkoan re #1086, just figured out that add_compiler_flag_if_available("-w") works as intended, i.e. silences all clang warnings. Not a permanent fix, but while you're debugging, it might make the output more legible...
[18:34:02] *** Quits: Celelibi (celelibi@2a01:e0a:587:56f0:215:afff:fe49:d5cd) (Read error: Connection reset by peer)
[18:50:13] *** Joins: Celelibi (celelibi@2a01:e0a:587:56f0:215:afff:fe49:d5cd)