[02:47] <sbeattie> okay, other than the android-platform-libnativehelper/android-framework-23 issue I mentioned, everything else in https://launchpad.net/~openjdk-11-transition/+archive/ubuntu/android-tools/ is +1 to go to bionic-proposed.
[02:51] <sbeattie> (and I'd be +1 on android-platform-libnativehelper and android-framework-23 if someone can explain to me why the dropped JniConstants are not an issue)
[03:27] <sbeattie> looks like https://android.googlesource.com/platform/libcore/+/72a46b24e87902908d544c4ba0bdd218368a280e and https://android.googlesource.com/platform/libcore/+/59cd43352e5fd4907303d06f9b92f01c943375f7 got rid of the references upstream.
[10:19] <doko> sbeattie: forwarded to #debian-java
[10:20] <doko> sbeattie: please review android-platform-build in android-tools. should be the last one there
[10:23] <doko> tdaitx: sweethome3d, running on my laptop, bionic, no chroot. so no bind-mount issue
[10:25] <doko> sbeattie: https://launchpad.net/~openjdk-11-transition/+archive/ubuntu/test/+build/16486471 has a a-f-23 test build
[10:26] <doko> https://launchpad.net/~openjdk-11-transition/+archive/ubuntu/test/+sourcepub/9968168/+listing-archive-extra
 @doko The build script for "android.jar" is not easy to maintain and updating it is so annoying. I have been experimenting a way to build it using upstream's build system.
 "android.jar" only contains declarations, the contents of methods are all erased. So it doesn't care about those missing APIs.
[12:01] <doko> sbeattie: ^^^
[13:02] <doko> sbeattie, tdaitx, huehner:feedback appreciated for the announcement mail to u-d-a: https://paste.ubuntu.com/p/nHvTqf5c8n/
[13:27] <doko> tdaitx: I copied the ppa packages to the cosmic pocket. first copied every package, then removed every package which showed a superseded version. So that should be safe. will you update the ti arch-any packages? but please don't re-upload those which I already removed
[13:56] <tdaitx> doko: sure
[16:31] <doko> sbeattie: any change to review the one missing package, and reply to the comments above?
[16:31] <doko> any chance even ...
[16:32] <sbeattie> doko: yes, looking now.
[16:32] <sbeattie> +1 on  android-platform-build
[16:39] <doko> sbeattie: and the issue about the "missing" references?
[16:42] <sbeattie> doko: to be clear, based on seamlik's comments, we'd keep things as-is and not backport those two commits like in the test build?
[16:42] <sbeattie> (just want to make sure I understand scrollback)
[16:43] <doko> yes
[16:43] <doko> yes, keept it as is
[16:49] <doko> sbeattie: ?
[16:49] <sbeattie> I'm still looking
[16:49] <sbeattie> (trying a local unmodified test rebuild)
[16:50] <doko> sorry to pester you, but I'm trying to get this ready, so we can copy these packages, and then send the announcement email
[16:52] <sbeattie> yeah, got you.
[16:52] <sbeattie> okay, I'm +1 on android-platform-libnativehelper and android-framework-23, so +1 on everything in the android-tools ppa.
[16:52] <sbeattie> doko: ^
[16:54] <doko> sbeattie: ta. and I checked, that the android.jar is the same in both packages
[16:54] <doko> sil2100: ^^^ please copy ppa:android-tools
[16:55] <sil2100> doko: whole PPA?
[16:55] <doko> sil2100: yes
[16:55] <sil2100> ACK
[17:08] <doko> tdaitx_: looking at your list of of openjdk 10 ftbfs: libgpars-groovy-java apache-log4j2. Should we update these to disco versions as well?
[17:09] <tdaitx_> libgpars-groovy-java is in stage3
[17:10] <doko> ok
[17:10] <tdaitx_> let me look at apache-log4j2
[17:15] <doko> sil2100: please also update the hints for the packages uploaded earlier today, and add those for the new ppa
[17:17] <sil2100> Ok, will do, for now I'm slowly approving the newly copied packages
[17:18] <tdaitx> so apache-log4j2 has a single fix for its own FTBFS + it adds a java 9 module
[17:18] <tdaitx> of its reverse (b-)deps we only updates jabref, which works fine with the current apache-log4j2 version in bionic
[17:18] <tdaitx> thus the only reason to update it is to fix the FTBFS as far as I can tell
[17:20] <tdaitx> doko: ^ do you see any other problem besides apache-log4j2 FTBFS in bionic? are we updating non-runtime issues at this step?
[17:20] <tdaitx> if so, yes, we should update it
[17:20] <huehner> tdaitx: sorry to jump in here... upstream precompiled jar files of log4j2 use a feature called multi-release jar-files (several verions of .class file targeting different jvm versions auto-selected at runtime) and current packaged ant has some bug (missing support) causing failures at runtime
[17:21] <huehner> tdaitx: not sure if can issue also with your package
[17:21] <huehner> https://bugs.launchpad.net/ubuntu/+source/ant/+bug/1808383
[17:21] <huehner> fixed upstream by ant but no release done yet :(
[17:21] <tdaitx> oh yeah, did the ant we backported included this fix?
[17:21] <tdaitx> I forgot to look at that
[17:21] <tdaitx> it would be a good time to include this
[17:21] <huehner> tdaitx: afaik not... we tied to push upstreadm (ant) to make a 1.10.6 release but they maybe forgot (weeks after not yet)
[17:22] <huehner> so afaik unfixed everywhere in ubuntu
[17:23] <tdaitx> huehner: any change this patch can introduce issues or require other packages to be updated or rebuild?
[17:23] <doko> tdaitx: yes, updating anything else would be ok, but not a priority for today's email
[17:23] <doko> tdaitx: I just checked for these openjdk-10 ftbfs in our progress document
[17:25] <doko> huehner: please attach a patch and subscribe ubuntu-sponsors
[17:28] <huehner> tdaitx: from reading upstream commit should be quite safe, as essentially switches to newer openjdk api understand new java feature, however they did bit of code churn +adding test-cases to diff not super small
[17:28] <huehner> i'll check if upstream commit applies on top of 1.10.5 (we have packaged) and attach it
[17:28] <huehner> will be in evening... outside dayjob
[17:31] <tdaitx> huehner: thanks, is this something that affects other packages builds or basically helps users that have to deal with mr-jars?
[17:37] <huehner> tdaitx: only mrjar when loaded via ant-classloader (your custom ant task running inside ant jvm)
[17:37] <huehner> tdaitx: so very like that for 95% of people they never touch it
[17:37] <tdaitx> ack
[17:37] <huehner> tdaitx: there is workaround flag for ANT_OPTS... but i would prefer fixing it inside ant if we can
[17:38] <huehner> tdaitx: just forgot to prod upstream more to do a 1.10.6 release including it (they wanted in january already but maybe forgot)
[17:51] <sil2100> Damn those LP timeouts
[17:54] <doko> the issue is too big
[17:54] <doko> but we'll have more fun adding the cosmic tasks ...
[17:58] <sil2100> Ok, all copied and approved, will update the hints
[18:05] <sil2100> Should be updated
[21:16] <huehner> tdaitx, doko: i am preparing ant patches we talked about: so far, 2 new quilt patches (1 per upstream commit) just with quilt import, 1patch mofified to remove hunk for patching WHATNEWS (does not apply as 1.10.6 section does not exist in 1.10.5) small changelog entry examples, will be tested by us on top of bionic-proposed 1.10.5 (idea to add explanation of commit + risk comments in lp bug)
[21:16] <huehner> does that fix what you need or you prefer something different?