[05:09] <nehaljwani> Hi! I am creating a debian package and adding it to a repo like this: https://paste.ubuntu.com/p/rVDBRDYpCk/  But when I install it , apt list --installed shows: <my-package>/unknown . What could be the reason?
[05:09] <nehaljwani> How do I remove 'unknown'?
[05:27] <nehaljwani> Found the problem. Suite: section was missing from conf/Distributions . Shouldn't apt fall back to use codename? Eh
[08:38] <rbalint> ginggs, how about passing --jobs 4 in mercurial's debian/tests/testsuite to speed it up and let it pass on arm64, too?
[08:41] <ginggs> rbalint: did you try it?
[08:41] <rbalint> ginggs, not yet, but should work
[08:41] <rbalint> ginggs, i'm just asking since you touched the package last and did not want to upload without a ping
[08:42] <rbalint> ginggs, if you pass i'm giving it a try
[08:43] <ginggs> rbalint: go ahead :)
[08:43] <rbalint> ginggs, ack
[08:50] <ginggs> rbalint: please fix the breaks while you are there
[08:51] <ginggs> mercurial-git (<< 0.8.12-1~) as in changelog instead of mercurial-git (<= 0.8.12-1~)
[08:51] <rbalint> ginggs, ok
[08:51] <ginggs> rbalint: thanks!  i only noticed the typo now
[11:36] <abeato> seb128, hey, any ETA for https://bugs.launchpad.net/ubuntu/+source/network-manager/+bug/1781597 in bionic?
[12:25] <seb128> abeato, hey, we need to clear off the current SRU first, see https://community.ubuntu.com/t/call-for-testing-network-manager-1-10-14-in-bionic/9424 ... maybe you can help verifing that one? :)
[12:26] <rbasak> bdmurray: did you miss doing SRU paperwork for your update-manager upload, or is this  somehow special?
[12:48] <abeato> seb128, sure think - had not seen that
[12:49] <seb128> abeato, thx
[13:32] <xnox> doko, rbasak - any volunteers to figure out why nodejs builds in debian, but not ubuntu? v10 is the LTS with openssl 1.1.1 so it would be nice to have it.....
[13:33] <xnox> and we are getting a tonne of node things in, which i guess target the new nodejs =/
[15:31] <bdmurray> rbasak: it doesn't do anything per se just provides a new variable which ubuntu-release-upgrader will use and I still need to upload that
[15:33] <rbasak> bdmurray: it should still have SRU paperwork before uploaded then though, no?
[15:33] <rbasak> Separately, I suggest that it be uploaded/reviewed at the same time as the ubuntu-release-upgrader SRU then. If for example it turns out further changes are needed, then we'll have wasted an SRU.
[15:33] <rbasak> (and regression risk, etc)
[15:35] <bdmurray> rbasak: yes, it should still have the paperwork
[15:36] <rbasak> OK, I'll leave it there for now, thanks. Sorry, I thought your first reply was implying that it wasn't needed.
[15:42] <LocutusOfBorg> xnox, doko rbasak I might have a nodejs clue
[15:42] <LocutusOfBorg> let me see if I can prove it
[15:43] <LocutusOfBorg> tl;dr: upload nodejs with testsuite disabled, and then reupload with testsuite enabled
[15:43] <LocutusOfBorg> it might be due to it using something old in the system
[15:44] <LocutusOfBorg> it might be a chicken and egg bootstrap issue
[15:44] <LocutusOfBorg> testsuite not using everything from the build directory but picking up something in the system
[15:46] <juliank> How did my email from @gmail get to ubuntu-devel? That's odd
[15:46] <juliank> I mean, that list is subscriber only I thought
[15:47] <xnox> LocutusOfBorg, i'm testing that locally at the moment, on a fast ppc64le builder
[15:47] <rbasak> AIUI posting is restricted to Ubuntu developers onlyu
[15:47] <rbasak> Presumably a Launchpad team.
[15:47] <xnox> LocutusOfBorg, please don't abuse our builders for this, especially if we don't know if it will work =)
[15:47] <rbasak> Does Launchpad know about your gmail address?
[15:47] <LocutusOfBorg> xnox, it is in a ppa, lower priority I would hope
[15:48] <LocutusOfBorg> juliank, how did you upload a build2 package without a build1 one?
[15:48] <xnox> LocutusOfBorg, also possibly better to do it in a bileto PPA, such that archive is not exposed to the testsuite-less build.
[15:48] <xnox> LocutusOfBorg, ppa is fine, cool =)
[15:48] <LocutusOfBorg> xnox, sure, of course
[15:48] <LocutusOfBorg> but meh, I plan to redo in bileto in case it works :)
[15:48] <juliank> LocutusOfBorg: the build1 was in the bileto ppa
[15:48] <LocutusOfBorg> ah ok so you followed my advice
[15:49] <LocutusOfBorg> I don't see the "grouped with ppa:" in update excuses page... strange
[15:54] <juliank> LocutusOfBorg: Yeah, because I published without doing the bileto building stuff I guess
[15:57] <LocutusOfBorg> ah ok
[15:57] <LocutusOfBorg> nice to know
[15:57] <LocutusOfBorg> thanks
[16:04] <LocutusOfBorg> nice the check override didn't work
[16:04] <LocutusOfBorg> xnox, how did you do it?
[16:05] <LocutusOfBorg> -	-ln -s /usr/bin/node node
[16:05] <LocutusOfBorg> -	mkdir -p $(NODE_TEST_DIR)
[16:05] <LocutusOfBorg> -	make $(DEB_MAKE_CHECK_TARGET)
[16:05] <LocutusOfBorg> -	rm -f node
[16:05] <LocutusOfBorg> -	rm -rf $(NODE_TEST_DIR)
[16:05] <LocutusOfBorg> +	echo "no testsuite, everybody happy!"
[16:05] <LocutusOfBorg> +	#-ln -s /usr/bin/node node
[16:05] <LocutusOfBorg> +	#mkdir -p $(NODE_TEST_DIR)
[16:05] <LocutusOfBorg> +	#make $(DEB_MAKE_CHECK_TARGET)
[16:05] <LocutusOfBorg> +	#rm -f node
[16:05] <LocutusOfBorg> +	#rm -rf $(NODE_TEST_DIR)
[16:05] <LocutusOfBorg> this didn't work
[16:05] <LocutusOfBorg> I'm adding this: -DEB_MAKE_CHECK_TARGET = $(exp-relax-check) test-ci-js
[16:05] <LocutusOfBorg> +#DEB_MAKE_CHECK_TARGET = $(exp-relax-check) test-ci-js
[16:09] <teward> did you intend to use a pastebin there LocutusOfBorg or no?
[16:09] <teward> just saying.
[16:09] <LocutusOfBorg> yep :/
[16:09] <LocutusOfBorg> bad copy-paste, lots of the lines were useless
[16:10] <teward> cool cool.
[16:10] <LocutusOfBorg> I just wanted to copy the -make +#make line :)
[16:10] <teward> :)
[16:12] <LocutusOfBorg> why cdbs sucks that much?
[16:12] <LocutusOfBorg> :/ maybe its just me, but I feel lost everytime
[16:13] <LocutusOfBorg> or maybe, debhelper, why you so simple?
[16:13] <xnox> LocutusOfBorg, i didn't bother patching the sources, and i'm just inside a schroot shell. However, no builds yield.
[16:13] <xnox> LocutusOfBorg, at the moment doing a stock rebuild in sid-amd64 fails. So I think it is now busted in Debian too.
[16:13] <xnox> LocutusOfBorg, once this finishes I will file a FTBFS bug report.
[16:14] <LocutusOfBorg> lol
[16:14] <LocutusOfBorg> interesting
[16:16] <LocutusOfBorg> http://debomatic-amd64.debian.net/distribution#unstable/nodejs/10.15.0~dfsg-8/buildlog
[16:16] <LocutusOfBorg> lets see
[16:17] <xnox> LocutusOfBorg, i also wonder if we should just wait for 10.15.1
[16:17] <xnox> https://github.com/nodejs/node/commit/d3c07bd9a81f5ea45a26325be74c93a94b60b243
[16:17] <LocutusOfBorg> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/nodejs.html
[16:18] <LocutusOfBorg> I scheduled new builds there, so you can use them in your bug report
[16:18] <LocutusOfBorg> xnox, seems like there are no commits wrt our issue
[16:18] <LocutusOfBorg> we had 58 failing tests, with acorn bootstrapped we are down to 3
[16:18] <xnox> LocutusOfBorg, there are dozens of tests: commits
[16:18] <LocutusOfBorg> and I already looked upstream with git log... nothing interesting wrt async
[16:18] <LocutusOfBorg> and exceptions
[16:19] <xnox> https://github.com/nodejs/node/pull/25346
[16:19] <xnox> better overview of changes
[16:19] <LocutusOfBorg> at least nothing obvious
[16:19] <xnox> 155 "test:"
[16:19] <LocutusOfBorg> and nothing with async :)
[16:20] <LocutusOfBorg> something with assert is there...
[16:21] <xnox> i'll try building that branch in a second.
[16:21] <LocutusOfBorg> that would be awesome
[16:22] <LocutusOfBorg> https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/locutusofborg-ppa/+packages
[16:22] <LocutusOfBorg> s390x build is already there
[16:22] <LocutusOfBorg> the notest one
[16:22] <LocutusOfBorg> as soon as everything is published I upload the new one
[16:27] <LocutusOfBorg> xnox, http://debomatic-amd64.debian.net/distribution#unstable/nodejs/10.15.0~dfsg-8/buildlog
[16:27] <LocutusOfBorg> debian is good
[16:27] <LocutusOfBorg> test-asser is ok
[16:27] <LocutusOfBorg> it fails another test, fs related (maybe its DebOmatic to blame here)
[16:28] <xnox> LocutusOfBorg, what is DebOmatic btw? never used it before.
[16:28] <LocutusOfBorg> an sbuild instance for ubuntu/debian/arch
[16:28] <LocutusOfBorg> you just ask an account and dput
[16:28] <LocutusOfBorg> http://debomatic-amd64.debian.net/distribution#unstable/nodejs/10.15.0~dfsg-8/buildlog
[16:28] <LocutusOfBorg> it does blhc, autopkgtest, publish
[16:28] <LocutusOfBorg> for a lot of architectures
[16:28] <LocutusOfBorg> (via qemu)
[16:29] <LocutusOfBorg> you can also setup an instance locally btw
[16:29] <LocutusOfBorg> https://packages.qa.debian.org/d/debomatic.html
[16:30] <xnox> cute
[16:31] <LocutusOfBorg> it is sbuild for dummies with a lot of javascript and automation via web interface :d
[16:41] <xnox> LocutusOfBorg, so, to me the check target is bad.
[16:41] <xnox> instead of testing just-built node, it tests system node, during normal package builds.
[16:42] <LocutusOfBorg> it works in debian
[16:42] <xnox> but that bit is added, for autopkgtests, where we do want to test the system node
[16:42] <xnox> but it should be guarded with a test for $$AUTOPKGTEST_TMP
[16:42] <xnox> variable for example.
[16:42] <LocutusOfBorg> so my test disable/enable might work then?
[16:42] <LocutusOfBorg> I see something wrt experimental in rules file
[16:42] <xnox> and if you look further up in the debian/rules, you will see that test results are ignored when building for experimental or unreleased.
[16:42] <LocutusOfBorg> different testsuite
[16:43] <LocutusOfBorg> yes exactly
[16:43] <xnox> it add '-i' to ignore results.
[16:43] <LocutusOfBorg> -i
[16:43] <LocutusOfBorg> that one yes
[16:43] <LocutusOfBorg> I got it too
[16:43] <LocutusOfBorg> I was going to add "disco" :)
[16:43] <xnox> my premise, is that well the just compiled node is never tested, which is bad.
[16:43] <LocutusOfBorg> but it would have been slower
[16:43] <xnox> i'd want to fix that.
[16:43] <LocutusOfBorg> xnox, sure, worth a bug, let me see if re-enabling and reuploading against this one works
[16:43] <LocutusOfBorg> so we are sure about our claims
[16:44] <LocutusOfBorg> it makes sense since the debian no-change rebuild on debomatic works well
[16:44] <LocutusOfBorg> (modulo the fs error, that is unrelated)
[16:46] <xnox> i wonder if my builds fail a lot more due to aggressive parallelism
[16:46]  * xnox was building & testing with -j 120; trying with -j 80 now
[16:47] <xnox> time for starbucks
[16:49] <LocutusOfBorg> LOL
[16:51] <teward> *sips his 6th cup of coffee*
[16:53] <LocutusOfBorg> https://bileto.ubuntu.com/#/ticket/3599
[16:53] <LocutusOfBorg> xnox, ^^
[16:55] <xnox> i still don't want to upload like that. i really want for the just built node to be tested, rather than always testing the wrong node
[16:55] <xnox> but i guess we want to build _a_ nodejs somehow
[17:02] <LocutusOfBorg> but at least we can unblock a lot of stuff
[17:02] <LocutusOfBorg> and then with some more time, fix it properly
[17:02] <LocutusOfBorg> I'm pretty sure debian maintainer already tried this, and they might failed to implement for good reasons
[17:08] <jbicha> it'll give autopkgtest something to do too :) :)
[17:09] <Laney> be kind to the test infrastructure please
[17:09] <Laney> don't queue pointless things at the moment
[17:09] <Laney> you can run tests locally...
[17:12] <LocutusOfBorg> Laney, this is why I'm using bileto
[17:12] <Laney> what is?
[17:12] <LocutusOfBorg> I did upload nodejs on silo so we don't run the testsuite twice
[18:30] <smoser> xnox: around ? or maybe rharper knows.
[18:30] <smoser> i want to make systemd-resolve search a network device tun0 for multiple domains
[18:30] <smoser> (basically wanting to set a dns-search)
[18:31] <smoser> but
[18:31] <smoser>  systemd-resolve  --interface=tun0 "--set-domain=cisco.com insieme.local"
[18:31] <xnox> smoser, sure. use resolvctl or the systemd-resolved command line.
[18:31] <xnox> smoser, `cisco.com insieme.local` is not a valid search domain.
[18:31] <smoser> but that then escapes the ' '
[18:31] <xnox> (single)
[18:31] <xnox> i think you want:
[18:31] <xnox> systemd-resolve  --interface=tun0 --set-domain=cisco.com --set-domain=insieme.local
[18:32] <xnox> but let me double check the code
[18:32] <smoser> where does resolvectl come from ?
[18:32] <xnox> new-ish replacement ofr systemd-resolve
[18:33] <xnox> in more recent releases resolctl is the new world order.
[18:33] <smoser> ah. /me no longer runs development release. back on 18.04
[18:33] <smoser> multiple --set-domain does seem to work
[18:33] <smoser> but i'm not sure that that is plugged correctly trhough network-manager settings for the vpn.
[18:34] <xnox> smoser, network-manager is broken and sends literal "foo.com,example.com" instead of a /list/ of values.
[18:34] <smoser> right
[18:34] <xnox> smoser, i think this has been fixed in disco now, but needs an SRU.....
[18:35] <smoser> oh nice.
[18:36] <smoser> i think you're just trying to get me to go back to development release though
[18:36] <smoser> :)
[18:44] <juliank> Seems there is a bug in the unbound package, just installing it enables the apparmor profile, but the profile is enabled only after the service is started, meaning the service does not have an apparmor profile loaded on first start. hmm
[18:53] <juliank> yup, reported to Debian :)
[19:24] <cornfeedhobo> heeeyyy, i just found a broken line in the AMI released.txt
[19:24] <cornfeedhobo> http://uec-images.ubuntu.com/query/trusty/server/released.txt
[19:24] <cornfeedhobo> search for "paravirtualtrusty"
[19:24] <cornfeedhobo> there is a missing new line
[19:40] <sarnold> cornfeedhobo: heh, nice find
[19:40] <sarnold> Odd_Bloke: /lastlog cornfeedhobo   -- http://uec-images.ubuntu.com/query/trusty/server/released.txt < cornfeedhobo> search for "paravirtualtrusty"