/srv/irclogs.ubuntu.com/2017/10/20/#ubuntu-devel.txt

=== Unit193 changed the topic of #ubuntu-devel to: Artful Released | Archive: closed | Devel of Ubuntu (not support or app devel) | build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of trusty-artful | #ubuntu-app-devel for app development on Ubuntu http://wiki.ubuntu.com/UbuntuDevelopment | Patch Pilots:
Unit193(To further ammend Laney's pointing to #ubuntu, precise isn't supported anymore.00:00
Unit193)00:00
daxUnit193: so you could say you made the topic more precise by making the topic have less precise00:02
sarnold*groan*00:02
Unit193Hahaha. :D00:02
ArMedicIs there a way to display the battery meter with a numerical output in 17.10?03:36
sarnoldArMedic: you may have more luck in #ubuntu03:41
ArMedicsarnold, Thanks, but thats why I came in here, no luck there.03:42
sarnoldoh. bummer.03:43
=== Foxtrot is now known as foxtrot
RAOFOh, huh.05:48
RAOFI appear to have TIL-status on capnproto. Oops.05:48
=== infinity0_ is now known as infinity0
RAOFDear lord!06:06
RAOFSince when did C++ libraries start downloading code from github as a part of the build?!06:06
juerghTrevinho, did you see my gjs_dumpstack() trace  in LP: #171498906:10
ubottuLaunchpad bug 1714989 in GNOME Shell "gnome-shell crashed with SIGSEGV in g_type_check_instance_cast() from st_label_set_text() (dash-to-panel specific?)" [Critical,Incomplete] https://launchpad.net/bugs/171498906:10
Unit193"Well golang does it!"06:10
Trevinhojuergh: yes, and I submitted a fix on the extension for that right now06:11
Trevinhojuergh: https://github.com/jderose9/dash-to-panel/pull/26306:11
Trevinhoif you can test that change in your local extension it would be cool06:11
Trevinho(just apply that diff to your file should work)06:11
juerghTrevinho, sweet!06:11
juerghTrevinho, I'll give it a try06:11
Trevinhojuergh: however I was askinf for other dumps to people having such issues with not that extension06:11
Trevinhojuergh: thanks06:11
juerghTrevinho, I have another dash-to-panel dump. You don't need that, right?06:12
Trevinhojuergh: oh, why not06:13
Trevinhojuergh: feel free to send it06:13
juerghTrevinho, http://kernel.ubuntu.com/~juergh/1714989/06:16
Trevinhojuergh: oh, that was the full crash i thought you menat a jhs_dumpstack06:17
Trevinhobut ok, this should be enough06:17
juerghTrevinho, oh. :-P06:18
Trevinhojuergh: for reference, you were getting the crash ina repeatible way, or just... it happens?06:28
juerghTrevinho, not repeatable but very frequent. there were actually two different types of crashes. 'simple' shell resets and real crashes that resulted in a dump.06:43
juerghI just had another crash and a new dump file. I'll upload it.06:44
Trevinhojuergh: ok, using the patch I did or not?06:44
juerghThis is with your patch. So there might be different issues.06:44
Trevinhook, fair enough, let me know once you've it uploaded06:44
Trevinhojuergh: you reloaded gnome-shell after updating the file, right?06:46
juerghTrevinho, yes06:46
Trevinhojuergh: [or alt+f2 -> r -> enter if in X]06:46
Trevinhook06:46
juerghand it just crashed again. oh boy.06:47
juerghTrevinho, http://kernel.ubuntu.com/~juergh/1714989/_usr_bin_gnome-shell.2000.crash-20171020084206:48
juerghoh hold on.06:48
Trevinhojuergh: oh, if you can attach to gdb would probably be helpful06:48
juerghI think that dump was replaced while I was uploading06:48
juerghTrevinho, ack. let me attach gdb.06:49
Trevinhojuergh: I'm getting 403 on it though06:49
Trevinhoin both06:49
Trevinho(even the old one)06:49
Trevinho+r might be nice :)06:49
juerghTrevinho, maeh. sorry about that. fixed now. but I removed the latest one, not sure if it's in a coherent state.06:59
Trevinhojuergh: well, is that unpacking with apport-unpack?07:06
=== JanC_ is now known as JanC
juerghTrevinho, I don't know what apport-unpack is, so I guess the answer is no? :-)07:07
Trevinhoit just unpacks your crash into usable data07:08
Trevinhojuergh: i mean you can check if it is valid at elast07:08
Trevinholeast*07:08
juerghbtw, I just uploaded a new dumpfile and stacktrace.07:08
juerghthis seems very unstable now...07:09
juerghTrevinho, I reverted your patch. I just had another crash.07:12
Trevinhojuergh: so, the one I tested that was uploaded had the same issue..07:14
Trevinhojuergh: http://kernel.ubuntu.com/~juergh/1714989/stacktrace.201710200903 this was referred to the one with my patch or not?07:16
juerghTrevinho, yes. with your patch.07:17
Trevinhojuergh: can you pass me the file too, so I have proper lines?07:18
Trevinhohome/juergh/.local/share/gnome-shell/extensions/dash-to-panel@jderose9.github.com/windowPreview.js07:18
Trevinhojuergh: oh, i just noticed a typo in my patch07:33
Trevinhojuergh: so the patch you should actually apply is https://pastebin.ubuntu.com/25777475/07:39
Trevinhofor your version I mean07:40
Trevinhojuergh: full file will be https://pastebin.ubuntu.com/25777478/07:40
Trevinhoas extension stock version is different from master in GH07:40
juerghTrevinho, that looks vastly different from github master.07:57
Trevinhojuergh: that's what is on gnome extension store07:58
Trevinhojuergh: when you installed the extension you should have got that version (minus the fix) no?08:01
juerghTrevinho, at first yes. but when it kept crashing I switched to github master.08:03
juerghTrevinho, what do you want me to test?08:21
Trevinhojuergh: oh, ok then try to use git master again with my patch08:32
Trevinhojuergh: and see how it goes08:32
Trevinhojuergh: (new patch, first one had a wrong leftover)08:32
Trevinhowhich didn't cause any fix for sure08:32
juerghTrevinho, just to be sure, you want me to test https://pastebin.ubuntu.com/25777475 ?08:46
=== morphis_ is now known as morphis
=== Logan_ is now known as Logan
dokorbasak, jamespage: are there any transitions you would like to start with for b?09:54
=== morphis__ is now known as morphis
Trevinhojuergh: that if you use the gnome-shell extensions website version, otherwise is https://github.com/jderose9/dash-to-panel/commit/1dfcd3e4b9e23fc2cd104be7a1e12c6fd4f18a46.diff10:53
juerghTrevinho, this looks promising. I haven't had a single reset/crash since 11am'ish.11:45
andreashi, I want to SRU ubuntu-advantage-tools into trusty, xenial and zesty. The current version in artful is just "10", and t/x/z have "2". What should the version "template" be like for t/x/z? For example, considering x, 10~ubuntu0.16.04.1? Or 10~0ubuntu0.16.04.1? Summary: http://pastebin.ubuntu.com/25780812/19:49
andreasslangasek: do you have a tip? ^19:50
johnnyfiveI am writing an implementation of the dpkg-scanpackages and other tools necessary for creating a full index in Go as a library. I am confused in one area. I can't seem to figure out which category/component (xenial-security/main) a package belongs to by reading the meta data from a .deb file. (19:54
sarnoldjohnnyfive: that information is recorded in the apt lists19:55
naccjohnnyfive: it's ot part of the deb19:55
naccyou were told this yesterday19:55
nacc(or whenever it was you asked the same thinng in #ubuntu)19:56
johnnyfivenacc, lol, yes I was, my question has moved on to where that data actually exists, which nobody had an answer for.19:56
sarnoldjohnnyfive: /var/lib/apt/lists/19:56
naccjohnnyfive: you were told where then, too19:56
johnnyfivety19:56
naccjohnnyfive: it's stored in the package lists19:56
johnnyfivenacc, no I was not. sarnold, thanks!19:56
naccjohnnyfive: https://irclogs.ubuntu.com/2017/10/18/%23ubuntu.html#t19:4319:58
nacctook me a second to find it19:58
slangasekandreas: https://wiki.ubuntu.com/SecurityTeam/UpdatePreparation#Update_the_packaging as a guide; I think 10~ubuntu0.16.04.1 follows, but either of those numbers would work and I have no strong feeling20:00
johnnyfivenacc, I stand corrected. I did not see that. However there has to be something upstream that makes those lists, but it's obvious that information is not available here20:01
naccjohnnyfive: aiui, packages in main are seeded (see the seed files) or a dependency of a package in main.20:05
naccjohnnyfive: i believe germinate is what figures out that full list20:06
naccuniverse is basically everything else, multiverse is non-free (from debian) and partner is a separate archive20:06
sarnoldwhat's the goal here? if you want to make your own archive mirror but after rebuilding all the packages, you might have more success with https://www.aptly.info/ or the apt-ftparchive package20:08
naccjohnnyfive: as i said the other day, i thikn you're still better off workingn with ubuntu rather than forking if you think you are solvig a real security problem20:08
naccsarnold: i believe johnnyfive is doig an archive rebuild internally to do something like source-address-randomization20:08
naccwhere by source- i mean compile-time20:08
sarnoldnacc: yeah, thanks for the log link :) quite handy20:08
naccsarnold: aptly is a good reference, though, i forgot about that oe20:09
johnnyfivenacc and sarnold, thanks for the answers. Let me try again. My repo serves 100s of variations of the same package, to different customers. At some point, there will be a need for creating indexes of an entire repo on the fly. This is part of my job, to create these repos.20:10
johnnyfiveand by "create these repos", I mean I'm given an output of /pool/main/<packages>, and I need to then create the appropriate index20:11
johnnyfivehowever, as you stated, there is no meta information on the deb to tell me whether the file should be indexed in /dists/xenial/main or /dists/xenial-updates/main20:11
johnnyfiveif you run dpkg-scanpackages on /pool/main, it just shoves it all in the same index, without breaking it up into the correct categories20:12
naccthe package lists only tell you the components ... you would need to follow the publishing iformation, as i mentioned the other day too20:12
naccor in your mirror + rebuild solution, save off the pocket things are in20:12
sarnoldjohnnyfive: right. the same binary package can be included in lucid, precise, trusty, xenial, zesty, artful, etc.20:12
johnnyfiveso how does ubuntu/debian handle this? is there a build file when they are recreating an entire repo that tells the indexers which files to add to which pocket?20:13
johnnyfiveI don't know what you mean by "publishing information", are you talked about the apt sources list data?20:14
nacchttps://launchpad.net/ubuntu/+source/php7.0/+publishinghistory20:14
nacce.g20:14
sarnoldjohnnyfive: if this were my problem to solve I thnk I'd skip launchpad, skip germinate, etc., and work strictly from the InRelease and Packages files from the mirrors20:15
sarnoldjohnnyfive: stuff in your per-customer md5, sha1, sha256, hashes for all the per-customer rebuilds20:16
johnnyfivesarnold, that's a def possibility. We have a few solutions, just really trying to understand how upstream does it.20:17
naccjohnnyfive: also, just to be clear, you very clearly indicate to your customers they are not running Ubuntu? :)20:18
sarnoldheh I'm sure they pay enough to be sure of that :)20:18
johnnyfiveI assume that during the build process, some database is queried to figure out which indexes it needs to be added to20:18
naccsarnold: :)20:18
johnnyfiveyes, they know exactly what they are getting20:18
sarnoldjohnnyfive: I think https://git.launchpad.net/germinate/tree/README?id=54c933eecf57cfa10526432740cc27f7b7671446 is the starting point..20:19
johnnyfiveThank you both20:21
johnnyfivei'll begin more digging. cheers20:21
sarnold(germinate -might- only do the .iso images! I've been happy to let others manage this ;)20:22
cjwatsonjohnnyfive: germinate is very probably too far back in the chain of events for your needs; for republishing an archive, you're better off looking at the override files in /ubuntu/indices/20:31
cjwatson(germinate is involved, but as the input to human decisions, so you don't want to use it here)20:32
cjwatsonhowever, override files will tell you which component (main, restricted, universe, multiverse) things live in, but not which versions go in which suites (xenial, xenial-updates, zesty, etc.)20:33
cjwatsonfor that, I agree with sarnold - if you want to know what suites packages are published in in bulk, you should look at the published Packages etc. files20:34
sarnoldcjwatson: oh cool, thanks. I've never actually looked in /ubuntu/indices/ before :)20:34
cjwatsonit is possible to derive this from Launchpad queries, but it will be excruciatingly slow to try it that way via the webservice API when you're doing things in bulk20:34
cjwatsonI mean, the Launchpad database is the ultimate source of truth on this and that's what we use, but it will be orders of magnitude faster to get that information from the Packages etc. files instead20:35
cjwatsonbecause you don't have direct database access20:35
Trevinhojuergh: cool, is it going smoothly still?20:37
cjwatsonyou don't need the overrides given the Packages files, though it may make life easier to feed them to apt-ftparchive20:37
johnnyfivecjwatson, awesome, ty for the info21:39

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!