[00:38] bdmurray: where does 'DpkgHistoryLog.txt' come from normally? [00:38] bdmurray: as in what file from the local user's system, or what command generates it? [00:45] slangasek: i was collecting/perusing the bugs we've gotten on php7.0 with failed upgrades. I think in all cases where we see the debconf backtrace ("Use of uninitialized value $reply in scalar chomp at /usr/share/perl5/Debconf/FrontEnd/Passthrough.pm line 66"), the invocation is from aptdaemon and running apt manually at the terminal fixes it. Is aptdaemon used only by update-manager? Or is it used [00:45] eventually by all of the various tools? [00:46] nacc: only by update-manager [00:46] slangasek: ok, good to know [00:46] maybe some other GUI frontends, but certainly not from apt on cli [00:47] slangasek: i tried to reproduce it in a VM and I was unable to in a few attempts, so i'm just trying to think of different approaches [00:47] I seem to regularly hit it here on my laptop whenever I apply updates but have not had the patience to debug it [00:48] slangasek: ok, that's good to know -- and do you see it for multiple packages? I can leave this VM around and see if it triggers it eventually [00:49] nacc: I see it for any package that uses debconf in its maintainer scripts, regardless of whether that would have triggered a debconf prompt to be displayed. [00:49] slangasek: interesting [00:51] this includes kernel upgrades :P [00:51] (which is why I don't have much patience for debugging it) [00:51] slangasek: and that would correspond to what we see as this seems to trigger for some people on the MRE SRUs, where no debconf prompts should occur [01:05] nacc: $ grep DpkgHistory /usr/share/apport/package-hooks/* [01:06] nacc: /var/log/apt/history.log so its incorrectly labelled [01:06] bdmurray: thanks! === salem_ is now known as _salem [06:47] good morning [06:47] o/ [08:16] mdeslaur, i didnt modify anything on purpose, what happened ? [08:41] chrisccoulson: any opinion on bug 1632870 please? [08:41] bug 1632870 in pepperflashplugin-nonfree (Ubuntu Yakkety) "Package is broken since Google stopped shipping Flash with Chrome 54 for Linux" [High,In progress] https://launchpad.net/bugs/1632870 [09:35] rbasak: I was reading the bug you linked and I'm puzzled [09:35] rbasak: I think I get all what was outlined there on pepperflashplugin-nonfree vs partner based adobe-flashplugin [09:36] rbasak: but I'm missing one bit (surely out of ignorance on the history of flash plugins) [09:36] rbasak: but what I did for years is install flashplugin-installer - how does that fit into the puzzle there? [09:37] note: I'm still reading through all the length of this [09:37] if it is down somewhere [09:39] rbasak: found it at about 60% of the bug - that is from ubuntu-restricted-addons [09:47] rbasak: wow, that is a complex maze in this bug - your current question is for your SRU duty to consider accepting them into proposed or not? [10:07] pitti, hi, sync systemd from experimental? :p [10:07] btw seems that you are fixing the autopkgtest failures, right? I see activity con git [10:16] cpaelzer: yeah, I thought chrisccoulson might be an appropriate person to review or comment as he usually uploads adobe-flashplugin. [10:16] But it does look entirely appropriate for an SRU to me. We shipped the package, it's broken, we should fix it. [10:17] And decisions about what to recommend or remove should probably be limited to the development release. [10:17] (where possible) [10:20] rbasak: yeah my opinion (very unimportant) would also to fix it especially as it is a security issue [10:21] rbasak: removing in relase - no-no, removing in zesty or later devel maybe, but even there I don't see the full reason [10:21] rbasak: there are more things which have multiple possible implementations [10:21] Thanks. I'm definitely +1 on fixing it. I'm asking chrisccoulson in case he has any comment or suggestion on the approach taken. [10:21] rbasak: OTOH (for the furture removal in devel) I agree that a removal might simplify this mess at least a bit [10:22] Because as you point out, it's a pretty complex area right now :-/ [10:22] jbicha: thanks for the icoutils sync - had been going to do that this morning :) [10:22] rbasak: sure, waiting for chrisccoulson is better [10:22] rbasak: yet I didn't want to leave your open question for opinions end up totally unconsidered :-) [10:24] LocutusOfBorg: I won't sync it for zesty (feature freeze); if someone wants it and wants to do the FFE, be my guest of course :) [10:25] LocutusOfBorg: I'm only aware of the root-unittests failure on armhf, talked about that with xnox yesterday (no hw access any more, so can't reproduce/investigate myself) === CRogers_________ is now known as CRogers === s1aden is now known as sladen [12:07] ogra_: our techboard meeting keeps getting modified when people modify other meetings on the phone [12:07] ogra_: I assume you meant to modify something else [12:21] rbasak, I don't really have an opinion on pepperflashplugin-nonfree, and that bug is quite long (I've not read the whole thing) [12:22] ideally the package wouldn't exist at all (flashplugin-installer exists so that ubuntu-restricted-addons can recommend it) [12:22] chrisccoulson: that's fine, thank you for looking. I just wanted to check if you had an opinion :) [12:22] chrisccoulson: OOI, how do adobe-flashplugin and flashplugin-installer fit together? [12:23] in fact, I think it's fairly bad that we have a flash plugin installer in the archive that requires users run a command outside of the normal update process to upgrade it [12:24] I wasn't clear on why that is. [12:24] I wonder whether I should just add the ppapi plugin to flashplugin-installer - it has to download it anyway [12:45] doko: llvm 4 & snapshot ftbfs on powerpc, but build fine on debian.. https://launchpadlibrarian.net/305798779/buildlog_ubuntu-zesty-powerpc.llvm-toolchain-4.0_1%3A4.0~+rc2-1_BUILDING.txt.gz === _salem is now known as salem_ === CRogers_________ is now known as CRogers [12:58] doko: any ideas if it's a bug in llvm or something else? [13:31] tjaalton: no, I didn't look. and we'll drop powerpc this cycle [13:31] should have happened in Feb afaik [13:33] ahh [13:33] ok [13:33] forgot about that [14:15] jbicha, hi, what do you mean by tagged the wrong commit? The last commit is "Remove the debian/ directory" in ukui-menu. [14:17] handsome_feng: hi! [14:18] https://github.com/ukui/ukui-menu/releases shows that you tagged 0f027ad but that's not the last commit [14:18] https://github.com/ukui/ukui-menu/commits/master [14:20] I remove the previous releases and update it, since it has the same tag "v1.0.0", and then it shows 10 commits to master since this release... [14:21] handsome_feng: you could use a new tag like 1.0.1? [14:23] jbicha, I have updated the release just now, :) [14:32] handsome_feng: thanks, can I upload to zesty from https://github.com/ukui/debian-packages/tree/master/ukui-menu/debian now? [14:33] jbicha: sure! thank you! [14:39] cyphermox: jbicha: happyaron: Can we talk about the patch to network-manager-applet? It is a single line change to src/libnma/Makefile.am which fixes the ability to use libnma from Python [14:39] jackpot51: don't need to discuss this, let's jfdi ;) [14:40] awesome - then Zesty's installer won't crash for students at university, or corporate customers [14:40] XD [14:40] well, it already shouldn't crash, no? [14:40] It crashes on any WPA 2 enterprise network [14:41] My patch to ubiquity fixes that crash by implementing WPA 2 enterprise [14:42] afaik that doesn't crash, it just doesn't work [14:42] (well, it doesn't work if you try to use the network panel in ubiquity, you already could use nm-applet directly instead [14:42] It will say something like "Installer has crashed. Entering recovery mode" and then go to a desktop [14:42] ok [14:43] There is no try...except around connecting to the Wifi network, so a DbusException from network manager is uncaught [14:44] My patch adds the try..except block in the gtk and kde components, as well as using libnma to support WPA 2 enterprise in gtk components [14:44] Line 63 here: https://git.launchpad.net/~network-manager/network-manager/+git/ubuntu-applet/tree/src/libnma/Makefile.am#n63 [14:44] Needs to change to this: NMA_1_0_gir_INCLUDES = NM-1.0 Gtk-3.0 [14:48] cyphermox: What do I need to do to "jfdi"? :P [14:48] nothing, I have your patch I'll upload it [14:48] Thanks very much! [14:50] cyphermox: thanks [14:51] hi is their any ubuntu devs around [14:52] usually, that's what this channel is [14:53] I've been directed to the right place ty [18:46] hey. ibus seems broken in zesty (can't switch back from layout after changing it). is it officially being phased out? [18:57] salty-horse: no [18:59] jbicha, is any integration done by ubuntu developers, or should I take it up with the ibus team? [19:02] salty-horse: start with a Launchpad bug, but if you think the issue is upstream you can try talking to them too [19:31] rbasak: found a few issues with the -devel pointer changes, testing them now === ddstreet_away is now known as ddstreet === JanC_ is now known as JanC [23:16] slangasek: been doing some digging, cjwatson's comment 10 in LP: #500175 is interesting [23:16] Launchpad bug 500175 in software-center (Ubuntu) "Canceling an installation in Software Center crashes debconf with "Use of uninitialized value $reply in scalar chomp at /usr/share/perl5/Debconf/FrontEnd/Passthrough.pm line 66,"" [Medium,Confirmed] https://launchpad.net/bugs/500175 [23:16] i think it's the most detailed analysis i've found so far [23:18] cjwatson: in case you're interested, context is LP: #1597001 and several php7.0 bugs with the same symptom (all using update-manager) [23:18] Launchpad bug 1597001 in linux (Ubuntu) "package linux-image-4.4.0-28-generic (not installed) failed to install/upgrade: subprocess new pre-installation script returned error exit status 128" [Medium,Confirmed] https://launchpad.net/bugs/1597001 [23:58] rbasak: awesome, so nice to be able to verify i didn't break anything by my fixed changes now [23:58] rbasak: so i just compared a from-scratch import of tftp-hpa to what is already pushed by you [23:59] rbasak: all the series branches match and the devel pointers are different (because they now have a merge-based history)