[00:01] <Bashing-om> rebooting for testing a command in the TTY environment .
[05:02] <lotuspsychje> good morning to all
[05:03] <oerheks> hi lotus
[05:03] <lotuspsychje> hey oerheks
[05:03] <lotuspsychje> where is your O
[05:04] <oerheks> oh, didn' t bother to correct it
[05:04] <lotuspsychje> like it better
[05:05] <oerheks> with the big O ?
[05:06] <lotuspsychje> Oerheks
[06:04] <lotuspsychje> hey EriC^^
[06:04] <EriC^^> hey lotuspsychje
[06:05] <lotuspsychje> all good there EriC^^ ?
[06:05] <EriC^^> yup you?
[06:06] <lotuspsychje> yeah fine here
[06:06] <lotuspsychje> coffee & chill before work
[06:07] <lotuspsychje> !info firefox xenial
[06:07] <ubot5> firefox (source: firefox): Safe and easy web browser from Mozilla. In component main, is optional. Version 57.0+build4-0ubuntu0.16.04.5 (xenial), package size 44608 kB, installed size 170740 kB
[06:07] <lotuspsychje> new logo
[06:18] <EriC^^> yeah firefox quantum
[06:19] <EriC^^> seems ok yeah?
[06:19] <lotuspsychje> yeah im testing it
[06:25] <lotuspsychje> love the loading tabs
[06:51] <sirru5h> Howdy Everyone
[06:51] <jink> Morning.
[06:51] <sirru5h> Hey there jink
[06:52] <sirru5h> Looks like we got firefox 57.0
[06:53] <lotuspsychje> hi jink
[06:53] <lotuspsychje> hi sirru5h
[06:54] <sirru5h> Hey there lotuspsychje how are you doing?
[06:55] <lotuspsychje> http://www.omgubuntu.co.uk/2017/11/firefox-57-quantum-better-faster-stronger
[06:55] <lotuspsychje> great here sirru5h
[06:56] <sirru5h> I just hope it is better than 56 man I had a ton of trouble with that one
[07:01] <sirru5h> At one point i actually removed and reinstalled firefox it was so buggy, weird because it was fine on my ubuntu 17.04 machine
[07:02] <ducasse> good morning, all
[07:02] <sirru5h> Hey there ducasse
[07:03] <ducasse> \o sirru5h
[07:07] <lordievader> Good morning
[07:10] <ducasse> hi lordievader
[07:10] <ducasse> how are you guys?
[07:10] <lordievader> Hey ducasse (IRC)
[07:10] <lordievader> Doing good here, how are you?
[07:12] <jink> Met je rare nick-completion.
[07:12]  * lordievader blames weechat plugin
[07:12] <ducasse> doing well, still waking up :)
[07:12] <jink> lordievader: :P
[07:12] <jink> lordievader: Get a real IRC client, like irssi. ^__^
[07:13] <lordievader> That doesn't support Matrix.
[07:40] <jink> lordievader: wtf is Matrix?
[07:40] <lordievader> No need to swear... https://www.matrix.org/
[07:46] <ducasse> i set up a riot account, but haven't looked at it much - maybe i should...
[07:46] <lotuspsychje> hey lordievader & ducasse
[07:47] <ducasse> good morning lotuspsychje
[07:47] <ducasse> all well?
[07:47] <lotuspsychje> yeah, you?
[07:47] <lotuspsychje> ducasse: new phone tweaked already?
[07:48] <ducasse> still new things to play with :)
[07:50] <lotuspsychje> hmmz every time there's a new firefox it scrambles up my speed dial addon grr
[07:53] <lotuspsychje> and i dont like top websites, as less visited websites are also important
[07:59] <lotuspsychje> oh well, yet another speed dial addon for a change
[07:59] <lotuspsychje> until next version doesnt support it
[08:03] <ducasse> lets hope they don't change the addon stuff again in a while now
[08:05] <lotuspsychje> yeah
[08:12] <lotuspsychje> bbl guys work ; )
[08:12] <lotuspsychje> have a nice day
[08:21] <Bashing-om> !ping
[08:21] <ubot5> pong!
[08:23] <lordievader> Hey lotuspsychje (IRC)
[08:23] <lordievader> How are you doing?
[08:24] <Bashing-om> Ouch . forum suddenly went down - hard !
[08:32] <Bashing-om>  .. back up ,, was not even able to ping :(
[08:40]  * Bashing-om done - eyes too heavy to maintain . good nite \o
[12:37] <BluesKaj> Hiyas all
[15:21] <Guest45412> hi
[15:39] <nicomachus> !info firefox
[15:39] <ubot5> firefox (source: firefox): Safe and easy web browser from Mozilla. In component main, is optional. Version 57.0+build4-0ubuntu0.17.10.5 (artful), package size 45043 kB, installed size 171518 kB
[15:39] <nicomachus> ayyyyy got that update
[16:07] <nicomachus> I guess I should get rid of that firefox-new ppa and revert to the repo version...
[16:31] <BluesKaj> trying out the new kde browser falkon ...reminds me of konqueror by the method used to import bookmarks .... impossible to figure out how to add them to the bookmarks toolbar ...so far it's got me beat..managed to import them, but adding them seems impossible
[16:35] <BluesKaj> the browser is nice tho, it doesn't buffer on some difficult sites due to my slow dsl connection
[16:36] <BluesKaj> buffer video that is
=== Guest7945 is now known as Spydar007
[19:41] <nicomachus> "there is no apt"
[19:41] <nicomachus> -_-
[19:56] <daftykins> o0
=== TJ_Remix is now known as TJ-
[20:54] <TJ-> nacc: can I sound you out on a theory regarding the sudo/nss issue from earlier?
[20:54] <nacc> TJ-: sure
[20:58] <TJ-> I think there may be a bug in the libnss-mdns .postinst script but I want a 2nd opinion. read "man nsswitch.conf", the section on the STATUS and ACTION, in particular for [NOTFOUND=return] ... do you agree "return" suggests no further lookups are done for that database (hosts in our case) e.g. if we have "hosts: mdns4_minimal [NOTFOUND=return] dns" it'll never reach 'dns' if 'hostname' was not found
[20:58] <TJ-> by mdns4_minimal?
[20:59] <TJ-> Because in /var/lib/dpkg/info/libnss-mdns:amd64.postinst it inserts that into the existing nsswitch.conf - I'd expect it to be "[NOTFOUND=continue]"
[21:00] <TJ-> which explains why jmft2 had  'hosts: files mdns4_minimal [NOTFOUND=return] dns myhostname'
[21:00] <daftykins> :)
[21:03] <nacc> TJ-: makes sense, readinng still
[21:03] <TJ-> that doesn't explain why 'files' didn't get it from /etc/hosts (but I assume the default is read as "files [success=continue] or "files [success=merge]"
[21:03] <nacc> TJ-: also, [NOTFOUND=continue] appears to b the default
[21:04] <nacc> TJ-: so i'm not sure why the =return is there at all
[21:04] <TJ-> I'm trying to determine what the default status=action settings are; it's explained in the man page
[21:04] <nacc> TJ-: yep, that's what i'm reading now
[21:04] <nacc> looks to be
[21:04] <nacc> success=return
[21:04] <nacc> notfound=continue
[21:04] <TJ-> nacc: right, you'd expect it'd be [FOUND=return]
[21:04] <nacc> unavail=continue
[21:04] <TJ-> sorry, [success=return]
[21:04] <nacc> tryagain=continue
[21:04] <TJ-> yeah
[21:05] <daftykins> borked=callTJ
[21:05] <nacc> heh
[21:05] <daftykins> :>
[21:05] <TJ-> so, in principle, jmft2's entry with "files" first should have read /etc/hosts, found the entry that matched /etc/hostname, and done [success=return]
[21:05] <TJ-> but it seems like it may have done [success=continue] and then failed on a later look-up
[21:06] <TJ-> if it had reached 'dns' and that called systemd-resolve, then the hostname would have resolved correctly even offline - according to my tests in an 17.10 container when doing "ip link set down dev eth0"
[21:08] <nacc> TJ-: were youa ble to determine for sure that the hostname it was trying to resolve was the one in /etc/hosts?
[21:08] <nacc> TJ-: if that mismatched, for some reason, i could see it falling over to dns
[21:09] <nacc> TJ-: unfort. i think to get any debugging, yhou have to get all the LIBC debugging
[21:09] <TJ-> so... seems like 3 possibilities: 1) nsswitch.conf is being ignored/is corrupt 2) libnss-mdns .postinst script is inserting the wrong [STATUS=action] or 3) there's a bug in the internal glibc nss implementation in 17.10
[21:09] <nacc> TJ-: was this a regression for them on upgrade?
[21:09] <TJ-> fresh install so no idea
[21:09] <nacc> ok
[21:10] <TJ-> the user said they're going to install 16.04 so no ability to further debug it there
[21:10] <TJ-> seems to me there should be a unit-test for this someplace for offline lookups of local hostname
[21:13] <TJ-> grrr! would you believe, for an NSS module NSS_STATUS_NOTFOUND can have two meanings, depending on whether errorno is set to ENOENT or SUCCESS !
[21:13] <TJ-> reading https://www.gnu.org/software/libc/manual/html_node/NSS-Modules-Interface.html#NSS-Modules-Interface
[21:14] <nacc> TJ-: lol
[21:16] <TJ-> aha. default status=action options: https://www.gnu.org/software/libc/manual/html_node/Actions-in-the-NSS-configuration.html#Actions-in-the-NSS-configuration
[21:16] <TJ-> default is "notfound=continue"
[21:17] <TJ-> so that puts suspicion on libnss-mdns .postinst
[21:19] <TJ-> if the nsswitch.conf cannot be read for whatever reason the default is: "For the hosts and networks databases the default value is dns [!UNAVAIL=return] files"
[21:19] <TJ-> which could be what we were seeing - DNS *is* available so it's answer is definitive and 'files' is never used
[21:19] <nacc> yeah, so *possibly* you are getting a parse error?
[21:20] <nacc> and that is coming across as "nsswithc.conf cannnot be read" ?
[21:20] <TJ-> possibly... there were /r/n line endings in /etc/hosts, but the user said /etc/nsswitch.conf didn't have them
[21:20] <nacc> hrm
[21:20] <TJ-> nacc: could be... it's only read once on start-up anyhow
[21:21] <TJ-> oh wow, ouch! just had an apport dialog telling me that ycmd thing has crashed!
[21:21] <TJ-> weird since I've not got any vim sessions going
[21:23] <TJ-> i can't think of a way to prove nsswitch.conf get's read. it'll be too early to set an inotify watch on it, unless it gets read once per process rather than once per boot
[21:25] <TJ-> oh, it's fine, it's once per process. "inotifywait -m /etc/nsswitch.conf" works
[21:28] <nacc> TJ-: yeah, i think something inthe manpage says a given rpocess will only read it once
[21:28] <nacc> so it wo't see updates at runtime
[21:29] <TJ-> "sudo -i" and it reads twice; I assume once for sudo and once for $SHELL
[21:29] <TJ-> it does mean we can experiment by altering the file and re-running commands, rather than needing to reboot
[21:29] <nacc> yeah
[21:29] <TJ-> I've set up "inotifywait -m --timefmt "%H:%M:%S" --format "%T %e %f" /etc/nsswitch.conf" so I get timestamps
[21:30] <TJ-> it's being read but no indication if there's a parse error; not sure about how to detect that
[21:30] <TJ-> I suppose the only way is to write my own debug module of libnss and include it so it logs queries
[21:34] <TJ-> !info libnss-wrapper
[21:34] <ubot5> libnss-wrapper (source: nss-wrapper): NSS wrapper library. In component universe, is optional. Version 1.1.3-1 (artful), package size 26 kB, installed size 86 kB
[21:35] <TJ-> aha
[21:35] <nacc> fun :)
[21:36] <TJ-> fab: "LD_PRELOAD=libnss_wrapper.so  NSS_WRAPPER_HOSTS=/path/to/host ./myapplication"
[21:38] <TJ-> and it also supports NSS_WRAPPER_DEBUGLEVEL=ERROR|WARNING|DEBUG|TRACE
[21:38] <nacc> ah that's great
[21:39] <nacc> so somoene thought of it before :)
[21:39] <TJ-> it seems so; now I need to find a way to reproduce the original failure with sudo so I can use it!
[21:40] <nacc> yeah, i was wondering if you had a testcase already or not
[21:40] <TJ-> first test I guess is to change mdns4_minimal to [NOTFOUND=continue]
[21:41] <TJ-> I doubt it'll happen in a container so I'm trying a VM
[21:41] <nacc> yeah
[21:41] <nacc> and yeah, i think you'd need as close to a real machine env as possible
[21:41] <TJ-> I'll give that a shot in the morning
[21:42] <TJ-> apparently it was a fresh install of the desktop so it should be possible to recreate the same environment. The only thing I didn't find out was how the network was being configured.
[21:48] <TJ-> well B**** me!!! Bug #1295229
[21:48] <ubot5> bug 1295229 in nss-mdns (Ubuntu) "With 'hosts: mdns4' in nsswitch.conf, getaddrinfo() returns -5 (EAI_NODATA) when network interface is down" [Undecided,New] https://launchpad.net/bugs/1295229
[21:49] <daftykins> Bind9? ;)
[21:50] <nacc> TJ-: interesting that it was filed so long ago
[21:50] <nacc> and no traction
[21:50] <nacc> i wonder why it's not seen more commonly
[21:51] <TJ-> I wonder if that's the bug that prompted [NOTFOUND=return] ... but the meaning of that was misunderstood and the maintainer meant to have it do [NOTFOUND=continue]
[21:52] <TJ-> oh, no, they show it in their description.
[21:52] <TJ-> nice; someones written me a testcase :)
[21:53] <TJ-> I'll test against 14.04 and 16.04 now I've seen that
[21:54] <TJ-> Tackle it tomorrow with some positive pointers