=== maskie [~maskie@196-30-108-239.uudial.uunet.co.za] has joined #ubuntu-meeting | ||
=== pitti [~martin@box79162.elkhouse.de] has joined #ubuntu-meeting | ||
=== pitti [~martin@box79162.elkhouse.de] has joined #ubuntu-meeting | ||
=== pitti [~martin@box79162.elkhouse.de] has joined #ubuntu-meeting | ||
=== jiyuu0 [~jiyuu0@219.95.210.165] has joined #ubuntu-meeting | ||
=== maskie [~maskie@196-30-109-144.uudial.uunet.co.za] has joined #ubuntu-meeting | ||
=== jiyuu0 [~jiyuu0@219.95.32.148] has joined #ubuntu-meeting | ||
=== ogra_ [~ogra@p508EB646.dip.t-dialin.net] has joined #ubuntu-meeting | ||
=== Keybuk [scott@descent.netsplit.com] has joined #ubuntu-meeting | ||
=== zul [~chuck@zul.developer.gentoo] has joined #ubuntu-meeting | ||
=== jiyuu0 [~jiyuu0@219.95.32.148] has joined #ubuntu-meeting | ||
=== mvo__ [~egon@ip181.135.1511I-CUD12K-01.ish.de] has joined #ubuntu-meeting | ||
=== ajmitch [~ajmitch@chc11-port78.jetstart.maxnet.co.nz] has joined #ubuntu-meeting | ||
=== elmo [~james@83-216-141-215.jamest298.adsl.metronet.co.uk] has joined #ubuntu-meeting | ||
=== lamont_r [~lamont@dsl-140-170.dynamic-dsl.frii.net] has joined #ubuntu-meeting | ||
=== mdz [~mdz@69-167-148-207.vnnyca.adelphia.net] has joined #ubuntu-meeting | ||
fabbione | (sc | 04:52 |
---|---|---|
fabbione | argh | 04:52 |
=== ..[topic/#ubuntu-meeting:mdz] : Tuesday 25 January 2005 16:00 UTC: Community Council meeting -- https://www.ubuntulinux.org/wiki/CommunityCouncilAgenda. Tuesday 18 January 2005 1600UTC: Technical Board -- http://www.ubuntulinux.org/wiki/TechnicalBoardAgenda. || This is NOT #ubuntu, nor #ubuntu-devel | ||
fabbione | hi * | 04:52 |
ajmitch | hello | 04:53 |
lamont_r | morning | 04:53 |
=== ogra [~ogra@p508EB646.dip.t-dialin.net] has joined #ubuntu-meeting | ||
zul | hey fabbione | 04:54 |
=== lamont_r makes a note to not reschedule holidays onto tuesdays ever again. :-( | ||
=== sabdfl [~mark@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-meeting | ||
=== silbs [~sbsm0084@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-meeting | ||
=== haggai [~halls@i-83-67-20-196.freedom2surf.net] has joined #ubuntu-meeting | ||
=== amu [~amu@amu.developer.debian] has joined #ubuntu-meeting | ||
=== silbs [~sbsm0084@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-meeting | ||
mdz | welcome, everyone | 04:59 |
mdz | time to get started | 04:59 |
mdz | hmm, we lost sabdfl | 05:00 |
Keybuk | heh | 05:00 |
=== jbailey [~jbailey@CPE000ded9d787c-CM014260028338.cpe.net.cable.rogers.com] has joined #ubuntu-meeting | ||
Keybuk | quick, let's finish before he gets back! :D | 05:00 |
mdz | Riddell: are you here? | 05:00 |
Riddell | hi | 05:00 |
Riddell | should we wait for sabdfl? | 05:01 |
mdz | Riddell: so you are currently a Member? | 05:01 |
ogra | rather commiter | 05:02 |
Riddell | mdz: I'm currently a member with special privilages to upload to universe | 05:02 |
Kamion | ogra: member's the relevant bit; implies he's passed CC | 05:02 |
mdz | ok, so CC signoff already happened | 05:02 |
elmo | not for full maintainership | 05:02 |
ogra | Kamion: yeah, true | 05:02 |
mdz | I don't think we need to go back to CC | 05:02 |
mdz | but mako has been slack in documenting the process ;-P | 05:03 |
elmo | err, yeah we do | 05:03 |
elmo | becoming a universe person requires sign off from only two CC | 05:03 |
elmo | becoming a maintainer is meant to require sign off from full CC and TB, AIUI | 05:03 |
mdz | Member is a community council vote | 05:03 |
=== sabdfl [~mark@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-meeting | ||
sabdfl | hi all | 05:04 |
mdz | welcome back | 05:04 |
Riddell | I'm working on kubuntu so if kdelibs and other bits go into main it would be useful if I could keep working on them | 05:05 |
mdz | we were just arguing about the still-undocumented new membership process | 05:05 |
sabdfl | hmm... i spec'd this out and mailed launchpad and mako | 05:05 |
sabdfl | happy for it to go onto the wiki for further refinement | 05:05 |
sabdfl | what's the question? | 05:05 |
ogra | Riddell for main | 05:05 |
mdz | it's already in the wiki | 05:06 |
mdz | it's been refined, and I think it needs to be properly documented on the main website | 05:06 |
mdz | the notes are here: http://www.ubuntulinux.org/wiki/NewDevelopersAndMaintainers | 05:06 |
mdz | they do not include the modifications for the accelerated universe-only process | 05:06 |
mdz | which Riddell has been through, iirc | 05:06 |
sabdfl | neither did my launchpad spec | 05:06 |
ogra | wasnt that a one timer (riddell aqnd me for universe) ? | 05:07 |
sabdfl | i tihnk we agreed that uploading permission could be component and package | 05:07 |
ogra | at least thats how i understood it | 05:07 |
Kamion | ogra: no, we were trying to define a longer-term process | 05:07 |
sabdfl | so riddell could hapily be given permission to upload to universe *and* specific packages, even if those migrate to main | 05:07 |
ogra | ah, ok, so i misunderstood then | 05:07 |
sabdfl | of course, i hope riddell would qualify for main uploading soon too | 05:07 |
Riddell | (I have been making quite a few mistakes in the packages I've uploaded) | 05:08 |
ajmitch | what is the accelerated universe-only process? I'm sure there are plenty of others (like me) who'd like to do that too | 05:08 |
sabdfl | ajmitch: find two TB / CC / MOTU folks to approve you | 05:08 |
ajmitch | ok | 05:08 |
sabdfl | ajmitch: sign the CoC | 05:08 |
amu | sabdfl: i would suggest Riddell and give him some help | 05:09 |
sabdfl | andyou'reoff | 05:09 |
ogra | ajmitch: http://www.ubuntulinux.org/wiki/NewDevelopersAndMaintainers | 05:09 |
mdz | that wiki page is just a bunch of notes from the BoF | 05:09 |
ajmitch | ogra: isn't that for full membership? | 05:09 |
mdz | and doesn't document the process as it stands | 05:09 |
ogra | mdz: but gives some hints at least | 05:09 |
Riddell | ogra: masters of universe is not mentioned there | 05:09 |
ogra | true | 05:09 |
amu | so i vouch for Riddell :) | 05:10 |
sabdfl | i'll second riddell | 05:11 |
ogra | am i allowed to vote ? | 05:11 |
sabdfl | riddell, can you get a signed CoC to elmo please? | 05:11 |
=== fabbione thumbs up | ||
Riddell | sabdfl: I sent one to mako last week | 05:11 |
elmo | sabdfl: he's sent a signed CoC to mako | 05:11 |
ogra | if so +1 for riddell | 05:11 |
elmo | sabdfl: he can already upload to universe | 05:11 |
elmo | sabdfl: what we're talking about here is full maintainership | 05:11 |
sabdfl | ok rock, thank you | 05:11 |
sabdfl | ah | 05:11 |
elmo | as in, unrestricted uploads to main | 05:12 |
ogra | sabdfl: you approved us both last week ;) | 05:12 |
sabdfl | Riddell: can we stick with universe + specific packages for a while? | 05:12 |
Riddell | sabdfl: sure | 05:12 |
sabdfl | cool | 05:12 |
mdz | if it comes to a vote today, I'd have to plead insufficient information | 05:12 |
sabdfl | can we move on? | 05:12 |
Riddell | anything beginning with a k would be fine | 05:12 |
ogra | heh | 05:12 |
fabbione | Riddell: Kernel? | 05:12 |
mdz | sabdfl: I think so | 05:13 |
lamont_r | fabbione: starts with an L | 05:13 |
Riddell | fabbione: it's called linux in ubuntu | 05:13 |
sabdfl | :-) | 05:13 |
fabbione | damn.. it didn't work :P | 05:13 |
sabdfl | mail storage standardisation | 05:13 |
sabdfl | everyone read the proposal? | 05:13 |
=== lamont_r read it, kind of favors the counter proposal | ||
mdz | who submitted the proposal? | 05:14 |
sabdfl | i did | 05:14 |
=== Kamion also favours the counter-proposal; my local experimentation with maildir didn't endear me to it | ||
=== sladen reads it two days ago | ||
fabbione | sabdfl: this is a religious topic | 05:14 |
sabdfl | right now we have no answer | 05:14 |
sabdfl | even removing postfix doesn't solve the deeper problem | 05:14 |
sabdfl | someone who WANTS mail delivery doesn't get something that Just Works | 05:14 |
lamont_r | sabdfl: true. | 05:14 |
Kamion | mutt is missing several advanced features when forced to use maildir | 05:14 |
sabdfl | it doesn't even Just Work badly | 05:14 |
mdz | the informal discussions that we have had on this subject lead to the irrelevance of Unix-style mail in the desktop | 05:14 |
pitti | I think we should implement the proposal | 05:14 |
lamont_r | mdz++ | 05:14 |
sabdfl | ubuntu is not jsut for desktops | 05:15 |
=== pitti votes for proposal _and_ removing postfix from default install | ||
Kamion | however, the proposal and counter-proposal seem orthogonal to me | 05:15 |
lamont_r | sabdfl: and apache2 isn't installed in base | 05:15 |
lamont_r | Kamion: I have to agree with you | 05:15 |
sabdfl | i have had concerns with postfix since before warty | 05:15 |
mdz | it is straightforward to have SMTP and IMAP services Just Work after installing dovecot and postfix | 05:15 |
sabdfl | this is orthogonal, as kamion points out | 05:15 |
mdz | using a traditional mbox spool | 05:15 |
lamont_r | I don't think that a crippled postfix install in base is the right answer | 05:15 |
lamont_r | which is what we have in warty | 05:15 |
sladen | Even maildir were promoted, I like the naming conventions in the proposal. But not sure about the wider topic of Maildir vs. mbox | 05:16 |
sabdfl | lamont_r: agreed | 05:16 |
Kamion | note that shifting to .maildir has implications for NFS-home folks | 05:16 |
lamont_r | sabdfl: but uncrippling postfix requires taht it ask questions, which is why we got where we are | 05:16 |
sabdfl | there are three scenarios on a typical box: | 05:16 |
sabdfl | - no mail needed locally | 05:16 |
mdz | shifting to delivery-into-homedir has impliications for NFS-home folks | 05:16 |
Kamion | although I do like having SOME kind of maildir naming conventions | 05:16 |
sabdfl | - basic mail needed locally | 05:16 |
sabdfl | - sophisticated requirements | 05:16 |
Kamion | mdz: indeed | 05:16 |
sabdfl | at the moment we address none of these | 05:17 |
sabdfl | Kamion: nfs automatically promotes you to sophisticated | 05:17 |
lamont_r | sabdfl: at the moment, the requirement to not listen on any ports by default gives you only 'no mail needed locally' | 05:17 |
sabdfl | anyone who is doing that can setup the configs to suit them | 05:17 |
lamont_r | so why do we install a daemon by default? | 05:17 |
fabbione | sabdfl: we can't address them in a clean way. as soon as we push one way the other 2 will suffer imho | 05:17 |
pitti | What do you think about Keybuk's local-only delivery script? I think that's a nice default installation | 05:17 |
sabdfl | once again, this does not solve the problem | 05:17 |
sladen | lamont_r: except root messages -> first user | 05:17 |
Kamion | sabdfl: yes; but we still have to not gratuitously break things for those folks | 05:17 |
lamont_r | pitti: I like that | 05:17 |
sabdfl | if i drop three pieces of mail infrastructure down right now it does not just work | 05:18 |
pitti | then postfix could install as a full MTA by default | 05:18 |
lamont_r | sladen: that's Keybuk's local delivery thing | 05:18 |
lamont_r | pitti: yes | 05:18 |
sabdfl | and mbox, while it has some advantages for a single app environment, does not work as well when you have multiple apps all sharing the data | 05:18 |
sladen | pitti: like that | 05:18 |
mdz | sabdfl: why doesn't it work? is there a problem other than postfix being crippled by default? | 05:18 |
sabdfl | Kamion: nothing i'm proposing prevents local configuration to suit a sysadmin | 05:18 |
Keybuk | the trouble is, if you want local mail delivery, you almost certainly have an unwavering love of one particular MDA | 05:18 |
sabdfl | mdz: yes | 05:19 |
mdz | (with dropping the three pieces together) | 05:19 |
Kamion | sabdfl: ok | 05:19 |
Keybuk | and you're going to want to remove postfix and install that instead | 05:19 |
sabdfl | i would like it to work as follows | 05:19 |
lamont_r | Keybuk: iff it's different, of course. | 05:19 |
pitti | mdz: can we agree to discuss both things separately? | 05:19 |
sabdfl | 1. we agree on format and naming | 05:19 |
lamont_r | pitti++ | 05:19 |
mdz | pitti: I'm trying to determine whether they're closely related | 05:19 |
sabdfl | 2. we setup ALL our mail components in main (and possibly the bigger ones in universe) to work that way out of the box | 05:20 |
pitti | thing 1: default installation; thing 2: make installations work out of the box | 05:20 |
sabdfl | this allows someone to remove postfix and install exim and have everything else continue to Work | 05:20 |
pitti | mdz: they are completely orthogonal | 05:20 |
sabdfl | then remove dovecot and install UW-Imap and have it Just Work | 05:20 |
mdz | pitti: not if the only reason it doesn't work out of the box is that postfix is crippled | 05:20 |
Keybuk | sabdfl: except if they do that, they have to remove ubuntu-desktop | 05:20 |
mdz | because this works in Debian. trivially. | 05:20 |
sabdfl | then replace mutt with Evo and have it Just Work | 05:20 |
Kamion | Keybuk: ubuntu-base | 05:20 |
pitti | sabdfl: ++ | 05:20 |
Keybuk | sorry, ubuntu-base | 05:20 |
sabdfl | i've been digging into this and the trend towards maildir is clear | 05:21 |
mdz | Keybuk: which is exactly the right thing, if you think about it | 05:21 |
Keybuk | I agree Maildir is better than mailbox, and if we go Maildir we should use $HOME/Maildir | 05:21 |
sabdfl | Keybuk: have you read the proposal? | 05:21 |
mdz | because if we decide to switch to a different MTA, they don't want it | 05:21 |
Keybuk | sabdfl: yes, I did | 05:21 |
sabdfl | ~/Maildir is b0rked, because it makes something visible that should not be so | 05:22 |
Keybuk | sabdfl: we could ship a default .hidden file so it's not visible in Nautilus | 05:22 |
mdz | sabdfl: what about ~/Desktop? | 05:22 |
sabdfl | Keybuk: would you like chewing gum with that? <duck> | 05:22 |
mdz | ~/public_html... | 05:22 |
pitti | The only thing I did not like in the proposal were the dot-file folders | 05:22 |
sabdfl | mdz: remember Oxford? i still feel the same way :-) | 05:22 |
Keybuk | sabdfl: it's better than changing *every* *single* MTA/MDA and MUA to respect ~/.maildir | 05:22 |
sabdfl | Keybuk: trivial changes | 05:23 |
sabdfl | i've done it on my own system and it's straightforward | 05:23 |
sladen | Keybuk: trivial changes x10 && have everything work Just So vs. No changes and have things sometimes work (I prefer the former) | 05:23 |
Keybuk | I'd rather we didn't ship an MDA in base at all | 05:24 |
Keybuk | and let people who want local delivery install one | 05:24 |
sabdfl | Keybuk: agreed, however, when you install it it should work | 05:24 |
pitti | Keybuk: but many packages depend on an mta | 05:24 |
Keybuk | sabdfl: oh, most definitely | 05:24 |
Keybuk | why doesn't postfix? I thought Debian policy required both MTAs and MUAs to respect /var/mail/$USER ? | 05:25 |
Kamion | if we change MUAs to look in ~/.maildir we should arrange that they still look in ~/Maildir if ~/.maildir isn't there | 05:25 |
pitti | cron, for example | 05:25 |
sabdfl | Kamion: +1 | 05:25 |
Keybuk | pitti: trivial to change, make it not mail if there's no sendmail | 05:25 |
mdz | if postfix doesn't simply work out of the box, we've broken it, because it works in Debian | 05:25 |
Keybuk | it seems to deliver properly here? | 05:25 |
lamont_r | mdz: no listening leads to a configuration that most people don't want | 05:25 |
sladen | Keybuk: does your local-delivery provide a sendmail -t work a like? | 05:26 |
mdz | lamont_r: but local mail delivery works | 05:26 |
lamont_r | yes | 05:26 |
Keybuk | if I type "mutt", I get a mailbox full of debconf notes | 05:26 |
pitti | Keybuk: agreed | 05:26 |
mdz | likewise | 05:26 |
=== seb128 [~seb128@ANancy-151-1-28-86.w83-194.abo.wanadoo.fr] has joined #ubuntu-meeting | ||
mdz | and surely dovecot exports /var/mail by default | 05:26 |
elmo | dovecot does automagic | 05:27 |
elmo | to try and determine what you're doing | 05:27 |
mdz | I'd like to consider separately the "make IMAP service just work" and "move to maildir" items | 05:27 |
pitti | so does anybody actually want an installed default MTA on a desktop? | 05:27 |
mdz | I don't | 05:27 |
lamont_r | pitti: no | 05:27 |
sladen | no | 05:27 |
mvo__ | no | 05:27 |
lamont_r | local MDA has uses, but no MTA | 05:28 |
pitti | so the first step seems to be obvious, right? | 05:28 |
pitti | remove postfix from desktop and uncripple it again | 05:28 |
Keybuk | mdz: yup (just tried it) | 05:28 |
Kamion | removing postfix from base without replacing it with anything means no /usr/sbin/sendmail, which I'm fairly sure will break some things during installation | 05:29 |
mdz | so we seem to have three distinct issues: | 05:29 |
Kamion | (I might be wrong though) | 05:29 |
=== doko [~doko___@dsl-082-082-208-171.arcor-ip.net] has joined #ubuntu-meeting | ||
mdz | 1. Standardize on maildir?, 2. Have no MTA in base?, 3. have mail service just work? | 05:30 |
pitti | Kamion: Keybuk's MDA would work for this | 05:30 |
mdz | but 3. seems to be a non-issue at present | 05:30 |
mdz | Kamion: we would fix those things | 05:30 |
pitti | mdz: 3 for bendy? | 05:30 |
mdz | the most problematic, I think, is LSB | 05:30 |
mdz | LSB requires an MTA | 05:30 |
pitti | really? | 05:30 |
pitti | not just an MDA? | 05:30 |
mdz | lsb Depends: mail-transport-agent | 05:31 |
Kamion | pitti: indeed, if it existed :-) | 05:31 |
fabbione | Kamion: i am pretty sure a bunch of pkgs will fail to install without an MTA | 05:31 |
sabdfl | 3 is an issue | 05:31 |
mdz | sabdfl: using the standard spool and installing dovecot seems to work fine | 05:31 |
sladen | any idea what Apple did. They ship Postfix but I've no idea how it's configured | 05:31 |
sabdfl | mdz: you only use mutt | 05:31 |
pitti | Keybuk: "existed" -> Keybuk's MDA? it does exist, I already saw it | 05:31 |
sabdfl | how can we work towards widespread recognition of ubuntu as a mail server if our mail components don't work together? | 05:31 |
mdz | sabdfl: meaning that you are considering the case where the system is both a mail server and a desktop system where people read mail? | 05:31 |
sabdfl | yes | 05:32 |
mdz | rather than a traditional mail server | 05:32 |
mdz | is that a case we need to optimize for? | 05:32 |
sabdfl | fundamentally, we should make sure that any combination of components has the best chance of working, and working well | 05:32 |
sabdfl | of course some experts will have preferences and will customise their systems | 05:32 |
Keybuk | sabdfl: have you got any two examples where mail components *don't* work together? Because it's Debian policy that they must, so if they don't, we've broken it | 05:32 |
mdz | Keybuk: I think he's talking about, e.g. evolution not reading the spool | 05:33 |
Kamion | Debian policy requires that stuff works together on the inbox, but not on other folders | 05:33 |
Keybuk | hmm, that's actually really really hard to do with Evolution | 05:34 |
Keybuk | because Evolution doesn't behave like that | 05:34 |
Keybuk | the "Local" thing is actually a special thing built-in to evo | 05:34 |
sabdfl | evo does do maildir | 05:34 |
Keybuk | and you really can't persuade it to read /var/mail/$USER instead, without adding an account | 05:34 |
Keybuk | sabdfl: yeah, but *only* if you add an Account | 05:34 |
Keybuk | and that has to have things like a real name, and e-mail address associated with it | 05:34 |
ogra | and a outgoing mailserver | 05:35 |
=== silbs [~sbsm0084@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-meeting | ||
sladen | Keybuk: could those be created by adduser/useradd when copying skel ? (GECOS and $USER@localhost) | 05:37 |
Keybuk | sladen: meh, it'd involve adding .gconf and stuff | 05:37 |
=== sabdfl [~mark@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-meeting | ||
Keybuk | welcome back Lunchpad | 05:38 |
sabdfl | ;-) | 05:38 |
mdz | sabdfl: should we be considering this in a Hoary context? | 05:39 |
sabdfl | any further comments? | 05:39 |
sabdfl | mdz: your call, bendy for certain | 05:39 |
=== jiyuu0 [~jiyuu0@219.95.215.221] has joined #ubuntu-meeting | ||
mdz | if I have a choice in the matter, this would be a bendy feature | 05:40 |
mdz | we don't seem to have a consensus yet | 05:40 |
mdz | we should have more discussion, and if we can do it for hoary, we will | 05:40 |
mdz | but I don't think we can commit to it yet | 05:41 |
sabdfl | ok | 05:41 |
mdz | I think that maildir is sane | 05:41 |
pitti | mdz: hmm, at least the default MTA should be settled for Hoary IMHO | 05:41 |
lamont_r | mdz: is this your #1 & 3? | 05:41 |
Keybuk | I like the idea of moving from /var/mail/$USER to $HOME/Maildir, I don't like the idea of changing that hugely-standard location (though I can see why people might desire it) | 05:41 |
mdz | pitti: agreed | 05:41 |
mdz | lamont_r: this is regarding #3 | 05:41 |
mdz | I have no concrete argument with maildir as a standard mail store | 05:42 |
sabdfl | ok. should we move on to ogra? | 05:42 |
lamont_r | how about mdz's #2? | 05:42 |
mdz | as to whether we should go forth and modify the defaults of all mail software in Ubuntu, that needs to be considered as a release goal in my opinion | 05:42 |
sabdfl | mdz: will need quite a bt of testing to make sure we don't break the GoldenRule | 05:42 |
mdz | #2 was: should we remove the MTA from base in hoary? | 05:42 |
sabdfl | new conf only | 05:43 |
lamont_r | yes | 05:43 |
mdz | I think we should | 05:43 |
lamont_r | mdz: yes it was, and yes we should | 05:43 |
mdz | anyone else have an opinion or comment? | 05:43 |
sabdfl | Kamion: are you happy we can remove postfix and have the installer still work? | 05:43 |
Keybuk | I think no MTA or MDA (not even my bit of Python), fix cron and debconf to only send mail if there's sendmail on the system -- gives server folk free reign to pick an MTA/MDA of choice without conflict | 05:43 |
lamont_r | the question that then comes up is: do we provide some thing (ala Keybuk's code), or not? | 05:43 |
mdz | Keybuk: to be honest, I think mail is the wrong thing for debconf anyway | 05:44 |
sabdfl | Keybuk: does your code have a life outside of birmingham? | 05:44 |
mdz | Keybuk: we should turn those into the hooks that mvo and I discussed | 05:44 |
Keybuk | http://people.ubuntu.com/~scott/sendmail.py | 05:44 |
sabdfl | in the sense of a community, outside interest, outside use? | 05:44 |
pitti | Keybuk: two issues: LSB and how people get notified about failed crons and debconf stuff= | 05:44 |
Keybuk | was written in Matar, and I mailed it to -devel when we first had this discussion :) | 05:44 |
sabdfl | right | 05:44 |
sabdfl | what about nullmail and similar local-only mailers? | 05:45 |
Keybuk | pitti: well, the problem there is people don't, because they don't know they even have local mail until they exit ssh or something :p | 05:45 |
Kamion | sabdfl: I'd only be confident about that with something else providing /usr/sbin/sendmail in its place | 05:45 |
Keybuk | nullmailer is a transfer agent, not a local-delivery agent fwir | 05:45 |
sabdfl | ok | 05:45 |
mdz | the basis for my opinion about removing it is that having local-only delivery isn't really useful | 05:45 |
lamont_r | mdz++++ | 05:45 |
mdz | sendmail.py doesn't really address that | 05:45 |
mdz | for Warty, we considered ourselves to be "stuck" with having local mail, because things like cron, at, debconf, etc. relied on it | 05:46 |
lamont_r | well, it provides local-only delviery, which we agreed isn't really useful | 05:46 |
mdz | but for hoary, we have the freedom to consider those bugs if needed | 05:46 |
lamont_r | bugs | 05:46 |
pitti | Keybuk: okay, agreed. But still, LSB cries for sendmail | 05:46 |
Kamion | Keybuk: debconf just calls /usr/bin/mail -s, and tolerates it being absent | 05:46 |
Keybuk | pitti: LSB cries for RPM too, doesn't it? | 05:46 |
lamont_r | mdz: although I have personal religious issues with all the packages that Depend: exim4 | mail-transport-agent | 05:47 |
pitti | Keybuk: we have alien :-) | 05:47 |
mdz | Kamion: perfect | 05:47 |
pitti | Keybuk: it's installed by default just to satisfy LSB, I think | 05:47 |
sabdfl | lamont_r: we can fix 'em ;-) | 05:47 |
Kamion | although I think the GNOME frontend might pop up an error dialog box in that case; we'd want to tweak that | 05:47 |
mdz | pitti: yes, indirectly | 05:47 |
lamont_r | sabdfl: thanks. I'll quote you on that. :-) | 05:47 |
azeem | "use exim4 or a real MTA" | 05:47 |
mdz | Kamion: if and when the GNOME frontend becomes viable | 05:47 |
sabdfl | mdz: isn't it useful to think that a local mailer allows debconf etc to continue as they currently do, and that if our mail components work together then ANY mail client will immediately see that mail? | 05:48 |
mdz | so it sounds like at most, we should only need to fix cron and at | 05:48 |
mdz | and possibly sudo | 05:48 |
lamont_r | mdz: changing Depends: mta to Recommends: mta??? | 05:49 |
ogra | is all this easy addressable on upgrades from warty ? | 05:49 |
lamont_r | mdz: and then postfix* move from base to ship | 05:49 |
mdz | sabdfl: I don't have a use case which isn't better satisfied by other functionality in the system | 05:50 |
mdz | debconf notes _should_ be blackholed by default unless a sysadmin has configured mail, because they only confuse users | 05:50 |
mdz | ones which are relevant to users should become post-install hooks | 05:50 |
mdz | ogra: warty upgrades would retain postfix | 05:50 |
mdz | though the release notes would recommend removing it if it's unused | 05:50 |
ogra | ah, ok...but that means i get a crippled hoary then | 05:50 |
mdz | cron errors _should_ be blackholed unless a sysadmin has configured mail, because those confuse users even more | 05:50 |
Kamion | erk, users wouldn't get errors from their crontabs? | 05:51 |
Kamion | I think that will be a support burden | 05:51 |
sabdfl | Kamion: agreed | 05:51 |
Kamion | "where did the errors go?" "uh ... nowhere" | 05:51 |
mdz | I think you guys are talking about a different kind of "user" | 05:51 |
Keybuk | Kamion: if users are writing crontabs, they're likely to be the kind of user who's installed an mda of their own | 05:51 |
lamont_r | ogra: how so? If you prefer a different MTA, then once you upgrade to hoary, installing it (and removing postfix) would not remove ubuntu-base | 05:51 |
mdz | a user who can write crontabs can configure mail, too | 05:51 |
Kamion | I think throwing away errors is a bug, period | 05:51 |
Kamion | mdz: yes but if they're coming from another system they might not realise that they need to | 05:52 |
ogra | erm, dont cron errors already go in syslog | 05:52 |
jbailey | A user who can write crontabs might not have permission to install mail. | 05:52 |
Kamion | ogra: news to me if they do | 05:52 |
mdz | jbailey: a user without permission is a client on a server, and needs to ask their admin to set things up | 05:52 |
mdz | ogra: no, they don't | 05:52 |
Kamion | if we're disabling mail, we need to come up with somewhere else to put cron errors | 05:52 |
jbailey | Just refering back to the support nightmare. | 05:52 |
mdz | Kamion: crontab could emit a warning | 05:52 |
Kamion | syslog would be fine | 05:52 |
mdz | "If you want to receive errors, install a mail-transport-agent" | 05:53 |
Kamion | although potentially an information leak, actually | 05:53 |
sabdfl | Kamion, jbailey: agreed, I think a minimal local-delivery-only agent is the answer | 05:53 |
mdz | syslog would need to be root-only | 05:53 |
Kamion | ugh | 05:53 |
Kamion | systems like that are such a pain | 05:53 |
Kamion | you can never find out what's going on | 05:53 |
sladen | AFAICT, the wording is ''requires /usr/sbin/sendmail *if* the system provides a sendmail compatible MTA'' ... doesn't say anything about /not/ providing a sendmail-compatible | 05:53 |
mdz | local unix mail is a wart on our desktop, and an annoyance on our server | 05:53 |
Kamion | mdz: crontab emitting a warning would be a reasonable option | 05:53 |
lamont_r | crontab emit a warning | 05:53 |
jbailey | sladen: Where are you looking? Chapter 3 seems to say that Table 3-1 lists the Commands and Utilities required to be present on a conforming system. sendmail is in the list. | 05:54 |
sabdfl | mdz: any serious MTA would replace it, right? | 05:54 |
sabdfl | this gives us crontab behaviour one would expect | 05:54 |
sabdfl | we can improve the debconf msgs so they are less confusing, or hook them as you described | 05:54 |
mdz | sabdfl: a serious unix mail user either wants postfix (and is thwarted by our crippled default configuration) or wants something else anyway | 05:55 |
sabdfl | thwarted? surely it would be a drop-in replacement? | 05:55 |
lamont_r | sabdfl: crontab would only emit the warning if /usr/sbin/sendmail was absent | 05:55 |
mdz | the hook stuff, by the way, is: http://www.ubuntulinux.org/wiki/InteractiveUpgradeHooks | 05:55 |
Kamion | mdz: s/user/admin/; I've never had a need to care what MTA was installed on a system where I was a mere user | 05:55 |
lamont_r | sabdfl: they expect the config to work, and the crippled default config confuses the hell out of most of the less expert users | 05:55 |
Kamion | well, not beyond knowing where the logs were | 05:55 |
mdz | Kamion: agreed | 05:55 |
sladen | jbailey: that was FHS, I guess LSB ch.3 supersedes that... | 05:56 |
mdz | mail logs aren't generally readable by unprivileged users anyway | 05:56 |
Kamion | that's a bug :-) | 05:56 |
ogra | from my woody server: /var/log/syslog.0:Jan 18 06:23:01 aleph /USR/SBIN/CRON[24496] : (mail) CMD ( if [ -x /usr/sbin/exim -a -f /etc/exim/exim.conf ] ; then /usr/sbin/exim -q ; fi) | 05:56 |
ogra | it obviously is possible | 05:56 |
mdz | sabdfl: if they want postfix, postfix is already there (but broken), so they can't just install it | 05:56 |
mdz | sabdfl: if they want something else, they need to install it anyway | 05:56 |
mdz | so the presence of a crippled postfix in base doesn't satisfy either use case | 05:56 |
sabdfl | mdz: i was proposing a simple local-only mta after install, and a proper postfix on request | 05:57 |
mdz | so if it isn't relative to a sysadmin, and isn't relevant to a desktop user, I think we should remove it | 05:57 |
mdz | sabdfl: that addresses those two problems, yes | 05:57 |
sabdfl | local-only being keybuks or similar | 05:57 |
mdz | but not the problem of local-only mail in the desktop | 05:57 |
mdz | which is that it grows until it consumes the disk | 05:57 |
lamont_r | sabdfl: the issue there is that installing _any_ mta would cause ubuntu-base to be removed | 05:57 |
sabdfl | hm... | 05:58 |
mdz | lamont_r: if that's the only issue, we can address it | 05:58 |
lamont_r | mdz: well, one of several, as you note | 05:58 |
sabdfl | lamont_r: even if ubuntu-base depends on something provided by both the local-only and postfix / exim / sendmail? | 05:58 |
lamont_r | sabdfl: that could work | 05:58 |
Kamion | mdz: applet that pops up when /var/mail/$USER is non-empty? | 05:58 |
Keybuk | Kamion: what would it give you to read it? | 05:59 |
Kamion | you could call it "system events" ;) | 05:59 |
Kamion | Keybuk: no idea | 05:59 |
Kamion | gnome-terminal -e mail ;) | 05:59 |
ogra | hehe | 05:59 |
mdz | Kamion: the messages the user would find in there are entirely cryptic to the average user | 05:59 |
=== Keybuk beats Kamion with a spoon | ||
mdz | meaningless and/or frightening | 05:59 |
lamont_r | Kamion: or snarf it into whatever format evo wants it in... | 05:59 |
mdz | they should not be presented to the user at all, in my opinion | 05:59 |
lamont_r | mdz++ | 05:59 |
Keybuk | most of the messages we're talking about are really cryptic debconf notes or random output from cron.monthly | 06:00 |
mdz | if the man page database isn't being regenerated properly, the user doesn't need to hear about that; we just need to fix it | 06:00 |
mdz | Keybuk: exactly | 06:00 |
mdz | scrollkeeper blowing up | 06:00 |
jbailey | If they're kept but not displayed, they need to be pruned over time. | 06:00 |
mdz | debconf saying "if you modified this random configuration file, you need to do X" | 06:00 |
mdz | files in /lost+found | 06:00 |
mdz | these are the only kinds of mail found in the default install | 06:00 |
lamont_r | no mta by default, crontab emits a warning. | 06:01 |
mdz | and so it doesn't seem worth the effort of developing tools to help the user see them | 06:01 |
Kamion | mdz: we'll never find out that we need to fix it, of course | 06:01 |
mdz | Kamion: we will, they won't | 06:01 |
Kamion | we won't get bug reports | 06:01 |
lamont_r | Kamion: we will from any users that install an mta | 06:01 |
mdz | Kamion: we won't get bug reports from most users anyway | 06:01 |
lamont_r | which is many | 06:01 |
mdz | they would read the message, not understand it, delete it, and move on | 06:01 |
lamont_r | and we don't want the bug reports from the rest^U | 06:02 |
lamont_r | ok - untrue, and all that | 06:02 |
sladen | Kamion: surely debbugs won't function without an MTA :) | 06:02 |
lamont_r | sladen: if it hits a web interface it will | 06:02 |
lamont_r | or deliveres directly to a named host's smtp port | 06:02 |
Kamion | sladen: surely Ubuntu doesn't use debbugs | 06:03 |
lamont_r | (yes, either of those is problematic) | 06:03 |
lamont_r | not installed on my system... | 06:03 |
mdz | if we remove postfix from base, we solve: 1) the religious MTA issue, letting a sysadmin choose the one they want, 2) the huge, under-acknowledged bug of the local mail spool growing without bound, 3) potential local security issues in the default install (though mitigated by postfix's design) | 06:03 |
=== lamont_r accepts mdz's bone in #3 | ||
mdz | replacing it with Keybuk's local-only agent solves #3 | 06:04 |
Kamion | mdz: TBH I'd be interested to see a system where it grows without bound more quickly than the rate of attrition to changelog.gz, but yeah | 06:04 |
lamont_r | 4) the number of complaints,etc on #ubuntu about the mta being broken | 06:04 |
ogra | mdz: 5) bend all daemons to write to a central log instead of mailing (to not loose the msgs) ? | 06:05 |
lamont_r | ogra: that's not something we fix, that's somethign that maybe needs to be done | 06:05 |
ogra | ubuntu_desktop.log :) | 06:05 |
mdz | ogra: that is a problem to be solved? | 06:05 |
Keybuk | cron seems to already syslog when it starts a cron, should be trivial to also log if it fails | 06:05 |
Kamion | we could then apply logrotate to those logs | 06:05 |
=== silbs [~sbsm0084@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-meeting | ||
Kamion | and solve the disk-usage problem | 06:06 |
ogra | Kamion: thats the idea :) | 06:06 |
sabdfl | Keybuk: how confident are you in your code for hoary? | 06:06 |
=== sivang [~sivang@box79162.elkhouse.de] has joined #ubuntu-meeting | ||
lamont_r | ugh. is pop-con in desktop? | 06:06 |
mdz | Keybuk: it could log a message saying "it failed", but not stderr | 06:06 |
mdz | which is fairly pointless | 06:06 |
mdz | lamont_r: it is, but thom modified it to use HTTP POST | 06:06 |
lamont_r | woot | 06:07 |
mdz | and it's disabled by default anyawy | 06:07 |
lamont_r | I prefer deleted entirely | 06:07 |
lamont_r | (mta, not popcon) | 06:07 |
mdz | because a) enabling it by default is a privacy issue, and b) we didn't want to ask an extra question | 06:07 |
Keybuk | sabdfl: *shrug* it's trivial code, it's mostly just Python's standard modules joined together | 06:07 |
sabdfl | ok | 06:07 |
fabbione | time to change the pix? | 06:08 |
mdz | change its diaper? | 06:08 |
=== lamont_r thinks he clicked on the wrong spot | ||
ogra | lol | 06:08 |
ogra | mdz: its ?? | 06:08 |
elmo | it's his BT ADSL line that keeps spassing out, not the PIX | 06:08 |
lamont_r | we should really set him up a proper nat box you know... | 06:09 |
fabbione | time to change ADSL? | 06:09 |
Keybuk | fabbione: seems to be a UK-wide problem | 06:09 |
Keybuk | mine's been doing it all day too | 06:09 |
fabbione | ah ok | 06:09 |
lamont_r | time to change countries? | 06:09 |
ogra | Keybuk: no cable in the uk ? | 06:09 |
Keybuk | Mark is the only other person I know with the same Efficient-based line | 06:09 |
=== sabdfl [~mark@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-meeting | ||
lamont_r | wb | 06:10 |
sabdfl | erk | 06:10 |
sabdfl | connection issues | 06:10 |
mdz | I don't have any real problem with moving to Keybuk's MTA as an incremental improvement | 06:10 |
mdz | but it only addresses a small portion of the problem in my opinion | 06:10 |
sabdfl | my vote is for (1) simple local-only keybuk solution, and (2) MailStorageStandardisation | 06:10 |
mdz | and it introduces some new ones | 06:10 |
mdz | sabdfl: do you feel that we should present cron errors and debconf notes to the user by default? | 06:11 |
mdz | because that would seem to be the result of implementing MailStorageStandardisation without doing something to suppress those | 06:11 |
sabdfl | mdz: your call | 06:12 |
sabdfl | yes, if we can make them easy to delete | 06:12 |
lamont_r | mdz: we could do that for bendy, and ship no-mta for hoary? | 06:12 |
mdz | and add a "what is all this confusing crap in my mailbox" entry to the FAQ? | 06:12 |
sladen | only if you can also present them with a button that says ''click here to let me fix it for you'' | 06:12 |
sabdfl | we should review the most common messages for, erm, friendliness | 06:13 |
Kamion | it would be easy to merely log debconf notes somewhere rather than shipping them, and I think we should do that | 06:13 |
sabdfl | Kamion: +1 | 06:13 |
Kamion | er, s/shipping/mailing/ | 06:13 |
lamont_r | Kamion: pruning over time? | 06:13 |
Kamion | lamont_r: logrotate | 06:13 |
ogra | cron as well.... | 06:13 |
lamont_r | 'k | 06:13 |
mdz | cron is not so straightforward, being multi-user | 06:13 |
Kamion | root cron jobs could be logged, and others mailed | 06:13 |
mdz | ideally errors would get written to ~user/.cron/log or something | 06:13 |
ogra | mdz: as i showed before, on woddy it writes to syslog by default | 06:14 |
sladen | lamont_r: would it be possible to never show the same error twice (eg, if they saw it yesterday, don't show it them again today) | 06:14 |
Kamion | that would get rid of the "cron.monthly said stuff I don't care about" problem | 06:14 |
sabdfl | mdz: this one i'd like to make visible :-) | 06:14 |
sivang | or a "system notification" buddy :) | 06:14 |
mdz | ogra: show me again? | 06:14 |
sabdfl | ~/cron.log | 06:14 |
=== mdz gapes | ||
ogra | /var/log/syslog.0:Jan 18 06:25:01 aleph /USR/SBIN/CRON[24665] : (root) CMD (test -e /usr/sbin/anacron || run-parts --report /etc/cron.daily) | 06:14 |
sabdfl | when it's created, put a baner at the top with instructions | 06:14 |
mdz | ogra: that is not what we are talking about | 06:14 |
Kamion | ogra: I think that's logging the job that got run, not its stderr output | 06:14 |
mdz | ogra: we are talking about when a command fails and outputs an error | 06:14 |
mdz | currently, those are sent via email | 06:15 |
sabdfl | mdz: the reason being, it's something we want the user to notice and are happy for him to delete | 06:15 |
ogra | ahh, ok...so bend stdout to /var/log/syslog then | 06:15 |
Kamion | mdz: it might also be worth getting rid of the "add initial user to /etc/aliases" thing | 06:15 |
mdz | ogra: we discussed some time earlier why that isn't feasible (security) | 06:15 |
ogra | ok | 06:15 |
sladen | what about a panel-applet that is installed but not enabled by default---right-click + add if required | 06:15 |
mdz | Kamion: leave root mail in root's box? | 06:15 |
sivang | sladen: my thought exactly | 06:15 |
Kamion | mdz: yes, you've already made it pretty clear that you don't want the user bothered by it | 06:15 |
Kamion | and the disk space thing is no worse that way | 06:16 |
Kamion | plus it's more clearly separated | 06:16 |
mdz | it's just shuffling the cruft around | 06:16 |
ogra | configuring logrotate to rotate /var/mail/root ? | 06:16 |
Kamion | sure but I think it's a better place for the cruft pending us fixing it | 06:16 |
Kamion | and easier to find the cruft | 06:16 |
mdz | ogra: please, be serious :-P | 06:16 |
ogra | :) | 06:16 |
sladen | I suspect they only message the user is bothered about is ''excessive IDE TimeOuts detected, your disk is shagged''---and that's something that's not reported at the moment anyway | 06:17 |
mdz | I maintain that there is nothing currently sent via email in the default install that a user (not a sysadmin) should see | 06:17 |
sabdfl | ok | 06:18 |
Kamion | hence let's remove the alias as a simple one-stop guarantee of that, first off | 06:18 |
mdz | Kamion: both root's spool and the user's spool are equally invisible in the desktop | 06:18 |
sabdfl | mdz: if that's the case let's go with kamion's loggin proposal for debconf notes | 06:18 |
mdz | we'd need to educate sysadmins to read that log | 06:19 |
sivang | sladen: so a special "admin only applet" to collect all of those msgs out of the mail spool file... | 06:19 |
Keybuk | meh, a mail client panel applet ?! | 06:20 |
=== Keybuk flees the insanity :p | ||
sabdfl | sivang, mdz: we could expose that log through Desktop -> Administration -> Logs | 06:20 |
mdz | considering this from a hoary perspective, the biggest bang/buck we get is by removing postfix from base | 06:20 |
Keybuk | sabdfl: we already have Applications -> System Tools -> System Log Viewer | 06:20 |
lamont_r | mdz: all else sounds rather bendy-timeframe | 06:20 |
sabdfl | Keybuk: right, the debconf stuff could link into that? | 06:20 |
ogra | sabdfl: any log could | 06:21 |
Keybuk | sabdfl: yeah, we make debconf and cron spit to syslog so they appear there | 06:21 |
sabdfl | mdz: i'd like to do the MailStorageStandardisation for at least core apps | 06:21 |
mdz | Keybuk: the one which helpfully shows only auth.log? | 06:21 |
sladen | Keybuk: sudo csplit '^From ' /var/mail/root | tail -1 | 06:21 |
Keybuk | mdz: that's just a silly default | 06:21 |
ogra | which should be changed :) | 06:21 |
Kamion | I don't think that syslog is a good place for debconf notes | 06:21 |
mdz | it's an application in search of a use case | 06:22 |
sabdfl | Keybuk: syslog isn't the right place for that though | 06:22 |
sabdfl | we need something else | 06:22 |
sabdfl | system events? | 06:22 |
Keybuk | why isn't syslog the right place? | 06:22 |
Kamion | they have a much longer lifetime of potential interest than most stuff in syslog, and are generated much less often | 06:22 |
sladen | System Log Viewer can probably be taught to read other logs by just passing them on the command line? | 06:22 |
sabdfl | also, they are messages, not lines | 06:22 |
Kamion | my syslog is full of junk, debconf notes are only generated on upgrades | 06:22 |
mdz | we're running long on time here; is there a reason we need to discuss this during the TB meeting rather than on ubuntu-devel? | 06:22 |
sabdfl | nope | 06:22 |
Kamion | I want to be able to go back to the last few upgrades | 06:22 |
mdz | we'll get wider feedback there | 06:22 |
Keybuk | sladen: it has a File -> Open menu item | 06:23 |
sabdfl | let's go on to ogra's proposal | 06:23 |
mdz | agreed | 06:23 |
lamont_r | mdz: is concensus then that we remove postfix from base? and do we have consensus on whether or not to provide a default mda? | 06:23 |
mdz | the question seems to be, should universe uploaders be able to add new packages to universe? | 06:24 |
mdz | ogra: is that an accurate characterization? | 06:24 |
ogra | non debian packages | 06:24 |
Keybuk | don't non-debian packages go into multiverse? | 06:24 |
ogra | selfmade or foreign packaged ones that are not in debian yet | 06:24 |
Keybuk | or am I confused? | 06:24 |
lamont_r | Keybuk: no | 06:24 |
pitti | yeah, dbus-1 :-) | 06:24 |
mdz | Keybuk: no, non-ubuntu-license-compliant packages go there | 06:24 |
pitti | pop up a dialog on a debconf note? | 06:24 |
sladen | Keybuk: I was under the impression multiverse was just non-free stuff | 06:24 |
mdz | ogra: I don't think it's relevant where the packages originate | 06:25 |
mdz | besides, new packages from Debian are automatically added anyway | 06:25 |
ogra | as an example i have the graveman package since some weeks on my server, i would like to upload it to have a gtk burning app.... | 06:25 |
mdz | I think that universe uploaders should absolutely be able to upload new packages | 06:25 |
ogra | ...but currently it would involve agreement of jdum or mdz | 06:25 |
mdz | they should of course be reviewed before being accepted | 06:25 |
ogra | jdub even | 06:26 |
Riddell | I have already added one new package to universe and have another one awaiting review | 06:26 |
ogra | so i think there should be a written policy | 06:26 |
ogra | that also prevents us from someone loading up a "nazi shooter" for instance | 06:26 |
mdz | ogra: what would the policy cover? | 06:26 |
sabdfl | reviewed by? | 06:26 |
mdz | sabdfl: currently, elmo | 06:27 |
sabdfl | ok | 06:27 |
mdz | with appeal to others where appropriate | 06:27 |
jbailey | What happens when a new universe package conflicts a new Debian package? | 06:27 |
sabdfl | let elmo make the licence call on universe vs multiverse, for one | 06:27 |
mdz | (security review by pitti, etc.) | 06:27 |
ogra | i would like to have it in a way that only critical things steal your time mdz | 06:27 |
ogra | or the others... | 06:27 |
mdz | ogra: my position is that anyone with upload privileges should be able to submit new packages for review | 06:27 |
sabdfl | how can we reduce the impact on pitti/mdz/elmo? | 06:27 |
ogra | thts my point sabdfl :) | 06:28 |
mdz | I don't think I need to be involved at all | 06:28 |
pitti | well, we talk about universe packages, right? | 06:28 |
elmo | mdz: i think we should concentrate on security reviewing main first? | 06:28 |
ogra | pitti: yup | 06:28 |
mdz | elmo: yes | 06:28 |
pitti | so I don't think we need an overly extensive review | 06:28 |
mdz | elmo: I meant in the context of "this looks fishy to me, pitti should look at it first" | 06:28 |
elmo | ok | 06:28 |
mdz | elmo: you'll shout if you're inundated with packages and we need to distribute that duty more, right? | 06:29 |
elmo | sure | 06:29 |
mdz | but I think it'll be a trickle for some time yet | 06:29 |
pitti | would it be possible to put new packages into a public queue? | 06:29 |
pitti | so that anybody can review it? | 06:29 |
ogra | or ist it probably a job for the MOTU master ? | 06:30 |
sladen | universe-staging? | 06:30 |
pitti | as soon as two members/motus give thumbs up, it's approved to universe? | 06:30 |
mdz | pitti: mentors.debian.net-a-like | 06:30 |
pitti | right | 06:30 |
mdz | I think that's a good idea, but there's no pressing need for it yet | 06:30 |
elmo | longer term launchpad solves this | 06:30 |
mdz | good point | 06:31 |
pitti | mdz: I think looking at the list of files shipped in the debs (correct paths, no setuid etc.) would be a good first and quick test | 06:31 |
elmo | I don't think we should invest any energy into solving this until the number of NEW packages becomes a problem for me | 06:31 |
mdz | ogra: are you satisfied with the solution that has been described? | 06:31 |
mdz | elmo++ | 06:31 |
ogra | yup :) | 06:31 |
sabdfl | is christoph haas around? | 06:31 |
ogra | nope | 06:31 |
mdz | doesn't seem to be | 06:31 |
mdz | he's ChrisH I believe | 06:32 |
ogra | i pinged him in -de | 06:32 |
sivang | I've called on him let's see if he responds | 06:32 |
sabdfl | elmo: can i ask you to ping us if you aren't able to approve / decline a package within two working days? | 06:32 |
mdz | yes, we did decide to make that type of commitment in Mataro | 06:32 |
sabdfl | we should promise the MOTU that level of service | 06:32 |
elmo | *shrug* k | 06:33 |
sabdfl | if we can't it's a bug on our processes, and we'll need to decentralise further | 06:33 |
mdz | but if someone unknown uploads a huge and complex package, there is no guarantee that we will be able to devote resources to the review within that time | 06:33 |
ogra | sabdfl: i think MOTU should bother canonical as less as it can... | 06:33 |
sabdfl | i'd like universe / multiverse to be relatively liberal | 06:33 |
sabdfl | ogra: agreed | 06:33 |
sabdfl | perhaps if two MOTU have approved the package it goes straight in? | 06:33 |
sabdfl | barring CoC issues ("nazi shooters") | 06:34 |
elmo | and license issues | 06:34 |
sabdfl | elmo: +1 | 06:34 |
sabdfl | mdz: that's reasonable, yes | 06:34 |
sabdfl | ok ogra I think that's an answer for your question | 06:34 |
ogra | yup... | 06:35 |
mdz | sladen: are you applying for member status or maintainer/uploader status? | 06:36 |
ogra | the prob is that we currently are only two MOTUs ..... | 06:36 |
ogra | ahh, kdz was faster again :) | 06:36 |
ogra | mdz | 06:36 |
sabdfl | sladen: or MOTU status :-) | 06:36 |
sladen | mdz: I think member, (immediately followed by committer/maintainer) ? | 06:36 |
mdz | ogra: that will provide incentive for you to recruit new MOTU ;-) | 06:36 |
ogra | yay ! | 06:36 |
mdz | members need a CC vote according to the wiki process | 06:37 |
sabdfl | i think we have cc present | 06:37 |
sabdfl | mako? elmo? kamion? | 06:37 |
elmo | sorry what are we voting on? universe only, or everything? | 06:37 |
mdz | we have 3/4 | 06:37 |
sladen | I think I'm going to need sysvinit, mkinitrd-tools, initrd-tools, grub(?), $xorg_resolution_detect, gdm, usplash-*, in the next couple of weeks (all of which are in main) | 06:38 |
sabdfl | for sladen, membership | 06:38 |
sabdfl | tb will do upload | 06:38 |
elmo | uh | 06:38 |
sivang | people who get approved by CC are eligable for voting afterwards in CC meetings? | 06:38 |
sabdfl | sivang: vote to confirm CC appointments | 06:38 |
ogra | sabdfl: we are in TB...so you could do both ;) | 06:38 |
sivang | sabdfl: ok. | 06:39 |
mdz | ogra: we will | 06:39 |
sabdfl | yes | 06:39 |
elmo | I wish these damn procedures would stabilize, they seem to change everytime I blink | 06:39 |
sabdfl | elmo, kamion? | 06:39 |
sivang | elmo: I get that feeling also ;-) | 06:39 |
sabdfl | membership for sladen. recognises a "significant and ongoing contribution" | 06:39 |
mdz | A person who wants to become a member should: | 06:39 |
mdz | Be contributing. Very wide definition of contributing: Coding, writing, art-work, etc. Everybody can be pa rt of the community. Contributions should be significant. | 06:39 |
mdz | Have good testimonials / recommendations from others in the community. | 06:39 |
mdz | Have a wiki-page with stuff he is working in / has worked on. | 06:39 |
mdz | Community Council votes whether a person can become a member. | 06:39 |
haggai | elmo: stop blinking so slowly then :) | 06:39 |
Kamion | sorry, just a sec | 06:40 |
sabdfl | thpecial | 06:40 |
mdz | I can speak for sladen's contributions to date being significant | 06:40 |
elmo | sabdfl: does that mean CC no longer have any say on whether a person can upload to main? | 06:40 |
sabdfl | elmo: yes | 06:40 |
sabdfl | cc is membership, which votes on cc appointments and procedural amendments etc | 06:40 |
sabdfl | tb is upload, and votes on tb appointments | 06:40 |
Kamion | I'm fine with membership for sladen on the basis of ongoing usplash work and other contributions | 06:41 |
mdz | wiki page is present and linked from the agenda | 06:41 |
sabdfl | possibly also votes on technical items | 06:41 |
sabdfl | elmo? | 06:41 |
elmo | *shrug* fine | 06:41 |
ogra | ahh, which answers my question from the beginning... | 06:41 |
sabdfl | sladen gets my vote too, welcome aboard | 06:41 |
sivang | do we have a list of the wiki of CC members already | 06:41 |
sivang | ? | 06:41 |
mdz | sladen: congratulations | 06:42 |
=== sladen gives nods of Thanks all round | ||
sivang | s/of/on | 06:42 |
sabdfl | sivang: no, we should have, could you create that please? | 06:42 |
ogra | yay, hi sladenMOTU | 06:42 |
sivang | sladen: congerts. | 06:42 |
sabdfl | we have a spec which the launchpad team will implement | 06:42 |
sabdfl | that will publish the membership list on launchpad.ubuntu.com | 06:42 |
sabdfl | Ubuntites unite! | 06:42 |
mdz | sivang: CC members, or Ubuntu members? | 06:42 |
mdz | CC members are on the website | 06:42 |
sivang | mdz: link? | 06:43 |
mdz | sivang: http://www.ubuntulinux.org/community/council | 06:43 |
mdz | but I think you're talking about Ubuntu members | 06:43 |
mdz | since there is no list that I know of | 06:43 |
sabdfl | ok, i think we are done with the agenda | 06:43 |
mdz | and there definitely should be! | 06:43 |
sabdfl | any other business? | 06:43 |
mdz | sabdfl: not quite | 06:43 |
sivang | yes, | 06:43 |
sladen | sladen upload? | 06:43 |
mdz | sladen applied for upload privileges | 06:43 |
sabdfl | true | 06:43 |
ogra | sladen: go | 06:43 |
sivang | I was adviced by mako and jdub and now pitti, to request a package for hebrew fonts to get in main, | 06:43 |
sivang | namel culmus. | 06:43 |
sivang | *namely | 06:44 |
mdz | I have authenticated sladen's gpg key myself, if I'm not mistaken, so I'm satisfied with that requirement | 06:44 |
=== sladen wonders if he should /part #ubuntu-meeting | ||
ogra | culmus was also requested in ubuntu-devel ML | 06:44 |
mdz | we don't have a lot of guidelines for upload status at this point | 06:44 |
sabdfl | mdz, keybuk, upload for sladen to universe plus sysvinit, mkinitrd-tools, initrd-tools, grub(?), $xorg_resolution_detect, gdm, usplash-* | 06:45 |
Keybuk | I don't have any problem with sladen | 06:45 |
mdz | or simply unlimited? | 06:45 |
elmo | guys, this "per package" thing is completely out of left-field | 06:45 |
Keybuk | (in general) | 06:45 |
mdz | I think having a huge list of packages is a bit silly | 06:45 |
elmo | no one told me about it --> katie has no support for it | 06:45 |
sabdfl | elmo: don't worry, we'll manage it socially for the moment | 06:45 |
sivang | mdz: if I Was approved by the CC , am I on the CC , or and Ubuntu member? :) | 06:45 |
sabdfl | soyuz will automate it | 06:45 |
mdz | if we trust someone to upload an essential package, we trust them to upload anything | 06:46 |
mdz | sivang: you are an Ubuntu member | 06:46 |
mdz | sivang: sabdfl appoints new CC members | 06:46 |
sabdfl | nominates | 06:46 |
sivang | mdz: ah ok :) now that made clea some stuff :) | 06:46 |
elmo | sabdfl: I'd really rather we didn't | 06:46 |
sabdfl | sivang votes to confirm :-) | 06:46 |
mdz | sabdfl: website says appoints | 06:46 |
sivang | sabdfl: hehe , cool. | 06:46 |
sabdfl | ok, i should change it | 06:46 |
sabdfl | can't appoint till its confirmed :-) | 06:47 |
sabdfl | we need a *picture* not another document | 06:47 |
mdz | ok, so the matter at hand | 06:47 |
mdz | as I said, there aren't any guidelines in place for us regarding confirmation of upload status | 06:47 |
sivang | sabdfl: nominates sounds more prestigues...;-) | 06:47 |
sabdfl | is just awaiting keybuk's vote | 06:47 |
ogra | cant sladen just agree he limits his uploads to the named packages and we all belive him ? | 06:47 |
sabdfl | yes | 06:47 |
mdz | but in my opinion, it's a matter of trusting the person not to exceed their limits | 06:47 |
Keybuk | sabdfl: sorry, which? (having slight connectivity problems and getting things in bursts) | 06:48 |
sabdfl | mdz: in future we will automate this, for now, it's just social | 06:48 |
sabdfl | sladen, upload to universe plus a bunch of break-my-ubuntu packages ;-) | 06:48 |
Keybuk | <Keybuk> I don't have any problem with sladen | 06:48 |
Keybuk | :p | 06:48 |
mdz | I don't think it's sensible to saddle sladen with a list of packages in main that he is allowed to upload | 06:48 |
=== ChrisH [~chaas@gw.workaround.org] has joined #ubuntu-meeting | ||
mdz | because it means revisiting the issue when that list changes with his work | 06:48 |
mdz | Keybuk: unrestricted main? | 06:49 |
Keybuk | mdz: I don't have a problem with main in general | 06:49 |
mdz | based on his participation so far, my feeling is that he won't do anything silly in main | 06:49 |
sladen | granted. It would be inconvient---there maybe some others than come into play. I'm happy to ping socially before doing so | 06:49 |
sabdfl | great | 06:50 |
ogra | mdz: else we wll find out where he lives :p | 06:50 |
Keybuk | ogra: oh, it's easy to find sladen ... just say the words "free bar", and he mysteriously turns up <g> | 06:50 |
ogra | lol | 06:50 |
mdz | and on that basis I have no problem with him uploading packages in main | 06:50 |
sladen | ogra: other side of town from sabdfl. I'm sure it'll pop on the Victoria Line and bring the Heavies if needed | 06:50 |
sladen | he'll | 06:51 |
ogra | *g* | 06:51 |
mdz | sabdfl: you voted? | 06:51 |
sabdfl | yes, +1 to sladen, unlimited | 06:51 |
ogra | <sabdfl> is just awaiting keybuk's vote | 06:52 |
mdz | ok, sounds like we're done | 06:52 |
sabdfl | sladen, congrats and thanks | 06:52 |
ogra | err, one last question.... | 06:52 |
sivang | anybody saw my fonts package question? ;) | 06:52 |
ogra | what turned out with the MOTU Master ? | 06:52 |
ogra | do we have one ? | 06:52 |
mdz | sivang: on the agenda? | 06:52 |
mdz | ogra: that's a CC matter | 06:53 |
ogra | ok | 06:53 |
sivang | mdz: hrm no, I was away for quite some time and pitti noted this to me only when the meeting was one hour in. | 06:53 |
mdz | ogra: and I thought it was decided at the CC meeting | 06:53 |
sabdfl | thans everybody | 06:53 |
ogra | mdz: i thought there was no answer from either haggai or ChrisH | 06:53 |
mdz | sivang: we have been here two hours; add it to the agenda for the next meeting | 06:54 |
sivang | mdz: sure, no prob, and I apologize. | 06:54 |
mdz | ogra: I think they were appointed, pending their approval of themselves ;-) | 06:54 |
mdz | thanks, everyone | 06:54 |
ChrisH | ogra: No public answer at least. :) | 06:54 |
mdz | adjourned | 06:54 |
sabdfl | i will get the current version of our membership / upload / commit permissions up on a wiki today | 06:54 |
haggai | mdz: I wrote to mako & sabdfl last week | 06:54 |
ogra | yup...ChrisH, but since i'm a MOTU now i'd like to know it | 06:55 |
sivang | sabdfl: does somebody has a list of already ubuntu members? | 06:55 |
ChrisH | ogra: haggai agreed to do it. I don't have the time currently. Already talked with him about it. | 06:55 |
mdz | sivang: it could at least be found in the CC transcripts | 06:56 |
ogra | a, k | 06:56 |
sabdfl | sivang: put up the wiki page with your best recollection, then well ask members to correct it | 06:56 |
sladen | sabdfl: would it be add a repositary of ''signed CoC's'' I think it would be a useful (and public) "role call" to have | 06:56 |
sivang | sabdfl: ok, sure. | 06:56 |
sabdfl | sladen: yes, elmo has them | 06:56 |
elmo | no I don't | 06:56 |
sabdfl | ? | 06:56 |
sladen | make does | 06:56 |
sabdfl | mako? | 06:56 |
elmo | that was another random change - they all got sent to mako instead | 06:56 |
sivang | lol | 06:56 |
ogra | regarding the CoC, there should be a text file for download to easily sign it | 06:56 |
sabdfl | elmo: sorry, my mistake, you should be getting them | 06:57 |
sabdfl | ogra: could you produce that? | 06:57 |
ogra | yup | 06:57 |
sabdfl | put a version on it, because we will need to rev it over time | 06:57 |
ogra | (since i already have one i signed) | 06:57 |
ogra | hmm, who is allowed to change that website ? | 06:59 |
Treenaks | ogra: oh you fixed it? | 06:59 |
Treenaks | ogra: I just signed a w3m -dump :) | 06:59 |
mdz | sladen: yes, that was the plan, to post them publicly | 06:59 |
ogra | Treenaks: i made a ps and used ps2txt ;) | 06:59 |
Treenaks | ogra: hmmm.. :) | 06:59 |
ogra | Treenaks: so you will _finally_ join MOTU....thats great news :) | 07:00 |
sivang | ogra: are you the new MOTU leader? | 07:00 |
Treenaks | MOTU? | 07:00 |
ogra | sivang: it appears that haggai is :) | 07:00 |
sladen | Masters Of The Universe | 07:00 |
Treenaks | oh wait yes | 07:00 |
=== ChrisH [~chaas@gw.workaround.org] has left #ubuntu-meeting [] | ||
=== pitti [~martin@box79162.elkhouse.de] has left #ubuntu-meeting [] | ||
=== Keybuk [scott@descent.netsplit.com] has left #ubuntu-meeting ["Leaving"] | ||
=== sabdfl [~mark@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-meeting | ||
mako | sabdfl: i'm here | 07:17 |
=== silbs [~sbsm0084@host217-37-231-28.in-addr.btopenworld.com] has left #ubuntu-meeting [] | ||
=== maskie [~maskie@196-30-109-23.uudial.uunet.co.za] has joined #ubuntu-meeting | ||
=== jiyuu0 [~jiyuu0@219.95.57.189] has joined #ubuntu-meeting | ||
=== sivang [~sivang@box79162.elkhouse.de] has left #ubuntu-meeting [] | ||
=== ogra [~ogra@p508EB646.dip.t-dialin.net] has left #ubuntu-meeting ["Verlassend"] | ||
=== doko [~doko___@dsl-082-082-213-108.arcor-ip.net] has joined #ubuntu-meeting | ||
=== elmo [~james@83-216-141-215.jamest298.adsl.metronet.co.uk] has left #ubuntu-meeting ["Leaving"] |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!