2004-09-21
Ubuntu-Meeting
Log 9-21-2004
03:10 jdub so many channels...
05:54 mdz_ 5 minutes
06:00 mdz_ ok, time to begin
06:00 mdz_ the first item on the agenda is a discussion of the new software to be brought into universe
06:01 mdz_ my understanding is that this includes Debian contrib, Debian non-free and a subset of Christian Marillat's packages
06:01 Mithrandir does this mean there will be no way for people to see whether they install free or non-free software from universe?
06:02 mdz_ that is correct
06:02 mdz_ universe would be a mixture
06:02 mdz_ or, we could use sections to delineate it
06:02 Kamion there's a comment on the agenda from jdub suggesting we avoid non-free in universe
06:02 Keybuk are we still going for a "you're installing software from universe, are you crazy?!" kinda warning?
06:02 mdz_ yes
06:02 mdz_ virtual jdub says, let's not do it
06:02 mdz_ Keybuk: yes
06:02 elmo actually, it's everything in Marillat's - he doesn't have decss or anything flat out indistributable
06:03 mdz_ I have patches for synaptic waiting
06:03 Keybuk what do we class as non-free? cf. mplayer
06:03 mdz_ elmo: he doesn't have java in there?
06:03 Mithrandir I think we should have non-free universe or at least sections in it
06:03 elmo mdz: no that I saw, no
06:03 mdz_ oh, ok. strike 'a subset of', then
06:03 elmo Mithrandir: the current plan for that is to change the section to universe/non-free for non-free software
06:04 elmo the alternate/backup plan is to create universe-non-free
06:04 Mithrandir I would prefer the latter
06:04 elmo (plan in the sense, of what I'd plan to do, if we go ahead, sort of thing)
06:04 mdz_ if the need is to let the user know whether they are installing free or non-free software, then a section is no good
06:05 Mithrandir also think of derivatives, they might want to get something off universe, and if they have to go through a lot of licenses, that's kinda wasted work.
06:05 mdz_ if the goal is just to classify it so that people can look and find out at a glance
06:05 sabdfl Mithrandir: universe is "you're on your own"
06:05 mdz_ then sections are useful
06:05 mdz_ I think it is preferable that universe live up to its name, and be inclusive by nature
06:05 mdz_ so if we want further classification, it should be in subdivisions of universe
06:05 mdz_ that could mean that universe becomes a separate archive, having components, or that we use sections
06:06 Mithrandir I guess we can't have packages that exists in both universe and free-universe?
06:06 Mithrandir or something along those lines?
06:06 sabdfl Mithrandir: i'd prefer we do NO classification of universe software at all
06:06 Keybuk what kind of non-free is likely to end up in universe?
06:06 elmo mdz: dude, we can't have two archives - we have packages with source in main and binaries in universe
06:06 sabdfl that way we don't have to argue semantics, it's either main, or it's not
06:06 mdz_ elmo: good point
06:07 mdz_ it would be useful to have a marker for derivative distributions
06:07 Mithrandir Keybuk: no-commercial stuff, for instance?
06:07 mdz_ if they know that something came from Debian main, they don't need to be concerned about the license
06:07 elmo sabdfl: yeah, but it's really harsh taking 9000 packages of free software and hiding 100 packages of very non-free software in there
06:07 elmo sabdfl: debian doesn't do a good job of checking licenses - expecting users too, is just, well harsh
06:07 mdz_ we'd be mixing things in Debian main with non-free stuff, forcing them to review the licenses individually
06:07 Mithrandir Keybuk: some commercial apps, like opera would be interesting (to me)
06:08 sabdfl Mithrandir: i've asked elmo to look into opera in universe
06:08 sabdfl elmo: 9x% of users, given the choice to add non-free during install, say yes
06:08 mdz_ limi will be pleased :-)
06:08 Mithrandir sabdfl: it's non-redistributable atm, afaik. they promised to fix it a few years back, but I figure I need to kick some butt.
06:08 elmo keybuk: mark's said we can ship anything that is distributable (even if it's "non-commercial only") and not likely to invoke DMCA type lawyer/police visits
06:09 sabdfl not "ship" on the cd
06:09 elmo err, sorry
06:09 elmo s/ship/add to universe/
06:09 sabdfl but certainly make accessible via universe, for people who want it
06:09 mdz_ elmo: and also doesn't contaminate other software we distribute (i.e., java)
06:09 elmo mdz: right, that too
06:09 sabdfl the idea is simply to get rid of all the "go over there, add this line to that file, then compile that bit of code...." hoops
06:10 elmo sabdfl: I don't think we should ignore the x% of users who _do_ actually care about freeness just because they're a minority tho
06:10 sabdfl mdz_: my reading of the java licence is not that it contaminates other java implementations, but just that those have to be clearly labeled as replacing bits of java
06:10 Keybuk is there going to be a contrib-alike "depends may not be complete" loop-hole, or do we want supported+universe to be depends-complete ?
06:10 sabdfl there are a ton of warnings when you add universe to your sources list, for them specifically
06:11 sabdfl Keybuk: depends-complete
06:11 elmo ??
06:11 sabdfl erm, actually, i don't know what i'm talking about there
06:11 elmo dude, you told me to import contrib, it can never be depends-complete if we do that
06:11 mdz_ contrib is permitted to depend on things that aren't packaged at all
06:11 Keybuk sabdfl: packages in Debian contrib are allowed to depend on packages that don't exist
06:11 sabdfl exactly. i don't know what i'm talking about there :-)
06:11 mdz_ or which are randomly downloaded from the Internet
06:12 sabdfl we should include the "randomly downloaded" stuff in universe too, if we can
06:12 mdz_ elmo: did Debian ever get an official-ish opinion on the java license?
06:13 Mithrandir sabdfl: I'm a bit wary of this -- it'll make universe include all kinds of _completely_ not-quality-controlled crap.. people treat it as "ok, not supported, but at least it's done by DDs so it might be decent" today.
06:13 elmo sabdfl: I think you're conflating stuff.. the warnings are primarily there because the software is unsupported.. we're going to have to check the license on stuff we import that isn't from debian/main anyway, so we'll know which is free and which isn't. if we choose not to separate them out, it's not because it costs us anything to do so
06:14 elmo mdz: official from who? AFAIC, the ftp-master position was that it was indistributable in non-free, and we invited -legal to tell us otherwise and they didn't
06:14 sabdfl Mithrandir: most people will only install something from there that they are explicitly looking for. if they are explicitly looking for it, they will install it wherever they find it
06:14 mdz_ Mithrandir: true, currently people moving from Debian see it that way, as a sort of compatibility layer
06:14 elmo there's also the issue of derivtatives, it's not just users
06:14 mdz_ elmo: official-ish as in consensus from people who enjoy arguing about licenses (i.e., debian-legal)
06:15 Kamion yeah, if we want to do something like gnubuntu, it'll be a real issue
06:15 sabdfl i'm committed to making it obvious that installing software from universe is well labelled as "scary"
06:15 sabdfl but then making it easy for people to do that if they want
06:15 Kamion although they may want to audit licences themselves anyway
06:15 Kamion if we could avoid throwing away the information about the origin component in Debian, that might help
06:15 elmo mdz: I REJECTed it, asked for follow ups to legal - no clear consensus emerged to dispute the contamination issues, AFIACR
06:16 Mithrandir we have a few issues here -- we have the free-vs-non-free argument, and we have the "put random stuff downloaded off the net" discussion
06:16 Mithrandir should we do one at a time?
06:17 mdz_ I don't think the latter is much of an issue at this point
06:17 Kamion elmo: is there any way we can have Debian-Section: or something in the Packages files?
06:17 Kamion elmo: that way, we can rearrange ourselves as we see fit without throwing away the information
06:17 Mithrandir most users don't give a shit about licenses -- derivatives will, to a lesser or bigger degree.
06:18 mdz_ right
06:18 Keybuk Kamion: or some kind of origin field in general?
06:18 mdz_ free-vs-non-free is an issue for philosophers and derivatives
06:18 Kamion and derivatives will be quite happy wandering about through the Packages files
06:18 mdz_ the question is, how much work are we willing to do in order to make life easier for those two groups?
06:18 mdz_ I think that a marker to say where the package came from, as Kamion is suggesting, is about the right amount
06:18 Kamion mdz_: we==Canonical are doing a hell of a lot of work to make life easier for derivatives :-)
06:18 mdz_ Kamion: on the specific issue of licensing :-)
06:19 Mithrandir throwing away information is silly, but if we can preserve it like Kamion suggests, it sounds good to me
06:19 sabdfl we can stick some of that stuff in the launchpad database, which would make it possible for the derivatives to use it in the same interface where they are selecting packages anyway
06:19 mdz_ don't we already preserve the Section: from Debian?
06:19 sabdfl all of that stuff, actually :-)
06:20 Keybuk I guess we need to keep track of where we get $random debs from anyway, to get the newer versions
06:20 elmo mdz: usually but not always - I override them sometimes, e.g. nothing in universe is >> optional
06:20 Mithrandir elmo: er, section != priority?
06:20 sabdfl i don't want to add any work to the archive structure and files, that's a next-gen archive problem
06:20 elmo that's priority, actually, but you get the idea, i.e. that we have our own overrides which are seeded from debian's
06:20 Kamion mdz_: I think the part before the / needs to be squashed ...?
06:20 mdz_ Kamion: doesn't it stick around in the .deb?
06:21 mdz_ and the .dsc
06:21 sabdfl i'll agree that we'll capture this stuff in launchpad, where it can be part of the automatic filtering of packages
06:21 Mithrandir mdz_: it is in the debian/control file at least.
06:21 Kamion mdz_: hm, true, assuming that the original .deb's section was right
06:21 Kamion although you have to get at least contrib/non-free right, I guess
06:21 sabdfl "don't show me packages that have licence XYZ"
06:21 elmo kamion: for Debian it has to be - debian doesn't have component overriding
06:21 mdz_ Kamion: I would at least hope that the original .dsc section is correct with respect to main/non-free/contrib
06:21 Kamion elmo,mdz: yeah, brain just catching up with fingers
06:22 elmo sabdfl: or more likely "don't show me non-free software" :-P [hell, that should be the default IMO]
06:22 Keybuk elmo: do you get to save your particular non-free religion in a cookie to do that? {ducks}
06:23 mdz_ ok, so if we're doing it fancy for the next-gen stuff, the question would seem to be, what can we do today in order to make it easy to seed that data when it's ready?
06:23 sabdfl mdz, we'll capture it in launchpad, then discuss for next-gen-archive
06:23 elmo at the very _least_ we should override the section to be universe/non-free
06:23 Keybuk mdz_: stick a field in the database and record where we get things from? "origin = debian non-free" or something
06:23 mdz_ Keybuk: the existing database?
06:23 Keybuk elmo: that makes some sense too
06:24 elmo sabdfl: we really shouldn't just put this all off till next-gen - we have to redo our archive anyway for universe, things like adding another section is mickey mouse
06:24 mdz_ sabdfl: is launchpad ready to receive it?
06:24 Keybuk mdz_: if it's no real work to capture it in both launchpad and katie, let's do that?
06:24 sabdfl no, i really DON'T want us to be clasifying this stuff. if we do it, we have to do it totally for everything, and that's going to require non-scalable review processes
06:25 sabdfl mdz_: no
06:25 mdz_ sabdfl: the sort of classification being diiscussed was a "where this came from", which is no work for us because we already know
06:25 mdz_ i.e., Debian main, Debian contrib, Debian non-free, and "other"
06:25 elmo sabdfl: dude, I can not sanely import non-free/*, there _is_ going to be software in there that says "only Debian may redistribute" this
06:25 elmo sorry, "Debian and it's mirrors"
06:26 mdz_ aren't we a de facto Debian mirror? :-)
06:26 elmo not when we do 1.0-2ubuntu1 uploads :-P
06:26 sabdfl we have to store that somewhere on our side for ongoing syncing purposes anyway, there is no need to stick it in the archive
06:26 Keybuk mdz_: dodgy, decididly
06:27 mdz_ so, the needs of derivatives are expected to be met by the launchpad db
06:28 mdz_ what about the philosophers?
06:28 mdz_ they want that information now, while derivatives can wait
06:28 elmo "philosophers" pfft
06:28 elmo guys, it's not being a philosophers to not want to unwittingly install some "no commercial" use software at work
06:28 sabdfl the philosophers have the following information: YOU ARE ON YOUR OWN. now philosophize.
06:29 elmo sabdfl: due respect, that's crap
06:29 mdz_ that's a big step down from Debian
06:29 mdz_ especially for the server market
06:29 elmo exactly
06:29 sabdfl mdz_: it's a different approach. why is the server market different?
06:29 elmo because home users aren't affected by "no commercial use" clauses
06:30 elmo (usually)
06:30 mdz_ sabdfl: today, I can put Debian on a corporate box, and know that anything I can intsall I am legally allowed to use
06:30 sabdfl same with ubuntu. if you dip into universe, you need to check the licence of anything you install.
06:30 Kamion this would basically make universe totally useless for anyone except home users
06:31 Kamion other people aren't likely to have time to check the licences
06:31 mdz_ sabdfl: the difference is that Debian has everything you could possibly want on a server
06:31 mdz_ sabdfl: while ubuntu doesn't, not by a long shot
06:31 mdz_ you don't need to dip outside Debian
06:31 elmo sabdfl: dude, no one (not even us) is going to run just 'main' on their servers
06:31 mdz_ every server has its dirty laundry :-)
06:32 elmo why on earth should we actively make things hard for people, when we already know whether a piece of software is free or not?
06:32 sabdfl because if we say some things are non-free, and some aren't, we expose ourselves to risks in other regards
06:32 Keybuk sabdfl: in what way?
06:33 mdz_ which is why I couched it in terms of origin, rather than licensing
06:33 sabdfl what about universe/known-free and universe/black-hole
06:33 mdz_ if we record where it came from, we're providing almost all the information the user needs, without making a licensing decision
06:33 elmo what about universe/debian-main, universe-REALLY-fucking-scary ? :P
06:34 Kamion and we can do 'Origin => Marillat', which is the sort of thing some people may be interested in anyway
06:34 sabdfl Keybuk: someone might argue that we are recommending things to people based on that classification
06:34 mdz_ universe-you-must-be-this-tall-to-ride-this-ride
06:34 sabdfl which is why i prefer NO classification
06:34 Kamion sabdfl: it's not a classification, it's informational
06:34 sabdfl mdz_: so LA
06:34 sabdfl Kamion: so is the label on herbal tea, and that can get you sued
06:34 Kamion the reason it isn't a classification is because the packages aren't segregated on that basis
06:34 Keybuk sabdfl: is that inherently much different to the nv driver being in main and the nvidia being in restricted?
06:35 mdz_ sabdfl: how about we mark things which came from Debian main with an Origin: tag
06:35 sabdfl but there we have done a really thorough job of the analysis.
06:35 mdz_ sabdfl: and simply omit that tag from everything else
06:35 Kamion sabdfl: if something's going to be getting us sued based on what random fields we put next to it in a Packages file, we should be running away from it, not distributing it
06:35 sabdfl and I don't WANT to do a really thorough job on some of the stuff that I want in universe
06:35 Keybuk sabdfl: yeah, I guess ... simply saying where it came from doesn't have that problem though? "Origin: debian contrib" for example?
06:36 Kamion that's a quotation more than a classification
06:36 mdz_ it wouldn't be necessary, except that it's not particularly easy to determine from the package itself whether or not it came from Debian
06:37 mdz_ or even possible
06:37 mdz_ sabdfl: does a Debian/non-Debian distinction give you any bad feelings?
06:38 sabdfl mdz_: yes, i'd prefer "known-free" and "something else"
06:38 elmo known-free is just as dangerous liability-wise
06:38 mdz_ sabdfl: how about if known-free happens to contain only stuff from Debian main?
06:38 elmo if Debian screws up (and it does - a lot) we're in the same boat
06:39 elmo if you're worried about liability, Debian-main/everything-else are better IMHO
06:39 sabdfl mdz_: it will end up carrying more than that, inevitably
06:39 sabdfl we will end up with potentially a lot more software in universe than in all of debian
06:39 mdz_ sabdfl: it will only carry more than that if we are reviewing licenses of things going into universe
06:40 mdz_ sabdfl: which we won't(?)
06:40 elmo err, we so will
06:40 elmo guys, there's 178 packages in non-free
06:40 elmo it'll take me like an hour or two at most to check those licenses, and I think it'd be unutterably insane to not do so
06:40 elmo there's avoiding liability and then there's gross negligence
06:41 Kamion a lot more in universe than in all of debian? it'll take even a crack team of packagers performing no QA at all quite a while to get that far
06:42 Kamion I can see a few special cases, but it seems we should be encouraging free software wherever possible, so there's no reason why the bulk of things shouldn't be in Debian too
06:45 mdz_ what if we lump everything together
06:45 mdz_ and some random schmo on the Internet publishes a web page with a list of what he thinks is free and what isn't
06:46 sabdfl mdz_: we'll do some internal review, but the less we advertise and make claims about what's there, the better
06:46 sabdfl elmo: of course i am happy for that review to happen, so we can find any real problems and weed them out
06:47 mdz_ sabdfl: agreed
06:47 mdz_ sabdfl: someone from the free software community could do such a classification, independently, though
06:47 sabdfl yes
06:48 sabdfl and we can put a very rich classification into launchpad, to make it easy for derivatives
06:48 mdz_ which I think would meet the needs of the zeal^Wphilos^Wfreedom fighters
06:48 mdz_ sabdfl: won't the rich classification in launchpad have the same issues?
06:49 sabdfl i'd prefer "known-free" and "unknown", where in launchpad, unknown trivially lets you read the licence and decide for yourself whether to include that package in your derivative
06:49 sabdfl i do see the point in "known-free" for 90% of the no-brainer cases out there
06:50 sabdfl is the text in sources.list suitably scary?
06:50 mdz_ no
06:50 elmo it needs to be FIGLET scary
06:50 mdz_ ## Uncomment the following two lines to fetch updated software from the network
06:50 mdz_ ## and be able to use more than 12000 unsupported packages from the universe archive.
06:50 sabdfl my install must be old, there's nothing there at all
06:51 mdz_ it sounds almost tempting
06:51 sabdfl eek, that sounds ENCOURAGED
06:51 sabdfl it should be something like:
06:51 mdz_ sabdfl: I have patches from mvo to add scary warnings to synaptic
06:51 elmo what about dselect/aptitude?
06:51 elmo or even apt it self?
06:51 mdz_ elmo: comments in sources.list
06:51 elmo let's offer a bounty to the maintainer of apt!
06:51 sabdfl # uncomment the following line to add software from the "universe" repository.
06:52 elmo mdz: that's it ???
06:52 mdz_ elmo: oh, you mean when installing stuff
06:52 mdz_ that's so Hoary
06:52 sabdfl # NB software from this repository is ENTIRELY UNSUPPORTED by the Ubuntu team
06:52 elmo I'm going to populate hoary soon, and you guys will have to starty saying "that's so Perky" :-P
06:52 Keybuk For following NEWS universe packages WHICH ARE ENTIRELY UNSUPPORTED YOU FOOL will be installed:
06:52 mdz_ elmo: you're bluffing
06:52 Keybuk elmo: grumpy
06:53 Keybuk hmm @ auto-typing
06:53 Mithrandir sabdfl: "they are not supported in any way, including security updates"
06:53 sabdfl # and may also not be under a free licence. Please satisfy yourself as to
06:53 Kamion sabdfl: if you could file a base-config bug with the text you want, that'd be good ...
06:53 Mithrandir but scarier.
06:54 sladen http:/.../ universe non-free non-commerical java distributable ?
06:54 sabdfl # to your rights to use the software. Also, please note that software in universe
06:54 mdz_ (transcribing this into a bug report btw)
06:54 sabdfl # WILL NOT receive any review and updates from the Ubuntu security team.
06:54 sabdfl how's that sound?
06:54 mdz_ s/and/or/
06:54 mdz_ and I'm happy
06:55 Mithrandir sabdfl: I would have it even scarier, if possible.
06:55 sabdfl Kamion: i'll file a bug on base-config
06:55 mdz_ is "under a free licence" sufficiently descriptive?
06:55 Kamion sabdfl: thanks
06:55 mdz_ sabdfl: I already filed one
06:55 Mithrandir and synaptic needs to pop that up before letting you enable the component
06:55 mdz_ https://bugzilla.ubuntu.com/show_bug.cgi?id=1564
06:56 sabdfl Mithrandir: agreed
06:56 elmo is synaptic what you get when you use the menus in gnome?
06:56 sabdfl really, we are saying "adults only"
06:56 sabdfl elmo: yes
06:57 mdz_ Keybuk: we'll have the infrastructure for that stuff in apt/aptitude/etc. in Hoary
06:57 sabdfl new school, kinda
06:57 Kamion note that we aren't localizing that text in any way
06:57 sabdfl synaptic will also show ALL universe packages with a clear label (skull and crossbones, i think, was mvo's choice)
06:57 Kamion that's so NOTWARTY it's untrue, though
06:58 elmo ?? "get your warez" here? that's not very _discouraging_ to most people :p
06:58 sabdfl we didn't manage to get the aptitude guy interested in being able to set a colour based on component though
06:59 sabdfl now let's be bold about what goes into universe
07:00 mdz_ ok, so the resolution on classification was known-free and "other"?
07:00 mdz_ and we'll update the comments in the default sources.list
07:00 mdz_ and bring in the patches to synaptic with the same text
07:01 mdz_ and also (in synaptic) the supported/unsupported classification in the UI
07:01 Keybuk sounds reasonable to me
07:01 mdz_ one hour, one agenda item down :-)
07:01 elmo so, hang on, sorry, am I not even mutilating the section now?
07:02 mdz_ elmo: implementation details :-P
07:02 mdz_ section? origin?
07:03 mdz_ section seems appropriate
07:03 mdz_ since it's _completely useless_ for anything else
07:03 sabdfl so universe/free and universe/other?
07:03 sabdfl or universe/free and universe/unknown?
07:04 sabdfl grrrr...
07:04 mdz_ unknown > other
07:04 Kamion that's gonna make everything show up in one humungous list in synaptic
07:04 mdz_ Debian sections are not useful for finding software
07:04 Kamion it won't even be categorized to the fairly minimal extent it is in Debian, and we don't have debtags yet
07:04 sabdfl please can the standard line just be "universe" and let the philosophers use the free/unknown filter
07:05 mdz_ sabdfl: yes, I don't think the line in sources.list needs to change at all
07:05 sabdfl Kamion: universe isn't designed to be useful for substantial package selection
07:05 Kamion the section pretty much has to be universe/something I think
07:05 sabdfl something?
07:05 mdz_ "something' being the original Debian section
07:05 Kamion mdz_: that's the current state
07:06 sabdfl so where's the "free" indicator go, for those that want it?
07:06 mdz_ admin base comm devel doc editors electronics embedded games gnome graphics hamradio interpreters kde libdevel libs mail math misc net news oldlibs otherosfs perl python science shells sound tex text utils web x11
07:06 mdz_ USELESS
07:06 Kamion mdz_: more useful than one huge list
07:07 Kamion it's certainly not very good, but it's better than the absence
07:07 mdz_ is there any point in having the freeness indicator in the archive?
07:08 mdz_ I don't think we have any package management UI in Ubuntu which lets you wear freeness blinders based on anything but sources.list
07:08 mdz_ so it'd be purely informational anyway
07:08 mdz_ X-Ubuntu-Known-Free: yes
07:09 Keybuk mdz_: the only thing would be the Section ... at least that shows up in the UI
07:09 elmo yes, AFAICS, we have 3 choices: (a) split universe into universe-known-free and universe-universe, (b) abuse Section, (c) add some random meta-data tag
07:10 mdz_ so is the freeness classification worth giving up the benefits of sections, however marginal?
07:10 elmo mark hates (a), (b) and (c) are kind of useless, (b) being marginally less so
07:10 mdz_ "what's the use case?"
07:10 elmo only (a) would come close to actually helping any poor sod who wants to use ubuntu in a commercial environment
07:10 mdz_ if it's "I only want free software on my computer"
07:10 mdz_ then none of those except (a) actually do anything
07:10 elmo that's not true
07:10 elmo (b) helps if you don't use apt-get directly
07:11 sabdfl elmo: how would they enforce that, would they use archive.ubuntu.com/ubuntu warty universe/free ?
07:11 mdz_ (b) helps if you never ever use the command line
07:11 elmo sabdfl: no, unfortunately, the only way to make it sources.list based is option (a)
07:11 elmo sources.list/apt-get has no concept of sections, only components
07:12 Kamion elmo: components can't be multi-level?
07:12 Keybuk Kamion: not in sources.list ... apt doesn't filter
07:12 elmo err.. well, I dunno, to be honest.. I suppose katie could - would they work with apt
07:12 elmo +?
07:12 mdz_ universe/free, you mean?
07:12 Kamion it'd have to be hyphen rather than slash if you wanted it to be sources.list based then
07:12 mdz_ stable/non-US works fine
07:13 Keybuk mdz_: but if you put just "stable", you don't get "stable/non-US"
07:13 mdz_ Keybuk: right
07:13 mdz_ if we switch to that, we need to convert everyone's sources.list
07:13 mdz_ which is not out of the question, if it's the right solution otherwise
07:14 Keybuk so the question really does come, do we want users to be able to select in their sources.list to only get known-free universe packages? yes/no?
07:15 sabdfl yes, that's the only point i can see to making the distinction in the archive not in launchpad
07:15 mdz_ agreed
07:15 Keybuk elmo: is it possible for a package to be simultaneously in two or more sections?
07:15 Keybuk components, sorry
07:16 sabdfl is there another place we can put a flag that will be displayed in aptitude and synaptic?
07:16 sabdfl no
07:16 elmo keybuk: package/version/arch tuple? no
07:16 elmo keybuk: package, yes
07:16 elmo across releases, for example
07:17 mdz_ there's currently noplace we can put a flag which will be visible across all apt frontends
07:17 Keybuk ok, so if we do it in sources.list we're looking at splitting universe into universe and universe-maybe-not-free/multiverse/whatever? there's no other way to do that?
07:17 mdz_ if we want to support a useful freeness distinction for the user in universe, it needs to be at the sources.list level
07:17 sabdfl universe and blackhole
07:17 sabdfl forget it, this is the mess i was trying to avoid :-)
07:18 sabdfl (and yes, i know it means there's a whole other kind of mess :-)
07:19 Keybuk elmo: any suggestions?
07:19 mdz_ we provide a freeness distinction within our supported set
07:19 mdz_ (only)
07:20 sabdfl how much work would it be to add a header item to the packages file which is displayed in both aptitude and synaptic when you view the package detail?
07:20 elmo Keybuk: not really? there is only really creating a separate component available to us as an option at this point in time.. and if sabdfl doesn't want to do to that
07:20 sabdfl Free-Software: Confirmed
07:21 Keybuk elmo: having a magic universe-known-free subset component isn't possible?
07:21 sladen If a package is non-free/non-commerical, I'd prefer if dpkg/apt didn't even know about it; ...rather than having to get them to filter it out.
07:22 mdz_ sladen: the answer would be, "then don't use universe"
07:23 sabdfl hmm...
07:23 elmo keybuk: how is a "magic subset" component different from a simply separate component
07:23 elmo +?
07:23 mdz_ Debian is in the business of collecting gobs of software and dividing it into free and not free
07:23 mdz_ Ubuntu is in the business of collecting gobs of software and dividing it into supported-free, supported-non-free, and unsupported
07:23 mdz_ that's how it stands today
07:24 mdz_ if our supported set expands to contain a reasonably complete subset of useful software, the distinction becomes less important
07:24 mdz_ but as it stands, Ubuntu supported << Debian main
07:25 Keybuk yeah, we could possibly add a universe note that if people are continually installing things from universe to discuss with us about making them supported.
07:25 Keybuk we may have missed something glaringly obvious
07:25 mdz_ Keybuk: making popularity-contest work would be a good start
07:25 mdz_ been meaning to file a bug about that
07:28 sabdfl my hope is that derivatives will commit to getting stuff that is in universe, that is important to them, supported to the level we require for inclusion in main
07:28 mdz_ Keybuk: it's certain that we've missed some glaringly obvious things; a few have come up already
07:28 sabdfl then, as mdz points out, main expands to be more useful
07:29 mdz_ if we get to a point where most people can use only supported software, then that will satisfy those who want the distinction as well
07:29 mdz_ they could remove 'restricted'
07:29 mdz_ RMS will never be happy with it
07:30 mdz_ but that's almost a given
07:30 Keybuk did you read his correspondance with Jeff? He was actually quite civil
07:31 Keybuk http://lists.gnome.org/archives/foundation-list/2004-September/msg00067.html
07:32 mdz_ I blacked out at "GNU slash Linux..."
07:32 mdz_ anyway, there is a decision on the table about whether to subdivide universe
07:32 mdz_ at least, whether to subdivide universe today
07:33 sabdfl just to clarify, the old division that makes sense, if we do it, is two components
07:33 Keybuk I don't think we can do it in a useful way today, by the sounds of it
07:33 sabdfl "universe" and "multiverse"
07:33 sabdfl s/old/only/
07:34 Keybuk not without two components anyway
07:34 elmo what the heck's multiverse?
07:34 sabdfl elmo: non-free universe
07:34 mdz_ given the lack of a straightforward solution, and hazy user need, I think we can safely defer the decision until post-Warty
07:34 elmo meh
07:34 mdz_ if we want, we can set up the Hoary archive differently
07:34 mdz_ can we not?
07:35 elmo what's not straightforward about two components?
07:35 sabdfl then the apt/sources.list can at least contain meaningful options w.r.t. known-free stuff
07:35 elmo if we don't want to do it, fine, but let's at least just say that
07:36 sabdfl elmo: is two components going to take more than an extra day to arrange?
07:36 mdz_ what would we call the components?
07:36 sabdfl universe and multiverse
07:36 mdz_ by definition there can only be one universe
07:36 elmo sabdfl: no, not at all, it's just a change in the planto import non-free
07:37 mdz_ multiverse doesn't tell the user anything about what they're asking for
07:37 mdz_ universe-known-free and universe-somethingelse would be my preference
07:37 sabdfl mdz_: i'm expressly trying to avoid warranties in this regard
07:38 mdz_ sabdfl: known-free was what you said you preferred
07:38 sabdfl i really think that text should still say "be an adult and check this stuff for yourself"
07:38 sabdfl i prefer known-free to non-free, since known-free is easier to assure
07:39 mdz_ what I suggested was Debian/non-Debian, which at the same time says nothing explicitly about freeness whatsoever, and provides the same utility for people who expect it from Debian
07:39 mdz_ which is trivial to assure
07:39 mdz_ but, I suppose, that's not useful in a sources.list context because it's tied to Debian
07:40 elmo universe, universe-non-debian ?
07:40 sabdfl no, i don't want to have a component based on source
07:40 sabdfl origin, sorry
07:40 Keybuk universe-unchecked ?
07:40 elmo universe-fuckitall
07:41 sabdfl universe and multiverse. the sources.list text can do the explaining
07:42 sabdfl i *really* prefer just universe, but let's get a sense of opinions
07:42 elmo from where/how?
07:42 sabdfl technically, this is just about mdz, keybuk and me, buti'd like to hear from others
07:42 sabdfl (a) just universe
07:42 sabdfl (b) universe and multiverse
07:43 Mithrandir multiverse is "possibly and surely non-free"?
07:43 sabdfl what else?
07:43 mdz_ (c) sections
07:43 sabdfl Mithrandir: yes
07:43 Mithrandir I'd like b) to happen
07:44 sabdfl Keybuk:
07:44 sabdfl ?
07:44 elmo I [obviously] vote (b), for whatever that's worth
07:44 Keybuk b, I think
07:44 mdz_ (b) is OK with me, except for the multiverse name
07:45 elmo mdz: is it really any worse than "universe"?
07:45 mdz_ elmo: universe and multiverse side-by-side? yes, absolutely worse
07:45 npmccallum the two do seem entirely ambiguous
07:46 mdz_ I think that universe-foo and universe-bar would be clearer
07:46 mdz_ but once a universe is subdivided, it's not a universe anymore :-)
07:46 elmo that's why multiverse is so cute
07:46 mdz_ anyway, my argument is semantic; I don't have any problems with the component implementation of the distinction
07:47 sabdfl i think the answer is that universe is "unsupported", period
07:47 sabdfl launchpad can have exact licence terms
07:47 Mithrandir mdz_: agree, we can use pwgen-based names for all I care. ;)
07:48 mdz_ as I mentioned before, launchpad answers the needs of derivative distributions, but not of users, at least not directly
07:48 mdz_ but I suppose someone could then create a "frubuntu"
07:48 mdz_ and go about classifying all the software there
07:49 elmo [are planning on supporting dervitatives who want to dervive "all of universe"? ]
07:49 mdz_ and they could call it "Frubuntu GNU/Linux", and maybe it really _would_ be possible to please everyone
07:49 mdz_ elmo: derivatives should be able to select any packages they want from universe, up to and including all of it, no?
07:50 elmo mmk
07:50 sabdfl elmo: sure
07:51 mdz that lets us leave universe alone for Warty, and also lets someone else do the work of classifying licenses
07:51 mdz and leaves room for more than one group of users to make their own classifications
07:51 elmo they can do that regardless of what we do
07:51 mdz (d) consider this a use case for a derived distribution
07:52 mdz (and not for Ubuntu itself)
07:53 sabdfl i can live with (b)
07:53 sabdfl i think it's simpler to go with (a) for warty and review for hoary
07:54 Keybuk yeah, I'm happy with that
07:54 sabdfl Keybuk: which?
07:54 mdz it all depends on what fraction of the Debian community considers the freeness bit important to them
07:54 mdz that subset won't look twice at Ubuntu if it doesn't have a distinction
07:55 Keybuk sabdfl: I prefer b, but would be happy with a
07:55 mdz but, I suppose we have a partial answer in main/restricted
07:56 mdz so a secondary decision is whether or not to do anything about the issue _for Warty_
07:56 sabdfl mdz: i think everyone considers it important
07:56 elmo I do think there's community backlack potential when we ship the 178 non-free packages from Debian in 'universe'.. mitigated maybe by a promise to review it for hoary.. *shrug*
07:56 sabdfl mdz: let's go with b, and just do the work to get the apt sources.list text right
07:57 mdz ok
07:57 Mithrandir sabdfl: as well as fixing synaptic. :)
07:57 sabdfl i think it's an unnecessary risk
07:57 mdz elmo says it's no more work, but if the list of elmo's tasks in my head is half as long as the complete one...
07:57 mdz elmo: speaking of which, remind me later to nag you about access to buildd logs
07:58 elmo remind you to nag me to do stuff I haven't done yet? yeah, 'cos _that's_ going to happen :-P
07:58 elmo (but yes, I know)
07:59 sabdfl so let's move on guys, time is limited
07:59 mdz so the consensus is that elmo should carry out (b)
08:00 sabdfl yes please
08:00 mdz ok, moving on then
08:00 sabdfl question: what are the criteria for "universe"? everything else is in "multiverse"?
08:00 mdz should be the same criteria as for main vs. restricted
08:00 mdz or, main and not-main
08:01 Keybuk sabdfl: universe is known to be free, I guess
08:01 elmo what about contrib?
08:01 sabdfl yes but how do you define free? dfsg? rms?
08:01 elmo sabdfl: ours
08:01 sabdfl ok
08:01 elmo it's our universe, damn it ;-)
08:01 Keybuk didn't we have our own guidelines for that?
08:01 Keybuk same as supported, I'd say
08:01 sabdfl ok, moving along
08:02 Keybuk universe is free enough for us, just as supported is ... people with stricter requirements make a perfect derivative distro
08:02 mdz the ESD issue, I think, was _mostly_ resolved in bugzilla overnight
08:02 Keybuk mdz: bug#?
08:02 mdz we should move everything to ESD, except for things which don't support it, and those things just suffer
08:02 Keybuk the totem/gstreamer thing?
08:02 mdz https://bugzilla.ubuntu.com/show_bug.cgi?id=1481
08:02 mdz gnomemeeting was pointed out as an app which doesn't support esd
08:03 elmo sorry, -ECONTEXT, why do we want to use ESD?
08:03 sladen esddsp
08:03 mdz elmo: #1481
08:03 Keybuk elmo: most alsa drivers don't allow concurrent access to /dev/dsp
08:03 mdz as a means to share the audio device
08:03 Keybuk if you have sound events on, esd hogs /dev/dsp
08:04 sabdfl is there a sense of momentum in the sound community on this issue?
08:04 sabdfl if the world+dog are moving to ALSA-native, we should not go in the other direction
08:04 mdz I think the momentum is in favour of sound hardware which can do mixing on its own, but it's not widespread yet
08:05 mdz it's a hardware issue that needs a software solution
08:05 sabdfl hardly seems fair to require hardware changes
08:05 mdz I don't think that ALSA-native buys us anything over OSS-ALSA in this case
08:05 Mithrandir oss-alsa is actually a fairly nice solution, IME.
08:05 Keybuk sabdfl: I think you get the same problem alsa-native anyway
08:06 Mithrandir Keybuk: that depends on your hardware.
08:06 Keybuk esd sounds like a reasonably warty solution to me; it may be worth getting someone to talk to (e.g.) the Fluendo guys on the hoary timeframe to see if there's a "better way" ?
08:06 mdz there are some vague rumblings about libasound doing software mixing, but I can't see how it could be implemente dalready
08:06 Mithrandir mdz: it's also rumored to suck badly
08:06 sladen mdz: (Its ugly), but have gnomemeeting depend on esound-clients and wrap gnomemeeting with esddsp gnomemeeting
08:06 Keybuk Mithrandir: certainly I get the same problem with alsa-native on my desktop ... but don't get either problem on my laptop :(
08:06 seb128 GNOME guys would like to kill esd I think
08:06 mdz it would require some IPC facility which doesn't exist
08:07 mdz esd can't die until something replaces it, though
08:07 Mithrandir Keybuk: my desktop has a decent sound card; I have hardware mixing on all my desktops, but not my laptop, I think.
08:07 mdz software mixing in kernelspace sounds unlikely to happen
08:07 mdz so when it is re-solved in userspace, it will probably be via the alsa API
08:08 mdz people who have hardware mixing aren't hurt by ESD
08:08 mdz and it helps people who don't
08:08 mdz my laptop seems to have hardware mixing
08:08 mdz at least, it allows multiple opens of /dev/dsp
08:08 elmo mdz: that's not true, esd does interesting things to my desktop machine
08:08 mdz elmo: such as what?
08:09 seb128 the problem is that esd is not really in a good state and not really maintained
08:09 elmo mdz: hmm, it's been a while since I stopped using it.. but initially, I think it just didn't work at all.. even after Ryan hacked around that, it would hang on signficant load
08:09 Mithrandir seb128: got a better suggestion?
08:10 mdz seb128: what is upstream's recommendation?
08:10 seb128 no, we are stucked with it for the moment ...
08:10 mdz ok, so we have no choice. sounds like we don't need to take an explicit decision
08:10 sabdfl done. next.
08:10 seb128 mdz: keep using esd for the moment since there is nothing else in this area
08:10 mdz a more pressing issue is what to do about the effect that the automounting magic has had on desktop usability
08:10 seb128 if you want to have sound events
08:10 mdz and our other desktop changes
08:10 mdz https://bugzilla.ubuntu.com/show_bug.cgi?id=980
08:10 seb128 but I think that most of the GNOME guys just don't use sound events
08:10 mdz https://bugzilla.ubuntu.com/show_bug.cgi?id=1234
08:11 mdz there is a usability issue in that users have no discoverable way to unmount things which get automounted
08:11 mdz and there is a functionality issue in that it interferes with CD burning
08:11 mdz the former issue arose because we removed the volume icons on the desktop without providing a replacement
08:12 sabdfl guys i have to step away at this point, please send me the logs
08:12 mdz sabdfl: before you go, your opinion for the record on restoring the volume icons on the desktop when things are automounted?
08:13 mdz that is by far the simplest solution
08:13 mdz ok, guess I missed him
08:14 mdz other opinions?
08:14 Keybuk tricky ... we seem to be getting positive feedback for not having the desktop icons ... but I do feel we need some immediate visual indication somewhere on screen that there's a CD mounted
08:14 mdz the consensus long-term solution in Oxford was to have a tray icon
08:15 mdz but unless someone can build that and have it stable for Warty final, we need an interim solution
08:15 mdz (tray icon? notification icon? whatever they're called)
08:15 Keybuk on a three week timescale? I don't think that's doable :(
08:15 npmccallum mdz: it won't be able to be stable by final
08:15 seb128 no, that's not
08:15 Keybuk mdz: "applets" if you listen to Novell <g>
08:15 mdz any proposals other than restoring the volume icons?
08:16 seb128 no ...
08:16 npmccallum no
08:16 mdz auto-unmounting when nautilus is closed?
08:16 mdz menu item in nautilus window?
08:16 npmccallum too risky at this stage
08:16 npmccallum (the first one that is)
08:17 seb128 menu item no
08:17 Keybuk not enough testing :-/ stuff like that needed to be done by freeze to get enough wide-testing for warty
08:17 seb128 agreed
08:17 mdz thing is, I think if sabdfl were here, he would veto the icons
08:18 Kamion I'm not sure auto-unmounting is very good UI either
08:18 seb128 if somebody has idea for the umount when all the windows are closed ...
08:18 Keybuk Kamion: mounting/unmounting in *general* is not good UI :-(
08:18 mdz Kamion: agreed, it makes for an awkward transition to the long-term solution too
08:18 Kamion seb128: that would only be good UI if there were a discoverable way to mount it again without having to eject/reinsert
08:18 sladen volume items on the desktop---for automounted items only---seems logical for me as a user; and sensibile given the time constaints and having already been tested
08:18 seb128 Kamion: true
08:19 mdz consensus, given release constraints, is that the icons need to come back
08:19 Keybuk Kamion: indeed, it would need to be unmount & eject for CDs
08:19 mdz but we need a fallback position given that sabdfl hasn't spoken on this
08:19 Keybuk which is actually kinda cute <g> you don't need to unmount usb devices because they can be pulled out manually
08:19 seb128 just having computer place for this is that bad ?
08:19 Keybuk but anyway
08:19 mdz and I know he feels strongly about this subject
08:19 Keybuk seb128: computer/disks is a bit hard to find
08:19 mdz seb128: the problem is that users can't find it
08:20 mdz because there is no indication on the screen when they insert something
08:20 seb128 write in BIG in the doc :)
08:20 Keybuk and until you find it, you really can't eject a CD
08:20 mdz there is nothing to lead them there
08:20 Keybuk seb128: who reads them?
08:20 mdz and ejecting a CD is a damn common use case :-)
08:20 seb128 Keybuk: we don't have any good solution ...
08:20 Keybuk mdz: yeah, I wish Mark hadn't wandered off at this exact point because he does have quite strong ideas about this area :-
08:21 seb128 so just umount when all the windows are closed
08:21 sabdfl voice of sabdfl from a distance.... "can we get them on the desktop for automounted ones only? so nothing there till you automount it..?"
08:21 seb128 so at least you can eject CD with the button
08:21 npmccallum yes
08:21 mdz sabdfl: yes
08:21 sabdfl i don't know how you eject a cd from windows, without pressing the hardware button
08:21 Keybuk sabdfl: as I understand it, yes; you'd only see CD/USB devices on the desktop
08:22 npmccallum sabdfl: right click, eject
08:22 mdz sabdfl: that's precisely what we're proposing
08:22 sabdfl "ugly, this is"
08:22 mdz empty desktop except when you insert a CD or hotplug something
08:23 sabdfl how do you feel about going with that for warty and fix it for panel in hoary?
08:23 npmccallum mdz: if there is a floppy drive in fstab, I'm pretty sure it will show up on the desktop
08:23 mdz that would be my preference
08:23 Keybuk yeah, I'd say that's the best we can do
08:23 seb128 mine too
08:23 mdz npmccallum: oh? we can probably fix that
08:23 mdz seb128: what's the gconf magic to turn it on? I'll try right now
08:23 Kamion seb128: not to be a powermac weenie, but what button? :)
08:23 mdz Kamion: the "F12" button
08:24 sladen npmccallum: floppy drives are a pain, since on x86 there's no way to get insert/eject notification
08:24 Kamion mdz: in theory ...
08:24 npmccallum sladen: yes, I hate them
08:24 seb128 mdz: /apps/nautilus/desktop/volumes_visible
08:24 mdz I'm more than happy to hide the floppy away
08:24 sabdfl i agree on the understanding the desktop is empty on boot, until something is explicitly automounted
08:24 npmccallum we can hide the floppy I'm sure
08:24 npmccallum thats a lot less intrusive than the other stuff
08:24 seb128 yes, should be easy to do
08:24 sabdfl cheers guys. community council next week.
08:24 Keybuk hrm, I don't have a floppy icon on my desktop until I mount the floppy
08:24 npmccallum Keybuk: is it in fstab?
08:25 seb128 that's the right behaviour
08:25 Keybuk npmccallum: yes
08:25 seb128 only mounted devices
08:25 Keybuk /dev/fd0 /floppy auto rw,users,noauto,sync 0 0
08:25 sladen only _user_ mounted devices
08:25 seb128 yes
08:25 mdz seb128: I did gconftool-2 --type boolean -s /apps/nautilus/desktop/volumes_visible 1
08:25 mdz and it doesn' tseem to have had an effect
08:26 seb128 nautilus manages your desktop ?
08:26 mdz stock Warty
08:26 Keybuk mdz: unless you have something mounted, you shouldn't see anything
08:26 mdz Keybuk: I inserted a CD
08:26 mdz and nautilus opened with it
08:26 mdz but no icon
08:26 sladen mdz: s/1/true/ ?
08:27 seb128 no, 1 is fine
08:27 mdz do I need to log out or something?
08:27 seb128 no
08:27 mdz ah, there it is
08:27 mdz I re-ran it with true
08:27 mdz ejected, inserted
08:27 mdz and there it is
08:27 mdz ok
08:27 mdz and no floppy
08:27 mdz so, let's do that for Warty
08:27 seb128 ok
08:27 mdz seb128: will you take care of it?
08:27 npmccallum I already did it
08:28 mdz ok
08:28 seb128 mdz: yes. Could you just open a bug to keep the trace ?
08:28 npmccallum just waiting for upload
08:28 seb128 oh ok
08:28 seb128 fine
08:28 mdz npmccallum: go ahead, if that's the only change
08:28 npmccallum yup
08:28 mdz just removing the patch from debian/patches, right?
08:29 npmccallum edited it -- cut out the section refering to volume icons only
08:29 Keybuk yeah, it works for me too here; only the mounted stuff shows up, there's no icon for the cd or floppy when not mounted
08:29 mdz ok
08:29 mdz moving on, we have the CD burning issue
08:29 mdz if you've burned CD-RWs with nautilus in Warty, you know what I'm talking about
08:30 Keybuk is this issue when you insert a blank CD? or when you want to write over top of a CD you've just inserted that's got stuff on it?
08:30 mdz Keybuk: only when the CD has stuff on it
08:30 mdz hence CD-RW
08:30 Keybuk can't nautilus' cd burner unmount the thing itself?
08:30 mdz you insert the CD, and it gets mounted and nautilus opens
08:30 mdz dunno
08:30 mdz seb128, npmccallum: ?
08:31 Keybuk that seems the most logical ... user-wise you probably *do* want the window open to check you're about to write over the right CD
08:31 npmccallum it can unmount it, the problem is it automatically gets mounted again on burn completion (if that is a problem)
08:31 Keybuk then when you go to write, the burner app should unmount it
08:31 mdz the other problem is that, even if you unmount it, then it gets mounted again _during_ and after the burning process
08:31 mdz at least for me
08:31 Keybuk npmccallum: getting unmounted after burning sounds right? check it burnt ok?
08:32 seb128 my cd writer is out of order, I've not really played with n-c-b ...
08:32 Keybuk during would be broken though?
08:32 sladen mdz: unmount + eject
08:32 Keybuk npmccallum: does the window pop up after it's finished burning?
08:32 mdz ideally, it should unmount at start, prevent mountitng during, and eject after
08:32 Kamion even if you *can* unmount it ... last time I tried, I only had an "Eject" entry on the right-click menu, no "Unmount"
08:32 npmccallum Keybuk: I believe so
08:32 mdz if the user wants to check it, they can close the CD tray
08:32 Keybuk Kamion: eject is friendlier anyway
08:32 mdz Kamion: my sr0 has an 'unmount'
08:33 mdz and when i unmount, the icon goes away
08:33 Kamion Keybuk: not if you're trying to burn (say I want to do something strange and therefore want to use cdrecord)
08:33 mdz so there's no place for an eject
08:33 npmccallum keeping it from mounting during burn is the bugger
08:33 mdz there is a separate issue of the command line
08:33 mdz it is very irritating to mount something on the command-line and have it pop up in nautilus on top of your terminal
08:34 Keybuk Kamion: if you're doing that, I'd expect you'd know to unmount it first
08:35 Keybuk what decisions need to be made here? before we wander too much
08:35 mdz the decision to be made is what to do about CD burning
08:35 mdz because that is a real use case which is in bad shape right now
08:35 mdz options:
08:35 mdz (a) work around it in nautilus-cd-burner as best we can
08:36 mdz (b) implement a locking facility in pmount (pitti has written the code for this already)
08:36 mdz (c) use locking in HAL, which upstream has in CVS
08:36 mdz (b) and (c) are sketchy given the release
08:36 Keybuk seb128, npmccallum: you're suggestions?
08:36 mdz locking never gets implemented correctly the first time
08:37 seb128 I can have a look on (a) tomorrow
08:37 npmccallum c is out, I don't want to pull cvs in at this point
08:37 Keybuk this is an upstream-affecting bug as well? there's nothing too magic about our g-v-m/nautilus other than we open windows?
08:38 npmccallum could we have whatever calls n-cd-b stop gvm and restart gvm when finished?
08:38 npmccallum Keybuk: the open windows is upstream now
08:38 seb128 sorry guys I've to go now
08:38 npmccallum seb128: see ya
08:38 Keybuk seb128: thanks! have fun
08:39 seb128 I'll read the log, I can talk with n-c-b upstream about this tomorrow if you want
08:39 mdz seb128: I'll assign the bug to you
08:39 seb128 ok fine
08:39 sladen npmccallum: what's the case when another volume is inserted whilst gvm is stoppped? Would gvm then be out of sync after it's restarted?
08:39 seb128 good evening guys
08:39 npmccallum sladen: it only handles events
08:39 Keybuk I don't have any major suggestions, I don't own a burner so haven't ever really tried that stuff out and seen the issues :-/
08:39 npmccallum sladen: so if gvm is stopped, it won't receive the event
08:40 npmccallum sladen: the events aren't cached
08:40 Keybuk it sounds to me like there's a whole pile of hal "I'm using this" embargo stuff that needs to be done here though
08:40 mdz ok, the only other thing I had on the agenda is a bit of an excuse to get several people together to talk about what is otherwise not a technical-board-scoped sort of bug
08:40 mdz Keybuk: that's happening upstream
08:40 mdz anyway, we have a mess with spell checking and dictionary stuff in warty at the moment
08:41 mdz spell checking doesn't seem to work in openoffice, evolution and such out of the box
08:41 mdz there are about 40 different spell checking libraries all with their own dictionaries
08:41 mdz and we need someone to own that problem and figure out what needs to be done
08:41 Keybuk spell checking in *general* is a bit of a mess
08:41 mdz volunteers?
08:43 mdz Mithrandir: if you want it, I'm more than willing to give it, but it's a bit out of scope
08:43 Mithrandir I could do it.. I have a bit of experience with it as I maintain the norwegian package in Debian
08:43 mdz npmccallum: will you work with Mithrandir on this?
08:43 Mithrandir mdz: "want", isn't exactly the word. :P It's a hairy issue.
08:44 npmccallum mdz: I have a bunch of bugs right now. I'm happy to take any patches to ooo, but can't really work on it per se
08:44 sladen what's the long-term goal here? Hack the various spelling libraries all to use the same backend database?
08:44 mdz npmccallum: currently, you and martin have the fewest RC bugs
08:45 sladen or have one database file and transcode that into the format for each library?
08:45 mdz let's not talk about long term just yet :-)
08:45 mdz for Warty, we just need for spell checking to work
08:45 Mithrandir sladen: it's fairly easy to generate word lists if you have ispell lists.
08:45 Mithrandir for the different brands of checkers
08:46 sladen is the simple answer of just shipping the 20 different libraries and their backends okay. Might work for English, not sure about others as there is unlikely to be a full set of data files for each library in each language
08:46 mdz npmccallum: are you working on issues which you feel are RC but not reflected that way in bugzilla?
08:46 npmccallum mdz: I have no idea how the spell checking even works, though I can try to learn it
08:46 Mithrandir sladen: : tfheen@yiwaz ~ > apt-cache showsrc norwegian | grep ^Binary
08:46 Mithrandir Binary: inorwegian, myspell-nb, aspell-no, myspell-nn, wnorwegian
08:47 mdz I don't think anyone has a handle on it at this point; that's the problem
08:47 npmccallum mdz: I also have a batch of initscripts which mark found that have to get done (not filed in bugzilla yet)
08:47 Kamion there's the issue of installing the proper set of dictionaries for the selected language, too
08:47 mdz npmccallum: please file them
08:47 npmccallum mdz: will do
08:47 mdz Kamion: yes, but if we could have english spell checking working, that't be a start
08:47 Kamion mdz: right
08:47 sladen Mithrandir: okay, next question, were you able to automatically generate those (from one source) when you created the packages?
08:48 mdz ok, not everyone is represented here; we'll take this offline
08:48 npmccallum mdz: let me get those initscripts whittled down and I'll be glad to work on the dictionary stuff, I just don't want someone else to take if they have the know how and the time
08:48 mdz Mithrandir: I want someone full-time to work on the issue with you
08:48 Mithrandir sladen: yes. I'm not fucking crazy out of my mind. Only slightly so.
08:48 Mithrandir mdz: ok
08:48 mdz any other issues to raise before we close the meeting?
08:48 sladen Mithrandir: ah ha! So there's already code to do that?
08:48 Mithrandir sladen:
08:48 Mithrandir cat norsk.mch | ispell -d ./norsk -e | \
08:48 Mithrandir perl -lane 'for (@F) {s/"(.)/$1$1-/g; print; s/($1)$1-/$1/g; print }'|\
08:49 Keybuk mdz: we're meeting every two weeks at about the same time?
08:49 Mithrandir aspell --local-data-dir=`pwd` --lang=nb create master ./nb.rws
08:49 Mithrandir and
08:49 Mithrandir ( wc -l words.norsk; cat words.norsk ) > nb_NO.mydict
08:49 Mithrandir ispellaff2myspell --charset=latin1 --split=200
08:49 Mithrandir ugly but works.
08:49 mdz Keybuk: it sounds like we should adjust the time
08:49 mdz given that Jeff wasn't able to show
08:49 mdz but yes, every two weeks
08:49 mdz hopefully most will be shorter than this
08:50 Keybuk should we have a mailing list for pre-suggesting stuff, or just stay with a wiki page?
08:50 Mithrandir it would be nice if we could schedule them a bit more than ten minutes in advance
08:50 mdz wiki and/or ubuntu-devel seems fine for now
08:50 mdz Mithrandir: yes, we'll get into a regular schedule
08:50 mdz probably on the same day and time every two weeks
08:50 mdz and announced to ubuntu-devel
08:51 Mithrandir great.
08:51 Keybuk *nods*
08:51 Mithrandir should we have an ubuntu-devel-announce?
08:51 mdz ubuntu-devel is very low-traffic right now; maybe later
08:51 Mithrandir ook
08:51 mdz I think it has something to do with the fact that we haven't announced it anywhere except the mailman page ;-)
08:52 mdz but the signal/noise is delicious right now
08:52 Keybuk heh, I admit that I can't keep up with -users at the moment
08:52 mdz Keybuk: yeah, it didn't take long, did it?
08:52 mdz it's up to about 200 messages/day already
08:52 mdz anyway, meeting adjourned, further chitchat to #ubuntu.*
08:52 mdz thanks, everyone
08:53 Keybuk yay, I can get up off the kitchen floor now and head for the gym <g> MeetingLog/Ubuntu/2004-09-21 (last edited 2008-08-06 16:39:27 by localhost)