Meeting:Packaging IRC log 20061128

From FedoraProject

Jump to: navigation, search
[Tue Nov 28 2006]  [11:00:03]  <tibbs>	So....
[Tue Nov 28 2006]  [11:00:22]  <lutter>	tibbs: what's up ?
[Tue Nov 28 2006]  [11:00:45]  <tibbs>	Meeting should be about now.
[Tue Nov 28 2006]  [11:00:57]  <tibbs>	Odd, what happened to the topic for this channel?
[Tue Nov 28 2006]  [11:02:18]  Join	racor has joined this channel (n=rc040203@Tb0f6.t.pppool.de).
[Tue Nov 28 2006]  [11:03:47]  Join	abadger1999 has joined this channel (n=abadger1@090.164-78-65.ftth.swbr.surewest.net).
[Tue Nov 28 2006]  [11:03:56]  <lutter>	doesn't look good for a quorum again
[Tue Nov 28 2006]  [11:04:11]  <abadger1999>	Hey all
[Tue Nov 28 2006]  [11:04:24]  <racor>	hi
[Tue Nov 28 2006]  [11:04:30]  <tibbs>	I see four.  I guess spot doesn't ever set himself as being away.
[Tue Nov 28 2006]  [11:04:57]  <tibbs>	And abadger1999's away even though he's here.
[Tue Nov 28 2006]  [11:05:10]  <spot>	i'm here
[Tue Nov 28 2006]  [11:05:20]  <spot>	sorry, distracted by internal RHAT stuff happening
[Tue Nov 28 2006]  [11:05:24]  <tibbs>	Wow, spot!
[Tue Nov 28 2006]  [11:05:37]  <abadger1999>	I'm here.
[Tue Nov 28 2006]  [11:05:41]  	 * spot has been neck deep in work and Aurora lately
[Tue Nov 28 2006]  [11:06:30]  <spot>	so, before i start trolling the depths of the todo list, any business items?
[Tue Nov 28 2006]  [11:06:52]  <tibbs>	There was question about thimm's membership.
[Tue Nov 28 2006]  [11:07:09]  <tibbs>	Probably best discussed on the list, though, so that he can participate.
[Tue Nov 28 2006]  [11:07:24]  <spot>	yeah, i think thimm wants to step aside
[Tue Nov 28 2006]  [11:09:06]  <tibbs>	Obviously he'll need to be replaced if he does leave the committee.
[Tue Nov 28 2006]  [11:09:52]  <tibbs>	I don't recall that we have an actual policy for how to replace him.
[Tue Nov 28 2006]  [11:11:22]  <abadger1999>	Does at large nominations followed by package committee vote not cover this instance?
[Tue Nov 28 2006]  [11:11:44]  <tibbs>	Probably.
[Tue Nov 28 2006]  [11:12:03]  <tibbs>	But does anyone want to spend time on this today?
[Tue Nov 28 2006]  [11:12:28]  <abadger1999>	List is fine.
[Tue Nov 28 2006]  [11:12:34]  <tibbs>	spot?
[Tue Nov 28 2006]  [11:13:23]  <spot>	sorry. internal stuff keeps getting in my face.
[Tue Nov 28 2006]  [11:14:02]  <spot>	all of the items on the Todo list which say "Ratify" need to be written up
[Tue Nov 28 2006]  [11:14:06]  <tibbs>	Indeed, that's the case for everyone.
[Tue Nov 28 2006]  [11:14:47]  <tibbs>	We should see if thimm wants to write up his approved item or if someone else should do it.
[Tue Nov 28 2006]  [11:15:15]  <spot>	do we want to reconsider making the Group tag optional for FC7+ ?
[Tue Nov 28 2006]  [11:15:38]  <tibbs>	Will rpm in rawhide deal with it?
[Tue Nov 28 2006]  [11:15:46]  <spot>	with a one line patch it will.
[Tue Nov 28 2006]  [11:16:06]  <spot>	paul said he'd take the patch in devel if the committee passes the "optional" resolution
[Tue Nov 28 2006]  [11:16:47]  <tibbs>	I seem to recall this committee agreed that we wanted to move in that dirction.
[Tue Nov 28 2006]  [11:16:55]  <spot>	yeah, but we never voted on it
[Tue Nov 28 2006]  [11:17:02]  <spot>	thus, its been sitting on the list for a while
[Tue Nov 28 2006]  [11:17:28]  <tibbs>	I don't think it has any downsides besides not currently being supported by rpm.
[Tue Nov 28 2006]  [11:17:38]  <tibbs>	Chicken and egg problem, that.
[Tue Nov 28 2006]  [11:18:02]  <tibbs>	It would have to be a unanimous vote today, though.
[Tue Nov 28 2006]  [11:18:10]  Join	Rathann|work has joined this channel (n=rathann@gw.icm.edu.pl).
[Tue Nov 28 2006]  [11:18:14]  <spot>	well, lets see if it passes
[Tue Nov 28 2006]  [11:18:30]  <tibbs>	Make Group: tag optional: +1
[Tue Nov 28 2006]  [11:18:37]  <abadger1999>	+1
[Tue Nov 28 2006]  [11:18:37]  <spot>	+1
[Tue Nov 28 2006]  [11:19:17]  <abadger1999>	lutter, racor?
[Tue Nov 28 2006]  [11:19:33]  <lutter>	abadger1999: sorry .. baby started crying .. lemme read back
[Tue Nov 28 2006]  [11:19:58]  <racor>	confused - what are we voting on?
[Tue Nov 28 2006]  [11:20:24]  <tibbs>	Making the Group: tag optional.
[Tue Nov 28 2006]  [11:20:24]  <f13>	oh crap, a meeting :/
[Tue Nov 28 2006]  [11:20:28]  	 * f13 was on another call.
[Tue Nov 28 2006]  [11:20:34]  <spot>	making the Group tag optional in FC7+
[Tue Nov 28 2006]  [11:20:35]  <lutter>	+1 for making the group tag optional
[Tue Nov 28 2006]  [11:20:50]  <f13>	spot: doesn't that require RPM changes?  Didn't we go through this before?
[Tue Nov 28 2006]  [11:20:57]  <spot>	f13: a one line patch
[Tue Nov 28 2006]  [11:21:06]  	 * f13 could have sworn we voted on this already once, and it was blocked by rpm changes that Paul wouldn't apply
[Tue Nov 28 2006]  [11:21:07]  <spot>	paul said he'd take it if the committee approved the optional resolution
[Tue Nov 28 2006]  [11:21:19]  <spot>	he also didnt want to slide it into fc6
[Tue Nov 28 2006]  [11:21:26]  <spot>	which was the original thought
[Tue Nov 28 2006]  [11:21:43]  <f13>	ok, so we're just taking the vote we already approved and moving it to say FC7+ ?
[Tue Nov 28 2006]  [11:21:49]  Join	rdieter has joined this channel (n=rdieter@sting.unl.edu).
[Tue Nov 28 2006]  [11:21:51]  <racor>	why?
[Tue Nov 28 2006]  [11:22:04]  <spot>	racor: because Comps is what matters, Group is a throwaway tag
[Tue Nov 28 2006]  [11:22:26]  <racor>	comps is irrelevant to pure rpm
[Tue Nov 28 2006]  [11:22:44]  <spot>	racor: nothing uses Group anymore. anaconda doesnt. yum doesnt.
[Tue Nov 28 2006]  [11:22:51]  <racor>	apt does
[Tue Nov 28 2006]  [11:22:53]  <f13>	spot: not nothing.  smart does and apt.
[Tue Nov 28 2006]  [11:23:09]  <f13>	but they like to duplicate grouping information.  because you know, its not confusing or anything
[Tue Nov 28 2006]  [11:23:14]  	 * rdieter looks in late...
[Tue Nov 28 2006]  [11:23:22]  <spot>	f13: they use both comps and group?
[Tue Nov 28 2006]  [11:23:28]  <f13>	spot: I do believe so
[Tue Nov 28 2006]  [11:23:32]  <spot>	eww.
[Tue Nov 28 2006]  [11:23:36]  <f13>	indeed
[Tue Nov 28 2006]  [11:23:49]  <spot>	well, the alternative to making Group optional is to try to standardize on a set of Groups.
[Tue Nov 28 2006]  [11:23:50]  <f13>	racor: thats the big problem here, we've got two different ways of expressing what group a package belongs in.
[Tue Nov 28 2006]  [11:24:03]  <spot>	i personally feel this is akin to rearranging the deck chairs on the titanic
[Tue Nov 28 2006]  [11:24:10]  <f13>	racor: comps gives way more control over this than the rpm itself.  and as such the rpm itself is often incorrect or doesn't match comps
[Tue Nov 28 2006]  [11:24:20]  <spot>	but if the committee feels that way, i'll do it.
[Tue Nov 28 2006]  [11:24:24]  <racor>	f13, so far comps is RH proprietary, rpm is not
[Tue Nov 28 2006]  [11:24:35]  <f13>	spot: don't we already have a standard set of groups?
[Tue Nov 28 2006]  [11:24:41]  <f13>	racor: and this is a Fedora packaging committee.
[Tue Nov 28 2006]  [11:24:47]  <spot>	racor: how is an xml file any more or less than proprietary than groups?
[Tue Nov 28 2006]  [11:24:48]  <f13>	racor: of which, comps exists and is used extensively
[Tue Nov 28 2006]  [11:24:56]  <spot>	especially since Fedora ships with it
[Tue Nov 28 2006]  [11:25:18]  <spot>	(and every Fedora cvs user has commit access to it)
[Tue Nov 28 2006]  [11:25:32]  <f13>	spot: So I think there are 3 options here.
[Tue Nov 28 2006]  [11:25:46]  <spot>	f13: not really. we have a list of "loose recommendations"
[Tue Nov 28 2006]  [11:25:46]  <f13>	1) Make groups optional, don't really pay attention to it during review
[Tue Nov 28 2006]  [11:26:05]  <f13>	2) Make groups required, have a strict set of possible groups
[Tue Nov 28 2006]  [11:26:18]  <f13>	3) Ignore groups all together
[Tue Nov 28 2006]  [11:26:35]  <racor>	4) derive comps from rpm's groups etc.
[Tue Nov 28 2006]  [11:26:53]  <f13>	racor: doesn't work.
[Tue Nov 28 2006]  [11:26:57]  <racor>	why
[Tue Nov 28 2006]  [11:27:05]  <f13>	racor: a package could live in multiple groups in comps, plus there are group requirements in comps
[Tue Nov 28 2006]  [11:27:18]  <spot>	comps is more intelligent/comprehensive than Group
[Tue Nov 28 2006]  [11:27:23]  <f13>	racor: like package foo is in language group bar, but only make it a default if group baz is already selected.
[Tue Nov 28 2006]  [11:27:38]  <f13>	racor: also, Group: has no way to define 'default', 'manditory', 'optional'
[Tue Nov 28 2006]  [11:27:51]  <racor>	f13: like rpm
[Tue Nov 28 2006]  [11:28:42]  <spot>	we can't derive comps from groups, nor can we use the comps values to populate Group
[Tue Nov 28 2006]  [11:28:50]  <f13>	which is why none of our package management applications use rpm by itself.
[Tue Nov 28 2006]  [11:29:04]  <racor>	f13: bug?
[Tue Nov 28 2006]  [11:29:08]  <spot>	the fedora direction is towards comps, as far as i can see.
[Tue Nov 28 2006]  [11:29:08]  <f13>	racor: no.
[Tue Nov 28 2006]  [11:29:16]  <f13>	racor: rpm is a very low level tool.
[Tue Nov 28 2006]  [11:29:18]  <racor>	f13: i disagree
[Tue Nov 28 2006]  [11:29:23]  <f13>	racor: thats nice.
[Tue Nov 28 2006]  [11:29:34]  <spot>	racor: i'm sure that paul would take a patch to enable a "Comps" field in rpm
[Tue Nov 28 2006]  [11:30:04]  <rdieter>	I'd have to say, until Groups go away completely (which afaict isn't soon), we're stuck with option 2.
[Tue Nov 28 2006]  [11:30:18]  <f13>	rdieter: not really
[Tue Nov 28 2006]  [11:30:35]  <f13>	rdieter: since all the tools we use to express groupings use comps, we can really just ignore what is in the Group: field of a package.
[Tue Nov 28 2006]  [11:30:41]  <f13>	rdieter: it becomes useless data
[Tue Nov 28 2006]  [11:30:50]  <spot>	f13: except apt and smart
[Tue Nov 28 2006]  [11:30:59]  <rdieter>	f13: useless to you maybe...
[Tue Nov 28 2006]  [11:31:04]  <racor>	and synaptic
[Tue Nov 28 2006]  [11:31:09]  <f13>	spot: I'd consider it a bug that they use the Group: tag (:
[Tue Nov 28 2006]  [11:31:11]  <rdieter>	I think it's the best compromise is all.
[Tue Nov 28 2006]  [11:31:29]  <spot>	alright. i'll work on a list of "approved groups" and we can go over it.
[Tue Nov 28 2006]  [11:31:37]  <racor>	f13: RH living on an island and "cooperating" with the community
[Tue Nov 28 2006]  [11:31:46]  <racor>	:(
[Tue Nov 28 2006]  [11:31:50]  <tibbs>	So what do these nonstandard package managers do when they see a group of "(none)" or whatever RPM stores when Group: is missing?
[Tue Nov 28 2006]  [11:31:52]  <f13>	racor: no, hte Fedora project has picked its package management tools.
[Tue Nov 28 2006]  [11:32:04]  <racor>	correction: RH has
[Tue Nov 28 2006]  [11:32:07]  <f13>	racor: if people want to write their own and go off in different directions, well, bummer.
[Tue Nov 28 2006]  [11:32:21]  <spot>	tibbs: presumably the same thing that they see when they look for "Packager"
[Tue Nov 28 2006]  [11:32:25]  <f13>	racor: no, Fedora has.  THe decision to use it in the installer and everywhere else had community input.
[Tue Nov 28 2006]  [11:32:29]  <spot>	or any of the other optional fields
[Tue Nov 28 2006]  [11:32:31]  <racor>	FYI: I actively participate in apt-repomd development
[Tue Nov 28 2006]  [11:32:33]  <tibbs>	So exactly what gets broken?
[Tue Nov 28 2006]  [11:32:47]  <spot>	tibbs: presumably it would confuse the grouping it does
[Tue Nov 28 2006]  [11:32:56]  <tibbs>	People don't sensibly choose Group: as is, so it's currently useless already.
[Tue Nov 28 2006]  [11:33:12]  <f13>	racor: the Fedora project isn't going to bend over backwards to support every possible package mangement tool that understands rpm.
[Tue Nov 28 2006]  [11:33:19]  <spot>	but as i never used apt/smart for more than a few minutes without being annoyed, i'm not qualified to speak in depth
[Tue Nov 28 2006]  [11:33:30]  <tibbs>	It's not something that's checked on review.  There is no standardization.  All of the effort is in cleaning comps up.  That should be the way forward.
[Tue Nov 28 2006]  [11:34:14]  <spot>	ideally, i'd like to see Group reworked in rpm to enable packagers to store the Comps groups (note plural) in that field
[Tue Nov 28 2006]  [11:34:34]  <spot>	but thats above and beyond the mandate for this committee
[Tue Nov 28 2006]  [11:34:47]  <tibbs>	That would be semi-reasonable, except that you'd be duplicating that data.
[Tue Nov 28 2006]  [11:34:52]  <spot>	indeed.
[Tue Nov 28 2006]  [11:34:56]  <f13>	and youd still have to hand edit comps.xml
[Tue Nov 28 2006]  [11:35:08]  <spot>	f13: or that field could be autopopulated by the buildsystem
[Tue Nov 28 2006]  [11:35:11]  <spot>	from comps.xml
[Tue Nov 28 2006]  [11:35:21]  <tibbs>	Or the other way around, I guess.
[Tue Nov 28 2006]  [11:35:46]  	 * f13 refrains from going further down the rabbit hole
[Tue Nov 28 2006]  [11:35:57]  <spot>	either way, we need to do something around group now
[Tue Nov 28 2006]  [11:36:16]  <f13>	why?
[Tue Nov 28 2006]  [11:36:19]  <spot>	as it is, we've got a mixed standard of reviewers insisting on the listed Groups, and some just ignoring group entirely
[Tue Nov 28 2006]  [11:36:20]  <tibbs>	How about just picking a primary location in comps and stuffing that into Group:
[Tue Nov 28 2006]  [11:36:26]  <f13>	can't it continue to be just as irrelevant as before?
[Tue Nov 28 2006]  [11:36:39]  <spot>	f13: because its not clear that it is irrelevant.
[Tue Nov 28 2006]  [11:36:43]  <f13>	ah.
[Tue Nov 28 2006]  [11:37:03]  <spot>	and if possible, i'd like to avoid making apt/smart convoluted.
[Tue Nov 28 2006]  [11:37:33]  <f13>	isn't that too late?
[Tue Nov 28 2006]  [11:37:45]  <rdieter>	yeah, let's just standardize it, enforce it, and be done with it.
[Tue Nov 28 2006]  [11:37:45]  <spot>	well, if we standardize on a set of valid Group tags...
[Tue Nov 28 2006]  [11:38:07]  <f13>	then we _still_ have duplicate info
[Tue Nov 28 2006]  [11:38:12]  <spot>	yeah.
[Tue Nov 28 2006]  [11:38:23]  <spot>	fix rpm, then we'll revisit making that duplicate info go away.
[Tue Nov 28 2006]  [11:38:33]  <tibbs>	I don't see what those alternate package managers lose by making Group optional.
[Tue Nov 28 2006]  [11:38:41]  <f13>	and if apt/smart want to use the same things that say anaconda/pirut use for groupings, they're going to have to either ignore group, deal with conflicts, or have lots of duplication
[Tue Nov 28 2006]  [11:39:14]  <racor>	tibbs: apt doesn't use comps at all, but can filter on many other rpm-details
[Tue Nov 28 2006]  [11:39:20]  <tibbs>	So?
[Tue Nov 28 2006]  [11:39:32]  <tibbs>	I still don't see what is lost by making Group: optional.
[Tue Nov 28 2006]  [11:39:34]  <racor>	tibbs: synaptics uses rpm groups to list packages
[Tue Nov 28 2006]  [11:39:38]  <tibbs>	So?
[Tue Nov 28 2006]  [11:39:41]  <f13>	hurray for a package manager not showing the same options as the isntaller.  _thats_ not confusing at all.
[Tue Nov 28 2006]  [11:39:46]  <tibbs>	You're not getting it.
[Tue Nov 28 2006]  [11:39:55]  <tibbs>	It already contains useless information.
[Tue Nov 28 2006]  [11:40:18]  <rdieter>	tibbs: but it's not useless...
[Tue Nov 28 2006]  [11:40:27]  <tibbs>	People don't put useful things there.  They just make something up.
[Tue Nov 28 2006]  [11:40:40]  <racor>	f13: because comps is an isolated, yum/anaconda/pirut proprietary solution
[Tue Nov 28 2006]  [11:40:52]  <rdieter>	tibbs: that's why we standardize the darn thing, and be done with it.  End of discussion.
[Tue Nov 28 2006]  [11:40:58]  <tibbs>	Instead of forcing them to make up something random thing, just allow them to leave the tag off.
[Tue Nov 28 2006]  [11:40:59]  <f13>	racor: yum/anaconda/pirut are all opensource, how is this proprietary?
[Tue Nov 28 2006]  [11:41:19]  <abadger1999>	tibbs: People do put useful things in there.  Some people put non-useful things in there.  But getting rid of Groups would get rid of the useful ones as well.
[Tue Nov 28 2006]  [11:41:32]  <racor>	f13: OK nitpicking on wordings, again
[Tue Nov 28 2006]  [11:41:38]  <f13>	abadger1999: he didn't say get rid of Group.  He said making it optional.
[Tue Nov 28 2006]  [11:41:44]  <tibbs>	abadger1999: How does "making optional" equal "getting rig of"?
[Tue Nov 28 2006]  [11:41:52]  <racor>	yum etc. are not used by anybody else but RH based distros
[Tue Nov 28 2006]  [11:41:54]  <f13>	racor: the only reason apt/smart/whatever doesn't use comps is because they choose to not support it.
[Tue Nov 28 2006]  [11:41:57]  <f13>	racor: lazy.
[Tue Nov 28 2006]  [11:42:09]  <rdieter>	imo, making it optional is almost as bad as getting rid of it.  Either use it, and use it *properly*, or not at all.
[Tue Nov 28 2006]  [11:42:10]  <racor>	RH has comps , SuSE has their own config.languages
[Tue Nov 28 2006]  [11:42:29]  <tibbs>	rdieter: One thing at a time, though.
[Tue Nov 28 2006]  [11:42:35]  <racor>	f13: No. because these are using rpm, not comps
[Tue Nov 28 2006]  [11:42:46]  <f13>	rdieter: frankly I'd rather go with the last option, since duplication and non-insync data is more harmful.
[Tue Nov 28 2006]  [11:43:01]  <f13>	racor: again, their choice.
[Tue Nov 28 2006]  [11:43:02]  <racor>	f13: because they assume a packageing system is complete
[Tue Nov 28 2006]  [11:43:20]  <f13>	racor: if they assumed that, why did they have to write somethign to manage rpm?
[Tue Nov 28 2006]  [11:43:21]  <racor>	comps is "band-aid" on top of rpm
[Tue Nov 28 2006]  [11:43:33]  <f13>	if a 'packaging system' is complete, why do we need things like yum and apt to handle it?
[Tue Nov 28 2006]  [11:43:51]  <rdieter>	racor: yeah, but it's a needed bandaid.
[Tue Nov 28 2006]  [11:44:08]  <racor>	i've not used comps ever since
[Tue Nov 28 2006]  [11:44:08]  <f13>	racor: you could claim that apt is a bandaid on top of rpm and deb
[Tue Nov 28 2006]  [11:44:32]  <abadger1999>	tibbs: That is partially true.  Compared to the current situation, allowing people to just not specify Group doesn't add to the problem.  But compared to standardizing on a set of Groups it is a losing proposition (for apt).
[Tue Nov 28 2006]  [11:45:00]  <abadger1999>	Because a non-existent group is as bad as a made up category.
[Tue Nov 28 2006]  [11:45:08]  <racor>	f13: Yes, in fact it is, all such tools implement deptracking, etc. because rpm lacks it
[Tue Nov 28 2006]  [11:45:17]  <abadger1999>	It doesn't help the end user find a package that can help them do hat they want.
[Tue Nov 28 2006]  [11:45:35]  <f13>	racor: rpm lacks it, deb lacks it, slackware tgzs lack it, etc...
[Tue Nov 28 2006]  [11:45:36]  <spot>	i see two options. Either we standardize on a set of groups or we make group optional.
[Tue Nov 28 2006]  [11:46:02]  	 * rdieter nods
[Tue Nov 28 2006]  [11:46:03]  <f13>	racor: so no packaging system is 'complete' until you consider the entire stack needed.  Some people build on top of rpm+comps, some people do their own thing.
[Tue Nov 28 2006]  [11:46:27]  <spot>	So, we should vote for one of those two options.
[Tue Nov 28 2006]  [11:46:41]  <f13>	my feelings are we shouldn't waste effort on data that isn't used by the installer, the default package updater, the default package installer, etc...
[Tue Nov 28 2006]  [11:46:51]  <tibbs>	Groups optional +1.
[Tue Nov 28 2006]  [11:46:56]  <racor>	f13: rpms contain the datasets to populate a database, yum/apt work on it.
[Tue Nov 28 2006]  [11:47:19]  <f13>	Groups optional (and unreviewed) +1
[Tue Nov 28 2006]  [11:47:21]  <racor>	comps is an additional dataset, outside of the rpmdataba
[Tue Nov 28 2006]  [11:47:35]  <rdieter>	standardize Groups: +1
[Tue Nov 28 2006]  [11:47:49]  <f13>	racor: becuase package grouping is highly dependant on who is grouping the packages.
[Tue Nov 28 2006]  [11:47:58]  <racor>	standardize Group: +1
[Tue Nov 28 2006]  [11:48:03]  <f13>	racor: I may want to take the build packages and group them in one way for one product, an entirely different way in another product
[Tue Nov 28 2006]  [11:48:05]  <racor>	Groups optional -1
[Tue Nov 28 2006]  [11:48:16]  <f13>	assuming how the end user would want to group thigns at package build time is forcing their hand.
[Tue Nov 28 2006]  [11:48:21]  <racor>	f13: symbolgrounding problem
[Tue Nov 28 2006]  [11:48:29]  <f13>	package grouping belongs OUTSIDE of the rpm itself.
[Tue Nov 28 2006]  [11:48:34]  <rdieter>	racor: you only get 1 vote. (:
[Tue Nov 28 2006]  [11:48:47]  <spot>	lutter/abadger1999 ?
[Tue Nov 28 2006]  [11:48:53]  <abadger1999>	spot: I vote +1 to either one with preference to Group optional.
[Tue Nov 28 2006]  [11:48:58]  <racor>	rdieter: this is absurd
[Tue Nov 28 2006]  [11:49:09]  <spot>	abadger1999: we could do a hybrid.
[Tue Nov 28 2006]  [11:49:10]  	 * rdieter is won't disagree.
[Tue Nov 28 2006]  [11:49:21]  <racor>	-1
[Tue Nov 28 2006]  [11:49:21]  <spot>	Group is optional, but here is the approved list if you want to use it
[Tue Nov 28 2006]  [11:49:36]  <f13>	useless
[Tue Nov 28 2006]  [11:49:57]  <f13>	spot: thats just saying standardized groups, where one of the groups is blank.
[Tue Nov 28 2006]  [11:49:59]  <tibbs>	Frankly I don't want to waste five more meetings coming up with an approved group list.ll
[Tue Nov 28 2006]  [11:50:47]  <rdieter>	tibbs: then stop arguing... (:
[Tue Nov 28 2006]  [11:50:47]  <racor>	tibbs: Then come up with a proposal for comps and copy it one to one into rpm groups
[Tue Nov 28 2006]  [11:51:05]  <tibbs>	So now it's my job?
[Tue Nov 28 2006]  [11:51:08]  <spot>	racor: doesn't work, can't have more than one rpm group
[Tue Nov 28 2006]  [11:51:16]  <spot>	lots of things are in multiple comps groups
[Tue Nov 28 2006]  [11:51:45]  <f13>	and really, package grouping depends on the person creating the repodata, not the person building the package.
[Tue Nov 28 2006]  [11:51:53]  <tibbs>	rdieter: You honestly think I'm arguing?
[Tue Nov 28 2006]  [11:52:23]  <f13>	what if I want to take all the fedora packages, put them in a different repo, group them differently and build a distro off that?  why should I have to rebuild every package and change every package just to get different groupings?
[Tue Nov 28 2006]  [11:52:42]  <spot>	i think group is a relic, a piece of the past, and we should make it optional
[Tue Nov 28 2006]  [11:52:50]  <spot>	+1 to group optional
[Tue Nov 28 2006]  [11:52:55]  <f13>	grouping of packages should _not_ be in the package.  It belongs in the repodata, which can be expressed differently
[Tue Nov 28 2006]  [11:53:19]  <spot>	lutter: need you to vote here.
[Tue Nov 28 2006]  [11:54:23]  	 * spot looks at the clouds
[Tue Nov 28 2006]  [11:54:51]  <f13>	12:24 <lutter> +1 for making the group tag optional
[Tue Nov 28 2006]  [11:55:06]  <spot>	yeah, but that was on the old vote. we don't let votes recycle
[Tue Nov 28 2006]  [11:55:20]  <spot>	and i'm not going to assume.
[Tue Nov 28 2006]  [11:55:26]  <f13>	'old vote"
[Tue Nov 28 2006]  [11:56:04]  <f13>	I haven't seen anything form lutter since then.  Did he wander off?
[Tue Nov 28 2006]  [11:56:06]  <spot>	ok. i don't see quorum for either item.
[Tue Nov 28 2006]  [11:56:18]  <spot>	thus, we'll continue to do nothing on group.
[Tue Nov 28 2006]  [11:56:40]  <spot>	new topic
[Tue Nov 28 2006]  [11:56:53]  <spot>	racor: are you interested in writing guidelines for packaging cross compilers?
[Tue Nov 28 2006]  [11:57:09]  <tibbs>	Recall that I was trying to do so.
[Tue Nov 28 2006]  [11:57:10]  <spot>	you seem to have a vested interest in this area, as well as experience
[Tue Nov 28 2006]  [11:57:21]  <racor>	spot: I would have been, but tibbs wrote something
[Tue Nov 28 2006]  [11:57:41]  <spot>	racor: i seem to recall you being rather opposed to his draft
[Tue Nov 28 2006]  [11:57:54]  <racor>	spot: well, yes.
[Tue Nov 28 2006]  [11:58:06]  <spot>	so, i'm interested in seeing a draft from you.
[Tue Nov 28 2006]  [11:58:17]  <tibbs>	Yes, please do be constructive.
[Tue Nov 28 2006]  [11:59:00]  <racor>	I'd have to check details again, but IIRC, his proposal was "beyond reality" (doesn't match with practice esp in GCC)
[Tue Nov 28 2006]  [11:59:17]  <tibbs>	And you of course suggested improvements.
[Tue Nov 28 2006]  [11:59:17]  <spot>	racor: would you take a try at writing a proposal that matches reality?
[Tue Nov 28 2006]  [11:59:26]  <tibbs>	Unfortunately I don't recall seeing what those improvements were.
[Tue Nov 28 2006]  [12:00:24]  <lutter>	f13: sorry (to all) .. had a bit of an emergency with the baby going nuts
[Tue Nov 28 2006]  [12:00:24]  <racor>	tibbs: I had edited some remarks into your wiki, but you preferred to shout at me when I pointed you at them
[Tue Nov 28 2006]  [12:00:33]  <f13>	lutter: completely understandable 9:
[Tue Nov 28 2006]  [12:00:34]  <f13>	(:
[Tue Nov 28 2006]  [12:00:48]  <spot>	lutter: you can vote on the group issue if you'd like
[Tue Nov 28 2006]  [12:00:48]  	 * f13 sees this going nowhere.
[Tue Nov 28 2006]  [12:01:03]  <spot>	lutter: either standardize groups or make Group optional
[Tue Nov 28 2006]  [12:01:10]  <f13>	racor: can you create a proposal based on your experience/thoughts?  May the best draft win?
[Tue Nov 28 2006]  [12:01:26]  <f13>	spot: I think its important that he reads backlog a bit
[Tue Nov 28 2006]  [12:01:27]  <spot>	racor: i would like to see your proposal
[Tue Nov 28 2006]  [12:01:50]  <rdieter>	I think we all would.
[Tue Nov 28 2006]  [12:01:50]  <lutter>	spot: +1 for making groups optional (and eventually getting rid of them)
[Tue Nov 28 2006]  [12:02:02]  <tibbs>	My draft was only an attempt to get something done.
[Tue Nov 28 2006]  [12:02:11]  <spot>	ok. thats +5 for making groups optional.
[Tue Nov 28 2006]  [12:02:15]  <lutter>	the rpm groups tag doesn't really have much value by itself IMHO .. you really need some overarching mechanism like comps
[Tue Nov 28 2006]  [12:02:32]  <racor>	spot: A black day in RH's history :(
[Tue Nov 28 2006]  [12:02:43]  <tibbs>	Finally some progress.
[Tue Nov 28 2006]  [12:02:58]  <f13>	its the doom and gloom show!
[Tue Nov 28 2006]  [12:03:17]  <rdieter>	it's more a light shade of grey, imo.
[Tue Nov 28 2006]  [12:03:19]  <racor>	f13: yes, you finally managed to kick out apt
[Tue Nov 28 2006]  [12:03:20]  <spot>	ok, thats an hour. anything anyone else wants to throw in before i close this out?
[Tue Nov 28 2006]  [12:03:59]  <spot>	ok. thanks for coming.
[Tue Nov 28 2006]  [12:04:05]  <abadger1999>	We should think of some better way to reach consensus.
[Tue Nov 28 2006]  [12:04:06]  <tibbs>	Nothing from me; license tag can wait until next week.
[Tue Nov 28 2006]  [12:04:11]  <f13>	racor: apt could easily make use of comps files, since it's the _actual_ _way_ that Fedora is defining package groupings.
[Tue Nov 28 2006]  [12:04:14]  <racor>	Final last words from me: I'll look into the cross-compiler stuff
[Tue Nov 28 2006]  [12:04:19]  <spot>	racor: thanks.
[Tue Nov 28 2006]  [12:04:26]  Quit	racor has left this server ("Leaving").