--- Log opened Thu Dec 14 12:57:08 2006
12:57 ::: f13!i=jkeating@fedora/ender has joined: #rpm
12:57 ::: Irssi: #rpm: 34 nicks (@/0 +/0  -/0)
12:57 ::: Irssi: Join to #rpm was synced in 0 secs
12:57 ::: jeremy!i=katzj@nat/redhat/x-0bdf990b8cbc523b has left #rpm: "Ex-Chat"
12:59 ::: clumens!i=chris@nat/redhat/x-33c9f910829904c9 has joined: #rpm
13:00 ::: clumens!i=chris@nat/redhat/x-33c9f910829904c9 has left #rpm: 
13:00 ::: jeremy!i=katzj@nat/redhat/x-0bdf990b8cbc523b has joined: #rpm
13:04 ::: cra_!n=cra@about/networking/255.255.255.4/cra has joined: #rpm
13:08 <jasonc> https://www.redhat.com/archives/fedora-announce-list/2006-December/msg00003.html
13:08 ::: rdieter!n=rdieter@sting.unl.edu has joined: #rpm
13:33 ::: laroche!n=laroche@p54A0915C.dip0.t-ipconnect.de has joined: #rpm
13:42 ::: msw!n=msw@rdu-nat.rpath.com has joined: #rpm
14:06 <rizzo> I need to have rpm manage my symlinks
14:06 <rizzo> Right now I create symlinks in my %post section
14:07 <rizzo> and those are always left broken when I upgrade/remote a package
14:08 <rdieter> then don't do it that way.  Create symlinks in the rpm, and have it own them.
14:08 <rdieter> unowned files/symlinks/dirs (whatever) are bad, mm-kay?
14:09 <rizzo> that's more or less what I was asking
14:09 <rizzo> I got errors using 'ln' during %install
14:09 <rizzo> hm
14:10 <rizzo> although my test may have been flawed
14:10 <rizzo> rpm.org eh
14:10 <rizzo> "RPM = RPM Package Manager" how clever ;)
14:11 <fray> boy this RPM buisness is going to get harry with the Fedora announcemnet.. ugh
14:12 <Rathann> what announcement?
14:12 <fray> http://lwn.net/Articles/214255/rss
14:13 * fray likes the direction jbj is going.. now I have to possibly fight fanatics who don't want to move forward
14:14 <Rathann> meh
14:14 <Rathann> I wonder why not just take 4.4.8...
14:14 <f13> fray: nothing prevents jbj from continuing his fork.  NOthing prevents the unified upstream from picking up features and such from jbj's fork, if its agreed upon to be useful and sane.
14:15 <fray> nothing prevents.. correct..
14:15 <fray> but like I said.. it's going to get harry..
14:16 <fray> my biggest concern is compatiblity..  will a package produced w/ one version work on the other... and if there is an issue.. whose problem is it?  (my guess is the distro/package creators)
14:16 ::: mode/#rpm: +o cra_ by ChanServ
14:17 ::: mode/#rpm: -o cra_ by cra_
14:22 ::: rdieter is now known as rdieter_away
14:22 <rsc> jbj: hm what do you think about the announcement of Max?
14:28 <rizzo> so jbj is the developer who left?
14:28 <rsc> jbj, the one and only one ;)
14:29 <fray> I laugh, cause frankly jbj is one of THE only true RPM developers left..
14:29 <rizzo> so I should just be able to do 'ln -s' in %install and then list it in %files?
14:29 <fray> I do local development and submit back when I have the time.. but frnakly I understand less then 10% of the code..
14:29 * rizzo uses google already
14:29 <fray> I doubt there is anyone who understands the code better then jbj..  James Olin? is probably the only other person I know of who understands it enough to work on the internals regularly!
14:30 <rsc> fray: IIRC there is nobody else except jbj. Paul Nasrat mostly backports e.g.
14:43 <rizzo> huh doing ln -s in %install and %files worked just great
14:43 <rizzo> wonder what the hell I was doing the first time I tried this
14:55 ::: pzb!n=pzb@gw-ott1.byward.net has joined: #rpm
14:56 <pzb> the wiki's password reset doesn't seem to be working :(
14:57 <rsc> blame Max ;)
14:59 <f13> pzb: hrm, could be a DNS issue if your mailer isn't seeing the new MX for lists.rpm.org
14:59 <f13> pzb: in which way is it not working?
15:00 <pzb> f13: I get the email, but it contains no password or links
15:00 <f13> hrm.
15:00 <pzb> f13: it says "use the data below", but there is no data
15:00 <f13> k, reporting to sysadmin
15:02 <pzb> I signed up for an account a couple of months ago, but forgot what I used for a password
15:05 <f13> oh wait.
15:05 <f13> what wiki?
15:07 <pzb> wiki.rpm.org
15:07 ::: rdieter_away is now known as rdieter
15:08 <f13> pzb: I'd try recreating the account?  or ping skvidal
15:10 <pzb> f13: I'm chatting with seth
15:12 <rizzo> THE seth?
15:19 <rsc> I only know one Seth and it's one too much ;)
15:24 <rizzo> I don't know any
15:34 ::: donavan!i=daemon@centos/slackers/donavan has quit: Excess Flood
15:35 ::: donavan!i=daemon@4wx.net has joined: #rpm
15:36 ::: donavan!i=daemon@centos/slackers/donavan has quit: Excess Flood
15:37 ::: donavan!i=daemon@4wx.net has joined: #rpm
15:54 <jasonc> I love how rpm.org has moved to duke so that no corporate entity controls it... but what happens when duke shitcans seth?
15:55 <f13> an interesting question.
15:55 <f13> duke probably would hand over the content to whomever.
15:55 <jasonc> and one nobody cared to ask, the only real concern was that this great new world order would be sans jbj
15:55 <jasonc> what a waste
15:55 <f13> but more to the point, Red Hat doesn't own it, Novell doesn't own it, Mandriva doesn't own it, etc...
15:56 <jasonc> f13: yeah I read the announcement I just didn't drink the kool-aid
15:56 <msw> except that Red Hat still owns the majority of the copyright...
15:57 * fray loves GPL.. doesn't matter who owns the copyrights as long as you abide by the terms of the GPL/LPGL
15:57 <fray> 'er.. LGPL
15:57 <pzb> jasonc: why sans jbj?  in some ways I think that this might be more to his liking
15:57 <pzb> jasonc: a very small core RPM that the distros use that he can build on, withouth getting lots of people yelling at him
15:57 ::: nasrat!n=nasrat@91.84.17.49 has joined: #rpm
15:58 <msw> hi nasrat!
15:58 * fray just knows this is likely to give him [fray] heartburn for a while..
15:58 <fray> nothing like diverging versions of RPM.. :P  LSB was at least half-way understandable.. (even though not terribly sane)
15:59 <f13> jasonc: actually, the question was asked re seth leaving Duke.
15:59 <nasrat> fray: as opposed to the already micro forks of all the distros
15:59 <f13> the infrastructure is moved to somewhere that wants to host it, whether that be rh, novell, bu, etc.
15:59 <jasonc> pzb: utopian vision aside, there was a clear choice made to start with what was there only up to jbj's leaving of RH (or termination depending on who you ask), RH repeatedly passed off the "no upstream maintainer" despite the fact that people are using jbj's version and that he is still working on it
15:59 <nasrat> anyway I'm off the clock
15:59 ::: nasrat!n=nasrat@91.84.17.49 has left #rpm: "Leaving"
15:59 <fray> nasrat, no doubt lots of microforks
16:00 * fray just wants to make sure rpmrc dies.. ;)
16:00 <jasonc> microfork definitely true, but if they really had any interest in reaching out to the people who are currently working on rpm, they wouldn't have started with 4.4.2 imho
16:00 <msw> fray: that's more likely to happen in a jbj tree than the "stable" tree
16:00 <fray> msw, already happened.. which is why I'm probably going to pick it up.. (jbj's tree)
16:01 <fray> jasonc ya.. that disappointed me a lot as well
16:01 <jasonc> basically the announcement said "we don't consider anything greater than 4.4.2 to be real RPM or worthwhile development"
16:01 <msw> jasonc: that's not the point
16:01 <msw> jasonc: there's worthwhile stuff in there
16:01 * jasonc waits
16:01 <f13> jasonc: the clear choice was to tkae what was left 'upstream'.  jbj created his own fork, by his own admittence.  The major stakeholders (RH/Novell) were both using 4.4.2, and this makes sense to use as a starting point.
16:01 <pzb> jasonc: I think you are seeing differing goals
16:01 <msw> jasonc: it's just really hard to get a grapple onit
16:02 <pzb> jasonc: Red Hat and Novell like slow and boring for something like RPM in their enterprise distros
16:02 <jasonc> f13: jbj's comment which people have been trumpeting around was "if you want to call my work a fork go ahead and enjoy"
16:02 <msw> things might be different if there were regression tests. :-/
16:02 <jasonc> hardly the same
16:02 * fray notes Wind River is still using 4.4.2.. but that's because we had a release cycle..  now that's over with.. time to move forward
16:02 <pzb> jasonc: I know I have had customers say they want to see a roadmap and other "enterprisy" features
16:03 <jasonc> pzb: do those same customers ask about a yum roadmap?
16:03 <jasonc> yum todo list?
16:03 <jasonc> hrm, that's not an issue, I wonder why
16:03 <pzb> jasonc: we (Novell) do not ship yum
16:03 <pzb> jasonc: and yes, they do routinely ask about system management roadmaps; I'm sure they push RH for RHN roadmaps
16:03 <fray> jasonc, no but I have heard them ask about Anaconda and soft-update.. ;)  (not sure if they ever got any answers though)
16:03 <f13> jasonc: thats not the quote I saw on rpm-devel
16:03 <jasonc> fray: heh
16:04 <jasonc> pzb: the point is the justification for this new world order is about as believable as iraq having wmds
16:04 <jasonc> you don't like jbj and you don't want to work with him is fine, but have the balls to say so (not you == pzb but you in the general sense)
16:04 <f13> jasonc: https://lists.dulug.duke.edu/pipermail/rpm-devel/2006-August/001374.html  thats the mail in question.
16:05 <pzb> jasonc: I like jbj a lot; we have met numerous times, and get along fine
16:05 <jasonc> pzb: like I said, the you wasn't meant to be you :)
16:06 <pzb> jasonc: but with my corporate hat on, Jeff introduced a lot of code into RPM that was not needed
16:06 <jasonc> pzb: and yet pld and cAos use those features... so which is it?
16:06 <msw> jasonc: all of them?
16:07 <fray> pzb, on the same token he has removed a LOT of code that is no longer needed.. i.e. rpmrc
16:07 <jasonc> pzb: open source development model and "I don't want to change anything for fear of borkage" don't mix
16:07 <f13> jasonc: whether or not jbj's work will be used in the upstream RPM is left up to the interested vendors that make up the upstream.
16:07 <jasonc> msw: dunno, ask them :)
16:07 <jasonc> yeesh I seem to have fired off a shitstorm, not my intent
16:07 <f13> hehe,
16:07 * fray notes sqlite support is HIS fault.. since MV couldn't use bdb..
16:07 ::: sekhmet!n=pez@ppp-70-226-146-102.dsl.mdsnwi.ameritech.net has joined: #rpm
16:07 ::: bress!n=bress@redhat/bress has joined: #rpm
16:08 <msw> fray: some sqlite support is in 4.2.2, no ?
16:08 <pzb> fray: why not bdb?
16:08 <jasonc> pzb: the thing I never saw in all the discussions of a fork is what features were added or removed that were so objectionable or useless
16:08 <fray> msw yes..
16:08 ::: avaia!n=avaia@yosemite.cellcom.com has joined: #rpm
16:08 <msw> pzb: sleepycat license is bad for embedded where sleepycat isn't getting paid
16:08 <fray> pzb, because bdb does 1) work reliably over NFS [even in a single user environment], 2) doesn't work reliably cross architecture
16:08 <jasonc> only thing I saw that was close was this: http://fedoraproject.org/wiki/rpm-devel
16:08 <fray> msw, actually license was the ONE place we didn't care
16:08 ::: sargeantd!n=mike@adsl-68-76-147-55.dsl.akrnoh.ameritech.net has joined: #rpm
16:08 <jasonc> and that hardly justified a fork given the input
16:08 <fray> it was purely functional reasons
16:08 <msw> fray: interesting
16:09 <fray> You create a db on x86_32, and try to use it on MIPS32_be once..  you'll find the db becomes read-only.. which makes upgrading a package a bitch
16:09 <jasonc> heh
16:09 <fray> use it over NFS (even if you are the one reader/writer) and locking turns to crap
16:10 * pzb admits he has never put RPM on NFS
16:10 <jasonc> pzb: if the discussion of supposedly cracktacular features occurred somewhere public I'd love to see it
16:10 <fray> switch to using sqlite (slightly slower performance, BUT) you suddenly can use the DB cross platform.. you can now work over NFS (assuming single reader/writer)..  multiple reader/write has a slight chance of corruption.. but multireader works fine
16:11 <fray> pzb, biggest hurtle was always managing the DB locking corruption.. as the maintainer/support person at MV.. I finally said f' you BDB.. and switched to sqlite.. suddenly all of the locking problems went away
16:11 <f13> jasonc: I don't think the features themselves (well many of them) are cractacular.  I think its more that the tools built on top of rpm are by far not ready for those.  To jump straight to that would be a mistake.  Starting with a common shipping stable tree, and going from there makes more sense.
16:12 <jasonc> f13: generally speaking in a debate you have to pick a side to argue and stick with it ;)
16:12 <f13> I'm sorry?
16:13 <jasonc> f13: are we talking about forking rpm or the tools? rpm maintainer's job won't be to fix yum's code too will it?
16:14 <jasonc> the discussion I've seen is about rpm, any other crap is fluff imho thrown on to justify our hunt for wmds
16:14 <f13> jasonc: rpm is a base part of the entire OS.  Calling a rpm codebase with features that the OS can't possibly handle the 'stable upstream' doesn't make sense at all.
16:15 <jasonc> f13: ok what features are those?
16:15 <f13> the hg.rpm.org tree is set to be teh stable rpm that is being shipped.  Other trees are forks or future development which may or may not get pulled in, depending on consenting decision
16:15 <f13> jasonc: off the top of my head, the soft deps
16:15 <rsc> f13: on decision of whom?
16:15 <pzb> jasonc: as I wrote a long time ago, I am hoping to see a very simple rpm
16:16 <pzb> jasonc: one that has a very well defined file format, one that offers a good stable library for higher level apps
16:16 <jasonc> f13: you mean the ones that are listed as "Use could be forbidden by policy AKA Packaging Guidelines" on the fedora wiki?
16:16 <jasonc> hardly sounds like a serious concern 
16:17 ::: AedinAZ!n=aedinaz@gmk.caoslinux-alliance.net has joined: #rpm
16:17 <fray> well.. I think there is an issue between RPM file format and RPM application..
16:17 <f13> rsc: the leadership that grows out of the upstream.  RH is putting resources behind it, as is Novell.  Others have been invited too
16:17 <pzb> fray: and rpm library and rpm build tools
16:17 <fray> I think the RPM file format should be documented at an agreed upon stable level... the RPM application is a seperate issue
16:18 <f13> jasonc: the rest of the OS is not ready for those features.  For that reason, we don't allow them in the Fedora specs.
16:18 <pzb> fray: could not agree more
16:18 <f13> (and we don't impliment them in our rpm)
16:18 <rsc> f13: mhm. Hopefully Red Hat itself is not involving itself to much, because the last Red Hat year regarding RPM was _worse_
16:18 <fray> rsc, agreed
16:18 <rsc> and the Red Hat opinions are most _worse_
16:18 ::: mspevack!i=mspevack@fedora/mspevack has joined: #rpm
16:19 <f13> Red Hat _wants_ more than just Red Hat making these decisions
16:19 <jasonc> f13: so they say
16:19 <fray> I remember nuerious bug reports I had filled while working at MV (against RH 9 or so) and nobody would fix it because "it was too destabalizing"
16:19 <f13> and Red Hat is dedicating people to work on the Upstream of rpm, not what Red Hat uses of rpm.
16:19 <rsc> f13: ha, but not accepting jbj ones? ;)
16:19 <jasonc> f13: the 2.4 kernel wasn't ready for NPTL, that didn't stop RH9 
16:19 <f13> rsc: if jbj wants to participate, why can't he?
16:20 <jasonc> f13: well for one the announcement has made quite clear that all of his current work is persona non grata
16:20 <rsc> f13: Red Hat didn't accept the opinions of jbj the last year or so. So hopefully Red Hat is only less involved into the new stuff.
16:20 <f13> his work is not desired for the shared stable tree that RH/Novell is trying to create out of what they're currently shipping.
16:20 <rsc> and they were refused for dubious and obious reasons.
16:20 <f13> jasonc: absolutely NO decisions have been made regarding Future Development of rpm.
16:20 <rsc> lol, Novell you can smoke.
16:20 <rsc> they stood for years at RPM 3.x ;)
16:21 <f13> I say Novell, but I really mean SuSE
16:21 <rsc> same stuff ;)
16:21 <jasonc> f13: so did rsc
16:21 <f13> and Mandriva and whomever else happens to be shipping RPM
16:21 <jasonc> f13: bs, the decision has been made to pretend like rpm never moved on, that's a development decision and a bad one at that
16:22 <pzb> jasonc: those decisions are made every day
16:22 <pzb> jasonc: you look at what you are shipping and see if you want to move forward or not
16:22 <rsc> yeah, but Red Hat is getting a sloppy mess :(
16:22 <rsc> pzb: but it's unbelieveable to stay at kernel 2.4, but not to stop at rpm 4.4.2
16:23 <jasonc> pzb: yes but typically your path forward involves putting the upgrade in the next release, not creating some brave new world that ignores the new releases
16:23 ::: gregdek!i=gdk@nat/redhat/x-c82e8e8094945a91 has joined: #rpm
16:23 <msw> greg!
16:23 <gregdek> msw!
16:23 <rsc> so you can smoke Red Hat the same way like Novell/SuSE ;)
16:23 <rsc> but it maybe tastes somehow stronger *g*
16:23 <f13> jasonc: its not a 'brave new world', it's an agreement on what both have already been shipping as a stable point, to look forward to that new release.
16:23 <gregdek> I smoke Red Hat every day.
16:23 <pzb> jasonc: depends on customer needs
16:23 <f13> evening gregdek 
16:24 <rsc> pzb: Fedora depends on customer needs?
16:24 <pzb> jasonc: I can say for myself (not Novell) I do not see much new that is in demand
16:24 <gregdek> f13: Hope you're maintaining your sanity.  ;-)
16:24 <jasonc> and since when has fedora given a shit about stable anything?
16:24 <rsc> pzb: or is this an agreement, that Fedora is the pre-release/testing area for Red Hat Enterprise Linux? ;)
16:24 <fray> customers need a package manager that works for the distro..  that's really not the issue IMHO..  the issue is the distro creaters need a version of RPM that allows them to create their distro..  (I think a lot of people miss that point)
16:24 <pzb> rsc: sure, users are customers
16:24 <rsc> jasonc: hey, the kernel at Fedora can be buggy, but RPM must be old and stable (s/stable/broken/) ;)
16:24 * fray honestly has never seen a customer say "I need version X of rpm".. they always say "I need a distro based on RPM packages.. or that USE rpm packages"
16:25 <rsc> fray: then I'm the first customer ;)
16:25 <jasonc> plenty of bugzilla entries at RH that ask for a newer rpm
16:25 <pzb> fray: a few say "I need feature X"
16:25 <fray> rsc, if you asked for a specific version of RPM, I'd be shocked..
16:25 <f13> jasonc: Fedora cares very much about self stability.  The software versions we choose to ship with we try very hard to ensure they are as stable as we can make it, often times forgoing newer versions of items if they are not stable enough.
16:25 <jasonc> fray: I read the bug report, he wasn't the only one
16:25 <fray> pzb, that is correct..  they ask for a feature..
16:25 <rsc> f13: self stability was a good joke...
16:26 <pzb> jasonc: the real question, you always want to ask is "what feature do you need?"
16:26 <fray> rsc, that I agree with
16:26 <f13> rsc: I'm sorry you feel that way.
16:26 <pzb> jasonc: not "what product" or "what version"
16:26 <jasonc> pzb: sure, but the feature was part of the bug report... I need feature X implemented in version Y of RPM
16:26 <jasonc> so clearly that wasn't the hold up
16:27 <jasonc> f13: I maintained a handful or so of fedora.us packages before RH absconded with the name and warren, I'm not entirely unfamiliar with the fedora way
16:27 <rsc> f13: where's the difference between shipping a broken kernel or a broken rpm? IMHO a broken kernel is more worse rather a broken rpm ;)
16:27 <rsc> f13: but Red Hat likes a broken kernel rather a broken rpm. Wrong world.
16:27 <rdieter> rsc: man, are you going to say anything that's not flamebait?
16:28 <jasonc> rdieter: sadly I can verify what rsc has said with a rhel bug report
16:28 <fray> rdieter I see rsc playing more of a devils advocate role..  as do I see myself..
16:28 <fray> I think this intential fork is a mistake.. vendor forks I can deal with
16:28 <jasonc> rdieter: lets look at https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=207244 shall we?
16:28 <f13> rsc: We try very hard to not ship broken kernels, or at least broken kernels that would effect a sizable portion of our userbase.  We've delayed releases quite long becuase of this.
16:28 <rsc> rdieter: where's the flame. Just look at what Red Hat produced about the last year ;)
16:29 <jasonc> default mta shipped with rhel4 causes a monotonically increasing load that eventually causes the box to be unusable
16:29 <f13> jasonc: fedora.us and early Fedora was a long time ago.
16:29 <rdieter> one bug != not caring about stability.  Puh,leaze.
16:29 <jasonc> rdieter: RH has a patch from kernel.org, chooses not to push it
16:29 <jasonc> f13: and?
16:29 <rsc> f13: right, but there's also something like Rawhide where things can be broken, right? But this isn't used for rpm somehow - unfortunately.
16:29 <jasonc> rdieter: we can go through bugzilla together if you really feel the need for more justification 
16:29 <rsc> jasonc: +1
16:29 <f13> jasonc: honestly lots has changed.  Mindsets have changed.  A battle we continue to fight and continue to take ground on daily.
16:30 <jasonc> f13: the battle was fought and lost a long time ago, imho you're arguing about $RPM_BUILD_ROOT vs. %{buildroot}, not anything useful
16:30 <rdieter> we're talking overall direction/intentions of the project here.  Taking individual bad examples (and yes, bad ones exist), and drawing sweeping conclusions, well...
16:31 <f13> jasonc: yes, trying to standardize on how our specs look is completely fruitless.  Why did we spend any time on it.
16:31 <fray> rdieter I'm honestly still not clear on the direction/intention other then "pick an arbitrary version and make stall it until the end of time"
16:31 <jasonc> rdieter: I tried waiting on fedora to clean up their act, you ran out of time around fc3
16:31 <fray> that looks dumb to me
16:31 <jasonc> f13: standardization is great, but only if you bother to actually do anything with it
16:31 <jasonc> f13: maybe when core is merged into extras it'll be better, but every spec I've grabbed from fedora this past month doesn't come close to being standardized
16:32 <f13> jasonc: with FC6 we forced every new core package to go through the package review, ditto RHEL5.  Guidelines and review process created by the FEdora community, not Red Hat.
16:32 <jasonc> f13: hah!
16:32 <f13> jasonc: with the merger, we're forcing every package as of yet unreviewed to get reviewed.
16:32 <f13> are there lots of crappy core packages?  Yes
16:32 <rsc> hm, Core is not really following Guidelines.
16:32 <f13> will it take al ot of work to fix them, lots.
16:32 <gregdek> f13: That is, in fact, one of the primary goals for merging Core and Extras: imposing the strong packaging standards developed by the community on to the RH developers.
16:32 <f13> rsc: there are still lots of grandfathered in Core packages, and yes, lots of them stink badly.
16:33 <jasonc> f13: again, I read the announcement, I'm just not drunk on kool-aid
16:33 <f13> rsc: we prevented further stink from happening by reviwing all new packages, this was a good step.  Further steps involve reviwing all the rest.
16:33 <rsc> f13: and a rpm version which is currently unwanted in Fedora helped to file many bugs (e.g. orphaned directories)
16:34 <f13> rsc: and the fix for that bug may very likely be pulled from jbj's tree into the new upstream stable tree.
16:34 <jasonc> f13: it makes little sense to impose a ruleset that you yourself don't abide by, fedora has a long way to go before the stink of hypocrisy wears off
16:34 <mspevack> jasonc: the goal of the announcement was not to get you (or anyone else) drunk on cool-aid :-)
16:34 <f13> jasonc: which is why every package is going to be reviewed and fall under compliance.
16:34 <fray> mspevack no, I read it as to piss of the few folks using jbj RPM, so as to switch to this new holy annointed version
16:34 <jasonc> mspevack: it did manage to piss me off, so it at least accomplished something sort of like getting drunk
16:35 ::: sargeantd!n=mike@adsl-68-76-147-55.dsl.akrnoh.ameritech.net has left #rpm: 
16:35 <jasonc> f13: kool-aid, tell me when it's done, not what you think you'll be doing
16:35 <rsc> ah mspevack is around here, too
16:35 * fray notes he's involved in a number of multi-vendor linux projects..  in his experience 1) announcement before code = death... 2) code before announcement may be useful..
16:35 <jasonc> until then I will point to the eclipse package, or aide in extras, or any number of other packages that have supposedly gone through strict review for conformance and I will say BS loudly
16:36 <fray> to me I read this as "hey look over here and avert your eyes to code in the other corner"
16:36 <jasonc> fray: yeah that was my reading as well
16:36 <f13> jasonc: unfortunately many of hte early extras packages went straight from core to extras without review.  We will be cleaning those up too.
16:36 <gregdek> jasonc: And you will submit patches to bring the packages into compliance, yes?
16:36 <mspevack> fray: then either we will *actually produce something useful* or we won't.  So clearly we have something to prove.  And we shall see what happens.  But i'm not trying to *force* anyone to do anything.
16:36 <gregdek> jasonc: Since, by virtue of knowing why it is they're screwed up, you'll also have a pretty good idea of how to fix them, yes?
16:36 <fray> gregdek, point me at the "compliance" doc.. and in about 3 weeks I will start to work on making core packages compliant..
16:37 <fray> but so far when I've brough up minor annoyances I've been ignored.. so I said screw it
16:37 <f13> fray: http://fedoraproject.org/wiki/Packaging/Guidelines
16:37 <gregdek> http://fedoraproject.org/wiki/Packaging/Guidelines
16:37 <gregdek> Continually updated.
16:37 <jasonc> gregdek: yup I could do that, but I'm more interested in distros that are aligned with what I see as more sensible goals... the fedora center of the universe mindset pisses me off, and I'm not playing in your sandbox anymore
16:38 <jasonc> I have to work with RHEL, that's as close as I'll get at this point
16:38 <f13> said "compliance" doc is living too.  Many of hte guidelines were created out of existing and considered best practices.  If there are errors or missassumptions, there is a packaging committee to review these issues and make changes accordingly.
16:38 <jasonc> mspevack: nor will you force anyone to do anything, but you are clearly doing some hand waving to pretend like nothing has happened since 4.4.2
16:39 <gregdek> jasonc: Then what's your rationale for complaining about the shortcomings of a community you've no interest in joining, and no interest in comsuming the output of?
16:39 <pzb> jasonc: I don't think anyone is saying nothing has happened
16:39 <pzb> jasonc: the announcement didn't say that
16:39 <fray> pzb, basically I agree with the way jasonc read it
16:39 <rsc> pzb: but why was (is) strictly ignored what has happend?
16:39 <pzb> jasonc: it said Red Hat and Novell are using 4.4.2, and it is easiest to start from a known point
16:40 <fray> I read it as "I'm taking the ball and going home.. if you wish to play you now have to come to my house"
16:40 <jasonc> gregdek: my rationale is if fedora is going to swallow projects I care about and apply a package nazi mindset to those packages, why shouldn't the same be true for fedora
16:40 <jasonc> gregdek: fedora is far too fork happy
16:40 <fray> I'd have preferred "We are going to do development on RPM, and will submit back all relevant changes"..  that way you are playing "nicely"
16:40 <mspevack> jasonc: the document clearly says that work has always been happening on RPM.  It doesn't pretend that stuff never happened.  It just says what the starting point that pnasrat and co are going to use is.
16:40 <jasonc> once again, center of the universe... this is where we are so this is where everyone should start from
16:41 <pzb> fray: that is the objective here; right now both RH and Novell have heavily patched rpm 4.4.2 versions
16:41 <pzb> fray: and jbj, by his own statement, has a fork
16:41 <mspevack> jasonc: no, it's where we are starting from.  And hopefully what it will *lead to* is something that others will be interested in.
16:41 <fray> pzb, as does Wind River (and MontaVista).. I can no longer speak for MontaVista.. but I can speak for WR.. We're moving forward for a newer version
16:41 <pzb> fray: so lets get a neutral place to try to merge bits
16:42 <fray> pzb, I don't believe this new "group" is neutral.. sorry
16:42 <rsc> hm, pnasrat himself said some time ago, he's absolutely busy and never had time to look detailed into rpm > 4.4.2, does he _now_ have time? Strange... ;-)
16:42 <jasonc> pzb: why isn't wraptastic.org "neutral" enough of a spot to merge at?
16:42 <f13> rsc: he now has time due to RH hiring more resources.
16:42 <jasonc> oh right, it's because jbj is there
16:42 <rsc> jasonc: +1 :)
16:42 <f13> rsc: and repurposing paul's day job.
16:43 <pzb> fray: heck of a lot more neutral than the old RH days; no feature requirements published, code going in based on RH internal requirements without discussion
16:43 <rsc> f13: ah, one positive aspect, but jbj has been fired before...it could have been much easier
16:44 <fray> pzb, ohh no doubt..
16:44 <f13> jasonc: "fedora is far too fork happy" would you care to expand upon that?
16:44 <jasonc> f13: sure, how many packages in fedora have zero patches?
16:44 <jasonc> f13: I'm a committer for jpackage, RH and Fedora folks are nice about not contributing patches back upstream
16:44 <f13> jasonc: quite a few.
16:44 <gregdek> s/fedora/debian s/fedora/ubuntu s/fedora/suse
16:45 <f13> jasonc: what?  our jpackage folks are doing all their work upstream, and just importing the srpms into our package CVS
16:45 * fray notes in his daily work there are generally three types of packages..  Features, Bug Fixes and Integration.. the first two go upstream the third rarely does
16:45 <jasonc> f13: oh really, where did that pesky eclipse package come from?  funny, I never saw that committed back to us
16:45 <jasonc> f13: java-1.4.2-gcj-compat was only added when requested, and has since languished 
16:46 <jasonc> f13: you ain't gonna win this argument, I've been with jpackage since before it had a name
16:46 <bress> jasonc: jpackage is the eclipse upstream?
16:46 <f13> jasonc: these are two examples in HOW many jpp packages?
16:46 <gregdek> jasonc: If you believe that Fedora works less well upstream than other distros, I'd be happy to see you put us publically to that Pepsi challenge.
16:46 <gregdek> Because I'm legitimately interested in the results, package by package.
16:46 * f13 too
16:46 <jasonc> bress: jpackage was where the eclipse package in fedora came from
16:47 <pzb> gregdek: I would suggest that distros that have a target to be stable generally are at odds with most upstreams
16:47 <jasonc> bress: I particularly liked that overholt stripped out all jpackage bits from the %changelog
16:47 <pzb> gregdek: I don't care if you call it "enterprise" or not, distros invariably end up forking almost everything
16:47 <pzb> gregdek: Fedora is, if anything, better than most, about staying close to upstream
16:47 <bress> jasonc: That hardly justifies it as a fork though.  I won't disagree with you about the innapropriateness of those actions, but you've certinaly not justified your statement about Feodra forking things.
16:47 <gregdek> pzb: Yep.  So do we do it more than most?  Less?  About the same?
16:47 <jasonc> f13: is there a number of packages I could provide as examples that would make my point, or is the number of examples needing to be all of jpackage
16:48 <jasonc> bress: oh eclipse 3.2 was never contributed back to jpackage, only jpp stripped off, changelog removed, and off they went
16:48 <pzb> gregdek: but that is just my feeling; I'm guessing openSUSE is or will be close
16:48 <gregdek> Now, I will be the first to admit that Java packaging for Fedora is probably pretty insane.  Show me the better model.
16:49 <pzb> gregdek: the enterprisy long term suport distros are horrid; see RH 7.2EE kernel for a scary thing ;)
16:49 <f13> jasonc: its my belief that the overwhelming number of Fedora packages are much closer to upstream then they have ever been before.  Patches that we have now are either unacceptable upstream or unsuitable upstream.
16:49 <jasonc> gregdek: we had a great multi-distro model with people from RH, SuSE (prior to Novell), and others with no distro affiliation
16:49 <gregdek> jasonc: At jpackage, you mean?
16:49 <jasonc> gregdek: yes
16:49 <jasonc> f13: sure, progress has been made
16:50 <f13> jasonc: to the point where the exceptions are what makes the rule.
16:50 <jasonc> f13: not so sure I agree with that
16:50 <gregdek> jasonc: I think it's fair to say that RPM packaging and Java packaging don't go terribly well together, and everybody has different ideas about how to fix those variances.
16:50 <f13> and you're welcome to your opinion (:
16:50 <jasonc> f13: gee thanks! :P
16:51 <f13> gregdek: boy you got that right.  multilib is a NIGHTMARE with java, especially when we're doing byte compiled java stuff
16:51 <jasonc> funny, multilib isn't an issue if you use a jdk...
16:51 <f13> jasonc: which we couldn't ship
16:51 <gregdek> So to tar Fedora with the "sux upstream" brush is only fair, IMHO, if you can point to a Linux distro that does it better, and tell me why.
16:51 <f13> and now that we can, or soon can, many of these problems will go away.
16:52 <jasonc> gregdek: ok, cAos, pld both are pretty patch light
16:52 <msw> gregdek: does what better?  upstreaming?
16:52 <msw> gregdek: rPath Linux is extremely unpatched
16:52 <msw> gregdek: by design
16:52 <f13> er..
16:52 <jasonc> f13: so at that point fedora will move back to using jpackage work?
16:52 <f13> isn't rpaths point that you _can_ patch and stay with that patch?
16:53 <f13> jasonc: depends on the java team.
16:53 <f13> jasonc: depends on how things shake out with what Sun is doing with java code.
16:53 <msw> f13: it's so that *other* people can patch and manage the patch
16:53 <msw> f13: we wanted to make a base distro that was extremely easy to patch
16:53 <f13> probably nothing different by the next Fedora release.
16:53 <f13> msw: gotcha
16:53 <msw> f13: so one of the things we had to do was clean *out* the patches, so you don't have to patch juggle
16:53 <jasonc> f13: might not ever, my faith in sun doing the whole kit and kaboodle is low
16:53 <gregdek> Courses for horses.
16:54 <f13> jasonc: as is mine :/
16:54 <f13> jasonc: which means we'll continually have to forgo a jdk, and thus differ from jpackage in our the packages are built.
16:54 <jasonc> f13: ah but our packages have aot and gcj build options, all you need to do is actually use them
16:55 <jasonc> f13: we've bent over backwards to work with every vendor interested, fedora (and particularly fedora-java) has been outwardly hostile for some time
16:55 <jasonc> f13: go through fedora-java list and find all the postings that say "don't use jpackage"
16:55 <jasonc> since apparently any argument I make will be required to have 75 supporting items
16:56 <bress> jasonc: If that's true, this is certainly not the forum to air your grievances.
16:56 <rdieter> jasonc: it's not hostility, its breakage from mixing/matching gcj/native-java pkgs.
16:56 <f13> jasonc: I know that content is in the spec.  It makes the spec very ugly but necessary.
16:56 <jasonc> rdieter: gcj by the way it works has to ship the jar, breakage only exists because it's wanted
16:57 <rdieter> f13: apparently fedora's java packagers have some work to do.
16:57 <f13> its entirely possible they do
16:57 * f13 isn't very fond of Fedora's java packagers
16:57 <jasonc> bress: I wasn't intending to air a grievance here, simply offering supporting examples for my statements
16:58 <jasonc> bress: I've publicly (on jpackage lists) commented on my concerns, and on fedora-java I was soundly ignored
16:58 <bress> jasonc: And what is your argument?  I'm not understanding where this conversation is going, or came from at this point.
16:58 <jasonc> bress: this was justification for my statement that fedora is entirely too fork happy atm
16:59 <jasonc> whether publicly announced or de facto fork, result the same
16:59 <bress> jasonc: You've not justified that.  Using someone elses spec file hardly qualifies as a patch.
16:59 <gregdek> jasonc: Seems like all of your (probably well-founded) arguments center around Fedora's inability to come up with coherent policy around Java, and RPM is really a side issue.
16:59 <bress> s/patch/fork/
16:59 <pzb> can fedora discussion move over -> there, please?
16:59 <jasonc> bress: but using someone elses spec file, removing attribution, and not contributing back to upstream?
16:59 * pzb points at #fedora
16:59 <gregdek> jasonc: Yep.  That's uncool.
17:00 ::: rdieter is now known as rdieter_away
17:00 <bress> jasonc: I'm not saying you don't have a point being pissed about that, but your sole example hardly justifies Fedora as being patch happy, nor should you complain about such things here.
17:00 <jasonc> gregdek: rpm is a similar problem in my book, just a different source with some personal vitriol mixed in to the decision making
17:00 <bress> s/patch/fork/
17:00 <bress> why the hell do I keep doing that.
17:01 <jasonc> sigh, it was an example, not meant to be the point of discussion nor the only data point
17:01 <bress> jasonc: But you keep refering to it.  You need more than one example to make such a broad statement.
17:01 <bress> jasonc: It sounds to me like you have a personal beef and are bitching about it here.
17:02 <jasonc> bress: that's fine if you see it that way, I see the rpm announcement as pissing on the work done on rpm since 4.4.2
17:03 <f13> I don't think any amount of discussion will change that either, fwiw.
17:03 <gregdek> jasonc: Some would argue that we've already "pissed on" that work by not rolling it back into our own RPM.  That ship has sailed.  What we're doing now is explaining why.
17:03 <gregdek> But I take your point.
17:03 <jasonc> bress: and again it's referred to simply because it's the example I can talk about off the top of my head... I guess I didn't realize I'd be required to provide as much justification as has been asked for 
17:03 <bress> jasonc: Broad statements should be backed up with broad examples :)
17:04 <jasonc> bress: I gave you an example, you didn't like it, not my fault
17:04 <bress> one != broad
17:04 <bress> jasonc: It is your fault.  You made a broad statement, then completely failed to back it up.
17:04 <jasonc> bress: maybe I just don't understand what you're looking for, why don't you explain to me the scenario under which you would believe my argument 
17:05 <jasonc> s/argument/statement/
17:05 <f13> jasonc: you've unfortunately touched a nerve though, as one of the goals of the Fedora project is to stick much closer with upstream.  Its something we try very hard to do, so broadly saying that we're failing at this is going to cause some irritation, and going to get you asked for proper justification.
17:05 <bress> jasonc: Why do you feel fedora is fork happy, using an example that is not taking your eclipse spec file.
17:05 <jasonc> bress: I gave numerous examples from jpackage
17:05 <jasonc> rpm is another
17:05 <jasonc> you could argue kernel.org
17:05 <f13> *BLINK*
17:05 <bress> I work on this crap all the time, I can promise you Fedora does a much better job of working with upstream than any other distro I've ever worked with.
17:06 <f13> jasonc: you're kidding right?
17:06 <fray> bress, IMHO Debian has been better.. but Fedora is one of the better distros
17:06 <f13> jasonc: what are we possibly carrying in our kernel that A) hasn't tried to be submitted upstream B) isn't upstream or C) is explicitly not acceptable upstream?
17:07 <msw> (this (kernel patches) has gotten *way* better than it was in my days at RHT)
17:07 <f13> indeed.
17:07 <jasonc> yeah, kernel has improved greatly in that regard
17:07 <pzb> f13: umm, RHEL 2.1 kernel isn't very pretty
17:07 <jasonc> which is why I said "you could argue kernel.org"
17:08 <f13> pzb: umm, RHEL2.1 is by far not fedora.
17:08 <f13> jasonc: I see, I misunderstood the statement.
17:08 <pzb> f13: sorry, I thought you were saying Red Hat
17:08 <f13> pzb: and even still RHEL2.1 is ancient.  RHEL5 would be a better 'Red Hat' argument (:
17:10 <jasonc> f13: generally I'm not that hard to read, irritible, argumentative, and abusive yes, but not hard to read :P
17:10 <f13> jasonc: yeah (:  I didn't get the context switch of 'you' in that statement.  sorry for going off there.
17:12 <jasonc> f13: heh, no problem
17:13 <f13> I'll freely admit that our java stuff is um... special.  If I had my way, we wouldn't have any java stuff in our distro, but alas...
17:13 <f13> and rpm is well rpm.
17:13 <jasonc> I really had no intention of firing off this massive storm, I have a nasty habit of doing that
17:13 <f13> aside from these, I really don't think we're 'fork happy' and in fact we're very anti-fork, very very strongly workign to get any fixes we need at least submitted upstream.  Whether or not upstream takes them is well...
17:15 <jasonc> sure, each project has its own goals
17:15 * fray just hopes Fedora is as receptive of his patches as he hopes Fedora's upstream is....
17:16 <f13> "Fedora" also does leave a _lot_ to the maintainer, without a ton of overhead preventing them from doing work.  We hope that the individual maintainer is following our goals, and sometimes we have to stop the car and having a talking to, but Fedora is largely about the individual maintainer doing what they think is best.
17:26 <jasonc> and on that note I need to head out
17:26 * jasonc apologizes to jbj for the massive scrollback he inadvertently caused
17:29 <f13> jasonc: have a good evening
17:29 <f13> or day, or whatever.  
17:30 ::: mspevack!i=mspevack@fedora/mspevack has left #rpm: "Leaving"
17:33 ::: JuanDRay!n=jrfuller@65.215.14.158 has quit: Read error: 110 (Connection timed out)
17:44 ::: Pingoomax!n=Maxime@lal69-1-82-67-13-150.fbx.proxad.net has quit: "Quitte"
17:55 ::: AedinAZ!n=aedinaz@gmk.caoslinux-alliance.net has quit: Remote closed the connection
18:07 <bernardl> is there a chan for yum on freenode?
18:07 <rizzo> doesn't look that way
18:07 <bernardl> hmmm
18:07 <rizzo> I would have guessed #yum anyway
18:08 <bernardl> nobody there
18:08 <bernardl> same with yum-devel
18:08 <rizzo> right
18:08 <bernardl> actually i was just wondering if there is a way to tell yum to just use ONE repository specified on the command prompt
18:14 <bress> bernardl: Read the man page: --disablerepo is the option you want.
18:17 <bernardl> bress: yes but i need to disable everything else, but rather i just want to choose one :)
18:42 ::: Rathann!n=rathann@debi.pekin.waw.pl has quit: "Gone postal."
19:09 ::: majorowl!n=matt@host-136-22.arin.net has quit: "Leaving"
20:05 ::: ixs!n=andreas@lacht.ueber.gattinnen-im-netz.de has joined: #rpm
20:19 ::: niemeyer!n=niemeyer@201.14.38.176 has quit: Read error: 110 (Connection timed out)
20:53 <jasonc> bernardl: bress steered you in the right direction, check the man page for --enablerepo
20:54 <bernardl> jasonc: doesn't --enablerepo only enable repo that is disabled?
20:56 <jasonc> bernardl: don't think so, but I'm speaking from memory here, might be totally off base
21:00 <bernardl>        --enablerepo=repoid
21:00 <bernardl>               Enables a specific repository that has been disabled in the con-
21:00 <bernardl>               figuration file using the enabled=0 option.
21:00 <bernardl>               Configuration Option: enabled
21:00 ::: Foolish!n=foolish@171.84-48-54.nextgentel.com has quit: "Ex-Chat"
21:01 <jasonc> ok, totally wrong, nevermind :)
21:02 <jasonc> actually now that I see that I think I had this same issue a while back and ending up using disablerepo for everything I didn't want... kinda verbose command
21:13 <bernardl> yes, that would work, but it's a bit verbose as you mentioned
21:13 <bernardl> i want to update rpms and i know they're coming from my own local repo, so i just want it to talk to my own repo instead of the other updates and stuff
21:13 <bernardl> i even tried -c blah.repo
21:13 <bernardl> that didn't fly :(
21:19 <z00dax> bernardl: have you looked at protectbase/ repo priorities plugins for yum ?
21:21 <bernardl> hmmm
21:22 <bernardl> that's plug-ins right
21:23 <bernardl> actually that probably doesn't help me
21:24 <z00dax> in 15 words, what are you trying to achieve ?
21:27 <bernardl> i want yum to just use a repo i specify on the command line
21:29 <f13> jasonc: bernardl: enable/disable repo takes globs
21:30 <f13> --disablerepo=\* --enablerepo=repoIwant
21:30 <bernardl> REALLY
21:31 <bernardl> let me try that
21:32 <jasonc> f13: good to know
21:33 <bernardl> f13: which version of yum supports that?
21:33 <z00dax> 2.2+
21:33 <bernardl> ok
21:33 <z00dax> not sure about 2.0, but i can check if you want ?
21:33 <bernardl> well it's good to know that it works ;-)
21:33 <bernardl> 2.1? :)
21:34 <z00dax> umm no, centos-3 has 2.0, centos-4 has 2.4
21:34 <bernardl> fedora core 3 has 2.1
21:34 <z00dax> humm ? fc3 never got 2.2 as an update ?
21:34 <bernardl> well it's possible i haven't updated it ;-)
21:35 <z00dax> good move
21:35 <bernardl> heh dag has 2.4
21:42 ::: JuanDRay!n=jrfuller@user-0ccemri.cable.mindspring.com has joined: #rpm
23:00 ::: Daveman!n=dave@pool-141-150-44-49.mad.east.verizon.net has quit: Read error: 104 (Connection reset by peer)
23:27 ::: laroche!n=laroche@p54A0915C.dip0.t-ipconnect.de has quit: Read error: 110 (Connection timed out)
23:35 ::: Matrix9!i=MiniMe@s142-179-197-109.ab.hsia.telus.net has joined: #rpm
--- Log closed Fri Dec 15 00:00:56 2006