--- Log opened Mon Jun 20 00:00:25 2005
00:02 < dmacks> miga: If you change the Depends in libgnome2 from "esound-shlibs" to "esound", does the circular-dep go away? That's the only dependency change I made.
00:03 < miga> Wait a moment, I'll make the change.
00:03 < theid> dmacks: I fixed the uint32_t thing by adding <stdint.h> - might have been you that pointed me there, even
00:03 < dmacks> theid: Ah right...I remember a discussion about uint vs u_int, didn't know why we were discussing it:)
00:04 < theid> so yeah, that was all it took
00:05 < theid> I'm currently having trouble with the latest gnome-vfs2-ssl, but I think I might be getting close to a solution if I don't keep getting distracted by things, like accepting job offers and moving
00:05 < miga> dmacks, no now it happens on libgettext3-dev.
00:06 < dmacks> miga: Starting to sound like a more general engine problem that vasi was mentioning.
00:06 < vasi> miga, which fink are you using?
00:07 < miga> Yes, but now, it is gettext-dev. The latest one from cvs HEAD.
00:07 < vasi> if it's HEAD, and you've updated it in the last couple of days, then these are spurious and my fault
00:07 < vasi> i suspect it has to do with get_splitoffs
00:07 < miga> I've updated maybe an hour ago or so.
00:09 < dmacks> theid: libgnome2 checks for the presence of stdint.h, but then never bothers to include it if it's present:(
00:10 < dmacks> Let's see if gtk+2 is any smarter.....
00:11 -!- xhrl [~ThomasW@24.80.39.250] has joined #fink
00:11 < miga> dmacks: the discussion on uint was for gpdf, so that means that probably, it should have the same include.
00:12 < dmacks> Ayup, gtk+-2.6.7 also checks for it and ignores the check result.
00:12 < miga> All gtk applications also.
00:13 < theid> looking at my gnome directory I see gtk+2, gnome-vfs2-ssl, and nautilus all needing the stdint.h patch
00:13 < vasi> ok, i see where it's going wrong in the dep engine...in Engine:1675 it looks for selected relatives, but with new Fink it doesn't find them
00:14 < miga> there are also gtk applications which are in crypto or editors or whatever.
00:14 < dmacks> So whenever we #include sys/types.h we should also include stdint.h?
00:15 < miga> Maybe not systematically, but does it hurt if it is included and not needed?
00:16 < vasi> there's maybe a package that defines things on its own
00:16 < vasi> but generally it shouldn't hurt
00:16 < dmacks> I wonder if something else in /usr/include that used to include stdint.h on 10.3 no longer does on 10.4.
00:17 -!- xhrl [~ThomasW@24.80.39.250] has quit [Remote closed the connection]
00:17 -!- xhrl_ [~ThomasW@24.80.39.250] has joined #fink
00:23 < vasi> ooh i think i found the problem
00:24 < pogma> panther has int32_t etc in /usr/include/machine/types.h
00:24 < pogma> well. ppc
00:26 < dmacks> On panther, uint16_t appears to only be in stdint.h
00:26 < dmacks> (which includes machine/types.h)
00:27 < pogma> Ah, his error was 16
00:27 < dmacks> Yeah...that one seems special. Dunno why.
00:32 < vasi> heh, i found a logic bug that happens to be in Fink::Engine while debugging this
00:32 < dmacks> heh.
00:32 < vasi> dmacks, in Engine around line 1690 it says $candcount++.....there should be a 'last;' after that, right?
00:33 < vasi> or 'next SIBCHECK'
00:34 < vasi> i'm not extremely familiar with the dep engine, so i'd like confirmation :-)
00:34 < pogma> also 1654 ?
00:35 < dmacks> Ah, because more than one sibling could satisfy (somehow I guess) so $candcount could be >1 and then $found won't get set?
00:35 < vasi> yep dmacks
00:35 < vasi> pogma, i don't see anything relevant on 1654, we're probably using slightly different revisions
00:36 < vasi> gimme some context
00:37  * dmacks hates huge variable scopes. *grrrr*
00:37 < vasi> fink's old policy was to predeclare all vars (like C)
00:38 < vasi> eventually you and I put an end to that <g>, but chrisp and max coded that way
00:38 < dmacks> Yeah. It's confusing as hell here.
00:39 < vasi> pogma?
00:39 < vasi> dmacks, i'm going to move the pixbuf loader in librsvg2 to the -shlibs splitoff
00:39 < vasi> cuz it doesn't belong in a BDO package
00:39 < dmacks> vasi: I'm not sure what should happen if more than one sib satisfies. Clearly $found should be be set, but what should $dname be?
00:39 < vasi> and the gtk+2 pixbuf loaders are in shlibs
00:40 < vasi> $dname = $cand
00:40 < vasi> $cand does not depend on $splitoff
00:40 < pogma> vasi: Sorry, my mistake
00:40 < vasi> pogma, s'ok :-)
00:40  * pogma fails to understand real_install
00:41 < vasi> pogma, i don't think anyone understands all of it
00:41 < vasi> it BADLY needs refactoring....next time TheSin needs a project ;)
00:41  * dmacks confused, not sure what the heck is supposed to happen here. Ask TheSin.
00:41 < dmacks> re:moving pixbuf-loader: sounds good.
00:42  * dmacks hesitates to ask why I'm associated with that pkg:)
00:42 < vasi> er, cuz it's gnome-core
00:42 < vasi> er, fink-gnome
00:42 < vasi> whatever :-)
00:42 < vasi> and i don't know who else to ask about that stuff!
00:42 < vasi> miga didn't know
00:43 < miga> vasi: I know very, very little.
00:43 < dmacks> Does the .dylib use the loaders&engines, or vice versa?
00:43 < vasi> it's ok miga, fink has vast areas where nobody knows anything
00:43 < vasi> i think the loaders use the dylib
00:44 < dmacks> Technically they should prolly stay in %N and move all the BDO to a new %N-dev pkg, but that's a pain in the ass.
00:45 < vasi> hmm, did i ask you already for your thoughts about autosplitting the Fink .pm's?
00:46 < dmacks> Yeah, just move 'em to -shlibs, and if %N ever changes lib major version and screws up their namespace, we can fix it then:)
00:46 < vasi> i could move the loaders and executables to a librsvg2-common package like debian does
00:47 < dmacks> Anything other than move-to-%Nshlibs is gonna create Work For You to fix other pkgs' deps:)
00:47 < dmacks> You asked about autosplit, but then we got sidetracked on cirdan's bug.
00:48 < vasi> ah...so?
00:49 < dmacks> What's the advantage? Do we usually not-need most functions in a given .pm?
00:49 < vasi> well fink list for example has to load all of PkgVersion
00:50 < vasi> which can take as much as a full second
00:50 < vasi> it uses almost none of it, just a few accessors
00:51 < vasi> (only two packages are improperly depending on librsvg2, so moving stuff to -common should be ok...)
00:52 < dmacks> I think autoloader won't work for some of our functions, but most shuold be okay in PV.
00:53 < dmacks> Using -common seems "non-obvious" to Joe Packager.
00:53 < dmacks> (unless -shlibs depends on it; then it's irrelevant)
00:55 -!- chdevers [~cdevers@pool-68-163-235-132.bos.east.verizon.net] has left #fink []
00:55 -!- rudy__ [~rudy@ip68-224-184-250.hr.hr.cox.net] has quit ["gone to sleep"]
00:56 < dmacks> To start, try with SelfLoader instead of AutoLoader...much easier migration, no changes needed to inject/install scripts.
00:58 < vasi> well already every package is leaving out the loaders, because they're in a BDO package....so it can't get worse :-)
00:58 < dmacks> True.
00:59 < dmacks> A while back, people whined that "sometimes %N is headers, sometimes %N-dev (the latter makes more sense); sometimes runtime tools are in %N, which is confusing since sometimes that's the headers"
01:00 < vasi> yeah, well what can we do, we're stuck with legacy
01:00 < dmacks> If you want legacy, Shlibs Policy says it should go in -bin :p
01:00 < vasi> the Right Thing To Do is to provide a set of "install targets" for users, so that most users have no idea about our crazy package names
01:01 < vasi> if a user actually cares where the headers are and such, they can look it up themselves :-)
01:01 < vasi> though i did suggest to msachs that after a buildfink we list what-files-are-in-what-package and make that publically available
01:01 < dmacks> "ThisThingUses:foo" which propagates BDep (and knows about BDepInherited) to %N and whatever Dep to all pkgs in the .info?
01:01 -!- xhrl_ [~ThomasW@24.80.39.250] has quit [Client Quit]
01:02 < vasi> heh, or just DoWhatIMean: true
01:02 -!- xhrl [~ThomasW@24.80.39.250] has joined #fink
01:02 < dmacks> Clef had a workingish implementation of a pkg contents query, but doesn't seem to have ever activated it.
01:02 < vasi> yeah, i remember that
01:02 < vasi> only worked for stable unfortunately
01:03 < dmacks> Hehe yeah...right now DWIM is an enum(false) :(
01:04 -!- xhrl [~ThomasW@24.80.39.250] has quit [Client Quit]
01:06 < dmacks> A not-crappy user interface gets pretty difficult to implement when the buildfink lags behind the cvs/rsync versioning.
01:06 < vasi> miga, please update your Fink, it should work now
01:06 < miga> OK.
01:07 < vasi> it would be nice to eventually have a testing distro that is never committed to directly
01:07 < vasi> instead commits to unstable would be built as they come in, and tested for validation and such
01:08 < vasi> and if they're ok, automatically dumped into testing
01:08 < dmacks> I guess we could have a separate "package technical details" page (!= package.php) that gave direct access to the .deb and package description files, and file lists? Or just a web interface to 'dpkg -S'?
01:08 < vasi> that way we could actually ensure that most things work, as opposed to unstable where lots of things work for you if and only if you have the same setup as the maintainer
01:09 < dmacks> yeah
01:09 < vasi> dmacks, just steal the stuff from packages.debian.org ;-)
01:09 < vasi> it seemed that msachs wanted something like that to happen, i don't know if there are plans though
01:09 < dmacks> Yeah, we already stole dpkg and apt-get, hell we even took dselect for some reason! May as well take their website too:)
01:09 < vasi> hehe dselect
01:10 < vasi> the main problem with auto-building and -testing is that we have no build farm :-(
01:10  * dmacks strongly disagrees with packages becoming available in any official form without human interaction.
01:11 < vasi> uh, there's a human committing them to unstable
01:11 < vasi> then a script does unstable -> testing
01:12 < dmacks> Right. I'm just saying don't want to go too far towards automation "why not have submissions-queue feed a 10.x/testing tree that everyone can use?" for example
01:12 -!- statico [~statico@langworth.com] has quit [Remote closed the connection]
01:12 -!- statico [~statico@langworth.com] has joined #fink
01:12 < vasi> yeah, it still has to be someone with a commit bit who puts it in unstable
01:12 < dmacks> (/me was thinking of "testing" as less-stable than "unstable")
01:13 < vasi> well debian uses stable > testing > unstable
01:13 < vasi> so it would be confusing as hell if we did stable > unstable > testing
01:13 < dmacks> So it's technically "test_ed_" :)
01:14 < dmacks> (I know I know, "go tell debian devels to fix [whatever]")
01:14 < vasi> <g>
01:16 < vasi> do you know if msachs is doing another buildfink atm?
01:16 < dmacks> I don't think one is running at this time.
01:16 < vasi> hmmm...which methods couldn't we have auto-loaded?
01:17 < dmacks> (corrections for the BuildAsNobody are still coming in)
01:17 < dmacks> IIRC, methods that have lexically scoped variables (caches, for example).
01:18 < dmacks> (at least it used to be that way.../me hasn't played with Autoloader in a few yrs)
01:19 < vasi> oh Package lexicals, yeah
01:19 < vasi> but we could just turn those into 'our' globals
01:20 < dmacks> True.
01:21 < dmacks> Broadly-scoped scratch vars with generic names kinda sucks though, and using specific names gets cluttery.
01:21 < dmacks> I think it's only 2-3 methods...not a big deal.
01:22 -!- mccbaka [~Snak@71.98.81.176] has quit ["-"]
01:25 < vasi> hmm...i also want to not load growl until we try to notify
01:25 < vasi> cuz it takes too long
01:25 < dmacks> Yes! /me has hated that default.
01:25 < vasi> and i have another way to speed up PDB loading
01:25 < vasi> so hopefully i can cut a few seconds off in total
01:25 < dmacks> overclock your CPU?
01:27 < vasi> heh, with a 133 MHZ system bus what's the point?
01:27 < dmacks> So overclock the bus driver. C'mon man, *think*!
01:27 < theid> please, think of the children!
01:28 < dmacks> ha!
01:30 < vasi> hmm, SelfLoader is apparently slower to startup cause it has to parse __DATA__ all at once
01:31 < dmacks> Right.
01:33  * dmacks just committed a probable fix for 10.4T gtk+2
01:37 < theid> to unstable or experimental?
01:37 < dmacks> unstable
01:37 -!- eno-away [~eno-away@adsl-68-123-123-227.dsl.snfc21.pacbell.net] has quit ["BitchX: a modern client for a old world"]
01:37 < dmacks> Appears to be fixed upstream in the just-released 2.6.8 also.
01:45 < miga> dmacks: would have it been possible to use libgettext3 instead of gettext in 10.4-transitional for all gnome?
01:46 < dmacks> miga: I assume so. I was aiming for "minimal changes"...this upgrade was painful enough already.
01:46 < vasi> damn, now librsvg2 doesn't build?
01:47 < vasi> miga, dmacks, have you noticed that the Help in most gnome apps doesn't work? do you know why?
01:47 < dmacks> I have not tried it.
01:47 < miga> Yes, never worked.
01:47 < dmacks> woohoo--not my fault!
01:48 < miga> That's not new, it has never worked. There is some hardcoded path I don't remember where, maybe glib2 or something like that.
01:48 < vasi> miga, ok just making sure it's not my fault either ;-)
01:48 < vasi> any of you ever seen this? /sw/bin/openjade:../sgml/rsvg.sgml:217:191:E: character ":" is not allowed in the value of attribute "LINKEND"
01:48 < dmacks> New to me.
01:48 < miga> No, vasi, absolutely sure it's not your fault.
01:49 < miga> That's probably a too old doc.
01:49 < miga> Wait, I'll try it.
01:51 -!- Feanor [~astrange@feanor.developer.opendarwin] has quit [Read error: 104 (Connection reset by peer)]
01:52 < vasi> that's trying to build librsvg2....i think it's a missing builddepend
01:53 < miga> maybe also some of docbook, sgml-entities, docbook-utils?
01:54 < vasi> gtk-doc is there
01:55 < vasi> those are all installed
01:55 -!- shreyas [~sshreyas@202.144.86.147] has joined #fink
01:55 < shreyas> Morning all 
01:56 < dmacks> howdy shreyas 
01:57 < shreyas> I am trying to build Evolution on Mac. At the end of the build process i want to build/ submit fink packages such that it is easy for mac users to do something like ap-get install evolution 
01:57 < shreyas> ofcourse i have no expert knowledge of how easy/ tough the above process is 
01:58 < dmacks> Fink doesn't have a way for users to submit compiled packages for redistribution at this time.
01:58 < shreyas> could somebody educate me please. Evolution has a whole load of dependencies on gnome packages as most of u would knoe
01:58 < vasi> yeah, you should be able to just do 'fink install evolution'
01:58 < vasi> we already have a package for it i believe
01:58 < vasi> and fink will take care of all the dependencies
01:58 < dmacks> Yup.
01:59 < vasi> if you have a group of people who you'd like to distribute to unofficially, that's ok, we have instructions on our site for making your own bindist
01:59 < vasi> !bindist
01:59 < Melian> i heard bindist is coming. Soon. No we don't know exactly when. It will be well-announced when the time comes.
01:59 < vasi> damn, not the right bindist
01:59 < vasi> http://fink.sourceforge.net/doc/advanced/bindist.php?phpLang=en
01:59 < dmacks> Melian: forget bindist
01:59 < Melian> dmacks: i forgot bindist
02:00 < vasi> Melian, bindist is http://fink.sourceforge.net/doc/advanced/bindist.php?phpLang=en
02:00 < Melian> vasi: okay
02:00 < vasi> shreyas?
02:01 < shreyas> vasi: well, The fact is eventually after i master the art i would like to see current versions of evolution into fink. Who currently updates the evolution fink packages 
02:02 < vasi> 'fink info fink' will show who the maintainer is
02:02 < vasi> you might want to check our experimental distribution to see if they're already working on a new version
02:03 < shreyas> Its my full time job currently to work on the mailer ( i work in novell ) also part of my job description is evolution on mac. So i am trying to work out the best way to distrubute probably weekly builds and so on 
02:03 < vasi> shreyas, it looks like version 2.0.1 is in experimental/gnome/2.8 right now
02:04 < vasi> you can start from there if you like
02:04 < dmacks> I'd love to have someone with some skill take an active role in maintaining that pkg. Lots of users complain about how old ours is:(
02:05 < vasi> shreyas, if you'd like to work on it and take over the package, feel free to submit it to the package tracker on our website
02:05 < shreyas> dmacks: well i dont know abt skill but i would like to get the experimental/gnome/2.12 like rest of the gnome community 
02:06 < dmacks> Yeah. Our whole gnome section has fallen into disrepair:( This past 2 weeks we've gotten many of the core libs up to 2.11
02:06 < vasi> so based on that, you can probably get most of the deps evolution is needed
02:08 -!- broeken [~chatzilla@fswfirewall.fss.uu.nl] has joined #fink
02:08 < dmacks> If you need some new version of [some lib], let us know. Otherwise everyone kinda coasts along with no priorities.
02:09 < shreyas> dmacks: yes, i will write up a nice list and send it to the mailing list
02:09 < shreyas> dmacks: hopefully we can get everything upto date and working nicely 
02:09 < dmacks> That'd rock!
02:09 < vasi> that would be great
02:10 < vasi> one good thing to do would be to put our progress on a webpage somewhere
02:10 < vasi> maybe the metapkg wiki?
02:11 < dmacks> "just check the pdb" har har har
02:12 < htodd> gtk+2 builds now that I installed xcode-2.1
02:12 < dmacks> The pdb, which until 2 weeks ago still thought we were at kde3.0, gnome1.4.
02:13 < shreyas> What i would like to do right now is first build the current cvs of evolution there were some problems on the mailing list wrt that and correspondingly i would work out the list of update packages we would need 
02:13 < vasi> shreyas, sounds like a good plan :-)
02:13 < shreyas> does that sound ok to u guys ?
02:13 < vasi> note that we prefer release versions usually
02:13 < dmacks> htodd: when did you selfupdate? (i.e., does your gtk+2.patch end with a block about stdint.h?
02:13 < vasi> but if CVS has a lot more features or something, then we can use CVS versions for now
02:13 < htodd> no, this was all before your patch
02:13 < dmacks> shreyas: Sounds good.
02:13 < dmacks> htodd: Interesting!
02:14 < shreyas> vasi: even in the unstable branch ?. surely you guys would want the latest evolution somewhere ... and 2.4 is not that far ;)
02:14  * dmacks wonders if we can shake down apple.com for a changelog of their /usr/include :)
02:15 < vasi> shreyas, usually latest release -> unstable, CVS -> experimental
02:15 < vasi> just because CVS versions often have many extant bugs
02:16 < dmacks> If CVS works well, we can have a snapshot. Technically, we need a URL to a tarball with a stable MD5.
02:16 < shreyas> right, thats interesting choice of words :-)
02:16 -!- eno-away [~eno-away@adsl-68-123-123-227.dsl.snfc21.pacbell.net] has joined #fink
02:19 < dmacks> Okay, I gotta be at work "human hours" tomorrow. So g'nite all.
02:20 -!- dmacks [~dmacks@dmacks.active.supporter.pdpc] has quit ["leaving"]
02:23 < vasi> toodles
02:26 -!- shreyas [~sshreyas@202.144.86.147] has quit ["Gnite"]
02:29 < Clef> hmm? my pkg contents thing works for all builds
02:29 < Clef> not just stable
02:29 < Clef> only reason its not activated is i have no SQL server to use
02:29 < Clef> don't trust OD's or SF's
02:31 < Clef> rustle me up one...
02:35 -!- shreyas [~sshreyas@202.144.86.147] has joined #fink
02:37 -!- spundun|sleep [~spundun@c-24-126-63-203.hsd1.ca.comcast.net] has joined #fink
02:38 < lisppaste> Spundun pasted "error when trying to install gnome-desktop through fink commander" at http://paste.lisp.org/display/9235
02:38 -!- spundun|sleep is now known as Spundun
02:39 -!- Fang [~Fang@2002:56c1:e559:1:0:0:0:1] has joined #fink
02:39 < Spundun> can someone checkout the above paste? 
02:39 < Spundun> I just installed Fink on my Tiger yesterday, and trying to install gnom-desktop as the first install..
02:52 < vasi> Clef, sqlite?
02:53 < vasi> Spundun, i suspect you have to scanpackages then apt-get update
03:04 < Spundun> vasi: ok.. lemme try that.
03:07 < lisppaste> Spundun annotated #9235 with "ok tried a commandline version" at http://paste.lisp.org/display/9235#1
03:07 < Spundun> see the above paste... how do I install X.org?
03:07 < Spundun> maybe I should read docs
03:08 -!- kane_ [~kane@perl.xs4all.nl] has joined #fink
03:08 < Spundun> But still for a simple user, something like installing fink. and installing finkcommander and clicking on gnome-desktop should work.. which doesnt... is that easy to fix?
03:12 < Spundun> fink install xorg returns "Failed: no package found for specification 'xorg'!" ... I have already done fink scanpackages and apt-get install, in that order...
03:18 < lisppaste> michga annotated #9235 with "See pdb" at http://paste.lisp.org/display/9235#2
03:20 < miga> You may install Apple's X11.
03:21 < miga> And while you're at it X11SDK also.
03:21 < miga> It's on the Tiger DVD, as optional installation.
03:22 -!- theid is now known as theid_away
03:24 < miga> vasi: for librsvg2, two options may work: first once, installing the previous version of gtk-doc and see it compiles (it seems that the new one is more strict), or trying to install the new upstream version of librsvg2. Anyway as it is dependencies are OK, the only problem is the doc which is pure sgml, the new upstream version has xml doc.
03:31 < Spundun> miga: or enable fink unstable?
03:32 < miga> yes, also.
03:32 < Spundun> miga: thanks...
03:33 < miga> well, I don't know anything about FinkCommander and apt-get, so...
03:35 -!- shreyas [~sshreyas@202.144.86.147] has quit ["Gnite"]
03:41 -!- cianhughes [~cian@cian.ws] has joined #fink
03:42 < gecko2> hmmm
03:42 < gecko2> Failed: Can't resolve dependency "growl (>= 0.7-1)" for package "mac-growl-pm586-0.7-10" (no matching packages/versions found)
03:42 < gecko2> (10.4)
03:56 -!- Albie [~ambs@eremita.di.uminho.pt] has joined #fink
04:08 < vasi> gecko, it's a virtual dependency
04:09 < vasi> you need to have installed growl first
04:09 < vasi> http://growl.info
04:10 < miga> vasi: how could I say foo >= x.y and foo <= x.z where z>y in Depends or BuildDepends?
04:11 < vasi> miga, i think this works, Depends: foo (>> x.y), foo (<< x.z)
04:11 < vasi> at least dpkg/apt-get respect that
04:11 < vasi> i don't know if fink's dep engine deals with it properly
04:11 < miga> OK, I'll test it with librsvg2.
04:12 < vasi> are you working on that now? i was about to try
04:12 < miga> As you want the problem is the header part in docs-rsvg.sgml;
04:12 < vasi> i thought i'd just get the new version instead
04:12 < gecko2> vasi: hmm, mayby the error should be more exact
04:12 < miga> Currently I'm rebuilding gtk+2.
04:13 < miga> Probably it's better.
04:13 < miga> I have recompiled bundle-gnome, and all compiles in 10.4-transitional.
04:15 < vasi> should i use the 2.9.x or 2.8.x version for librsvg2?
04:16 < miga> 2.8.x I'll guess.
04:16 < vasi> ok
04:17 < miga> I think that's OK for all dependencies.
04:17 < miga> From the configure.in.
04:20 < miga> 2.9 requires libcroco3, which is not in Fink.
04:34 -!- KraMer [~mark@adsl-70-240-193-143.dsl.hstntx.swbell.net] has left #fink ["Leaving"]
04:37 -!- shreyas [~sshreyas@202.144.86.147] has joined #fink
04:38 -!- KraMer [~mark@adsl-70-240-193-143.dsl.hstntx.swbell.net] has joined #fink
04:38 -!- shreyas [~sshreyas@202.144.86.147] has quit [Client Quit]
04:39 -!- shres [~sshreyas@202.144.86.147] has joined #fink
04:41 -!- ringerc [~craig@dsl-202-72-144-62.wa.westnet.com.au] has joined #fink
05:38 < vasi> miga, are you around?
05:38 < miga> Yes
05:38 < vasi> 2.8.1 also needs croco3, so i ported it
05:38 < vasi> do you think i should set Maintainer: gnome, or me?
05:38 < miga> Oh, perfect. 
05:39 < miga> You, it's better, otherwise, it has worse as Maintainer: none.
05:39 < vasi> yeah, ok :-)
05:39 < miga> s/has/is.
05:39 < vasi> by the way, if you ever decide to take over a gnome package, that's ok by me
05:40 < miga> vasi, I'd like, but frankly, I'm so null :-(
05:40 < vasi> don't say that miga, you've been a lot of help
05:40 < miga> jfm wants me to take mozilla.
05:40 < vasi> oooh wow
05:40 < vasi> hope you have a fast Mac :-)
05:40 < miga> already he has succeeded to make me work on epiphany.
05:41 < miga> not especially fast an imac.
05:41 < miga> but it's better than the dead g3.
05:41 < vasi> hehe
05:42 < miga> The problem for me is not the time, I have plenty of time, the problem is skills and knowledge.
05:43 < vasi> well you seem to be doing pretty well so far
05:43 < vasi> none of us know everything :-)
05:43 < miga> No, don't think so, people helped me a lot.
05:43 < vasi> i mean i look back at the stuff i did a few years ago when i just started working with fink, man i was soooo dumb ;-)
05:44 < vasi> and eventually i learned some of the tricks...but i still make a lot of mistakes
05:44 < miga> and woman peor :-)
05:44 < miga> just kdding (could not resist).
05:45 < miga> See, I wanted to begin with library with devhelp, which is really very easy, I cannot.
05:45 < vasi> if you want to show me what's going wrong, maybe we can work through it together?
05:46 < miga> That would be very nice, yes, maybe I can understand.
05:46  * vasi thinks we need a Fink mascot to go around cheering us up when things get hard :-)
05:46 < miga> You can tell her to be here all day long for me :-)
05:49 < vasi> so about that library...
05:49 < miga> At the moment, the info file is not correct, it's missing dependencies, but that's not the problem when you have a minimum of gnome stuff installed and mozilla it works (I make the true true dependencies at the end).
05:49 < miga> lisppaste:url?
05:49 < lisppaste> To use the lisppaste bot, visit http://paste.lisp.org/new/fink and enter your paste.
05:50 < lisppaste> michga pasted "devhelp.info" at http://paste.lisp.org/display/9240
05:51  * vasi is lagging...
05:52 < vasi> hmm...i'm not sure if gconf files are supposed to go in ConfFiles? i really don't know!
05:53 < vasi> and it's usually better to have gcc4, otherwise msachs will send you email ;-)
05:53 < vasi> but otherwise it looks ok....should i build it?
05:55 < miga> Woah, it works now. I'm fully astonished.
05:55 < miga> Yes, please build it. But do you have mozilla installed?
05:56 < vasi> i think so, hopefully it's the new version
05:56 < miga> I think the installation is fully wrong: some things go in parent, they should go in shlibs or vice-version.
05:56 < miga> s/version/versa
05:57 < miga> well the otherway around :-)
05:57 < vasi> versa :-)
05:57 < vasi> does it take a long time to build?
05:57 < miga> mozilla? or devhelp?
05:58 < miga> mozilla quite a few hours, devhelp about 2 minutes, if you have the dependencies installed.
05:59 < vasi> ah, ok, i hope my mozilla is up to date then
05:59 < vasi> it's already on the master mirror!
06:00 < vasi> is there already a devhelp package?
06:00 < miga> Oh, maybe someone puts it?
06:00 < vasi> aaaah....darwinports :-)
06:00 < vasi> :-(  it wants to rebuild mozilla
06:01 < vasi> are you on a fast connection? if you can send me the mozilla .deb's it would make this much faster :-)
06:01 < miga> yes, darwinports has 1.7.2, and I think the devhelp version is lower, no?
06:01 < miga> yes, why not, where?
06:01 < vasi> dports has devhelp 0.9.2 :-)
06:01 < vasi> can you DCC?
06:02 < miga> how do I do that?
06:02 < vasi> it depends on your IRC client
06:02 < miga> xchat-aqua here.
06:02 < vasi> probably right-click on my nick or something :-)
06:03 < miga> There are several files.
06:03 < vasi> yay, well that's one of them :-)
06:04 < vasi> i imagine the other ones are bigger
06:06 < vasi> miga, do you use fink from CVS HEAD all the time?
06:06 < miga> yes.
06:07 < vasi> can i ask what you think of it? anything you like, don't like? anything you want added?
06:08 < miga> Yes, I should think of it quietly. But I have no idea what could be added.
06:09 < vasi> 'think of it quietly'? i think that got lost in translation...
06:09 < miga> I should take some time to think about it in depth.
06:09 < vasi> oh ok
06:10 < vasi> i'm going to try to improve startup time some more, that's my next task
06:10 < vasi> (well, after fixing bugs)
06:10 < miga> how do you say that?
06:11 < miga> Oh, yes it takes some time to reindex.
06:11 < vasi> just the way you explained it :-)
06:11 < vasi> to reindex?
06:12 < vasi> do you mean 1) startup when no changes have been made, 2) startup when there are new/modified .info files or 3) startup when upgrading to a new fink?
06:12 < miga> Maybe not reindex, at the end of the inject script when it reads shared libraries.
06:12 < miga> No I mean when injecting.
06:12 < vasi> ah ok, #3 :-)
06:12 < vasi> yeah, that's a problem i can't really fix unfortunately
06:12 < vasi> the only way to fix that is to put a version number on the fink database format
06:13 < miga> That's not a problem for me.
06:13 < vasi> and the other devels don't like that idea, so i can't do anything
06:14 < vasi> if you want to save time, you can just copy *.pm to /sw/lib/perl5/Fink....if there has been a format change that will break and be confusing :-)  but usually it will work
06:14 < vasi> (just don't tell anyone i told you to do that)
06:14 < miga> I think you have all the deps now.
06:14 < vasi> ok
06:15 < miga> No, I prefer to do it the right way, otherwise I will not be able to un-confuse things, if I dare say so.
06:16 < miga> You should have parent, browser, chatzilla, dev, dom-inspector, js-debugger, mailnews, psm, and shlibs.
06:16 < vasi> i'm trying to install debhelp now, let's see if this works
06:16 < miga> all 1.7.5-2 version.
06:18 -!- _BleedAway [whocares@saus04.usc.es] has joined #fink
06:18 < vasi> ok, i think it's working now...a lot of build dependencies to install
06:19 < miga> yes, about the same as conglomerate.
06:19 < vasi> oooh i didn't know about conglomerate...is it good?
06:20 < miga> It all depends what you think "good" is.
06:20 < vasi> :-) ok, do you like it?
06:20 < miga> That is it is very interesting, but very unpracticable.
06:21 < miga> Yes, I'd like it the day I understand how their specification work.
06:22 < miga> Problem is the team project is very short in people and do not have much time to work on it. Hence even less to explain how it works, hence no much people comes to help.
06:22 < vasi> hmm, the screenshots look pretty
06:22 < vasi> when i was working on Tangram before, there was only one developer, and he lived in New Zealand so we were never awake at the same time
06:23 < vasi> very hard to coordinate things
06:23 < vasi> oh sorry, that was "OOPS", not Tangram
06:23 < vasi> i always get them confused
06:24 < miga> That's about the same with conglomerate, one or two people on it coding, another managing the docs.
06:26 < miga> What is interesting in it is that you can work with any kind of xml specification, you have only to build the one you want and it should work more or less. The problem then is to understand how it works.
06:26 < vasi> does it work ok for fink's xml documentation?
06:26 < miga> But for a gtk coder, it should be quite clear, not for me.
06:26 < vasi> i always hate doing that by hand
06:28 < miga> It lacks some docbook stuff, and precisely I offer to work on it some time ago, but as nobody takes time to explain me how it works, I cannot change what is needed, and the team does not know docbook very well. That is we have not find a way to conjuguate our knowledge.
06:28 < vasi> :-(
06:28 -!- BleedAway [whocares@saus04.usc.es] has quit [Read error: 110 (Connection timed out)]
06:28 -!- _BleedAway is now known as BleedAway
06:32 < miga> For fink documentation, I use bluefish. It works well.
06:34 < miga> But not wysiwyg.
06:41 < miga> Does it still compile, vasi?
06:42 < vasi> i'm having a bit of .deb trouble, working it out
06:42 < miga> Oh, the ones I sent you were not good?
06:44 < vasi> no they're ok, i have some others sitting in bad places
06:44 < vasi> which confuses fink because apt can find them but fink can't
06:45 < miga> Oh, ok. Do you want me to send you another debs?
06:47 < vasi> nope, i just wrote a script that takes a list of debs and finds the right directory for them to be in so fink can find them :-)
06:50 < vasi> ok, the debs are working now
06:50  * miga hides, fearing the slaps on the head.
06:51 < vasi> Can not initialize XPCOM Glue
06:51 < vasi> Can not aquire component registra
06:51 < vasi> is that ok?
06:51 < miga> Yes, all normal here
06:51 < vasi> *whew*
06:51 < vasi> ok, now it's finally building!
06:51 < miga> That's not ok, but it's a problem with xprint which is not in X11.
06:53 < vasi> missing builddep on pango
06:53 < miga> I've seen that xorg has made a change in cvs two or three days ago. I don't know if it includes the xprint server, but they said that now firefox does not complaint any more. Hence the same for mozilla (they rely both on a xprint server). And xfree86 does not include either the config files for the server, nor the headers.
06:54 < miga> Oh, well in Depends, a pango-xft2-shlibs maybe.
06:55 < vasi> does fink have a separate thing for xprint?
06:55 < miga> No, it is not in fink.
06:56 < miga> The client is here, but not the server.
06:57 < vasi> woops more missing builddeps
06:58 < vasi> before you commit the package, you should try building from an empty fink installation to catch those :-)
06:58 < miga> That's likely. It's just a rough info file.
06:58 < vasi> yeah, it's ok for now
06:58 < miga> Yes, of course, but as it did not run...
06:59 < miga> But I think I don't understand where to put the Depends and BuildDepends when building a package with a library.
07:00 < miga> In shlibs splitoff or in parent, and how can I know which one in parent, which one in shlibs.
07:00 < vasi> ok now it's compiling
07:01 < vasi> usually i look at the output of the make install, and then figure out how to arrange that into packages
07:05 < vasi> do you know where i can find db2html?
07:06 < miga> docbook-utils I think.
07:10 < vasi> yes i think that's it
07:19 < vasi> hmmm, i keep running into missing libs
07:19 -!- Miaz [~ben@h-68-164-241-34.snvacaid.dynamic.covad.net] has joined #fink
07:19 < miga> which ones?
07:20 < vasi> audiofile, gnome-keyring
07:21 < miga> yes audiofile-shlibs, gnome-keyring-shlibs and maybe gnome-keyring-dev
07:21 < shres> guys, i have glib2 v2.4.6 installed on my system. Now i want to update to 2.6.5. The 2.6.5 package is in my local tree. I wrote a .info file for it. Do i just put it in the old glib2.info location and run fink update glib2 ? 
07:21 < vasi> we have 2.6.4 in unstable right now, why not use that?
07:23 < shres> vasi: you do ? http://fink.sourceforge.net/pdb/package.php/glib2. Am i missing something here ?
07:23 < vasi> the PDB isn't always up-to-date, Sourceforge has a slightly broken server that they make us use
07:24 < vasi> but trust me, if you enable unstable and do 'fink list glib2', you'll see 2.6.4
07:31 < vasi> libjpeg...
07:31 < miga> yes, also, libpng3; libtiff, probably.
07:33 < vasi> is there a way to make a splitoff that only exists for a certain variant?
07:33 < vasi> it doesn't seem like it!
07:33 < miga> Wait a moment. I think baba does that for gdal.
07:34 < shres> hmmm... i dont see it. Wonder what the goof up is !
07:34 < miga> No, maybe I'm wrong.
07:36 < miga> But there is a way, if you do not put anything in the depends, buildepends, etc fields for a particular variant, it does not build it.
07:43 < vasi> shres, have you done a fink selfupdate recently?
07:43 < vasi> miga, which freetype should i link to?
07:43 < vasi> the one in /usr/X11R6, or the one in /sw?
07:43 < miga> freetype2 (not 2.19).
07:44 < miga> the same as mozilla.
07:44 < miga> I think the one in /usr should be good. as anyway it is the same right now.
07:45 < miga> but now that gnome2.10 is here, it is possible to install mozilla 1.8, then freetype2.19.
07:46 < miga> Actually I don't grasp what to do with freetype, it seems very complicated and each maintainer hass his/her own idea about it. I have absolutely none.
07:50 < miga> By the way, from what I've poorly done with devhelp, one can deduct what should be added to the packaging manual so that a beginner maintainer should be able to make a package with library :-)
07:53  * miga hopes pogma's hearing/reading.
08:01 -!- shres [~sshreyas@202.144.86.147] has quit ["Gnite"]
08:02 < vasi> woops, missing scrollkeeper too
08:03 < miga> yes.
08:03 < miga> shame on me.
08:03 < vasi> it's ok :-)
08:04 < vasi> alright, so i did a test install of devhelp
08:04  * miga crosses the thumbs.
08:05 < vasi> oh, i have no idea yet if it works, i'm looking at the layout first
08:06 < vasi> aha, see this package needs (at least) 3 parts
08:06 < miga> It's likely not to be good. Another thing that I don't understand in the manual.
08:06 < vasi> -shlibs for just the libraries
08:06 < vasi> -dev for the headers
08:06 < miga> ah so, it is not clear in the manual.
08:06 < vasi> and .la and pkconfig stuff goes in dev too
08:06 < miga> ok.
08:08 < vasi> some things it's hard to know...for example, the share/locale stuff may only be needed for the executable, or maybe the library also
08:08 < miga> needed at run time.
08:09 < miga> how can I know if it is needed for the library?
08:10 < vasi> well, you can either search the library code for things that use gettext
08:10 < vasi> or, you can just be safe and assume that it is needed for the library, and then include the locales in shlibs
08:11 < vasi> BUT if you do include the locales in shlibs, you have to make sure they have a version number!
08:11 < vasi> cuz otherwise if there's later a devhelp2 library it would conflict with the locales
08:11 < vasi> (it's complicated, eh?)
08:12 < miga> far too much for me.
08:12 < vasi> well this package is a hard one, because it links to sooo much things
08:12 -!- shreyas [~sshreyas@202.144.86.147] has joined #fink
08:13 < miga> as far as I've seen, there is no simple package with library which does not have already a maintainer.
08:14 < miga> And another advantage is that it is no more updated, hence no risk to have conflicts with update.
08:16 < shreyas> vasi: yes, i did that first up. I can see the packages but apparently i need to bootstrap fink or use gcc 4.0 to install them 
08:16 < vasi> there's plenty of packages people want, many of which have libs...
08:16 < vasi> shreyas, explain?
08:16 < miga> If you know of a simplistic one...
08:18 < shreyas> vasi: i try to install the new 2.6.4 packages i get http://pastebin.com/302167 that error
08:18 < vasi> miga, i don't know offhand, but check freshmeat or the package suggestion tracker
08:19 -!- chris01 [~chris01@212.126.165.246] has joined #fink
08:20 < vasi> miga, in this case you can look in po/fr.po
08:20 < vasi> and you see that some files there, like dh-parser.c, are in the library
08:20 < miga> what?
08:20 < vasi> in devhelp
08:20 < miga> oh, ok :-)
08:21 < vasi> so that means the library needs the po files
08:21 < miga> OK.
08:21 < vasi> shreyas, did you try running gcc_select 4.0?
08:21 < miga> Yes, now that you tell me, that's obvious, those are the headings.
08:22 < shreyas> vasi: i dont want to do that unless its the only option  ;-)
08:23 < vasi> you do translation, i'm sure you know more about that than i do
08:23 < vasi> shreyas, why don't you want to?
08:23 < vasi> cuz it kinda is the only option :-)
08:23 < miga> I do not always connect all neurons.
08:24 < vasi> miga, so now if you put share/locale in -shlibs just like it is, that would be bad, do you understand why?
08:24 < shreyas> vasi: i want to build evolution with 3.3 and not 4.0. Is the gcc_select only for fink itself or for packages it builds too ?
08:25 < vasi> shreyas, it's for your whole system
08:25 < vasi> why do you want to build with 3.3?
08:26 < vasi> shreyas, if you tell me what you're trying to accomplish from a high-level view, i can help you better :-)
08:28 < shreyas> vasi: sure, I am trying to build evolution 2.3.xx. But i am slightly sceptical about it building with 4.0. So thats the high-level view. Obviously i cant build glib with 4.0 and then evolution itself with 3.3 ;)
08:28 < cirdan> morning all
08:28 < vasi> shreyas, uh yes you can!
08:28 < vasi> it's only C++ that you can't mix-and-match, but both glib and evolution are C
08:28 < miga> If the locale are updated, that would be bad. But if they are not, is it ok?
08:28 < cirdan> shreyas: the pnly problsms when switching gcc is when building c++ code
08:29 < cirdan> fink takes care of the rest
08:29 < vasi> miga, well that's not quite it...this is a bit hard to understand, so bear with me
08:29 -!- knghtbrd_ [~knghtbrd@d172-210.uoregon.edu] has joined #fink
08:29 < miga> that's more the other way around, no?
08:29 < vasi> the idea of -shlibs packages is that they should NEVER conflict with each other
08:29 < miga> ok
08:30 < miga> so I should not assume that in the future they would not.
08:30 < miga> but how can I version the locales???
08:30 < vasi> so if you have devhelp1-shlibs and devhelp2-shlibs at some point in the future, devhelp1-shlibs will contain libdevhelp.1.dylib and devhelp2-shlibs will have libdevhelp.2.dylib
08:30 < vasi> yeah, so you need to version the locales
08:31 < cirdan> miga: foo-i8n?
08:31 < vasi> well you can set GETTEXT_PACKAGE in the makefiles to something else
08:31 < miga> oh, ok.
08:31 < vasi> er, actually in configure
08:31 < miga> yes, of course. That's more familiar to me.
08:31 < vasi> GETTEXT_PACKAGE=devhelp => GETTEXT_PACKAGE=devhelp1 (or whatever the current version is)
08:32 < vasi> grrr libtool sucks
08:32 < cirdan> hehe
08:32 -!- knghtbrd [~knghtbrd@d172-210.uoregon.edu] has quit [Read error: 145 (Connection timed out)]
08:33 < vasi> it's reordering my -L flags!
08:33 < cirdan> libtool knows best ;-)
08:33 < pogma> vasi: what version?
08:33 < miga> the current version of the package is 0.9.2, but the libdevhelp is 1.0.
08:33 < vasi> 1.5.6 apparently
08:34 < cirdan> miga: are you versioning the locales into the -shlibs deb?
08:34 < vasi> ooh it's not reordering it, it's adding a new -L in front of mine
08:34 < pogma> vasi: Hmm, put the -L flags that must be first in the CC env var before you call configure
08:34 -!- akh_laptop [~akhansen@216.75.154.114] has joined #fink
08:34 < vasi> pogma, tricksy! :-)
08:35 < miga> No, I just say what the upstream version has done; they have a number 0.9.2 for the package, but the libdevhelp has always a 1.0 version from the beginning.
08:35 < vasi> er actually, now that more people are here...what freetype should i be linking to?
08:35 < cirdan> the new one i think
08:35 < cirdan> freetype219 or so
08:35 < vasi> what about the xorg one? avoid that?
08:35 < pogma> vasi: I don't know, wish there were only one?
08:35 < miga> no, it does not work with mozilla 1.7.5.
08:36 < pogma> s/?/!/
08:36 < cirdan> what will be the version number for the deb?
08:36 < miga> cirdan: to me?
08:36 < cirdan> miga: yes
08:36 < miga> the deb are 0.9.2, the dylib 1.0.
08:37 < miga> the latest and last version of deb is 0.9.3. The project leader announces last week he will not continue it.
08:37 < akh_laptop> Makes it easy to maintain.  ;-)
08:37 < miga> Yes, exactly.
08:38 < miga> But it is still interesting as an exercise for me and for easy browsing of the gtk api.
08:39 < akh_laptop> Sure.
08:39 -!- Netsplit orwell.freenode.net <-> irc.freenode.net quits: mee_bot
08:40 < vasi> just wait...next week someone else will decide they need more development, take over the project, and number it 2.0 :-)
08:40 < miga> but, obviously, as vasi said, I should not take that into account. Nobody knows if somebody will not continue it in the future.
08:41 < vasi> akh, were you able to investigate that sqlite3 issue more?
08:41 -!- mee_bot [~mee_bot@dragontooth.servers.class.csupomona.edu] has joined #fink
08:41 -!- chris01 [~chris01@212.126.165.246] has quit [Read error: 104 (Connection reset by peer)]
08:41 < akh_laptop> Not as such.
08:41 < cirdan> miga: if the package version is 0.9.2/3, then version it to that version
08:41 < cirdan> so that 0.9.2 and 0.9.3 -shlibs can be installed at the same time
08:42 < cirdan> in theory
08:42 < cirdan> err...
08:42 < cirdan> wait, nm
08:42 < miga> So maybe I put a devhelp-shlibs with version 0.9.2 and almost nothing in it. and another one, with the locale?
08:42 < vasi> no, the locale and shlibs go together in this case
08:42 < vasi> because the shlibs needs the locale
08:42 < cirdan> the -shlibs replace each other until the lib version increases a major version
08:43 -!- shreyas [~sshreyas@202.144.86.147] has quit ["Bye then"]
08:43 < cirdan> so devhelp1-shlibs and if it ever appears devhelp2-shlibs need to be installed at the same time
08:43 < cirdan> need to be able to be installed at the same time
08:44 < miga> yes, I understand that, what I don't understand is how to make locales versioned.
08:47 < vasi> i told you, you just change GETTEXT_PACKAGE in configure
08:47 < akh_laptop> biab
08:47 -!- akh_laptop [~akhansen@216.75.154.114] has quit []
08:47 < vasi> add a number to the end to represent the version, that's it :-)
08:47 < vasi> then it should make all the locale files with numbers at the end
08:47 < miga> Oh, that's that simple.
08:48 < miga> I look always for complicated things.
08:48 -!- shres [~sshreyas@202.144.86.147] has joined #fink
08:48 < vasi> :-)
08:48 < vasi> it felt complicated the first time i ran into it...tonyarnold and i spent a long time figuring it out
08:48 < miga> Then the problem with the lib version. Can I change the one they put in configure?
08:49 < vasi> what's wrong with the version?
08:49 < cirdan> what problem?
08:49 < miga> They put 1.0, put the package is 0.9.2
08:49 < cirdan> thats ok
08:49 < vasi> that's ok :-)
08:49 < vasi> heh
08:49 < vasi> the library version and package version need not be the same
08:50 < miga> You told me the shared libs should have the same version number.
08:50 < vasi> i did? woops
08:50 < vasi> i didn't mean to
08:50 < miga> not you vasi, cirdan, but maybe I did not understand.
08:51 < miga> OK, that seems to be clear for this part.
08:52 < cirdan> no, the locals version should patch the package version
08:52 < miga> what do you mean here?
08:53 < cirdan> when u do the configure call to version the locale files, use the package version
08:53 < miga> you mean the GETTEXT-PACKAGE-0.9.2?
08:54 < vasi> er
08:54 < miga> OK, I thought you spoke about the libs version, clear now
08:54 < vasi> you should have GETTEXT_PACKAGE = devhelp-1  or something
08:54 < miga> yes.
08:55 < pogma> damn mailer, sorry, "can't send the message" So I hit send again
08:55 < vasi> and then for all the other files, you have to figure out if the shlibs needs it, or just the binary
08:55 < miga> GETTEXT_PACKAGE=devhelp-0.9.2, ok?
08:55 -!- ringerc [~craig@dsl-202-72-144-62.wa.westnet.com.au] has quit ["zzzzzzzzzz"]
08:56 < vasi> no, use 1
08:57 < vasi> because if 0.9.3 still has libdevhelp.1.dylib, you don't want to change the locale version
08:57 < miga> So, the gettext version should match the library version.
08:57 < vasi> yes
08:57 < miga> ok.
08:58 < miga> and the package version still is 0.9.2.
08:58 < vasi> yes
08:58  * miga needs a handtool.
08:59 < miga> OK; now I will probably take the rest of the week to determine the right dependencies and their order. Thanks very much.
09:00 < miga> By the way it's fabulous now, there is no more weird symbols in the doc.
09:01 < vasi> :-)
09:04 < miga> I'd like to know: when a software is using tetex, what part of tetex should it Depends on?
09:07 -!- akh_laptop [~akhansen@68-118-244-23.dhcp.oxfr.ma.charter.com] has joined #fink
09:12 < vasi> oy
09:12 -!- ctd [ctd@ctd.user] has quit [Read error: 54 (Connection reset by peer)]
09:31 < cirdan> heheh
09:33 -!- baba [~baba@YahooBB220041001026.bbtec.net] has joined #fink
09:37 < miga> hi baba.
09:37 < baba> hi miga
09:38 < baba> any solution to your XML problem?
09:38 < miga> I don't know, unless you put a new news.en.xml :-)
09:39 < miga> Maybe it was just temporary.
09:40 < baba> i had such problem when i was hacking Makefile
09:40 < baba> did you do anything with Makefiles?
09:40 < miga> I promise I did not hack any Makefile.
09:40 < baba> i guessed so
09:41 < baba> maybe some new package(s) doesn't work with the current make system
09:41 < miga> That's only with news. The other files are ok.
09:41 < baba> ic
09:42 < baba> maybe because i did something to news.dtd/news.xslt
09:42 < miga> bad baba :-)
09:42 < baba> oh no
09:42 < miga> But did you do it very recently?
09:42 < baba> im more good than bad
09:42 < baba> nope
09:43 < baba> not in the last few days
09:43 < miga> I was jokking with many many :)))
09:43 < miga> of course, you're not bad.
09:44 < baba> hehe
09:45 < miga> So that's anything else probably, because I've updated news on 10.6 without this dump.
09:46 < miga> There is also those strings outside of range from Japanese, but I'll guess those are the ones with three bytes encoding.
09:48 < baba> that shouldn't happen on mac, though
09:48 -!- newmanbe [~newmanbe@503ad105b351b07d.session.tor] has joined #fink
09:49 < miga> Maybe that's because I have not the right encoding as locale? Do you see them?
09:50 < baba> ah yes, indeed
09:50 < akh_laptop> miga, baba:  I get the "outside of range" messages but no dumps.
09:50 -!- akh_laptop [~akhansen@68-118-244-23.dhcp.oxfr.ma.charter.com] has left #fink []
09:50 < baba> outside of range is another issue, i remember
09:50 < baba> hi akh
09:50 < baba> he's gone
09:50 < cirdan> heh
09:50 -!- akh_laptop [~akhansen@68-118-244-23.dhcp.oxfr.ma.charter.com] has joined #fink
09:50 < miga> sorry, hi akh
09:50 < cirdan> !wb akh_laptop 
09:52 < akh_laptop> oops
09:52 < baba> wb akh
09:52 < newmanbe> !wb akh
09:52 < Melian> Welcome back akh, o lonely traveller amongst the TCP/IP packets of chaos.
09:52 < miga> excellent!
09:52 < baba> yeah, "outside of range" happens because the make tries to make a URL for each news for RDF files, but Japanese chars cannot be encoded
09:52 < Melian> Welcome back akh_laptop, o lonely traveller amongst the TCP/IP packets of chaos.
09:52 < baba> or something like that
09:52 < miga> They can, but not if you reduce the encoding to latin1 or ascii.
09:52  * newmanbe is finally read the new bindist announcement.
09:53 < newmanbe> In some places it uses http://www.finkproject.org/ and in other places, http://fink.sourceforge.net/
09:53 -!- Melian [~blootbot@pcp04356153pcs.glstrt01.nj.comcast.net] has quit ["adios amigos"]
09:53 < baba> http://fink.sourceforge.net/ is still the only official, afaik
09:53 < baba> am i wrong?
09:53 < newmanbe> It has both.
09:53 < newmanbe> That is the weird thing.
09:54 < cirdan> baba: well, finkproject is owned by me, but we dont usually use it
09:54 < newmanbe> Except in Melian's factoids. :)
09:54 < cirdan> well, not my doing :-)
09:54 < newmanbe> Some of them are.
09:55 < pogma> fink.sourceforge.net is the only official site
09:55 < akh_laptop> (for now_
09:55 < akh_laptop> )
09:55 < baba> yo pogma
09:55 < pogma> hi baba 
09:55 < newmanbe> Oh, I think the finkproject.org is part of akh's signature (actually header).
09:56 -!- Melian [~blootbot@pcp04356153pcs.glstrt01.nj.comcast.net] has joined #fink
09:56 < akh_laptop> mine? 
09:56 < newmanbe> You are Alexander K. Hansen, aren't you?
09:56 < akh_laptop> Yah
09:56 < newmanbe> The made scientist. :)
09:56 < newmanbe> mad
09:57 < akh_laptop> wish I were 'made' ;-)
09:57 < newmanbe> !lisppaste
09:57 < Melian> [lisppaste] a bot that lets you post large chunks of text without flooding the channel. Use it at http://paste.lisp.org/new/fink
09:58 < lisppaste> newmanbe pasted "akh's header" at http://paste.lisp.org/display/9247
09:58 < newmanbe> Yours.
09:59 < pogma> looks erroneous :)
09:59 < akh_laptop> oops
10:00 < akh_laptop> my bad
10:00 < pogma> no problem, the message got to -announce
10:00 < cirdan> hehe
10:00 < pogma> dmalloc ought to have done it
10:00 < cirdan> !seen dmalloc
10:00 < Melian> dmalloc <~mule@gw.uptime.at> was last seen on IRC in channel #fink, 3d 2h 46m 40s ago, saying: 'well it sinot only this netowrk start happening a lot now all over the place'.
10:00 < akh_laptop> And it's not worth a resend for something that trivial.
10:01 -!- TheSin [~TheSin@iphost-64-56-130-194.edm.wiband.net] has joined #fink
10:02 < newmanbe> If dmalloc would have sent it, it would have been a signed message.
10:03 < akh_laptop> I sent it.
10:03 < newmanbe> I know.
10:03  * pogma wonders what signing has to do with it
10:03 < newmanbe> It would have been magically better.
10:04 < newmanbe> GPG makes everything better.
10:05 < newmanbe> If you have a cold, use GPG.
10:05 < newmanbe> Stomach ache: use GPG.
10:05  * newmanbe stops giving more examples.
10:06 < akh_laptop> !comfort newmanbe
10:06 < Melian> There, there, newmanbe.  It's OK.  I'm here for you.
10:06 < akh_laptop> We'll get you some help  :-P
10:06  * cirdan prescribes 1 lobotomy :)
10:07 < pogma> aw, can't we just kick him and tell him not to cime back until he's better? :-p
10:07  * newmanbe hacks the SF.net's Mailman to make it only work if messages are signed.
10:07 < cirdan> sure...
10:07 < cirdan> mmm
10:07 < akh_laptop> That'd keep the post volume down.  ;-)
10:07 < cirdan> would cut down on bugreports and whiners :-)
10:07 < cirdan> newmanbe: only on -beginners and -users :-)
10:08 < newmanbe> I suppose.
10:09 < cirdan> although it would slow message processing by a few factors of 10 :-)
10:09 < newmanbe> Or increase usage of GPG.
10:10 < vasi> grr
10:10 < vasi> how am i supposed to find out whether or not the compatibility version was *really* supposed to change
10:10  * pogma runs from vasi, thinking he's grr'ing at libtool
10:10 < cirdan> hehe
10:11 < vasi> no, i don't think it's libtool's fault again
10:11 < vasi> (for once)
10:11 < vasi> next time i'm sure it will be :-)
10:11 -!- jtyler [~jtyler@iphost-64-56-130-194.edm.wiband.net] has joined #fink
10:11 < pogma> vasi: In general, we go with upstreams versioning, even if we think it broken
10:12 < vasi> well if compat version changes, it's a whole new lib...and i don't think it's supposed to be
10:13 < cirdan> not for a small change
10:13 < pogma> vasi: if api is added to the library then the compat version should increase and the install_name stay the same
10:13 < cirdan> only major compat version increase changes installname
10:14 < vasi> old things will still work? ah
10:14 < pogma> vasi: yeah, but for packaging it becomes difficult
10:15 < pogma> vasi: old things rebuilt using new library will not work if old one is installed
10:15 < vasi> it's ridiculous how it starts at 2.8.1, then libtool sees 10:1:8, and turns it into major 2, current-version 11:1:0
10:15 < vasi> so things need to use >= depends
10:15 < vasi> damn
10:15 < cirdan> pogma: that is where shlibs: comes in 
10:16 < vasi> cirdan, i don't think TheSin's code includes forwards compatibility
10:16 < cirdan> it's fine
10:17 < TheSin> fowards?
10:17 < pogma> vasi: Well, imo, things that use the new api need to use >= depends
10:17 < cirdan> the thing is maintainers need to doc in Shlibs when the abi changes
10:17 < pogma> vasi: other things, well if the situation arises, telling the user to update foo-shlibs is not too bad
10:17 < vasi> if i have foo, compat 1.0.0 installed, and build bar against it
10:17 < vasi> then bar is still ok when i update foo to 2.0.0, right?
10:17 < pogma> vasi: it is not often that someone updates foo-shlibs, rebuilds bar, and then downgrades foo-shlibs
10:18 < pogma> vasi: yeah
10:18 < vasi> TheSin, but i think your Shlibs code says that's not ok
10:18 < TheSin> it isn't
10:18 < TheSin> as it would be foo2-shlibs
10:18 < TheSin> which is not foo-shlibs
10:18 < vasi> no, this is compat version not major version
10:19 < TheSin> compat version won't matter
10:19 < TheSin> it uses >=
10:19 < TheSin> for the depends line
10:19 < TheSin> so if foo-shlibs changes the compat it won't care
10:19 < TheSin> so long as it's still foo-shlibs
10:20 < TheSin> the problem with that is the deb will differ
10:20 < vasi> will the old .shlibs file stay around?
10:20 < TheSin> no
10:20 < vasi> so then it won't find it
10:20 < TheSin> yes it will
10:20 < TheSin> cause it doesn't care about the compat
10:20 < TheSin> the shlibs system depends on the pkg name for that
10:21 < TheSin> or it should
10:21 < TheSin> now that you mention it I think it oes compare compats
10:21 < TheSin> though it shouldn't
10:21 < vasi> $deb = $self->get_shlib($lib, $compat);
10:21 < vasi> it uses the compat
10:21 < TheSin> I think we needed to add that because if we didn't it would make different debs
10:21 < TheSin> which is against policy
10:22 < cirdan> which needs to be modified somehow
10:22 < vasi> :-)
10:22 < TheSin> no
10:22 < TheSin> it can't be
10:22 < TheSin> it'll break policy
10:22 < vasi> i think you have to define a comparator for compatibility version?
10:23 < cirdan> it should be the shlibs doic the 1st version of the package that has that exact compat # in the shlibs field
10:23 < TheSin> that'll still break policy
10:23 < vasi> no, that still doesn't work
10:23 < cirdan> and that package version is used as the >= when building a deb
10:23 < vasi> crap, this really don't look goof
10:23 < TheSin> if the compat number differs the md5sum of the deb differ
10:23 < vasi> good
10:23 < cirdan> policy needs to be changed
10:23 < cirdan> we already break policy
10:23 < vasi> i'll think about it more, but i think we may need to maintain a Shlibs history
10:23 < newmanbe> Abolish policy!
10:23 < newmanbe> Fink anarchy!
10:23 < TheSin> vasi, I agree
10:23 < cirdan> TheSin: the md5 changes each time the deb is built
10:23 < vasi> cirdan, where is it broken?
10:24 < vasi> cirdan, that's ok, it's just dates and prebinding
10:24 < TheSin> cirdan, right because for file creation dates
10:24 < vasi> Depends are more serious
10:24 < TheSin> totally
10:24 -!- broeken [~chatzilla@fswfirewall.fss.uu.nl] has quit ["Chatzilla 0.9.68a [Firefox 1.0.4/20050511]"]
10:24 < cirdan> i know but that was why we can't care about md5 anything
10:24 < vasi> because then i can't take a deb from my machine and put it on yours
10:24 < cirdan> veos x auto-depends
10:24 < cirdan> vasi: yes you can
10:24 < vasi> ve-who?
10:25 < TheSin> cirdan, we can care, we can't depend on it
10:25 < cirdan> u just need to have the corrrect min version of foo-shlibs
10:25 < TheSin> you need the same compat
10:25 < cirdan> vasi: os x auto depends
10:25 < TheSin> to be safe
10:25 < cirdan> if revision is the same we have the same looking deb in 2 different trees
10:25 < TheSin> if you have the newer one and compile something
10:25 < TheSin> then I have the old compat and try to install it
10:25 < TheSin> I could have missing symbols
10:26 < TheSin> even if it's forwards
10:26 < cirdan> TheSin: if you build foo against foo-shlibs 1 3.0, and install it with foo 1.2.0, it will crash
10:26 < cirdan> TheSin: no, forwards is guaranteed to be compat woth older stuff, until the major version changed
10:26 < vasi> forwards is ok Thesin
10:26 < cirdan> at that point the install name changes too
10:27 < cirdan> sorry, meant foo-shlibs-1.2.0
10:27 < cirdan> above
10:28 < cirdan> vasi: the sholibs field can be used to have a history
10:28 < TheSin> yes forward is fine
10:28 < vasi> the problem is if you have a package bar which links to libfoo.dylib with compat >= 1.0.0
10:28 < TheSin> but that would need a builttime compat
10:28 < TheSin> and a runtime compat
10:28 < TheSin> just to make sure
10:29 < cirdan> instead of changing the line for the new version, we can add a new line for the same file w/the new info
10:29 < vasi> you want to make sure that even if you have libfoo compat=2.0.0 installed, package bar still "knows" it's compatible with 1.0.0
10:29 < cirdan> vasi: no, it's not compat
10:29 < cirdan> 1.x.y is not compat with 2.x.y
10:29 < TheSin> well I gotta get some work done @ $JOB
10:29 < TheSin> but I am against changing to allow >=
10:30 < cirdan> but 1.x.y is compat with 1.(x+i).(y+j)
10:30 < TheSin> because you can't make sure what a deb is compiled against then
10:30 < vasi> cirdan, if you build against 1.0.0 it still work agains 2.0.0
10:30 < TheSin> unless it take the current compiled version and adds >=
10:30 < cirdan> TheSin: builddep foo-dev foo=shlibs
10:30 < TheSin> so the builtime version is the lowest allowed
10:30 < cirdan> depends on foo-shlibs >= min package version with this abi or greater
10:31 < cirdan> vasi: no
10:31 < TheSin> cirdan, right
10:31 < TheSin> but that means droping how shlibs work presently
10:31 < TheSin> since hte shlibs file will be useless
10:31 < cirdan> majot version changes in the abi are *only* when compatability breaks
10:31 -!- Feanor [~astrange@mp1-249-176.dialup.emory.edu] has joined #fink
10:32 < cirdan> TheSin: not dripping, just modifing
10:32 < cirdan> dropping
10:32 < TheSin> and adding
10:32 < TheSin> not dropping
10:32 < TheSin> yes I know this
10:32 < vasi> cirdan, i just asked pogma a few screens back
10:33 < vasi> and he said the does work
10:33 < vasi> *that
10:33 < TheSin> it's the adding that will break things backwards
10:33 < cirdan> no, minor and micro versions only
10:33 < TheSin> anyhow $JOB bbiab
10:33 < cirdan> TheSin: right, which is why the shlibs field needs to be changed whenever the abi version changes
10:34 < vasi> cirdan, the first number in compat is NOT the library major version
10:34 < TheSin> that will happen
10:34 < TheSin> based on the shlibs files
10:34 -!- bbraun [~bbraun@bbraun.core.opendarwin] has left #fink []
10:34 < TheSin> since when you update your -shlibs file
10:34 < TheSin> it'll replace the shlibs
10:34 < TheSin> file
10:35 < TheSin> then it'll use that one
10:35 < TheSin> that is why proper shlibs lines are soo important
10:35 < cirdan> vasi: sure it is, but libtool make libbfoo's install name 1+major abi version
10:35 < TheSin> the way it is now
10:35 < cirdan> iirc
10:35 < TheSin> it'll work forwards
10:36 < TheSin> because the dep on a pre built dep will be blah-shlibs >= x.y.z
10:36 < TheSin> which is gets from blah-shlibs' .shlibs file
10:36 < TheSin> which will update to the new compat
10:36 < TheSin> and next build it'll update the dep line in the deb
10:36 < vasi> cirdan, no....if you have a library versioned upstream as X.Y.Z, you should get libfoo.X.dylib compat=(X+Y+1).Z.0
10:37 < vasi> the first number of compat increases with Y
10:37 < cirdan> hat doesnt sound right
10:38 < vasi> i'm pretty sure that's how it works
10:39 < cirdan> compat version is oildest it's compat with, and current is what it's currently at now
10:39 -!- dsias [~dsias@68-235-246-26.atlsfl.adelphia.net] has quit [Read error: 145 (Connection timed out)]
10:41 < vasi> right, there's always forwards compatibility with compat version
10:41 < cirdan> when something is linked against libfoo, it need a min of the current version of the lib
10:41 < vasi> *nods*....major version breaks forwards AND back....compat breaks only back
10:41 < cirdan> yes
10:41 < vasi> so if you compile against compat 1.0.0, you'll still accept compat 2.0.0
10:42 < vasi> but not vice versa
10:42 < vasi> but if you compile against libfoo.1.dylib, you won't accept libfoo.2.dylib
10:42 < cirdan> no, then yes
10:42 < pogma> this is why the shlibs field has a compat version in it
10:43 < pogma> and the install_name
10:44 < vasi> pogma, our debs do end up different though, eh? if i rebuild against foo with a higher compat, my pkg has exactly the same version and revision but there's a difference in what it will accept being runtime-linked to
10:44 < cirdan> yes
10:44 < cirdan> which is why we really needs autoshlibs :-)
10:44 < vasi> autoshlibs = ?
10:45 < cirdan> thesin's code for auto deps
10:45 < cirdan> it only does -=shlibs for now, iirc
10:45 < vasi> but it still gives us differnt debs
10:45 < vasi> which breaks in the long run
10:46 < cirdan> no, it doesnt break if the dep lines are correct
10:46 < cirdan> apt/dpkg know how to read the deps line and will make sure everything is ok
10:46 < pogma> vasi: yes, but as I noted, it is very unusual to update foo-shlibs, build stuff, then downgrade foo-shlibs
10:46 < vasi> yeah
10:46 < cirdan> pogma: unless you have an old install and apt-get new pkg
10:47 < vasi> cirdan, i'm worried about rpm-hell though
10:47 < cirdan> but if the deps are correct apt-get will also update foo-shlibs, w/o the user even noticing
10:47 < vasi> where the same deb that works fine on my system doesn't work on yous
10:47 < vasi> i gotta go
10:47 -!- vasi [~vasi@modemcable214.145-70-69.mc.videotron.ca] has quit []
10:48 < cirdan> well, this is why debian has a nice binary archive :-)
10:48 < cirdan> it's unavoidable w/o it though
10:58 < neo> hello
10:58 < neo> can you tell me how i can see if X11 is installed on my tiger
10:58 < newmanbe> Neo: Hi.
10:58 < newmanbe> Does the folder /usr/X11R6 exist?
10:59 < neo> yes
10:59 < neo> (so it is ... ;) )
10:59 < newmanbe> Then you probably have X11 installed.
10:59 < neo> ok
11:00 < newmanbe> You can start it by starting either X11 or XDarwin which are probably in /Applications/.
11:01 < neo> they aren't
11:02 < newmanbe> Then it looks like you have a broken installation.
11:02 < neo> :(
11:02 < newmanbe> Please rm -rf /etc/X11/ and /usr/X11R6/ and try again. :)
11:02 < cirdan> no
11:02 < cirdan> wait
11:02 < neo> i didn't install anything
11:02 < newmanbe> No?
11:03 < newmanbe> Wait?
11:03 < cirdan>  /Applications/Utilities/X11.app
11:03 < newmanbe> Oh yeah, I forgot Apple moved it.
11:03 < neo> cirdan: doesn't exist
11:03 < cirdan> it's always been there
11:03 < cirdan> iirc
11:03 < cirdan> then rm -rf and isntall xorg :-)
11:03 < newmanbe> I think it was in /Applications/ in Mac OS X 10.2.
11:05  * newmanbe hasn't used Apple's X11 for a while if you couldn't tell. :)
11:06 < neo> how can i install xorg ?
11:06 -!- jtyler [~jtyler@iphost-64-56-130-194.edm.wiband.net] has quit ["Leaving"]
11:06 < newmanbe> !xorg
11:06 -!- jtyler [~jtyler@iphost-64-56-130-194.edm.wiband.net] has joined #fink
11:07 < newmanbe> !x11
11:07 < Melian> extra, extra, read all about it, x11 is at http://www.finkproject.org/doc/x11/
11:07 < Albie> :-|
11:08 < newmanbe> Albie: What?
11:08 < Albie> nothing ;)
11:13 -!- miga [~miga@easyconnect2121138-139.clients.easynet.fr] has quit ["Bye, I'm leaving"]
11:15 -!- regeya [~shane@adsl-sp3-cdale176.micgi.com] has joined #fink
11:23 -!- chris01 [~chris01@212.126.165.246] has joined #fink
11:28 -!- ronwalf [~ronwalf@138.88.110.164] has quit []
11:30 < chris01> could somebody on 10.4 verify if perl581-core installs /sw/bin/perl5.8.1?
11:31 -!- mGiff [~mGiff@ottawa-hs-209-217-124-92.d-ip.magma.ca] has joined #fink
11:31 -!- gzl [~np@hole.macnn.com] has quit [Read error: 145 (Connection timed out)]
11:37 < akh_laptop> chris01: checking...
11:38 < akh_laptop> It looks like it's supposed to.
11:38 -!- gzl [~np@hole.macnn.com] has joined #fink
11:39 < chris01> akh_laptop: thanks for checking. What do you mean by "supposed" in this context?
11:41 < akh_laptop> I fetched it from the bindist but didn't install it (dpkg -c)
11:52 < Albie> 581? we need 587 :)
11:52 < akh_laptop> legacy support
11:53 < akh_laptop> At least it's not 560  ;-)
11:53 < Albie> :-D
11:57 -!- mdmonk [~mr_twitch@64.81.110.110] has joined #fink
12:00 -!- dsias [~dsias@adsl-068-153-207-210.sip.bct.bellsouth.net] has joined #fink
12:19 -!- jtyler [~jtyler@iphost-64-56-130-194.edm.wiband.net] has quit ["Leaving"]
12:22 < chris01> Albie: yes, i know, 581 is old. But somebody tried to install html-parser-pm581, and it didn't work.
12:22 -!- jtyler [~jtyler@iphost-64-56-130-194.edm.wiband.net] has joined #fink
12:22 < chris01> akh_laptop: would you be so kind and try to install html-parser-pm581?
12:34 < Albie> well.. need to go
12:34 -!- Albie [~ambs@eremita.di.uminho.pt] has quit ["I'm not here!"]
12:39 < akh_laptop> chris01: OK
12:42 < chris01> aha? Good. Thanks a lot!
12:42 < chris01> akh_laptop: could you past me what you get from 'fink list perl58' ?
12:42 < chris01> paste
12:43 < akh_laptop> chris01: just a sec--I'm still downloading html-parser-pm581
12:43 < akh_laptop> (installing now)
12:43 < chris01> akh_laptop: aha, oops, ok. :)
12:45 < lisppaste> akh pasted "perl" at http://paste.lisp.org/display/9251
12:45 < akh_laptop> There
12:46 < akh_laptop> (oops--forgot my nick)
12:46 -!- akh_laptop is now known as akh
12:48 -!- Spundun [~spundun@c-24-126-63-203.hsd1.ca.comcast.net] has quit ["Leaving"]
12:50 < chris01> akh: thanks.
12:53 -!- chris01 [~chris01@212.126.165.246] has quit ["bye"]
13:05 -!- att [~att@rrcs-24-123-50-242.central.biz.rr.com] has joined #fink
13:06 -!- ronwalf [~ronwalf@gandalf.mindlab.umd.edu] has joined #fink
13:20 -!- jtyler [~jtyler@iphost-64-56-130-194.edm.wiband.net] has quit ["Leaving"]
13:21 -!- Fang [~Fang@2002:56c1:e559:1:0:0:0:1] has quit ["Be incomprehensible.  If they can't understand, they can't disagree."]
13:28 -!- akh is now known as akh_afk
13:32 -!- shres [~sshreyas@202.144.86.147] has quit ["Gnite"]
13:40 < TheSin> !lisppaste
13:40 < Melian> i heard lisppaste is a bot that lets you post large chunks of text without flooding the channel. Use it at http://paste.lisp.org/new/fink
13:41 < lisppaste> TheSin pasted "Gnome/Pango Problem, CVS HEAD this morning and selfupdate this morning" at http://paste.lisp.org/display/9255
13:42 < cirdan> whee
13:42 < TheSin> thought I'd report that :D
13:43 < TheSin> looks like it needs a -ljpeg, not sure why though
13:51 < TheSin> man
13:51 < TheSin> hear is an other one
13:51 < lisppaste> TheSin pasted "gtk+2 bad patch!" at http://paste.lisp.org/display/9256
13:52 -!- newmanbe_ [~newmanbe@0be10570c2fd8ab9.session.tor] has joined #fink
13:54 < cirdan> dm
13:54 < cirdan> !punish errors
13:56 -!- Murr [~neeri@A17-202-20-71.apple.com] has joined #fink
13:57 < newmanbe_> !fix broken dishwasher
13:57 -!- newmanbe [~newmanbe@503ad105b351b07d.session.tor] has quit [Nick collision from services.]
13:57 -!- newmanbe_ is now known as newmanbe
13:57 -!- newmanbe [~newmanbe@0be10570c2fd8ab9.session.tor] has quit [Nick collision from services.]
13:58 -!- knghtbrd_ is now known as knghtbrd
13:58 -!- newmanbe [~newmanbe@bd8a5b68a3202e3e.session.tor] has joined #fink
14:00 -!- dsias [~dsias@adsl-068-153-207-210.sip.bct.bellsouth.net] has quit ["Leaving"]
14:09 -!- Leafw [~LeafWind@FW-171-241.go.retevision.es] has joined #fink
14:09 < Leafw> Hi. Why installing the gcc3.3.3 alone fails to provide a working gcc for macosx. Does anyone knows what else is needed ?
14:11 < Leafw> the macosx 10.3 SDK ?
14:12  * newmanbe doesn't get what you are asking.
14:12 < cirdan> all the headers and all?
14:13 < Leafw> I updated from 10.2.8 to 10.3.9
14:13 < Leafw> and well, everything got wiped out (as I wanted)
14:13 < Leafw> so now I installed xcode
14:13 < Leafw> which did not install any gcc, so I installed gcc3.3, and it fails to compile a simple hello world program
14:14 < Leafw> the linker complains: 
14:14 < newmanbe> Did you install the November 2004 Update?
14:14 < cirdan> xcode should have installed gcc
14:14 < cirdan> if not, you got problems
14:14 < Leafw> I thought so
14:14 < Leafw> I do
14:14 < Leafw> xD
14:14 < Leafw> man
14:14 < cirdan> you can't not install gcc
14:14 < cirdan> it forces it
14:15 < cirdan> gcc4 anyway
14:15 < cirdan> on 10.4
14:15 < cirdan> and gcc3 on 10.3
14:15 < Leafw> 10.3.9 here
14:15 < cirdan> /Developer.Tools/uninstall--devtoolspl
14:15 < cirdan> uninstall-devtools.pl
14:15 < Leafw> and then reinstall ?
14:16 < Leafw> what should be different ..
14:17  * newmanbe goes to investigate 100% CPU usage on this computer.
14:17 < cirdan> do a custom install
14:17 < cirdan> install gcc's
14:17 < cirdan> all
14:18 < Leafw> 'all' meaning what, sorry, all the packages in the installation disk, or what exaclty
14:20 < cirdan> all the avail gcc's
14:20 < Leafw> just to be sure: Xcode is the XcodeTools.mpgk ?
14:20 < cirdan> yeah
14:20 < Leafw> good
14:30 -!- KraMer [~mark@adsl-70-240-193-143.dsl.hstntx.swbell.net] has quit [Remote closed the connection]
14:33 -!- dreamind [~dreamind@C2107.campino.wh.tu-darmstadt.de] has joined #fink
14:34 < Leafw> now works
14:34 < Leafw> thanks for the directions
14:34 < Leafw> 1.2 Gb reinstalled ... what could have gone wrong in the first run, I wonder
14:39 < cirdan> sometimes the installer is still a little tarded
14:43 < Leafw> who knows. In any case, thanks. CU
14:43 -!- Leafw [~LeafWind@FW-171-241.go.retevision.es] has quit ["Leafw has no reason"]
14:47 -!- dmacks [~dmacks@netspace.org] has joined #fink
14:54 < cirdan> bbl
15:01 < dmacks> TheSin: Any idea what's up with that gtk+2.patch bug?
15:02  * newmanbe guesses it was stolen from DarwinPorts and blames that port's "maintainer".
15:02 < TheSin> didn't look into it, I'm gonna try and other rsync
15:02 < TheSin> maybe the patch and info file got out of sync some how
15:03 < dmacks> Nah, it's a cleanroom implementation. But I'd be happy to blame them anyway:)
15:04  * dmacks should just bump to 2.6.8, which doesn't need that #include at all.
15:04  * akh_afk dubs dmacks "gnome-core"
15:04 -!- akh_afk is now known as akh
15:06 < dmacks> Yeah, you'd *better* change your nick so I can't find you after saying something like that!
15:06 < akh> heh
15:06 < TheSin> hehe
15:06 < akh> kidding, of course.
15:07 < akh> Need someone who's not here to object, of course. ;-)
15:07 < akh> hmmm....
15:08 < TheSin> that file doesn't exist dmacks 
15:08 < TheSin> justin@Diode [/sw/src/fink.build/gtk+2-2.4.9-9/gtk+-2.4.9/gtk/xdgmime]$ ls xdgmime*
15:08 < TheSin> xdgmime.c  xdgmimeglob.c  xdgmimeint.c  xdgmimemagic.c
15:08 < TheSin> xdgmime.h  xdgmimeglob.h  xdgmimeint.h  xdgmimemagic.h
15:08 < TheSin> no xdgmimecache.c
15:08 < dmacks> gtk+-2.4.9? Your .info isn't updating
15:08 < TheSin> 2.4.9 yup
15:09 < dmacks> unstable is at 2.6.7
15:09 < TheSin> the patch file says 2.6.7 though
15:09 < TheSin> see look like the info and patch got up of sync
15:11 < dmacks> It's correct in sf cvsweb (2.6.7-2)
15:11 -!- akh is now known as akh_powerbook
15:11 -!- dmacks is now known as dmacks_away
15:12 < TheSin> just did an other rsync
15:12 < TheSin> still didn't get it
15:12 < TheSin> rsync://master.us.finkmirrors.net/finkinfo
15:12 < TheSin> forcing an other mirror
15:12 < TheSin> great
15:12 < TheSin> rsync -az -q  rsync://dkx.tn.us.finkmirrors.net/finkinfo//TIMESTAMP /sw/fink/TIMESTAMP.tmp
15:12 < TheSin> The timestamp of the server is older than what you already have.
15:13  * TheSin considers going back to CVS *shivers*
15:14 -!- KraMer [~mark@adsl-70-240-193-143.dsl.hstntx.swbell.net] has joined #fink
15:15 -!- akh_linux_box [~akhansen@68-118-244-23.dhcp.oxfr.ma.charter.com] has joined #fink
15:15 < TheSin> got an other mirror dmacks
15:15 < TheSin> still 2,4,9
15:15 < TheSin> rsync://msn.wi.us.finkmirrors.net/finkinfo
15:15 -!- akh_powerbook [~akhansen@68-118-244-23.dhcp.oxfr.ma.charter.com] has quit []
15:17 < TheSin> rsync://bru.be.eu.finkmirrors.net/finkinfo
15:18 < TheSin> 2.4.9
15:18 < TheSin> lets blame sf
15:20 -!- Albie [~ambs@bl5-165-189.dsl.telepac.pt] has joined #fink
15:21  * akh_linux_box is shocked that you'd blame sf.
15:21 < akh_linux_box> :-)
15:21 < newmanbe> !lart SourceForge.net
15:21  * Melian throws SourceForge.net's poor little doggy off a cliff
15:22 < TheSin> i learnt it from you akh_linux_box
15:22 < TheSin> :P
15:22 < newmanbe> TheSin: Took you long enough to blame them.
15:22 < akh_linux_box> heh
15:22 < TheSin> I was countinh on you newmanbe to bring us to light :P
15:28 < akh_linux_box> grrr...can't log in to my box at work.  Stupid TIger power management.
15:31 < newmanbe> No; stupid U.S. Government and it's Energy Star program.
15:32  * akh_linux_box isn't 100% sure, though--I could log in earlier today.
15:35 < akh_linux_box> Last thing I tried to do was update mozilla via Fink.
15:35 < akh_linux_box> Fink crashed my computer!
15:35 < akh_linux_box> ;-)
15:37 < dmacks_away> Fink killed my father on two separate occasions!
15:39 < akh_linux_box> Fink is driving Apple away from PowerPC architctecture!!
15:41 < akh_linux_box> (had to increase the implausibility)
15:42 < das_> Fink raped my dog :(
15:42 < das_> Oh, no, wait, that was just this: http://www.fox21.com/Global/story.asp?S=3456745
15:43 < cirdan> das_: www.fu-fme.com
15:43 < dmacks_away> Why is it "like a man would do to a woman" and not "like a male dog would do to Princess"?
15:43 < cirdan> think we can get a darwin driver?
15:43 < Albie> !seen dmacks
15:43 < Melian> dmacks <~dmacks@dmacks.active.supporter.pdpc> was last seen on IRC in channel #fink, 32m 31s ago, saying: 'It's correct in sf cvsweb (2.6.7-2)'.
15:44 < Albie> erm
15:44 < cirdan> dmacks_not_so_away :-)
15:44 -!- dmacks_away is now known as dmacks_really_aw
15:44 < dmacks_really_aw> crap
15:44 < Albie> lol
15:44 -!- dmacks_really_aw is now known as dmacksReallyAway
15:45 -!- dmacksReallyAway is now known as dmacks
15:45 < akh_linux_box> So you were lying about being ReallyAway?
15:45 < dmacks> (it was pissing off nickserv)
15:46 < dmacks> Are you really a linux box?
15:46 < dmacks> Whatcha want, Albie?
15:47 < Albie> dmacks, to ask what involved to be a package maintainer :)
15:47 < akh_linux_box> Blood, sweat, and tears.  ;-)
15:47 < cirdan> a pulse
15:47 < Albie> s/involved/involves/
15:47 < dmacks> Mostly just being a contact person to provide updated .info/.patch files as needed.
15:47 < cirdan> some things need major work
15:48 < cirdan> but very few
15:48 < cirdan> fix k3b :-)
15:48 < Albie> I might help with Perl modules
15:48 < Albie> if there is a need for them O:-)
15:48 < dmacks> cirdan: This is a followup to a specific -submissions item, which needs I think no work except a Version/Revision bump.
15:48 < Albie> dmacks, right
15:49 < dmacks> But yeah, lots of the -pm are unmaintained:(
15:49 < cirdan> oh
15:49 < cirdan> grep -r None /sw/fink/dists/*/finkinfo | grep devel
15:49 < cirdan> :-)
15:49 < dmacks> http://fink.sourceforge.net/pdb/nomaintainer.php
15:49 < dmacks> :p
15:50 < cirdan> mine's more accurate
15:50 < cirdan> :-p
15:50 < Albie> heh
15:50 < dmacks> Yeah. I update the sf db myself and I still don't trust it:/
15:50 < cirdan> hehe
15:52 < dmacks> Albie: So if you want to claim that pkg, attach a new .info (not a diff against the old pls) with yourself as Maintainer, and we'll get it updated.
15:52 < Albie> dmacks, ok.
15:53  * dmacks really going away now.
15:53 -!- dmacks is now known as dmacks_away
15:55 < Albie> erm... why the stable .info file has a maintainer and the unstable one does not?
15:56 < cirdan> older in stable
15:56 < Albie> k
15:56 < akh_linux_box> And it was inadvertenly not de-maintained.
16:02 -!- Murr [~neeri@A17-202-20-71.apple.com] has quit [Read error: 145 (Connection timed out)]
16:22 -!- ambs_ [~ambs@bl6-36-60.dsl.telepac.pt] has joined #fink
16:30 -!- Albie [~ambs@bl5-165-189.dsl.telepac.pt] has quit [Read error: 145 (Connection timed out)]
16:31 -!- ambs_ is now known as Albie
16:43 -!- Murr [~neeri@A17-202-20-71.apple.com] has joined #fink
16:53 -!- att [~att@rrcs-24-123-50-242.central.biz.rr.com] has quit []
16:53 -!- akh_linux_box [~akhansen@68-118-244-23.dhcp.oxfr.ma.charter.com] has quit [Read error: 104 (Connection reset by peer)]
17:07 -!- stephano [~stephano@wb9.powerschool.com] has joined #fink
17:11 -!- mdmonk [~mr_twitch@64.81.110.110] has left #fink []
17:13 -!- stephano [~stephano@wb9.powerschool.com] has quit ["Leaving"]
17:17 -!- Albie [~ambs@bl6-36-60.dsl.telepac.pt] has quit ["Leaving"]
17:47 -!- baba [~baba@YahooBB220041001026.bbtec.net] has quit ["Leaving"]
17:55 -!- Leafw [~LeafWind@FW-171-241.go.retevision.es] has joined #fink
17:56 < Leafw> Hi. What is the current compatibility of apple's X11 with other windowmanagers (not quartz-wm) like windowmaker ?
17:56 -!- You're now known as RangerAway
17:58 < newmanbe> They should all work if you get them to compile.
17:58 < TheSin> Leafw, shoudl be fine
17:58 < Leafw> ok, sounds good
17:58 < Leafw> through fink, I guess ... manually compiling wmx didn't work
17:58 < Leafw> claimed X11 is not there
17:58 < Leafw> I bet the headers may not be in th expected places
18:04 < TheSin> did you install the X11SDK from the xcode cd
18:04 < TheSin> you need to choose custom install
18:04 < TheSin> they don't install with the basic install
18:04 < Leafw> ok, then I'll put it in manually
18:04 < Leafw> thank you
18:04 < Leafw> I thought the 1.2 GB included all
18:04 < TheSin> if you picked custom then selected all it should be there
18:05 < TheSin> but if you didn't pick custom it wouldn't be installed
18:05 < Leafw> didn't do that unfortunatelly
18:07 -!- Murr [~neeri@A17-202-20-71.apple.com] has quit [Read error: 131 (Connection reset by peer)]
18:08 -!- Murr [~neeri@A17-202-20-71.apple.com] has joined #fink
18:08 < Leafw> I'll be back.
18:08 -!- Leafw [~LeafWind@FW-171-241.go.retevision.es] has quit ["Leafw has no reason"]
18:11 < TheSin> I might be back :D
18:11 -!- TheSin [~TheSin@iphost-64-56-130-194.edm.wiband.net] has quit ["Client exiting"]
18:15 -!- leafw [~albertca@FW-171-241.go.retevision.es] has joined #fink
18:16 < leafw> more questions ... I can't paste text from aqua apps to Apple's X11 windows. Is there a need for a special setup?
18:17 < leafw> hum it may be control+v and not cmd+v ..
18:20 -!- theid_away [~theid@207.177.103.77] has quit [Read error: 110 (Connection timed out)]
18:24 < leafw> oh man, 'option' and 'cmd' to emulate middle and right. Well, sounds easy ..
18:29 < dmacks_away> leafw: Investigate quartz-wm "--only-proxy" option, the 'pbcopy' and 'pbpaste' commands, perhaps the 'autocutsel' package. In X11, it's still cmd-keys, but if you've got a 1-button mouse and use cmd for emulation...heh:) Disable cmd-keys in X11 prefs and you can use the pull-down menu items.
18:29 < leafw> dmacks_away : thanks. But: which pulldown menu items, you mean the standard apple top menu ?
18:30 < dmacks_away> Yup.
18:30 < leafw> ok
18:30 < leafw> I like keys over mouse
18:30 < leafw> I see that option-click is middle-click and will paste in most X11 apps, specially in xterms
18:30 < dmacks_away> Google for [x11 cut paste apple] or somesuch...you'll get *lots* of useful hints.
18:30 < leafw> which is what I need
18:30 < leafw> yes, I'm realizing that, thanks :)
18:31 < leafw> http://developer.apple.com/qa/qa2001/qa1232.html
18:31 < leafw> this page specially useful
18:31 < dmacks_away> If you use quartz-wm, I think the x11 pasteboard and the aqua clipboard are merged, so you just gotta use "whatever is the correct" way to cut'n'paste for whichever environment you're using at the moment.
18:31 -!- ronwalf [~ronwalf@gandalf.mindlab.umd.edu] has quit []
18:31 < leafw> right
18:32 < leafw> the pbcopy and pbpaste commands, though, may be used as cmd+something commands though right ?
18:32 < dmacks_away> Those are unix (command-line) things.
18:32 < leafw> I am a freebsd user struggling in this mac box
18:33 < leafw> I understand ... 
18:33  * dmacks_away wasn't sure exactly what you were doing, didn't have time to read the scrollback, am off to a meeting now:)
18:33 < leafw> no prob
18:38 -!- lisppaste [~lisppaste@common-lisp.net] has quit [Read error: 104 (Connection reset by peer)]
18:41 < leafw> hummm mwindowmaker crashes apple's X11
18:42 < leafw> too bad, that's the one I wanted
18:42 < newmanbe> !xorg
18:42 < leafw> actually, what I really wanted was the ability to shade xterms in this X11 (apple's), is this possible .
18:51 -!- lisppaste [~lisppaste@common-lisp.net] has joined #fink
18:53 -!- theid_away [~theid@207.177.103.77] has joined #fink
19:03 -!- leafw [~albertca@FW-171-241.go.retevision.es] has quit ["leafw has no reason"]
19:07 -!- Murr [~neeri@A17-202-20-71.apple.com] has quit [Read error: 131 (Connection reset by peer)]
19:08 -!- Murr [~neeri@A17-202-20-71.apple.com] has joined #fink
19:33 -!- vasi [~vasi@modemcable214.145-70-69.mc.videotron.ca] has joined #fink
19:46 -!- spundun [~spundun@mermaid.isi.edu] has joined #fink
19:46 < spundun> hmmm....
19:46 < spundun> vim doest use gtk....
19:46 < spundun> on fink/tiger
19:46 < spundun> I wonder why, gtk would be better
19:47 < Feanor> use textwrangler
19:47 < spundun> Feanor: :)... does it work with X?
19:48 < Feanor> oh, if you want X11 forwarding it won't work
19:48 < Feanor> but it's definitely a better text editor
19:48 < spundun> yeah I have heard good stuff about it..
19:49 < spundun> Btw, the crypto versions of gnome packages are crashing... some symbol not found in libcrypto.dylib or somethin
19:49 < spundun> 3 lines incoming... 
19:49 < spundun> dyld: lazy symbol binding failed: Symbol not found: _EVP_idea_cbc
19:49 < spundun>   Referenced from: /sw/lib/libssl.0.9.7.dylib
19:49 < spundun>   Expected in: /usr/lib/libcrypto.0.9.7.dylib
19:54 < vasi> known bug
19:54 -!- Murr [~neeri@A17-202-20-71.apple.com] has quit [Read error: 145 (Connection timed out)]
19:55 < vasi> which i wish i could fix, but i still can't reproduce it on my box
19:55 < ralfWORK> erm ok
19:55 < ralfWORK> should I move to unstable?
19:56 < ralfWORK> seems someone was saying that unstable was actually fairly stable
19:56 < vasi> yes it is
19:57 < vasi> if you need something from unstable, move to unstable :-)
19:57 < vasi> otherwise, don't
19:57 < vasi> pretty simple
19:57 < ralfWORK> what's the actual unstable branch called? :P
19:58 -!- akh_la [~akhansen@68-118-244-23.dhcp.oxfr.ma.charter.com] has joined #fink
19:58 -!- akh_la is now known as akh
19:58 < akh> stupid IRC client
20:02 < vasi> ralfWORK, just unstable
20:02 < vasi> yeah, Debian and Ubuntu got names, we're not that cool
20:02 < vasi> if you want to make one up for us, go ahead :-)
20:03 < ralfWORK> hahahah
20:03 < ralfWORK> so there is "stable" and "unstable"
20:03 < ralfWORK> and thats it?
20:03 -!- ottoaim__ [~ottoaim@h-66-167-223-180.chcgilgm.dynamic.covad.net] has joined #fink
20:04 < vasi> well there's also experimental
20:04 < vasi> but it's not really a dist
20:04 < vasi> it's just a collection of things that various fink people are working on
20:04 < ralfWORK> ohh right I think experimental was what I was thinking about
20:04 < ralfWORK> ok gotcha
20:04 < vasi> so you can't really "enable experimental"
20:04 < ralfWORK> ok
20:05 -!- Murr [~neeri@A17-202-20-71.apple.com] has joined #fink
20:07 < akh> You wouldn't want to "enable experimental" anyway.  
20:07 -!- ottoaim [~ottoaim@h-64-105-104-202.chcgilgm.dynamic.covad.net] has quit [Read error: 145 (Connection timed out)]
20:08 < ralfWORK> heh
20:08 < ralfWORK> and so I shall not
20:08 < ralfWORK> is there a list of packages on the site?
20:09 < akh> For the regular trees:  http://fink.sourceforge.net/pdb/list.php
20:11 < newmanbe> akh: Forget regular trees
20:11 < akh> newmanbe:  I've forgotten regular trees.
20:11 < newmanbe> akh: regular trees is http://www.finkproject.org/pdb/list.php
20:11 < newmanbe> :-p
20:12  * akh can't commit to that until it's official.
20:12 < spundun> vasi: ok, thanks for responding...
20:12 < akh> (yeah, I'm inconsistent)
20:13  * akh is a creature of habit.
20:13 < vasi> grr, don't use the PDB
20:13 < vasi> it's consistently borked
20:13 < spundun> vasi: I will tell youexactly what I did, install tiger, install Dev tools, install fink, install fink commander, enable unstable, fink install gnome-desktop, fink install bundle-gnome .. thats all..
20:13 < newmanbe> vasi: Hmm, you seemed motivated enough to move it to sancho.
20:13 < spundun> ;)
20:13 < akh> It looks OK right this minute--who knows what'll happen in an hour though.
20:14 < vasi> newmanbe, i'm unfortunately not a Web Person
20:14 < newmanbe> Well that just won't do. :(
20:14 < vasi> Spundun, what's the trouble again?
20:16 -!- pell [~pellican@artemis.gaelicwizard.net] has joined #fink
20:18 < vasi> i'm going to take over librsvg2 from gnome-core, if anyone objects, speak now or forever hold your peace :-)
20:19 -!- theid_away is now known as theid
20:19 < akh> Spundun:  use 'env DYLD_FALLBACK_LIBRARY_PATH=: <application_name>'
20:21 -!- pell [~pellican@artemis.gaelicwizard.net] has quit []
20:22 < akh> (don't include the ', of course)
20:22 -!- Murr [~neeri@A17-202-20-71.apple.com] has quit [Read error: 131 (Connection reset by peer)]
20:23 -!- Murr [~neeri@A17-202-20-71.apple.com] has joined #fink
20:27 -!- nahamu [~nshalman@tiresias.cs.brandeis.edu] has joined #fink
20:38 < akh> vasi:  I guess librsvg2 is yours then.
20:43 < cirdan> we really should add to our official instructions to uninstall any devtools when upgrading
20:46 < akh> I can go for that (but it won't be tonight)
21:05 < vasi> uninstall devtools? why?
21:05 < cirdan> old devtolls can cause problems
21:06 < cirdan> bad headers are leftover
21:06 < cirdan> been happening on some of the apple lists
21:06 -!- muesli [~muesli@mail.muehlhaeuser.de] has quit [Read error: 104 (Connection reset by peer)]
21:06 < vasi> oops
21:06 < cirdan> only needed for an upgrade, not an archive & install/format
21:06 < cirdan> yeah, subtle problems
21:06 < cirdan> like gcc3.1 hanging around...
21:06 < cirdan> even 2.95
21:07 -!- muesli [~muesli@mail.muehlhaeuser.de] has joined #fink
21:07 < vasi> man gcc has had some version number inflation lately
21:07 < vasi> it took like 10 years to go from 2 to 2.95
21:07 < cirdan> grr, took me like 2-3 hrs to d/l all the sources i need to upgrade
21:07 < vasi> and now it's at 4
21:07 < cirdan> oh, dist-uograde works :-)
21:08 < vasi> uh, which version of it?
21:08 < cirdan> vasi: well, getting closers to standards and specs too :-)
21:08 < cirdan> umm
21:08 -!- newmanbe [~newmanbe@bd8a5b68a3202e3e.session.tor] has quit ["Leaving"]
21:08 < cirdan> the one i'm running
21:08 < cirdan> :-)
21:08 < cirdan> i made a few logic oopses
21:08 < vasi> lemme rephrase...what does the current version of dist-up do?
21:09 -!- dreamind [~dreamind@C2107.campino.wh.tu-darmstadt.de] has quit []
21:09 < vasi> just block compiles until the user runs dist-up, then when they run dist-up do symlink munging?
21:09 < cirdan> doesn't let users on mixed systems build, doen't auto change the dists line, and changes the setting for them
21:09 < cirdan> yes
21:09 < vasi> ok, not the whole dummy-packages thing
21:09 < cirdan> also allows puredarwin users to do it too
21:09 < cirdan> no
21:10 < vasi> cuz i still think we need much more discussion about that
21:10 < cirdan> i wanna add questions about selfupdate and maybe even cleaning the devtools too
21:10 < cirdan> right
21:10 < vasi> ok, that's doable
21:10 < vasi> can you post what you have to the list, and explain it for everybody, we'll get some testing
21:11 < cirdan> i was gonna, but didnt yet cause i wasn't sure which functs to use :-)
21:11 < vasi> and make sure that everybody's ok for it
21:11 < cirdan> for user input
21:11 < vasi> CLI.pm
21:11 < vasi> prompt_boolean is what you want i think
21:12 < cirdan> i'm not sure how to wrap a selfupdate w/o it killing the current fink
21:12 < cirdan> fork maybe?
21:12 < vasi> you want to wrap a selfupdate?
21:12 < vasi> don't do that
21:12 < vasi> post-selfuupdate, we always want to use the new fink
21:12 < vasi> use that update_finish method, or soemthing like that
21:12  * vasi never remembers the precise naming
21:13 < cirdan> i wanna ask the user if they wanna do a selfupdate, if they say yes, do it, if new fink build new fink and then exit, telling the user that there is a new fink and to run dist-upgrade again
21:18 -!- mGiff [~mGiff@ottawa-hs-209-217-124-92.d-ip.magma.ca] has quit [Client Quit]
21:20 -!- Murr [~neeri@A17-202-20-71.apple.com] has quit [Read error: 145 (Connection timed out)]
21:24 -!- shres [~sshreyas@202.144.86.147] has joined #fink
21:25 < vasi> hmm
21:26 < vasi> i think the right way to do that is a modified selfupdate
21:27 < vasi> where you set some config flag somewhere...and then selfupdate 1) Doesn't update anything but fink and 2) when it re-execs and runs 'fink selfupdate-finish' re-starts the dist-upgrade
21:27 < vasi> rather than wrapping it in another fink
21:27 < vasi> but dmacks would be better to ask there, in fact probably max would be better if he ever shows up
21:32 -!- __jt__ [~james@69-162-30-40.stcgpa.adelphia.net] has joined #fink
21:32 -!- akh [~akhansen@68-118-244-23.dhcp.oxfr.ma.charter.com] has quit []
21:33 < cirdan> i modified fink already to only builds fink on selfupdate if dist-upgrade needs to be done
21:33 < cirdan> vasi: my main concern is when selfupdate fails
21:34 < vasi> if it fails, then in should tell the user to re-run distup i guess....but you can do that in SelfUpdate.pm
21:35 < vasi> or in fact, even in Engine.pm....look at the stuff we do (in process i think?) about detecting failure so we can notify
21:37 -!- shres [~sshreyas@202.144.86.147] has left #fink ["away"]
21:38 -!- ronwalf [~ronwalf@138.88.110.164] has joined #fink
21:39 -!- ronwalf [~ronwalf@138.88.110.164] has left #fink []
21:52 -!- vasi is now known as TVasi
22:06 -!- shres [~sshreyas@202.144.86.147] has joined #fink
22:10 < spundun> TVasi: thanks
22:11 -!- spundun [~spundun@mermaid.isi.edu] has quit ["Leaving"]
22:17 < dmacks_away> This is great! I can push updates to the gnome pkgs and then set (okay "leave") gnoem-core on auto-ignore and let others deal with it:)
22:17 < dmacks_away> 1) be a jackass, 2) ???, 3) profit!
22:19 -!- dmacks_away is now known as dmacks
22:21 < TVasi> :-)
22:24 < dmacks> Anyone know a supplier for lithium cyanide?
22:33 < TVasi> i've always preferred KCN myself
22:33 < TVasi> ;-)
22:33 < dmacks> Already tried it, didn't work. (go ahead, make your "we know...you're still here dan" joke)
22:48 -!- xhrl [~ThomasW@24.80.39.250] has joined #fink
22:50 < TVasi> dmacks, any objection if i take over librsvg2 from gnome-core?
22:51 < dmacks> Take it!
22:53 < dmacks> Whoop, gnome just released new %v of lots of stuff:/
22:53 -!- miga [~miga@easyconnect2121138-139.clients.easynet.fr] has joined #fink
22:55 < miga> Hi everybody.
22:55 < dmacks> hello miga
22:55 < miga> dmacks: hello.
22:55 < TVasi> hi miga
22:56 < miga> hello TVasi.
22:56 < TVasi> er yeah, show's over
22:56 -!- TVasi is now known as vasi
22:56 < vasi> i always forget to change my nice
22:56 < vasi> nick
22:56 < miga> :-)
22:59 < miga> gtk 2.6.8 has been released, it is backward compatible with 2.6.5; and it fixes the issue we had with uintxx_t, but I'm not sure this the right fix. Could you have a look at it? Now that I've been approved as gnome translator, I will make me a pleasure to fill as many bugs as you want :-).
22:59 -!- __jt__ [~james@69-162-30-40.stcgpa.adelphia.net] has quit [Remote closed the connection]
22:59 < miga> Here is the bug: < http://bugzilla.gnome.org/show_bug.cgi?id=304924>
23:00 < vasi> miga, the fix looks good
23:01 < dmacks> miga: The 2.6.8 patch is certainly a viable fix.
23:01 < vasi> has it been tested?
23:01 < dmacks> Our pkg is already patched (not the same patch though) no?
23:01 < vasi> yeah, but it's usually preferable to bump the version rather than patching an old version
23:02 < dmacks> Right. That's why I wrote: < dmacks> Whoop, gnome just released new %v of lots of stuff:/
23:02 < vasi> (tbh, i think our fix is better though)
23:02 < miga> but, it makes not the gnome team aware that on Tiger, whenever you use uintxx_t, you should make something after having defined the HAVE_STDINT stuff. What do you think if I fill in a general bug on it?
23:02 < dmacks> (yeah:)
23:02 < miga> They have also released gtk2.7.
23:03 < vasi> ooh, new beagle
23:03 < miga> but it is unstable right now.
23:03 < miga> No, our patch is not the same, it's better as far as I'm concerned.
23:03 < vasi> yeah, probably best for us to refrain from package odd-numbered releases
23:03 < dmacks> Their servers are *swamped* right now. I'm getting <14K/sec on the new tarballs.
23:03 < cirdan> heh
23:04 < vasi> miga, which do you think is better?
23:04 < miga> I think our patch is better, since it is more general and uses the D_HAVEXXX mechanism completely.
23:05 < miga> I mean that kind of funny to test for a header, define a D_HAVE and not do anything with it.
23:05 < vasi> heh, well i think it's just standard auto* to test for stdint
23:05 < dmacks> Are the uintXX_t a newer standard or an older one than C99?
23:06 < vasi> i'm not sure whether they are actually standard
23:06 < miga> it is the newest I think. Should be checked. I mean don't trust me too much here.
23:06 < vasi> just convention, possiibly
23:06 < miga> they are standard.
23:06 < miga> sure.
23:06 < vasi> oh sorry, i mean they ARE C99 i think
23:07 < vasi> ie: stdint.h is a req of c99
23:08 -!- spundun [~spundun@c-24-126-63-203.hsd1.ca.comcast.net] has joined #fink
23:08 -!- eno-away is now known as eno
23:08 < miga> I have only the c++ standard by me, so I cannot check.
23:08 < dmacks> Okay, so uintXX_t is C99, which apparently isn't available everywhere.
23:08 < vasi> just google C99 uint_16t
23:08 < vasi> er, uint16_t
23:09 < vasi> and you'll find a bunch of folks claiming it's part of c99
23:09 < vasi> well it *is* available on OS X, you just need to include stdint
23:09 < miga> And moreover, it ensures that whichever is the architecture it is the good uintxxx which is defined (see future).
23:10 < vasi> http://www.nongnu.org/avr-libc/user-manual/group__avr__stdint.html
23:10 < lisppaste> Spundun pasted "Out of nowhere now I am getting connection refused errors on startx..." at http://paste.lisp.org/display/9267
23:11 < spundun> can someone checkout the above error? previously I was getting some other error... I dont know what happened... but it looks link the thing is breaking earlier now.
23:12 < dmacks> Right, but if it [sometimes? often?] requires a nonstandard #include (for example, platforms that have to be told "use C99 standard #includes" like we seem to have here), gotta either handle that situation or not use advanced standards.
23:13 < vasi> Spundun, i think you need 'startx -- -quartz' or something?
23:13 < miga> They can check the uname, no? darwinxxx
23:13 < spundun> vasi: why? I want to run a gnome-session ...
23:13 < vasi> Spundun, i think the quartz driver is still necessary
23:13 < dmacks> miga: As a rule, one should check for the actual features one wants, not for a metafeature like "platform" and try to guess what that platform has.
23:14 < spundun> And I was not getting this error before..
23:14 < vasi> if you want to be logged into OS X GUI also
23:14 < vasi> Spundun, alias startx='startx -- -quartz' :-)
23:15 < dmacks> (for example, there are some users who have a broken uname, but otherwise a completely normal OS X system:)
23:15 < spundun> vasi: that didnt help either... :(
23:15 < vasi> hmm, Spundun, lemme check
23:16 < lisppaste> Spundun annotated #9267 with "even with -- -quartz it doesnt work." at http://paste.lisp.org/display/9267#1
23:16 < spundun> vasi: I think the problem is something else.
23:16 < vasi> ooh i think it's a permissions issue
23:17 < vasi> Spundun, what does this give? ls -ld ~/.X*
23:17 < spundun> vasi: oooh... between my previous tries and this one I enabled ssh though sys prefs... and enabled X11 forwarding through the sshd_config file... 
23:18 < lisppaste> miga annotated #9267 with "xinitrc" at http://paste.lisp.org/display/9267#2
23:18 < miga> spundun: make a .xinitrc in your home directory and past in it what I'll paste in lisppaste, then double click on xdarwin, xorg, x11, whatever.
23:18 < spundun> vasi: mermaid:~ spundun$ ls -ld ~/.X*
23:18 < spundun> -rw-------   1 spundun  spundun  122 Jun 20 18:34 /Users/spundun/.Xauthority
23:18 < spundun> mermaid:~ spundun$ 
23:18 < spundun> vasi: lemme remove that file..
23:18 < dmacks> spundun: The error message tells you that's wrong:)
23:19 < spundun> vasi: cool.. removing that file... I got back to my previous error.. about dyld: Symbol not found: _EVP_idea_cbc
23:20 < lisppaste> miga annotated #9267 with "profile" at http://paste.lisp.org/display/9267#3
23:20 < miga> then in your .profile or whatever, paste this.
23:20 < miga> an normally it should work.
23:20 < vasi> miga, please don't export that
23:21 < miga> what should not I export?
23:21 < vasi> do instead:   DYLD_FALLBACK_LIBRARY_PATH=: exec /sw/bin/gnome-session
23:21 < spundun> vasi: yeah.. that worked :)(
23:21 < spundun> I forgot about that.. got distracted by this new error..
23:21 < vasi> miga, it might affect something else the user decided to link
23:21 < miga> Otherwise the gnome-session will not work if the user has installed ssl libs.
23:22 < spundun> vasi: I think its gonna work now... 
23:22 < vasi> miga, i know, but don';t export it...just use it for the required command
23:22 < spundun> vasi: Its running slowly because I am also brning a DVD through iDVD..
23:22 < vasi> miga: it's safer generally, though in this case it won't matter unless the user modifies his xinitrc later
23:22 < spundun> dmacks: vasi: miga: thanks for the help
23:23 < vasi> np Spundun
23:23 < miga> Not sure if it will work with gtk app. No.
23:23 < vasi> miga, if what will work?
23:23 < dmacks> You're welcome, spundun.
23:24 < miga> I mean if you only have gnome-session with the fallback, then you want to launch say mozilla, I'm not sure mozilla start.
23:24 < vasi> miga, i think it works, because things launched by gnome-session should inherit gnome-session's env
23:24 < vasi> but i haven't tested
23:24 < vasi> so don't take my word for it
23:25 < vasi> anyway, the proper fix is for us to auto-build wrappers for affected apps....if i could actually reproduce the problem on my system i'd try that
23:27 -!- shres [~sshreyas@202.144.86.147] has quit ["Gnite"]
23:27 < dmacks> The proper fix is for $SOME_GUY at apple.com to fix the damn linker.
23:27 < cirdan> right
23:27 < miga> Yes, it works, but make the loading time much more long.
23:28 < vasi> that's true :-)
23:28 < vasi> dmacks, is there any good reason why DocFiles isn't inherited by SplitOffs
23:28 < vasi> ?
23:28 < miga> Yes, but meanwhile I will not certainly wait on an hypothetic fix/
23:29 < dmacks> vasi: That field was intended for actual documentation, not just copyright/changelog. No reason to stuff 5meg of API docs into the -shlibs.
23:30 < vasi> but actual documentation is usually installed by make install anyhow
23:30 < vasi> and the SplitOff could still always override if it's inherited
23:30 < dmacks> I'm sayin' the reason and rationalizing, not saying I agree or that there mightn't be better ways:)
23:31 < vasi> ah
23:31 < vasi> cuz i'm getting annoyed at having to add the same line to every splitoff
23:31 < cirdan> before Files:
23:31 < spundun> what is this DYLD_FALLBACK_PATH problem?
23:32 < vasi> Spundun, apple made a boo-boo
23:32 < vasi> if you really want the long explanation, i can do that
23:32 < vasi> but it's probably not gonna help you much
23:33 < vasi> cuz it just comes down to "apple screwed up, nothing we can really do, hope they fix it"
23:33 < dmacks> Inheritting does make some sense, especially since we inherit License, which attaches fink-policy requiremwnts to DocFiles
23:34 < dmacks> (It's a common submissions-tracker mistake too)
23:34 -!- shres [~sshreyas@202.144.86.147] has joined #fink
23:34 < spundun> vasi: I would like to know more about it, if you dont mind.
23:35 < dmacks> (OTOH, there are cases where InstallScript installs the doc files, so one still needs DocFiles in SplitOffs).
23:35  * dmacks stops rambling
23:35 < vasi> Spundun: when dyld (the dynamic linker) loads a program, it's supposed to look at the things it says it links to (otool -L), and load them
23:36 < vasi> and then if it can't find one in the path that otool -L gives, it should look in the FALLBACK path
23:36 < vasi> which defaults to /usr/lib
23:36 < vasi> in Tiger, it doesn't do things the way it says it should...it's looking in the fallback path *first*
23:36 < vasi> before looking for the install names that otool gives
23:37 < spundun> vasi: hmmmm...
23:37 < spundun> ok, thats definately a skrewup
23:37 < vasi> so when a fink package links to /sw/lib/libssl, dyld looks in fallback and loads /usr/lib/libssl instead
23:37 < vasi> which is a different version and has different symbols, so *boom*
23:38 < vasi> dmacks, couldn't we give /sw/lib/libssl a different install_name?
23:38 < spundun> vasi: This is going to hurt a *LOT* of apps I have a feeling.
23:39 < dmacks> vasi: I proposed that in jest the other night--In Jest! I say--and there was already a lynch mob forming.
23:39 < vasi> dmacks, well it wouldn't hurt anything if done right...i think :-)
23:39 < pogma> msachs was meant to poke nick on the bug
23:39 < dmacks> Yeah, just gotta recompile every file that links to it.
23:40 < vasi> dmacks, nope...you still keep the current libssl
23:40 < dmacks> ...and pretend nobody in the world links to fink libs with manually-compiled stuff.
23:40 < vasi> just treat it as a new major version
23:40 -!- ringerc [~craig@dsl-202-72-144-62.wa.westnet.com.au] has joined #fink
23:41 < vasi> that way old stuff still works (modulo the dyld bug)
23:42 < dmacks> To solve the bug still requires rebuilding whatever links to it; and perhaps whatever links to them?. 
23:42 < dmacks> You're just proposing an easy way to force the rebuilding:)
23:42 < vasi> er, i think just one level, not sure
23:42 < vasi> yeah :-)
23:42 < spundun> BTW I copied /etc/X11/xinit/xinitrc to ~/.xinitrc and appended DYLD_FALLBACK_LIBRARY_PATH=: gnome-session
23:43 < dmacks> (that brings us back to one of drm's favorite issues with foo|foo-ssl, and whether -lfoo causes linkage to libfoo if no symbols are used)
23:44 < vasi> dmacks, sorry?
23:44 < miga> spundun: better to do what vasi said. that is DYL... exec gnome session on a unique line.
23:44 < dmacks> pogma: do we have live examples of things that have incorrect or non-absolute paths encoded, such that they rely on FALLBACK to /usr/lib?
23:44 < spundun> miga: so I put "exec" in the middle?
23:45 < miga> yes, spunden and remove the gnome-session line.
23:45 < spundun> miga: should I export DYLD... ?
23:45 < spundun> or can I just set it?
23:46 < pogma> dmacks: be back in a minute
23:46 < lisppaste> miga annotated #9267 with "better xinitrc" at http://paste.lisp.org/display/9267#4
23:46 < dmacks> Phone ringing...will be back in half hr.
23:46 -!- dmacks is now known as dmacks_Away
23:48 < miga> spundun: instead of copying /etc/xinit... just use the first line I give to you. This way it executes /etc/xinit and, what is very important, makes addition so that fink is happy :-)
23:49 < vasi> miga, there's a mozilla plugin in librsvg2 :-)
23:49 < miga> oh, cool.
23:50 < miga> so, that means it is interesting to install librsvg2?
23:51 < vasi> it's going to be in its own package, librsvg2-mozilla
23:52 < miga> No need to change anything to mozilla? 
23:52 < miga> So that it loads, I mean
23:52 < vasi> since i don't want to force average people to have mozilla in order to use librsvg2 
23:52 < spundun> miga: ok... thanks a lot.
23:53 < vasi> nope, it loads automatically...i'll be committing in a few minutes i think
23:53 < miga> my pleasure, spundun.
23:53 < miga> I'll test it when it comes, thanks, vasi.
23:54 < vasi> i'm using a ton of aux packages here, got librsvg2, -shlibs, -bin, -gtk, -mozilla, -mozilla-dev
23:54 < miga> So, that's a good example for me (for devhelp).
23:55 < vasi> maybe not, i think maybe i over-do it
23:56 < miga> I've made some progress with it. Just a question: should I use devhelp-1.0.0.dylib or devhelp-1.1.1.dylib? Is there a policy here when it is the first time the library appears in fink?
23:56 -!- spundun [~spundun@c-24-126-63-203.hsd1.ca.comcast.net] has quit ["Leaving"]
23:57 < miga> By the way, porting and the packaging manual are not very uptodate in library matters for 10.4.
--- Log closed Tue Jun 21 00:00:12 2005