[02:39] <zul> infinity: i have a bunch of patches to push to you 
[03:09] <lamont> fabbione: kernel-package is b0rked
[03:10] <lamont>        ifneq ($(strip $(architecture)),$(strip $(DEB_BUILD_ARCH)))
[03:10] <lamont> -        #KERNEL_CROSS:=$(architecture)-$(strip $(DEB_HOST_ARCH_OS))-
[03:10] <lamont> -        KERNEL_CROSS:=$(DEB_HOST_GNU_TYPE)-
[03:10] <lamont> -        ifeq ($(architecture), amd64)
[03:10] <lamont> -          KERNEL_CROSS:=$(architecture)-$(strip $(DEB_HOST_ARCH_OS))-
[03:10] <lamont> -        endif
[03:10] <lamont> +        KERNEL_CROSS:=$(architecture)-$(strip $(DEB_HOST_GNU_SYSTEM))-
[03:10] <lamont> +        #KERNEL_CROSS:=$(DEB_HOST_GNU_TYPE)-
[03:10] <lamont> +        #ifeq ($(architecture), amd64)
[03:10] <lamont> +        #  KERNEL_CROSS:=$(architecture)-$(strip $(DEB_HOST_ARCH_OS))-
[03:10] <lamont> +        #endif
[03:10] <lamont>        endif
[03:16] <lamont> hppa is the only one that actually cross-compiles, yes?
[03:22] <lamont> fabbione: at any rate, hppa is ftbfs unless KERNEL_CROSS:=$(architecture)-$(strip $(DEB_HOST_GNU_SYSTEM))-
[03:22] <lamont> and then there's the question of what, if anything, sparc needs...
[03:33] <zul> what did i do now?
[03:33] <zul> ....its not time for school...
[05:19] <lamont> zul: was hoping for an answer on cross-compiling... but it was just a forlorn hope.
[05:38] <fabbione> morning
[05:38] <fabbione> lamont: in theory all arches do crosscompile
[05:38] <fabbione> the sections in kernel-package should be per arch.
[05:39] <fabbione> but if it can't find the compiler, i think it's because the gcc symlinks are missing from gcc-*
[05:39] <fabbione> because it builds in hoary
[05:39] <lamont> no.  it's because the compiler is named hppa64-linux-gnu-gcc, not hppa64-linux-gcc
[05:40] <lamont> and that was a 3.3->3.4 change, I expect
[05:40] <lamont> or rather, it's a breezy binutils change
[05:40] <fabbione> nope.. that was hory -> breezy change..
[05:40] <fabbione> doko needs to add the symlinks to gcc
[05:40] <fabbione> kyle was telling that in #parisc
[05:40] <lamont> it'll be the only package in existance that has them....
[05:41] <fabbione> ah
[05:41] <lamont> Jul 31 15:10:45 <lamont>         /bin/sh: hppa64-linux-objdump: command not found
[05:41] <lamont> Jul 31 15:10:52 <lamont>        dpkg --contents pool/main/b/binutils/binutils-hppa64_2.16.1-2ubuntu2_hppa.deb | grep objdump
[05:41] <lamont> Jul 31 15:10:52 <lamont>        -rwxr-xr-x root/root    262376 2005-07-27 10:15:56 ./usr/bin/hppa64-linux-gnu-objdump
[05:41] <lamont> Jul 31 15:10:55 <lamont>        which one should change?
[05:41] <lamont> Jul 31 15:11:58 <doko>  linux-gnu is the new / correct one
[05:41] <fabbione> than ok, the part in kernel-package is arch specific
[05:41] <fabbione> so if you change that it should be ok
[05:41] <lamont> right, so does my diff look OK then?
[05:42] <lamont> kernel build with that change is running now on hppa, figured I'd let it finish before uploading in any case
[05:43] <fabbione> lamont: i need to look at it... gimme a sec :)
[05:43] <fabbione> i just woke up because my wife forgot to remove the alarm clock at the first day of vac :)
[05:44] <lamont> hehe
[05:44] <fabbione> where is the offset for that patch?
[05:47] <fabbione> lamont: what is the offset for the kernel-package patch you have?
[05:48] <lamont> 2765 or so
[05:48] <lamont> 2765 is the first occurance of KERNEL_CROSS in the file
[05:48] <lamont> and ${destination_architecture}-${DEB_HOST_GNU_SYSTEM} is the correct prefix for the cross compiler everywhere
[05:49] <lamont> then it's just a question of whether or not a cross compiler is needed.
[05:49] <lamont> note that the ppc64 ifneq there breaks cross compiling ppc64 kernels on any arch other than ppc.
[05:49] <lamont> but I'm not going to change that today.. :0)
[05:49] <fabbione> neither we care :)
[05:51] <fabbione> lamont: KERNEL_CROSS:=$(architecture)-$(strip $(DEB_HOST_ARCH_OS))-
[05:51] <fabbione> so this one should be the only correct one...
[05:51] <fabbione> for all arches..
[05:51] <fabbione> and return?
[05:51] <fabbione> $arch-linux- ?
[05:52] <fabbione> sparc-linux-gnu-gcc-3.4
[05:52] <fabbione> i have the -gnu- in the middle
[05:52] <fabbione> root@vultus5:~ # dpkg-architecture -qDEB_HOST_ARCH_OS
[05:52] <fabbione> linux
[05:53] <fabbione> what adds -gnu ?
[05:53] <lamont> DEB_HOST_GNU_SYSTEM is the right one
[05:53] <doko> fabbione: DEB_BUILD_GNU_TYPE or DEB_BUILD_GNU_SYSTEM
[05:53] <doko> good morning, btw ...
[05:53] <fabbione> oh i see
[05:53] <fabbione> ok
[05:54] <fabbione> it seems fine for sparc
[05:54] <fabbione> lamont: go ahead..
[05:54] <lamont> so my change was to uncomment the line at 2765, and comment out the ones following
[05:54] <lamont> :-)
[05:54] <lamont> fabbione: ok.  I'll do that after the kernel build finishes in a few hours...
[05:54] <lamont> and I wake up again
[05:54] <fabbione> lamont: bump the B-D for 6.7
[05:54] <fabbione> just to ensure to use the proper one
[05:55] <fabbione> the baz dance is:
[05:55] <fabbione> 1) start from playgroud
[05:55] <fabbione> 2) merge into mainline--2.6.12
[05:55] <fabbione> 3) tag
[05:55] <fabbione> 4) branch the new pre
[05:55] <fabbione> 5) switch to the new pre
[05:55] <fabbione> 6) ./debian/rules startnewrelease (or something)
[05:55] <fabbione> 7) baz lint
[05:56] <fabbione> 8) add new files
[05:56] <fabbione> 9) check the diff
[05:56] <fabbione> 10) copy the abi
[05:56] <fabbione> 11) commit
[05:56] <fabbione> if there is no ABI change: baz mv debian/abi/2.6.12-6.6 debian/abi/2.6.12-6.7
[05:57] <fabbione> because you are working on the new 6.8 
[06:00] <fabbione> doko: dude.. i really really need a fixed binutils..
[06:00] <fabbione> do you know if there is any progress?
[06:01] <lamont> fabbione: will bump
[06:03] <lamont> hehe..  branch with one arg switches.. :)  so 4,5 is one step
[06:04] <lamont> fabbione: and i'm not worried about a -6.x ABI transition on ia64, since non-booting->booting is a good change.. :0)(
[06:04] <fabbione> lamont: the point is that if there is an abi change, it will FTBFS because of the checker
[06:04] <fabbione> the files must be there
[06:04] <fabbione> otherwise even make clean will fail
[06:05] <lamont> ah, right
[06:06] <lamont> how do I know there's no abi change? test builds everywhere?
[06:06] <fabbione> so if there is an abi change use the nice target in debian rules: bumpabi (needs fakeroot)
[06:06] <fabbione> lamont: you did change only one CONFIG_ on ia64, right?
[06:06] <lamont> yrd
[06:06] <lamont> yes
[06:06] <fabbione> so just build on ia64
[06:06] <lamont> CONFIG_PRINTK_TIMING or some such was fatal
[06:06] <lamont> ah, ok
[06:07] <fabbione> if it goes trough all of the 4 without screaming and yelling, there is no ABI change
[06:07] <fabbione> otherwise it will tell you that there are symbols mistmatches
[06:07] <lamont>  /bin/sh: hppa64-linux-objdump: command not found
[06:07] <doko> fabbione: no, drow just '*sigh*'ed about it
[06:07] <lamont> WTF?
[06:07] <fabbione> it's missing -gnu-
[06:08] <lamont> yeah.  that's with my change
[06:08] <fabbione> doko: ok thanks
[06:08] <lamont> fabbione: so that's "it's _still_ missing -gnu"
[06:08] <fabbione> good night lamont
[04:25] <zul> that dvb card patch looks correct
[07:38] <dilinger> fabbione: ping
[07:38] <dilinger> fabbione: actually, nm
[07:48] <lamont__> sigh... hppa's build issue is, well, hppa.  OTOH, kernel-package should still be fixed,dammit
[07:48] <lamont__> doko: the ${ARCH}-linux-gnu-objdump change, when does that hit debian?
[07:51] <doko> lamont__, it already did
[07:51] <lamont__> way cool
[07:59] <lamont__> hrm... any patches from anyone for -6.7?
[07:59] <jbailey> lamont__: You doing a new kernel?
[08:00] <lamont__> yes
[08:00] <jbailey> lamont__: I buillt a new kernel on the weekend with the initramfs/dsdt patch in it, but I haven't tested it yet.  When do you need it by?
[08:00] <lamont__> so far, it's one semi-non-bug, and hppa/ia64 fixes...
[08:00] <lamont__> I want a better excuse... :)
[08:01] <jbailey> This will be a good enough excuse.  Lemme test the patch. =)
[08:01] <lamont__> hrm... 1600 UTC or so... but that's past already... so we have time. :-(
[08:02] <lamont__> before midnight UTC-0600 would be way cool
[08:11] <zul> i have a bunch of patches in my baz
[08:13] <jbailey> Hmm, well the acpi patch doesn't keep it from booting anyway.  Now let's try a DSDT replacement.
[08:17] <jbailey> Old style DSDT replacement still works.
[08:25] <lamont__> zul: anything significant?
[08:25] <lamont__> anything that wants/needs to get into -6.7?
[08:49] <zul> lamont__: acerhk support, bug fix, and a couple of patches i stole from git
[08:52] <zul> everything is signifacant ;)
[08:55] <zul> heh...redhat is a bit behind the times
[09:05] <jbailey> Dear keybuk, why did you taunt us with Wigg&Penn?  
[09:16] <lamont__> zul: if you want to merge from the playground, and test your stuff, i expect I can include it... I'd like to avoid an ABI bump though...
[09:16] <zul> lamont__: did that yesterday i didnt notice an abi bump it wouldnt build if it did would it?
[09:17] <lamont__> right.  but I really don't want to follow -6.7 with -7.8 just because of that...
[09:17] <zul> true
[09:17] <zul> did you commited anything to the playground today?
[09:17] <lamont__> playground has all my stuff
[09:18] <zul> ok ill re-build my stuff in a bit and ill let you know 
[09:19] <zul> with the playground+my patches
[09:19] <lamont__> thans
[09:19] <lamont__> if you have a reachable archive somewhere, I can toss it at the DC machines to do test builds around
[09:20] <zul> yep its at http://zulinux.homelinux.net/arch/zulcss@gmail.com--2005
[09:21] <zul> crap i still have *.orig in my patches
[09:21] <lamont__> coolness - holler once your merge is done, eh?
[09:22] <zul> ok
[09:24] <zul> bbl...im building right now i need to finish off the house work