**** BEGIN LOGGING AT Thu Apr 19 02:59:57 2007
Apr 19 03:03:44 <sn9> could be a mistake in your menuconfig
Apr 19 03:05:25 <overseer> i'm trying the same build (r6962) on 3 hosts -- all centos-4.4.2 -- two x86, one x86_64.   same menuconfig steps after 'make distclean ; rm .config'  -- change to broadcom 2.6 main, and wgt634u subtype, thats it.
Apr 19 03:06:04 <overseer> only the 64bit host compiles the whole tree without stopping in that same g++spec spot
Apr 19 03:06:11 <sn9> make clean before make distclean
Apr 19 03:06:44 <overseer> i've gone as far as rm -rf trunk followed by svn checkout to try and figure this out.
Apr 19 03:07:01 <Bartman007> overseer: centos 4.4.2?
Apr 19 03:07:06 <overseer> y
Apr 19 03:07:07 <sn9> and did that work?
Apr 19 03:07:21 <overseer> nope -- same behavior -- 64bit host compiles clean, others fail
Apr 19 03:07:45 <overseer> centos-release-4-4.2 :)
Apr 19 03:07:47 <sn9> you're using the same .config each time?
Apr 19 03:08:29 <overseer> same default config, which is none, followed by menuconfig to change to 2.6 broadcom kernel and wgt634u type build.
Apr 19 03:08:37 <Bartman007> arg...
Apr 19 03:08:48 <sn9> hmm
Apr 19 03:08:59 <overseer> i have V=99 logs from all hosts now :)
Apr 19 03:09:06 <Bartman007> I forgot, create an x86 domU.
Apr 19 03:10:15 <Bartman007> can't create*.
Apr 19 03:11:38 <sn9> r6996 built fine on ia32 for me
Apr 19 03:11:59 <Bartman007> same here.
Apr 19 03:12:03 <Bartman007> not centos though.
Apr 19 03:12:11 <sn9> right
Apr 19 03:12:12 <overseer> i'm with ya there -- i don't believe this is a systemic issue with the build tree but rather a subtler issue somehow not caught by the buildreqs or ... something else weird
Apr 19 03:12:28 <sn9> overseer: make --version
Apr 19 03:12:51 <overseer> GNU Make 3.80 on all platforms
Apr 19 03:12:56 <sn9> bingo
Apr 19 03:13:06 <overseer> including the host that builds sucessfully :)
Apr 19 03:13:25 <sn9> the buildroot is known-incompatible with anything below 3.81
Apr 19 03:13:37 <overseer> except apparently on x86_64 ... wow
Apr 19 03:14:46 <overseer> so minimum version is > 3.81 ?
Apr 19 03:14:51 <sn9> yep
Apr 19 03:15:30 <sn9> the sanity checks even say so, but they don't check well enough to know when to say it
Apr 19 03:16:15 <overseer> uh, how about at the toplevel default 'make world' when its checking for awk and friends?
Apr 19 03:16:27 <sn9> that's the one
Apr 19 03:17:11 <sn9> you'll see what i mean if you read the makefile
Apr 19 03:17:32 <overseer> checking over my logs of the build process (including stderr) i see nothing related to make version
Apr 19 03:18:00 <sn9> it's the very first line
Apr 19 03:19:26 <overseer> first line of the script or first line output from 'make' when run ?
Apr 19 03:19:32 <overseer> s/script/Makefile/
Apr 19 03:19:33 <sn9> output
Apr 19 03:20:28 <overseer> Checking 'non-root'... ok.  is my first line output
Apr 19 03:20:51 <sn9> huh? that's the last sanity check, not the first
Apr 19 03:21:50 <overseer> ok, so how to i reset those checks, make clean ; make distclean didn't do it...
Apr 19 03:22:14 <sn9> it checks every time, regardless of anything else
Apr 19 03:22:40 <sn9> maybe V=99 suppresses it partially
Apr 19 03:22:57 <overseer> maybe i'm forgetting V=99 in my stupor
Apr 19 03:23:21 <sn9> if you don't V=99, you see all the sanity checks
Apr 19 03:23:40 <sn9> unfortunately, the first one is brain-damaged
Apr 19 03:24:15 <overseer> Checking 'non-root'... ok.  --- that's my first line from 'make clean; make V=99'
Apr 19 03:24:26 <sn9> leave off V=99
Apr 19 03:24:42 <overseer> second line is Checking 'working-make'... ok.
Apr 19 03:25:01 <sn9> hmm, it was the other way when i built
Apr 19 03:25:13 <overseer> i was too focused to even notice the second line
Apr 19 03:25:26 <sn9> then again, i use FORCE=1 because i'm root
Apr 19 03:26:11 <overseer> i see the check in include/prereq-build.mk now -- but the test is lame
Apr 19 03:26:26 <sn9> that's what i've been saying
Apr 19 03:26:33 <overseer> *groan* -- i agree
Apr 19 03:27:49 <sn9> 3.80 passes it with flying colors and blows up later
Apr 19 03:28:28 <sn9> took me some time & whining to figure that out
Apr 19 03:29:43 <overseer> can't there be a simple test for 3.80 that just checks the version# and exits with failure?
Apr 19 03:29:56 * overseer grovels rpm repos for a make-3.81
Apr 19 03:30:00 <sn9> but since i went to 3.81 to fix it anyway, i could no longer come up with a patch that i could try against 3.80
Apr 19 03:30:31 <overseer> logic: string match, version # == bad ?? fail
Apr 19 03:31:13 <sn9> logic is one thing, empirical demonstration quite another
Apr 19 03:31:42 <overseer> i'll try to submit one now that i have 3 platforms to try it on
Apr 19 03:32:36 <sn9> also test against 3.79.1 and 3.79, or is that too much to ask?
Apr 19 03:37:31 <overseer> i'm surprised there isn't a blacklist of versions in the test
Apr 19 03:37:59 <sn9> the test as is made no sense to me at all
Apr 19 03:39:22 <overseer> finding a fedora/centos/redhat package of make-3.81 is causing me trouble
Apr 19 03:41:03 <sn9> you can find 3.79.1 in rh7.3
Apr 19 03:41:25 <overseer> thanks :)
Apr 19 03:42:50 <sn9> fc6 has 3.81
Apr 19 03:43:19 <overseer> i just found EL5 has it too
Apr 19 03:43:27 <overseer> rpmfind didn't have it
Apr 19 03:43:33 <sn9> rpmseek did
Apr 19 03:43:41 <sn9> i checked both
Apr 19 03:44:52 <sn9> but they don't list el5, only el4, which has 3.80
Apr 19 03:45:14 <overseer> i used search 'make' on rpmfind and the latest listed was 3.80
Apr 19 03:45:29 <overseer> so, my guess is it hasn't indexed the EL5 tree
Apr 19 03:45:44 <sn9> rpmfind didn't even have fc6
Apr 19 03:45:49 <sn9> rpmseek did
Apr 19 03:46:07 <overseer> ah -- didn't notice the different sitename
Apr 19 03:47:00 <sn9> where did you check el5?
Apr 19 03:50:42 <overseer> hmph?
Apr 19 03:51:02 <sn9> here, use this with rpmbuild --rebuild: http://mirrors.kernel.org/centos/5.0/os/SRPMS/make-3.81-1.1.src.rpm
Apr 19 03:55:22 <overseer> the centos-5-key.asc i got from keyserver.net was borked
Apr 19 03:56:10 <sn9> http://mirrors.kernel.org/centos/5.0/os/i386/RPM-GPG-KEY-CentOS-5
Apr 19 03:56:17 <overseer> i'm good
Apr 19 03:56:24 <overseer> installing built rpm
Apr 19 03:56:57 <sn9> you really should make your own with rpmbuild --rebuild
Apr 19 03:57:03 <overseer> i did
Apr 19 03:57:32 <overseer> and not as root ;)
Apr 19 03:57:46 <sn9> i think you have to be root
Apr 19 03:58:01 <overseer> to build what?
Apr 19 03:58:06 <sn9> an rpm
Apr 19 03:58:20 <overseer> you're on some good drugs, do you have any extra to share?
Apr 19 03:58:49 <sn9> when i dealt with rpm's, they never built right as non-root
Apr 19 03:58:53 <overseer> trust me when i say this: buildling rpms as root is asking for trouble.
Apr 19 03:59:46 <overseer> would you like my .rpmmacros file that lets anyone build rpms as non-root?
Apr 19 03:59:50 <sn9> rpm's are trouble without asking, thanx
Apr 19 03:59:54 <overseer> lol
Apr 19 04:00:12 <overseer> 11-lines of yummy rpm goodness
Apr 19 04:00:54 <sn9> yum is slow
Apr 19 04:00:55 <overseer> oh, sn9: btw -- after upgrading make, the first sanity check is now:  Checking 'working-make'... ok.
Apr 19 04:01:03 <overseer> s/yummy/delicious/
Apr 19 04:01:34 <overseer> and non-root got moved to last
Apr 19 04:02:05 <sn9> it's bizarro-make!
Apr 19 04:02:12 <overseer> it really is
Apr 19 04:03:01 <overseer> ok, it'll take 10-15min to get to the breakpoint... thanks for your suggestions
Apr 19 04:03:07 <sn9> yw
Apr 19 04:03:10 <overseer> bbl maybe
Apr 19 04:03:29 <overseer> hopefully with an effective sanity check for the future
Apr 19 04:04:27 <sn9> against r7000, preferably
Apr 19 04:04:42 <overseer> is that a milestone ;)
Apr 19 04:04:53 <sn9> it's also the current rev
Apr 19 04:05:54 <sn9> my patches are proudly included
Apr 19 04:07:29 <sn9> my persistence turned the rdc-2.6 port from "broken" to "supported" very quickly
Apr 19 04:07:50 <overseer> my patches are proudly included in the 2.4 series kernel tcp.c
Apr 19 04:07:56 <overseer> whats the rdc-2.6 ?
Apr 19 04:08:59 <sn9> i was instrumental in getting the RDCPort (that's the wiki page title) off the 2.4 dependency
Apr 19 04:09:52 <overseer> usb devices?
Apr 19 04:10:08 <sn9> has not been tested with usb
Apr 19 04:10:22 <overseer> curious what's out there that replaces the wgt634u
Apr 19 04:11:22 <sn9> asus wl-500g-premium is good
Apr 19 04:11:47 <sn9> faster cpu than the 534u
Apr 19 04:11:52 <sn9> *634u
Apr 19 04:12:33 <overseer> i have the turtle beach audio usb dongle on one of my 634's which sounds great and I highly recommend
Apr 19 11:42:40 <CIA-3> nbd * r7005 /trunk/package/madwifi/files/lib/wifi/madwifi.sh: don't try to set up wpa if hostapd is not installed
Apr 19 12:06:42 <CIA-3> kaloz * r7006 /trunk/package/wireless-tools/Makefile: we don't need the shared library there
Apr 19 13:19:38 <CIA-3> florian * r7007 /trunk/package/admswconfig/ (. Makefile): Add admswconfig until we add support for adm5120 switch in kmod-switch, thanks Thomas Langer for the makefile.
Apr 19 13:24:14 <CIA-3> florian * r7008 /trunk/package/admswconfig/Makefile: Fix a typo
Apr 19 13:27:08 <CIA-3> florian * r7009 /packages/utils/io/ (. Makefile src/ src/io.c): Add io, thanks Thomas Langer !
Apr 19 15:14:26 <loswillios> after flashing r7006 my /etc/config/wireless is empty on a WGT634U
Apr 19 15:17:51 <loswillios> uh. wlanconfig ath create wlandev wifi0 wlanmode monitor -> wlanconfig: ioctl: No such device
Apr 19 15:18:10 <loswillios> Maybe my wlan-card is b0rked :/
Apr 19 23:48:03 <[mbm]> damn, cable modem is doing over 700k/s on the ubuntu torrent
Apr 19 23:48:14 <[mbm]> not used to seeing downloads go much above 300
Apr 19 23:48:53 <sn9> it was released today
Apr 19 23:49:15 <[mbm]> no, I mean, I have a crappy isp that's actually working foronce
Apr 19 23:49:48 <[mbm]> hmm and now the traffic shaping kicks in
Apr 19 23:49:54 <[mbm]> and drops my speed
Apr 19 23:49:59 <sn9> it's more demand, too
Apr 19 23:52:01 * [mbm] wants to see what they've done with virtualization in the server distro
Apr 19 23:55:59 <sn9> not that much
Apr 19 23:56:15 <sn9> the 64-bit edition sucks ass
Apr 19 23:56:32 <[mbm]> ?
Apr 19 23:56:49 <sn9> it can't run 32-bit binaries, period
Apr 19 23:57:11 <[mbm]> actually yes it can, you just need to install the 32bit libs
Apr 19 23:57:18 <sn9> they don't work
Apr 19 23:57:29 <[mbm]> worked in edgy
Apr 19 23:57:37 <sn9> they work on all other distros
Apr 19 23:57:55 <sn9> you got them to work in edgy? first i've heard of that
Apr 19 23:58:10 <[mbm]> enough that I could run google earth
Apr 19 23:58:18 <sn9> doesn't count
Apr 19 23:58:29 <[mbm]> ?
Apr 19 23:58:41 <sn9> google earth carries its own libs, remember?
Apr 19 23:58:49 <[mbm]> is there a specific 32bit app you want?
Apr 19 23:59:59 <sn9> the general capability outweighs specific apps unless the virtualization is mature, which *buntu hasn't put enough emphasis on
Apr 20 00:00:27 <[mbm]> umm is there any linux distribution with a mature virtualization at this point?
Apr 20 00:00:42 <sn9> most others, including debian
Apr 20 00:01:14 <[mbm]> so how are you defining mature?
Apr 20 00:01:40 <sn9> "works without having to rewrite half of it"
Apr 20 00:02:51 <[mbm]> haven't seen any distributions with virtualization working out of the box
Apr 20 00:03:28 <sn9> not yet
Apr 20 00:03:42 <[mbm]> even less with a polished interface
Apr 20 00:03:49 <sn9> except maybe gentoo to some degree, but i wouldn't know
Apr 20 01:27:31 <overseer> sn9: last night we discussed my build failures and the fix suggested was 'upgrade make' -- i tried upgrading make to 3.8.1 but the x86 hosts still fail where the x86_64 host succeeds -- any more thoughts?
Apr 20 01:28:19 <sn9> clearly, there is an additional problem
Apr 20 01:29:59 <overseer> x86_64 has gcc-3.4.6-3.1, tried both gcc-3.4.6-3 and gcc-3.4.6-3.1 on the x86 hosts, without change
Apr 20 01:30:35 <sn9> next thing i would check is the JLEVEL
Apr 20 01:30:43 <sn9> in the .config
Apr 20 01:30:53 <overseer> not familiar with that -- what to look for?
Apr 20 01:30:57 <sn9> but you already said that was default
Apr 20 01:31:28 <overseer> ya, i start without a .config, and only change to 2.6brcm and wgt634u wifi
Apr 20 01:34:34 <overseer> is it strange that it fails during the toolchain/gcc-install ?  i'd expect it to fail in gcc-compile
Apr 20 01:35:16 <sn9> ccache is off, right?>
Apr 20 01:35:33 <overseer> don't know how to check that
Apr 20 01:36:11 <sn9> .config, again
Apr 20 01:58:50 <overseer> CONFIG_JLEVEL=1 on all hosts
Apr 20 02:05:20 <sn9> you have single cpu's on everything, right?
Apr 20 02:33:48 <overseer> SMP kernels on everything
Apr 20 02:34:08 <sn9> but how many cores?
Apr 20 02:34:19 <overseer> some dual some just have HT turned on
Apr 20 02:34:48 <sn9> try increasing the JLEVEL to 2 and see what happens
Apr 20 02:35:09 <overseer> ok , make clean , make distclean , make V=99 ?
Apr 20 02:35:22 <sn9> yeah
Apr 20 02:35:52 <overseer> s/,/&&/
**** ENDING LOGGING AT Fri Apr 20 02:59:57 2007