[12:10] <mjg59> Scott's going to be the only one without any at this rate
[12:13] <jdub> mjg59: spread some love by convincing mdz/fabbione to accept the dsdt-initrd patch ;)
[12:13] <mjg59> Haha
[12:13] <stuNNed> jdub, schweet
[12:13] <jdub> i'm going to be forever building kernels :|
[12:13] <stuNNed> jdub, disabling agp, installing custom acpi script from wiki and editing a bit fixed acpi here (for the last 3 days)
[12:13] <jdub> i need a custom dsdt
[12:13] <stuNNed> does ubuntu installer have ntfsresize and vfatresize utils for users deciding to install but not wipe?
[12:13] <mako> you will be in mataro and want to participate in the keysinging: http://people.ubuntulinux.org/~mako/ksp-mataro/
[12:13] <mako> the keySIGNing will be directly before the singing i think
[12:13] <mako> or directly after, i can never keep it straight
[12:13] <mdz> need to shut off power to the whole house for a bit
[12:13] <mdz> need to do this while there's sunlight :-)
[12:13] <jdub> heh
[12:13] <mdz> back in hopefully under an hour
[12:13] <jdub> ciao
[12:32] <Matt|> guys what is the difference between the linux-kernel-headers package, and the linux-headers packages?
[12:38] <mako> mdz: so what were you doing that required unplugging your whole house?
[12:39] <mdz> mako: replacing a circuit breaker
[12:40] <mdz> call me funny, but I have this hangup about poking about with a screwdriver inside a live panel
[12:40] <jdub> seriously dude
[12:40] <jdub> what could go wrong?
[12:40] <jdub> the worst that could happen is that you turn into a superhero
[12:40] <jdub> like ELECTRODE MAN
[12:40] <mako> mdz: jdub is right
[12:40] <jdub> or THE BLUE SMOKE
[12:41] <mako> i saw just that happen in ernest goes to jain
[12:41] <mako> sorry, jale
[12:41] <mako> jail
[12:41] <mako> damnit
[12:41] <jdub> through cargo cult logic, whenever someone has an outlandish accident, they turn into a superhero
[01:30] <mdz> Kamion: hoary/powerpc installation with 20041201.1 was a success
[02:19] <lamont>   evolution: Depends: libgtkhtml3.6-12 (>= 3.5.0) but it is not installable
[02:19] <lamont> (it's libgtkhtml3.6-13 now...)
[03:03] <melazyboy> VLC is out of date, they have a new packages that fixes numerous bugs
[04:38] <jdub> hahahahah
[04:38] <jdub> Most of the distributions built on top of Debian, such as Linspire, Xandros, Skolelinux, LinEx, or Ubuntu, apply some discretion in the packages they select. They are unlikely to include tools like hot-babe, and, thus, may be considered safer versions to use in situations where somebody may get offended.
[04:38] <jdub> Well, OK, perhaps we can't be too sure with Ubuntu. 
[04:39] <jdub> ^ from this week's lwn
[04:39] <jdub> 
[08:46] <fabbione> morning guys
[09:12] <Keybuk> mdz: around?
[10:08] <Kamion> stuNNed: the hoary installer includes ntfsresize, and the partition manager knows how to call out to it to resize NTFS partitions; parted supports FAT natively
[10:08] <Kamion> stuNNed: the warty installer doesn't include ntfsresize
[10:42] <Keybuk> definite need for apt-get install pxeboot-loving
[11:24] <loogaroo> hi
[11:25] <loogaroo> is somewhere a guide for developing configuration tools for ubuntu?
[11:25] <Kamion> gnome-system-tools is probably the place to start
[11:26] <loogaroo> But, I prefer python
[11:27] <azeem> loogaroo: you also /might/ want to look at the debian-desktop archives. They seem to work on something connection debconf, config4gnu and g-s-t
[11:27] <Kamion> g-s-t has python integration I thought
[11:28] <ross> g-s-t is C frontend and perl backend
[11:32] <loogaroo> and what is/are the prefered languages/toolkits of the ubuntu devs?
[11:32] <ross> python and gtk are blessed
[11:32] <bob2> {py,}glade is 31337
[11:33] <loogaroo> hmm, sorry what do you mean with 31337, should this be leed?
[11:33] <Kamion> he's joking
[11:34] <Kamion> slang for "elite", parody of script-kiddie talk
[11:34] <loogaroo> :( and I do not understand it
[11:34] <loogaroo> oh, ok
[11:34] <bob2> = it's quite a nice way to write gui code in python
[11:35] <loogaroo> then my next questition, are there any libs for python which handles the linux config files?
[11:35] <ross> loogaroo: "the"? they are all formatted differently :)
[11:36] <ross> python will let you parse most files natively
[11:36] <loogaroo> ok
[11:37] <loogaroo> is it possible to render a gtk gui at the console?
[11:37] <loogaroo> like curses
[11:38] <ross> there is a hack, but no
[11:38] <ross> if you want to do curses, use curses
[11:38] <azeem> or debconf perhaps
[11:39] <Kamion> I would avoid debconf outside the installer and packages' maintainer scripts
[11:40] <azeem> I didn't say it was a good idea
[11:40] <azeem> at least, not until there is something like the GNOME HIG for debconf-UI
[11:42] <loogaroo> I like gtk, but I think gtk config tools are not so handy for remote work
[11:43] <bob2> that's when you use an editor :)
[11:44] <loogaroo> hmm
[11:45] <ross> loogaroo: or remote X :)
[11:45] <loogaroo> VNC :)
[11:45] <ross> why export a desktop over the network when the application you want will do the job?
[11:46] <loogaroo> ok, X is better
[11:47] <loogaroo> but you are right it makes not much difference, if curses or gtk forwarted
[11:48] <loogaroo> so gtk is nice
[11:54] <Keybuk> gtk is sex
[11:54] <ross> this isn't sex: http://gemal.dk/misc/nsb05.png
[11:55] <ross> "How Not To Make A Web Browser"
[11:56] <Keybuk> Whoah!
[11:56] <jdub> isn't that astoundingly bad?
[11:56] <thom> holy mother of god
[11:56] <jdub> i wonder who they got to do that
[11:57] <Kamion> my God, that's a busy-looking screen
 My eyes couldn't actually stay on it, which is why it took me so long to realise how bad it was.
[11:58] <ross> rofl
[11:59] <ross> everyone seen the ubuntu comment in this weeks LWN?
[11:59] <jdub> yeah :)
[11:59] <jdub> haw haw
[11:59] <jdub> that's going to stick
[11:59] <Keybuk> Ubuntu, it's a little bit wooo, a little bit wayyy
[11:59] <ross> haha
[11:59] <ross> jdub: get the ubuntu models to do some more artwork for hot-babe
[12:00] <thom> ubuntu-hotter-babe
[12:00] <smurfix> ross: ubuntu'll never live that down anyway, so go with it ;-)
[12:00] <daniels> Kamion: MUST ADD MORE SHINY
[12:01] <ross> i'm still wondering what all the buttons on the tab bar actually do
[12:02] <thom> i have closed the screenshot because it made my brain close
[12:04] <robatylor|away> mako: happy birthday, btw :)
[12:04] <jdub> does anyone have bluetooth input hardware?
[12:05] <jdub> keyboard, mouse, etc.
[12:05] <Treenaks> didn't edd have that?
[12:05] <azeem> on a related note, is it safe to buy just any bluetooth USB dongle, or should I be careful about the model WRT Linux/Ubuntu compatibility?
[12:06] <daniels> azeem: seems to be fairly easy to just grab random crap
[12:06] <azeem> daniels: great, thanks
[12:06] <edd> azeem: you're unlikely to buy one there's no drivers for these days
[12:06] <ross> azeem: i bought the cheapest i could find and it works fine
[12:06] <jdub> edd: coming to matar?
[12:06] <Treenaks> ross: ubuntu comment on LWN? (I don't have a subscription..)
[12:06] <edd> jdub: no. family time this month, etc.
[12:06] <daniels> i just got a 100m-range belkin one and it worked fine
[12:06] <jdub> ahr.
[12:06] <daniels> edd: btw, my k700i magically kinda fixed itself, so sweet dela
[12:06] <daniels> deal, also
[12:07] <edd> daniels: sick
[12:07] <edd> (i heard that on radio 1, it means 'good')
[12:07] <Mithrandir> Treenaks: They [Debian derivatives]  are unlikely to include tools like hot-babe, and, thus, may be considered safer versions to use in situations where somebody may get offended.
[12:07] <Mithrandir> Well, OK, perhaps we can't be too sure with Ubuntu.
[12:07] <daniels> edd: ill
[12:07] <azeem> Mithrandir: heh, yeah
[12:07] <Mithrandir> jdub: my phone and hopefully my shinyshinyshiny x40.
[12:07] <daniels> Mithrandir: word
[12:07] <jdub> Mithrandir: input hardware, like keyboards/mouse
[12:08] <sjoerd> jdub: i've got a bluetooth mouse
[12:08] <Mithrandir> jdub: oh, that kind of hardware.
[12:08] <bob2> jdub: lu has a bt keyboard, iirc
[12:08] <thom> certainly a bt mouse
[12:11] <jdub> cool
[12:12] <Treenaks> What about GPS integration btw? :)
[12:12] <Treenaks> I have a USB GPS device.. shouldn't the weather applet etc. derive the 'default location' from that? :)
[12:13] <jdub> heh
[12:13] <Mithrandir> Treenaks: patches accepted :)
[12:13] <Treenaks> Mithrandir: hmm.. we'd first need to separate gpsd from the gpsdrive package, probably
[12:13] <Keybuk> Treenaks: interesting way to seed /etc/timezone
[12:14] <Treenaks> Mithrandir: and then autodetect GPS dongles
[12:14] <Mithrandir> Treenaks: wouldn't be against me :)
[12:14] <Treenaks> (which can be Hard because they're just serial convertors..)
[12:14] <Mithrandir> Treenaks: autodetecting USB stuff is a bit icky, yes.
[12:15] <Treenaks> oh well, I guess I can use it to learn python :)
[12:16] <Mithrandir> uhm, s/USB/GPS/
[12:16] <sjoerd> Treenaks: you probably want ngpsd btw (if it has it's dbus interface by now)
[12:16] <jdub> what on earth do you guys use gps for?
[12:16] <Treenaks> sjoerd: if dbus and HAL? :)
[12:16] <Mithrandir> Keybuk: that should work fine, if you can detect that you have a GPS.
[12:16] <Mithrandir> jdub: play.
[12:16] <Treenaks> jdub: wardriving/cycling/walking :)
[12:16] <sjoerd> jdub: geek toys :)
[12:17] <sjoerd> and time 
[12:17] <Treenaks> jdub: and as toys :)
[12:17] <sjoerd> a gps device delivers extremely acurrate time information if it has a fix
[12:17] <Treenaks> sjoerd: accurate time + bluetooth delay + usb delay = inaccurate time :P
[12:17] <sjoerd> bluetooth delay isn't that big
[12:18] <Mithrandir> Treenaks: USB delay shouldn't be too bad, should it?
[12:18] <Treenaks> Mithrandir: depends on the amount of hubs involved
[12:18] <jdub> i tend to carry about four or five hubs when cycling, yeah
[12:18] <sjoerd> Treenaks: none (apart from the root)
[12:19] <Treenaks> jdub: You are jdub of borg?
[12:19] <sjoerd> and my laptop has bt built in.. so dedicated usb controller
[12:19] <Mithrandir> jdub: but you're weird.
[12:20] <sjoerd> jdub: why the interest in bluetooth input devices btw
[12:20] <jdub> sjoerd: bluetooth integration in hoary
[12:20] <sjoerd> ah
[12:20] <Treenaks> jdub: /all/ kinds of bt devices? :))
[12:21] <edd> jdub: you'll my tracking my recent frenzy of uploads then :)
[12:21] <jdub> Treenaks: sure
[12:21] <sjoerd> i can bring an bt headset too, but there is no good code for it yet..
[12:21] <jdub> edd: :-)
[12:21] <Treenaks> sjoerd: there is time to hack..
[12:21] <jdub> sjoerd: that hardware in particular?
[12:21] <Treenaks> sjoerd: I have a BT headset as well, but it doesn't even integrate nicely with my phone
[12:21] <Treenaks> sjoerd: (nokia headset + nokia phone..
[12:22] <sjoerd> jdub: no general bluetooth audio stuff
[12:22] <Treenaks> I have a Nokia HS-3W I could bring..
[12:23] <jdub> Treenaks: bt incompatibility seems endemic now, particularly with phones
[12:23] <Treenaks> jdub: I need to get new firmware on my phone.. it should fix some of the BT problems
[12:23] <jdub> hrm
[12:23] <jdub> i haven't even thought to see if there were upgrades for mine
[12:24] <Treenaks> jdub: which one?
[12:24] <jdub> t630
[12:25] <sjoerd> jdub: if you know how to put new nice firmware on that, lemme know 
[12:25] <Treenaks> http://www.mobiledia.com/forum/topic7181.html ?
[12:26] <Treenaks> http://www.esato.com/archive/t.php/t-68451 sorry
[12:28] <Treenaks> sjoerd: like my 3650..
[12:28] <Treenaks> smurfix, sorry
[12:28] <smurfix> Treenaks: what for?
[12:29] <Treenaks> smurfix: oh, it tends to disconnect after some AT command on BT
[12:30] <smurfix> Ah, the wrong nick. I first thought you were sorry about my phone. ;-)
[12:30] <Treenaks> smurfix: well, that too :P
[12:31] <jdub> hrm
[12:31] <jdub> i have latest non-developer firmware according to that page
[12:31] <smurfix> Anyway it also has infrared, so the BT-challenged can do some stuff too.
[12:31] <Treenaks> my 3650 has 2.50.. latest released version: somewhere around 4.17
[12:32] <Treenaks> 4.13
[12:32] <Treenaks> no 4.17 :)
[12:35] <bob2> does linux tend to usefully support talking to phones via ir?
[12:36] <azeem> depends on how you define 'talking'
[12:37] <bob2> can I send sms? read them?
[12:38] <azeem> I do that for my Siemens S45 via scmxx
[12:38] <azeem> if you have a phone which supports IrMC (like the S45*i*, drat), you can easily sync contacts and calendar, too
[12:41] <bob2> bah, nokie.
[12:41] <bob2> er, nokia.
[12:41] <azeem> there are apps for nokia phones as well, but I don't know them
[12:41] <azeem> gnokii or somethign
[12:41] <bob2> yeah
[12:42] <stratus> heh
[12:42] <seb128> ah ah
[12:42] <sjoerd> bob2: most are just serial ports, so you can send normal AT command to them
[12:42] <smurfix> azeem: define "easily". Between my old Psion and the phone, for instance, *somebody* messed up the timezones. All my all-day events now start at 2am instead. :-/
[12:43] <jdub> smurfix: multisync has some funny timezone options
[12:43] <stratus> seb128, FYI #3634 is reproducible with evolution 2.1.1
[12:43] <azeem> smurfix: "easily" as in: "You do not have to download stuff from the phone and sync maunally via vi" :p
[12:44] <azeem> anyway, off for lunch, laters
[12:44] <seb128> stratus: that why it's not closed :)
[12:46] <stratus> seb128, hmm have you reproduced this bug? AFAIK it's just a problem here.
[12:47] <thom> ahr, readahead package
[12:47] <smurfix> jdub: Ah, thanks
[12:47] <seb128> stratus: it doesn't crash here, but the bug is still open upstream and has 7 dups
[12:47] <jdub> thom: hrrrmmm?
[12:47] <thom> for great justice^Wspeed
[12:47] <thom> jdub: packaging readahead, currently
[12:47] <jdub> thom: i looked at doing that the other week -> we don't have glibc support for it, though
[12:47] <thom> yeah we do
[12:48] <thom> i'm just ignoring nfs
[12:48] <stratus> seb128, ah i didn't read the 7 dups...reported at upstream or ubuntu ?
[12:48] <seb128> upstream
[12:48] <stratus> thanks.
[12:48] <seb128> np
[12:49] <thom> jdub: in our typical case, i don't think it's a problem
[12:49] <jdub> mmm
[12:49] <thom> i'm warning all and sundry in README.Debian anyway
[01:02] <Treenaks> smurfix: Psion + timezone trouble.. I have that now on my 3650 :(
[01:02] <Treenaks> ok.. to the nokia center! let's go!
[01:05] <daniels> Kamion: er, what's responsible for creating /etc/default/rcS?  appears to be neither sysvinit nor base-config AFAICT
[01:06] <daniels> Kamion: er, nevermind me
[01:06] <Kamion> prebaseconfig fiddles with it
[01:06] <Kamion> may be something else too
[01:18] <Treenaks> ARGH
[01:22] <daniels> Kamion: initscripts has this beautiful little stanza where it says 'if /etc/default/rcS doesn't exist, then copy it from /usr/share' in postinst
[01:22] <daniels> HELLO CONFFILES
[01:22] <daniels> so I had my way with it
[01:22] <daniels> RCSMD5="$(md5sum /etc/default/rcS | awk '{ print $1; }')"
[01:22] <daniels> if [ "$RCSMD5" = "9d2fb3f6d13039109c02de2092f880f5" -o \
[01:22] <daniels>      "$RCSMD5" = "68ed0d1fdeee47c682f66c335f58b65e" -o \
[01:22] <daniels>      ! -f /etc/default/rcS ] 
[01:24] <thom> it even appears to work, shockingly
[01:26] <Kamion> daniels: is that your code or existing code?
[01:26] <jdub> he found it on dailywtf
[01:26] <Kamion> if it's your code, could you change '-o' to ']  || ['?
[01:26] <daniels> Kamion: my code
[01:26] <daniels> Kamion: oh, is this an XSIism?
[01:26] <Kamion> yep
[01:26] <Kamion> or some other kind of non-POSIXism anyway
[01:26] <Kamion> XSI I think
[01:26] <daniels> brer
[01:27] <daniels> ber, also
[01:28] <ross> holy crap
[01:28] <ross> today's dailywtf is great
[01:37] <thom> ross: the indexed arrays done manually one is genius too
[01:37] <ross> it surely is
[01:39] <edd> jdub: what do you mean, "what happened" ?
[01:39] <jdub> edd: diveintomark.org
[01:40] <edd> ah, i see. sounds like some big argument or other.
[01:40] <jdub> it's been like that for quite a while now
[01:41] <azeem> "Why specs matter
[01:41] <azeem> Most developers are morons, and the rest are assholes."
[01:41] <azeem> heh
[01:49] <zul> heylo
[01:49] <Treenaks> zuul?
[01:49] <zul> yep
[01:57] <lupus_> is it just me
[01:57] <lupus_> or is evolution in hoary broken?
[01:58] <Treenaks> a bit
[01:58] <lupus_> hehe
[01:58] <lupus_> no workaround?
[02:00] <azeem> lupus_: use mutt
[02:01] <lupus_> thunderbird it is :p
[03:13] <pitti> sjoerd: thx for packaging the new hal
[03:13] <pitti> sjoerd: btw, any interest in tarball.mk for this? :-)
[03:13] <sjoerd> decided not to do that, because the amount of patches isn't really worth it imho
[03:14] <pitti> okay
[03:14] <pitti> yeah, it's more important for gvm
[03:14] <sjoerd> and to package it it was basically get new tarball, remove merged patches, build, upload :)
[03:14] <pitti> so again I will steal it from incoming, merge and have it in Ubuntu earlier than in Sid :-)
[03:15] <sjoerd> next time i'll upload just before the archive run stuff :)
[03:18] <pitti> hehe
[03:25] <sjoerd> and commit after the upload ...
[03:26] <bob2> lupus_: if no one has filed a bug yet, please do
[03:37] <pitti> sjoerd: don't bother, I see upload messages earlier than your commits :-)
[03:39] <sjoerd> k
[04:08] <lupus_> bob2, k filled in
[04:09] <bob2> lupus_: thanks!
[04:15] <bob2> omg
[04:15] <bob2> that is actually netscape?
[04:15] <daniels> Kamion: ping
[04:15] <bob2> I thought it was some horrible hack a 3 year old had knocked up
[04:16] <Kamion> daniels: pong
[04:24] <daniels> Kamion: what's the best way to migrate runlevels for an init script?
[04:24] <daniels> i.e. dbus-1 used to be S20, I'd like to make it S12
[04:25] <Kamion> does the old prerm do update-rc.d remove?
[04:25] <Kamion> (or postrm - something called on upgrade anyway)
[04:26] <daniels> Kamion: only on purge
[04:27] <Kamion> hm, in that case I don't know, I'm afraid
[04:27] <Kamion> I've never done this ...
[04:29] <gicmo> sladen, ping (I assume your are not here .. but its worth a shot :)
[04:34] <daniels> Kamion: ok, so just testing for rc2.d/s12dbus-1
[04:34] <daniels> Kamion: and removing if so, and unconditionally adding with 12 later
[04:34] <daniels> Kamion: sound decent?
[04:37] <sladen> gicmo: I am here :-)
[04:37] <gicmo> sladen, ha .. :) .. I cant believe it :)
[04:39] <gicmo> sladen, I am totaly keen on coding on that graphical stuff .. any chance you can send me that code .. or tell me where to help/start
[04:40] <lupus_> daniels, working on making gdm start earlier? :)
[04:40] <thom> yeah, he is
[04:46] <daniels> lupus_: sure am
[04:50] <lupus_> nice :)
[04:51] <Kamion> daniels: can't test for rc2.d/S12* I'm afraid, breaks with file-rc
[04:52] <daniels> Kamion: arse
[04:53] <daniels> Kamion: any ideas on how to migrate it there?
[04:53] <daniels> bbiab
[04:59] <mxpxpod> has anyone tried to run beagle on ubuntu?
[05:00] <zul> yeah didnt compile for me
[05:02] <mxpxpod> zul: how did you get the dbus mono bindings?
[05:02] <zul> followed the instructions in the wiki
[05:26] <daniels> Kamion: for now, I'm just uploading sysvinit/dbus/alsa-driver/gdm with migration logic for sysvinit, we can sort out file-rc later
[05:40] <Kamion> daniels: I'm sure there must be a better way; maybe update-rc.d remove and update-rc.d (whatever, to re-add) with a dpkg --compare-versions check
[05:40] <daniels> Kamion: dpkg --compare-versions is no good if it doesn't make it back to debian
[05:41] <daniels> Kamion: i'm using the update-rc.d remove/update-rc.d (readd) thingy, but the problem is telling how it's already at 20 with file-rc
[05:41] <Kamion> aren't you the Debian maintainer anyway? just do it there ...
[05:50] <daniels> er, neuro maintains gdm in debian :P
[05:50] <daniels> this touches alsa/dbus (-> 12) and gdm (-> 13)
[05:53] <Kamion> oh, I see
[05:53] <Kamion> ask debian-policy@ ?
[05:57] <daniels> yah, think I will
[06:32] <mdz> morning
[06:45] <daniels> mdz: 'morning
[07:01] <fabbione> what a day
[07:02] <Treenaks> what about it?
[07:03] <fabbione> Treenaks: have been really sick all day 
[07:03] <fabbione> 2 computers broke down
[07:03] <fabbione> my gf broke down the car
[07:03] <fabbione> the washing machine exploded
[07:03] <fabbione> not bad for less that 24 hours
[07:03] <daniels> ugh :(
[07:03] <Treenaks> ok.. that qualifies as a Bad Day
[07:03] <fabbione> and i still feel shitty
[07:04] <mako> robtaylor: thanks :)
[07:10] <mdz> fabbione: ouch
[07:11] <fabbione> i am going back to sleep
[07:11] <fabbione> perhaps something will autofix itself during the night
[07:11] <mdz> fabbione: hopefully the bed will not collapse beneath you
[07:11] <fabbione> SHHH
[07:11] <fabbione> don't even think about it!
[07:21] <mako> someone is asking me how to do ubuntu OEM, i don't think i understand what that means
[07:25] <Treenaks> mako: Pre-installed ubuntu?
[07:25] <Treenaks> maybe?
[07:26] <mako> yeah, i think so
[07:26] <Treenaks> Like I can't buy "OEM windows" without buying a complete PC
[07:26] <mako> that is pretty cool i suppose :)
[07:26] <mako> well, i'm not sure what he wants my help with :)
[07:27] <mako> "yes, you install it on a computer before you sell it"
[07:27] <Treenaks> yeah, but we'd probably need a special installation mode that drops you at the "create a user" screen
[07:27] <mako> i guess it's culture shock coming to free software
[07:27] <Treenaks> :)
[07:33] <zul> mako: for most people it is
[07:34] <Treenaks> I had that shock when I first read the GPL back in '96 :)
[07:35] <mako> Treenaks: yeah, i guess i did too.. :)
[07:35] <mdz> mako: what's the word on ubuntu traffic?
[07:35] <Treenaks> "WTF? This is a COOL idea! I want this!"
[07:37] <mako> mdz: UT13 will be done in an couple hours i think.. i'm going to try to get the remaining one out today or tomorrow
[07:37] <mako> mdz: i've read everything for UT13 but haven't finished writing it
[07:38] <Kamion> mako: Jeff and I were talking about that, some kind of "do everything but the user-customisation bits of the install"
[07:39] <mdz> Kamion: yeah, so that on the first boot, you get a few questions and then gdm
[07:41] <Kamion> right
[07:41] <mako> Kamion: yeah, totally that would be interesting
[07:42] <Treenaks> and of course it should be "imageable" or something
[07:42] <Kamion> well, only onto identical disks, but sure
[07:42] <Treenaks> so you can make 600 disks with the same almost-ready-to-use Ubuntu on it
[07:42] <Kamion> not as if we do product activation ;)
[07:42] <Kamion> although ... hmm, hostname
[07:42] <Kamion> that's gonna suck, no netcfg.deb
[07:42] <Kamion> so, depends on post-reboot network config, sigh
[07:43] <Treenaks> do dhcp by default, /change/ the hostname and net config in the "first boot" screen?
[07:44] <Kamion> still depends on post-reboot network config one way or another :-)
[07:44] <Treenaks> uh yes
[07:45] <zul> why not do a wizard when you are doing a first boot, the wizard can do things like detect your soundcard network, users etc
[07:45] <Treenaks> which is needed anyway I think?
[07:45] <Kamion> basically we need to be able to do netcfg-a-like functionality in base-config when it wasn't done pre-reboot (e.g. not netboot)
[07:45] <Kamion> Treenaks: depends on your point of view
[07:45] <Kamion> zul: sound card should be detected by hotplug; if that doesn't work we should just fix hotplug
[07:45] <Treenaks> Kamion: well, say I have my ubuntu PC connected to my (DHCP) cable modem, and I decide I want DSL (PPPoE)
[07:45] <Treenaks> Kamion: then I need to change..
[07:46] <Treenaks> Kamion: so I need a post-install network configurator
[07:46] <Kamion> Treenaks: oh, sure, for reconfiguration you want gnome-system-tools though
[07:46] <Kamion> Treenaks: I'm talking about base-config kinda level here
[07:46] <zul> Kamion: sound card is just an example
[07:46] <Kamion> zul: applies to all hardware
[07:47] <Kamion> zul: note that we already have base-config which is run at first boot and deals with stuff like adding users
[07:47] <Kamion> zul: for OEM, we'd just do most of base-config and leave the job incomplete
[07:47] <Kamion> that's one approach, anyway
[07:48] <Treenaks> the easiest one, probably
[07:48] <Kamion> could even do it with the GNOME frontend come grumpy
[07:48] <Treenaks> (what a name.. grumpy)
[07:59] <mako> only 23 keys for the party in mataro.. i need more!
[08:01] <daniels> elmo: ping
[08:04] <mdz> mako: where did you announce it?
[08:04] <mdz> mako: is it linked from the conference agenda?
[08:05] <mako> mdz: announced on -devel, -users, warthogs
[08:06] <mako> it is on the conference agenda but not linked (yet).. doing that now
[08:08] <lamont_r> so is it safe to upgrade to current hoary?
[08:08] <stratus> daniels, i'll reboot with gdm at S13 can i kill you if it break my ubuntu?
[08:09] <daniels> stratus: you can try if you like ...
[08:09] <daniels> works fine for me
[08:10] <stratus> daniels, i see that you've uploaded ubuntu6 i'll wait for it...
[08:11] <stratus> daniels, was needed only change S99 to S13 ?
[08:12] <daniels> stratus: and dbus and alsa to go to S12 and to change DELAYLOGIN to no in /etc/default/rcS
[08:13] <stratus> oh, i see previous uploads.
[08:13] <stratus> thanks.
[08:58] <bronson_> What are the chances of getting some mountpoints for CF cardreaders added to /etc/fstab?
[08:58] <azeem> bronson_: zero, I'd guess
[08:58] <Treenaks> bronson_: you don't need those
[08:58] <Treenaks> bronson_: unless you boot from them
[08:59] <bronson_> Really?
[08:59] <Treenaks> bronson_: hal/pmount/gnome-volume-manager/nautilus mount them automatically
[08:59] <bronson_> Then I'm doing something wrong...
[08:59] <bronson_> https://www.ubuntulinux.org/wiki/CardReaderHowto
[08:59] <bronson_> Treenaks: wasn't working for me.
[09:00] <Treenaks> bronson_: it is for me.. and lots of other people
[09:00] <bronson_> So I added my cardreader to udev.  Now I get the /dev files.
[09:00] <bronson_> How do I find out what's going wrong?
[09:01] <bronson_> Does udev automatically create the /dev nodes for you?
[09:01] <azeem> bronson_: check hal-device-manager, whether your reader is mentioned there
[09:01] <bronson_> azeem: yes it is.
[09:02] <Treenaks> bronson_: yes
[09:03] <bronson_> Treenaks: can you point me to the rule that creates them?
[09:03] <bronson_> My card reader was being ignored until I added a rule for it.
[09:04] <Treenaks> bronson_: uh.. it's automatic
[09:04] <Treenaks> with the default setup
[09:05] <bronson_> udev requires a rule for everything it creates.
[09:05] <bronson_> How about this question: what node does it create for the CF reader?
[09:05] <bronson_> From that, I could grep the rule.
[09:06] <Treenaks> it's broken atm because it has restarted while booted
[09:07] <bronson_> Does it just create the default /dev/sda1?
[09:07] <bronson_> Or does it create something like /dev/cf1?
[09:08] <bronson_> Since it's USB mass storage, I would guess that it just uses the /dev/sd* nodes...?
[09:11] <bronson_> Does anybody here have a CF reader that works?
[09:12] <bronson_> I'd like to know how it's automounted since most USB CF readers don't issue a media inserted event.
[09:12] <bronson_> Without that, how does the udev node get created?
[09:12] <sjoerd> bronson_: check if you have a /sys/block/sd*
[09:12] <Treenaks> bronson_: I have it working now..
[09:12] <bronson_> sjoerd: yes.
[09:12] <sjoerd> do you also have a sd* subdir in that dir
[09:13] <bronson_> sjoerd: sda1, yes
[09:13] <sjoerd> what happens if you run pmount /dev/sda1 as your user
[09:13] <Treenaks> bronson_: what does "lshal" say?
[09:13] <Treenaks> bronson_: (about linux.sysfs_path_device = '/sys/block/sda')
[09:14] <bronson_> sjoerd: I don't have a /dev/sda1.
[09:14] <bronson_> Treenaks: looking...
[09:15] <Treenaks> bronson_: it shuold give you a huge list
[09:15] <Treenaks> bronson_: important parts:
[09:15] <Treenaks>   storage.automount_enabled_hint = true  (bool)
[09:15] <Treenaks>   storage.media_check_enabled = true  (bool)
[09:15] <sjoerd> bronson_: what's the output of udevinfo -q all -p /sys/block/sda/sda1  
[09:15] <sjoerd> Treenaks: your looking much too high on the stack if that device doesn't exist
[09:16] <bronson_> sjoerd: not much...  3 lines.
[09:16] <Treenaks> sjoerd: if /sys/block/sda exists, it shuold be in lshal
[09:16] <bronson_> pasc: /block/sda/sda1
[09:16] <bronson_> N: cf
[09:16] <bronson_> seb128: 
[09:16] <sjoerd> ah so could you do pmount /dev/cf
[09:17] <bronson_> I'm afraid that's my custom udev rule triggering.
[09:17] <bronson_> /dev/cf is my own invention.  :)
[09:17] <sjoerd> that should just work (tm)
[09:17] <bronson_> I should probably change my rule to get it to create sda1.
[09:17] <sjoerd> bronson_: did the pmount work though ?
[09:17] <bronson_> yes.
[09:17] <bronson_> Excellent.
[09:18] <bronson_> I'll change my udev rule to create the sd* nodes.
[09:18] <sjoerd> bronson_: no, don't
[09:18] <sjoerd> could you tail your .xsession-errors and show the output when you plug in cf stuff
[09:20] <bronson_> sjoerd: nothing.  Did you mean /var/log/messages?
[09:20] <bronson_> Doesn't matter.  Nothing appears in /var/log/message either.
[09:20] <sjoerd> is gnome-volume-manager running ?
[09:20] <bronson_> no?!
[09:21] <sjoerd> :)
[09:21] <bronson_> but gnome-volume-manager still needs a /dev node to do its magic doesn't it?
[09:21] <sjoerd> yeah, but you have /dev/cf right ?
[09:21] <bronson_> Only due to my hacked up udev rule.
[09:22] <sjoerd> doesn't matter
[09:22] <sjoerd> hal knows 
[09:22] <bronson_> So gnome-session is supposed to start gnome-volume-manager?
[09:22] <sjoerd> to get it working extra nice, make it readable by group plugdev
[09:23] <sjoerd> bronson_: btw, this is ubuntu stuff, not ubuntu-devel imho
[09:24] <bronson_> sjoerd: you're right.  It started with a question about /etc/fstab.
[09:24] <bronson_> Sorry bout that.  :)
[09:24] <bronson_> Even with gnome-volume-manager running, nothing happens when I plug/unplug cards.
[09:25] <sjoerd> what does it say in your .xession-errors ?
[09:25] <bronson_> nothing.
[09:25] <sjoerd> is this warty or hoary ?
[09:26] <bronson_> hoary
[09:26] <Treenaks> is dbus-daemon running?
[09:26] <bronson_> 3 of them are.
[09:26] <Treenaks> maybe that's the problem
[09:26] <Treenaks> I have only 1
[09:27] <sjoerd> you should at least have two (one for the system one for the user)
[09:27] <bronson_> At the risk of a flood, here are the 3 lines...
[09:27] <sjoerd> but gvm && hal use the system one, so it doesn't matter
[09:27] <bronson_>  4036 ?        Ss     0:01 /usr/bin/dbus-daemon-1 --system
[09:27] <bronson_>  4867 ?        Ss     0:00 dbus-daemon-1 --fork --print-pid 8 --print-address 6 --session
[09:27] <bronson_>  8990 ?        Ss     0:00 dbus-daemon-1 --fork --print-pid 8 --print-address 6 --session
[09:27] <bronson_> So it appears I have one too many...
[09:27] <Treenaks> 2 session-daemons is a bit much I think..
[09:28] <sjoerd> doesn't matter
[09:28] <bronson_> Criminy.  Something's really wrong on my machine.
[09:28] <bronson_> Does gnome-volume-manager poll?  I'm pretty sure my CF reader doesn't offer media-inserted events.
[09:29] <sjoerd> no, but hal does
[09:30] <sjoerd> try dbus-monitor --system and replug your card.. Should give a lot of output
[09:30] <bronson_> nothing.
[09:31] <sjoerd> that's odd
[09:31] <sjoerd> what does replugging the whole device do ?
[09:32] <bronson_> erm, that's going to be somewhat difficult...  gotta get the screwdriver.
[09:32] <bronson_> Plugging an unrelated USB device showed a bunch of stuff.
[09:33] <bronson_> Is there any way of getting hal to tell us what it's polling?
[09:33] <sjoerd> it polls stuff with storage.media_check_enabled = true
[09:34] <sjoerd> bronson_: btw if sda1 == /dev/cf, then what is sda ?
[09:37] <bronson_> Something just broke.  I can't run lshal anymore.  I'm going to get rid of my udev rule, mv ~/.* dotfiles/ and reboot.
[09:38] <sjoerd> bronson_: is your hal process D ?
[09:38] <mxpxpod> who can I talk to to get pbbuttonsd and powerprefs updated to where debian is?
[09:38] <bronson_> It appears that using the home directory from my old debian system is causing some trouble.
[09:39] <mdz> mxpxpod: https://bugzilla.ubuntu.com/show_bug.cgi?id=3785
[09:39] <bronson_> mxpxpod: https://bugzilla.ubuntu.com/show_bug.cgi?id=3785
[09:39] <mdz> bronson_: :-)
[09:39] <bronson_> Jinx 12345678910
[09:40] <mdz> mxpxpod: there are instructions there so that you can help review the merge, which is the work which needs to be done
[09:41] <mdz> mxpxpod: for powerprefs, it should be synched soon
[09:41] <mxpxpod> mdz: cool
[09:41] <mdz> no merge necessary
[09:41] <mxpxpod> mdz: well, the new powerprefs needs the new pbbuttonsd
[09:45] <mdz> Kamion: what's your feeling on the proposal to seed totem, rather than totem-gstreamer?
[09:46] <Kamion> it would make a lot of user problems go away, I suspect; you'd have to seed both totem and totem-gstreamer to make it unambiguous of course
[09:46] <Kamion> but you've caught me just before going out
[09:46] <mdz> if we seed both totem and totem-gstreamer, it defeats the purpose
[09:47] <seb128> Kamion: totem = totem-gstreamer | totem-xine 
[09:47] <seb128> we can't seed both, they conflict
[09:47] <Kamion> uh-huh, but how's germinate supposed to know which you want?
[09:47] <mdz> the one in main
[09:47] <Kamion> I guess it can randomly pick the first but that seems a bit unstable
[09:47] <seb128> Kamion: the first in the | ?
[09:47] <Kamion> mdz: er ... that's circular, germinate defines main
[09:47] <Kamion> seb128: yeah, I think it does that anyway
[09:48] <Kamion> it just feels ambiguous to me :)
[09:48] <mdz> me too
[09:48] <Kamion> anyway, I'm going out for now
[09:48] <seb128> better solution ?
[09:48] <seb128> have a good evening Kamion :)
[09:48] <mdz> the ideal solution is to have the ubuntu-meta packages use Recommends rather than Depends
[09:49] <Kamion> lamont: I know about the various parted-related build failures BTW
[09:52] <bronson> Wow that was scary.  My home dir just vanished.  Mount showed nothing out of the ordinary, df -h showed normal usage, but /home was empty.
[09:53] <bronson> Banged on reset, booted to single, and everything appears OK agian.
[09:53] <bronson> I'm erasing all my dotfiles and starting afresh.
[10:05] <bluefoxicy> does anyone know what compiler options ubuntu is built with?
[10:16] <pasc> bronson: ?
[10:18] <bronson> pasc: yo
[10:19] <pasc> bronson:  /block/sda/sda1 ?
[10:25] <lamont> Kamion: do I just need to give-back d-i?
[10:29] <mdz> bluefoxicy: it's in the FAQ
[10:29] <mvo> hey mdz
[10:29] <mdz> mvo: hi
[10:30] <mvo> I got a replay from daniel burrows today. he said he may have time to work on apt-secure support for aptitude in feb/march 
[10:30] <mvo> he's working on his thesis right now
[10:30] <bluefoxicy> mvo: http://www.ubuntulinux.org/support/documentation/faq/optimised-packages is not sufficient information.
[10:30] <mdz> mvo: ok, so I guess we'll need to do it
[10:31] <bluefoxicy> err
[10:31] <bluefoxicy> mdz, not mvo
[10:31] <mvo> mdz: all right, I'll have a look
[10:32] <bluefoxicy> I'm on amd64, and I use (aside from proper -march) "-O2 -pipe -ftracer  -fweb -funit-at-a-time -fomit-frame-pointer -mfpmath=387 -mno-sse" (Gentoo, gcc 3.4)
[10:32] <bluefoxicy> I'm curious about what optimizations are used aside from just optimizing for a specific CPU
[10:34] <mdz> -O2
[10:34] <bluefoxicy> k
[10:34] <mdz> lamont: ping
[10:34] <bluefoxicy> -fweb and -funit come from -O3, but I avoid -O3 due to the way it interacts with stack smash protection (it optimizes the protection out)
[10:35] <bluefoxicy> hmm
[10:35] <bluefoxicy> "A comprehensive experiment to evaluate the effectiveness of more aggressive processor optimisations for all Ubuntu packages is planned after the Warty release."
[10:35] <lamont> mdz: yo
[10:35] <bluefoxicy> I'm interested in detailed results from this experiment
[10:35] <lamont> mdz: actually, the only flag we force is -pipe
[10:36] <lamont> and we allow -O0 through -O3
[10:36] <lamont> -O6 will get you -O3
[10:36] <lamont> -march=486 -mcpu=pentium4
[10:37] <bluefoxicy> ah
[10:37] <mdz> lamont: Debian policy is -O2
[10:37] <mdz> and we inherit that
[10:39] <lamont> mdz: unless the package overrides that
[10:39] <lamont> we don't force that
[10:39] <lamont> broke too many things.
[10:39] <bluefoxicy> lamont: Lorenzo has been working on altering the debian tools and such to allow building things with stack smash protection, among other things, to produce a more secure distro.  Much of this is suitable for general consumption (i.e. suitable for Ubuntu), but requires some extra considerations
[10:39] <bluefoxicy> for example, -O3 will still break SSP AFAIK
[10:40] <bluefoxicy> (it optimizes away the code that checks for buffer overflows)
[10:40] <mdz> lamont: right
[10:40] <bluefoxicy> lamont:  in the future, would it be possible to force -O2 on -O3 packages if Ubuntu had taken such enhancements as SSP?
[10:41] <mdz> I've updated the FAQ entry to be a bit more specific in that regard
[10:41] <mdz> bluefoxicy: it would be inadvisable
[10:41] <mdz> oh, you mean force it down to -O2 if the package tries to use -O3?
[10:41] <mdz> yes, that would be possible
[10:42] <bluefoxicy> yes
[10:42] <mdz> very little builds with -O3, though, and where it does, it's probably the right choice
[10:44] <trulux> hey guys
[10:46] <bluefoxicy> hey
[10:46] <bluefoxicy> mdz, lamont:  this is the guy that's been working on the modified debian tools
[10:47] <trulux> bluefoxicy, i'm doing some jobs with the libssp, give me a few minutes to make a debian sarge-ready pkg
[10:47] <trulux> and provide some more stuff
[10:47] <bluefoxicy> ok
[10:47] <trulux> the rest of things are done and ready
[10:47] <lamont> trulux: does it handle both up and down stack growth checking?
[10:47] <bluefoxicy> trulux:  as a proof of concept, a glibc package should be workable, but if you can get libssp up in a timely manner, that's good :)
[10:47] <lamont> that's been a shortfall of many of the stacksmash checks...
[10:48] <bluefoxicy> lamont:  http://www.trl.ibm.com/projects/security/ssp/main.html
[10:48] <lamont> bluefoxicy: trivial to force -O3 to turn into -O2
[10:48] <Sturzflut> it does not support that
[10:48] <bluefoxicy> lamont:  *nod*
[10:48] <bluefoxicy> I must go now
[10:48] <bluefoxicy> being called out.
[10:48] <trulux> bluefoxicy, timely manner means less than 20/30 minutes? sure ;D
[10:48] <Sturzflut> like you have on hppa
[10:48] <bluefoxicy> side note
[10:48] <lamont> the only wierdness there is that (for purposes of checking) -Os is treated identicallyo -O2
[10:48] <bluefoxicy> growing up is a trivial issue security wise
[10:49] <bluefoxicy> since stack smashes can't affect important data in *most* (there ARE CASES!) cases
[10:49] <trulux> lamont, i'm have some patches for GCC that would be nice to be used inside ubuntu gcc packages
[10:49] <bluefoxicy> at least, to my understanding.
[10:49] <Sturzflut> stacks that grow up can't overwrite the return address can they?
[10:49] <lamont> bluefoxicy: buffers in the current frame passed to a function that then overruns the buffer will result in arbitrary return points from the called function
[10:49] <lamont> that is, it's not really any different from the stack growing down.
[10:50] <Sturzflut> actually, I guess you could
[10:50] <Sturzflut> lamont: do you have access to such hardware?
[10:50] <lamont> Sturzflut: I've only ever written two exploits for hppa, overwritting the stack with code and rp.
[10:50] <lamont> Sturzflut: have hardware
[10:51] <Sturzflut> I don't think the SSP patch does hppa yet but I could be wrong about that
[10:51] <lamont> I'm also aware of a case where hp-ux was exploited writing only the return pointer, not executing code from the stack... (they found a useful chunk of code already in place...)
[10:51] <Sturzflut> right
[10:51] <lamont> Sturzflut: of course, that's not an ubuntu discussion.  yet... :-)
[10:51] <Sturzflut> heh
[10:52] <Sturzflut> I would think Ubuntu would focus where the money is first :)
[10:52] <lamont> both exploits I wrote were in response to R&D labs saying "no one could exploit that vul"
[10:52] <lamont> the first one took me about 3 hours, since I wrote some tools.  The second one took about an hour.
[10:52] <lamont> Sturzflut: that's a low blow... :[)
[10:52] <Sturzflut> that's why you should never say never
[10:52] <lamont> :-)
[10:54] <lamont> yeah - the discussion then went "but you have an exceptional knowledge of the architecture" to which I replied, "yes, that's why it only took me an hour - remember that the folks writing the attacks have _NOTHING_BETTER_TO_DO_.."
[10:54] <lamont> "oh".
[10:55] <Sturzflut> ha
[10:56] <Sturzflut> SSP is a good idea for things like IRC and IM clients
[10:56] <lamont> then they got mad when I wouldn't give them the exploit code.
[10:56] <Sturzflut> since they seem to have lots of exploits :)
[10:56] <|trey|> Hey, who deals with site accounts?  I don't recall my password, but its saying the account is not valid, however is saying that it does exist... lunitik@gmail.com, if possible, please send me my account info...
[10:56] <Sturzflut> and for many daemons too
[10:56] <lamont> Sturzflut: yeah.  shoddy coding is the real issue.
[10:57] <mdz> seb128: gnome-volume-manager just crashed on me after a hal/dbus upgrade; is this a new bug, or did the old one get un-fixed?
[10:57] <Sturzflut> the only cure for bad programmers is a pink slip, but that doesn't work for OSS
[10:57] <seb128> mdz: old one not fixed imho
[10:58] <mdz> oh, I thought that bug was fixed back in Warty
[10:58] <lamont> Sturzflut: actually, embarassment helps more with OSS.
[10:59] <Sturzflut> I think I agree with that
[10:59] <|trey|> No one can help with the site issue?
[11:00] <lamont> |trey|: I can't remember which person is the vicitm for site issues...
[11:00] <|trey|> s/the/my/
[11:00] <Sturzflut> I wish Debian had more focus on security than they have now
[11:00] <Sturzflut> they seem to be too focused on dealying sarge release
[11:00] <lamont> but if my guess is correct, he should be around in a couple of hours or so...
[11:00] <lamont> or do you mean bugzilla?
[11:00] <Sturzflut> delaying
[11:00] <|trey|> lamont: ugh... are you able to look through the accounts and send a request for info at least?
[11:00] <Mithrandir> Sturzflut: that's a holdup which is due to a small number of people being able to do the magic needed for the freeze to start.
[11:01] <Sturzflut> Mithrandir: the security stuff, isn't it?
[11:01] <Sturzflut> security infrastructure
[11:02] <|trey|> lamont: like I said, when I request, it says the account isn't valid... but when I try to re-sign up, it says the account already exists... what would the issue be do you think?
[11:02] <Sturzflut> Mithrandir: the DPL should get the whip going, imho
[11:02] <|trey|> lamont: I will be leaving soon... I suppose I could wait till later...
[11:03] <Mithrandir> Sturzflut: I'm not sure, it's both the security infrastructure and the autobuilders for testing-proposed-updates.. I know a lot of people would just like to release as soon as possible, even without security support in place for testing.
[11:06] <lamont> |trey|: I have no admin access
[11:07] <Sturzflut> Mithrandir: I don't understand why those with access haven't delegated these tasks to the willing
 bluefoxicy: buffers in the current frame passed to a function that then overruns the buffer will result in arbitrary return points from the called function
[11:07] <bluefoxicy> lamont:  true, I said it was less likely, not impossible
[11:07] <bluefoxicy> lamont:  protection is possible though, by guarding functions that are passed char* pointers.  I don't think SSP does this; but I'll ask etoh
[11:08] <Sturzflut> if it's possible, it will be done
[11:08] <bluefoxicy> (once one function is not passed a pointer to a buffer, no further called function can get that buffer, unless it's stored in a global variable, which is already brainfuck)
 bluefoxicy, timely manner means less than 20/30 minutes? sure ;D
[11:10] <bluefoxicy> trulux:  I meant more like within the next few weeks but the next few minutes is good too  :D
[11:11] <trulux> bluefoxicy, libssp is done
[11:11] <trulux> gcc needs to be recompiled, and here i can talk a bit to ubuntu devs if they are proud ;-)
[11:12] <lamont> trulux: "if they are proud"?
[11:13] <trulux> it's just a joke
[11:13] <trulux> lamont, in order to use the SSP first you must decided what to changem the glibc, the libgcc or use a libssp
[11:13] <trulux> libgcc is crappy
[11:14] <Sturzflut> heh
[11:14] <trulux> ssp inside glibc has some advantages, better performance and also backwards compatibility if you change to libssp
[11:14] <Sturzflut> __guard symbols used for SSP can be in three places
[11:14] <trulux> libssp has some advantages like more soft implementation and easier use
[11:14] <trulux> stuNNed, yep
[11:14] <Sturzflut> the default for the protector patch is inside of libgcc
[11:14] <trulux> oops, nick autocompletion
[11:14] <trulux> Sturzflut, yes
[11:15] <Sturzflut> libc is best place for it
[11:15] <lamont> ah, ok
[11:15] <Sturzflut> having it in libgcc can cause problems
[11:15] <trulux> lamont, but make it inside glibc is difficult if maintainers are not "open" for this stuff
[11:15] <trulux> Sturzflut, s/can/will/
[11:16] <Sturzflut> heh, but glibc is already a bloated mass, I don't see why one more thing is going to hurt :)
[11:17] <trulux> yep
[11:17] <trulux> Sturzflut, pappy said one thing that is truth and important: think before doing things
[11:17] <Sturzflut> they should add addition fixes to glibc too
[11:18] <trulux> imagine we used ssp inside libgcc since the project start, now we could be fscked up
[11:18] <trulux> Sturzflut, i maintain my own glibc with all of that stuff
[11:18] <Sturzflut> not really, a simple recompile of all packages fixes that
[11:18] <trulux> hopefully waiting for somebody to test and package it
[11:18] <Sturzflut> all C packages that got SSP that is
[11:18] <lamont> trulux: the rules are: (1) never get involved with someone more messed up than you are, (2) think _BEFORE_ you act, and (3) if it shakes the house (and you don't own the house) get permission before doing it.
[11:19] <Sturzflut> heh
[11:19] <trulux> lamont, yeah
[11:19] <Sturzflut> you forgot (4) Always have an exit plan
[11:19] <Sturzflut> when the ship goes down, I won't be on it
[11:19] <trulux> and also the 5) Ignore the rest of rules and disappear
[11:20] <lamont> Sturzflut: s/exit/backup/
[11:20] <trulux> &) if you can't disappear just change your name into something else more difficult to guess: Dead Beef
[11:20] <trulux> :)
[11:21] <Sturzflut> naw, I really did mean exit plan
[11:21] <lamont> heh
[11:21] <Sturzflut> applies to the corporate world well
[11:22] <Sturzflut> when you see things are about to go bad, give yourself a bonus and retire :)
[11:23] <trulux> http://www.research.ibm.com/vali/
[11:23] <trulux> Sturzflut, that's for my regression tests suite
[11:25] <Sturzflut> Vali sounded familiar, but I don't think I've seen that project before :)
[11:32] <bluefoxicy> well I must go to spanish class
[11:32] <trulux> bluefoxicy, suerte ;-)
[11:33] <bluefoxicy> If possible, i'd love to see a developer's release of Hoary (before public release time, not side by side ;) that has SSP in it so that I can throw it at my laptop and check for regression :)
[11:33] <bluefoxicy> trulux:  suerte?
[11:34] <trulux> good luck
[11:34] <bluefoxicy> ah thanks
[11:34] <trulux> bluefoxicy, i'm going to add SSP regression test to my rts
[11:35] <bluefoxicy> trulux:  the regressions I was most worried about are programs that overflow buffers by 2 or 3 bytes (not fatal, but they DO set ssp off) on their own
[11:35] <bluefoxicy> and things like someone accidentally using -fstack-protector-all instead of -fstack-protector (-all has bugs that break things like Mozilla)
[11:37] <trulux> i see
[11:37] <trulux> i can implement that easily in my regression t. suite
[11:39] <bluefoxicy> not really; it's per-program specific (except the -all one, which if you can find what causes it would be great; etoh could fix it then :)
[11:41] <trulux> i was meaning about a simple SSP regression test
[11:41] <trulux> did if SSP is alocated in the toolchain