#emc-devel | Logs for 2009-04-20

[02:16:16] <cradek> holy crap, a release
[02:17:55] <cradek> yay!
[02:18:10] <cradek> thanks to everyone who worked on this. very cool.
[02:18:48] <jmkasunich> seconded
[02:18:52] <fenn> * fenn fires an ak-47
[02:19:02] <jmkasunich> now I can upgrade the shoptask without using a beta ;-)
[02:19:19] <jmkasunich> or, I can once the dapper packages are built
[02:21:01] <cradek> I'll get to it shortly
[02:21:06] <cradek> hope that machine is on...
[02:21:52] <jmkasunich> didn't realise it was you that built those
[02:24:34] <cradek> yay, it's on
[02:33:19] <cradek> jmkasunich: packages ready
[02:34:02] <jmkasunich> thanks
[02:34:12] <jmkasunich> (will do it tomorrow, getting late today)
[11:51:28] <jepler> * jepler kicks his DSL provider
[11:51:33] <jepler> .. cvs service should now be restored
[12:52:57] <cradek> jepler: working on the missing -all now
[12:55:50] <cradek> fixed, I think
[12:58:30] <cradek> I guess nobody with mesa (or maybe nobody at all) tried the betas in dapper
[12:59:00] <skunkworks> cradek: how was names?
[12:59:51] <cradek> nice - I got there for just the last several hours
[13:00:30] <cradek> dale and greg's setup was really nice, and dale was good at answering questions and showing people things.
[13:04:33] <skunkworks> That is great
[13:08:10] <cradek> while I was sitting there someone came up and asked: I don't know anything about cnc or computers. How hard will it be for me to learn EMC?
[13:08:44] <cradek> while I was stumbling around explaining why it's impossible for me to give a meaningful answer to that question, Dale intercepted and said "let me show you it" and did a little demo
[13:09:52] <cradek> so, as you can see, he's much better at it.
[13:20:28] <skunkworks> heh
[14:20:35] <christel> [Global Notice] Hi all, I'm going to need to take services down for a few minutes -- this is to let you know that the network will be without Chan, Nick, MemoServ et al for the next few minutes. They'll be back up again soon and apologies for the inconvenience.
[16:57:29] <skunkworks> alex_joni: is the iso anywhere yet?
[16:57:47] <alex_joni> skunkworks: aether
[16:58:29] <skunkworks> aether?
[16:59:12] <alex_joni> the personification of the "upper sky", space and heaven
[16:59:28] <ehj> alex_joni: Did you see my post on emc-developer?
[16:59:34] <skunkworks> ah - I don't think I have actually seen that word in writing.. ;)
[17:00:16] <alex_joni> ehj: reading it now
[17:01:36] <alex_joni> ehj: right
[17:01:50] <ehj> k
[17:01:54] <alex_joni> ehj: debian/control needs to reference rtaismp
[17:02:01] <alex_joni> it probably doesn't right now for you
[17:02:24] <alex_joni> simply changing debian/control won't work, because it's generated from debian/control.in (or something like that)
[17:02:39] <alex_joni> debian/control.in is also generated by reading the flavours you have defined
[17:02:59] <alex_joni> hang on 5 mins to boot my hardy box, and I'll find the name of the files you need to change
[17:03:19] <ehj> k, thanks
[17:03:21] <alex_joni> (or you can use grep -r -e rtai debian/*, and see where rtai is, and rtaismp isn't)
[17:03:33] <ehj> k
[17:06:12] <ehj> Not particularly helpful, it returned a couple thousand lines. :)
[17:07:59] <alex_joni> back
[17:08:33] <ehj> k
[17:08:45] <alex_joni> ok.. so the main file is debian/control
[17:08:52] <alex_joni> which gets generated from debian/control.stub
[17:09:11] <alex_joni> and the control.stub file gets generated by reading rules.d/i386.mk
[17:09:13] <ehj> Yes, I have it open. It is fairly large.
[17:09:19] <alex_joni> and rules.d/0-common-vars.mk
[17:09:33] <alex_joni> you only care about the last 2 mk's
[17:09:42] <ehj> I modofied the last two as you can see from the tutorial link,
[17:09:53] <alex_joni> can you link that again?
[17:10:00] <alex_joni> it stayed on my windows box :)
[17:10:07] <ehj> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC_With_Custom_Kernel
[17:10:54] <alex_joni> that looks ok
[17:11:10] <alex_joni> did you do debian/rules debian/control.stub ?
[17:11:17] <alex_joni> and debian/rules debian/control ?
[17:11:27] <ehj> yes
[17:12:30] <alex_joni> * alex_joni suggest doing that until debian/control.stub contains rtaismp
[17:12:55] <ehj> First time complained about kernel-wedge not being installed. I then installed it and it seemed to run properly.
[17:13:36] <alex_joni> right, it also builds udebs
[17:13:40] <alex_joni> although you don't need those
[17:15:28] <ehj> running debian/rules debian/control.stub says make ... is up to date.
[17:16:06] <alex_joni> then grep control.stub for rtaismp to see if it's there
[17:16:28] <ehj> running debian/rules debian/control lists a script ending with kernel-wedge gen-control > debian/control
[17:17:22] <ehj> No, not found.
[17:18:35] <alex_joni> can you paste the command you used?
[17:19:03] <ehj> I opened control.stub i gedit and did a search for rtaismp
[17:19:09] <ehj> *in
[17:21:43] <ehj> I searched control.stub for just rtai to see if I made a typo in one of the other files. All I see are references to rtai, none to rtaismp.
[17:22:34] <alex_joni> then probably you made a typo in one of the mk's
[17:22:44] <alex_joni> open a terminal, go to debian/ and write"
[17:22:49] <alex_joni> grep -r -e rtaismp *
[17:22:52] <ehj> Yea, that is what I was thinking, checking now.
[17:27:48] <ehj> Well that returns many lines, but rechecking i386.mk and 0-common-vars.mk in rules.d I see no typos.
[17:28:53] <ehj> Running grep from rules.d shows just the two instances.
[17:31:23] <alex_joni> I only get 64 for rtai
[17:31:30] <alex_joni> for rtaismp you should get < 10
[17:32:43] <ehj> I am getting hundreds to thousands.
[17:33:09] <alex_joni> in debian/ ?
[17:33:12] <ehj> Recall that the better part of the build did run.
[17:33:19] <ehj> yes
[17:33:24] <alex_joni> debian/rules clean :)
[17:34:49] <ehj> I get an error, "you must run this a root (or use fake root)".
[17:35:11] <ehj> dh_testroot
[17:35:12] <alex_joni> I'm sure it says (or use fakeroot)
[17:35:31] <alex_joni> fakeroot debian/rules clean
[17:35:32] <ehj> How do I use fakeroot?
[17:35:35] <ehj> k
[17:36:00] <alex_joni> ehj: I thought you built these packages already?
[17:36:54] <ehj> I am building them again for 2.3. Plus I wanted to check the tutorial to see that I really did have the steps right. Obviously I do not.
[17:37:14] <alex_joni> 2.3 has nothing to do with the kernel packages
[17:38:03] <ehj> I only got two instances of rtaismp after rules clean.
[17:38:14] <alex_joni> those 2 mk's
[17:38:15] <ehj> The two I should.
[17:38:18] <ehj> ys
[17:38:19] <alex_joni> right
[17:38:19] <ehj> yes
[17:38:25] <alex_joni> and next you should get about 50-60
[17:38:34] <alex_joni> after a proper control.stub and control have been generated
[17:38:56] <ehj> I did not think I needed to build the kernel packages again, but I wanted to go through the steps for the tutorial.
[17:39:09] <ehj> k
[17:40:29] <ehj> No, I just get the same two.
[17:40:45] <alex_joni> ehj: maybe try removing those
[17:40:49] <alex_joni> then generate them again
[17:40:57] <ehj> k
[17:43:13] <ehj> By generate them again, do you mean re-edit or run debian/rules... agains the unmodified mk files?
[17:43:30] <ehj> *against
[17:44:12] <alex_joni> when you run debian/rules debian/control.stub
[17:44:21] <alex_joni> a new control.stub file is generated
[17:44:46] <alex_joni> you never edit the control.stub file itself, it gets generated
[17:44:59] <ehj> BTW, when I run debian/rules debian/control, is it supposed to display the entire script, including things like "if" statements?
[17:45:08] <alex_joni> no
[17:45:21] <alex_joni> obviously you need to be one directory hierarchy up
[17:45:24] <alex_joni> not in debian/
[17:45:26] <ehj> I did not think so, but that is what I am getting.
[17:45:50] <ehj> k, just a sec. Starting over.
[17:48:29] <ehj> Ok, from the base folder "ubuntu-hardy" I run debian/rules debian/control.stub and it responds "make: 'debian/control.stub' is up to date.
[17:49:34] <alex_joni> even if the file doesn't exist?
[17:49:43] <ehj> Then I run debian/rules debian/control and get the text of the script displaying.
[17:49:58] <alex_joni> forget debian/control
[17:50:03] <ehj> Which file?
[17:50:08] <alex_joni> until control.stub isn't ok, no need to go further
[17:50:42] <ehj> k, why isn't it running properly?
[17:52:15] <alex_joni> ehj: I have no idea, you're at the machine
[17:52:32] <alex_joni> I can only tell you how it's supposed to work
[17:52:43] <alex_joni> you'll need to figure out why it doesn't
[17:52:55] <ehj> What am I looking for, an file not edited properly or a package not installed?
[17:52:59] <alex_joni> do you get any 'rtai' in the control.stub ?
[17:53:07] <alex_joni> this has nothing to do with packages
[17:53:28] <alex_joni> it (debian/rules debian/control.stub) simply reads a couple mk files, and generates debian/control.stub
[17:53:51] <ehj> rtai yes, rtaismp no.
[18:00:10] <alex_joni> can I see (on pastebin) the output from "grep -r -e rtaismp debian/*" ?
[18:08:09] <ehj> yea, just a sec, I got a phone call.
[18:10:01] <ehj> Right now it is just the two lines I just put back into 0-common-vars.mk and i386.mk
[18:10:41] <alex_joni> then paste them in here
[18:12:08] <ehj> Can't, it is on a different machine with nothing set up.
[18:13:21] <alex_joni> then check for spelling spaces, etc
[18:13:50] <ehj> http://pastebin.com/m15891514
[18:14:47] <alex_joni> I'm surprised debian/binary-custom.d/rtaismp doesn't come up in the grep
[18:16:28] <ehj> I just checked to see if I did something stupid like hit a cap. But it is there, spelled correctly and in the correct case.
[18:25:27] <ehj> Here is the folder listing: http://pastebin.com/m6c42046b
[18:26:06] <alex_joni> ehj: looks ok from here
[18:26:32] <ehj> Yea, if I made a typo some place I sure don't see it.
[18:27:58] <ehj> I could do a touch, but the dates all look good too.
[18:31:11] <alex_joni> simply removing debian/control.stub
[18:31:26] <alex_joni> then generating it with debian/rules debian/control.stub should be all you need to do
[18:32:38] <ehj> k
[18:33:11] <ehj> What about control.stub.in?
[18:35:44] <ehj> k, it did it and there is now a reference to rtaismp
[18:36:09] <alex_joni> control.stub.in is the original file, which you never edit or change
[18:37:15] <ehj> k, No difference when making debian/control, should I delete that too?
[18:38:23] <alex_joni> you can try
[18:39:10] <jepler> fwiw this patch shows all the differences from the original ubuntu kernel to the rtai kernel package I built, perhaps something in it will cause a bell to ring.
[18:39:14] <jepler> http://emergent.unpy.net/files/sandbox/ubuntu-hardy-rtai.patch.bz2
[18:40:04] <ehj> I did. There are now references to rtaismp in control as well. However it is still dumping the text of the script.
[18:40:17] <alex_joni> ehj: don't mind that
[18:40:19] <ehj> Should I try 'build.sh'?
[18:40:25] <ehj> k
[18:40:30] <alex_joni> yup
[18:40:50] <ehj> talk to you in an hour or so. :)
[18:43:09] <alex_joni> ok
[18:47:28] <alex_joni> come ON!!! this isn't fair:
[18:47:30] <alex_joni> 360837 extents written (704 MB)
[18:51:56] <jepler> hah
[18:51:57] <jepler> poor alex_joni
[18:56:35] <jepler> hm, that's about 1 cdrom + 1 floppy
[18:56:36] <alex_joni> hmm.. this is odd
[18:56:48] <jepler> perhaps .. nah, forget it
[18:56:50] <alex_joni> emc2: Depends: libpango1.0-0 (>= 1.20.5) but 1.20.1-1 is to be installed
[18:57:08] <alex_joni> jepler: found an initrd.back which is ~7MB, so maybe that's enough
[18:57:15] <alex_joni> but I notice it didn't want to upgrade emc2
[18:57:27] <jepler> yeah, you'll have to install the newer libpango
[18:57:39] <alex_joni> http://packages.ubuntu.com/hardy/libpango1.0-0
[18:57:44] <alex_joni> is there a newer one?
[18:57:55] <jepler> http://packages.ubuntu.com/hardy-updates/libpango1.0-0
[18:58:15] <alex_joni> ah, from updates
[19:01:06] <alex_joni> fsck, another 259 upgrades from hardy-updates :/
[19:01:22] <jepler> oh, excuse me while I laugh at your expense
[19:01:28] <alex_joni> and another 10.3MB diskspace :/
[19:01:42] <alex_joni> yeah, yeah.. rub it in
[19:02:13] <alex_joni> jepler: lets do it like this, I'll be the release manager for 2.4, and you'll build the livecd :P
[19:02:13] <jepler> it's a thankless task, so I'm not going to say "thank you".
[19:02:25] <alex_joni> * alex_joni is only kidding
[19:02:38] <alex_joni> it's not such a pita.. I just like to complain a bit :)
[19:07:04] <jepler> hm, I wonder how to get dpkg or apt to purge old configuration files
[19:07:43] <jepler> nearly 2 megs in /etc/emc2/sample-configs, for example
[19:07:51] <jepler> I think an upgrade to 2.3 leaves them
[19:12:12] <alex_joni> yeah, I'll remove them by hand
[19:13:20] <alex_joni> jepler: but that's before it's compressed, and since it's mostly text I doubt it's more than 1-200 kB
[19:49:35] <ehj> alex_joni: FYI, looks like the build worked this time. Not sure what I did wrong. I will add to the tutorial to just delete those two files (control.stub and control).
[20:02:24] <alex_joni> sounds good
[20:51:12] <cradek> hi all
[20:51:37] <alex_joni> hi chris
[20:51:53] <cradek> has anyone tried the dapper install mesa problem yet?
[20:52:29] <alex_joni> hmm.. not me, but I could in a "shortish" while
[20:52:41] <jepler> cradek: I don't think it's fixed
[20:52:51] <cradek> dang.
[20:53:19] <jepler> http://pastebin.ca/index.php
[20:53:21] <jepler> argh
[20:53:26] <jepler> http://pastebin.ca/1397681
[20:53:32] <cradek> I'll look again tonight.
[20:53:35] <jepler> ^^ list of emc2-related packages I have after updating
[20:54:02] <cradek> can't look at that very easily now.
[20:54:55] <jepler> understood
[20:54:58] <jepler> are you on the road tomorrow too?
[20:55:26] <cradek> hopefully back tonight
[20:56:04] <cradek> brb
[21:03:19] <alex_joni> yay, it fits :)
[21:10:07] <cradek> the lady in the next booth reports that she doesn't have enough red blood cells.
[21:10:24] <alex_joni> email some
[21:11:24] <ehj> alex_joni: I am up to building rtai and am getting an unmet dependency for linux-headers-2.6.24-16-rtai which does not exist, but linux-headers-2.6.24-16-rtaismp does. I see from my notes that last time you had me do "sudo apt-get build-dep linux-image-2.6.24-23-generic", which I did, but it made no difference.
[21:17:26] <alex_joni> you need to edit rtai
[21:17:38] <alex_joni> especially the debian/control part
[21:17:55] <alex_joni> change 2.6.24-16-rtai with 2.6.24-16-rtaismp
[21:18:40] <alex_joni> also in debian/rules there's a KERNEL_VER=
[21:19:15] <ehj> I got the debian/rules. I don't have any other changes in my notes.
[21:20:33] <ehj> I don't see a debian/control under rtai.
[21:21:07] <ehj> Never mind, wrong folder.
[21:23:45] <ehj> That got it.
[21:55:57] <alex_joni> jepler: I just had a thought
[21:56:19] <alex_joni> we don't really need the devel tools on the CD itself, right?
[21:56:48] <alex_joni> we could just specify the packages needed in the desktop manifesto, and they "should" be installed during the process
[21:57:08] <alex_joni> (although on PCs without network access nothing changes :?)
[21:57:41] <jepler> alex_joni: the point is to help people who don't have network access
[21:58:03] <alex_joni> yeah, just realized that... it's probably a bit late here :)
[21:58:31] <alex_joni> although I noticed build-essential and gcc are on the CD
[21:59:03] <jepler> oh really? I didn't know that.
[21:59:21] <alex_joni> they are in the pool/ included on the CD (not in the LiveCD)
[21:59:33] <alex_joni> so you need to apt-cdrom add, then you can apt-get install them
[22:00:44] <alex_joni> it's g++-4.2_4.2.3-2ubuntu7_i386.deb
[22:00:54] <alex_joni> and matching libstdc++
[22:01:03] <jepler> man it's windy out there
[22:01:22] <jepler> does that take less space than installing them in the filesystem? interesting.
[22:01:37] <alex_joni> I would expect it takes up more
[22:01:43] <alex_joni> it's not compressed
[22:02:20] <alex_joni> odd things included there.. like fakeroot and dpkg-dev
[22:54:57] <alex_joni> if anyone else wants to take a look at the livecd it's on www.linuxcnc.org/hardy/
[22:55:18] <alex_joni> (note: haven't tried it yet myself, except in a qemu)