#emc-devel | Logs for 2009-04-21

[00:47:09] <cradek> home!
[01:20:31] <jmkasunich> yay
[02:06:15] <jepler> cradek: yay
[02:06:43] <jepler> cradek: have time to take a look at the dapper package thing?
[02:14:19] <jmkasunich> jepler: I have an axis question
[02:14:31] <jmkasunich> there are some special g-code comments that axis understands
[02:15:05] <jmkasunich> are those only "understood" when axis is generating the preview, or does axis see them during program run too?
[02:16:58] <jepler> jmkasunich: when generating the preview
[02:17:20] <jepler> for instance, I use (AXIS,hide) / (AXIS,show) to hide tool changes and tool length probes in the preview .. but they show up in the backplot
[02:18:52] <jmkasunich> darn
[02:19:20] <jmkasunich> I was hoping that you could add one (AXIS, clear), which would do the equivalent of ctrl-K
[02:19:42] <jmkasunich> for when you are running the same program dozens of times in a row - the backplot buffer will get full
[02:19:55] <jepler> that's an interesting idea
[02:20:35] <jmkasunich> I've got a job coming up that will be 48 pcs
[02:20:53] <jmkasunich> and hopefully after the first few I won't be tweaking the g-code and reloading all the time
[02:22:26] <jepler> I can make axis-remote able to clear the backplot, and then you could use an M1xx to do it. For 2.3, you could directly incorporate the magic in the M1xx script instead of calling axis-remote.
[02:22:30] <jepler> * jepler codes and tests
[02:22:43] <jepler> there's also a constant you could change somewhere to modify the max backplot size..
[02:24:14] <jmkasunich> the size fix isn't as elegant
[02:24:33] <jmkasunich> it would be nice to clear the plot at the beginning of each run, so you can clearly see the progress on the part
[02:25:02] <jmkasunich> and maybe clear it at other key points, like when roughing is done and the finish passes are about to start
[02:26:42] <jepler> http://emergent.unpy.net/index.cgi-files/sandbox/axis-remote-clear.patch http://emergent.unpy.net/index.cgi-files/sandbix/axis-clear.tcl
[02:29:03] <jmkasunich> one or the other of those, not both, right?
[02:29:37] <jepler> right
[02:29:54] <jepler> the first I'm going to commit on TRUNK, the second is probably better to use on 2.3 particularly if you want to use the packaged version
[02:31:57] <jmkasunich> given that the second one is 3 lines and works on existing versions, and the first is much more complex, why would the first one ever be preferred?
[02:32:22] <CIA-2> EMC: 03jepler 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis-remote.py: add axis-remote --clear to clear the backplot
[02:32:22] <CIA-2> EMC: 03jepler 07TRUNK * 10emc2/docs/man/man1/axis-remote.1: add axis-remote --clear to clear the backplot
[02:32:40] <jepler> because 'axis-remote --clear' will be documented and intended to work forever
[02:32:46] <jepler> the other is touching what I consider implementation details
[02:32:54] <jmkasunich> ah, and the tcl hack might break someday
[02:33:09] <jepler> (the exact name of the command to do that thing might change if there's a good reason in the future .. not likely, to be honest, but possible)
[02:33:34] <jmkasunich> the Mxx script for the portable version would just be one line? "axis-remote --clear"
[02:34:13] <jmkasunich> well, plus the "#! /usr/bin/bash" part
[02:34:20] <jepler> yeah pretty much
[02:38:26] <jmkasunich> thanks
[02:41:46] <jepler> 'night
[02:41:51] <jepler> have fun making chips
[02:59:11] <cradek> any dapper folks around to help me try to fix this?
[03:02:43] <cradek> hm, I still don't see a problem.
[03:03:04] <cradek> binary-all has the Packages files, and those are listed in Release
[06:52:27] <fenn_> fenn_ is now known as fenn
[08:50:19] <fenn_> fenn_ is now known as fenn
[12:10:18] <jepler> cradek: I am now
[12:10:23] <jepler> cradek: http://pastebin.com/m607995eb
[12:11:02] <jepler> I don't see what is wrong either -- the -all is listed in my Release file, but no corresponding Packages file has been downloaded by apt
[12:15:05] <alex_joni> * alex_joni tries it too
[12:15:47] <alex_joni> jepler: btw, do you have a couple machines you could boot the LiveCD on?
[12:16:03] <jepler> alex_joni: there are two I can try with only small amounts of difficulty
[12:16:18] <alex_joni> no amount of difficulty would be even better
[12:16:31] <jepler> at the very least, you have to wait for tonight
[12:16:48] <alex_joni> sure, no hurry.. I'll keep booting it where I get a chance
[12:17:03] <alex_joni> the first machine I had to use irqpoll (didn't detect the HDD chipset)
[12:17:27] <jepler> this is different from the old livecd? why would that be?
[12:17:34] <alex_joni> not different
[12:17:41] <alex_joni> I think I did the same with the old one too
[12:17:58] <alex_joni> (was just pointing it out in case you have a machine where it doesn't want to boot)
[12:19:03] <jepler> hmmm
[12:19:08] <jepler> us.archive.ubuntu.com_ubuntu_dists_dapper_universe_binary-i386_Packages:Package: libio-all-perl
[12:19:11] <jepler> us.archive.ubuntu.com_ubuntu_dists_dapper_universe_binary-i386_Packages:Filename: pool/universe/libi/libio-all-perl/libio-all-perl_0.33-2_all.deb
[12:19:30] <jepler> it looks like on ubuntu's dapper repositoryies, the "all" packages are listed in binary-i386, contrary to the way hardy does it
[12:19:40] <jepler> (repositoryies?!)
[12:19:57] <jepler> no -all_Packages files are retrieved
[12:19:59] <alex_joni> reposistoryies
[12:21:51] <alex_joni> http://www.linuxforums.org/forum/ubuntu-help/81793-apt-get-reports-malformed-release-file.html
[12:24:08] <jepler> nothing's being reported as malformed in this case..
[12:24:19] <alex_joni> only half of that report
[12:24:35] <alex_joni> he had to add a line to apt.conf in order to see binary-all packages
[12:24:37] <jepler> oh, "I broke it and *then* I got this error"?
[12:24:57] <alex_joni> I "fixed" it, and then I got the error
[12:25:10] <jepler> I am nearly certain that specifying "all" as your architecture is wrong
[12:25:24] <alex_joni> probably
[12:25:36] <alex_joni> anyways, assuming our users will change anything in apt.conf is unrealistic
[12:25:48] <alex_joni> so probably the only way is adding the packages in binary-i386
[12:26:07] <jepler> right, I think what cradek should try next is moving the _all packages into the i386 dir and rebuilding again
[12:26:39] <alex_joni> so no more binary-all dir for dapper
[12:45:41] <alex_joni> hi samco
[12:46:52] <skunkworks> Hi alex - downloading now http://www.linuxcnc.org/hardy/ubuntu-8.04-desktop-emc2-aj07-i386.iso
[12:47:06] <skunkworks> I mean 10
[12:47:08] <alex_joni> that's the old one :)
[12:47:10] <alex_joni> ok then
[12:47:12] <skunkworks> heh
[12:51:37] <micges> skunkworks: how much time left of download ?
[12:53:11] <jepler> this *does* work on hardy, right?
[12:53:14] <jepler> (binary-all)
[12:54:10] <jepler> .. I don't think it does
[12:55:00] <skunkworks> 12 minutes
[12:56:15] <micges> me next
[12:56:21] <jepler> micges: I don't think there's any need to take turns
[12:56:25] <skunkworks> heh
[12:56:28] <jepler> www.linuxcnc.org is a grown-up webserver and can handle it
[12:56:34] <skunkworks> dreamhost should have enough bandwidth.
[12:56:37] <jepler> it's cvs.linuxcnc.org that is on a depressingly slow internet connection
[12:57:29] <jepler> $ lsb_release -c; apt-cache show emc2-firmware-mesa-7i43-hostmot2 | grep Version
[12:57:32] <jepler> Codename:hardy
[12:57:33] <jepler> binary-all doesn't work on hardy either
[12:57:35] <jepler> Version: 1:2.3.0~beta1
[13:01:18] <alex_joni> crap
[13:01:47] <alex_joni> micges: I have a mirror that should work better for you
[13:02:09] <jepler> alex_joni: I fixed this with symlinks
[13:02:22] <alex_joni> jepler: cool
[13:02:26] <alex_joni> micges: http://dsplabs.upt.ro/~juve/emc/ubuntu-8.04-desktop-emc2-aj10-i386.iso
[13:02:32] <jepler> $ ls -l */emc2-firmware-mesa-7i43-hostmot2_2.3.0_all.deb
[13:02:32] <jepler> -rw-r--r-- 1 jepler jepler 409914 Apr 19 12:34 binary-all/emc2-firmware-mesa-7i43-hostmot2_2.3.0_all.deb
[13:02:36] <jepler> lrwxrwxrwx 1 jepler jepler 60 Apr 21 07:58 binary-amd64/emc2-firmware-mesa-7i43-hostmot2_2.3.0_all.deb -> ../binary-all/emc2-firmware-mesa-7i43-hostmot2_2.3.0_all.deb
[13:02:39] <jepler> lrwxrwxrwx 1 jepler jepler 60 Apr 21 07:58 binary-i386/emc2-firmware-mesa-7i43-hostmot2_2.3.0_all.deb -> ../binary-all/emc2-firmware-mesa-7i43-hostmot2_2.3.0_all.deb
[13:02:42] <jepler> (ugh, ugly paste, sorry)
[13:03:35] <alex_joni> looks ok if I widen the irssi enough
[13:04:17] <micges> alex_joni: works better thanks
[13:10:38] <skunkworks> downloaded
[13:12:30] <skunkworks> boy - there are a lot of odd answers on the list for the install problem.
[13:13:01] <skunkworks> burning
[13:13:42] <alex_joni> bbl
[13:15:00] <skunkworks> burned
[13:16:00] <skunkworks> booting - come on - you guys love the play by ply
[13:16:03] <skunkworks> play
[13:24:23] <skunkworks> http://imagebin.org/46324
[13:24:29] <skunkworks> now lets see if it installs
[13:38:07] <cradek> jepler: should I move the -all pacakges into binary-i386?
[13:55:40] <jepler> cradek: yes, I think that's the right fix for dapper
[13:55:46] <jepler> there aren't two arches, so a move is as good as a symlink
[13:59:13] <cradek> will do.
[14:00:16] <jepler> alex_joni: are the firmwares on the cd?
[14:02:22] <cradek> dapper/emc2.3/binary-i386/emc2-firmware-mesa-4i65-hostmot2_2.3.0_all.deb
[14:02:29] <cradek> is this _all.deb problematic?
[14:02:45] <jepler> I think that is OK
[14:02:59] <cradek> it's done then. can you retest?
[14:03:02] <jepler> yes
[14:03:40] <jepler> yes, after apt-get update the packages are now installable
[14:03:46] <jepler> in fact I just installed the 5i20 one
[14:03:52] <cradek> yay
[14:04:38] <jepler> thanks
[14:05:45] <cradek> and thank you
[14:06:04] <cradek> I should install it on the lathe today
[14:06:26] <skunkworks> your not running 2.3 on your lathe?
[14:06:41] <cradek> cvs, not the released package
[14:07:07] <skunkworks> ah
[14:07:53] <cradek> bbl
[14:15:09] <jepler> I finally got around to installing 2.3 on zenbot, but haven't run with it yet
[14:42:56] <cradek> is it dapper too?
[14:47:23] <jepler> zenbot is dapper
[14:47:47] <jepler> I don't run emc2 packages on any hardy machines, I just do sim development there..
[15:23:26] <skunkworks> http://imagebin.org/46331
[15:23:49] <skunkworks> now I will boot it on a bunch of machines and see if I have any issues
[15:31:07] <skunkworks> I like the new icon for the shortcut. :)
[15:31:30] <skunkworks> or - atlest it looks different to me
[16:15:52] <steves_logging> steves_logging is now known as steve_stallings
[16:17:52] <steve_stallings> Hi guys, I am back and beginning to recover from the NAMES show. It looks like 2.3 is released now. Will there be a new install disk? I ask because the Ubuntu LTS 8.04.02 maintenance release adds drivers that I need for my new Thinkpad laptop.
[16:20:56] <cradek> there will be a cd eventually, but you do not have to wait for it; just use the 8.04.02 you already have and add emc2 afterward.
[16:22:14] <steve_stallings> But the 8.04.02 is not an RTAI kernal and the existing 8.04 EMC install disk will not install due to the missing drivers.
[16:23:49] <steve_stallings> Missing stuff is I think for the new Intel matrix storage control scheme and without it there is no hard disk access.
[16:23:55] <cradek> yes I guess the kernel will be replaced with the older rtai one when you install emc2
[16:24:24] <cradek> oh so you're really needing a newer rtai kernel. I am pretty sure alex is not building a new kernel.
[16:24:53] <cradek> do you want to control machinery with this laptop, or just simulate?
[16:25:19] <steve_stallings> Hoping for control. I purchased the discrete video version.
[16:26:07] <steve_stallings> Will the older 8.04 see the drivers if they are already there by virtue of an 8.04.02 install?
[16:26:15] <steve_stallings> Guess I can just try it and see...
[16:26:16] <cradek> darn, sounds like it may be painful to get to work.
[16:26:32] <cradek> no, not if they are kernel drivers/modules.
[16:27:39] <steve_stallings> Dang, it seems I get caught every time because of the unfortunate timing of LTS releases and Intel's new "better" gadgets. I got caught with 6.06 driver problems on my latest desktop.
[16:29:48] <cradek> steve_stallings: I guess I always win by using slightly old hardware.
[16:30:36] <cradek> before you do too much, you could boot the old live cd (no hard disk support is ok) and run the latency test. if the results are bad, you may as well not bother going further.
[16:31:22] <cradek> (laptops usually just don't work acceptably)
[16:31:44] <steve_stallings> That works well for me many times. I bought a couple of 3 GHz Pentium 4 HP small footprint boxes just to get Win XP to support critical software that required XP. Cost of box with "refirb" XP license was cheaper that just a new XP license.
[16:32:32] <cradek> * cradek is typing on a 300 MHz machine
[16:32:58] <steve_stallings> If I recall the T400 Thinkpad would not even boot the EMC disk. I will try again.
[16:33:42] <steve_stallings> My office machine is still a 433 MHz with 128 meg, but it is due to be replace any day now because of the same issue with XP.
[16:34:36] <steve_stallings> New box will be dual boot for now, probably VMware later.
[16:44:26] <steve_stallings> steve_stallings is now known as steves_logging
[17:25:32] <alex_joni> jepler: err, you're right, no firmware on the LiveCD
[17:25:35] <alex_joni> I'll redo it :/
[17:32:21] <skunkworks> alex_joni: I have booted it on several computers - installed on 1. No issues (I did not test the mesa firmware) ;)
[17:36:23] <alex_joni> skunkworks: can you check the sources.list on the machine where it's installed?
[17:36:47] <alex_joni> jepler: should we install the firmware on the CD, but not to the target machine?
[17:36:50] <skunkworks> sure where is it?
[17:36:58] <alex_joni> in /etc/apt/sources.list
[17:40:09] <skunkworks> http://pastebin.ca/1398437
[17:45:39] <alex_joni> looks ok
[17:52:48] <jepler> alex_joni: I think if we can avoid forcing the user to do stuff like apt-cdrom it is better
[17:53:23] <alex_joni> hmm.. ok.. then I'll install them all, and install them for the installed desktop
[17:53:24] <jepler> do you know how to apt-cdrom? can you describe it over IRC so an impatient linux newbie with $300 sunk in mesa cards can do it? I don't and I can't.
[17:53:51] <jepler> also the emc2-install.sh maybe should pull them in automatically as well
[17:53:58] <alex_joni> probably so
[17:54:03] <alex_joni> install emc2-firmware?
[17:54:10] <alex_joni> or each one of them?
[17:54:31] <alex_joni> there is probably some foo to apt to install suggested packages aswell
[17:55:31] <alex_joni> space is on the CD for all of them, so no worries there
[17:55:35] <alex_joni> maybe even emc2-dev
[17:57:33] <jepler> this installs all of them, but I don't know if aptitude is installed by default (right now, we use apt-get install): sudo aptitude install --with-recommends emc2-firmware
[17:58:25] <jepler> is aptitude on the cdrom?
[17:59:46] <jepler> aha sudo apt-get -o Apt::Install-Recommends=true install emc2-firmware
[18:01:19] <jepler> * jepler modifies emc2-install.sh accordingly
[18:28:14] <alex_joni> After this operation, 8806kB of additional disk space will be used.
[18:29:14] <alex_joni> emc2-dev pulls in about 34MB of extra stuff
[19:38:19] <alex_joni> yay, nice stuff.. it's 4MB too big again :D
[19:42:01] <skunkworks> yecky
[19:42:18] <alex_joni> building a new one ;)
[19:42:23] <jepler> what did you remove this time?
[19:42:50] <alex_joni> pkgcache.bin & srcpkgcache.bin
[19:42:57] <alex_joni> from /var/cache/apt/
[19:43:03] <jepler> ah
[19:43:10] <alex_joni> sudo rm extract-cd/casper/filesystem.squashfs
[19:43:10] <alex_joni> sudo mksquashfs edit extract-cd/casper/filesystem.squashfs -nolzma
[19:43:13] <alex_joni> argh
[19:43:23] <alex_joni> this paste is driving me crazy :D
[19:43:37] <alex_joni> I keep cleaning the touchpad, and new squashfs's appear :)
[19:44:47] <jepler> example-content might be something to consider removing -- installed-size 11052
[19:44:54] <jepler> maybe you dared to do that already
[19:45:11] <alex_joni> what's example-content?
[19:45:19] <alex_joni> package?
[19:45:27] <jepler> yes, it's a package
[19:46:37] <alex_joni> hah, "These examples should be small but meaningful, and easily discoverable. "
[19:53:49] <alex_joni> wheee.. 695MB
[21:38:18] <alex_joni> good night all
[21:39:14] <alex_joni> (the aj12 version is uploading now, it's already on my mirror, if anyone wants to try it out: http://dsplabs.upt.ro/~juve/emc/ubuntu-8.04-desktop-emc2-aj12-i386.iso )
[21:39:26] <alex_joni> md5sum at linuxcnc.org/hardy/livecd.errata
[21:50:43] <fenn_> fenn_ is now known as fenn
[22:45:21] <cradek> hm, look at Program Extents... http://linuxcnc.org/docs/2.3/html/gui_axis.html#r1_3_3
[22:46:07] <cradek> (specifically, the numbers)