Back
[05:38:22] <jthornton> on my test computer for uspace I've installed wheezy with no root password and following the instructions for installing on debian wheezy with preempt rt kernel
[05:39:03] <jthornton> and sudo apt-get install linux-image-rt-686-pae returns unable to locate package linux-image-rt-686-pae
[05:39:10] <jthornton> did the name change?
[08:03:55] <jepler> that is the correct package name for wheezy when you started with a 32-bit install.
https://packages.debian.org/wheezy/linux-image-rt-686-pae
[08:04:14] <jepler> if you started with a 64-bit install, it would be
https://packages.debian.org/wheezy/linux-image-rt-amd64
[08:06:59] <jepler> otherwise, check whether you can install any packages with apt; maybe the installer left you with a busted apt installation. at least at some times in the past, if the network wasn't configured properly during installation this would happen.
[08:08:27] <jepler> .. do additional troubleshooting, such as use apt-cache search to find out package names that are available, apt-cache policy to show where the package would be downloaded from, etc
[08:12:09] <JT-Shop> yes it's a 32bit install
[08:12:57] <JT-Shop> do I need to say apt-get install
https://packages.debian.org/wheezy/linux-image-rt-686-pae
[08:13:06] <jepler> no
[08:13:18] <jepler> that page is for humans
[08:13:34] <jepler> it is my way of saying, "yes there is a debian package with that name in wheezy, you can go look at information about it"
[08:13:37] <jepler> afk
[08:13:43] <JT-Shop> ok
[08:36:29] <micges> JT-Shop: any progress with sserial problem?
[08:39:39] <JT-Shop> got make to work, now I need to change the mask and run make again. Didn't have much free time yesterday
[08:52:46] <micges> ok
[09:12:43] <JT-Shop> zlog
[10:11:09] <cradek> jepler: the linuxcnc wheezy cd gives you a correct sources.list no matter how the network is while you install
[10:11:38] <cradek> where correct to me means it'll be right when you later get hooked up to the network
[10:25:23] <jepler> cradek: thanks for that info. I have lost track of whether JT-Shop is working on instructions starting with debian install media or linuxcnc.org install media.
[10:26:08] <cradek> I guess I don't know either
[10:27:01] <cradek> I dislike their assumption that you will never have a network if you don't have one at install time, so I hacked around it
[10:27:59] <cradek> speaking of which, is it time to try to make a new linuxcnc cd with jessie and seb's rtai kernel?
[10:41:08] <seb_kuzminsky> cradek: i'm not sure yet
[10:41:19] <seb_kuzminsky> it keeps crashing on the buildbot vm
[10:41:26] <cradek> heck
[10:54:34] <seb_kuzminsky> i've had some emails with rtai.org about it, but nothing's really come of it
[10:55:10] <seb_kuzminsky> i dont know how much of it is from my garage being cold (i had some non-rtai VMs flake out around the same time)
[10:55:38] <seb_kuzminsky> or how much is from it being a VM, and maybe not obeying the kinds of timing assumptions that rtai makes?
[10:56:25] <seb_kuzminsky> or how much is from my kernel config being somehow bad (though i think paolo looked at it, and he didnt tell me about any mistakes)
[10:56:31] <JT-Shop> I'm back to following the manual using stock debian wheezy
[10:57:01] <JT-Shop> micges: just got back and tried the mask and got
http://gnipsel.com/files/bp/1-13-16.txt
[10:57:55] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.6 7576756 06linuxcnc 03debian/extras/lib/udev/rules.d/99-shuttlexpress.rules 10debian/extras/linuxcnc.files 10docs/man/man1/shuttlexpress.1 include udev rule file for ShuttleXpress USB jog pendant * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=7576756
[11:04:38] <JT-Shop> this is the result from sudo mesaflash --device 5i25 --sserial
http://gnipsel.com/files/bp/5i25-info.txt
[11:08:11] <jepler> huh here's some weird behavior from limit3:
https://emergent.unpythonic.net/files/sandbox/limit3-badbehavior.png
[11:08:29] <jepler> I wonder how it totally misseshalf the transitions in the signal it's supposed to track
[11:09:10] <cradek> oh dear
[11:09:26] <cradek> is it only when the period just matches?
[11:10:27] <jepler> that screenshot was with maxa = 14
[11:10:41] <jepler> at maxa = 16, it never misses a direction change
[11:11:57] <maxcnc> Hi where is the .axis_preferences locadet in the system
[11:11:59] <jepler> with maxa still lower you can get other behaviors, like being nearly 180 degrees out of phase
[11:12:56] <jepler> maxcnc: with the git source, you can use "git grep" to begin to answer a question like that
[11:13:00] <jepler> jepler@parrot:~/local/src/linuxcnc$ git grep -F .axis_preferences
[11:13:00] <jepler> src/emc/usr_intf/axis/scripts/axis.py: self.fn = os.path.expanduser("~/.axis_preferences")
[11:13:54] <maxcnc> "~/.axis_preferences" is this the home location
[11:14:03] <maxcnc> is this hidden
[11:14:26] <jepler> did you mean "located" or "loaded"? I thought you were saying the latter.
[11:14:45] <jepler> ~ is shorthand for "the user's home directory". Many programs hide files when the names begin with ".", it's a common convention in unix.
[11:14:53] <maxcnc> i want to change a default value that is in the file but i cand find it
[11:15:18] <jepler> every setting in that file can be changed by a menu item or keypress in the AXIS program.
[11:16:10] <cradek> maxcnc: what problem are you trying to solve? it sounds like you are on the wrong track and maybe we can help
[11:16:23] <jepler> https://emergent.unpythonic.net/files/sandbox/limit3-more-weird-behavior.png
[11:16:44] <jepler> maybe limit3's algorithm is just not very good in the case where it can't fully track its input signal :-/
[11:17:37] <maxcnc> cradek: ive been asked to change the default preview to View_type value 2
[11:17:52] <maxcnc> there is no ini for that
[11:18:51] <cradek> I suggest changing AXIS so the selected view is saved across runs, like other preferences are saved. I'd happily review a patch that does this.
[11:19:01] <maxcnc> ok in axis there is vars.view_type.set(0)
[11:20:23] <cradek> jepler: is limit2 broken too?
[11:21:13] <jepler> cradek: good question. I was only testing limit3, and only varying maxa. maxv was unset, which I think means essentially unlimited..
[11:25:08] <jepler> limit2 (or limit3 with maxa infinite and varying maxv) doesn't misbehave as readily
[11:25:36] <jepler> it tracks the square wave as well as it can. if maxv is too small, it of course can never reach it, but it's always going in the right direction
[11:25:48] <jepler> and certainly not "cruising" while way off from the real value
[11:33:17] <jepler> when the position is steady but at the wrong value, it's switching between the rampa < 0 if and else clauses on alternate cycles
[11:34:22] <jepler> .. in the branch that says it's trying to ramp to match v
[11:35:42] <jepler> so yeah, limit3 has the case "can follow exactly" and the case "ramp to match input velocity', which is of course 0 in this case
[11:36:43] <jepler> it should have had some sort of "ramp to match input *position*" case
[11:39:08] <cradek> to me, limit3 seems very hard to write
[11:41:11] <jepler> looks like limit3 never had a tracks-position mode
[11:42:25] <jepler> if the input signal cannot be tracked exactly it's hard to say just what you should do. You would have to define what is desired before you could write it.
[11:48:15] <linuxcnc-build> build #1690 of 4017.deb-wheezy-amd64 is complete: Failure [4failed shell_1] Build details are at
http://buildbot.linuxcnc.org/buildbot/builders/4017.deb-wheezy-amd64/builds/1690 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[13:20:05] <pcw_home> and all errors except remote errors should go through the inc/dec/threshold filter so can be squelched
[13:20:36] <pcw_home> oops
[13:21:19] <andypugh> Someone working on the sserial errors thing?
[13:21:39] <pcw_home> chatting with micges about it
[13:21:54] <andypugh> Great, that means I don’t have to fix it?
[13:22:57] <pcw_home> I think Jepler is going to work on it once he has some hardware and a way to trigger errors
[13:23:26] <andypugh> Triggering errors is the part I was struggling to figure out
[13:23:46] <skunkworks> sounds like you just need a running vfd close by ;)
[13:23:59] <jepler> yes, I intend to take a look.
[13:24:16] <andypugh> Right, To The Shop! then
[13:24:26] <jepler> pcw_home had a suggestion about how to trigger the errors on the cheap, namely generating glitches with a stepgen
[13:24:41] <jepler> it's somewhere in scrollback/logs, maybe on monday..
[13:25:23] <pcw_home> not sure why JT has such severe problems, even with extrachar errs masked there seem like hundreds of errors
[13:27:12] <pcw_home> (not sure if that is with the filtered UART 5I25 firmware or not but if so it means 100 or >~1V noise spikes longer than 200 ns on the 7I77/5i25 cable)
[20:01:01] <malcom2073> PetefromTn_: You win powerball yet?
[20:03:40] <jepler> > Without the need to breathe, [sharks] can stay underwater for a long time
[20:03:43] <jepler> wow *THAT* is reporting
[20:37:34] <skunkworks> yikes - everyone knows they use electrolysis of water to make oxygen to breath..
