Jacky^ is now known as Jacky^afk
jblunzen is now known as cncuser
03paul_c * 10emc2-auto/wiki/ (15 files in 10 dirs): "Auto update wiki from a cron job. Mon Nov 7 05:30:02 GMT 2005 "
[10:15:32] <Jacky^> http://www.sjdjweis.com/david/pix/servers/penguin.jpg
[10:16:56] <Jacky^> http://vibeke.schou.dk/artwork/woodwork/penguin.jpg
hi everyone :)
how did the developers meeting go yesterday?
good morning #emc
another day, another dollar. I have to do some "microsurgery" on some sensors under the microscope.
easy on the coffee for me.
looking forward to seeing till's code huh ken?
les_w: Well, I read the paper and could more or less follow it, though I suspect I would screw up the details of the delays if I tried to implement it. Right now, I'd just like to get emc2 built so that I could run it. One of the things I hate about those 'meetings' like yesterday is that I find myself making commitments that I really shouldn't. (Although, I didn't commit a date, if it is...
...going to get done, it should be sooner rather than later.) I suspect that it should only take a few hours to put my interpreter changes into the emc2 source.
Yeah...I have to concentrate more on the TP since I now have an expensive spindle that I cannot fully use.
Here's the deal (I think):
We #ifdeffed out a segment that had a horrible math error
that part is specifically needed to allow proper velocity adaptation.
Need to find the bad math rather than just #ifdef it all out
I think I'll call Fred on that one.
les_w: I bet I could fix it if I had a description, like we have for segmentqueue, of what's supposed to be happening. If they wrote it looking at some paper or book, I'd love to know what it is.
chris: I'll call fred and ask.
les_w, where's the ifdef'd portion? (what file)
[18:10] <SWPadnos> morning, Alex
lerman's you around?
Goooooooooooooooooooooooooooooooooooood Morning Vietnam!!!!!!!!!!!!!!!1
well - it's a hot day out there, but it's a dry heat, the kind that makes your balls itch, you know what I mean?
SWPadnos: no, but don't get into details
SWPadnos: Are you sure that's the environment causing that condition of yours?
heh - just (nearly) quoting a movie :)
I have no itch.
well - there's one itch, but I need none of your help with that.
You won't even get your gf's help with that till the green puss is gone.
Lerman is back.
My gf won't help, but my wife may :)
hello.. was just composing an answer
but this makes it easier
lerman: it seems to me that it's an RTAI problem
what else did you run on this system?
SWPadnos: Maybe it was your gf that gave you the green puss.
or is it a new install?
ok.. so that runs ok?
did you try the rtai tests?
never mind if you don't know where those are...
Well, it did last time I tried it. -- Although it seems to go down overnight.
What rtai tests?
can you try some commands?
Sure... As long as I can do them from a remote login.
sure you can..
if/when scripts/realtime will start, emc2 will run
ok.. first make sure there's no realtime started (use lsmod to look at current modules)
Well, rtai_math is running, but unused.
same for adeos (if there is such a module)
ok.. so no more rt* modules in the kernel
maybe rtl8139 ,)
but rtai_* shouldn't be there..
-su-2.05b# lsmod | grep rt
parport 22144 0 [lp]
agpgart 41604 7
rtc 7228 0 (autoclean)
ok.. sounds good enough
now try insmod adeos
-su-2.05b# insmod adeos
hrmmm... could you check that it's the same version (file size) than /usr/realtime/modules/ ?
lerman: probably now emc could run (you probably want to change tkemc to usermot to allow it to work over ssh)
lerman:if you didn't export X over the ssh tunnel..
There is no /usr/realtime/modules/ad*
did you touch /usr/realtime/* ? or it came with BDI-Live?
I assume you're running rc46
I didn't touch it. How can I tell if I'm running rc46?
you should remember what you installed :D
but it's morphix..and rc46 is the only BDI I know that's morphix based
No.... Systems should identify themselves.
bdi-2.xx doesn't qualify (too old)
I just did a script/realtime start and it ran without error.
bdi-4.xx is 2.6 based
lerman: seems it's a problem with the adeos module
copy that to /usr/realtime/modules/
not sure why it's missing ...
isn't adeos compiled into the kernel?
alternative 2: Makefile.inc should read /lib/modules/2.4.25-adeos/kernel/adeos/ as the module dir
SWPadnos: not always, depends on the preference
-bash-2.05b$ cat /proc/version
Linux version 2.4.25-adeos (root@Morphix) (gcc version 2.95.4 20011002
(Debian prerelease)) #1 Wed Mar 31 12:50:23 UTC 2004
SWPadnos: same one
(that's a paste from lerman's email
but doesn't that imply that adeos is in there?
nope.. it doesn't
OK - nevermind :)
it implies that the kernel has the adeos patch applied
but furtheron the adeos part can be either compiled in or used as a module
sure, I guess I never noticed that there's a separate adeos module to load
* alex_joni prefers to compile it in the kernel
ah - I guess I always included it
lerman: re (version you have).. you should talk to paul about that
* alex_joni had nothing to do with the BDI
yet I am happy if I can help out on problems
alex_joni is now known as alex_joni_away
On a slightly different topic, lerman - you may want to remove the parport module. I think thius was causing some problems with parallel port machines (cups would reset the port in the middle of the night)
I copied adeos.o there. rmmod all of the rtmodules and did a scripts/realtime start and it worked.
I don't remember the exact fix, but I think you had to remove any printing-related stuff from crontab
cool. Does "it" mean emc2, or realtime/start?
I thought I had removed it, but I rebooted since then and it appears that I did not do everything necessary to keep parport out. realtime/start worked. I'm going out to the shop now to try emc2. Thanks for the help.
no problem (in case I helped at all :) )
YES! Thanks to SWPadnos and alex_joni -- emc2 runs. And puts up screens that don't look at all familiar. But I assume that I can use tkemc with it once I properly edit sim.ini.
you should be able to.
In fact, if you use remote X or vnc, you can do so from the comfort of the inside of your home (or office) :)
Yes. My family room desktop runs windoze (that' where I am now and do most of my work). My main office machine is pretty slow, is my server, and runs linux. The shop machine is in the garage. I guess I should setup vnc to work, but I haven't bothered with it yet.
cygwin-X works nicely
there's a wiki page titled something like "Viewing EMC on Windows"
ah - here it is: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMCDisplayOnWindows
I have a similar "problem". My main work machine runs Windows, and has 3 monitors. The emc box has an 800x600 LCD - nowhere enough real estate for development work. I use Cygwin-X and can run all of the devtools (I've run QT and KDE Designer, Kate, etc), and also EMC.
It was pretty cool the first time I saw the axis positions change on the Windows machine :)
SWPadnos: you mean running dual GUI's ?
not necessarily, though that works too
Yeah, I ran keystick on the emc box, and minimill on the win box
you can do multiple *G*uis, you know :)
I didn't want X running on the emc box
I've even had multiple displays on the windows machine (one that I was developing, which was just a DRO, plus tkemc)
what are the specs of your emc machine?
well - in that case - nevermind.
forget the model#, but for $85 who cares! lol
it has emc, XP and dos tribooting on it
no ug, works nicely
I've also got a laptop that dual-boots, but it's a pretty powerful one (for 5 years ago, anyway)
Well, if I REALLY cared, I'd use my Fuji Laptop.... P4 2.8GHz w/ 768MB serial and paraport
But I dont do testing/playing on that box...
gotta run - be back later.
SWPadnos is now known as SWP_Away
awfully nice here...fall colors, clear blue sky, 70 degrees. Takeing a walk for a break from work.
Fantastic. I really liked the fall colors here this year.
Maxine and I went for a three mill through the woods on Sat.
Leaves are all gone. Tamarac are bright yellow/orange.
les_w: Are you taking a walk for a break from work at the same time as you typed that? :-) BTW: where are you located?
Ken, just got back. Located in the blue ridge mountains at the georgia north carolina border.
But back to work. Suits coming day after tomorrow. Seeking continued research funding, and I have to get a gizmo ready for them to take back.
alex_joni_away is now known as alex_joni
sure away ?
* alex_joni just came back
so. what did I miss?
les_w: Dont forget the doohickey and the whatchamacallit
hey Jymmm you alive ? :P
Jacky^: No, please leave a message at the exorcism
alex_joni : G'Afternoon =)
if you wish...
Man, having such a hard time finding tooling < 1/8"
1/16" would be very nice.
alex_joni: apt-get install ntpdate
1/16 end mills with 1/8 shank are very standard
Jymmm: is eating a cow :P
at this time
cradek: single flute?
no, I don't know anything about single flute tools
cradek: Most of what I have seen the CEL is really shallow.
Eh, I got a call into Onsrud. We'll see what they have to say.
In the meantime... I'll just play around with some polycarb and see if that resolves the issues.... "HELP ME... I'm m e l t i n g!"
hehe, got my PCBs
but have forgotten to order the components :-)
did u read the devel list?
on which topic do you want to focus
that Markus guy.. working on a portal
Till said this half a year ago in the users list
sounds very much like what you want/need
don't know if it depends on EMC
but it'll be available next year only ...
it is done with emc
while i was waiting for the PCBs i lookt a bit into the code, to find the places which need to be changed
i need a very clever idea, otherwhise i have to edid a lot
the motion controller has a lot of states, if i forgott one, the machine is dead
Imperator_: have an idea of what you want to do?
maybe i do the most in do_homing and output to hal
let me knwo
and during normal running, just slave the two axes?
you probably need to touch trivkins
also the ferror code (to add abserror)
and take care of the difference between them, like folowing error
hmm.. that might be easy to do...
how do you like this idea:
you have position-set to axis 1
that can be done i output_to_hal and check_faults
position-is 1 = position-set 2
and position-set1 and position-is2 is used for ferrors
Heh... need to pickup some spare spreader clamps to CYA =)
Imperator_: get my point?
to compute ferror for the portal from the max. distance
its just coping
that can be done in check_softlimits
ohh.. john, hello ;)
* jmkasunich hides
big question to you :D
jmkasunich: perfect time to join :-)
yeah.. quick ..
* alex_joni goes first
jmkasunich: make install
* jmkasunich runs away
I have a version I really find ok
* Jymmm hands jmkasunich a jet pack
it's only administrative and CVS stuff I want to ask
so don't run just yet
* ValarQ skips and jumps
the code was in a branch (autoconf_install_0_1)
it was working... last time I touched it
* Jymmm give ValarQ his medication
now. in the meantime most of emc2 changed a lot
the diff from the branch and HEAD is about 1.6MB
* ValarQ sits still on his chair and smiles
merging would be truly messy
Damn... wrong medication
the biggest problem is that a lot of files (most involved in the change) have been moved
I checked out a clean HEAD, and reaplied all changes
Jymmm: everything looks so cosmic now
now.. where should I commit that?
back to the branch? then probably all files will be different
ValarQ: Better enjoy it... that was my last one and ... it was for me!!!
that would probably be messy when merging back to HEAD..
1) did anything you do change the functionality when it is not installed (just make and run)?
that was the first point I started from
it's a different run script that gets installed
and the run script gets adapted by configure
also minor changes to the tcl code (to be able to find emcsh on an installed system)
skipping the details, I guess what I'm asking is: will anyone who is currently using EMC2 by the "make and run locally" method gonna get busted if you commit to HEAD?
jmkasunich: good work
jmkasunich: not to my knowledge, and not according to my tests
then commit to head, and forget the branch
it has served its purpose
ok.. the Makefile will have make install commented out for now
and only who really needs to test it will try it
ok, sounds good
btw, I did make an emc2 deb package out of the make install ;)
one thing, if you don't mind...
well.. with the help of make install is more appropiate
either add another section to directory.map, or make a new file called install.map, that shows where things go
that's already in cvs
you need to keep up with the changes :D
and your work matches it? (those things tend to get out of date)
I just updated it a few days ago
and added the make install part
also touched abot 30 files last night (added the same header to all)
hope you don't mind..
* jmkasunich is reading Matt's declaration of surrender...
alex_joni: and others who might know: is the interpreter file rs274_errors.cc (in emc2) generated manually or automatically?
some of the comment lines should be touched up, I'll work on that a little
lerman: think it was generated manually... but I have little idea about interp
old line: This file, foo.c, does blah, blah
jmkasunich: lerman is the youngest emc2 user :D
this file, foo.c, does blah, blah
since the previous line has the filename, the sentence should be changed to simply say "does blah blah"
ok.. suit yourself.. :)
* jmkasunich is nitpicky sometimes
* alex_joni lost it after 2 hours of copy/paste through the sources :D
lots of concetration is needed in order to remember what you just did :)
I can understand that ;-)
but getting the standard headers in there is a good thing, so thanks
ok.. so I'll commit the make install stuff
and maybe the debian stuff too
although I have little knowledge of that yet :D
but I am very proud to have built emc2-0.0.1-i386.deb
damn... you've been very busy!
jmkasunich: well..gotta catch up on your work on the mazak :)
I really should do some documentation there, before I forget it all
yup.. did you follow: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl/emcinfo.pl?BoardElection
not looked at it since Thursday
did see the nominations of Matt and Paul
(on the list)
matt accepted, paul didn't
do we have someone to send out the ballots? do they need help or info from me (or anyone else who remembers how the last ballot was run)
do they need a list of email addresses from me or another list admin?
I put explicit instructions for updating AXIS / bdi4.30 on axis.unpy.net and updated the README in AXIS cvs.
jmkasunich: matt convinced chris to do it again this year
I sent him(matt) the email list
I didn't know you had the list, you are an admin too?
* alex_joni grins :D
should I tell that other guy that we don't need his help with the ballots?
fenn.. guess so, but keep a backup open :D
03alex_joni * 10emc2/ (13 files in 5 dirs): added make install target, changed some files to allow that, run-in place should still work. needs some testing,, for now the make install target is commented out
03alex_joni * 10emc2/scripts/install.run.in: added the install.run.in script. this one gets modified by configure, and installed during make install.
hmm.. this is a nice machine to control with emc :)
[23:33:00] <alex_joni> http://www.lemonzoo.com/funny_videos/504/cool_racing_simulator.html
you named a box 'bald' ?!