dang... anaconda.condig needs some additional work.
annaconda ... vaguely remmeberthat
I vaguely remember seeing it crash during RH installs anyway :)
A quick question for cradek or jepler: How demanding is AXIS as a system load? The box's I have available to try it on are Duron 600/Celeron 400, with Nvidia/TNT cards
robin_sz: I thought 'digimatic' calipers used a magnetic scale with a vernier method
Some are capacative, others inductive, and a few magnetic.
As a machinist, I hate 'em - they lose position, same goes for micrometers
They do have a use around abravisive contamination
A friend uses one, making pool queues - super micro-fine wood dust
But they're not good for production use - you never know when they're going to 'slip', so you have to calibrate and set the zero often
Back to cleaning [sigh]...cables and tools and debris, oh my
* robin_sz pays attention again
my christmas project is to pull everything out of the workshop into the yard
and then put it back
in some sort of useful order
my christmas project is to put a new cylinder gasket in the fork truck ... and then ...
sodding hting uses more oil and water than propane
03Zathras 07BDI build system * 10Babylon Cluster/comps.xml: File changed. New revision:1.14
* paul_c wonders if jmkasunich is still awake...
asdfqwega: These are the machines Chris or I have been developing AXIS on: 1.5GHz Pentium M, 2.8GHz Pentium4, "2400+" AMD Athlon. If you have accelerated OpenGL, duron 600 and celeron 400 should do OK, but not be silky smooth.
asdfqwega: I know cradek has actually tried axis on a a <400MHz Pentium II laptop with no 3d accel, and it runs but not well
asdfqwega: but opengl accel should make a big difference
difficult to wake up :-)
You disappeared a fraction too early last night
was going to let you download the beta test...
but now you changed your mind ;)
last night I was at home.. on dial-up
now I could do that...
daytime will knock my bandwidth.
ok.. I can cron the dl
I found a couple of "interesting" issues that need to be fixed first.
ok... I can wait
btw, before I forget.. I'll go away on Wednesday
going on a business trip to germany, I'll be back next week
so I guess I'll miss the chat on sunday
03Zathras 07BDI build system * 10Babylon Cluster/isolinux.cfg: File changed. New revision:1.2
* alex_joni is right back
* alex_joni is back
I had a very brief chat with zwisk last night
'bout install targets ?
yeah... a little, but he didn't have much time :(
maybe the next time
he was complaining about no thaving a working rtai.rpm
he was complaining about not having a working rtai.rpm
I had the same problem with rtlinux for bdi-2.xx
and to a lesser extent, with the rtai deb package.
I can imagine
There isn't much I can do to help zwisk on that front.... It is something he'll have to work on.
arrgg... Got rtai.spec from the TNG somewhere.... along with a tweaked kernel spec.
back in a bit.
* alex_joni is gone for today
what's the second worst thing you can type on irc?
Doesn't look like axis will run with the 2.6.9 build of EMC.
can you give more details?
cradek: Thanks for correcting my error in the abort bug fix in emc2
damned maze of switch statements...
I tested it, and I still had the bug
I would have never noticed it by looking at the patch
yeah it's nasty in there
what, did the patch helpfully apply to the wrong switch statement entirely?
As soon as I looked at the full source file, I saw the error in patching.
jepler: it was a meat-patch
jepler: (I think)
paul_c: give me more information about why axis isn't working for BDI
This build uses libnml from emc2 with the directory layout of emc1
that sounds strange but I don't see it as a big problem
agreed, it isn't a problem.
so what problem are you having with axis?
Haven't tried to run it yet, jist a quick look at the code indicated that there would be problems.
headers and libs for the C stuff
it just takes some changes to setup.py to make it build with emc2 only
yes, the names of the directories and libs are a little different, but that's no problem
yup. minor problem.
so is there anything else?
I'll help you fix it however I can
I think it's important for it to work on this BDI, otherwise many people won't be able to try it
nope - Just got to get emc to compile on the main dev. box
Once I get this to compile, I'll have a go at axis.
I'll be here until at least 2200UTC
I'll probably be around till much later than that.
cradek: would you happen to have a screenshot of AXIS?
[15:20:30] <cradek> http://axis.unpy.net
lots of info about it there
les: screenshots in the "about" link
paul_c: I'll try to be around too
resolved the compile errors, now building another CD ready to try an install.
wow the axis screenshots look great...I will want to try this out for sure
hope those old computers have enough system resources left to run it...
yw chris...great job
let me know if you have trouble building or installing it
there are a couple fixes already, mostly for older systems, described on the web page
I will just follow the instructions....I assume I do not have to rebuild emc?
you do need a built emc source tree
have that...several in fact...I keep old versions intact when making a change
often when I try to do individual makes I get garbage (don't know enough about gcc I guess) so I just build the whole thing
les: it's probably the build system, not you, but a full build doesn't take long
right..it's real quick so no problems
hmm got an email from another happy customer with his emc carved teak sign
looks like many thousands of customers will end up with products made with emc
I wonder if anyone else is using the program for production?
who was making those TPS sensors?
sorry went to the shop
heh the TPS was me too...
but that was just all the design protos...we have cut injection molds now
axis is running on 2.6.9
* paul_c is feeling smug
hacked python al by myself ;)
Half a cheer for Paul - Hip! Hip! Hoo-
if you don't want it to be the default, let's put a commented out DISPLAY=axis in the ini
it will be great if people can try it easily
Will add an entry in the run script for axis so that we don't have to symlink it
Now how do we create a deb (with the required deps) ?
I don't know, jeff did the first one
I plan on trying AXIS before the week is out - been doing to much computer fiddling of late
if you have a modern OS install it's really easy
if you have to build all the python stuff from source, not so easy
a suggestion for the build instructions
I'm using the Live rc46 - it should be easy enough
asdfqwega: there's a binary package on the web site
aptget python-dev, xlibmesa-gl-dev, & xlibmesa-glu-dev
jeff had xlibmesa-dev
is that wrong?
got the mail to the list?
about axis? yes
glad to hear you like it so far
OK... xlibmesa-dev is a meta package for gl & glu dev
paul_c: ok, I'll leave it as the simpler one then
paul_c: will this be called BDI-3?
Dragging axis across the screens, I have noticed it doesn't redraw on screens 2 & 3
we'll look at that as soon as we can run this BDI!
How many will be running multiple heads
It's a problem with openGL - I find most things will only run on screen 0
paul: 2 & 3 are multiple heads on your setup?
paul_c: ahh that would explain it
I thought you meant virtual desktops, which would be extremely strange, but you do actually mean screens
Have three monitors on one box here
paul_c: will this be called BDI-3?
sorry... Should have answerd that...
assumed you just missed it
Not sure what to call it.
I'm calling it "the upcoming BDI release" then
Had BDI-2.xx, BDI-TNG, umpteen BDI-Lives
How about BDI-6 ?
I dunno, you tell me :-)
TNG should have been -3.xx
* paul_c finds a file that doesn't plot fuly.
you mean the preview is incomplete?
it gets coordinate systems and a few arcs wrong
it will all be fixed in 1.0
(we are already making progress on 1.0)
This is previewing about 1/3 of the full plot.
send me the file?
does it use g92, g54....g59.3?
those will certainly display wrong
anything new and exiting?
we just heard AXIS will be on BDI-4
14833 lines, and just G0 & G1 moves
it's the first step to world domination!
paul_c: yeah I definitely want to see that
Ah.. jepler - Just the lad we need..
say for when is the new release due?
Can the "programs" dir default to emcsourcedir+programs ?
paul_c: would you make up a list of things you find
I will add them to the wishlist or buglist as appropriate
well... The only other question was how to build a deb
ok, well I can't help with that
but about the directories:
I already added a wishlist item that it should default to the directory you last loaded a file from
that's a separate issue from the first default directory, I know
but it will make it better
I'll add another wishlist item saying the initial directory should be programs/
what if the default program dir was read in from the ini file ?
that would be fine too
might be kind of nice
probably be better in the long run
we would have to fix all the guis
but that wouldn't be hard
where's the feature request page ?
right now, it's the TODO file in our cvs
* file open dialog should remember the directory from previous open
* file open's initial default directory should be read from the ini
it now has these two things
and needs to be able to handle ~/foo correctly (which emc doesn't at the moment
it already does
it's a smart dialog, you can type a dir name and hit enter, and it will show you that dir
or you can type ~/file.ngc and it opens it directly
One advantage of using a higher level language.
gotta go, I might be back in a few hours
I'll read /msg later if you have other thoughts
paul: wanna hardcode the path to the inifile?
I meant in the inifile
* paul_c waits for python-axis to add to the CD.
I think this will get of an issue when doing a make install (this path needs to be changed)
Does G64 (motion blending) operate only between G01/2/3 moves?
alex_joni: The alternative is to follow the example of many other packages and enforce an install in order to run.
I'm working with my image->gcode python script, and I'm working with G00/1 moves
asdfqwega: Would have expected G64 to work with G0 moves.
And I don't see any difference
paul_c: I don't think emc2 is there yet
emc2 has a long way to go.
Hm...would it be too much to hope that there's already an .ini variable to control the amount of blending?
what if the path in the inifile points to the installed path (emc2 developers who run-in-place can handle that)
asdfqwega: No such variable that I know of in the ini file
alex_joni: The answer is simple - Pass an ini parameter on the command line.
emc.run -ini /foo/emc.ini
paul_c: Probably wouldn't matter anyway...I need a better machine before G64 would really pay off for this
paul_c: so a different emc.ini than the one installed? (if I understand you right...)
On the next BDI release, there is only one master run script
all desktop icons call the runscript with an ini file name.
I agree there
what I was talking about was the programs path inside the ini file (emc.ini)
if you fix the programs path inside emc.ini it must get modified during install (not that good)
Not really... the run script would prefix any relative paths with $(pwd)
so the runscript reads the programs path from the ini and passes it to the DISPLAY?
Hmmm. The GUI program would be responsible for reading the ini file if it needs any variables of it's own.
One of the tenants of the EMC design is that each part can operate as a standalone program and use NML messaging.
yeah but the gui would only know of the inifile and the folder of the inifile
Only the task controller & the kerenl modules need to run as an "item"
and of course the location from where it was run (but that can easily be /usr/bin)
but all the locations are modifyable by ./configure, so that complicates it a little
anyways.. that's not that important right now.. I'll stop bugging you ;)
I still don't see it as an issue..
ottos...you running the motenc-100 board?
we'll address it when time comes
only thing I've wanted from the .run file would be to take a filename to open immediately. That way you could double-click a g-code file and AXIS will open with that file
well...still getting all the cables/etc
jepler: you could do that if you have one implicit inifile to open (and of course inifile-location)
ottos- maybe send me an email firstname.lastname@example.org so we can stay in touch
set a mime data-type for ngc files.
sure..you have the same?
I'm having problems and i suspect I may be the only one
ottos: I also have a Vital card (for testing)
well i thought you said something about the vsi motenc-100 board yesterday
do you have a good compile of vital.o and vitalmod.o yet?
questions about pinouts/etc...
dave-e: Have you tried messing with the offset values for the SCALE ?
paul...no i haven't
I think my problems are more fundamental
If you have drift in the DAC outputs, try a little bit of offset
well right now I have a stability problem.
even at low P I oscillate
enough to fault the amp
did bad things to the test servo motor
* paul_c could do with flying over to see the setup & play.
oh how i wish
Dec 20th for two weeks OK ?
wish had that kind of $$
just to restore my confidence i moved the stg setup from my small mill to the mazak. took a couple of hours but
it is running fine on two axes with minimill...now I need to add homing
* alex_joni is going to sleep
Dave , did you consider omron D5A/B for home switch?
the mazak has all the home switches, etc already...I do have some automation direct switches to use on the small mill
my servo amps have + and - inhibits on them which is really nice.
that way you have both a soft and hard limit
amps are servo dynamics
i've snatched some Yaskawas SGDA, want to see them in action...
hmmm...are those ac?
you have amps for them yet?
say for PID tunning any suggestios?
no really .... just bring up logging and pick your way thru.
the hard way..:D
send me and email and I'll send you some stuff on tuning.
gotta go pick at the system...............later
Bump the version number up on the emc.deb.
03Zathras 07BDI build system * 10Babylon Cluster/Make.rules: File changed. New revision:Makefile
03Zathras 07BDI build system * 10Babylon Cluster/emc.ini: File changed. New revision:generic.ini
03Zathras 07BDI build system * 10Babylon Cluster/Makefile: File changed. New revision:126.96.36.199
03Zathras 07BDI build system * 10Babylon Cluster/Makefile: File changed. New revision:iosh.cc
03Zathras 07BDI build system * 10Babylon Cluster/Makefile: File changed. New revision:dro.c
03Zathras 07BDI build system * 10Babylon Cluster/Makefile: File changed. New revision:188.8.131.52
03Zathras 07BDI build system * 10Babylon Cluster/Makefile: File changed. New revision:emctaskmain.cc
03Zathras 07BDI build system * 10Babylon Cluster/Makefile: File changed. New revision:184.108.40.206
03Zathras 07BDI build system * 10Babylon Cluster/- New directory: File changed. New revision:NONE
03Zathras 07BDI build system * 10Babylon Cluster/- New directory: File changed. New revision:NONE
03Zathras 07BDI build system * 10Babylon Cluster/emc.png: File changed. New revision:hello.gif
03Zathras 07BDI build system * 10Babylon Cluster/G code editor.desktop: File changed. New revision:Sherline_inch.desktop
03Zathras 07BDI build system * 10Babylon Cluster/comps.xml: File changed. New revision:1.15
03paul_c * 10emc2/src/emc/task/emctaskmain.cc: Didn't quite get all of Chris Radek's bug fix in.
That should kick off the farm in a little while ;)
cradek: is the test for 'cutting' something you added to emcplot3d?
03Zathras 07BDI build system * 10Babylon Cluster/picax.xml: File changed. New revision:1.12
jepler: What's the quick'n'easy way to package axis in a deb ?
brown paper and string
jepler: line #?
paul_c: I took some stuff from the web called bdist_deb, and put it in my local tree. It's not even in CVS yet
cradek: around 793
jepler: OK... I'll get a CD ready for Ray and we can add axis at a later stage.
I'm going to have to look in cvs
I think it doesn't work right anyway, but I'm not sure
paul_c: I could tar up the "debian" directory that bdist_deb created
That would do - I can create a package here with the tweaks needed for the 2.6.9 build
jepler: yes, in 1.12
jepler: but I'm not sure it always works right