EMC: 03jmkasunich 07TRUNK * 10emc2/configs/vismach/max5kins.ini: make kins version use its own var and tool files
maybe this is the real error:
Log line: !pdfTeX error: pdflatex (file 32_tmp_buildd_emc2-2_3_0_alpha1_docs_src_hal_parp
Log line: ort-block-diag.pdf): cannot find image file
Log line: ==> Fatal error occurred, no output PDF file produced!
(assuming this is even the same problem that tissf was having)
(hm, inside pbuilder the problem is that nothing pulled in ghostscript, which is required to convert eps files to pdf)
(I wonder why the package which contains epstopdf doesn't require gs, the program that makes it work !?)
(and after that, 'convert' from imagemagick is also needed)
EMC: 03jepler 07TRUNK * 10emc2/debian/configure: fix building in pbuilder on hardy/sim
EMC: 03jmkasunich 07TRUNK * 10emc2/src/hal/user_comps/vismach/hbmgui.py: sample configuration for a horizontal boring mill (uses vismach to model the machine)
EMC: 03jmkasunich 07TRUNK * 10emc2/configs/vismach/ (hbm.hal hbm.ini hbm.tbl hbm.var README): sample configuration for a horizontal boring mill (uses vismach to model the machine)
trying to get opengl to draw edges in vismach, but it didn't quite turn out like I wanted .. http://emergent.unpy.net/index.cgi-files/sandbox/vismach-lines-medium.png http://emergent.unpy.net/index.cgi-files/sandbox/vismach-lines-medium-2.png
(it draws lines all over everything, and I'm not entirely sure why)
2nd file = 404
oops, correct url is http://emergent.unpy.net/index.cgi-files/sandbox/vismach-lines-2-medium.png
what do you mean by "lines all over"?
lines that should be hidden are visible
oh, I thought it was supposed to be wireframe
well I want a kind of wireframe
but the lines that are on the far sides or that have a surface in front should be removd or hidden
EMC: 03jepler 07TRUNK * 10emc2/src/ (configure configure.in): testing shows that convert and gs are now necessary to build pdf documentation on hardy
hmmm. I bet those are the two packages in my list
in pumakins there are pins defined in main function that are used for runtime change of kinematics behavior ?
iirc kinematics modules are used as generic hal modules, and where are used kinematicsForward and kinematicsInverse functions ?
document explaining how kinematicsForward and kinematicsInverse are used: http://www.linuxcnc.org/docview/html//motion_kinematics.html
A2, A3, D3 and D4 are numbers which define something about the geometry of a specific puma-type machine. I am not sure what they mean, though
kinematics modules are free to define any parameters and pins that they find useful. For instance, 5axiskins defines a pin to be given the length of the tool, which is important when performing tool length compensation in 5 axes.
I believe this diagram shows what the meaning of the puma A2 and so on are: http://prt.fernuni-hagen.de/lehre/KURSE/PRT001/EXAMPLES/eqs2/puma_dh.gif
apparently it is called "Denavit-Hartenburg (DH) parameter assignment"
what I want to do: I have X linear and Y rotary table and I wonder if kins module can easly tace care of complex scale maths of Y
yes and no.
Y is a linear axis in G-code, so kins would have to translate between the G-code representation and linear+rotary joint hardware
if you can write the mathematical equations to find the position for the linear and rotary joints given an X and Y (and the inverse), then you can use kinematics to move to the correct locations.
the "no" is because emc doesn't presently obey joint velocity and acceleration constraints, only cartesian ones
which can make it difficult to actually run useful gcode
it is laser so velocity must be precise
so it can't be done now
you can get a precise velocity. what EMC2 doesn't do at the moment is restrict joints from accelerating too quickly
unless they happen o be trivkins joints
imagine that the current point is somewhere far from the center of rotation. If you want to move 10mm perpendicular to the line from there to the center of rotation, the amount of rotation is small.
if you are close to the center of rotation, the amount of rotation is large
right now emc will plan both moves at the same velocity and acceleration
but if you make the permitted velocity high enough that the first (distant) movement goes as fast as it can, then the second (close) movement will have a velocity much too high for the rotary joint.
if you do it the other way around, the first (distant) move will be much slower than necessary
this is one of the main problems remaining emc2's support of nontrivial kinematics machines
jepler: today I had sliced bread for breakfast and I still think the tool table touch off is better (I used it a lot last night)
making anything interesting?
those knobs - they are coming out really great
I should make a video - it's a neat setup.
I'm glad it didn't require much UI work -- your idea to use "touch off" for it was a good one.
yeah, I was very surprised to figure out that's how it should work, and it was mostly done already
now when the program cuts with tool 6 at .4375 and I measure the part at .433, I T6M6G43; G0X.4375; touch off X=.433
I don't see how it could be easier
it feels a little bit silly to have to move the machine to enter the new tool diameter
and load the tool?
a special screen could have Tool#___ programmed diameter___ measured diameter___
or radius or whatever
but not sure more ways to do a thing is always a win though
check tht you have ghostscript and imagemagick installed on your hardy system -- I found that without them installed I would get an error without any explanatory message building the pdf documentation..
jepler: yes I have the two
what is alex doing with the params-to-pins changes?
01:32:12 <jepler> (assuming this is even the same problem that tissf was having)
tissf: then you need to add "-dbg any" to the lyx commandline by editing docs/src/Submakefile and let me look at the output it makes
using pastebin or something
seb_kuzminsky: take PID tuning, for instance. It is not possible to make a pyvcp that lets you tune the pid coefficients, because they are on parameters which can't have signals connected to them
yes I have to past but that's a >3000lines !
tissf: yes I know -- but unless you think you can spot the relevant part, I need to see it all
why are there parameters at all? why not just all pins?
tissf: did you try a fresh cvs checkout?
seb_kuzminsky: you'd have to ask jmk that -- but my impression is that originally this kind of use of a vcp wasn't envisioned.
anyway, I think this conversion may be back on hold because I raised some issues regarding pin direction vs parameter direction
take stepgen for instance -- if it's set by halcmd setp to an impossible value, .maxvel is updated by stepgen to hold the maximum attainable speed
that means that technically it has to be an IO pin
but you can't connect that to an OUT pin, which is what you'd want to do to tune it by vcp
perhaps that's just a bizarre thing for stepgen to do
that's certainly another point of view
i dont understand why you can't connect multiple OUT and IO pins to the same signal
it'd be confusing if done wrong, sure
so dont do it wrong :-)
tissf: thank you, looking..
tissf: that is cut off at line 1366 which is the part I wanted to see
it cut the beginning
ok, you must not have done what I intended
make cut the beginning
in docs/src/Submakefile did you add the "-dbg any" flag to the lyx commandline?
but it didn't cause thousands of lines of output below "Building ../docs/src/Master_HAL.pdf"?
$(ECHO) Building $@
@rm -f $@
- @$(LYX) --export pdf2 $<
+ @$(LYX) -dbg any --export pdf2 $<
@test -f $@
let me double check, but yes I think that's what I did
mine prints about 44000 lines of output, starting with
|| Setting debug level to any
|| Debugging `info' (General information)
I changed the line like this:
-@$(LYX) --export pdf2 $<
+$(LYX) -dbg any --export pdf2 $<
! don't understand wy
(removing the @ causes make to print the commandline, instead of hiding it)
ummm wait -- is what you showed me a diff of the Submakefile, or the actual contents of the Submakefile?
it should look like this: http://pastebin.com/m29566b2c
next problem will be that the output is way too big to put on pastebin
maybe you'll have to e-mail it to me (firstname.lastname@example.org)
not the same error!
ok I test pastebin
I have exactly the same file, please wait I make clear
/usr/bin/mandb: attention: /home/francis/emc2-trunk/src/../docs/man/man1/CVS: nom de fichier erroné, ignoré
/bin/sh: Syntax error: word unexpected
make: *** [../docs/src/Master_HAL.pdf] Erreur 2
the make > debug.txt guive the same file
OK, I have no idea what is going on at this point.
you should use 'make > debug.txt 2>&1' to capture both "standard output" and "standard error"
but why you get "word unexpected" I do not know.
it seems likely to me that you've still edited the Submakefile incorrectly
$(ECHO) Building $@
@rm -f $@
@(LYX) -dbg any --export pdf2 $<
@test -f $@
You removed a "$" sign from before (LYX).
it cuts again the beginning
if you e-mail me that output I'll read it in an hour or so
sorry for your time, thanks for help
jepler: 55000 lines pastebin accepts
[17:19:20] <tissf> http://pastebin.com/d52b9ef19
jeple: The file of pastebin is not good! I e-mail the good
jepler: :( pffffffffff
tissf: still not the information I need on pastebin, and I didn't see an e-mail
yeah, I know -- but he sometimes looks at the scrollback
make: Échec de refabrication du makefile « Makefile ».
(that's the only error I can find in the log that tissf mailed to me ... time to get a fresh checkout and/or try the --debug=b,m flag described on the wiki http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?FailedToRemakeMakefile
jepler: can you look at that?
[20:56:19] <tissf> http://pastebin.com/d1d2020a1
It would seem to have a problem of missing images
"only" 10 échecs :)
EMC: 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/protocol_modbus_master.h: change for splitting of modbus i/0 page into 2
EMC: 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/zEMC_README.txt: update
jepler: it is the result of the make --debug=b, m after make clean
jepler: a portion of the result
tissf: do you have the directory docs/src/classicladder/images ?
if not, you need to make sure to specify the flags '-Pd' with 'cvs up'. You can do this automatically by putting this line in your ~/.cvsrc file: update -Pd
.. then cvs up again
you should tell us all your opinion on cvs again :-P
no images directory under dapper !
i'm not saying anything!
total succes... pfff I am an idiot I have shame!
thank you for your help guys
tissf: it works now? I'm glad we solved it
have a nice day, I'm going home
yes all works fine, thank you
good night all