linuxcnc blogs

fix a stupid bug in my previous commit

2010-03-13T07:16:59+00:00

fix a stupid bug in my previous commit

LIB_DIR is where we're planning to *install* the libs, not where we're
currently staging them.  We stage them always in "../lib".

(I didn't notice this because I stupidly had emc2.deb installed, so
LIB_DIR *did* indeed have the libs in it...  Third time's the charm
I guess/hope.)

[D H] src/Makefile

fix linking when compiling for installation

2010-03-13T06:45:09+00:00

fix linking when compiling for installation

When we're compiling for installation (manual installation or via the
debs), we want our executables and libraries to be linked against
our not-yet-installed libs, and we don't want our build dir in the
executables' rpath.  "rpath-link" does exactly this.

When we're compiling for run-in-place, we want our build dir in the rpath
(which also lets the linker find it), so nothing changes in this case.

[D H] src/Makefile

link librs274 against libemcini

2010-03-13T05:40:28+00:00

link librs274 against libemcini

librs274 uses a bunch of symbols from libemcini.

Before this commit, librs274 did not link against libemcini, instead
requiring that any executable that links against librs274 also link
against libemcini, even if that executable itself did not need libemcini.

This commit rectifies the omission in librs274, and also stops linking
the rs274 executable against libemcini (since that executable does not
itself use that library).

[D H] src/emc/rs274ngc/Submakefile
[D H] src/emc/sai/Submakefile

avoid linking to a couple of unused libraries

2010-03-13T05:05:54+00:00

avoid linking to a couple of unused libraries

[D H] src/hal/utils/Submakefile

get rid of unneeded import

2010-03-12T04:03:51+00:00

get rid of unneeded import

[D H] src/emc/usr_intf/axis/scripts/axis.py

correct error message

2010-03-12T03:43:20+00:00

correct error message

[D H] src/emc/rs274ngc/gcodemodule.cc

wrap long lines

2010-03-12T03:39:45+00:00

wrap long lines

[D H] src/emc/rs274ngc/gcodemodule.cc

Remove unneeded offsetof hack

2010-03-12T01:04:25+00:00

Remove unneeded offsetof hack

beacuse this file doesn't use offsetof(non_pod), the hack isn't needed
to avoid compiler warnings.

[D H] src/emc/rs274ngc/gcodemodule.cc

make it easier to profile axis

2010-03-11T22:16:29+00:00

make it easier to profile axis

[D H] bin/profile_axis
[D H] configs/sim/profile_axis.ini

move methods to base class

2010-03-11T13:32:46+00:00

move methods to base class

[D H] lib/python/glnav.py
[D H] lib/python/rs274/glcanon.py

reduce memory usage of display lists

2010-03-11T13:12:41+00:00

reduce memory usage of display lists

so that the program can be displayed without rapids, there are
two display lists.  Before, the display lists were "everything"
and "everything but rapids".  However, you can get the same
ability to draw with and without rapids using two display lists
"only rapids" and "everything but rapids".  Presumably having
less stuff in display lists reduces load time and memory usage.

[D H] lib/python/rs274/glcanon.py

get rid of dead code

2010-03-11T13:10:25+00:00

get rid of dead code

this method was replaced by the method directly below

[D H] lib/python/rs274/glcanon.py

make the preset view buttons react to a rotate

2010-03-10T23:40:03+00:00

make the preset view buttons react to a rotate

[D H] src/emc/usr_intf/axis/scripts/axis.py

rotate should snap into perspective mode

2010-03-10T23:38:41+00:00

rotate should snap into perspective mode

this behavior was inadvertently lost with the merge of glrefactor

[D H] lib/python/glnav.py

make my editor use the right settings

2010-03-10T23:26:15+00:00

make my editor use the right settings

[D H] lib/python/glnav.py
[D H] lib/python/rs274/glcanon.py
[D H] src/emc/rs274ngc/gcodemodule.cc

get rid of redundant code

2010-03-10T23:25:03+00:00

get rid of redundant code

open_file_guts does this, no need to do it here

[D H] src/emc/usr_intf/axis/scripts/axis.py

simplify since it's always OK to call extents_info

2010-03-10T23:24:08+00:00

simplify since it's always OK to call extents_info

[D H] lib/python/rs274/glcanon.py

make it always legal to call extents_info

2010-03-10T23:20:00+00:00

make it always legal to call extents_info

[D H] lib/python/rs274/glcanon.py

a redraw is required after the rotation changes

2010-03-10T23:03:07+00:00

a redraw is required after the rotation changes

[D H] src/emc/usr_intf/axis/scripts/axis.py

fix display of origin when rotated

2010-03-10T23:02:44+00:00

fix display of origin when rotated

the origin rotation can't be inside the display list, because the
rotation varies

[D H] lib/python/rs274/glcanon.py

remove redundant method

2010-03-10T22:55:23+00:00

remove redundant method

this method is defined in the base class, Translated.

[D H] lib/python/rs274/glcanon.py

remove unneeded initializations

2010-03-10T22:54:59+00:00

remove unneeded initializations

these fields will be initialized by the first call to set_origin_offsets
or set_xy_rotation, so these statements were not needed

[D H] lib/python/rs274/glcanon.py
[D H] lib/python/rs274/interpret.py

Merge remote branch 'origin/v2.4_branch'

2010-03-09T20:06:16+00:00

Merge remote branch 'origin/v2.4_branch'

Conflicts:
	src/emc/usr_intf/axis/scripts/axis.py

[D H] src/emc/usr_intf/stepconf/stepconf.py

fix a memory leak

2010-03-09T20:00:51+00:00

fix a memory leak

By binding a method of AxisCanon, the instance becomes
uncollectable.  This reformulation still has a small leak
(one Tkinter command binding) but it is not the several
megabytes that leaking an AxisCanon could cause.

[D H] src/emc/usr_intf/axis/scripts/axis.py

fix a memory leak

2010-03-09T19:58:55+00:00

fix a memory leak

By binding a method of AxisCanon, the instance becomes
uncollectable.  This reformulation still has a small leak
(one Tkinter command binding) but it is not the several
megabytes that leaking an AxisCanon could cause.

[D H] src/emc/usr_intf/axis/scripts/axis.py

add static qualifiers where appropriate

2010-03-09T19:49:10+00:00

add static qualifiers where appropriate

[D H] src/emc/rs274ngc/gcodemodule.cc

get rid of unneeded declarations

2010-03-09T19:48:57+00:00

get rid of unneeded declarations

[D H] src/emc/rs274ngc/gcodemodule.cc

get rid of unused variable

2010-03-09T19:48:37+00:00

get rid of unused variable

[D H] src/emc/rs274ngc/gcodemodule.cc

get rid of some emc1-compatibility defines

2010-03-09T19:45:48+00:00

get rid of some emc1-compatibility defines

[D H] src/emc/rs274ngc/gcodemodule.cc

simplify maybe_new_line with default argument

2010-03-09T19:39:27+00:00

simplify maybe_new_line with default argument

Based on a patch from Alexey Starikovskiy <astarikovskiy@suse.de>

[D H] src/emc/rs274ngc/gcodemodule.cc

avoid extra work when there's no rotation

2010-03-09T15:59:58+00:00

avoid extra work when there's no rotation

Based on a patch from Alexey Starikovskiy <astarikovskiy@suse.de>

[D H] lib/python/rs274/interpret.py

avoid unneeded work when sequence_number is unchanged

2010-03-09T15:59:09+00:00

avoid unneeded work when sequence_number is unchanged

Based on a patch from Alexey Starikovskiy <astarikovskiy@suse.de>

[D H] src/emc/rs274ngc/gcodemodule.cc

Fix period, maxvel calculations in axis test

2010-03-09T13:20:13+00:00

Fix period, maxvel calculations in axis test

The calculations of period and maxvel must account for systems with
long steplen values.

Closes SF#2924185, SF#2924173

[D H] src/emc/usr_intf/stepconf/stepconf.py

More clamping room

2010-03-08T08:53:23+00:00

I need to implement tags on the last cut, so that the part can be held in place.
This part separated when I wasn't expecting it.
It didn't break the tool, luckily.
I have enough done to experiment with drilling the holes on the side.
I have now got a 6mm slot drill, which will quarter the cutting time.

Delrin parts

2010-03-08T07:07:09+00:00

I have been given some step files from someone who wants some parts making from Delrin.
He wants 8 of this part. The first thing is to cut the circle out of the middle.
I haven't got the Delrin, yet, so I am having a go with some wood that I have found in the workshop.

Next to clear the lower level with a pocket operation.
I made the shapes to cut by doing Edge to Sketch, on some circular edges, then making circles with "centre" and "end" ticked in "digitizing". Then I made the circles to sketches, then to faces. Then I subtracted one face from another and made another face from a sketch across the middle, and subtractacted that. Finally I did Face to Sketch to get me the sketches I wanted.
I am doing it all with my 3mm slot drill, so I don't have to set the z height of the tool again. This does mean a lot of tool path, though.

Drop-cutter again

2010-03-05T17:39:37+00:00

This is about my third rewrite of this fairly simple cam-algorithm where the cutter is dropped from above until it touches a triangle. It’s now in C++ with Boost-python bindings and with visualization using vtk.

The cutter-location points are calculated by bringing the cutter into contact with the vertices of the triangle (green cl-points), the edges (red cl-points), and the facet (blue cl-points). Then the cl-point with the highest Z-value is selected as the final cutter location. At the white points we did not make contact with the triangle at all.

If you look closely enough, all surfaces in the world are made of triangles, even Tux:

Due to compression, the video might not show enough detail, so here’s a screenshot:

I wonder if anyone is still interested in this stuff? Given enough time I would like to develop waterline-paths and an octree-based cutting simulator also. It would help if these algorithms were incorporated in a CAD-program, or someone would develop a GUI.

Merge branch 'v2.4_branch'

2010-03-04T14:51:27+00:00

Merge branch 'v2.4_branch'

* v2.4_branch:
  Don't reference gpios that dont exist on the 7i43
  revive the firmware/driver interface documentation

[D H] configs/hm2-stepper/hm2-stepper.hal

Don't reference gpios that dont exist on the 7i43

2010-03-04T14:49:18+00:00

Don't reference gpios that dont exist on the 7i43

The 7i43 only has 48 GPIOs, so referencing gpios 48 and above was causing
this config to not load on the 7i43.  This commit fixes SF bug #2963158,
thanks to FlyingElectron for the bug report.

[D H] configs/hm2-stepper/hm2-stepper.hal

revive the firmware/driver interface documentation

2010-03-03T17:59:14+00:00

revive the firmware/driver interface documentation

This file is mesa-hostmot2/firmware/src/regmap from
961df911035af110e900a75164cbe01289a2532c, which was the last commit
before the file was removed.

[D H] src/hal/drivers/mesa-hostmot2/doc/regmap

revive the firmware/driver interface documentation

2010-03-03T17:59:14+00:00

revive the firmware/driver interface documentation

This file is mesa-hostmot2/firmware/src/regmap from
961df911035af110e900a75164cbe01289a2532c, which was the last commit
before the file was removed.

[D H] src/hal/drivers/mesa-hostmot2/doc/regmap

Merge commit 'origin/v2.4_branch'

2010-03-03T16:57:55+00:00

Merge commit 'origin/v2.4_branch'

* commit 'origin/v2.4_branch':
  Support kernels 2.6.29 and newer (ie Lucid Lynx)
  turn off some stray execute bits on config files
  fix another annoying lintian warning
  Switch a changelog entry to cradek's real email address
  Dont install the INSTALL file
  remove src/config.{log,status} on "make clean"
  fix a lintian deprecation warning
  let the debhelper tools run ldconfig for us on install
  Call the debhelper scripts in a better order
  let the user specify the branch name
  fix debian package version numbers on the buildbot

[D H] configs/common/configurable_options/pyvcp/m5i20panel.xml
[D H] configs/plasma-thc/mp1000A.odg
[D H] configs/plasma-thc/mp1000A_parport.hal
[D H] configs/plasma-thc/mp1000B_parport.hal
[D H] configs/plasma-thc/mp1000_serport.hal
[D H] configs/plasma-thc/thc300.ini
[D H] configs/plasma-thc/thc300.odg
[D H] configs/plasma-thc/thc300_parport.hal
[D H] configs/sim/sim-xyyz.hal
[D H] debian/changelog
[D H] debian/control.in
[D H] debian/emc2.postinst
[D H] debian/rules.in
[D H] src/Makefile
[D H] src/hal/drivers/mesa-hostmot2/bitfile.c
[D H] src/hal/drivers/mesa-hostmot2/bitfile.h
[D H] src/hal/drivers/mesa-hostmot2/hm2_7i43.c
[D H] src/hal/drivers/mesa-hostmot2/hm2_pci.c
[D H] src/hal/drivers/mesa-hostmot2/hostmot2.c

French translation update

2010-03-03T14:24:51+00:00

French translation update

Signed-off-by: Francis Tisserant <tissf@free.fr>

[D H] src/po/fr.po

French translation update

2010-03-03T14:24:51+00:00

French translation update

Signed-off-by: Francis Tisserant <tissf@free.fr>

[D H] src/po/fr.po

hostmot2-firmware 0.5

2010-03-03T13:22:41+00:00

I'm pleased to announce the next release of the hostmot2 firmware packages. These packages are compatible with emc 2.3.5 and above, and will be the standard firmware packages for emc 2.4.

Changes since the 0.4 release:

New firmwares: i22-*/SVST2_4_7I47, i20/SV12IM_2X7I48
Symlinks for compatibility with emc 2.3 firmware locations
.PIN files in documentation directory, not in /lib/firmware

Downloads

These files are also available as packages for Ubuntu "Hardy Heron" on the linuxcnc.org package server. If you use a different system than Hardy Heron, then these downloads may be useful to you.

Files currently attached to this page:

hostmot2-firmware-0.5.tar.gz	128.1kB
hostmot2-firmware-bin-3x20-1-0.5.tar.gz	296.8kB
hostmot2-firmware-bin-4i65-0.5.tar.gz	338.6kB
hostmot2-firmware-bin-4i68-0.5.tar.gz	402.2kB
hostmot2-firmware-bin-5i20-0.5.tar.gz	677.3kB
hostmot2-firmware-bin-5i22-1-0.5.tar.gz	404.6kB
hostmot2-firmware-bin-5i22-1.5-0.5.tar.gz	466.6kB
hostmot2-firmware-bin-5i23-0.5.tar.gz	400.0kB
hostmot2-firmware-bin-7i43-2-0.5.tar.gz	146.4kB
hostmot2-firmware-bin-7i43-4-0.5.tar.gz	253.7kB

RELEASED: emc 2.3.5

2010-03-03T13:15:49+00:00

I'm pleased to announce the next bugfix release in the emc 2.3 series, emc 2.3.5. A number of problems in earlier releases have been fixed.

The new packages are now available for Ubuntu 8.04 "Hardy Heron" in the package repositories. Packages for Ubuntu 6.06 "Dapper Drake" should follow within the next few days. If you have already installed an earlier 2.3.x version of emc, this will be offered as an update.

If you still have 2.2.8 or earlier installed, 2.3.5 will not be available as an automatic upgrade. For more information and installation/upgrade instructions, please see the wiki

git users can get this release by checking out the v2.3.5 tag (or follow bugfixes on v2_3_branch). Source tarballs are available from the package repository: http://linuxcnc.org/hardy/dists/hardy/emc2.3/source

Many thanks to the people who have contributed bugfixes for this release.

Change since 2.3.5:

fix jog increments for mini
interpreter: detect gouging in last corner before turning off cutter comp
interpreter: fix another case of a skipped full-circle arc
task: fix toolchange at quill-up and g30 positions (SF#2941688)
stepconf: fix direction of 2nd and 3rd parports
configs: add smithy 516 sample configuration
image2gcode: respect current TLO and fixture settings
image2gcode: "extend image" crash on non-square image (SF#2947390)
documentation improvements
hostmot2: Accept "idrom v3" firmwares. these are the firmwares that will be released with emc 2.4.x.
axis: mitigate a memory leak each time a program is loaded or reloaded

Support kernels 2.6.29 and newer (ie Lucid Lynx)

2010-03-02T21:22:07+00:00

Support kernels 2.6.29 and newer (ie Lucid Lynx)

Before 2.6.29, request_firmware() gave the driver the firmware in a
writable buffer.  The hostmot2 driver needs to reverse the bits of each
byte in the .BIT file before sending it to the FPGA.  Up until this commit
the driver re-wrote the buffer in place.  Kernels 2.6.29 and newer give
the driver the firmware in a read-only buffer, so the driver page-faults
when it tries to write to the buffer.

This commit makes it so it doesnt try to write to the firmware buffer,
instead it does the rewriting on the fly as it's sending the data to
the FPGA.

Here's the linux kernel commit that made struct firmware's data pointer
read-only:

    <http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=6e03a201bbe8137487f340d26aa662110e324b20>

[D H] src/Makefile
[D H] src/hal/drivers/mesa-hostmot2/bitfile.c
[D H] src/hal/drivers/mesa-hostmot2/bitfile.h
[D H] src/hal/drivers/mesa-hostmot2/hm2_7i43.c
[D H] src/hal/drivers/mesa-hostmot2/hm2_pci.c
[D H] src/hal/drivers/mesa-hostmot2/hostmot2.c

Fix case when no vector was selected

2010-03-01T16:54:06+00:00

Fix case when no vector was selected

Signed-off-by: Michael Geszkiewicz <micges@wp.pl>

[D H] lib/python/rs274/glcanon.py

Remove unused varables and minor redundant code

2010-03-01T16:47:32+00:00

Remove unused varables and minor redundant code

Signed-off-by: Michael Geszkiewicz <micges@wp.pl>

[D H] src/emc/usr_intf/axis/scripts/axis.py

Allow the user to specify a stepping waveform

2010-02-28T15:26:58+00:00

Allow the user to specify a stepping waveform

Slavko Kocjancic wanted a custom step type for his homemade stepper
driver.  This adds a (single) customizable step waveform.  These
waveforms have the same limitations as the existing ones: up to 5 output
phases, and up to 10 steps per cycle.

[D H] docs/man/man9/stepgen.9
[D H] src/hal/components/stepgen.c

turn off some stray execute bits on config files

2010-02-27T22:30:10+00:00

turn off some stray execute bits on config files

[D H] configs/common/configurable_options/pyvcp/m5i20panel.xml
[D H] configs/plasma-thc/mp1000A.odg
[D H] configs/plasma-thc/mp1000A_parport.hal
[D H] configs/plasma-thc/mp1000B_parport.hal
[D H] configs/plasma-thc/mp1000_serport.hal
[D H] configs/plasma-thc/thc300.ini
[D H] configs/plasma-thc/thc300.odg
[D H] configs/plasma-thc/thc300_parport.hal
[D H] configs/sim/sim-xyyz.hal

fix another annoying lintian warning

2010-02-27T05:46:02+00:00

fix another annoying lintian warning

The "Description" field of the control file is not supposed to be a
sentence, it's suposed to be a short descriptive phrase without a period
at the end:

    <http://emc2-buildbot.colorado.edu/buildbot/builders/deb-hardy-sim-binary-i386/builds/85/steps/shell_5/logs/stdio>

[D H] debian/control.in

Switch a changelog entry to cradek's real email address

2010-02-27T05:33:57+00:00

Switch a changelog entry to cradek's real email address

Debian policy requires that each entry in the changelog be "signed"
by a real email address.  Lintian caught the little problem:

    <http://emc2-buildbot.colorado.edu/buildbot/builders/deb-hardy-rt-binary-i386/builds/82/steps/shell_5/logs/stdio>

[D H] debian/changelog

Dont install the INSTALL file

2010-02-27T05:15:02+00:00

Dont install the INSTALL file

The INSTALL file tells how to compile and install (or run-in-place)
emc2, which is not useful in the .deb.  Also, lintian complains about it:

    <http://emc2-buildbot.colorado.edu/buildbot/builders/deb-hardy-rt-binary-i386/builds/82/steps/shell_5/logs/stdio>

[D H] src/Makefile

remove src/config.{log,status} on "make clean"

2010-02-27T01:22:47+00:00

remove src/config.{log,status} on "make clean"

These files are generated by configure.  This patch is so we don't have
these files polluting our source debs, as pointed out by lintian here:

    <http://emc2-buildbot.colorado.edu/buildbot/builders/master-package-rt-hardy-source/builds/83/steps/shell/logs/stdio>

[D H] src/Makefile

fix a lintian deprecation warning

2010-02-27T00:57:55+00:00

fix a lintian deprecation warning

I guess Source-Version is spelled binary:Version now...  The (now-fixed)
lintian warning is here:

    <http://emc2-buildbot.colorado.edu/buildbot/builders/master-package-rt-hardy-source/builds/83/steps/shell/logs/stdio>

[D H] debian/control.in

let the debhelper tools run ldconfig for us on install

2010-02-26T23:59:46+00:00

let the debhelper tools run ldconfig for us on install

dh_makeshlibs notices that we install shared libraries, and wants to add
a section to emc2.postinst to run ldconfig.  This commit puts the magic
marker-comment into emc2.postinst that gives it permission to add it.

This fixes a warning from lintian:

    <http://emc2-buildbot.colorado.edu/buildbot/builders/master-package-sim-hardy-source/builds/84/steps/shell/logs/stdio>

[D H] debian/emc2.postinst

Call the debhelper scripts in a better order

2010-02-26T21:45:02+00:00

Call the debhelper scripts in a better order

dh_makeshlibs can make metadata files used by dh_installdeb, so it should
be called first.  This fixes a warning pointed out by lintian:

    <http://emc2-buildbot.colorado.edu/buildbot/builders/master-package-rt-hardy-source/builds/81/steps/shell/logs/stdio>

[D H] debian/rules.in

let the user specify the branch name

2010-02-26T19:01:13+00:00

let the user specify the branch name

The buildbot interface to git does not record what branch it's working on
in the repo itself, so this info needs to be supplied from the outside,
via the commandline.

[D H] debian/update-dch-from-git

let the user specify the branch name

2010-02-26T19:01:13+00:00

let the user specify the branch name

The buildbot interface to git does not record what branch it's working on
in the repo itself, so this info needs to be supplied from the outside,
via the commandline.

[D H] debian/update-dch-from-git

fix debian package version numbers on the buildbot

2010-02-26T15:09:15+00:00

fix debian package version numbers on the buildbot

"git describe" shows the most recent (timewise) tag in the history.
When we merge a branch that has tags into another branch, for example
merging 2.4 into master, the tags from both branches get searched by
git-describe.

Since we use tags to make the version numbers of our debian packages,
this can lead to incorrect debian versions.  Debs in the master repos are
always builds of the master branch, but they sometimes would incorrectly
claim to be 2.4 debs....  Debs in the 2.4 repos are always builds of the
2.4 branch, and they always have correct 2.4 version numbers (since we
haven't merged any branch with tags into 2.4 yet).

This commit fixes the way we run git-describe so it only searches for
tags appropriate for the current branch.

[D H] debian/update-dch-from-git

fix debian package version numbers on the buildbot

2010-02-26T15:09:15+00:00

fix debian package version numbers on the buildbot

"git describe" shows the most recent (timewise) tag in the history.
When we merge a branch that has tags into another branch, for example
merging 2.4 into master, the tags from both branches get searched by
git-describe.

Since we use tags to make the version numbers of our debian packages,
this can lead to incorrect debian versions.  Debs in the master repos are
always builds of the master branch, but they sometimes would incorrectly
claim to be 2.4 debs....  Debs in the 2.4 repos are always builds of the
2.4 branch, and they always have correct 2.4 version numbers (since we
haven't merged any branch with tags into 2.4 yet).

This commit fixes the way we run git-describe so it only searches for
tags appropriate for the current branch.

[D H] debian/update-dch-from-git

Merge branch 'v2.4_branch'

2010-02-26T04:15:46+00:00

Merge branch 'v2.4_branch'

[D H] src/emc/rs274ngc/interp_internal.cc

fix silly warning

2010-02-26T04:14:41+00:00

fix silly warning

[D H] src/emc/rs274ngc/interp_internal.cc

Merge commit 'origin/v2.4_branch'

2010-02-26T02:50:39+00:00

Merge commit 'origin/v2.4_branch'

[D H] lib/python/rs274/glcanon.py
[D H] lib/python/rs274/interpret.py
[D H] src/configure.in
[D H] src/emc/iotask/ioControl.cc
[D H] src/emc/rs274ngc/gcodemodule.cc
[D H] src/emc/rs274ngc/interp_find.cc
[D H] src/emc/rs274ngc/rs274ngc_pre.cc
[D H] src/emc/task/emctaskmain.cc
[D H] src/emc/usr_intf/axis/scripts/axis.py
[D H] src/emc/usr_intf/pncconf/pncconf.py
[D H] src/hal/drivers/hal_parport.c
[D H] src/hal/drivers/probe_parport.c
[D H] src/hal/halmodule.cc
[D H] src/hal/utils/halcmd.c

set diameter to zero when unloading a tool

2010-02-26T02:39:16+00:00

set diameter to zero when unloading a tool

on a nonrandom machine this showed up as an incorrectly
persistent diameter after T0M6, as seen by a subsequent
G41.

[D H] src/emc/iotask/ioControl.cc

make (AXIS,stop) take effect immediately

2010-02-26T02:15:20+00:00

make (AXIS,stop) take effect immediately

before, the special comment would wait until the next call to
check_abort, which could be many lines of gcode later.  Now the
abort will happen right away, using the same exception as check_abort

[D H] lib/python/rs274/glcanon.py

redraw the preview when the characteristics of the tool change

2010-02-26T02:09:11+00:00

redraw the preview when the characteristics of the tool change

before this was based on tool number only which was wrong when
reloading the tool table or changing the tool with g10 l1

[D H] src/emc/usr_intf/axis/scripts/axis.py

fix preview of cutter comp

2010-02-26T02:01:33+00:00

fix preview of cutter comp

this was broken when random_toolchange was merged.  now axis will
do the tool table manipulation that iocontrol does, so that the
preview comes out right

[D H] lib/python/rs274/interpret.py
[D H] src/emc/usr_intf/axis/scripts/axis.py

make sure we have the right tool table after a tool change

2010-02-26T01:47:53+00:00

make sure we have the right tool table after a tool change

task happens to get this right (by calling synch) but axis doesn't.
it makes more sense for the interpreter to be aware of times its idea
of the tool table may be out of date, such as right after pockets have
been shuffled around.

[D H] src/emc/rs274ngc/rs274ngc_pre.cc

these parameters are now pockets, not tools

2010-02-26T01:46:22+00:00

these parameters are now pockets, not tools

[D H] src/emc/rs274ngc/gcodemodule.cc

clear interpreter list on error

2010-02-26T01:35:41+00:00

clear interpreter list on error

the case that came to my attention was with RFL, G1 / G2 starting on the
G2.  The linear move from the G1 wasn't cleared, so RFL went to its
endpoint after the "arc endpoints differ" message.  It also fixes a
problem Chris noticed that isn't related to RFL, where program
execution would continue after encountering the error "Specified tool
not found in tool table".

[D H] src/emc/task/emctaskmain.cc

Fix RFL on line that does not specify all axis words

2010-02-26T01:19:31+00:00

Fix RFL on line that does not specify all axis words

CANON_UPDATE_END_POINT is not enough to get the actual start point of the
first non-skipped line to the interpreter.  If the following line was
'G0Z0' it would go to Z0 and the X and Y location of the previous (skipped)
move.  The same was true of an M6 toolchange when it uses the quill-up
location

[D H] src/emc/task/emctaskmain.cc

make this error message less mysterious

2010-02-25T23:58:08+00:00

make this error message less mysterious

[D H] src/emc/rs274ngc/interp_find.cc

Fixed two typos in 924.ini.

2010-02-25T17:48:39+00:00

Fixed two typos in 924.ini.

[D H] configs/smithy/924.ini

Work around a numarray slicing bug (SF#2947390)

2010-02-21T15:18:39+00:00

Work around a numarray slicing bug (SF#2947390)

On some platforms, numarray slicing seems to be broken.  On these platforms,
a slice like [4:5] gives 5 items instead of 1 item:
    # Python 2.6.4, numarray 1.5.2, numpy 1.3.0, amd64 (Ubuntu 9.10)
    >>> l = [0]*10; z1 = numarray.zeros((10,)); z2 = numpy.zeros((10,))
    >>> len(l[4:5]), len(z1[4:5]), len(z2[4:5])
    (1, 5, 1)

On other platforms, numarray slicing behaves as expected:
    # Python 2.4.3, numarray 1.5.1, no numpy, i386 (Ubuntu 6.06)
    >>> l = [0]*10; z1 = numarray.zeros((10,))
    >>> len(l[4:5]), len(z1[4:5])
    (1, 1)

The same problem applies to slicing of 2D arrays, which is one of the
primary activities of image-to-gcode.

The reformulation of
    arr[l:l+w, t:t+h]
into
    arr[l:,t:][:w,:h]
gives a slice of the desired items even on systems with broken numarray.

[D H] src/emc/usr_intf/axis/scripts/image-to-gcode.py