linuxcnc.git
6 years agoRelease 2.5.5 v2.5_branch
Chris Radek [Thu, 12 May 2016 01:04:23 +0000 (20:04 -0500)]
Release 2.5.5

Signed-off-by: Chris Radek <chris@timeguy.com>
6 years agoincrease stepgen to 16
John Thornton [Wed, 10 Jul 2013 17:24:31 +0000 (12:24 -0500)]
increase stepgen to 16
Signed-off-by: John Thornton <jthornton@gnipsel.com>
(cherry picked from commit 61ebd8221eedad3d22b9e679c7857574b10c57d4)

6 years agoDocs: update location of iso files
John Thornton [Mon, 4 Apr 2016 12:00:29 +0000 (07:00 -0500)]
Docs: update location of iso files
some people with old hardware may need to restore with older versions
of EMC2.

Signed-off-by: John Thornton <bjt128@gmail.com>
7 years agoFix pin-1-in-not
Chris Radek [Tue, 8 Dec 2015 17:33:37 +0000 (11:33 -0600)]
Fix pin-1-in-not

... which apparently nobody has ever used

Signed-off-by: Chris Radek <chris@timeguy.com>
7 years agointerp: fix an old bug in canned cycle preliminary & in-between moves
Sebastian Kuzminsky [Sat, 24 Oct 2015 05:37:16 +0000 (23:37 -0600)]
interp: fix an old bug in canned cycle preliminary & in-between moves

The rules (for the XY plane, to simplify the terminology) are as follows:

1. (Preliminary) if OLD_Z is below R, rapid up to R
2. (In-between) rapid to XY location of next cycle
3. (In-between) rapid down to R, if needed
4. Perform cycle (leaves controlled point at CLEAR_Z)
5. If there are more cycles to do, goto 2.

Signed-off-by: Sebastian Kuzminsky <seb@highlab.com>
7 years agointerp: comment the "rapid up to R if needed" code
Sebastian Kuzminsky [Sat, 24 Oct 2015 05:29:40 +0000 (23:29 -0600)]
interp: comment the "rapid up to R if needed" code

This adds a comment that explains the previous bugfix commit (0b552aa5).

Signed-off-by: Sebastian Kuzminsky <seb@highlab.com>
7 years agoMerge remote-tracking branch 'origin/v2.4_branch' into v2.5_branch
Sebastian Kuzminsky [Sat, 24 Oct 2015 06:00:31 +0000 (00:00 -0600)]
Merge remote-tracking branch 'origin/v2.4_branch' into v2.5_branch

7 years agog81 test: add repeat cycles
Sebastian Kuzminsky [Sat, 24 Oct 2015 00:11:35 +0000 (18:11 -0600)]
g81 test: add repeat cycles

This commit adds coverage of L-word repeats on G81 drill cycles.

Signed-off-by: Sebastian Kuzminsky <seb@highlab.com>
7 years agofix canned cycles where old Z is below retract plane
Sebastian Kuzminsky [Thu, 22 Oct 2015 01:58:46 +0000 (19:58 -0600)]
fix canned cycles where old Z is below retract plane

Signed-off-by: Sebastian Kuzminsky <seb@highlab.com>
7 years agoupdate g81 tests for new init/shutdown Canon calls in 2.5
Sebastian Kuzminsky [Thu, 22 Oct 2015 01:57:19 +0000 (19:57 -0600)]
update g81 tests for new init/shutdown Canon calls in 2.5

Signed-off-by: Sebastian Kuzminsky <seb@highlab.com>
7 years agoMerge remote-tracking branch 'origin/v2.4_branch' into v2.5_branch
Sebastian Kuzminsky [Fri, 23 Oct 2015 04:53:56 +0000 (22:53 -0600)]
Merge remote-tracking branch 'origin/v2.4_branch' into v2.5_branch

Conflicts:
scripts/platform-is-supported (picked the version in 2.5)

7 years agotests: add an interpreter test of G81
Sebastian Kuzminsky [Thu, 22 Oct 2015 01:01:35 +0000 (19:01 -0600)]
tests: add an interpreter test of G81

This uses SAI to run G81 in different ways and verifies the resulting
Canon calls.

Tests G81 in G17, G18, and G19.

Signed-off-by: Sebastian Kuzminsky <seb@highlab.com>
7 years agoadd platform-is-supported script in 2.4
Sebastian Kuzminsky [Thu, 22 Oct 2015 23:38:45 +0000 (17:38 -0600)]
add platform-is-supported script in 2.4

This script lets the buildbot know what platforms to build the branch on.
2.4 doesn't build on newer distros.

Signed-off-by: Sebastian Kuzminsky <seb@highlab.com>
7 years agopackaging: dont build 2.5 on debian jessie
Sebastian Kuzminsky [Tue, 25 Aug 2015 03:03:21 +0000 (21:03 -0600)]
packaging: dont build 2.5 on debian jessie

Jessie uses tcl 8.6, which is too new for 2.5.

Signed-off-by: Sebastian Kuzminsky <seb@highlab.com>
7 years agoFix a dubious type cast, from machinekit dac83399
Chris Radek [Thu, 23 Apr 2015 19:24:46 +0000 (14:24 -0500)]
Fix a dubious type cast, from machinekit dac83399

Signed-off-by: Chris Radek <chris@timeguy.com>
7 years agohal/library/hal_link: fix fatal memory corruption bug on linking pin to a signal
Michael Haberler [Wed, 11 Mar 2015 07:51:37 +0000 (08:51 +0100)]
hal/library/hal_link: fix fatal memory corruption bug on linking pin to a signal

Original two log messages:

see also https://github.com/machinekit/machinekit/issues/524

hal/unlink_pin: assure pin value on unlink is reset properly

This was introduced in 51a0afc76, and while not fatal,
this line

  *(hal_data_u *)dummy_addr = *sig_data_addr;

potentially restores more bytes from the signal than intended
(eg on a bit pin it will still store sizeof(double) bytes)
which may cause dummysig to hold confusing values when
inspected with a debugger

I (cradek) had to massage these commits somewhat to apply after mk's
refactoring etc.

My how-to-reproduce case:

halcmd: newsig s1 u32
halcmd: newsig s2 u32
halcmd: newsig s3 u32
halcmd: sets s1 0xaaaaaaaa
halcmd: sets s3 0x55555555
halcmd: show sig
Signals:
Type          Value  Name     (linked to)
u32      0xAAAAAAAA  s1
u32      0x00000000  s2
u32      0x55555555  s3

halcmd: net s2 => conv-u32-s32.0.in
halcmd: show sig
Signals:
Type          Value  Name     (linked to)
u32      0xAAAAAAAA  s1
u32      0x00000000  s2
 ==> conv-u32-s32.0.in
u32      0x00000000  s3

Notice s3 is overwritten.

Signed-off-by: Chris Radek <chris@timeguy.com>
8 years agoFix shift-jog speed being too slow on inch configs displaying mm
Chris Radek [Wed, 29 Oct 2014 16:22:19 +0000 (11:22 -0500)]
Fix shift-jog speed being too slow on inch configs displaying mm

Fixes the second part of bug #397

Signed-off-by: Chris Radek <chris@timeguy.com>
8 years agoFix UVW jogs being too fast by 25.4x, on inch configs displaying mm
Chris Radek [Wed, 29 Oct 2014 16:01:30 +0000 (11:01 -0500)]
Fix UVW jogs being too fast by 25.4x, on inch configs displaying mm

Signed-off-by: Chris Radek <chris@timeguy.com>
8 years agoMerge remote-tracking branch 'jepler/signed-off-by' into v2.5_branch
Jeff Epler [Sun, 28 Sep 2014 01:02:30 +0000 (20:02 -0500)]
Merge remote-tracking branch 'jepler/signed-off-by' into v2.5_branch

8 years agoMerge remote-tracking branch 'origin/v2.5_branch' into jepler/signed-off-by
Jeff Epler [Sat, 23 Aug 2014 18:14:38 +0000 (13:14 -0500)]
Merge remote-tracking branch 'origin/v2.5_branch' into jepler/signed-off-by

8 years agoExplicitly give the right build dependency for insmod/rmmod/etc
Chris Radek [Wed, 26 Mar 2014 16:35:47 +0000 (11:35 -0500)]
Explicitly give the right build dependency for insmod/rmmod/etc

(cherry picked from commit 39e36ebd8fa5530ba87e162c17bac2c5ce57621f)

Conflicts:

debian/configure
debian/control.in

8 years agoFail on missing programs, instead of generating bogus scripts
Chris Radek [Tue, 25 Mar 2014 16:44:01 +0000 (11:44 -0500)]
Fail on missing programs, instead of generating bogus scripts

(cherry picked from commit 0b5fba46aba7b35c8c3bcb80de43dec0ba8e97da)

8 years agoThese are needed for building and running on all platforms
Chris Radek [Tue, 25 Mar 2014 23:59:02 +0000 (18:59 -0500)]
These are needed for building and running on all platforms

(cherry picked from commit 3b03cefb49d964adaa9a83dc0f1775c4fdb97770)

Conflicts:
debian/control.in

8 years agodeb: provide build-arch and build-indep targets
Sebastian Kuzminsky [Wed, 20 Aug 2014 15:09:15 +0000 (09:09 -0600)]
deb: provide build-arch and build-indep targets

This is required by Debian Policy Manual section 4.9:

    https://www.debian.org/doc/debian-policy/ch-source.html#s-debianrules

8 years agoFrench doc update: fix startup code example
Francis Tisserant [Tue, 12 Aug 2014 09:36:36 +0000 (11:36 +0200)]
French doc update: fix startup code example

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agoDCO: Customize for LinuxCNC
Jeff Epler [Thu, 31 Jul 2014 13:53:22 +0000 (08:53 -0500)]
DCO: Customize for LinuxCNC

.. noting that we require all submitted code to be licensed
compatibly with the GPLv2 with "or later" clause

Signed-off-by: Jeff Epler <jepler@unpythonic.net>
8 years agoDCO: Move docs to the right location for linuxcnc
Jeff Epler [Thu, 31 Jul 2014 13:35:44 +0000 (08:35 -0500)]
DCO: Move docs to the right location for linuxcnc

Signed-off-by: Jeff Epler <jepler@unpythonic.net>
8 years agodocs: remove some broken asciidoc markup
Sebastian Kuzminsky [Mon, 28 Jul 2014 06:10:27 +0000 (00:10 -0600)]
docs: remove some broken asciidoc markup

(Variables) looks like an typoed attempt at creating an asciidoc anchor,
but it's not valid asciidoc anchor syntax so it doesnt work, and it
passes through into the generated docs, where it makes no sense.

None of the other asciidocs tries to use an anchor named Variables,
so just remove it.

8 years agoDocs: fix startup code example.
John Thornton [Fri, 25 Jul 2014 12:42:29 +0000 (07:42 -0500)]
Docs: fix startup code example.
g1 without a feed rate is an error and is a bad example for the startup code
thanks to archivist for spotting this
Signed-off-by: John Thornton <jthornton@gnipsel.com>
8 years agopncconf -fix error with firmware with more then 5 sserial channels
Chris Morley [Thu, 17 Jul 2014 05:49:01 +0000 (22:49 -0700)]
pncconf -fix error with firmware with more then 5 sserial channels

Pncconf can only display 5 channels but the code checked for 8
in certain circumstances.

8 years agoFrench doc update: fix float value with old data
Francis Tisserant [Mon, 14 Jul 2014 20:00:26 +0000 (22:00 +0200)]
French doc update: fix float value with old data

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agoFrench translation update
Francis Tisserant [Mon, 14 Jul 2014 19:33:16 +0000 (21:33 +0200)]
French translation update

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agopncconf -fix an icompatibility between MESA XMLs and linuxcnc XMLs
Chris Morley [Sat, 12 Jul 2014 06:31:57 +0000 (23:31 -0700)]
pncconf -fix an icompatibility between MESA XMLs and linuxcnc XMLs

Linuxcnc used '???' to designate no secondary function, just GPIO,
where MESA use 'not used'. now both work
This made some GPIO pind not show in pncconf

8 years agoDocs: try to word synopsis to be a bit clearer in the description
John Thornton [Wed, 2 Jul 2014 13:38:52 +0000 (08:38 -0500)]
Docs: try to word synopsis to be a bit clearer in the description

Signed-off-by: John Thornton <jthornton@gnipsel.com>
8 years agoDocs: fix float value with old data
John Thornton [Wed, 2 Jul 2014 13:27:06 +0000 (08:27 -0500)]
Docs: fix float value with old data

Signed-off-by: John Thornton <jthornton@gnipsel.com>
8 years agoFrench doc update: Document G5,G5.1,G5.2,G5.3 NURBS/spline commands
Francis Tisserant [Wed, 18 Jun 2014 20:40:33 +0000 (22:40 +0200)]
French doc update: Document G5,G5.1,G5.2,G5.3 NURBS/spline commands

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agoFrench doc update and cleaning
Francis Tisserant [Wed, 18 Jun 2014 14:34:31 +0000 (16:34 +0200)]
French doc update and cleaning

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agolinuxcnctop: fix the crawling scrollbar in many cases
Jeff Epler [Wed, 11 Jun 2014 21:28:51 +0000 (16:28 -0500)]
linuxcnctop: fix the crawling scrollbar in many cases

.. by only changing the necessary part of the text, the 'crawl' is usually
avoided.  It might still occur in cases when a changing value is one that
is split across more than two lines (at least this is the pattern I
believed I saw when working on this problem).

It looks and smells like a Tk bug but oh well

8 years agoDocument G5, G5.1, G5.2, G5.3 NURBS/spline commands
Chris Radek [Wed, 4 Jun 2014 19:03:24 +0000 (14:03 -0500)]
Document G5, G5.1, G5.2, G5.3 NURBS/spline commands

This should close bug #369.

8 years agoNURBS: fix biarc G1 continuity
Chris Radek [Sun, 1 Jun 2014 22:08:45 +0000 (17:08 -0500)]
NURBS: fix biarc G1 continuity

The center of the second half of each biarc was wrong.

8 years agoClear saved state from previous NURBSes to avoid bogus motion
Chris Radek [Sun, 1 Jun 2014 20:25:40 +0000 (15:25 -0500)]
Clear saved state from previous NURBSes to avoid bogus motion

Previously if an incorrect NURBS specification was given, such as when
there aren't enough control points, an error would be generated, but a
second load would succeed because there were leftover control points
from last time.  This caused weird behavior such as every second program
load succeeding (when it should have failed every time), and then of
course giving bogus motion when run.

8 years agoDocs: add link
John Thornton [Fri, 30 May 2014 11:19:04 +0000 (06:19 -0500)]
Docs: add link

Signed-off-by: John Thornton <jthornton@gnipsel.com>
8 years agoDocs: add info on the servo axis calibration assistant
John Thornton [Fri, 30 May 2014 11:18:48 +0000 (06:18 -0500)]
Docs: add info on the servo axis calibration assistant

Signed-off-by: John Thornton <jthornton@gnipsel.com>
8 years agotests: verify that comp rejects .comp files whose names dont match the component...
Sebastian Kuzminsky [Thu, 29 May 2014 03:34:28 +0000 (21:34 -0600)]
tests: verify that comp rejects .comp files whose names dont match the component name

8 years agocomp: reject .comp files whose names dont match the component name
Sebastian Kuzminsky [Thu, 29 May 2014 03:35:20 +0000 (21:35 -0600)]
comp: reject .comp files whose names dont match the component name

8 years agoDocs: add info about file naming
John Thornton [Wed, 28 May 2014 11:35:44 +0000 (06:35 -0500)]
Docs: add info about file naming

Signed-off-by: John Thornton <jthornton@gnipsel.com>
8 years agomanpages -update pages on classicladder
Chris Morley [Tue, 27 May 2014 06:35:13 +0000 (23:35 -0700)]
manpages -update pages on classicladder

missing floats and pinnames were wrong for s32 pins

8 years agohal: fix minor bug in allocating hal memory for HAL_S32 and HAL_U32 signals
Michael Geszkiewicz [Wed, 21 May 2014 19:52:47 +0000 (21:52 +0200)]
hal: fix minor bug in allocating hal memory for HAL_S32 and HAL_U32 signals

Signed-off-by: Michael Geszkiewicz <micges@wp.pl>
8 years agohal/link/unlink: revise semantics of hal_link()/hal_unlink():
Michael Haberler [Mon, 17 Mar 2014 08:31:30 +0000 (09:31 +0100)]
hal/link/unlink: revise semantics of hal_link()/hal_unlink():

old behavior: a pin's default value would be overwritten by a signal value
on link, effectively making the default pin value invisible outside the
component, and potentially causing a jump in the pin's value on link.

an unlink would re-apply the pin's hidden default value regardless of the
signal value, also causing a potential jump in the pin's value on unlink.

new behavior: the very first link of a signal to a pin will make the
signal inherit the pin's value. On unlink, the pin's default value will
be set from the current signal value.

The result is:
- no sudden jumps in pin values as a result of link/unlink
- it is now possible to access a pin's default value from other pins
- it is possible to have the changed default value retained on unlink

Change of behavior: Code relying on signal values defaulting to zero
post-linking should be revised to consider:

So far, a new signal value defaulted to 0/0.0/false regardless of
the pin's values linked to it.

Now, the signal will inherit the first linked pin's value.

ref: [Emc-developers] PID bidirectional pins. Mar 16 2014

8 years agodocs: fix hm2 encoder.rawcounts pin name in manpage
Sebastian Kuzminsky [Tue, 20 May 2014 03:15:55 +0000 (23:15 -0400)]
docs: fix hm2 encoder.rawcounts pin name in manpage

The hm2 driver exports pins called "hm2_BLAH.X.encoder.YY.rawcounts",
but the manpage incorrectly called them ".rawcount".  Fix the manpage
to match the existing pin name.

8 years agocomp: remove an unused and confusing global variable
Sebastian Kuzminsky [Fri, 9 May 2014 04:14:09 +0000 (22:14 -0600)]
comp: remove an unused and confusing global variable

This commit removes an unused global variable named __comp_state, from
.c files generated by the comp(1) tool.  The global was never set,
so it never served any actual purpose.

The functions declared by "function" declarations in the .comp file,
and the optional EXTRA_SETUP() function, all take an argument which is
also named __comp_state.  Thus, in these functions, the global was masked
by the function argument of the same name.  The __comp_state variable
is used by the convenience defines to access the hal pins and params of
the component instance.  The functions all worked fine and accessed the
hal pins via their argument, just like they should.

The problem was that the global made the convenience defines compile-time
valid everywhere in the .comp, but at run-time any non-"function",
non-EXTRA_SETUP code would access the uninitialized global and crash.

We're better off without it.

8 years agobiquad: fix a crash-when-enabled bug
Sebastian Kuzminsky [Fri, 9 May 2014 03:56:26 +0000 (21:56 -0600)]
biquad: fix a crash-when-enabled bug

This fixes a bug in the biquad comp that would cause a crash with a NULL
pointer dereference immediately upon the user setting the .enable HAL
pin to True.  Thanks to forum user gaston48 for the bug report.

The biquad comp exports a hal function called "_" in the .comp and .c
files, called "biquad.0" in hal.  This function calls a function internal
to the biquad component (not exported to hal, not handled by comp(1))
called "Biquad_CalcCoeffs".  The Biquad_CalcCoeffs function accesses the
comp instance's hal pins and params, but this behavior is not supported
by the comp(1) tool.

Only functions handled by the comp tool can access the comp instance's hal
pins and params.  That means functions declared with a .comp "function"
declaration, and the EXTRA_SETUP() function.

This commit moves the Biquad_CalcCoeffs code from its separate function
into the main "_" aka "biquad.0" function, this avoiding the disallowed
"access hal from a called function" problem.

8 years agoadd a very basic test of the biquad component
Sebastian Kuzminsky [Fri, 9 May 2014 03:12:13 +0000 (21:12 -0600)]
add a very basic test of the biquad component

8 years agodocs: fix up some comp docs that had rotted a bit
Sebastian Kuzminsky [Fri, 9 May 2014 04:47:18 +0000 (22:47 -0600)]
docs: fix up some comp docs that had rotted a bit

The docs referred to an old name of an internal data structure.  The code
changed in a7a7fd39a but i forgot to update the docs.

The new name includes the underscore character "_", which is meaningful
to asciidoc.  This commit changes the markup delineating the name from
single-tick "'" to backtick "`", to tell asciidoc to pass through the
text without further substitution.  This also changes the rendering
from italic to monospace, which is arguably better anyway (though it's
somewhat inconsistent with the surrounding markup).

This commit also fixes a markup syntax blooper in the FOR_ALL_INSTS()
documentation.

8 years agoFrench doc: cleaning
Francis Tisserant [Tue, 29 Apr 2014 10:21:10 +0000 (12:21 +0200)]
French doc: cleaning

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agoFrench doc: add info about the direction arrow requirements
Francis Tisserant [Mon, 28 Apr 2014 22:10:34 +0000 (00:10 +0200)]
French doc: add info about the direction arrow requirements

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agoDocs: fix spello
John Thornton [Mon, 28 Apr 2014 18:15:42 +0000 (13:15 -0500)]
Docs: fix spello

Signed-off-by: John Thornton <jthornton@gnipsel.com>
8 years agoDocs: add info about the direction arrow requirements
John Thornton [Mon, 28 Apr 2014 18:08:29 +0000 (13:08 -0500)]
Docs: add info about the direction arrow requirements
Direction arrows must be a separate word with a space
between arrow and the pin name

Signed-off-by: John Thornton <jthornton@gnipsel.com>
8 years agoMoves are way too short to make tones, if in mm mode
Chris Radek [Sun, 27 Apr 2014 04:41:20 +0000 (23:41 -0500)]
Moves are way too short to make tones, if in mm mode

8 years agohalsh: initialize stubs library
Dewey Garrett [Fri, 18 Apr 2014 15:16:52 +0000 (08:16 -0700)]
halsh: initialize stubs library

problem appeared for tclStubsPtr symbol with tcl8.6 in (x)ubuntu 14.04

8 years agoedge.comp: fix output pulse width
Sebastian Kuzminsky [Sat, 19 Apr 2014 04:35:43 +0000 (22:35 -0600)]
edge.comp: fix output pulse width

The edge manpage claims that the output pulse width is controllable by
the .out_width_ns parameter, but there's a bug: the output pulse is one
thread period too long.

This commit fixes it.

8 years agoedge.comp: fix .out-invert pin on first invocation
Sebastian Kuzminsky [Sat, 19 Apr 2014 04:34:01 +0000 (22:34 -0600)]
edge.comp: fix .out-invert pin on first invocation

Before this commit, edge.*.out and .out_invert would both be 0 after the
first invocation of the component's function.  This commit fixes
.out_invert to always be the inverse of .out.

8 years agotests: add some tests of the edge component
Sebastian Kuzminsky [Fri, 18 Apr 2014 18:41:37 +0000 (12:41 -0600)]
tests: add some tests of the edge component

8 years agoRelease 2.5.4
Chris Radek [Thu, 17 Apr 2014 16:55:29 +0000 (11:55 -0500)]
Release 2.5.4

8 years agoChangelog for 2.5.4
Chris Radek [Mon, 7 Apr 2014 20:42:35 +0000 (15:42 -0500)]
Changelog for 2.5.4

8 years agopncconf -fix off by one error on pin numbering of 5i25 boards
Chris Morley [Thu, 17 Apr 2014 02:36:36 +0000 (19:36 -0700)]
pncconf -fix off by one error on pin numbering of 5i25 boards

8 years agopncconf -honour 7i43 address changes
Chris Morley [Thu, 17 Apr 2014 02:35:04 +0000 (19:35 -0700)]
pncconf -honour 7i43 address changes

priviously always used default address.

8 years agoadded new configs for ppmc with encoder velocity estimation
elson [Fri, 11 Apr 2014 04:22:04 +0000 (23:22 -0500)]
added new configs for ppmc with encoder velocity estimation

8 years agoedge.comp: eliminate false trigger upon startup
Frank Tkalcevic [Thu, 10 Apr 2014 15:04:48 +0000 (10:04 -0500)]
edge.comp: eliminate false trigger upon startup

Fixing http://sourceforge.net/p/emc/bugs/346/

The edge component, edge.comp, starts with the last input value (last_in) 0, so
if you start the component, and the input is already 1, you will get a false
edge trigger.

8 years agoComponent: fix incorrect calculation of velocity tolerance percent
John Thornton [Thu, 10 Apr 2014 12:09:28 +0000 (07:09 -0500)]
Component: fix incorrect calculation of velocity tolerance percent
the velocity tolerance would end up being much too small as you increased
the value.

Signed-off-by: John Thornton <jthornton@gnipsel.com>
8 years agonew pyvcp_widgets
Michael Haberler [Wed, 2 Apr 2014 18:10:48 +0000 (20:10 +0200)]
new pyvcp_widgets

this fixes http://sourceforge.net/p/emc/bugs/364/

Signed-off-by: Michael Haberler <git@mah.priv.at>
8 years agoFrench doc: cleaning typo
Francis Tisserant [Fri, 28 Mar 2014 19:27:01 +0000 (20:27 +0100)]
French doc: cleaning typo

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agoIgnore auto-generated manpages for two new comps
Chris Radek [Sun, 6 Apr 2014 00:32:54 +0000 (19:32 -0500)]
Ignore auto-generated manpages for two new comps

8 years agoAdd a component to access individual bits from a u32
Andy Pugh [Wed, 2 Apr 2014 21:48:49 +0000 (22:48 +0100)]
Add a component to access individual bits from a u32

Signed-off-by: Andy Pugh <andy@bodgesoc.org>
8 years agoAdd bitwise manipulations of u32.
Andy Pugh [Wed, 2 Apr 2014 21:46:54 +0000 (22:46 +0100)]
Add bitwise manipulations of u32.

Signed-off-by: Andy Pugh <andy@bodgesoc.org>
8 years agoComp: fix spelling error
John Thornton [Wed, 2 Apr 2014 16:29:29 +0000 (11:29 -0500)]
Comp: fix spelling error

Signed-off-by: John Thornton <jthornton@gnipsel.com>
8 years agoFrench doc. update to follow John
Francis Tisserant [Fri, 28 Mar 2014 18:52:38 +0000 (19:52 +0100)]
French doc. update to follow John

Docs: expand a bit on the description to clear up confusion

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agoDocs: expand a bit on the description to clear up confusion
John Thornton [Thu, 27 Mar 2014 22:38:19 +0000 (17:38 -0500)]
Docs: expand a bit on the description to clear up confusion

Signed-off-by: John Thornton <jthornton@gnipsel.com>
8 years agofix some broken links
Francis Tisserant [Wed, 26 Mar 2014 20:52:53 +0000 (21:52 +0100)]
fix some broken links

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agoMerge branch 'ilowpass-wrap' into v2.5_branch
Jeff Epler [Wed, 26 Mar 2014 16:35:28 +0000 (11:35 -0500)]
Merge branch 'ilowpass-wrap' into v2.5_branch

8 years agoilowpass: test for wraparound behavior
Jeff Epler [Wed, 26 Mar 2014 16:12:51 +0000 (11:12 -0500)]
ilowpass: test for wraparound behavior

8 years agoilowpass: handle s32 overflow better
Jeff Epler [Tue, 25 Mar 2014 16:40:07 +0000 (11:40 -0500)]
ilowpass: handle s32 overflow better

The conversion from double to int is undefined when the input operand
is out of range; in fact, gcc seems to return INT_MIN in this case.
By casting through __s64, and with -fwrapv specified, we get correct
signed integer wraparound until value exceeds 2^63.

8 years agoAsk for -fwrapv, we depend on it
Jeff Epler [Wed, 26 Mar 2014 16:21:50 +0000 (11:21 -0500)]
Ask for -fwrapv, we depend on it

We want to be sure that e.g., (INT_MIN) - (INT_MAX) is -1, so that encoder
overflows work better.  The C standard says that the result of this
is implentation-defined, and gcc provides -fwrapv to ask for the common
"wraparound arithmetic" behavior.  -fwrapv is in all compilers we are
likely to encounter; the internet says it was added in gcc 3.3.

-fno-strict-overflow prevents the compiler from assuming things such as
`i + 10 > i` always being true.  It's not clear whether -frwapv effectively
implies -fno-strict-overflow (though in this specific case it should,
since -fwrapv defines some values of i for which the comparison is false),
so go ahead and specify it if available; the internet says it was
added in gcc 4.2.

8 years agofix another bad paste
Francis Tisserant [Wed, 26 Mar 2014 13:27:35 +0000 (14:27 +0100)]
fix another bad paste

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agocleaning bad pastes
Francis Tisserant [Wed, 26 Mar 2014 12:25:34 +0000 (13:25 +0100)]
cleaning bad pastes

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agocleaning & fix broken links
Francis Tisserant [Wed, 26 Mar 2014 10:01:01 +0000 (11:01 +0100)]
cleaning & fix broken links

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agofix broken links
Francis Tisserant [Tue, 25 Mar 2014 20:54:58 +0000 (21:54 +0100)]
fix broken links

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agobad code delimiters previously entered by mystake
Francis Tisserant [Tue, 25 Mar 2014 20:18:08 +0000 (21:18 +0100)]
bad code delimiters previously entered by mystake

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agoCleaning for a better use of titles
Francis Tisserant [Tue, 25 Mar 2014 18:52:48 +0000 (19:52 +0100)]
Cleaning for a better use of titles

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agofix some asciidoc warnings
Francis Tisserant [Tue, 25 Mar 2014 15:28:56 +0000 (16:28 +0100)]
fix some asciidoc warnings

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agofix some broken links
Francis Tisserant [Tue, 25 Mar 2014 14:44:16 +0000 (15:44 +0100)]
fix some broken links

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agofix some broken links
Francis Tisserant [Tue, 25 Mar 2014 13:21:38 +0000 (14:21 +0100)]
fix some broken links

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agofix some broken links
Francis Tisserant [Tue, 25 Mar 2014 11:24:17 +0000 (12:24 +0100)]
fix some broken links

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agofix some broken links
Francis Tisserant [Mon, 24 Mar 2014 11:15:06 +0000 (12:15 +0100)]
fix some broken links

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agofix some broken links
Francis Tisserant [Mon, 24 Mar 2014 07:55:09 +0000 (08:55 +0100)]
fix some broken links

Signed-off-by: Francis Tisserant <tissf@free.fr>
8 years agoloadtool.ngc: provide options for g43 control
Dewey Garrett [Mon, 17 Mar 2014 21:24:39 +0000 (14:24 -0700)]
loadtool.ngc: provide options for g43 control

loadtool.ngc is a nc_files/ngcgui_lib/utilitysubs/ helper subroutine
for use with ngcgui_lib demonstration subroutines.

Add options for subroutine invocation:

   #<use_g43>
   #<h_for_g43>
   #<verbose>

to allow user to control behavior.

Also, allow toolno==0

Minor: use shorthand names for some call parameters since line length
limit is exceeded with additional items used.

8 years agoloadtool.ngc uncondtionally apply G43
Dewey Garrett [Sun, 16 Mar 2014 22:48:34 +0000 (15:48 -0700)]
loadtool.ngc uncondtionally apply G43

thanks to user pi.r for the report

8 years agoignore the new html docs stamp files
Sebastian Kuzminsky [Thu, 13 Mar 2014 14:28:10 +0000 (08:28 -0600)]
ignore the new html docs stamp files

8 years agobuild system: fix parallel build of docs
Sebastian Kuzminsky [Wed, 12 Mar 2014 03:03:21 +0000 (21:03 -0600)]
build system: fix parallel build of docs

Before this commit, each html file we built would copy its images from the
docs source directory to the docs html (target) directory.

Some of our html documents share images with other html documents, for
example many of the translations.

With 'make -j', and since all the html targets are built in parallel,
sometimes two make processes would try to copy the same image file at
the same time, resulting in failure.

This commit changes how images are handled.  All the html documents
are still handled in parallel, but the image copying is deferred until
after the html docs are all done, then all the images are copied by a
single thread.

8 years agodeb/rules: dont use cpio when cp will do
Sebastian Kuzminsky [Thu, 16 Jan 2014 20:32:39 +0000 (14:32 -0600)]
deb/rules: dont use cpio when cp will do

Debian Wheezy doesn't come with cpio by default, this commit lets us
not add a build-dependency on it.

(cherry picked from commit 7a1c034150ab1d42f014f85600709e69ac194678)

Conflicts:
debian/rules.in