Peter Clifton [Sun, 10 Jan 2016 23:58:59 +0000 (10 23:58 +0000)]
Attempt to fix GTK hid zooming/panning/flipping
Peter Clifton [Sun, 10 Jan 2016 18:47:17 +0000 (10 18:47 +0000)]
hid/gtk: Make SwapSides action operate with passed coordinates.
Previosuly we used the last known X,Y coordinates from a mouse event seen
by the event handlers.
Peter Clifton [Sun, 10 Jan 2016 17:21:22 +0000 (10 17:21 +0000)]
Revert "gtkhid-main.c: remove seemingly useless double conversion."
This reverts commit
f4b511daed4bcc07df7445a345f2d19105b2be70.
They don't do anything. The locals event_x/y are just copied
out then straight back into gport->pcb_x/y, none of which are
use at all in the interim.
Peter Clifton [Thu, 7 Jan 2016 03:18:57 +0000 (7 03:18 +0000)]
Delete the debian/ directory
Including disro-packaging upstream is discouraged by debian policy, as it
makes packaging harder.
Deleting at the request of our upstream maintainers.
Peter Clifton [Thu, 7 Jan 2016 03:29:30 +0000 (7 03:29 +0000)]
Revert "hid/ps/ps.c: avoid division by zero."
This reverts commit
6e24a51370c17eb99e8d823c670b77cbb9eba515.
Given a linewidth of zero, 'inf' was written as line width to the
PostScript file, which made the file invalid. Now we at least get
valid PostScript.
Not yet solved: why is there an attempt to draw a circle of zero
size?
This is related to bug LP1497963.
Peter Clifton [Thu, 7 Jan 2016 03:18:32 +0000 (7 03:18 +0000)]
Revert "hid/ps/ps.c: deal with zero sized circles even better."
This reverts commit
2eb76110abcc1bbed375b169808ef72c472cb604.
Instead of hardcoding a size, actually calculate the best fit.
This also makes the PostScript output of the circles test case
matching what's shown on the screen.
Peter Clifton [Sun, 10 Jan 2016 17:19:57 +0000 (10 17:19 +0000)]
Revert "gui-top-window.c: no need to check arguments if not used."
This reverts commit
92d4bab703cd0894421bc8c8fe1aa7a67dc1c679.
Peter Clifton [Sun, 10 Jan 2016 17:16:17 +0000 (10 17:16 +0000)]
Revert "ActionDisplay: show help on failure."
This reverts commit
0978828fcf27be98d4edee22e57a13756a684b62.
We need to be strict about argument checking.
Peter Clifton [Sun, 17 Jan 2016 13:09:57 +0000 (17 13:09 +0000)]
drc: Fix polygon-polygon clearance checks
IsPolygonInPolygon was applying 2x bloat to the line it uses to test with
against the second polygon, and IsLineInPolygon was applying a further
bloat offset - making the clearance used incorrect.
This caused false DRC violation reports of polygons being too close to
one-another, despite there being adequate clearance.
Note for future - a cleaner fix regarding "Bloat", and the DRC, is that
we should probably pass it more explicitly as a parameter where it is
to be used, in order to avoid surprises as to where it gets applied.
bert [Tue, 12 Jan 2016 19:21:20 +0000 (12 20:21 +0100)]
Added keywords to the appdata file.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
Anton Dubniak [Sun, 3 Jan 2016 23:00:37 +0000 (4 01:00 +0200)]
Add measurement units to route-styles string in preferences file.
This avoids unit conversion to cmil on save of preferences, and
avoids the rounding errors associated with that.
Anton Dubniak [Wed, 30 Dec 2015 15:53:00 +0000 (30 17:53 +0200)]
Fix Auto-enforce-DRC check to allow the exact minimum spacing
Removes an improper "+ 1" (COORD - e.g. nm) condition applied in the auto-drc
code, and similar from the XOR rendering code.
Reviewed-by: Peter Clifton <Peter.Clifton@clifton-electronics.co.uk>
Closes-bug: lp-
1013358
bert [Sun, 27 Dec 2015 19:33:19 +0000 (27 20:33 +0100)]
src/hid.h: Converted plain comments into doxygen comments.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Sun, 27 Dec 2015 19:30:01 +0000 (27 20:30 +0100)]
src/heap.[ch]: Converted plain comments into doxygen comments.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Sat, 24 Oct 2015 10:41:08 +0000 (24 12:41 +0200)]
src/global.h: correction of a typo in the Doxygen header.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Sat, 21 Nov 2015 12:21:34 +0000 (21 13:21 +0100)]
Added QFN24_5 and TQFN24_5.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Sat, 21 Nov 2015 11:54:15 +0000 (21 12:54 +0100)]
Added LQFP80-10.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
Milan Prochac [Wed, 14 Oct 2015 16:59:39 +0000 (14 18:59 +0200)]
Added escaping of Attribute value
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Sun, 27 Dec 2015 16:35:52 +0000 (27 17:35 +0100)]
Corrections in documentation.
Reported-by: dima
Signed-off-by: bert <bert.timmerman@xs4all.nl>
Dan McMahill [Thu, 17 Dec 2015 04:29:46 +0000 (16 23:29 -0500)]
Add some comments and improve default refdes silk screen
Dan McMahill [Thu, 17 Dec 2015 04:07:30 +0000 (16 23:07 -0500)]
Add generator for Aries Electronics, Inc. footprints
Added footprint generator and definitions for the
High-Temp Universal ZIF DIP Burn-in and Test sockets
from Aries Electronics.
See for example www.arieselec.com for information on their
ZIF sockets.
Martin [Tue, 1 Dec 2015 12:09:26 +0000 (1 13:09 +0100)]
Fix invalid jump to next VNODE in Collect1() function
When constructing polygon Collect1() function jumps to cur->next vertex
when calling Gather() which results in inclusion of invalid vertex into
resulting polygon.
Fixed by moving the test and the cur->next expression into parameters of
first call to Collect1() only where it is valid (in contrast to second
call after jump() test).
Signed-off-by: bert <bert.timmerman@xs4all.nl>
Markus Hitter [Wed, 30 Sep 2015 00:15:51 +0000 (30 02:15 +0200)]
unit_to_coord(): explicitely cast the result.
Previously 32-bit machines would return
999999999 as the result
of "1.0 m", making the unit test fail.
This introduces the DOUBLE_TO_COORD() macro which should be used
everywhere where doubles are used for calculating coordinates.
Wisdom about rounding found at:
https://www.cs.tut.fi/~jkorpela/round.html
bert [Thu, 29 Oct 2015 05:53:13 +0000 (29 06:53 +0100)]
Added RenumberBlock() and RenumberBuffer().
Converted from plug-in sources from DJ Delorie.
Source file adapted for nm resolution and added user documentation
and a file header for Doxygen.
Added to the build system, so it gets always compiled in.
Found on http://www.delorie.com/pcb/renumberblock.c
Reported-by: DJ Delorie <dj@delorie.com>
Signed-off-by: bert <bert.timmerman@xs4all.nl>
Kai-Martin Knaak [Sat, 24 Oct 2015 20:10:18 +0000 (24 22:10 +0200)]
Merge branch 'master' of ssh://git.geda-project.org:65/pcb
Kai-Martin Knaak [Sat, 24 Oct 2015 20:08:31 +0000 (24 22:08 +0200)]
let git ignore *.o and *.a files in the gts folder
Kai-Martin Knaak [Sat, 24 Oct 2015 19:48:30 +0000 (24 21:48 +0200)]
+ translation file de.po
Add German to the list of translation goals
Kai-Martin Knaak [Sat, 24 Oct 2015 19:47:20 +0000 (24 21:47 +0200)]
Update of the translation files
Kai-Martin Knaak [Sat, 24 Oct 2015 19:45:55 +0000 (24 21:45 +0200)]
Add translation infrastructure to the empty layout error message
Kai-Martin Knaak [Sat, 24 Oct 2015 17:48:50 +0000 (24 19:48 +0200)]
png HID should exit with warning when layout contains no data
Instead of segfault exit decently with an error message on stdout.
Original patch by Amit Dhingra
fixes bug #
1048256
https://bugs.launchpad.net/geda-project/+bug/
1048256
Kai-Martin Knaak [Sat, 24 Oct 2015 16:18:07 +0000 (24 18:18 +0200)]
sync my local master with upstream master
Britton Leo Kerin [Tue, 20 Oct 2015 00:41:26 +0000 (19 16:41 -0800)]
Use hypot() instead of sqrt(pow, pow)
Britton Leo Kerin [Tue, 20 Oct 2015 00:40:17 +0000 (19 16:40 -0800)]
Use some computed vars to make them used
Britton Leo Kerin [Tue, 20 Oct 2015 00:33:44 +0000 (19 16:33 -0800)]
Use wide types for comfortable Coord output
Britton Leo Kerin [Tue, 20 Oct 2015 00:03:21 +0000 (19 16:03 -0800)]
Added missing #include
bert [Thu, 15 Oct 2015 19:47:18 +0000 (15 21:47 +0200)]
Changed name of teardrop example pcb file as to avoid make interference.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Mon, 12 Oct 2015 16:58:26 +0000 (12 18:58 +0200)]
Added an example figure and more edits to the user doumentation.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Sun, 11 Oct 2015 19:41:59 +0000 (11 21:41 +0200)]
Added user documentation for teardrops.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Sun, 11 Oct 2015 19:07:48 +0000 (11 21:07 +0200)]
Renamed to teardrops.c.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Sun, 11 Oct 2015 18:56:10 +0000 (11 20:56 +0200)]
Added teardrop.c to POTFILES.in.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
Markus Hitter [Sun, 30 Aug 2015 13:22:11 +0000 (30 15:22 +0200)]
Add teardrops.c from DJ Delorie.
Source file unchanged, just prettyfied file header for Doxygen.
Added to the build system, so it gets always compiled in.
Found on http://www.delorie.com/pcb/teardrops/
bert [Mon, 12 Oct 2015 21:40:24 +0000 (12 23:40 +0200)]
Make compiler warning go away.
hid/common/hidgl.c:455: warning: format ‘%lu’ expects type
‘long unsigned int’, but argument 2 has type ‘unsigned int’
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Mon, 12 Oct 2015 19:12:35 +0000 (12 21:12 +0200)]
Remove the unwanted N_("..."); macro cruft from the user documentation.
This commit solves LP bug #
1505262.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Mon, 12 Oct 2015 17:16:51 +0000 (12 19:16 +0200)]
Added user documentatation for SmartDisperse.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Sat, 10 Oct 2015 18:30:37 +0000 (10 20:30 +0200)]
src/global.h: Converted plain comments into doxygen comments.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Thu, 24 Sep 2015 19:50:13 +0000 (24 21:50 +0200)]
src/gettext.c: Converted plain comments into doxygen comments.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Wed, 23 Sep 2015 05:45:06 +0000 (23 07:45 +0200)]
src/free_atexit.[ch]: Converted plain comments into doxygen comments.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
Markus Hitter [Sat, 3 Oct 2015 10:58:43 +0000 (3 12:58 +0200)]
gtkhid-main.c: swap sprintf() for snprintf().
It's good practice to use the 'n' variant even if it's reasonably
sure the string will never overflow.
This should satisfy bug
1069317.
Kai-Martin Knaak [Wed, 30 Sep 2015 02:45:31 +0000 (30 04:45 +0200)]
Gtk: right-aligned static strings in layer groups dialogue.
The static strings are a little more easy to the eye when right
aligned. A colon was added for extra legibility.
Kai-Martin Knaak [Wed, 30 Sep 2015 02:24:56 +0000 (30 04:24 +0200)]
Gtk: use radio buttons in layer groups dialogue.
Kai-Martin Knaak [Wed, 30 Sep 2015 01:00:42 +0000 (30 03:00 +0200)]
Gtk: more explicit comment in the layer group code.
Kai-Martin Knaak [Tue, 29 Sep 2015 09:55:20 +0000 (29 11:55 +0200)]
Gtk: change the order of top and bottom in preference dialogue.
Kai-Martin Knaak [Sun, 27 Sep 2015 22:44:33 +0000 (28 00:44 +0200)]
gtk/gui-config.c: swap the order of silk layer groups in GUI.
The silk layers double as layer groups for top and bottom
copper. Current pcb/gtk GUI shows them bottom above top in
File -> Preferences -> Layers -> Groups. This is inconsistent
with the general order of layers.
The patch swaps the order the silk layers are displayed in the
Groups dialogue.
Markus Hitter [Tue, 29 Sep 2015 12:45:16 +0000 (29 14:45 +0200)]
hid/png/png.c: more checks wether a layer actually exists.
Especially when considering the move to layer types there is no
guarantee a mask and/or silk layer actually exists.
Adrian Pardini [Mon, 13 Jun 2011 00:51:55 +0000 (12 21:51 -0300)]
hid/png/png.c: check wether a bottom layer actually exists.
This should fix LP796389: fixes segfault with --photo-mode
--photo-flip if the layer "solder" is
not in the "Solder side".
Markus Hitter [Thu, 1 Oct 2015 13:50:23 +0000 (1 15:50 +0200)]
global.h: eliminate GCC_VERSION macro.
It was used in this place, only, so no need to clutter the
namespace.
Markus Hitter [Thu, 1 Oct 2015 13:47:07 +0000 (1 15:47 +0200)]
global.h: remove forgotten comment.
RCS IDs and the macro described by the comment were removed with
commit
02a720280796e439015f in December 2011.
Markus Hitter [Sat, 26 Sep 2015 21:46:25 +0000 (26 23:46 +0200)]
hid/ps/ps.c: deal with zero sized circles even better.
Instead of hardcoding a size, actually calculate the best fit.
This also makes the PostScript output of the circles test case
matching what's shown on the screen.
Markus Hitter [Sat, 26 Sep 2015 21:42:56 +0000 (26 23:42 +0200)]
Tests:more challenging circles.pcb.
This time several circles of varying size, along with ruler lines
for measuring them on non-screen output, e.g. a PostScript file.
Note that the zero sized circle with 50 mil line width is drawn
differently in PostScript than on screen.
Markus Hitter [Fri, 25 Sep 2015 22:21:21 +0000 (26 00:21 +0200)]
Introduce regression tests for PostScript exporter.
Not the most fancy testing, but at least it recognizes invalid
PostScript. It triggers with the bug reported in LP1497963.
Markus Hitter [Fri, 25 Sep 2015 20:18:23 +0000 (25 22:18 +0200)]
hid/ps/ps.c: avoid division by zero.
Given a linewidth of zero, 'inf' was written as line width to the
PostScript file, which made the file invalid. Now we at least get
valid PostScript.
Not yet solved: why is there an attempt to draw a circle of zero
size?
This is related to bug LP1497963.
Kai-Martin Knaak [Wed, 30 Sep 2015 00:48:34 +0000 (30 02:48 +0200)]
swap the order of silk layer groups in GUI
The silk layers double as layer groups for top and bottom copper. Current
pcb/gtk GUI shows them bottom above top in File->Preferences->Layers->Groups.
This is inconsistent with the general order of layers.
The patch swaps the order the silk layers are displayed in the groups dialog.
Markus Hitter [Tue, 29 Sep 2015 23:18:59 +0000 (30 01:18 +0200)]
Follow autoconf's recommendation for including glu.h.
Markus Hitter [Tue, 29 Sep 2015 23:18:59 +0000 (30 01:18 +0200)]
Follow autoconf's recommendation for including gl.h.
Markus Hitter [Fri, 25 Sep 2015 12:52:15 +0000 (25 14:52 +0200)]
Debian: 'gerbv' isn't a package build dependency.
It's needed for running regression tests, only, and these tests
aren't run when building packages.
Tested with pbuilder.
Markus Hitter [Fri, 25 Sep 2015 11:53:48 +0000 (25 13:53 +0200)]
README.ppa: also set the actual uploader in debian/control.
... or at least provide instructions on how to do so.
Markus Hitter [Fri, 25 Sep 2015 11:04:43 +0000 (25 13:04 +0200)]
debian/control: remove obsolete field DM-Upload-Allowed.
This also fixes the corresponding warning while building the
package.
Markus Hitter [Thu, 24 Sep 2015 21:34:14 +0000 (24 23:34 +0200)]
Debian: add dependencies for building straight from Git.
When building from the bare repository there are quite a few
additional dependencies, almost all of them for documentation and
tutorials.
Markus Hitter [Thu, 24 Sep 2015 21:23:37 +0000 (24 23:23 +0200)]
README.ppa: add instructions on how to build in a chroot.
After many failures with many different tools, 'pbuilder' was
finally found to be the one to work as described.
Markus Hitter [Wed, 23 Sep 2015 20:10:55 +0000 (23 22:10 +0200)]
Debian: remove Quilt entirely.
All patches were removed earlier already, because there should
never be a need for packaging patches when packaging straight from
the Git repository. Accordingly there's no need to have a patch
system at all.
Markus Hitter [Wed, 23 Sep 2015 16:33:53 +0000 (23 18:33 +0200)]
Debian: add instructions on how to build for the PPA.
Markus Hitter [Wed, 23 Sep 2015 18:40:30 +0000 (23 20:40 +0200)]
Debian: remove all patches.
When building straight from the Git repo there should never be a
need for packaging patches.
Markus Hitter [Tue, 22 Sep 2015 09:43:53 +0000 (22 11:43 +0200)]
Debian: additional rules to build straight from Git.
Previous rules were set to build from a distribution tarball, so
'configure' was expected to exist already.
These fixes are just enough to get the packages building, no
or sanity checks happened.
Markus Hitter [Tue, 22 Sep 2015 09:13:39 +0000 (22 11:13 +0200)]
Debian: adjust build dependencies to recent development.
- Relax dependency on 'tk8.5' to just 'tk'.
- Add 'libgts-dev' and 'texlive-binaries'.
So far only changes to get it building. All these dependencies
likely need a review, like trying to build without them in a
chroot, one by one, and keeping only the neccessary ones.
Markus Hitter [Mon, 21 Sep 2015 19:20:36 +0000 (21 21:20 +0200)]
Debian: import debian/ as-is from Eugene Mikhantievs PPA.
Fetched from
https://launchpad.net/~mehanik/+archive/ubuntu/geda-unstable
Britton Leo Kerin [Sun, 27 Sep 2015 22:14:34 +0000 (27 14:14 -0800)]
Move gdk_window_process_all_updates to safest spot
Because it might be a tiny bit inefficient for zoom where it was. This
way nothing hits this except the case where we're about to do a pointer
warp
Britton Leo Kerin [Sun, 27 Sep 2015 21:38:34 +0000 (27 13:38 -0800)]
Do not warp pointer too easily...
* Pointer warping bug people so now you have to double-click to get it
* Tooltip for the drc window
Britton Leo Kerin [Fri, 18 Sep 2015 17:46:56 +0000 (18 09:46 -0800)]
Fix CenterWindow, DRC activation pointer warping.
CenterWindow didn't do what it's name said, it just warped the pointer.
Now it does both for the gtk hid, so DRC checking works consistently
when the violation is off-screen or not. It's still broken in lesstif
HID but no worse than it was, and there's a warning about the
pan-and-warp action not being implemented there.
Britton Leo Kerin [Fri, 18 Sep 2015 17:46:56 +0000 (18 09:46 -0800)]
gtkhid-main.c: remove seemingly useless double conversion.
They don't do anything. The locals event_x/y are just copied
out then straight back into gport->pcb_x/y, none of which are
use at all in the interim.
Kai-Martin Knaak [Sun, 27 Sep 2015 17:26:18 +0000 (27 19:26 +0200)]
hid/gtk/gui.h: fix definition of LAYER_BUTTON_SILK
LAYER_BUTTON_SILK should give the number of the layer after all copper
layers. Since the copper layers are enumerated starting with zero, this
is equal to the number of copper layers MAX_LAYER.
The previous definition erreously added LAYER_BUTTON_SILK to this number.
This only worked, because src/const.h sets this constant to zero.
Markus Hitter [Sun, 27 Sep 2015 14:52:24 +0000 (27 16:52 +0200)]
Fix the recent addition of smartdisperse.c.
Before committing something, please run a 'make distcheck'. This
is considered to be the most safe way we have to make sure there
is no regression.
If this isn't possible, please build with and without NLS and run
a 'make check' each time. Like:
./autogen.sh && ./configure --disable-nls --disable-doc && make
make check
./autogen.sh && ./configure && make
make check
This is a bit more manual work but gives similar confidence.
All these tests pass now again.
Keith Packard [Tue, 8 Jul 2014 22:05:00 +0000 (8 15:05 -0700)]
Ignore completely clipped polygons in PS export
This avoids dereferencing a null pointer and crashing.
Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Sun, 27 Sep 2015 09:16:35 +0000 (27 11:16 +0200)]
Added smartdisperse.c from Ben Jackson.
Removed bitrot, and prettyfied the file header for Doxygen.
Added to the build system, so it gets always compiled in.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
Markus Hitter [Sat, 26 Sep 2015 13:55:37 +0000 (26 15:55 +0200)]
configure.ac: report tests enablement in the summary.
Also moved CPP_FLAGS & Co. upwards to have the more important
part last.
Markus Hitter [Sat, 26 Sep 2015 13:35:29 +0000 (26 15:35 +0200)]
Consider build tests not run to be a failure.
Previously a missing test environment would go unnoticed, the
distinction between a successful test run and a test not run at
all was (almost) not noticeable.
bert [Tue, 22 Sep 2015 16:36:26 +0000 (22 18:36 +0200)]
src/fontmode.c: Converted plain comments into doxygen comments.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Tue, 22 Sep 2015 16:35:51 +0000 (22 18:35 +0200)]
src/flags.c: Converted plain comments into doxygen comments.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Tue, 22 Sep 2015 16:17:09 +0000 (22 18:17 +0200)]
src/find.[ch]: Converted plain comments into doxygen comments.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Fri, 18 Sep 2015 20:20:07 +0000 (18 22:20 +0200)]
src/file.[ch]: Converted plain comments into doxygen comments.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
Markus Hitter [Tue, 15 Sep 2015 10:02:44 +0000 (15 12:02 +0200)]
run_tests.sh: make sure the test candidate is writable.
This makes a distinction when running 'make distcheck', because
the file is written read-only to the distribution.
Markus Hitter [Mon, 14 Sep 2015 21:39:05 +0000 (14 23:39 +0200)]
Remove these almost 25 Makefiles for just copying tests.
Before there was a Makefile/.am/.in in each test directory,
just to define the EXTRA_DIST variable for this directory. A
rather complex approach.
The new strategy is to list all these files in one place, in
tests/Makefile.am. This almost doubles the speed of compilation
after small changes and reduces clutter sent to the command
line a lot, so a great improvement for developers.
Creation of new tests is also simplified, creating a Makefile.am
and editing configure.ac is no longer neccessary.
Previous versions of this commit used wildcards to include new
tests automatically. This strategy was discarded in favour of a
hand-edited list, because some developers aren't exactly excited
by using wildcards.
As we're on the topic, here are a few informative readings
regarding recursive makes and regarding wildcards:
"Recursive Make Considered Harmful"
Peter Miller, 2008
http://aegis.sourceforge.net/auug97.pdf
GNU automake about recursion:
http://www.gnu.org/software/automake/manual/html_node/Directories.html
GNU make about wildcards:
http://www.gnu.org/software/make/manual/html_node/Wildcard-Function.html
GNU automake about wildcards:
http://www.gnu.org/software/automake/manual/html_node/Wildcards.html
Markus Hitter [Wed, 16 Sep 2015 13:16:32 +0000 (16 15:16 +0200)]
w32/Makefile.am: refine the move to wildcards.
As per discussion in #geda IRC, liberal use of wildcards isn't
exactly a good idea. Some even prefer to use no wildcards at all.
Others want to make situations, where files are often added or
removed, less forget-prone.
As a compromise, move towards more strict, unambiguous wildcards.
Markus Hitter [Tue, 15 Sep 2015 14:08:37 +0000 (15 16:08 +0200)]
Makefile.am: refine the move from INCLUDES to AM_CPPFLAGS.
The key is apparently this sentence in the autoconf docs:
"AM_CPPFLAGS is ignored in preference to a per-executable
(or per-library) _CPPFLAGS variable if it is defined."
Accordingly we have to add -I$(top_srcdir) to every
target-specific *_CPPFLAGS.
This fixes a 'make distcheck' more, it builds now, but fails on
the checks.
Markus Hitter [Tue, 15 Sep 2015 12:02:41 +0000 (15 14:02 +0200)]
w32/Makefile.am: use wildcards to include minipack stuff.
This should end the need for the often required commits to let the
Makefile catch up with actual development.
This partially fixes 'make distcheck', it runs a lot further.
bert [Mon, 14 Sep 2015 19:20:20 +0000 (14 21:20 +0200)]
src/error.[ch]: Converted plain comments into doxygen comments.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Mon, 14 Sep 2015 18:36:13 +0000 (14 20:36 +0200)]
src/drill.[ch]: Converted plain comments into doxygen comments.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
bert [Mon, 14 Sep 2015 18:26:12 +0000 (14 20:26 +0200)]
src/draw.[ch]: Converted plain comments into doxygen comments.
Signed-off-by: bert <bert.timmerman@xs4all.nl>
Markus Hitter [Sun, 13 Sep 2015 20:28:49 +0000 (13 22:28 +0200)]
configure.ac: remove test for maintainer mode.
The test is now pointless since usage of maintainer mode was
removed.
Markus Hitter [Sun, 13 Sep 2015 19:42:10 +0000 (13 21:42 +0200)]
A README shouldn't be run as shell script.
Not that the file had the executable bit set, but having
a #!/bin/sh on top fools some syntax coloring editors.
Markus Hitter [Sun, 13 Sep 2015 21:13:46 +0000 (13 23:13 +0200)]
Fix a configuration warning.
This was the warning message:
src/Makefile.am:224: warning: 'INCLUDES' is the old name for
'AM_CPPFLAGS' (or '*_CPPFLAGS')