Force the symbol library links whenever the symbol library table has been
modified. This will use the cache as a fallback when a library has been
removed that contains links in the schematic rather than display.
Fix the SCH_COMPONENT symbol resolver when falling back to the cache. The
resolver was using the LIB_ID ':' notation which was failing. Replacing
':' with '_' fixed this issue. This was also an issue when generating the
symbol message panel information.
Convert wxLogDebug to wxLogTrace in symbol resolver code path. Add new
trace type KICAD_SYM_RESOLVE.
Fixes lp:1821606
https://bugs.launchpad.net/kicad/+bug/1821606
There is a rendering bug in GTK3, which appears to be
an upstream GTK issue.
This can be worked around by, when filtering, ensuring the
*parent* item of the selected item is visible. This will
not cause the selected item to not be visible, as the selected
item will be the first shown child. So it will be visible, as long
as the list box is greater than a single row high, which it will be
in all practical scenarios.
This is done on all platforms, as it has a beneficial side-effect:
the parent library of the selection is naturally shown to the
user, so they don't need to scroll up to see what library their
current filter selection was in.
Fixes: lp:1804400
* https://bugs.launchpad.net/kicad/+bug/1804400
This allows a const BITMAP_BASE to expose const image
data. This is currently done with const_casts, which is
"OK", only as long as the source image is not declared
const, in which case it's undefined behaviour.
Also immediately dereference the pointer to a reference
if it's not checked to make the non-null requirements
explicit at the point of access.
Both versions of EDA_DRAW_FRAME::isBusy have a useless const_cast.
GetScreen() const returns a non-const SCH_SCREEN already, it doesn't
need to be cast, and it's stored in const SCH_SCREEN anyway.
This is useful when working on code paths that rely on KiWay
communications, since you can run and debug the kicad application
without needing to run an install task first.
In OpenSSL 1.1.0 and newer, the locking callback macros do nothing.
This causes unused warnings for the static function lock_callback.
Ignore this function on newer OpenSSL builds, but keep
the infrastructure for now.
When OpenSSL's min version is 1.1 or greater, all the locking
infrastructure can be removed.
Prevent a divide-by-zero bug in SHAPE_ARC::ConvertToPolyline.
When the radius is zero, just use the initial angle (it makes
no different anyway, the result is the centre point, which is
the start point.
The wire image undo/redo option required special handling and was only
used in a single location. Replacing with single change handling
reduces the code complexity
Move repeated GAL config reading routines into GAL_DISPLAY_OPTIONS.
THe app-level config is in here already, do the same for the
common config.
This means that the configs are loaded consistently, which fixes
the symbol-chooser preview window, which previously didn't use the
same config routine as other GAL canvases.
Future work could move these functions to free functions that
act on the public interface of GAL_DISPLAY_OPTIONS to avoid
GAL_DISPLAY_OPTIONS having to know about wxConfig and wxWindow.
Fixes: lp:1824524
* https://bugs.launchpad.net/kicad/+bug/1824524
Hash collisions reduce the speed of unordered maps by increasing the
comparison time of underlying values. This implements the inverse
golden ration offset into the x value to ensure that consecutive points
are far from each other in the map.
This allows setting the config dir at build time, providing packagers a
method of controlling where the configuration files are placed.
Fixes: lp:1780601
* https://bugs.launchpad.net/kicad/+bug/1780601
On some platforms like MSVC, NaN prints as "-nan(ind)". This
is a bit needlessly ugly, so print "NaN" on all platforms
consistently.
This fixes a test failure on MSVC.
This allows us to make the various netlist and pcb update routines
more atomic and less reliant on carefully sequenced asynchronous
messages.
This is also a prelude to adding support for footprint testing
without a netlist.
Both Cairo and OpenGL had issues (different, though) with circles that are
thicker in line width than they have radius. This corrects the OpenGL
implementation (radius is calculated to the outer edge) as well as Cairo
(line width needs to be clamped to twice the radius)
Fixes: lp:1822765
* https://bugs.launchpad.net/kicad/+bug/1822765
Statically constructing wxCursors as used for the probe
and tune cursors crashes the qa_eagle_plugin test
instantly.
Fix this by introducing a new class CURSOR_STORE,
which allows to abstract the platform wierdness of wxCursors
and allow the simulator to laziliy init its own cursors
at runtime.
This code isn't properly tested, as these cursors are never
actually used. However, it does allow the eeschema unit test
to run.
Fixes: lp:1781191
* https://bugs.launchpad.net/kicad/+bug/1781191
Bus upgrades: core new connectivity code
Bus upgrades: eeschema integration and modifications
Bus upgrades: eeschema dialogs
Bus upgrades: netlist export
Bus upgrades: file format changes
Wayne Stambaugh
Jon Evans
John Beard
Tomasz Włostowski
Jeff Young
jean-pierre charras
Seth Hillbrand
Michael Kavanagh