Graphic lines and polygons are drawn to fit specific areas for which it
is helpful to have the endpoint controlled by the cursor even when
constraining the angle.
Fixes: lp:1805502
* https://bugs.launchpad.net/kicad/+bug/1805502
Append the appropriate file extension to the file name if the file name
does not have a file extension. This is a know issue with the GTK+ file
dialog.
Make the last 3D viewer screenshot last file used code more coherent.
Add the 3D viewer main frame trace string to the trace environment
variables doxygen group.
Doxygen comment and other minor code cleaning.
Fixes lp:1804980
https://bugs.launchpad.net/bugs/1804980
Under GTK3, the scrollbars fade in and out, causing a cascade of
re-paint events. These cause KiCad to be unresponsive while the repaint
events occur. By default this is 1s fade in 1s fade out but can be
configured by the user/wm.
Area mapping the repaint events to just scrollbar areas will be a good
solution but until/unless that happens the scrollbars in GTK3 are
problematic.
On GTK+3, the zoom events are smaller and faster, so
the trigger the zoom acceleration.
A simple fix is to use the non-accelerated Mac-style
zooming and an appropriate constant. This can be tweaked
further as needed.
Fixes: lp:1786515
* https://bugs.launchpad.net/kicad/+bug/1786515
GTK3 smooth scrolling is enabled by wxWidgets but not to the exclusion
of normal scroll up/down events. wxWidgets maps these both to the same
scroll handler and will fire them both if they are not handled before
being queued. Testing timestamps allows us to mark and ignore the
dupes.
This is set to GTK3 only for now as it isn't listed as a problem for
other platforms. But it shouldn't cause issues if it is enabled
elsewhere in the future.
The libeval grammar can now be rebuilt with "make libeval_grammar".
This is only possible when lemon is installed.
Also a couple of modifications to the lemon grammar to
account for external changes since the grammer was last generated.
Also move the grammar defines into grammar.h - this is how
Lemon produces them, and by manually copying into the target .cpp,
it make automated regeneration impossible.
This adds a check for contiguous board outlines to the DRC. It also
uses the calculated outline to ensure that traces are not crossing the
outlines.
Fixes: lp:1648055
* https://bugs.launchpad.net/kicad/+bug/1648055
The reduces a little bit of WX dependency, and makes
the timing code a bit more type-safe.
Also adds a more testable interface for the accelerated
zoom controller.
This is done to avoid a big chunk of conditionally-compiled code
in the middle of the event function.
Also separates the zoom logic from the WX_VIEW_CONTROLS object
and isolates it in a separate class behind a clearer interface.
Add some simple tests for sane steps on GTK+3-sized scroll
steps.
This syncs the "All" checkbox with the state of the report options such
that selecting "All" automatically selects the other boxes and
deselecting a box results in the "All" box also being unselected. This
allows a single click filtering of report messages.
Fixes: lp:1796992
* https://bugs.launchpad.net/kicad/+bug/1796992
When the error switching to OpenGL is triggered by the requested switch,
we want to fallback to the cairo GAL renderer instead of the
GAL_TYPE_NONE, which defaults to legacy.
Fixes: lp:1799017
* https://bugs.launchpad.net/kicad/+bug/1799017
This makes using Scrollbars more easy to use, especially for "small" page sizes like A or B.
Note also the working/drawing area size is bigger than the page size (3 times)
This parameter defines the working area (full page) size.
The fix is not perfect, because it does not take in account the page size.
However it is similar to the "old" initialization, before Eeschema GAL.
In Eeschema, a reasonable boundary size is used.
Due to recent changes in legacy canvas code, when starting a block command,
the block outline was not drawn in pcbnew and pl_editor during block definition.
Fixes: lp:1803668
https://bugs.launchpad.net/kicad/+bug/1803668
Factor out save current canvas view to image file code from symbol editor
code so it can be used anywhere.
Add ability to save to any image format supported by wxBitmapType. See
https://docs.wxwidgets.org/3.0/gdicmn_8h.html#a90a1eb6d85b5044a99b706fd979f27f5.
Currently only PNG output is implemented.
Please note that there is a minor bug that appears to be due to the
scroll bars which causes unfilled areas on the right and bottom edges
of the image. This always existed in the save symbol editor view
image but it was not as noticeable because by default the background
color is white. It is very noticeable in the footprint editor with
a black background.
The usual smattering of coding policy and comment fixes.
Fixes lp:1802127
https://bugs.launchpad.net/kicad/+bug/1802127
Due to the implicit floor of the cast from double to unsigned char,
there was a small rounding error in COLOUR4D's WX conversion function.
This fixes the failing tests.
Also make the cast to unsigned char explicit.
This avoids having to manually specify include/legacy_gal
in and legacy GAL targets, and harominizes with legacy_wx.
This also means .cpp files in common/legacy_gal do not
need to specify the legacy_gal subdirectory, so they
will continue to work as needed when legacy_wx is removed.
Display GALs had an incorrect world unit value set. Now the world unit
value says how many internal units are in an inch, in accord with the comments
in the GAL header. Bitmap drawing code relied on the information about DPI,
so scaling worked differently for display and print GALs.
CAIRO_PRINT_CTX provides a Cairo context created from wxPrintDC.
It allows one to prepare printouts using the Cairo library and
let wxWidgets handle the rest.
When reading (and writing) a file, we must switch the current locale to "C" for LC_NUMERIC.
Unfortunately, on Windows, a wxWidgets assert was shown when reading some items (bitmaps images).
This wxWidgets assert (related to decimal separator) is overzealous and is now hidden only when reading/writing files.
Fixes a crash when typing fast in the place footprint filter box.
Also adds a bunch more checking to GAL locking, including making
sure the same person unlocks as locked, and preventing piece-meal
calls (the RAII objects must be used).
The decision to include the TTL library make KiCad a derivative work
under the terms of the AGPL version 3 or later license.
Add both the GPL3 and AGPL3 license text files and add a readme file
about the KiCad licensing. Contributing to KiCad still falls under
the GPL3+ license.
Fixes lp:1797095
https://bugs.launchpad.net/kicad/+bug/1797095
It happens when the char is not found in the font used to display netnames/pad names.
When happens, a wxASSERT is raised.
Unfortunately it can crash the application if it is during a paint event or some critical time.
Linux does not handle the resize command with wide-character extended
table UTF-8. The solution did not work for W7-32bit. This is a
compromise, attempting first the preferred, previous solution and
falling back to the secondary solution.
(cherry picked from commit 6106210c87)
Replace all instances of wxLogDebug with wxLogTrace in the common and
kicad folders to prevent unwanted debugging output.
Add new trace flags for locale and screen object tracing.
The usual smattering of code policy fixes.
Invalid wchar characters would throw an exception. We need a defined
output from the += operator, even for unknown character.
Fixes: lp:1798144
* https://bugs.launchpad.net/kicad/+bug/1798144
(cherry picked from commit b37bc69476)
The connectivity files were unwieldy. This separates them logically
into data, algo and items where the items classes are those that hold,
surprise, surprise, the items, lists and clusters.
This prevents deadlocks when exceptions are thrown and the context
ends up not getting unlocked.
It also removes an earlier hack to try and minimize this which
didn't work anyway.
Dynamic casts should only be used when we explicitly check for the
resulting pointer to be NULL. Where we know the class is castable we
can use static_cast, save on overhead and ensure our resulting pointer
is non-null.
Block rotate and mirror now work like they do in Pcbnew (that is
they're treated as part of the block move/duplicate/whatever
rather than finishing it).
Fixes: lp:1780794
* https://bugs.launchpad.net/kicad/+bug/1780794
1) when duplicating don’t keep original hidden until end of drag
2) reset selectionArea when showing it so it doesn’t flash in its previous location
3) center a pasted block on the cursor
4) don’t draw the source selectionArea when pasting a block
5) implement selection-style highlighting for contents of blocks
6) add pasted items to view so they don’t disappear when the block is placed
Fixes: lp:1747197
* https://bugs.launchpad.net/kicad/+bug/1747197
Split antialiasing options out from display options. Move
antialiasing to common. Duplicate the rest of display options
for Eeschema.
Implement OnSelectGrid and hookup GAL canvas refresh to
SetPresetGrid.
Add Grid Settings... to View menu and move Show Grid from
preferences to View Menu to match Pcbnew.
5.0 doesn't set the GAL's line width. It gets away with this since
it's using the same GAL to measure as to set up the drawing, so
the width happens to be set correctly for other reasons. 5.1 uses
a separate GAL and so isn't so lucky.
Note: it's only the renderer that is changed from wxDC to GAL. Tools stay the same, consider this a temporary hack.
This commit splits the EDA_DRAW_FRAME, EDA_DRAW_PANEL and BLOCK_SELECTOR classes into two variants, each independently compiled into a static library.
- "legacy_wx" for PCBnew/Gerbview - wxDC-based legacy canvas. We have full GAL support there anyway so it makes no sense to introduce GAL rendering API to the legacy tools.
- "legacy_gal" for EEschema and the rest - GAL-based legacy canvas, using legacy tools but with a GAL renderer.
Such split ensures only a small part of the common library and eeschema is affected, without messing around with already GALified tools.
The commit also removes some header dependencies on class_drawpanel.h
It would appear that some platforms process the KILL_FOCUS event
after running TransferDataFromWindow(). This change makes sure
that the evaluation is done no matter the order.
Fixes: lp:1793911
* https://bugs.launchpad.net/kicad/+bug/1793911
LC_NUMERIC is the right option to use in LOCALE_IO because only the floating point separator must be modified
When using LC_ALL, the Env var expansion wxGetEnv() was not working when the env var contains non ascii7 codes after a call to LOCALE_IO.
When using setlocale( LC_NUMERIC, "C" ) in LOCALE_IO, wxGetEnv() works fine.
Fixes: lp:1795990
https://bugs.launchpad.net/kicad/+bug/1795990
The '/' and ':' are reserved and cannot be used in symbol or footprint
names. They will cause the LIB_ID parser and formatter to fail. While
it seems like they should be legal in symbol alias names, they will
trigger a symbol rescue the next time the schematic is loaded.
Use ID_SCH as in the Eagle schematic plugin rather than ID_ALIAS to
ensure symbol names do not need rescued the next time the schematic is
opened.
Remove ID_ALIAS since the rules for alias names are the same as the
rules for symbol names. Otherwise, allowing '/' and ':' in alias names
will force a symbol rescue on the next schematic load.
Fixes lp:1795600
https://bugs.launchpad.net/kicad/+bug/1795600
When querying the view, we do not want to including items that are not
on layers that are currently visible. This routine is currently only
used by the selection tool (pcbnew and gerbview) and the grid helper.
Both expected this logic but were indifferent to the additional items.
Fixes: lp:1796403
* https://bugs.launchpad.net/kicad/+bug/1796403
The guard isn't working on GTK (causing eval not to happen on
a kill focus), and I can't remember what issue I put it in to
solve. I've done a bunch of testing and it appears that we
don't need it, although I'm sure I put it in for something....
Fixes: lp:1793911
* https://bugs.launchpad.net/kicad/+bug/1793911
Patch modifies the way graphs are displayed: instead of drawing one
point for each X coordinate, there is a line segment joining min and max
values for continuous graphs and all unique points are displayed for
discrete graphs.
Fixes: lp:1674143
* https://bugs.launchpad.net/kicad/+bug/1674143
Graph rendering takes a lot of time, especially when there is a high number
of points to be drawn. The initial implementation drew all points/lines
even if the the coordinates were duplicates, in the new version duplicated
coordinates are skipped.
Fixes: lp:1674143
* https://bugs.launchpad.net/kicad/+bug/1674143
There are lot of places where constants are used in the KiCad UI
as "magic numbers". The most common one is "5", used in many
wxFormBuilder and manual UI constructions as the margin.
This commit provides a place for all UI to look up shared
constants and other functions, to help create a consistent UI using
functions that provide meaning and intent to these magic numbers.
This is in preparation for making this widget optionally read-only.
Major changes:
* Construct panel in code, not with wxFormBuilder. This make's it
easier to conditionally construct elements that won't be used
in a read-only mode (e.g. the buttons).
* Use a generic "button row panel" widget for the buttons, as the
sizing and layout logic is reusable in nearly all dialogs, and
it's simplifies layout in the higher-level dialog widget. This
widget is one example of many possible "reuable widgets".
Put the cursor and pan control TOOL_ACTIONs in ACTION,
and refer to them in that way.
The handlers are currently identical in Pcbnew and Cvpcb,
and Gerbview (and all canvases) can benefit from them as well
if they are in libcommon. Also saves duplicated code.
Fixes: lp:1795855
* https://bugs.launchpad.net/kicad/+bug/1795855
Adds Cut/Copy/Paste and Revert for footprints; introduces a new
shared Revert Changes? dialog; hooks up Add Library for footprints,
standardizes the Save As terminology.
This puts generic logic for KiCad environment vars in
one place.
Also updates the DIALOG_CONFIGURE_PATHS help to document
the new KICAD_USER_TEMPLATE_DIR and KICAD_TEMPLATE_DIR.
This version makes use of lots of things learned going down the
other rat hole. Avoiding the wxComboFocusHandler is key, as well
as specifying using the AltPopupWindow.
The key handler is still tricky with respect to those platforms
that use native controls, but the starting-key strategy is similar
to the one used with wxGrid text editors.
Happens only in Opengl mode.
So enforce controls in 3 group function: DrawGroup, ChangeGroupColor, ChangeGroupDepth.
Fixes: lp:1795044
https://bugs.launchpad.net/kicad/+bug/1795044
This is caused by:
* Not checking the hotkey data is not null when performing a
hotkey action
* Allowing hotkey actions on non-hotkey rows.
This fixes both of these, and adds an assert to warn if someone
does manage to fire a hotkey action on a non-hotkey row (but it
won't crash).
Fixes: lp:1794756
* https://bugs.launchpad.net/kicad/+bug/1794756
It was possible to get conflicting hotkeys when undoing and resetting hotkeys
to defaults.
This uses the same logic as when setting hotkeys to avoid conflcits in these
other two cases.
Fixes: lp:1794730
* https://bugs.launchpad.net/kicad/+bug/1794730
This separates the "ground truth" store of hotkeys from what is shown
in the dialog. This will allow us to filter the displayed hotkeys
while keeping the same underlying data structures.
Now, the UI data items interact with an intermediate set of data, which
represents the "original" hotkey data, and the "changed" data. The
ultimate aim here is to allow UI elements to come and go, but the
hotkeys that are "in-edit" are preserved.
This also allows us to abstract some bookkeeping complexity
out of the WIDGET_HOTKEY_LIST class into a separate non-GUI
class.
GTK uses a level of cooperative multitasking when dealing with
secondary event loops. This gives the wm enough leeway to process the
redraw before returning to the secondary loop
This commit finishes the removal of OpenMP from the KiCad codebase.
Removed in this commit are the OpenMP calls in 3d-viewer and
qa/polygon_triangulation as well as all references in CMakeLists.txt
std::thread is used instead for multithreaded computation
wxString constructor will determine string length using strlen() when it
is not explicitly specified, but there is no termination byte added when
reading file contents to 'buffer' variable.
Commit 73c229714 was a bit of a sledgehammer for the associated problem
of degenerate points. This commit replaces that one by only performing
additional simplification of the zone polygons on those polygons that
fail our initial triangulation attempt.
jean-pierre charras
Seth Hillbrand
Maciej Suminski
Jeff Young
Wayne Stambaugh
John Beard
qu1ck
Thomas Pointhuber
Simon Richter
Tomasz Włostowski
Ronnie Gaensli
Hal Gentz
Fabio Rossi