Between schematic and pcb editors, we want the following actions:
- Single click on ERC/DRC item should show the item in the ERC/DRC
window. But if the window is not visible, it should only update the
status bar
- Double click on ERC/DRC item should show and raid the ERC/DRC window
as well as select the line item
Fixes https://gitlab.com/kicad/code/kicad/-/issues/17383
ADDED: Horizontal/vertical zoom for simulator plots, via mouse wheel,
toolbar buttons, menu commands, and hotkeys.
ADDED: Simulator preferences panel, populated with mouse wheel
and trackpad settings that control pan and zoom of simulator plots.
ADDED: Zoom In/Out Horizontally/Vertically commands that can be bound
to hotkeys.
CHANGED: Simulator plot scroll wheel gestures are no longer hard-coded
and can now be configured via the new Simulator preferences panel.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/16597
Other unreported bugs that were fixed:
- Fixed wierd, jumpy simulator plot view limiting behavior.
- Fixed Zoom In Center and Zoom Out Center commands not preserving
the simulator plot center point.
- Fixed simulator plot nudging when exported as PNGs.
- Fixed rectangular selection zoom being able to exceed simulator plot
view limits.
Notes:
- Provided new SIM_PREFERENCES struct to be used for future
simulator preferences set via the simulator preferences dialog.
- Bundled pre-existing EESCHEMA_SETTINGS::SIMULATOR settings into
EESCHEMA_SETTINGS::SIMULATOR::VIEW.
- Replaced mpWindow::EnableMouseWheelPan with more general
SetMouseWheelActions.
- Refactored and tidied up wxMathPlot's mpWindow code involved with
fitting, zooming, and panning.
- Consolidated long lists of duplicated member variable initializers to
a new mpWindow private delegated constructor.
- Provided provisional Zoom In/Out Horizontally/Vertically toolbar
icons that need improvement by a graphics designer.
- Provided gitignore entries for the Qt Creator IDE
Prior to this change, the schematic connectivity was updated any time a
change was made to a connectable object. Now the connectivity is only
updated when an object change actually affects the connectivity. Other
properties like line width, fill type, custom fonts, etc. will not cause
the connectivity graph to be rebuilt.
The SCH_COMMIT flag SKIP_CONNECTIVITY has been removed. All schematic
objects can test if they are connectable and if there have been changes
to any connection properties that require a connectivity rebuild.
Remove duplicate rebuild connectivity calls from editor control tool.
This was causing the tangling end test to get called four times on every
undo and redo action because the dangling end test is already called in
the connectivity graph calculation code.
Update connectivity when changing label names which fixes an unreported
connectivity bug.
Short description:
Works for Symbol and Footprint Editor behind an advanced config option.
For Symbol Editor it is shown for a single item selection (library or symbol).
For Footprint Editor it is shown for a footprint selection.
(fp editor allows a single tree item selection only).
Option stays hidden if current frame has been modified.
Also small fix(?) for similar action to the project manager.
(Call for the Execution has moved inside the file loop.)
Fixes: https://gitlab.com/kicad/code/kicad/-/issues/15736
Avoid O(N^2) by spatial sorting, don't run checks if the bounding boxes
don't overlap.
A second copy is ordered by type to help classes that only want to check
a few types having to walk the whole list.
(Don't require the same text to be put into the tooltip.
It's error-prone and also disallows us from having a separate
tooltip for toolbar buttons.)
Fixes https://gitlab.com/kicad/code/kicad/-/issues/17009
(cherry picked from commit f4a085575a)
Apparently we forgot to handle schematic connectivity changes when redoing
undone changes. The same connectivity update logic that was added to the
commit object is now used in SCH_EDIT_FRAME::PutDataInPreviousState().
Please note the granularity of connectivity rebuilds on schematic object
property changes is at the object level which is overly aggressive. This
means that connectivity rebuilds will happen when connectivity changes
that have nothing to do with and objects connectivity are changed. Until
objects can handle their own connectivity state, this will have to suffice.
AppleClang seems to generate different typeinfo for forward declarations
of classes than the actual class object, so the any_cast for the pointer
parameter fails due to the types not matching (even when they look the
same when printed in the error).
Instead, we must always ensure we use the actual object when defining
the parameter type when placing the object into the event.