The log viewer tool lets you inspect all the intermediate stages of the routing algorithms. This patch:
- Adds source location tracking of the debug calls (need to use the PNS_DBG macro, sorry)
- Moves some wxLogTrace calls to DEBUG_DECORATOR::Message() so that messages can be displayed alongside the corresponding geometric shapes
- Fix walk failure if the input path starts on the hull boundary
(happens very often during tight walkaround)
- Input line can have a loop at the end now. Such condition happens when the routing
destination point lies inside the hull of a colliding primitive.
By 'root lines' we mean the oldest traceable ancestor of each track moved by the router
(i.e. after shoving for a while, the root of each shoved line is it's latest non-shoved version).
With this we can teach the OPTIMIZER more tricks, such as the LIMIT_CORNER_COUNT constraint. It ensures
the results of the optimization will not be less cornery than the original line, reducing the feeling
of the optimizer being too intrusive.
Fixes picking the last segment instead of the last vertex to drag (when one is requested).
I didn't notice any behaviour change of the dragger wrs to arcs.
From discussion with Tom, Jeff, and Wayne, it appears as though
we all agreed that Highlight Collisions should behave like it
did in 5.1, which is to say, always just highlight collisions,
never prevent them. The Allow DRC Violations checkbox just
controls whether or not you can fix/commit the head line if it
has violations; you are not prevented from moving the head
line to violating positions in any case.
Right now, there does not seem to be much demand for a separate
"Stop at First Obstacle" mode since Walkaround basically does that
but better, but we agreed that if there is demand for it in the
future, it should be implemented as a new router mode rather than
a behavior of Highlight Collisions mode controlled by the Allow
DRC Violations checkbox.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/7828
Using a synthetic via here doesn't quite let us use VIA::PushoutForce
because it will use the wrong clearance, and also doesn't quite have
the logic we want. I am not familiar enough with PushoutForce to know
if its logic is a bug in other cases, so instead I just brought in the
parts of its algorithm that are needed here.
Additionally, we prevent pushing more than once from a given obstacle,
which causes walkaround to be more successful when routing diff pairs
against a large collider such as a keepout area.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/8232
I'm not sure where the magic number of "4x worst" came from, but it's
been around forever. This is extremely inefficient as it negates much
of the power of r-tree filtering in dense designs. If we really trusted
it, we could set this just to worstClearance. Keeping it above the worst
clearance by a little bit seems to provide enough of a speed improvement
to resolve the test cases I have, so I'll go with that for now.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/7777
This allows fixing/committing what is on screen even if the current
cursor position doesn't have a valid solution (the old behavior
would result in the current trace disappearing)
The previous method fails if the primitive is an expanding or
contracting pair (diagonals going inward or outward) resulting in
incorrect orientation, which then leads to incorrect candidate
gateways being generated and no solution found.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/8185
TODO:
* The resulting line chain is broken with the arcs, and
this appears to be some subsequent issue I was not able to pin down.
* The requirement of a correction factor is not clear to me
VIEW::GetBoundary() returns the entire view area, not the visible area.
Surprisingly, we had no API for this, so I added one.
Also, changed the dragger behavior to toggle between optimizing just the
modified area and optimizing the visible area, i.e. we will now never
optimize off-screen portions of the dragged track.
CHANGED: The interactive router settings now include a switch to
enable or disable optimization of the entire dragged track,
which is now disabled by default. When enabled, the router
will reroute the entire track (from the dragged segment to
the closest pad/via in each direction) to be more optimal.
When disabled, only the area around the dragged segment
will be modified.
CHANGED: The "optimizer effort" slider is removed from the interactive
router settings dialog, as this setting did not have any
meaningful impact in most cases and was a source of confusion.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/5918
Bitmaps are now identified by an enum class instead of by pointers.
Bitmap loading and caching is now handled by a class in common, and
we no longer compile most bitmaps into the binary, so there is no
longer a bitmaps static library.
Instead, bitmaps are archived to a .tar.gz file which is installed
in ${KICAD_DATA}/resources/images.tar.gz
The source PNGs are checked in to Git as the original CPP files were,
so that people can build without the required dependencies to convert
SVGs to PNGs.
Initial support is also added for dark theme icons, although this
is not yet exposed in the GUI.
Stubs are present for multi-resolution image resources, but this is
not fully-baked yet and could use some refinement.
- Make DIALOG_SHIM::SetPosition working)
- rename PCB_BASE_FRAME::SelectLayer() to PCB_BASE_FRAME::SelectOneLayer()
- PCB_BASE_FRAME::SelectOneLayer(): make dismiss dialog by escape key working.
Fixes#7578https://gitlab.com/kicad/code/kicad/issues/7578
Keeps parent information such as UUID and pad removal options for
elements that are just updated in routing rather than created fresh or
fully destroyed.
Fixes https://gitlab.com/kicad/code/kicad/issues/7460
Moves rule area triangulation outside of layer iterator and keeps the
triangulation_valid flag false until the triangulation has been
completed.
Fixes https://gitlab.com/kicad/code/kicad/issues/7032
Show the dragging tools in the context selection menu when a single
component is selected. Prevent the dragging tools from appearing when
more than one object is selected, as they will not activate anyway.
Fixes https://gitlab.com/kicad/code/kicad/issues/7258
1) take hole plating thickness into account
2) build hulls specifically for via so pushing back to the hull doesn't
have to try and figure out how much extra to push back for the via
and/or hole
3) clear MK_HOLE flags when colliding with the pad itself so that we
don't get stuck in HOLE mode
4) move flag setting/clearing to updateNearest() so that it will keep
track of the state of the nearest collision
Fixes https://gitlab.com/kicad/code/kicad/issues/6913
The previous auto-posture algorithm worked great for pads
but not so great when starting routing from a segment.
Now we can optimize the posture based on the starting
segment if one exists, to maximize the "obtuseness" of the
posture.
CHANGED: When dragging a footprint, if the mouse isn't directly over the
footprint anchor when 'D' is pressed, the footprint center will jump to
the mouse position when dragging begins. This MR uses the current
mouse position instead of the footprint anchor as the inital reference
point when starting to drag.
Fixes https://gitlab.com/kicad/code/kicad/issues/6813
CHANGED: manually-placed (stitching) vias won't have their nets automatically updated
(unless the via is placed directly on a track segment)
CHANGED: stitching vias can be placed on footprint pads and pick up their nets
Fixes https://gitlab.com/kicad/code/kicad/-/issues/5484
Also removes EditToolSelectionFilter which was misused in more places
than it was used correctly. The original point of the client filter
was to move the logic to the point of use, which the
EditToolSelectionFilter sort of obviated anyway.
Fixes https://gitlab.com/kicad/code/kicad/issues/6751
Also, don't use ViewGetLayers() for visibility checking. For some
things (like vias) it deals with synthetic layers, not board layers.
Fixes https://gitlab.com/kicad/code/kicad/issues/6328
Two main changes: netclass values need to go through the DRC engine
so they can interact with other rules. They're also now dependent
on the layer being routed as well as the start object.
Also make the controls adjust to each other better. For instance,
copy-track-width needs to turn off when you select a particular
track width, and a particular track width needs to zero out when
you choose copy-track-width.
Fixes https://gitlab.com/kicad/code/kicad/issues/5951
Also checks the nets when checking for a pad being flashed on a particular
layer (so we don't end up thinking the currently-routing-collided track
is a connection).
When invoking "Break Track" from the RMB context menu, Pcbnew
gets and uses the current position of the mouse to figure out where to
break the track. The problem is that the mouse has to be moved to select
the "Break Track" menu item, and the mouse position used is wrong. This
can result in the break in the wrong location or not happening at all.
CHANGED: This commit determines if the "Break Track" was invoked from a
context menu or a hotkey. If a hotkey, it uses the current mosue position.
If a context menu, it uses the original postion of the mouse when the menu
was opened.
The previous way of displaying a context menu was being
usurped by the selection tool because it was displaying its
menu before DispatchContextMenu was called, and all other tools
explictly use the TOOL_MENU system.
Fixes https://gitlab.com/kicad/code/kicad/issues/5762
These objects can now be used in advanced DRC rules and
not just for keeping things out. Also remove the restriction
that at least one of the "basic" keepout rules must be set,
so that these areas can be used for more advanced rules.
You can now enable and disable snap to grid when drawing/editing across
all apps. You can also tie snap to grid to the visibility of the grid
to allow rapid enable/disable via grid display.
Edge.Cuts have no parent but cross all layers, so we need to handle the
"all layers" case. Also renames "IsPadOnLayer" to "IsOnLayer".
Fixes https://gitlab.com/kicad/code/kicad/issues/5351
The use of printf, wxLogDebug, and std::err/std::out causes excessive
debugging output which makes finding specific debugging messages more
difficult than it needs to be.
There is still some debugging output in test code that really needs to
be moved into a unit test.
Add debugging output section to the coding policy regarding debugging
output.
This option removes copper layers from pads and vias where they are not
connected to other board elements. This allows the inner layers to be
more closely routed if the via landing pad is not needed.
Fixes https://gitlab.com/kicad/code/kicad/issues/1835
Rather than extra splitting of the RTree that we use for collision
detection on types and direction, we limit this to checks on
layer-by-layer basis. This also allows for layer expansion by using
deque of RTrees rather than fixed index
This fixes the missing layernumbers and keeps the layers in a deque to
prevent re-allocation on expansion.
Replaces ad94b6205
Rather than extra splitting of the RTree that we use for collision
detection on types and direction, we limit this to checks on
layer-by-layer basis. This also allows for layer expansion by using
vector of RTrees rather than fixed index
The first one keeps a log of events (start routing, mouse motion, etc).
The second allows for adding temporary debug drawings and messages which are stored synchronously with the events in LOGGER.
The event stream together with the PCB design (now with UUIDs) can be used to deterministically replay routing bugs as the user sees them.
Various architecture upgrades to support this.
Creating a BOARD now requires a valid PROJECT, which caused
some (mostly transparent) changes to the Python API internals.
ADDED: Project local settings file
CHANGED: Board design settings are no longer stored in PCB file
CHANGED: Net classes are no longer stored in PCB file
CHANGED: Importing board settings now reads boards, not just projects
Fixes https://gitlab.com/kicad/code/kicad/-/issues/2578
Fixes https://gitlab.com/kicad/code/kicad/-/issues/4070
1) An actual distance of 0 is still a collision, even if the allowed
distance is 0.
2) Be consitent about edges and interiors. Everyone expect the edge
of a RECT to be part of the RECT; same with a CIRCLE. SHAPE_POLY_SET
shouldn't be any different. (And SHAPE_LINE_CHAIN was a split-
personality with the edge considered part of it for Collide() but not
for PointInside()).
There were a lot of plotters, exporters, etc. that were rolling their
own implementations.
This also introduces a lazily-built set of SHAPE objects for doing
collision detection and some forms of rendering (and later DRC).
It's currently only supported in the Footprint Editor. It could be
easily added to the board editor (all the code is there), but the board
editor is a little short on room in the drawing tools toolbar.
The arc shapes need to connect with their adjacent points. By storing
the relevant points, we allow exact point matching on both ends of the
arc as well as localize point storage.
Note: this is a basic implementation but it could be
improved once we include bus information in the netlist
and pcbnew can natively keep track of buses and nets
instead of just nets.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/4158
* Consolidate the measure tool into one tool (this gives cvpcb
unit changing and snapping capabilities in its measure tool)
* Transition cvpcb to use actions for the sketch modes
* Replumb how magnetic items settings are stored and used
Springback introduced a new PNS structure to check for valid placement
as well as separating the Fix/Commit routines. This updates the
remaining placers to utilize the new structure.
Fixes https://gitlab.com/kicad/code/kicad/issues/4008
Dragging using highlight caused an assertion as the duplicated lines
were being re-added with their old links still inplace. While this was
probably fine as the link references shouldn't be stale, we clear them
to allow for the possibibility that the dragger will adjust the links in
the future.
The bump direction should only be concerned with physical points to
calculate the heuristic for loop-back otherwise we get too many false
positives.
Fixes https://gitlab.com/kicad/code/kicad/issues/3971
Needs to handle changes to PNS_ROUTER. This does not implement
springback or backspace in DIFF_PAIR_PLACER, only allows committing
lines as before
Fixes https://gitlab.com/kicad/code/kicad/issues/3988
Not all footprint pad items are connected to elements in the PNS, so if
we don't find the parent element, we avoid later crashes/assertions by
skipping their addition
Fixes#3972 | https://gitlab.com/kicad/code/kicad/issues/3972
This is due to the fact the PNS_TUNE_STATUS_POPUP was not hidden
before calling its destructor.
It is now fixed by setting its virtual destructor calling Hide().
TODO: see if it could be used for any class derived from STATUS_POPUP.
(We already have this issue in other places)
CHANGED: Settings are now stored in versioned sub-directories
ADDED: First-run dialog for migrating settings from a previous version
CHANGED: Settings are now stored as JSON files instead of wxConfig-style INI files
CHANGED: Color settings are now all stored in a separate settings file
CHANGED: The symbol editor and footprint editor now have their own settings files
CHANGED: Color settings are no longer exposed through BOARD object
CHANGED: Page layout editor now uses Eeschema's color scheme
Settings are now managed through a central SETTINGS_MANAGER held by PGM_BASE.
Existing settings will be migrated from the wxConfig format on first run of each application.
Per-application settings are now stored in one class for each application.
Now that the tool framework gracefully shutdowns the tools, graphics
can be created on the stack inside a tool function.
This reverts commit f6881ce3de.
The new line collision search uses BBox() to check for colliding
objects. BBox in the SHAPE_LINE_CHAIN did not include width as the
chains were assumed to be zero-width. This is not the case for
PNS::LINE elements.
We mostly don't notice this because DRC checks for SEGMENT collisions
but it becomes obvious/annoying when we cannot place a track for unknown
reasons and the snap-back doesn't take line width into account.
Fixes#3776 | https://gitlab.com/kicad/code/kicad/issues/3776
When routing where we hit multiple obstacles, we choose the shortest
path to allow overlapping hits to choose the shorter path, thus giving a
better chance of returning true for the DRC-compliant path.
Fixes#3773 | https://gitlab.com/kicad/code/kicad/issues/3773
* Split up the thirdparty code into the thirdparty folder (#3637)
* Create a new kimath static library containing all the math functions
This is part of cleaning the build system for #1906.
This is the first step to allowing non-segments in the line chain.
External routines cannot be allowed to change the line chain without
going through the internal routines. To accomplish this, we remove the
Vertex() and Point() access routines and only leave the const versions.
Transformations are given for both points as well as the chain itself.
It happens on Windows, and wxWidgets 3.1.3.
It is created by PNS_TUNE_STATUS_POPUP instance used in tool.
I am pretty sure this crash is created by the stack switching
when managing events, due to some changes in wxWidgets code.
the fix creates the instance on the heap, instead of on the stack.
This is not the first time I see this kind of issue.
The IFACE for kicad refers to the board hidden state, not the
temporarily hidden state used by the router. We allow snapping to items
that have been removed by the temporary router state to provide snapping
to original track location.
Fixes#1827 | https://gitlab.com/kicad/code/kicad/issues/1827
(cherry picked from commit 1e64524afb)
This update replaces the existing uses of unique pointer creation with
the C++14 std::make_unique call that provides proper memory release in
event of an exception.
The pointer passing for display options is deprecated. This removes the
excess casting as the EDA_FRAME didn't need the base call with no value.
All requests for display options are now returned const and are updated
with a Set() routine after modification.
In Gerbview, this resolves an issue where the display options were not
stored because it was receiving the NULL from EDA_FRAME.
This adds the pad to die parameter from each pad to the total line
length of the tuned line for comparison with desired.
Fixes: lp:1711541
* https://bugs.launchpad.net/kicad/+bug/1711541
The previous fix for optimizer breakout length calculation was providing
better scores to longer traces rather than longer breakouts as intended.
This limits the length scoring to breakout only.
Flipped footprint pads may report their "primary" layer as not the front
layer in high contrast. We need to check whether the board item for
snaps is visible on the active layers.
Fixes: lp:1847877
* https://bugs.launchpad.net/kicad/+bug/1847877
The router items may not yet have a board item parent. In this case,
the reference to the parent needs to be guarded before dereferencing.
Fixes: lp:1847717
* https://bugs.launchpad.net/kicad/+bug/1847717
Items and layers that are hidden in the view should not be used as snap
points. This happens in multiple locations:
1) Grid Helper for normal tools
2) TOOL_BASE::snapToItem for router
Resolves KiPro Issue #116
Fixes: lp:1833128
* https://bugs.launchpad.net/kicad/+bug/1833128
Calling Init() in an action routine will reset the router
settings to the state they were in when first starting pcbnew.
The Init() function is automatically called when the tool is created.
Fixes: lp:1843810
* https://bugs.launchpad.net/kicad/+bug/1843810
Before, if the tools were activated from the context menu,
they would start drawing where the menu item was selected
instead of where the menu was opened.
The INT_MAX limit for most elements makes sense only for single-segment,
straight line elements. For elements that accumulate lengths, we should
utilize the long long int (64 bits) to allow for greater lengths.
Fixes: lp:1842367
* https://bugs.launchpad.net/kicad/+bug/1842367
This modifies that logic for optimizing pad connections.
1) Custom pads now prefer the connections to their longer side. This is
particularly important to match the behavior in a rectangular pad
2) Costs for straight lines are non-zero. This favors fewer connections
3) Custom pad breakouts are at the polygon boundary rather than offset
by 0.1mm. This also matches the circular and rectangular paradigm.
4) Line Heads contained in the pad itself are not optimized. This
prevents escaping the pad unnecessarily
Fixes: lp:1824182
* https://bugs.launchpad.net/kicad/+bug/1824182
If the layer is not shown, we should not be allowing it's items to be
selected and chosen as the start/end items in the router.
(cherry picked from commit fc1fb7a590)
Drag is used for multiple modes so should be initialized when starting.
The line that we generate however needs to be pristine in drag, so we
clear the links in the duplicate.
Fixes: lp:1839612
* https://bugs.launchpad.net/kicad/+bug/1839612
The problem is that wxEVT_CHAR_HOOK doesn’t do the key translation
properly. wxEVT_CHAR does, but we only get to that if we skip the
event at the end of the tool’s event processing loop, which most tools
don’t do. (Selection tools, point editors, pickers, and a couple of
others do skip, which is probably why this didn’t get reported earlier.)
I played around with a couple of ways to fix wxEVT_CHAR_HOOK. Most of
them don’t work, and the few egregious hacks I tried weren't cross-
platform.
So I’m changing it so that most tools now skip at the end of their
event loops. I left out a couple that I felt were high risk (length
tuning, for instance). But there’s still enough risk that I’m 100%
sure it will break something, I just haven’t a clue what.
Fixes: lp:1836903
* https://bugs.launchpad.net/kicad/+bug/1836903
This sets the clearance outline to a slightly better opacity to allow
viewing of items behind the routed track clearance. Future work should
include making this a configuration option.
Fixes: lp:1826719
* https://bugs.launchpad.net/kicad/+bug/1826719
(cherry picked from commit e0589d2908)
Includes the addition of an onSetCursor() handler which must be called
from both the GAL canvas AND the GAL backend (at least on OSX) to prevent
cursor flickering between (for instance) pencil and arrow.
Also includes new architecture for point editors which allows them to
coordiate cursors with the editing tools (so we can switch to an arrow
when over a point).
We were running into various corner conditions where a tool's event
loop would exit while the tool was still active, or the tool would
get popped while we were still in the event loop. (A lot of these
had to do with the POINT_EDITOR's, but not all of them.)
The new architecture:
1) tools always do a Push()/Pop()
2) everyone is responsible for their own pops; no more stack-clearing
on a cancel
3) CancelInteractive events go to all tools to facilitate (2)
It's a bit of a hack because they're statically initialized and
so we can't make use of the _() macro. We do still want it in the
code, however, because the string harvesting is based off of it.
Fixes: lp:1833000
* https://bugs.launchpad.net/kicad/+bug/1833000
* Circular gal<->common dependency bites again. Add 'common'
again to the 'kicad' target's link libraries. libgal.a depends
on SHAPE_POLY_SET in common, but common depends on gal. This is
not a real fix, but it unsticks Linux compilations. Fixing the
circular dep is an existing bug: https://bugs.launchpad.net/kicad/+bug/1832229
* Missing LEGACY_HK_NAME for ACT_EndTuning: this was probably accidentally
allowed on some other platform due to different wxString constructors.
On Linux/GCC, it breaks.
* Unused include from /kicad in export_vrml.cpp. This is not findable
on Linux, as pcbnew does not include the right dirs for this to be
found. But it's not used, so just remove it.
This removes the remaining hard-coded segments counts and replaces them
with the relative error calculation where the segments per arc is
determined by the maximum error we allow (smaller arcs = fewer segments)
When dragging, you will seldom want to snap to the original item. This
commit prevents the "size of track" snap to the original item and
replaces it with an auxilary grid origin for the snap. This means
effectively that the snap will happen to grid for the original item
unless the cursor is closer to the original item line than it is to the
grid crossing.
Fixes: lp:1820248
* https://bugs.launchpad.net/kicad/+bug/1820248
Allows 0 to 4 chamfered corners, not only one.
A custom shape allow this kind of shape. However because it is a primitive,
it is easier to edit and it support thermal reliefs.
When adding colors to length tuner, we depend on the window background
color, which can be set by the user, resulting in an unreadable text.
This uses the HSL conversion to detect which shade it should use for the
coloring.
Also adds test cases for round-trip HSL and HSV conversion
Fixes: lp:1814530
* https://bugs.launchpad.net/kicad/+bug/1814530
Libpolygon can provide its own includes via target_include_dirs PUBLIC.
This means any linking targets do not need to specifiy them manually.
As common requires polygon, the polygon dep is also now no longer
required downstream of libcommon, as it's transisitvely implied
by libcommon's target_link_libraries.
This resolves a circular dependency previously detected and also
simplifies CMakeLists.
The bitmap definitions (BITMAP_DEF and so on) do not
have any dependencies on other libs, including WX. This
means the bitmaps library can be isolated from the other
dependencies.
Common now depends on bitmaps, and libraries that depend
on common can pick it up from the common target_link_libraries,
as it is PUBLIC. This means a lot of targets no longer
need manual bitmap linkage.
This avoids a circular dependency that was previously reported
by static analysis.
Avoiding pulling in WX and other headers into the include
tree of each bitmap .cpp is a huge speed up (around 10x) in
compilation, and the generated static library is also 10x
smaller (20MB vs 200MB)
Add common as a link library to pnsrouter,connectivity.
THese library do still use common code (including bitmaps,
via base_screen.h) and this allows them to pick up the libcommon
includes correctly.
Fixes and issue where forces were calculated using the current track
width even in differential placer mode where the tracks are specified by
the differential pair width.
Fixes: lp:1814480
* https://bugs.launchpad.net/kicad/+bug/1814480
Snap/Grid modifier follow preferences from 05483a06c6. Brings Shift/Alt
modifiers into alignment with drawing tools. Shift modifier still
conflicts when choosing end item with the complete track click-modifier.
This will be addressed in v6 with a larger preference setting.
Snapping to item options were partially disconnected when unifying
framework. This re-attaches the options and adds the additional option
for graphical items.
Fixes: lp:1801377
* https://bugs.launchpad.net/kicad/+bug/1801377
The selection for interactive drag uses cursor position off grid to
choose the element for dragging. We pass the same value into the start
dragging routine to choose the correct end of the element.
Fixes: lp:1813665
* https://bugs.launchpad.net/kicad/+bug/1813665