After importing, some schmeatics are moved, item at a time. When moving
schematic items, we need to update the RTree bounding box to allow
searching on position
Fixes#3915 | https://gitlab.com/kicad/code/kicad/issues/3915
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.
When searching for fields, the code was sometimes comparing translated and not translated names.
This is an issue for mandatory fields, in non English languages.
Translated field names should be used only in messages.
Using a UTF8 char inside a C string can work on unix, but not on Windows.
It must be explicitly converted in a unicode value inside a wide string
(like a wxString)
When adding a .pretty library, if the actual .pretty folder were
selected then no folder would be added to the list of libraries.
This led to an assert when normalizing paths, and the library
not being added.
Add also a comment in common.cpp about using wxLocale instead of setlocale.
DIALOG_UPDATE_FIELDS has issues in non English languages, due to its design.
The fix for mandatory fields is only a partial fix that avoid major issues.
The immediate action option clears (or not) the position of the events.
We use this to determine if the command should start at the given
position or merely activate the tool. This was being checked in the
menu options, which only activated for tool commands in the context
menu. Moving to the process event, we catch hotkeys as well.
This also restores the previous logic in eeschema that used a static
variable for storing wires rather than the private class variable.
Starting the draw event now picks up from the existing wires when
activated in immediate mode.
Fixes#3891 | https://gitlab.com/kicad/code/kicad/issues/3891
If the point editor is active and a change is made, the tool was shut down after
deleting the board data, in the frame Dtor.
Therefore a crash happened because a call to GetScreen()->SetModify(); using a null pointer,
and after the test for modifications was done.
Now tools are stopped when closing the window, before any test for modifications.
The parents are the only items that live in the RTree, so the rebuild
check needs to recurse into the sub structure to check for selection
Fixes https://gitlab.com/kicad/code/kicad/issues/3858
Previously, OnModify() was called for a lot of items, even when not modified.
For large schematics, the calculation time was unreasonable ( several minutes)
It is now called only a few times.
Shared sheets can create issues in a complex hierarchy.
This is not related to sharing sheets between projects.
(Sharing sheets between projects is not supported by Kicad)
These help files are created from the .md source files and contain I18n strings,
but do not exist if kicad is not built from sources.
However they are needed by translators.
ADDED: Back annotation algorithm,
eeschema back annotation dialog
CHANGED: added some minor helper methods to SCH_REFERENCE_LIST and SCH_REFERENCE,
split SCH_REFERENCE_LIST::CheckAnnotation on 2 parts to reuse code
The RTree does not have a deterministic iterator, so extracting items
may be in arbitrary order, causing issues as the schematic appears to
change when comparing to previous revisions.
This uses the SCH_ITEM comparison operator to for ordering by type, then
by custom sorting within type.
For the netlist, we choose the first available unit in the sheet for
each component.
Fixes#3811 | https://gitlab.com/kicad/code/kicad/issues/3811
We need to ensure that the appended sheet does not already exist in the
parent hierarchy. We do this for new sheets but need to check for
existing ones as well.
Fixes#3806 | https://gitlab.com/kicad/code/kicad/issues/3806
When opening the symbol editor without loading the schematic editor the
default text size setting is not loaded and was defaulting to the old mils
setting. Convert the default text size to internal units in case the
symbol editor is launched first.
Set default line, bus, and selection highlight width to internal units.
Change new symbol dialog pin position offset control from a spin control
to a text control so the unit binder object can handle converting to and
from internal units.
Fixes#3802https://gitlab.com/kicad/code/kicad/issues/3802
The components are reloaded when libraries change and when loading a
schematic. The initial load does not contain pins and so bbox hits may
miss elements.
Fixes#3800 | https://gitlab.com/kicad/code/kicad/issues/3800
Convert minimum sheet width and height definitions from mils to internal
units.
Add comment to sheet minimum width and height definitions to indicate
sizes are in mils.
This file should really be included first from any source file, because it
contains platform specific workarounds.
This also groups includes a bit by source.
Bus submenus weren't sending events because tool wasn't set.
Net names need to be unescaped for display.
Unfolding wasn't drawing the new wire segments.
SCH_LINE_WIRE_BUS_TOOL could get caught in an infinite event handling loop.
Fixes https://gitlab.com/kicad/code/kicad/issues/3796
Scope: NETLIST_ITEM, CONNECTION_TYPE, ELECTRICAL_PINTYPE,
NET_CONNECTION, NETLIST_ITEM, GRAPHIC_PINSHAPE
Note, the pin type enum had PT_ added to the front to prevent
shadowing of the INPUT symbol on msys2 (see discussion at
c17c9960d8)
Spice lines defining circuit elements are contained in a .subckt ..
.ends block. The intervening lines should be exported to the netlist as
well as the control lines.
dynamic_cast could not resolve the cast chain leading to a null undo
point. Since we know the item to be LIB_ITEM, we can use static_cast
anyway and avoid the overhead.
Fixes#3771 | https://gitlab.com/kicad/code/kicad/issues/3771
This also removes vector cover types which do nothing except obfuscate
the underlying implementation.
Mainly changes SCH_SHEET_PINS and CONFIG_PARAM_ARRAY (which will soon
be replaced by Jon's new stuff).
* Ensure the lib edit undo checks for pointer validity
* Assert that the frames are the correct type (satisfies coverity)
* Convert C casts to C++ casts
It served us (mostly) well for more than a decade. It helped KiCad grow
before the std:: came into decent shape or speed. It was a good little
list.
RIP DLIST 2008-2020
This moves EESchema DLIST structures to rtree. These changes are more
fundamental than the pcbnew changes from 9163ac543888c01d11d1877d7c1
and 961b22d60 as eeschema operations were more dependent on passing
drawing list references around with SCH_ITEM* objects.
Add try/catch blocks when saving symbols using a plugin object directly
to prevent potential unhandled exception bugs. I'm not totally happy
using wxLogError to flag exceptions but I couldn't think of a better way
to handle them.
Fixes#3665https://gitlab.com/kicad/code/kicad/issues/3665
"INPUT" declaration in sch_text.h generate *a lot of* compil warnings.
So it is replaced by PS_INPUT.
For consistency, others member of enum class PINSHEETLABEL_SHAPE are also renamed.
For legacy reasons, it stored left and right "spin" as flipped integers in the file format.
But the code handled the flip in multiple locations rather than just doing it on file io.
This change unifies the internal code and does the mapping in the file I/O.
and switch to the paren sheet from the context menu.
Mainly, change the way the dialog is managed from the sch_edit_frame:
the pointer to find the created dialog is replaced by a call to wxWidows::FindWindowByName().
Using a pointer set and cleared by events is a recipe for crashes.
ADD: Allow drawing of current and phase traces with dotted style
for easier differentiation from voltage and magnitude. The option
is available in the View menu of the simulator.
* 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.
ADD: User can now decide to keep the hierarchy navigator open while working
on a schematic.
This behavior can be configured in eeschema->preferences->eeschema->Editing options.
New v6 can test multiple points not previously considered. This
corrects the test to ensure that junctions are correctly added when
lines are present.
Fixes#3729 | https://gitlab.com/kicad/code/kicad/issues/3729
Change default text size and anchor positions back to mils. Internal
units need to be converted at the source, otherwise the size will get
converted twice in the schematic editor configuration code which will
result in the default text size being off by the conversion factor.
Fixes#3730https://gitlab.com/kicad/code/kicad/issues/3730#
Derived symbols were causing the symbol properties grid to assume that
the reference could not be edited because derived symbols inherit the
reference from the parent symbol. The flattened symbols in the schematic
still have the parent set which cause the issue. Clearing the parent of
the flattened symbol resolves the issue.
Fix a minor bug in the symbol information of derived symbols show in the
symbol chooser dialog.
Fixes#3723https://gitlab.com/kicad/code/kicad/issues/3723
When dragging we'll create new line segments between selection-edge
objects and things outside the selection that they're attached to.
These new lines have their IS_NEW flags set, but shouldn't fool us
into thinking that the selection contains pasted (or placed) items.
Abandon the previous behavior of always creating a new copy of the
buffered symbol which required a re-parenting of every derived symbol
to prevent accidentally orphaning a derived symbol parent symbol
pointers.
This change required something that should have been done a long time
ago by adding an assignment operator to LIB_PART. The assignment
operator makes it possible to create a single copy of the symbol to
buffer which gets updated rather than deleting the last copy and
replacing it with a new copy.
Add a clear method to the MULTIVECTOR class so that existing LIB_PART
object draw items can be removed before assigning the draw items from
the part being assigned.
Fixes#3672https://gitlab.com/kicad/code/kicad/issues/3672
m_connected_bus_item and m_connected_bus_items pointers were not cleared when
rebuilding connection data.
So in some cases these pointers were not updated, and created a crash
once the old data was no longer available (deleted)
Create an IsContentModified() function in each frame
that provides if the content has been modified, and
use that instead of always copying the checks.
ADDED: a config option to select (when clicking on a pin) the parent symbol
or only the pin.
Previously, when cliking on a pin, the parent symbol was selected.
Allowing the pin selection is mandatory in cross probing, and to display the pin
characteristics in some cases, for instance if a pin is flagged by the ERC tooL.
Do not check the connection graph when editing the symbol library table
from KiCad because the connection graph object is no valid when the
schematic editor is not open.
Fixes kicad/code/kicad#3682
There is no need to create a new root symbol and re-parent all derived
symbols when renaming a root symbol. Just rename the buffered root
symbol.
Test for duplicate symbol names when validating the properties dialog
information to prevent broken symbol pointers.
Fix a field drawing bug when editing derived symbols.
Fix bug to insure duplicate symbol names cannot exist in a library.
Fixes kicad/code/kicad#3654
Fixes kicad/code/kicad#3672
The library manager update part function was orphaning the root symbol
of derived symbols when the root symbol was edited. Re-parent the
inherited symbols when updating a root symbol.
Update the mandatory field attributes whenever the parent is set for
derived parts. This will show the fields with the correct attributes
in the symbol editor and viewer.
Using std::unique_ptr to hold the current symbols was deleting the
pointer on symbol selection changes causing the pointer in the library
manager buffer to be stale in some cases. Revert back to using a
simple pointer and manual clean up as required.
Prevent derived symbols from being saved to a different library to
prevent orphaned symbols.
Fixes kicad/code/kicad#3649
The schematic symbol swap function still had the old pin map swapping
code which overwrote the correct pin maps that were rebuild after the
library symbol was update. Duh!
Fixes kicad/code/kicad#3661
Apparently the schematic and symbol library editors now have common
selection filtering so checking for a valid LIB_EDIT_FRAME when
filtering for the schematic editor frame prevented all schematic
object from being selected.
The symbol editor selection criteria did not include LIB_FIELD objects
which allowed the root symbol name to be changed causing broken derived
symbol links. Disable the symbols fields from being edited when a
derived part is shown in the editor to prevent this.
Add a missing warning that deleting a root symbol used to derive other
symbols would also delete all derived symbols from a library. Give the
user a chance to cancel the delete operation in this case.
Fixes kicad/code/kicad#3654
The new symbol library changes left stale pointers in the connection
graph because the SCH_COMPONENT object pins were not always getting
updated when a new copy of the library symbol was updated.
Force a full update of the connection graph whenever the schematic
library symbol links are refreshed. This seems like overkill but it
ensures that there a no stale libraries items left in the connection
graph.
Fixes kicad/code/kicad#3658
Some windows systems will not fall back correctly on our standard
initialization step, so we need the additional std:: error fall back to
catch them.
Fixes#2620 | https://gitlab.com/kicad/code/kicad/issues/2620
Use the root symbol when comparing against the cached symbol when
checking to see if a symbol changed. When the original symbol is
a derived symbol, the test will always fail.
Force a symbol link refresh on a rescue to prevent stale links from
crashing the connection graph refresh when running the project rescue
on demand.
Force a symbol link refresh on project load to prevent stale links from
symbol link changes from crashing the connection graph when the project
rescue in invoked.
Fixes kicad/code/kicad#3645
ADDED new option to set selection highlight thickness.
Also change selection shadow width constants to
make the selection thickness change less drastically
with the zoom level.
Replace some C casts with C++ dynamic_cast.
Fix iterator bug when deleting inherited symbols from legacy file format
symbol library cache.
Remove unnecessary const when return wxString object instead of reference.
Don't clobber value file when load aliases in legacy symbol library cache
parser.
Use actual symbol library LIB_PART pointer rather than a flattened copy
of the symbol. This fixed a bug when displaying the parent field in the
message panel for derived symbols.
Simplify the flatten code by copying the parent and updating the lesser
information from the inherited symbol.
This change completely removes the LIB_ALIAS design pattern an replaces
it by allowing LIB_PART objects to inherit from other LIB_PART objects.
The initial implementation only allows for single inheritance and only
supports the mandatory fields in the derived part because that is all
that the current symbol library file format will support. Once the new
file format is implemented and saving to the old file format is deprecated,
more complex inheritance will be added. The LIB_ALIAS information saved
in the document files was move into the LIB_PART object. This change
impacts virtually every part of the schematic and symbol library editor
code so this commit message is woefully incomplete.
REMOVE: Removed the symbol aliases concept from the schematic and symbol
editors and the symbol viewer.
NEW: Replace the symbol alias concept with simple inheritance that allows
a library symbol to be derived from another library symbol.
The wxArrayString doesn't have a standard assignment operator to
std::vector<wxString> so we iterate over the members.
Fixes: lp:1854562
* https://bugs.launchpad.net/kicad/+bug/1854562
The pin name defined in Eeschema is now available as pad info.
Useful for the board designer (the pin function is displayed in the message panel).
Needed for the Gerber P&P files.
It happens when the spice library model shown in dialog is a large file.
Looks like the wxTextCtrl used to displays the library content
creates the crash ( GKT bug?) for large texts.
The fix is to replace this wxTextCtrlt by a wxStyledTextCtrl.
It has advantages (does not crash... and much faster display) and
inconvenients (cannot disable editing).
Fixes: lp:1853161
https://bugs.launchpad.net/kicad/+bug/1853161
Apparently there are some issues with wxString::CmpNoCase() that was
causing the sheet file name case sensitivity test to incorrectly fail.
Converting the name strings to lower case before comparing resolved
the issue.
Add missing check for root sheet when searching sheet hierarchies for
already loaded schematics. This prevents the root sheet from being
omitted when adding new sheets using the root sheet file name.
CHANGED: Make file name tests case sensitive so that schematic sheet
file names on non-Windows systems can be uses as expected.
Warn users when attempting to use schematic file names that only vary
by case sensitivity that doing so will result in a project that is not
portable to Windows.
Fixes lp:1843415
https://bugs.launchpad.net/kicad/+bug/1843415
When dragging a bus the entry is connected to the bus so the wire
connected to the entry (if any) needs to stretch.
While an entry can connect in the middle of bus, it cannot connect
in the middle of a wire.
Fixes: lp:1849973
* https://bugs.launchpad.net/kicad/+bug/1849973
Remove not used files:
- markdown2html is not used (we are using "Sundown" tool)
- newstroke_font_without_CJK.cpp is removed.
Due to code optimization the full font works fine without memory issues
-fix also a few very minor compil warnings
CHANGED: No longer call schematics noname.sch in standalone mode,
instead don't given them any name so the user can choose on save.
Fixes: lp:1852607
* https://bugs.launchpad.net/kicad/+bug/1852607
We have different requirements than sheet paths for netlists,
so going through that code and then doing a lot of conversions
just ends up with a mess.
We now generate the path directly from the SCH_SHEET_PATH.
Fixes: lp:1852353
* https://bugs.launchpad.net/kicad/+bug/1852353
The sort routine requires the value to be true with a is strictly lower
than b. But just inverting the '<' doesn't yield strictly lower, it
includes '>='. This causes items to jump in the list.
Copy/Paste should keep the original items' designation for all units and
conversion type. If it was specified in the original copy, we should
keep the same specification when pasting (but update for the current
frame)
Handle symbols as well as strings in the "source" lists of
netlist files.
Handle Protex gerber file extensions.
Check for project_name-whatever pattern for files that we
don't recognize (such as project_name-NPTH-drl_map.ps)
NEW: The simulator plot now avoids colors already in use, if possible.
The palette is replaced with a larger one with less saturated colors.
Fixes: lp:1851372
* https://bugs.launchpad.net/kicad/+bug/1851372
NEW: A text entry allows adding signals by name. This includes vectors
not in the signal list, which is useful for plotting calculated vectors
like `let AB=V(a)*V(b)`. It's also a faster way to add a signal if you
know its name already.
Fixes: lp:1851373
* https://bugs.launchpad.net/kicad/+bug/1851373
The new connectivity algorithm had a tendency to cleanup
globally, but that could insert undo records from other sheets
into the current screen's undo stack. Needless to say, this
was a recipie for segfaults.
Fixes: lp:1846247
* https://bugs.launchpad.net/kicad/+bug/1846247
Mainly replacing std::string by wxString to manage filenames.
On Windows, a sdt::string cannot manage a filename, unless using
in many places TO_UTF8 and FROM_UTF8.
So the best way is to use a wxString for filenames and error messages.
Note that since the markup might exist for other reasons, it has
to be turned on with a preference setting. (It goes through a set
of bitflags so the same architecture can be used for other markup
structures that we might want to support in the future.)
Note also that this is more about engineering nomenclature than
visual formatting. In that respect it's more similar to overbar
than italic or bold.
When more than one item has the duplicated timestamp, the update code
doesn't accurately change the duplicates as it only checks that a single
sort has not adjacent duplicates.
This adds all items to a set, ensuring that updated timestamps are
unique to the schematic, which prevents reseting or mangling the units
for duplicated components.
Right click is a pointing operation combined with a request for the
context menu. We need to highlight the user's intended object when they
have moved their mouse to a new element instead of showing the context
menu for the selected object.
However, in cases where the selection is difficult to make, it is
sometimes helpful to select the item first, then right-click. In these
cases, we keep the selection where the mouse remains inside the selected
object's bounding box.
When placing a new component in the schematic, this allows the user to
iterate through all units in the selected component without re-opening
the choose part dialog. It also allows the optional addition of
multiple copies of the same component to be placed in the schematic.
Fixes: lp:1806217
* https://bugs.launchpad.net/kicad/+bug/1806217
Adding a field to a selection changes the bounding box and count for
rotation. This can cause odd offsets when the field is rotated against
the selection center rather than the bounding box as well as with the
component.
Fixes: lp:1849078
* https://bugs.launchpad.net/kicad/+bug/1849078
The break wire command will break all wires at a point. If there are
multiple crossing wires, we need to check to see if a junction might be
required to reflect the new connection.
Fixes: lp:1848450
* https://bugs.launchpad.net/kicad/+bug/1848450
* Fix not drawn shapes for clock low, inverted clock and broken shape for falling edge clock
* Especially, shape for falling edge clock was a not existing shape, therefore useless.
* now falling edge clock and clock low shape are similar.
Fixes: lp:1846901
https://bugs.launchpad.net/kicad/+bug/1846901
The previous converter (maddy) is still available, during some time.
We therefore have the time to choose between them.
sundown is better to convert a md text, but is written to C
maddy has a few issues to convert a md text, but is written to C++
Both have no dependency.
When checking for 3 or more connections, at least one of them must be
a pin. Otherwise we need to fall down to the next check which also
ensures that at least one of them is not parallel.
Fixes: lp:1841458
* https://bugs.launchpad.net/kicad/+bug/1841458
The spin style is stored as a numeric and can be corrupted. We handle
this by triggering assertions for debugging the bad value while
resetting to default if we have non-valid spin values that are outside
the array.
Fixes: lp:1843091
* https://bugs.launchpad.net/kicad/+bug/1843091
This makes the pcbnew multiline text editor natively accept tabs in the
same manner as eeschema and extracts key handling routines to the shim
to allow undo/redo.
Also allows Ctrl-Y under Linux in addition to other platforms. While
not as popular as Ctrl-Shift-Z, it is utilized for Redo in some contexts
When we add enough SHEET_PINs that the vector has to grow, it
re-allocates the vector causing our for-loop to get its knickers
tied in a knot.
Fixes: lp:1842394
* https://bugs.launchpad.net/kicad/+bug/1842394
1) When sheet pins are on 3 or more sides there's no point in switching
the orientation to vertical.
2) Careful that we don't cause the sheet to walk when rotating and the
sheet-name or file-name is longer than the side it's on.
Fixes: lp:1841714
* https://bugs.launchpad.net/kicad/+bug/1841714
Also, if it can't be loaded then at least initialize it so that we
don't crash later (trying to access a NULL screen).
Fixes: lp:1841801
* https://bugs.launchpad.net/kicad/+bug/1841801
Saving individual parts of a schematic can break the schematic. This
moves the schematic save option to save all sheets rather than just the
current sheet.
Fixes: lp:1840383
* https://bugs.launchpad.net/kicad/+bug/1840383
The highlight color chosen from the system highlight doesn't always show
against the schematic sheet background color. Allowing the user to
customize with their KiCad theme makes it configurable on par with other
colors. This is stop-gap until we get full color schemes from
lp:1678345
When editing symbol library reference, this ensures that all units are
modified by the same action and the component is not represented by its
subunits.
Fixes: lp:1834509
* https://bugs.launchpad.net/kicad/+bug/1834509
(cherry picked from commit 80e0d848f2)
They're now used for bus definition control characters.
Also fixes the sheet pin edit dialog to correctly escape/unescape
netnames.
Fixes: lp:1840834
* https://bugs.launchpad.net/kicad/+bug/1840834
While it's tempting to look at the changed flag that TestDanglingEnds
returns, it's insuficient as the connectivity algorithm might have
already updated it (and ignored the changed flag).
Fixes: lp:1835107
* https://bugs.launchpad.net/kicad/+bug/1835107
Items that have BG fill need to be plotted in the background but we
still need their stroked edges to be in the FG. This is a workaround
that overplots the edges ensuring visual similarity. Re-working the
plot engine to support layers is a larger project.
Fixes: lp:1833428
* https://bugs.launchpad.net/kicad/+bug/1833428
(cherry picked from commit 6796e8fff6)
Control ordering is critical to make dialog mnemonics work correct. The
static text with the mnemonic must immediately proceed the control and
no additional controls can be added between static text/control pair.
Fixing this required adding the units static text controls after all of
the mnemonic static text controls and their associated edit controls.
Mnemonics do not work for read only controls so all of the combobox
controls were moved after the text controls.
Fixes lp:1818930
https://bugs.launchpad.net/kicad/+bug/1818930
1) don't ASSERT; use exceptions (which can be caught) instead.
2) clean up LIB_LOGGER so it's only used during sync
3) fix typos in wxLogMessage() calls so they produce output.
4) throw on unrecognized token in loadContent().
5) if clipboard can't be pasted as content then paste as text.
Fixes: lp:1840172
* https://bugs.launchpad.net/kicad/+bug/1840172
* Push a function into CONDITIONAL_MENU that adds the item
* Modify the tooltip for close and exit items to have the
program name
Fixes: lp:1835454
* https://bugs.launchpad.net/kicad/+bug/1835454
The use of wxPATH_UNIX when calling GetFullPath() on windows drops the
volume identifier (c:\) from the file name which will cause comparison
failures.
1) Add highlighting to simProbe and simTune tools.
2) Probe tool should accept wires and pins, not wires and components.
3) Give simulator window a flat look to match rest of Kicad.
4) Collapse whitespace out of simulator window.
5) Add some error messages.
6) Add some whitespace to simulation setting dialog.
When removing a junction, we need to keep the next pointer for deleted
item, marking them with a flag until committed.
Fixes: lp:1837398
* https://bugs.launchpad.net/kicad/+bug/1837398
(cherry picked from commit 4363e6a213)
When selecting items where the reference point is off-screen, GTK
would default the cursor to jump to the bottom of the screen.
Fixes: lp:1838843
* https://bugs.launchpad.net/kicad/+bug/1838843
While we do not generate these arcs internally, some external tools
generate arcs that are technically valid but display correctly only
because of an oddity in the degeneracy mapping.
Fixes: lp:1838557
* https://bugs.launchpad.net/kicad/+bug/1838557
(cherry picked from commit cf949609b2)
Check for symbol libraries that somehow ended up in the project file
after remapping occurred and remove all of them to prevent potentially
broken symbol library links.
Add nagware dialog to warn the user and a checkbox to never show the
dialog again.
Fixes lp:1838185
https://bugs.launchpad.net/kicad/+bug/1838185
Images have a scale factor and this factor was taken in account twice
For image having a scale factor not equal to 1.0, the bounding box was too big (or too small)
Remove an existing sheet from the current page that requires a schematic
load from file to prevent a corrupt screen list from causing a segfault.
Merge the edit sheet file loading with the append schematic code since
they are functionally the same. This allows the sheet edit code to
take advantage of all of the added broken symbol library links added in
the recent append schematic fixes.
Fixes lp:1835841
https://bugs.launchpad.net/kicad/+bug/1835841
Also rewrites the PCBNew Find dialog to make use of the above, including:
1) searching in user-defined footprint fields
2) searching in pcb text
3) a history list in the search popup
4) case sensitive searching
5) word sensitive searching
6) the ability to turn wildcard searching on/off
7) better placement of the result when the dialog obscures part of the
window
Fixes: lp:1838006
* https://bugs.launchpad.net/kicad/+bug/1838006
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
The point editor expects rectangles with a height and width > 0.
Previously, this constraint did not exist,
so some LIB_RECTANGLE graphic items can create issues when trying to move a corner.
Now, the LIB_RECTANGLE graphic item is normalized before trying to move a corner.
This fixes the last suffix that was missed by 168fa09f5. The schematic
backup files should have the suffix .sch-bak to distinguish from
.kicad_pcb-bak or other backups.
(cherry picked from commit b6fe4ff637)
Previously, for instance after modifying preferences or language,
they were rebuilt but the tools state was not set due to a missing call to SyncToolbars() after rebuilding.
Do not add cache library from another project to old library list. This
is wrong and can cause all kinds of symbol library linking issues. It's
better to force the user to fix broken symbol library links than the false
security of using multiple cache libraries.
Use correct paths when appending schematic outside of the current project
path when hierarchical sheets are appended. If the path can be relative,
give the user the option to use either relative or absolute paths.
Do not change symbol library names in schematics appended from a source
outside the current project. This will almost ensure the symbol library
links will be broken in the source project.
When the appended schematic is from another project, attempt to check all
of the possible combinations of symbol library table importing that could
cause broken symbol library links and give the user the option of canceling
the append operation or live with the possibility that there may be broken
symbol library links.
When the append schematic is in the current project path but not part of
the schematic, check to see if there are any library nicknames that do
no exist in the project symbol library table and give the user a chance
to cancel the append process.
This reverts commit eadf6d93bc.
because the issue with net names containing '/' is not actually inside the netlist.
This commit eadf creates issues with ngspice 2.8 and older, and do not fix anything in ngspice
Selecting in Eeschema has major issues, as does highlighting in
pcbnew. So for now we do highlighting in eeschema and selecting
in pcbnew. Improving highlighting for pcbnew would be the next
logical step....
Fixes: lp:1836640
* https://bugs.launchpad.net/kicad/+bug/1836640
wxLog::SetActiveTarget() was initialised to libedit frame,
but the active target was not released after closing and deleting libedit, thus creating a crash on Kicad exit.
The full filename comparison to detect the reuse of a file was broken on Windows
because it mixed a native filename notation (containing the drive id),
and a unix notation (no drive in filename, different separator).
Because the pickers are called with an Activate()/Wait() pair, any
usurping tool gets in between the picker handling the cancel (due
to activation) and the picker client doing a PopTool(). The new
tool is therefore pushed before the old tool is popped.
Fixes: lp:1835907
* https://bugs.launchpad.net/kicad/+bug/1835907
This changes make_lexer() so that it no longer generates a custom target
but instead attaches the generated files to an existing one (so the first
argument now is the name of an existing library or executable, and it needs
to come after the add_library/add_executable call).
The generated source is no longer listed in the project sources, as it is
added by the function. The files are generated in the build tree rather
than the source tree, and the directory is added to the include path for
the respective project as well as exported to projects linking against it.
Generated files in subdirectories are somewhat supported, but need to be
referenced with the same name as they were generated (i.e. including the
subdirectory name).
Fixes: lp:1831643
* https://bugs.launchpad.net/kicad/+bug/1831643
Fixes: lp:1832357
* https://bugs.launchpad.net/kicad/+bug/1832357
Fixes: lp:1833851
* https://bugs.launchpad.net/kicad/+bug/1833851
If Symbol Editor isn't open then Eeschema needs to read in the
default line width.
Symbol Editor was failing to write line width parameter back to
properties.
Fixes: lp:1835051
* https://bugs.launchpad.net/kicad/+bug/1835051
Previously, the background was highlighted, but not the text.
On Windows, the text was unreadable, until it was selected.
Now the foreground color is also set to the text selected color.
Replace also in lib_manager.cpp some DisplayErrorMessage calls by wxLogMessage calls,
to avoid hanging or crashes, at least on Windows.
wxLogMessage also has the advantage of grouping all messages and displaying them during a idle time.
1) Use PenWidth() (which is clipped at 0) instead of Width()
2) Implement an interior HitTest when a PolyLine is filled
3) Brighten background objects as well as foreground
4) Don't HitTest last segment when PolyLine is not filled
Fixes: lp:1834703
* https://bugs.launchpad.net/kicad/+bug/1834703
* Implement proper natural sorting of the reference values
* Perform quoting of all items in BOM generators that use csv.writer
Fixes: lp:1833822
* https://bugs.launchpad.net/kicad/+bug/1833822
This is based on the notion that in Kicad the start and end-points
are far more important (as they usually connect to other segments)
than the center is.
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)
Prevent the symbol and footprint library table editors from substituting
the current project path (${KIPRJMOD}) when adding libraries to the global
library table. This will almost assuredly fail when opening a different
project. We still allow users to manually use ${KIPRJMOD} in global table
paths under the assumption that if they are typing this in then they most
likely understand environment variable substitution.
Fixes lp:1830694
https://bugs.launchpad.net/kicad/+bug/1830694
In singletop mode, all frames show the "Quit" option in the file menu
and will quit on Ctrl-Q. When launched from the main KiCad interface,
sub-programs show the "Close" option instead and will close with Ctrl-W.
In this mode, Ctrl-Q will instruct the main program to exit.
Fixes: lp:1779938
* https://bugs.launchpad.net/kicad/+bug/1779938
EEschema optimizes wires by merging colinear segments. If a schematic opened without a valid
cache library or missing installed libraries and later saved, this optimization can cause connectivity
errors. In order to fix that we check each pin-wire connection and junctions if necessary.
Netlists do not accept any char in netnames (especially spice).
They must use only "/" as root sheet path name.
Especially _( "<root sheet>" ) breaks netlists because:
- there is a space in name, and special chars (< and >)
- it is a translatable name. so the actual name cannot be managed.
- most of netlist code in Kicad expects a "/" as root path.
This adds the functionality to add multiple labels to your schematic at
once by using spaces in the original label name.
Fixes: lp:1095113
* https://bugs.launchpad.net/kicad/+bug/1095113
Prefix the sheet name with "Sheet: " and the sheet file name with "File: "
when plotting so that it matches the display sheet rendering.
Improve the root sheet plot file name.
Fixes lp:1832081
https://bugs.launchpad.net/kicad/+bug/1832081
The grid apply code was always preventing empty fields. Change this the
broken behavior to allow any field to be empty except the reference and
value fields.
Fixes lp:1832722
https://bugs.launchpad.net/kicad/+bug/1832722
Also changes the presentation of the string when the checkbox appears
so that the "Changes will be lost" warning isn't hidden by the turndown.
This last part may require conditional compilation as it's undocumented.
Lastly, regularizes the Unsaved Changes strings between the apps.
Drawing junction dots at the line width allows users who do not like
them in their schematics to avoid having them drawn.
(cherry picked from commit 75b5324762)
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
DIALOG_EXIT was not being laid out correctly after setting the dialog
message. Rather than fix our custom DIALOG_EXIT, wxMessageDialog and
wxRichMessageDialog were substituted as direct replacements for all
of the variants of DIALOG_EXIT. This make message dialogs appear more
consistent because wxMessageDialog uses the default platform message
dialog and wxRichMessageDialog uses the default platform rich message
dialog on windows.
Remove DIALOG_EXIT_BASE as it is no longer required.
Fixes lp:1832899
https://bugs.launchpad.net/kicad/+bug/1832899
If a tool called something like clearSelection while processing a
MOUSE_CLICK, the SELECTION_TOOL will pass the clearSelection
COMMAND_EVENT because it handles it as a transition, not as an
event. Because m_passEvent is effectively global, the tool manager
would then interpret that as passing the MOUSE_CLICK and we'd end
up processing the click by multiple tools.
A bug in LIB_MANAGER::LibraryExists() prevented a library removed from
the symbol library table from being removed from the tree view. Add the
proper check to SYMBOL_TREE_SYNCHRONIZING_ADAPTER::Sync().
This exposed another issue with synchronization between the library
table editor and the tree view which could lead to orphaned library
modifications and/or a segfault when the currently selected library was
removed from the symbol library table. Give the user a chance to save
or revert any changes before allowing changes to the symbol library
table.
Fixes lp:1821691
https://bugs.launchpad.net/kicad/+bug/1821691
These options were used to support the legacy canvas when drawing on
MacOS and GTK3. With the move to 100% GAL, they are extraneous. This
moves all DC over to "COPY" as we only use this for printing support in
Eeschema at the moment, so there is a single draw command (no erasing)
for the canvas.
There are a whole bunch of cases which all have to work:
regular move
move followed by drag
drag followed by move
move followed by duplicate
duplicate
paste
and all both with and without Warp Mouse on Move checked.
Return the EDA_DRAW_FRAME class back to common, whence it came (before it was
duplicated into legacy_wx and legacy_gal). Now there is only one
implementation (the GAL one), it doesn't need to be in a separate library.
This simplifies the dependencies for common lib users.
Since Eagle does not connect stacked NC pins but KiCad will if the pins
are coincident, we choose to only import the first of the stacked NC
pins. While this creates symbols that are slightly different in pin
count from the Eagle version, it keeps the schematic and netlist
functionally correct relative to the Eagle version.
Fixes: lp:1821319
* https://bugs.launchpad.net/kicad/+bug/1821319
(cherry picked from commit 370109b868)
Net names in eeschema that include '/' are escaped ('/' replaced by "{slash}")
Escaped netnames are only for internal use, and must be unescaped in spice netlist and dialogs.
It was escaped (replaced by "{slash}") in dialog.
This is right for labels, not for texts.
A consequence was the directives in spice simul can be broken.
Now the graphic text is not escaped.
Building custom targets that depend on custom commands such as our
keyword lexer requires two layers of indirection to pick up changes
properly and not overwrite the same file in parallel builds.
Fixes: lp:1831643
* https://bugs.launchpad.net/kicad/+bug/1831643
When loading a new file, we do not want to add junctions where they did
not exist before. This could generate invalid netlists by breaking a
wire at a pin connection point.
We do break/add connections when the user modifies the schematic but
this happens while the user is actively working with the component and
sees the action.
Fixes: lp:1830411
* https://bugs.launchpad.net/kicad/+bug/1830411
Also fixes the post-duplicate move to start at the cursor pos.
Also fixes the fact that duplicate wasn't getting its hotkey.
Fixes: lp:1830596
* https://bugs.launchpad.net/kicad/+bug/1830596
This fixes multiple issues when importing Eagle V6 files.
Crashes occured when arcs of 0-length were found and when parts were
referenced in the schematic that were not found in the library. This
could happen if the library and schematic were different cases.
Fixes: lp:1830564
* https://bugs.launchpad.net/kicad/+bug/1830564
This only applies to the root symbol when the datasheet field is not
empty and the datasheet for root symbol alias in the document file is
empty. Use the root symbol datasheet field value when adding a new
symbol that meets this criteria.
The root symbol datasheet field gets copied to the root symbol alias
datasheet member variable so if the symbol is modified and saved, the
datasheet information is saved in the .dcm file.
https://bugs.launchpad.net/kicad/+bug/1830078
Fixes lp:1830078
This only applies to the root symbol when the datasheet field is not
empty and the datasheet for root symbol alias in the document file is
empty. Use the root symbol datasheet field value when adding a new
symbol that meets this criteria.
https://bugs.launchpad.net/kicad/+bug/1830083
Fixes lp:1830083
This is done to allow access to the eeschema library
internals for purposes of test and script access, as the
DLL library has highly restrictive -fvisibility settings
that otherwise prevent the tests being able to access 99.9%
of the eeschema library functions (only a single function
is APIEXPORT'ed, therefore that's the only test we can do).
Using object libraries is a bit of a hack, and makes for
a slower link when done for multiple targets, but with the currently
supported CMake versions, it's about as good as we can get.
A better solution in the longer term may be to break eeschema_kiface(_objects)
into many smaller libraries, each of which has a much more defined scope,
rather than one big interlinked amorphous lump. This has the advantage that
each module is testable in isolation, and we get better organisation of
inter-dependencies in the codebase.
Then, the kiface DLL will gather these sub-libs and present what
is needed on the visible DLL API. Thus, we get both a testable
suite of library functions, and a restricted kiface DLL interface.
If you construct wxTimer statically, it will be constructed
before WX init, which means you might get assertions about
missing m_impl's within the timer class.
The solution is to construct the wxTimer at run time. In this
case, static within the function will be constructed only after
the function is called (and actually the scope only needed to be
this function anyway).
After importing Eagle files, we need to make sure that the connectivity
network is updated before the system requests a new netlist. Otherwise,
we end up with different results from the first to second updates.
This commit redirected too many menuitem IDs to the tool manager.
Only some context menu items have to be redirected, not all inside ACTION_EVENT events.
This hack is not perfect. It needs a better handling of ACTION_EVENT handling.
Add the N/Shift+N hotkeys to the eeschema hotkey file.
Also adjust the COMMON_TOOLS::GridNext/Prev to work when
the screen's grid definitions do not have consecutive command
IDs (they do not in eeschema).
This is related to lp:1811018 (but doesn't constitute a fix).
* https://bugs.launchpad.net/kicad/+bug/1811018
This is a stop gap fix. It should be possible to use the wxFrame title
string for this but some of the frame windows have appropriated the title
string for configuration file naming purposes which is less than ideal.
Fixes lp:1825977
https://bugs.launchpad.net/kicad/+bug/1825977
Otherwise we set item from nullptr back to the selection, and then
we think we're dragging again which causes all kinds of grief.
Fixes: lp:1828067
* https://bugs.launchpad.net/kicad/+bug/1828067
Prevent user from pasting text with carriage return and/or line feed
characters in field value edit control in the schematic symbol and symbol
library properties dialogs.
Make all objects derived from GRID_CELL_TEXT_BUTTON use validators
correctly.
Add validators to FIELDS_GRID_TABLE object for field values other than
the mandatory fields.
Create a validator to check that a string is a valid LIB_ID.
Fixes lp:1828214
https://bugs.launchpad.net/kicad/+bug/1828214
When activating the move from a drag extension, we need to add
additional components to the commit. This breaks for duplicate/copy
when we have already added the correct components
Also reformatted the file to make it easier to see what is there and
what is missing. Apologies to the 100-char-width folks, but we were
just missing to many command IDs in this file.
Fixes: lp:1828006
* https://bugs.launchpad.net/kicad/+bug/1828006
If you have, for instance a pin selected and do a cmd-E we'll filter
out the pin and then do nothing. We should instead treat it as a
hover selection if everything got filtered.
When, inside a complex hierarchy, one switch from an instance of a sheet to an other instance of this sheet,
the page number shown in the worksheet was not updated on screen, and the initial number was displayed.
Fixes: lp:1827981
https://bugs.launchpad.net/kicad/+bug/1827981
It was confusing that the primary frames registered their tools
differently than the other frames. In addition, since the other
frames also added their own tools, foo_actions::RegisterAllTools()
didn't really register all tool but rather those used by the
principal frame (PCB_EDIT_FRAME, SCH_EDIT_FRAME, etc.)
Don't edit sheet during AddToScreenAndUndoList() call. If it's
cancelled and we delete the item, callers will still own pointers
to the freed memory. Do it in New and Paste instead.
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
When Pcbnew's add footprint browser is run up, it constructs
the recently used list without checking for nulls.
Although the DoAddLibrary call does internally check for
null, it's easier to filter these out before placing into
the vector in the first place.
The same logic in the symbol tree is already handled in the
same way.
Currently, this widget doesn't read the common library settings.
This means it doesn't use the configured DPI scaling or anti-aliasing
settings.
Use the ReadConfig method on the GAL options to fix (rather than
only ReadAppConfig).
Fixes: lp:1825913
* https://bugs.launchpad.net/kicad/+bug/1825913
These dialogs had logic to avoid a table re-adjustment unless
the width changed. This was done to avoid spurious resizing
calls under GTK+3. This was commit
13249b723b, fixing bug lp:1817810.
However, by only calling event.Skip() when the width changed,
redraws were inhibited when only the height changed.
Placing the Skip() outside the width-change check fixes this,
and does not re-introduce the lp:1817810 bug (it is the column
adjust call that causes that).
Fixes: lp:1826615
* https://bugs.launchpad.net/kicad/+bug/1826615
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.
Jeff Young
Jon Evans
Seth Hillbrand
jean-pierre charras
Ian McInerney
Sylwester Kocjan
Wayne Stambaugh
Karl Zeilhofer
Jonathan Warner
Seth Hillbrand
Alexander Shuklin
Simon Richter
Mark Roszko
Jonatan Liljedahl
Franck Jullien
Jon Evans
Adam Wolf
Michael Kavanagh
Maciej Suminski
Tomasz Włostowski
Steven A. Falco
Seppe Stas
John Beard