The original algorithm picked the value from the last component having
non empty value for a given field, but the processing order was
dependent on the layout of the components in the memory. It means that
for each component, the field values could have been taken from any
unit, randomly.
The patch improves the algorithm, trying to get all values from the unit
with the lowest number and resorts to other units only when there are
field values left empty.
Values generated by NETLIST_EXPORTER_PSPICE::GetSpiceFieldDefVal()
and the actual value that was stored in a netlist differed when a
component has multiple units.
Due to that, the Spice model editor dialog incorrectly recognized
"alternate node sequence" value as the default one and cleared it
when the dialog was closed.
Code fixing Spice device names (prefixing reference with a character
corresponding to the assigned device model type) that was duplicated in
a few places has been moved to a function (NETLIST_EXPORTER_PSPICE::GetSpiceDevice()).
For some reason, when calling python and giving a full filename script, the last separator in the filename
must be '/', not '\'.
Otherwise the import command inside the script does not find files to import in the same folder as the script.
We cannot replace blindly '\' to '/' in command line because it does not work for file on a server (name starting by \\server_name\).
So the fix is just replacing one '\' in python script full filename.
This is not perfect, but at least it works for newly created plugin commands.
Spice netlist exporter processes all texts on the exported schematic
sheet to find spice directives and include them in the output file.
By default it also added ".title KiCad schematic" in the first line, so
if there was another .title directive in the exported schematic sheet,
the generated file would contain two .title directives.
This patch looks for .title directive and when one is found - it uses
the specified title in the first line.
This is the last of the object save/load code that was not moved into
the SCH_LEGACY_PLUGIN object. All schematic and library I/O is now
performed in the SCH_LEGACY_PLUGIN object and as been removed from the
schematic and library objects.
The old single symbol file format has been replaced with the normal
symbol library file format since there was no difference between them
except the SYMBOL token. The SYMBOL token was no longer being read
since the introduction of the SCH_LEGACY_PLUGIN symbol library loader.
Update the Doxygen comments in all of the modified files.
SEARCH_STACK is a deprecated method for getting the list of paths where
one could look for a file. Instead it tries the project path and
environmental variables.
Simplify some of the library editing code. There have been a few minor
changes in the behavior of the editor. If the current symbol is deleted
from the library, the next symbol in the library is not loaded. The
deleted symbol is cleared and the current symbol is empty.
Change component to symbol to align with the preferred terminology
discussed on the developer's mailing list.
Add separate update UI event for save library as to enable the menu entry
whenever a library is selected.
Change the select symbol list dialog to a single column using the LIB_ID
format LIB_NICKNAME:LIB_ITEM_NAME so that the selection can be parsed by
LIB_ID.
Add method to expand URI to LIB_TABLE_BASE.
Override wxApp::OnExceptionInMainLoop() in debug builds to make debugging
easier when an unhandled exception occurs in a wxUpdateUIEvent handler.
Change SCH_SCREENS::HasNoFullyDefinedLibIds() to return false when the
schematic has no symbols to prevent the remapping dialog from being run.
Use SCH_COMPONENT part reference when creating netlist rather than looking
up the library symbol.
(On windows, xsltproc.exe does not accept '\' in output filename, only '/' like on Unix, so this separator is converted when possible)
Minor code cleanup
- pad names are stored as wxString instead of a char[4] & integer union
- removed pad name to string conversion functions
- fixed pad & pin properties dialog restrictions regarding the name
length
This commit was too broad and not cognizant of the purpose of the class
UTF8.
Add MAYBE_VERIFY_UTF8() macro, which can trap non-UTF8 encoded strings in
debug builds.
Use that macro conditionally in class UTF8 to trap non-UTF8 encoded strings
being put into UTF8 instances.
Sadly, each unit of a component can have its own unique fields. This
change finds the last non blank field and records it. Last guy wins
and the order of units occuring in a schematic hierarchy is variable.
Therefore user is best off setting fields into only one unit. But this
scavenger algorithm will find any non blank fields in all units and use
the last non-blank field for each unique field name.
Fixes lp:1471417
https://bugs.launchpad.net/kicad/+bug/1471417
Allows to map nodes to pins for Spice models that have more pins than
corresponding schematic symbol. An example is a MOSFET model containing
gate, source, drain & body, whereas schematic symbol part has only gate,
source & drain with body and source connected inside the part.
Use LIB_ID instead of wxString for storing the library symbol information
in the schematic symbol in preparation for the upcoming symbol library table
implementation.
Change the FindLibAlias and FindLibPart functions in the PART_LIBS object
instead of wxString. Please note that only the library ID name is used to
search the list of libraries. The library nickname is ignored. Once the
symbol library table is implemented and full LIB_IDs are defined, the
library search code will no longer be used and will only be kept to load
older schematics that have not been converted.
Move SCH_LEGACY_PLUGIN_CACHE definition so that the legacy plugin knows
how to properly delete the cache object.
netlist generation: remove option "replace IC and U ref by X" (broken option).
All component refs are now prefixed by the Spice_Primitive field value.
Therefore the netlist used by Kicad ngsipce simulator and the spice netlist generated from dialog are the same.
(note: like previously, forbidden chars like ( and ) are replaced by _ in netnames)
This commit creates more issues than resolves. It could happen that
there are components with different reference types (e.g. U1 and IC1)
that would be later converted to X1, causing a conflict.
Jeff Young
Maciej Suminski
Wayne Stambaugh
jean-pierre charras
ludovic léau-mercier
Camille
Simon Richter
Dick Hollenbeck
Chris Pavlina
Martin Stoilov