This adds an io-benchmark case of the STRING_LINE_READER
class, which reads a file into a std::string, *then*
reads it line by line.
As expected, due to it all being in memory, this is very fast.
Also fixes an issue in io_benchmark where the input file
must be in the current dir.
This adds a test program which can be used to test the
parsing of a given KiCad PCB file. This interface is
useful for both manual or automated debugging of given
files, as well as providing an interface suitable for
fuzz-testing tools.
Also adds to the testing docs to detail how fuzzing can
be used.
Also moves some useful re-usable code from io-benchmark
to a new library qa_utils, which can contain code that
isn't need in the actual KiCad libs.
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)
This forces the compiler class specific features rather than borrowing
from the base class's std::string. In some cases prior to this,
wxString( std::string ) was being called rather than UTF8::operator
wxString() leading to garbled wxStrings.
Added function UTF8::wx_str() which is of great convenience also.
Implicit conversions still work as before, and hopefully more reliably.
This appears to normally use the move constructor, but on Jenkins Fedora
20, it tried to use the copy ctor, which is deleted.
Constructing directly fixes this, and is the right way, anyway.
This compares the performance of RICHIO line readers against other
implementations, and can be used for profiling and optimisation.
Current benchmarks provided:
* richio FILE_LINE_READER
* Raw std::ifstream (no LINE_READER wrapper), using getline (so no line
length limiting)
* LINE_READER wrapper around std::istream, with a std::ifstream
implementation
* Existing richio wxInputStream wrappers (with File and FFile
implemntations)
* Move add_directory()s in main CMakeList.txt after all find_package()
macros are run so all flags in the main CMakeList.txt are propagated
to the sub-folder CMakeList.txt files.
* Remove commented out include of config.h
* Make the link maps only build on linux as -${TO_LINKER},-cref
-${TO_LINKER},-Map=blah just gives warnings on osx w/ clang
* Make the link maps only build with flag -DKICAD_MAKE_LINK_MAPS is
defined during CMake configuration as they are highly specific.
* Moved the code for setting link maps into the main CMakeList.txt
file to avoid duplication.
* Removed -D__ASSERTMACRO__ from osx as its no longer needed
* Removed a couple of other OSX only things that wouldn't work anyway
* Moved set (BU_CHMOD_BUNDLE_ITEMS) to the main CMakeList.txt as
otherwise it would only work for the KiCad build not the other
applications
* Made KICAD_SCRIPTING_MODULES enable KICAD_SCRIPTING as currently if
you have modules enabled without the scripting base it will on build.
This could be changed to a fatal error saying you need to also enable
scripting but it seems unnecessary.
* Removed duplication of pcbnew.py install code under modules and
scripting since you can't have modules without scripting now
was used locally. Then comment it out in favor of a newer strategy for
filling in nicknames in cvpcb.
Add MODULE* FootprintLoadWithOptionalNickname( const FPID& aFootprintId )
throw( IO_ERROR, PARSE_ERROR );
from code found elsewhere.
! The initial testing of this commit should be done using a Debug build so that
all the wxASSERT()s are enabled. Also, be sure and keep enabled the
USE_KIWAY_DLLs option. The tree won't likely build without it. Turning it
off is senseless anyways. If you want stable code, go back to a prior version,
the one tagged with "stable".
* Relocate all functionality out of the wxApp derivative into more finely
targeted purposes:
a) DLL/DSO specific
b) PROJECT specific
c) EXE or process specific
d) configuration file specific data
e) configuration file manipulations functions.
All of this functionality was blended into an extremely large wxApp derivative
and that was incompatible with the desire to support multiple concurrently
loaded DLL/DSO's ("KIFACE")s and multiple concurrently open projects.
An amazing amount of organization come from simply sorting each bit of
functionality into the proper box.
* Switch to wxConfigBase from wxConfig everywhere except instantiation.
* Add classes KIWAY, KIFACE, KIFACE_I, SEARCH_STACK, PGM_BASE, PGM_KICAD,
PGM_SINGLE_TOP,
* Remove "Return" prefix on many function names.
* Remove obvious comments from CMakeLists.txt files, and from else() and endif()s.
* Fix building boost for use in a DSO on linux.
* Remove some of the assumptions in the CMakeLists.txt files that windows had
to be the host platform when building windows binaries.
* Reduce the number of wxStrings being constructed at program load time via
static construction.
* Pass wxConfigBase* to all SaveSettings() and LoadSettings() functions so that
these functions are useful even when the wxConfigBase comes from another
source, as is the case in the KICAD_MANAGER_FRAME.
* Move the setting of the KIPRJMOD environment variable into class PROJECT,
so that it can be moved into a project variable soon, and out of FP_LIB_TABLE.
* Add the KIWAY_PLAYER which is associated with a particular PROJECT, and all
its child wxFrames and wxDialogs now have a Kiway() member function which
returns a KIWAY& that that window tree branch is in support of. This is like
wxWindows DNA in that child windows get this member with proper value at time
of construction.
* Anticipate some of the needs for milestones B) and C) and make code
adjustments now in an effort to reduce work in those milestones.
* No testing has been done for python scripting, since milestone C) has that
being largely reworked and re-thought-out.
To use multiple working threads. This entailed adding KiCad typedefs:
*) Add typedefs for MUTEX and MUTLOCK which mask the actual choices for the project.
*) Add FOOTPRINT_LIST::DisplayErrors( wxWindow* ) which is a single strategy for
showing aggregated load errors. Although what's there is only scaffolding
and needs a volunteer who knows HTML pretty well.
*) Ensure all callers of ReadFootprintFiles() use the new DisplayErrors() function.
*) Push utf8.cpp and utf8.h into common library for open use.
bit string version of property_tree. The ram resident structure of the ptree is
mostly compatible with one created using the xml_parser from
boost::property_tree, with slight differences in the way atoms are stored. The
result is you can use Format() to convert from xml to s-expression, but not the
other way around. You can write a simple s-expression beautifier in just a few
lines of code.
The main value however is the s-expression parser, i.e. Scan(), which is an
alternative to crafting a custom recursive descent parser for a particular
grammar. The tipping point depends on whether you want to read only a small
portion of a much larger document. If so, then using the ptree will likely be a
"faster to code" route. Documentation on how to navigate a ptree can be found on
the boost website and there are a number of examples in the
pcbnew/eagle_plugin.cpp file in this project. Powerful path navigation support
makes it easy to extract a subset of a ptree.
New classes:
- VIEW - represents view that is seen by user, takes care of layer ordering & visibility and how it is displayed (which location, how much zoomed, etc.)
- VIEW_ITEM - Base class for every item that can be displayed on VIEW (the biggest change is that now it may be necessary to override ViewBBox & ViewGetLayers method for derived classes).
- EDA_DRAW_PANEL_GAL - Inherits after EDA_DRAW_PANEL, displays VIEW output, right now it is not editable (in opposite to usual EDA_DRAW_PANEL).
- GAL/OPENGL_GAL/CAIRO_GAL - Base Graphics Abstraction Layer class + two different flavours (Cairo is not fully supported yet), that offers methods to draw primitives using different libraries.
- WX_VIEW_CONTROLS - Controller for VIEW, handles user events, allows zooming, panning, etc.
- PAINTER/PCB_PAINTER - Classes that uses GAL interface to draw items (as you may have already guessed - PCB_PAINTER is a class for drawing PCB specific object, PAINTER is an abstract class). Its methods are invoked by VIEW, when an item has to be drawn. To display a new type of item - you need to implement draw(ITEM_TYPE*) method that draws it using GAL methods.
- STROKE_FONT - Implements stroke font drawing using GAL methods.
Most important changes to Kicad original code:
* EDA_ITEM now inherits from VIEW_ITEM, which is a base class for all drawable objects.
* EDA_DRAW_FRAME contains both usual EDA_DRAW_PANEL and new EDA_DRAW_PANEL_GAL, that can be switched anytime.
* There are some new layers for displaying multilayer pads, vias & pads holes (these are not shown yet on the right sidebar in pcbnew)
* Display order of layers is different than in previous versions (if you are curious - you may check m_galLayerOrder@pcbnew/basepcbframe.cpp). Preserving usual order would result in not very natural display, such as showing silkscreen texts on the bottom.
* Introduced new hotkey (Alt+F12) and new menu option (View->Switch canvas) for switching canvas during runtime.
* Some of classes (mostly derived from BOARD_ITEM) now includes ViewBBox & ViewGetLayers methods.
* Removed tools/class_painter.h, as now it is extended and included in source code.
Build changes:
* GAL-based rendering option is turned on by a new compilation CMake option KICAD_GAL.
* When compiling with CMake option KICAD_GAL=ON, GLEW and Cairo libraries are required.
* GAL-related code is compiled into a static library (common/libgal).
* Build with KICAD_GAL=OFF should not need any new libraries and should come out as a standard version of Kicad
Currently most of items in pcbnew can be displayed using OpenGL (to be done are DIMENSIONS and MARKERS).
More details about GAL can be found in: http://www.ohwr.org/attachments/1884/view-spec.pdf
John Beard
Seth Hillbrand
Simon Richter
jean-pierre charras
Dick Hollenbeck
Kristoffer Ödmark
Simon Wells
Marco Serantoni
Maciej Suminski
Miguel Angel Ajo