Contact: | info@riverbankcomputing.co.uk |
---|---|
Version: | 4.6 |
Copyright: | Copyright (c) 2007 Riverbank Computing Limited |
This is the reference guide for SIP 4.6. SIP is a tool for automatically generating Python bindings for C and C++ libraries. SIP was originally developed in 1998 for PyQt - the Python bindings for the Qt GUI toolkit - but is suitable for generating bindings for any C or C++ library.
This version of SIP generates bindings for Python v2.3 or later.
There are many other similar tools available. One of the original such tools is SWIG and, in fact, SIP is so called because it started out as a small SWIG. Unlike SWIG, SIP is specifically designed for bringing together Python and C/C++ and goes to great lengths to make the integration as tight as possible.
The homepage for SIP is http://www.riverbankcomputing.co.uk/sip/. Here you will always find the latest stable version, current development snapshots, and the latest version of this documentation.
SIP is licensed under the same terms as Python itself. SIP places no restrictions on the license you may apply to the bindings you create.
SIP, and the bindings it produces, have the following features.
- bindings are fast to load and minimise memory consumption especially when only a small sub-set of a large library is being used
- automatic conversion between standard Python and C/C++ data types
- overloading of functions and methods with different argument signatures
- access to a C++ class's protected methods
- the ability to define a Python class that is a sub-class of a C++ class, including abstract C++ classes
- Python sub-classes can implement the __dtor__(self) method which will be called from the C++ class's virtual destructor
- support for ordinary C++ functions, class methods, static class methods, virtual class methods and abstract class methods
- the ability to re-implement C++ virtual and abstract methods in Python
- support for global and class variables
- support for global and class operators
- support for C++ namespaces
- support for C++ templates
- support for C++ exceptions and wrapping them as Python exceptions
- the ability to define mappings between C++ classes and similar Python data types that are automatically invoked
- the ability to automatically exploit any available run time type information to ensure that the class of a Python instance object matches the class of the corresponding C++ instance
- full support of the Python global interpreter lock, including the ability to specify that a C++ function of method may block, therefore allowing the lock to be released and other Python threads to run
- support for the concept of ownership of a C++ instance (i.e. what part of the code is responsible for calling the instance's destructor) and how the ownership may change during the execution of an application
- the ability to generate bindings for a C++ class library that itself is built on another C++ class library which also has had bindings generated so that the different bindings integrate and share code properly
- a sophisticated versioning system that allows the full lifetime of a C++ class library, including any platform specific or optional features, to be described in a single set of specification files
- the ability to include documentation in the specification files which can be extracted and subsequently processed by external tools
- the ability to include copyright notices and licensing information in the specification files that is automatically included in all generated source code
- a build system, written in Python, that you can extend to configure, compile and install your own bindings without worrying about platform specific issues
- support for building your extensions using distutils
- SIP, and the bindings it produces, runs under UNIX, Linux, Windows and MacOS/X
SIP comprises a number of different components.
- The SIP code generator (sip or sip.exe). This processes .sip specification files and generates C or C++ bindings. It is covered in detail in Using SIP.
- The SIP header file (sip.h). This contains definitions and data structures needed by the generated C and C++ code.
- The SIP module (sip.so or sip.pyd). This is a Python extension module that is imported automatically by SIP generated bindings and provides them with some common utility functions. See also Using the SIP Module in Applications.
- The SIP build system (sipconfig.py). This is a pure Python module that is created when SIP is configured and encapsulates all the necessary information about your system including relevant directory names, compiler and linker flags, and version numbers. It also includes several Python classes and functions which help you write configuration scripts for your own bindings. It is covered in detail in The SIP Build System.
- The SIP distutils extension (sipdistutils.py). This is a distutils extension that can be used to build your extension modules using distutils and is an alternative to writing configuration scripts with the SIP build system. This can be as simple as adding your .sip files to the list of files needed to build the extension module. It is covered in detail in Building Your Extension with distutils.
SIP has specific support for the creation of bindings based on Trolltech's Qt toolkit.
The SIP code generator understands the signal/slot type safe callback mechanism that Qt uses to connect objects together. This allows applications to define new Python signals, and allows any Python callable object to be used as a slot.
SIP itself does not require Qt to be installed.
- The SIP_BUILD C preprocessor symbol has been removed.
- sipConvertToCpp(), sipIsSubClassInstance() and the old Generated Type Convertors have been deprecated. The functions sipCanConvertToInstance(), sipConvertToInstance(), sipForceConvertToInstance(), sipConvertFromInstance(), sipConvertFromNewInstance(), sipCanConvertToMappedType(), sipConvertToMappedType(), sipForceConvertToMappedType() and sipConvertFromMappedType() should be used instead. Handwritten %ConvertFromTypeCode and %ConvertToTypeCode now has the responsibility for using these to implement the Transfer and TransferBack annotations.
You can get the latest release of the SIP source code from http://www.riverbankcomputing.co.uk/sip/download.php.
SIP is also included with all of the major Linux distributions. However, it may be a version or two out of date.
You may also find more up to date pre-compiled binaries on SourceForge.
After unpacking the source package (either a .tar.gz or a .zip file depending on your platform) you should then check for any README files that relate to your platform.
Next you need to configure SIP by executing the configure.py script. For example:
python configure.py
This assumes that the Python interpreter is on your path. Something like the following may be appropriate on Windows:
c:\python25\python configure.py
If you have multiple versions of Python installed then make sure you use the interpreter for which you wish SIP to generate bindings for.
The full set of command line options is:
-h | Display a help message. |
-a | Export all symbols in any SIP generated module and the SIP module itself. This was the default behaviour of SIP prior to v4.2. Normally only a module's inititialisation function is exported. This option is deprecated as the ModuleMakefile class of The SIP Build System allows this to be specified on a per module basis. |
-b dir | The SIP code generator will be installed in the directory dir. |
-d dir | The SIP module will be installed in the directory dir. |
-e dir | The SIP header file will be installed in the directory dir. |
-k | The SIP module will be built as a static library. This is useful when building the SIP module as a Python builtin (see Builtin Modules and Custom Interpreters). |
-n | The SIP code generator and module will be built as universal binaries under MacOS/X. |
-p plat | Explicitly specify the platform/compiler to be used by the build system, otherwise a platform specific default will be used. The -h option will display all the supported platform/compilers and the default. |
-u | The SIP module will be built with debugging symbols. |
-v dir | By default .sip files will be installed in the directory dir. |
The configure.py script takes many other options that allows the build system to be finely tuned. These are of the form name=value or name+=value. The -h option will display each supported name, although not all are applicable to all platforms.
The name=value form means that value will replace the existing value of name.
The name+=value form means that value will be appended to the existing value of name.
For example, the following will disable support for C++ exceptions (and so reduce the size of module binaries) when used with GCC:
python configure.py CXXFLAGS+=-fno-exceptions
A pure Python module called sipconfig.py is generated by configure.py. This defines each name and its corresponding value. Looking at it will give you a good idea of how the build system uses the different options. It is covered in detail in The SIP Build System.
SIP, and the modules it generates, can be built with MinGW, the Windows port of GCC. You must use the -p command line option to specify the correct platform. For example:
c:\python25\python configure.py -p win32-g++
SIP, and the modules it generates, can be built with the free Borland C++ compiler. You must use the -p command line option to specify the correct platform. For example:
c:\python25\python configure.py -p win32-borland
You must also make sure you have a Borland-compatible version of the Python library. If you are using the standard Python distribution (built using the Microsoft compiler) then you must convert the format of the Python library. For example:
coff2omf python25.lib python25_bcpp.lib
The next step is to build SIP by running your platform's make command. For example:
make
The final step is to install SIP by running the following command:
make install
(Depending on your system you may require root or administrator privileges.)
This will install the various SIP components.
Bindings are generated by the SIP code generator from a number of specification files, typically with a .sip extension. Specification files look very similar to C and C++ header files, but often with additional information (in the form of a directive or an annotation) and code so that the bindings generated can be finely tuned.
We start with a simple example. Let's say you have a (fictional) C++ library that implements a single class called Word. The class has one constructor that takes a \0 terminated character string as its single argument. The class has one method called reverse() which takes no arguments and returns a \0 terminated character string. The interface to the class is defined in a header file called word.h which might look something like this:
// Define the interface to the word library. class Word { const char *the_word; public: Word(const char *w); char *reverse() const; };
The corresponding SIP specification file would then look something like this:
// Define the SIP wrapper to the word library. %Module word 0 class Word { %TypeHeaderCode #include <word.h> %End public: Word(const char *w); char *reverse() const; };
Obviously a SIP specification file looks very much like a C++ (or C) header file, but SIP does not include a full C++ parser. Let's look at the differences between the two files.
- The %Module directive has been added [1]. This is used to name the Python module that is being created and to give it a generation number. In this example these are word and 0 respectively. The generation number is effectively the version number of the module.
- The %TypeHeaderCode directive has been added. The text between this and the following %End directive is included literally in the code that SIP generates. Normally it is used, as in this case, to #include the corresponding C++ (or C) header file [2].
- The declaration of the private variable this_word has been removed. SIP does not support access to either private or protected instance variables.
If we want to we can now generate the C++ code in the current directory by running the following command:
sip -c . word.sip
However, that still leaves us with the task of compiling the generated code and linking it against all the necessary libraries. It's much easier to use the SIP build system to do the whole thing.
Using the SIP build system is simply a matter of writing a small Python script. In this simple example we will assume that the word library we are wrapping and it's header file are installed in standard system locations and will be found by the compiler and linker without having to specify any additional flags. In a more realistic example your Python script may take command line options, or search a set of directories to deal with different configurations and installations.
This is the simplest script (conventionally called configure.py):
import os import sipconfig # The name of the SIP build file generated by SIP and used by the build # system. build_file = "word.sbf" # Get the SIP configuration information. config = sipconfig.Configuration() # Run SIP to generate the code. os.system(" ".join([config.sip_bin, "-c", ".", "-b", build_file, "word.sip"])) # Create the Makefile. makefile = sipconfig.SIPModuleMakefile(config, build_file) # Add the library we are wrapping. The name doesn't include any platform # specific prefixes or extensions (e.g. the "lib" prefix on UNIX, or the # ".dll" extension on Windows). makefile.extra_libs = ["word"] # Generate the Makefile itself. makefile.generate()
Hopefully this script is self-documenting. The key parts are the Configuration and SIPModuleMakefile classes. The build system contains other Makefile classes, for example to build programs or to call other Makefiles in sub-directories.
After running the script (using the Python interpreter the extension module is being created for) the generated C++ code and Makefile will be in the current directory.
To compile and install the extension module, just run the following commands [3]:
make make install
That's all there is to it.
See Building Your Extension with distutils for an example of how to build this example using distutils.
[1] | All SIP directives start with a % as the first non-whitespace character of a line. |
[2] | SIP includes many code directives like this. They differ in where the supplied code is placed by SIP in the generated code. |
[3] | On Windows you might run nmake or mingw32-make instead. |
Let's now look at a very similar example of wrapping a fictional C library:
/* Define the interface to the word library. */ struct Word { const char *the_word; }; struct Word *create_word(const char *w); char *reverse(struct Word *word);
The corresponding SIP specification file would then look something like this:
/* Define the SIP wrapper to the word library. */ %CModule word 0 struct Word { %TypeHeaderCode #include <word.h> %End const char *the_word; }; struct Word *create_word(const char *w) /Factory/; char *reverse(struct Word *word);
Again, let's look at the differences between the two files.
- The %CModule directive has been added. This has the same syntax as the %Module directive used in the previous example but tells SIP that the library being wrapped is implemented in C rather than C++.
- The %TypeHeaderCode directive has been added.
- The Factory annotation has been added to the create_word() function. This tells SIP that a newly created structure is being returned and it is owned by Python.
The configure.py build system script described in the previous example can be used for this example without change.
In this last example we will wrap a fictional C++ library that contains a class that is derived from a Qt class. This will demonstrate how SIP allows a class hierarchy to be split across multiple Python extension modules, and will introduce SIP's versioning system.
The library contains a single C++ class called Hello which is derived from Qt's QLabel class. It behaves just like QLabel except that the text in the label is hard coded to be Hello World. To make the example more interesting we'll also say that the library only supports Qt v3.0 and later, and also includes a function called setDefault() that is not implemented in the Windows version of the library.
The hello.h header file looks something like this:
// Define the interface to the hello library. #include <qlabel.h> #include <qwidget.h> #include <qstring.h> class Hello : public QLabel { // This is needed by the Qt Meta-Object Compiler. Q_OBJECT public: Hello(QWidget *parent, const char *name = 0, WFlags f = 0); private: // Prevent instances from being copied. Hello(const Hello &); Hello &operator=(const Hello &); }; #if !defined(Q_OS_WIN) void setDefault(const QString &def); #endif
The corresponding SIP specification file would then look something like this:
// Define the SIP wrapper to the hello library. %Module hello 0 %Import qt/qtmod.sip %If (Qt_3_0_0 -) class Hello : QLabel { %TypeHeaderCode #include <hello.h> %End public: Hello(QWidget *parent /TransferThis/, const char *name = 0, WFlags f = 0); private: Hello(const Hello &); }; %If (!WS_WIN) void setDefault(const QString &def); %End %End
Again we look at the differences, but we'll skip those that we've looked at in previous examples.
- The %Import directive has been added to specify that we are extending the class hierarchy defined in the file qt/qtmod.sip. This file is part of PyQt. The build system will take care of finding the file's exact location.
- The %If directive has been added to specify that everything [4] up to the matching %End directive only applies to Qt v3.0 and later. Qt_3_0_0 is a tag defined in qtmod.sip [5] using the %Timeline directive. %Timeline is used to define a tag for each version of a library's API you are wrapping allowing you to maintain all the different versions in a single SIP specification. The build system provides support to configure.py scripts for working out the correct tags to use according to which version of the library is actually installed.
- The public keyword used in defining the super-classes has been removed. This is not supported by SIP.
- The TransferThis annotation has been added to the first argument of the constructor. It specifies that if the argument is not 0 (i.e. the Hello instance being constructed has a parent) then ownership of the instance is transferred from Python to C++. It is needed because Qt maintains objects (i.e. instances derived from the QObject class) in a hierachy. When an object is destroyed all of its children are also automatically destroyed. It is important, therefore, that the Python garbage collector doesn't also try and destroy them. This is covered in more detail in Ownership of Objects. SIP provides many other annotations that can be applied to arguments, functions and classes. Multiple annotations are separated by commas. Annotations may have values.
- The = operator has been removed. This operator is not supported by SIP.
- The %If directive has been added to specify that everything up to the matching %End directive does not apply to Windows. WS_WIN is another tag defined by PyQt, this time using the %Platforms directive. Tags defined by the %Platforms directive are mutually exclusive, i.e. only one may be valid at a time [6].
One question you might have at this point is why bother to define the private copy constructor when it can never be called from Python? The answer is to prevent the automatic generation of a public copy constructor.
We now look at the configure.py script. This is a little different to the script in the previous examples for two related reasons.
Firstly, PyQt includes a pure Python module called pyqtconfig that extends the SIP build system for modules, like our example, that build on top of PyQt. It deals with the details of which version of Qt is being used (i.e. it determines what the correct tags are) and where it is installed. This is called a module's configuration module.
Secondly, we generate a configuration module (called helloconfig) for our own hello module. There is no need to do this, but if there is a chance that somebody else might want to extend your C++ library then it would make life easier for them.
Now we have two scripts. First the configure.py script:
import os import sipconfig import pyqtconfig # The name of the SIP build file generated by SIP and used by the build # system. build_file = "hello.sbf" # Get the PyQt configuration information. config = pyqtconfig.Configuration() # Get the extra SIP flags needed by the imported qt module. Note that # this normally only includes those flags (-x and -t) that relate to SIP's # versioning system. qt_sip_flags = config.pyqt_qt_sip_flags # Run SIP to generate the code. Note that we tell SIP where to find the qt # module's specification files using the -I flag. os.system(" ".join([config.sip_bin, "-c", ".", "-b", build_file, "-I", config.pyqt_sip_dir, qt_sip_flags, "hello.sip"])) # We are going to install the SIP specification file for this module and # its configuration module. installs = [] installs.append(["hello.sip", os.path.join(config.default_sip_dir, "hello")]) installs.append(["helloconfig.py", config.default_mod_dir]) # Create the Makefile. The QtModuleMakefile class provided by the # pyqtconfig module takes care of all the extra preprocessor, compiler and # linker flags needed by the Qt library. makefile = pyqtconfig.QtModuleMakefile( configuration=config, build_file=build_file, installs=installs ) # Add the library we are wrapping. The name doesn't include any platform # specific prefixes or extensions (e.g. the "lib" prefix on UNIX, or the # ".dll" extension on Windows). makefile.extra_libs = ["hello"] # Generate the Makefile itself. makefile.generate() # Now we create the configuration module. This is done by merging a Python # dictionary (whose values are normally determined dynamically) with a # (static) template. content = { # Publish where the SIP specifications for this module will be # installed. "hello_sip_dir": config.default_sip_dir, # Publish the set of SIP flags needed by this module. As these are the # same flags needed by the qt module we could leave it out, but this # allows us to change the flags at a later date without breaking # scripts that import the configuration module. "hello_sip_flags": qt_sip_flags } # This creates the helloconfig.py module from the helloconfig.py.in # template and the dictionary. sipconfig.create_config_module("helloconfig.py", "helloconfig.py.in", content)
Next we have the helloconfig.py.in template script:
import pyqtconfig # These are installation specific values created when Hello was configured. # The following line will be replaced when this template is used to create # the final configuration module. # @SIP_CONFIGURATION@ class Configuration(pyqtconfig.Configuration): """The class that represents Hello configuration values. """ def __init__(self, sub_cfg=None): """Initialise an instance of the class. sub_cfg is the list of sub-class configurations. It should be None when called normally. """ # This is all standard code to be copied verbatim except for the # name of the module containing the super-class. if sub_cfg: cfg = sub_cfg else: cfg = [] cfg.append(_pkg_config) pyqtconfig.Configuration.__init__(self, cfg) class HelloModuleMakefile(pyqtconfig.QtModuleMakefile): """The Makefile class for modules that %Import hello. """ def finalise(self): """Finalise the macros. """ # Make sure our C++ library is linked. self.extra_libs.append("hello") # Let the super-class do what it needs to. pyqtconfig.QtModuleMakefile.finalise(self)
Again, we hope that the scripts are self documenting.
[4] | Some parts of a SIP specification aren't subject to version control. |
[5] | Actually in versions.sip. PyQt uses the %Include directive to split the SIP specification for Qt across a large number of separate .sip files. |
[6] | Tags can also be defined by the %Feature directive. These tags are not mutually exclusive, i.e. any number may be valid at a time. |
When a C++ instance is wrapped a corresponding Python object is created. The Python object behaves as you would expect in regard to garbage collection - it is garbage collected when its reference count reaches zero. What then happens to the corresponding C++ instance? The obvious answer might be that the instance's destructor is called. However the library API may say that when the instance is passed to a particular function, the library takes ownership of the instance, i.e. responsibility for calling the instance's destructor is transferred from the SIP generated module to the library.
Ownership of an instance may also be associated with another instance. The implication being that the owned instance will automatically be destroyed if the owning instance is destroyed. SIP keeps track of these relationships to ensure that Python's cyclic garbage collector can detect and break any reference cycles between the owning and owned instances. The association is implemented as the owning instance taking a reference to the owned instance.
The TransferThis, Transfer and TransferBack annotations are used to specify where, and it what direction, transfers of ownership happen. It is very important that these are specified correctly to avoid crashes (where both Python and C++ call the destructor) and memory leaks (where neither Python and C++ call the destructor).
This applies equally to C structures where the structure is returned to the heap using the free() function.
See also sipTransferTo() and sipTransferBack().
SIP v4.6 introduced support for wide characters (i.e. the wchar_t type). Python's C API includes support for converting between unicode objects and wide character strings and arrays. When converting from a unicode object to wide characters SIP creates the string or array on the heap (using memory allocated using sipMalloc()). This then raises the problem of how this memory is subsequently freed.
The following describes how SIP handles this memory in the different situations where this is an issue.
- When a wide string or array is passed to a function or method then the memory is freed (using sipFree()) after than function or method returns.
- When a wide string or array is returned from a virtual method then SIP does not free the memory until the next time the method is called.
- When an assignment is made to a wide string or array instance variable then SIP does not first free the instance's current string or array.
Python's Global Interpretor Lock (GIL) must be acquired before calls can be made to the Python API. It should also be released when a potentially blocking call to C/C++ library is made in order to allow other Python threads to be executed. In addition, some C/C++ libraries may implement their own locking strategies that conflict with the GIL causing application deadlocks. SIP provides ways of specifying when the GIL is released and acquired to ensure that locking problems can be avoided.
SIP always ensures that the GIL is acquired before making calls to the Python API. By default SIP does not release the GIL when making calls to the C/C++ library being wrapped. The ReleaseGIL annotation can be used to override this behaviour when required.
If SIP is given the -g command line option then the default behaviour is changed and SIP releases the GIL every time is makes calls to the C/C++ library being wrapped. The HoldGIL annotation can be used to override this behaviour when required.
The syntax of the SIP command line is:
sip [options] [specification]
specification is the name of the specification file for the module. If it is omitted then stdin is used.
The full set of command line options is:
-h | Display a help message. |
-V | Display the SIP version number. |
-a file | The name of the QScintilla API file to generate. This file contains a description of the module API in a form that the QScintilla editor component can use for auto-completion and call tips. (The file may also be used by the SciTE editor but must be sorted first.) By default the file is not generated. |
-b file | The name of the build file to generate. This file contains the information about the module needed by the SIP build system to generate a platform and compiler specific Makefile for the module. By default the file is not generated. |
-c dir | The name of the directory (which must exist) into which all of the generated C or C++ code is placed. By default no code is generated. |
-d file | The name of the documentation file to generate. Documentation is included in specification files using the %Doc and %ExportedDoc directives. By default the file is not generated. |
-e | Support for C++ exceptions is enabled. This causes all calls to C++ code to be enclosed in try/catch blocks and C++ exceptions to be converted to Python exceptions. By default exception support is disabled. |
-g | The Python GIL is released before making any calls to the C/C++ library being wrapped and reacquired afterwards. See The Python Global Interpreter Lock and the ReleaseGIL and HoldGIL annotations. |
-I dir | The directory is added to the list of directories searched when looking for a specification file given in an %Include or %Import directive. This option may be given any number of times. |
-j number | The generated code is split into the given number of files. This make it easier to use the parallel build facility of most modern implementations of make. By default 1 file is generated for each C structure or C++ class. |
-r | Debugging statements that trace the execution of the bindings are automatically generated. By default the statements are not generated. |
-s suffix | The suffix to use for generated C or C++ source files. By default .c is used for C and .cpp for C++. |
-t tag | The SIP version tag (declared using a %Timeline directive) or the SIP platform tag (declared using the %Platforms directive) to generate code for. This option may be given any number of times so long as the tags do not conflict. |
-w | The display of warning messages is enabled. By default warning messages are disabled. |
-x feature | The feature (declared using the %Feature directive) is disabled. |
-z file | The name of a file containing more command line options. |
A SIP specification consists of some C/C++ type and function declarations and some directives. The declarations may contain annotations which provide SIP with additional information that cannot be expressed in C/C++. SIP does not include a full C/C++ parser.
It is important to understand that a SIP specification describes the Python API, i.e. the API available to the Python programmer when they import the generated module. It does not have to accurately represent the underlying C/C++ library. There is nothing wrong with omitting functions that make little sense in a Python context, or adding functions implemented with handwritten code that have no C/C++ equivalent. It is even possible (and sometimes necessary) to specify a different super-class hierarchy for a C++ class. All that matters is that the generated code compiles properly.
In most cases the Python API matches the C/C++ API. In some cases handwritten code (see %MethodCode) is used to map from one to the other without SIP having to know the details itself. However, there are a few cases where SIP generates a thin wrapper around a C++ method or constructor (see Generated Derived Classes) and needs to know the exact C++ signature. To deal with these cases SIP allows two signatures to be specified. For example:
class Klass { public: // The Python signature is a tuple, but the underlying C++ signature // is a 2 element array. Klass(SIP_PYTUPLE) [(int *)]; %MethodCode int iarr[2]; if (PyArg_ParseTuple(a0, "ii", &iarr[0], &iarr[1])) { // Note that we use the SIP generated derived class // constructor. Py_BEGIN_ALLOW_THREADS sipCpp = new sipKlass(iarr); Py_END_ALLOW_THREADS } %End };
The following is a semi-formal description of the syntax of a specification file.
specification ::= {module-statement} module-statement ::= [module-directive | statement] module-directive ::= [%CModule | %Copying | %Doc | %ExportedDoc | %ExportedHeaderCode | %Feature | %Import | %Include | %License | %MappedType | mapped-type-template | %Module | %ModuleCode | %ModuleHeaderCode | %OptionalInclude | %Platforms | %PreInitialisationCode | %PostInitialisationCode | sip-option-list | %Timeline | %UnitCode] sip-option-list :: %SIPOptions ( option-list ) option-list ::= option [, option-list] statement :: [class-statement | function | variable] class-statement :: [%If | class | class-template | enum | namespace | opaque-class | operator | struct | typedef | exception] class ::= class name [: super-classes] [class-annotations] { {class-line} }; super-classes ::= name [, super-classes] class-line ::= [class-statement | %BIGetReadBufferCode | %BIGetWriteBufferCode | %BIGetSegCountCode | %BIGetCharBufferCode | %ConvertToSubClassCode | %ConvertToTypeCode | %GCClearCode | %GCTraverseCode | %TypeCode | %TypeHeaderCode | constructor | destructor | method | static-method | virtual-method | special-method | operator | virtual-operator | class-variable | public: | public slots: | protected: | protected slots: | private: | private slots: | signals:] constructor ::= [explicit] name ( [argument-list] ) [exceptions] [function-annotations] [c++-constructor-signature] ; [%MethodCode] c++-constructor-signature ::= [( [argument-list] )] destructor ::= [virtual] ~ name () [exceptions] [= 0] [function-annotations] ; [%MethodCode] [%VirtualCatcherCode] method ::= type name ( [argument-list] ) [const] [exceptions] [= 0] [function-annotations] [c++-signature] ; [%MethodCode] c++-signature ::= [ type ( [argument-list] )] static-method ::= static function virtual-method ::= virtual type name ( [argument-list] ) [const] [exceptions] [= 0] [function-annotations] [c++-signature] ; [%MethodCode] [%VirtualCatcherCode] special-method ::= type special-method-name ( [argument-list] ) [function-annotations] ; [%MethodCode] special-method-name ::= [ __abs__ | __add__ | __and__ | __call__ | __cmp__ | __contains__ | __delitem__ | __div__ | __eq__ | __float__ | __ge__ | __getitem__ | __gt__ | __hash__ | __iadd__ | __iand__ | __idiv__ | __ilshift__ | __imod__ | __imul__ | __int__ | __invert__ | __ior__ | __irshift__ | __isub__ | __ixor__ | __le__ | __len__ | __long__ | __lshift__ | __lt__ | __mod__ | __mul__ | __ne__ | __neg__ | __nonzero__ | __or__ | __pos__ | __repr__ | __rshift__ | __setitem__ | __str__ | __sub__ | __xor__] operator ::= operator-type ( [argument-list] ) [const] [exceptions] [function-annotations] ; [%MethodCode] virtual-operator ::= virtual operator-type ( [argument-list] ) [const] [exceptions] [= 0] [function-annotations] ; [%MethodCode] [%VirtualCatcherCode] operatator-type ::= [ operator-function | operator-cast ] operator-function ::= type operator operator-name operator-cast ::= operator type operator-name ::= [+ | - | * | / | % | & | | | ^ | << | >> | += | -= | *= | /= | %= | &= | |= | ^= | <<= | >>= | ~ | () | [] | < | <= | == | != | > | >>=] class-variable ::= [static] variable class-template :: = template < type-list > class mapped-type-template :: = template < type-list > %MappedType enum ::= enum [name] [enum-annotations] { {enum-line} }; enum-line ::= [%If | name [enum-annotations] , function ::= type name ( [argument-list] ) [exceptions] [function-annotations] ; [%MethodCode] namespace ::= namespace name { {namespace-line} }; namespace-line ::= [%TypeHeaderCode | statement] opaque-class ::= class scoped-name ; struct ::= struct name { {class-line} }; typedef ::= typedef [typed-name | function-pointer] ; variable::= typed-name [variable-annotations] ; [%AccessCode] [%GetCode] [%SetCode] exception ::= %Exception exception-name [exception-base] { [%TypeHeaderCode] %RaiseCode };` exception-name ::= scoped-name exception-base ::= ( [exception-name | python-exception] ) python-exception ::= [SIP_Exception | SIP_StopIteration | SIP_StandardError | SIP_ArithmeticError | SIP_LookupError | SIP_AssertionError | SIP_AttributeError | SIP_EOFError | SIP_FloatingPointError | SIP_EnvironmentError | SIP_IOError | SIP_OSError | SIP_ImportError | SIP_IndexError | SIP_KeyError | SIP_KeyboardInterrupt | SIP_MemoryError | SIP_NameError | SIP_OverflowError | SIP_RuntimeError | SIP_NotImplementedError | SIP_SyntaxError | SIP_IndentationError | SIP_TabError | SIP_ReferenceError | SIP_SystemError | SIP_SystemExit | SIP_TypeError | SIP_UnboundLocalError | SIP_UnicodeError | SIP_UnicodeEncodeError | SIP_UnicodeDecodeError | SIP_UnicodeTranslateError | SIP_ValueError | SIP_ZeroDivisionError | SIP_WindowsError | SIP_VMSError] exceptions ::= throw ( [exception-list] ) exception-list ::= scoped-name [, exception-list] argument-list ::= argument [, argument-list] [, ...] argument ::= [type [name] [argument-annotations] [default-value] | SIP_ANYSLOT [default-value] | SIP_QOBJECT | SIP_RXOBJ_CON | SIP_RXOBJ_DIS | SIP_SIGNAL [default-value] | SIP_SLOT [default-value] | SIP_SLOT_CON | SIP_SLOT_DIS] default-value ::= = expression expression ::= [value | value binary-operator expression] value ::= [unary-operator] simple-value simple-value ::= [scoped-name | function-call | real-value | integer-value | boolean-value | string-value | character-value] typed-name::= type name function-pointer::= type (* name )( [type-list] ) type-list ::= type [, type-list] function-call ::= scoped-name ( [value-list] ) value-list ::= value [, value-list] real-value ::= a floating point number integer-value ::= a number boolean-value ::= [true | false] string-value ::= " {character} " character-value ::= ` character ` unary-operator ::= [! | ~ | - | +] binary-operator ::= [- | + | * | / | & | |] argument-annotations ::= see Argument Annotations class-annotations ::= see Class Annotations enum-annotations ::= see Enum Annotations function-annotations ::= see Function Annotations variable-annotations ::= see Variable Annotations type ::= [const] base-type {*} [&] type-list ::= type [, type-list] base-type ::= [scoped-name | template | struct scoped-name | short | unsigned short | int | unsigned | unsigned int | long | unsigned long | float | double | bool | char | signed char | unsigned char | void | wchar_t | SIP_PYCALLABLE | SIP_PYDICT | SIP_PYLIST | SIP_PYOBJECT | SIP_PYSLICE | SIP_PYTUPLE | SIP_PYTYPE] scoped-name ::= name [:: scoped-name] template ::= scoped-name < type-list > name ::= _A-Za-z {_A-Za-z0-9}
Here is a short list of differences between C++ and the subset supported by SIP that might trip you up.
- SIP does not support the use of [] in types. Use pointers instead.
- A global operator can only be defined if its first argument is a class or a named enum that has been wrapped in the same module.
- Variables declared outside of a class are effectively read-only.
- A class's list of super-classes doesn't not include any access specifier (e.g. public).
SIP supports the use of ... as the last part of a function signature. Any remaining arguments are collected as a Python tuple.
SIP supports a number of additional data types that can be used in Python signatures.
This is both a const char * and a PyObject * that is used as the type of the member instead of const char * in functions that implement the connection or disconnection of an explicitly generated signal to a slot. Handwritten code must be provided to interpret the conversion correctly.
This is a PyObject * that is a Python callable object.
This is a PyObject * that is a Python dictionary object.
This is a PyObject * that is a Python list object.
This is a PyObject * of any Python type.
This is a PyObject * that is a Python slice object.
This is a PyObject * that is a Python tuple object.
This is a PyObject * that is a Python type object.
This is a QObject * that is a C++ instance of a class derived from Qt's QObject class.
This is a QObject * that is a C++ instance of a class derived from Qt's QObject class. It is used as the type of the receiver instead of const QObject * in functions that implement a connection to a slot.
This is a QObject * that is a C++ instance of a class derived from Qt's QObject class. It is used as the type of the receiver instead of const QObject * in functions that implement a disconnection from a slot.
This is a const char * that is used as the type of the signal instead of const char * in functions that implement the connection or disconnection of an explicitly generated signal to a slot.
This is a const char * that is used as the type of the member instead of const char * in functions that implement the connection or disconnection of an explicitly generated signal to a slot.
This is a const char * that is used as the type of the member instead of const char * in functions that implement the connection of an internally generated signal to a slot. The type includes a comma separated list of types that is the C++ signature of of the signal.
To take an example, QAccel::connectItem() connects an internally generated signal to a slot. The signal is emitted when the keyboard accelerator is activated and it has a single integer argument that is the ID of the accelerator. The C++ signature is:
bool connectItem(int id, const QObject *receiver, const char *member);
The corresponding SIP specification is:
bool connectItem(int, SIP_RXOBJ_CON, SIP_SLOT_CON(int));
This is a const char * that is used as the type of the member instead of const char * in functions that implement the disconnection of an internally generated signal to a slot. The type includes a comma separated list of types that is the C++ signature of of the signal.
In this section we describe each of the directives that can be used in specification files. All directives begin with % as the first non-whitespace character in a line.
Some directives have arguments or contain blocks of code or documentation. In the following descriptions these are shown in italics. Optional arguments are enclosed in [brackets].
Some directives are used to specify handwritten code. Handwritten code must not define names that start with the prefix sip.
%AccessCode code %End
This directive is used immediately after the declaration of an instance of a wrapped class or structure, or a pointer to such an instance. You use it to provide handwritten code that overrides the default behaviour.
For example:
class Klass; Klass *klassInstance; %AccessCode // In this contrived example the C++ library we are wrapping defines // klassInstance as Klass ** (which SIP doesn't support) so we // explicitly dereference it. if (klassInstance && *klassInstance) return *klassInstance; // This will get converted to None. return 0; %End
%BIGetCharBufferCode code %End
This directive (along with %BIGetReadBufferCode, %BIGetSegCountCode and %BIGetWriteBufferCode) is used to specify code that implements Python's buffer interface. See the section Buffer Object Structures for the details.
The following variables are made available to the handwritten code:
%BIGetReadBufferCode code %End
This directive (along with %BIGetCharBufferCode, %BIGetSegCountCode and %BIGetWriteBufferCode) is used to specify code that implements Python's buffer interface.
The following variables are made available to the handwritten code:
%BIGetSegCountCode code %End
This directive (along with %BIGetCharBufferCode, %BIGetReadBufferCode and %BIGetWriteBufferCode) is used to specify code that implements Python's buffer interface.
The following variables are made available to the handwritten code:
%BIGetWriteBufferCode code %End
This directive (along with %BIGetCharBufferCode, %BIGetReadBufferCode and %BIGetSegCountCode is used to specify code that implements Python's buffer interface.
The following variables are made available to the handwritten code:
%CModule name [version]
This directive is used to identify that the library being wrapped is a C library and to define the name of the module and it's optional version number.
See the %Module directive for an explanation of the version number.
For example:
%CModule dbus 1
%ConvertFromTypeCode code %End
This directive is used as part of the %MappedType directive to specify the handwritten code that converts an instance of a mapped type to a Python object.
The following variables are made available to the handwritten code:
The handwritten code must explicitly return a PyObject *. If there was an error then a Python exception must be raised and NULL returned.
The following example converts a QList<QWidget *> instance to a Python list of QWidget instances:
%ConvertFromTypeCode PyObject *l; // Create the Python list of the correct length. if ((l = PyList_New(sipCpp -> size())) == NULL) return NULL; // Go through each element in the C++ instance and convert it to a // wrapped QWidget. for (int i = 0; i < sipCpp -> size(); ++i) { QWidget *w = sipCpp -> at(i); PyObject *wobj; // Get the Python wrapper for the QWidget instance, creating a new // one if necessary, and handle any ownership transfer. if ((wobj = sipConvertFromInstance(w, sipClass_QWidget, sipTransferObj)) == NULL) { // There was an error so garbage collect the Python list. Py_DECREF(l); return NULL; } // Add the wrapper to the list. PyList_SET_ITEM(l, i, wobj); } // Return the Python list. return l; %End
%ConvertToSubClassCode code %End
When SIP needs to wrap a C++ class instance it first checks to make sure it hasn't already done so. If it has then it just returns a new reference to the corresponding Python object. Otherwise it creates a new Python object of the appropriate type. In C++ a function may be defined to return an instance of a certain class, but can often return a sub-class instead.
This directive is used to specify handwritten code that exploits any available real-time type information (RTTI) to see if there is a more specific Python type that can be used when wrapping the C++ instance. The RTTI may be provided by the compiler or by the C++ instance itself.
The directive is included in the specification of one of the classes that the handwritten code handles the type conversion for. It doesn't matter which one, but a sensible choice would be the one at the root of that class hierarchy in the module.
Note that if a class hierarchy extends over a number of modules then this directive should be used in each of those modules to handle the part of the hierarchy defined in that module. SIP will ensure that the different pieces of code are called in the right order to determine the most specific Python type to use.
The following variables are made available to the handwritten code:
The handwritten code must not explicitly return.
The following example shows the sub-class conversion code for QEvent based class hierarchy in PyQt:
class QEvent { %ConvertToSubClassCode // QEvent sub-classes provide a unique type ID. switch (sipCpp -> type()) { case QEvent::Timer: sipClass = sipClass_QTimerEvent; break; case QEvent::KeyPress: case QEvent::KeyRelease: sipClass = sipClass_QKeyEvent; break; // Skip the remaining event types to keep the example short. default: // We don't recognise the type. sipClass = NULL; } %End // The rest of the class specification. };
The SIP API includes the sipMapIntToClass() and sipMapStringToClass() functions that convert integer and string based RTTI to Python type objects based on ordered lookup tables.
%ConvertToTypeCode code %End
This directive is used to specify the handwritten code that converts a Python object to a mapped type instance and to handle any ownership transfers. It is used as part of the %MappedType directive and as part of a class specification. The code is also called to determine if the Python object is of the correct type prior to conversion.
When used as part of a class specification it can automatically convert additional types of Python object. For example, PyQt uses it in the specification of the QString class to allow Python string objects and unicode objects to be used wherever QString instances are expected.
The following variables are made available to the handwritten code:
The handwritten code must explicitly return an int the meaning of which depends on the value of sipIsErr.
If sipIsErr is NULL then a non-zero value is returned if the Python object has a type that can be converted to the mapped type. Otherwise zero is returned.
If sipIsErr is not NULL then a combination of the following flags is returned.
- SIP_TEMPORARY is set to indicate that the returned instance is a temporary and should be released to avoid a memory leak.
- SIP_DERIVED_CLASS is set to indicate that the type of the returned instance is a derived class. See Generated Derived Classes.
The following example converts a Python list of QPoint instances to a QList<QPoint> instance:
%ConvertToTypeCode // See if we are just being asked to check the type of the Python // object. if (!sipIsErr) { // Checking whether or not None has been passed instead of a list // has already been done. if (!PyList_Check(sipPy)) return 0; // Check the type of each element. We specify SIP_NOT_NONE to // disallow None because it is a list of QPoint, not of a pointer // to a QPoint, so None isn't appropriate. for (int i = 0; i < PyList_GET_SIZE(sipPy); ++i) if (!sipCanConvertToInstance(PyList_GET_ITEM(sipPy, i), sipClass_QPoint, SIP_NOT_NONE)) return 0; // The type is valid. return 1; } // Create the instance on the heap. QList<QPoint> *ql = new QList<QPoint>; for (int i = 0; i < PyList_GET_SIZE(sipPy); ++i) { QPoint *qp; int state; // Get the address of the element's C++ instance. Note that, in // this case, we don't apply any ownership changes to the list // elements, only to the list itself. qp = reinterpret_cast<QPoint *>(sipConvertToInstance( PyList_GET_ITEM(sipPy, i), sipClass_QPoint, 0, SIP_NOT_NONE, &state, sipIsErr)); // Deal with any errors. if (*sipIsErr) { sipReleaseInstance(qp, sipClass_QPoint, state); // Tidy up. delete ql; // There is no temporary instance. return 0; } ql -> append(*qp); // A copy of the QPoint was appended to the list so we no longer // need it. It may be a temporary instance that should be // destroyed, or a wrapped instance that should not be destroyed. // sipReleaseInstance() will do the right thing. sipReleaseInstance(qp, sipClass_QPoint, state); } // Return the instance. *sipCppPtr = ql; // The instance should be regarded as temporary (and be destroyed as // soon as it has been used) unless it has been transferred from // Python. sipGetState() is a convenience function that implements // this common transfer behaviour. return sipGetState(sipTransferObj); %End
When used in a class specification the handwritten code replaces the code that would normally be automatically generated. This means that the handwritten code must also handle instances of the class itself and not just the additional types that are being supported. This should be done by making calls to sipCanConvertToInstance() to check the object type and sipConvertToInstance() to convert the object. The SIP_NO_CONVERTORS flag must be passed to both these functions to prevent recursive calls to the handwritten code.
%Copying text %End
This directive is used to specify some arbitrary text that will be included at the start of all source files generated by SIP. It is normally used to include copyright and licensing terms.
For example:
%Copying Copyright (c) 2007 Riverbank Computing Limited %End
%Doc text %End
This directive is used to specify some arbitrary text that will be extracted by SIP when the -d command line option is used. The directive can be specified any number of times and SIP will concatenate all the separate pieces of text in the order that it sees them.
Documentation that is specified using this directive is local to the module in which it appears. It is ignored by modules that %Import it. Use the %ExportedDoc directive for documentation that should be included by all modules that %Import this one.
For example:
%Doc <h1>An Example</h1> <p> This fragment of documentation is HTML and is local to the module in which it is defined. </p> %End
This isn't a directive in itself, but is used to terminate a number of directives that allow a block of handwritten code or text to be specified.
%Exception name [(base-exception)] { [*header-code] raise-code };
This directive is used to define new Python exceptions, or to provide a stub for existing Python exceptions. It allows handwritten code to be provided that implements the translation between C++ exceptions and Python exceptions. The arguments to throw () specifiers must either be names of classes or the names of Python exceptions defined by this directive.
name is the name of the exception.
base-exception is the optional base exception. This may be either one of the standard Python exceptions or one defined with a previous %Exception directive.
header-code is the optional %TypeHeaderCode used to specify any external interface to the exception being defined.
raise-code is the %RaiseCode used to specify the handwritten code that converts a reference to the C++ exception to the Python exception.
For example:
%Exception std::exception(SIP_Exception) /PyName=StdException/ { %TypeHeaderCode #include <exception> %End %RaiseCode const char *detail = sipExceptionReference.what(); SIP_BLOCK_THREADS PyErr_SetString(sipException_StdException, detail); SIP_UNBLOCK_THREADS %End };
In this example we map the standard C++ exception to a new Python exception. The new exception is called StdException and is derived from the standard Python exception Exception.
%ExportedDoc text %End
This directive is used to specify some arbitrary text that will be extracted by SIP when the -d command line option is used. The directive can be specified any number of times and SIP will concatenate all the separate pieces of text in the order that it sees them.
Documentation that is specified using this directive will also be included by modules that %Import it.
For example:
%ExportedDoc ========== An Example ========== This fragment of documentation is reStructuredText and will appear in the module in which it is defined and all modules that %Import it. %End
%ExportedHeaderCode code %End
This directive is used to specify handwritten code, typically the declarations of types, that is placed in a header file that is included by all generated code for all modules. It should not include function declarations because Python modules should not explicitly call functions in another Python module.
See also %ModuleCode and %ModuleHeaderCode.
%Feature name
This directive is used to declare a feature. Features (along with %Platforms and %Timeline) are used by the %If directive to control whether or not parts of a specification are processed or ignored.
Features are mutually independent of each other - any combination of features may be enabled or disable. By default all features are enabled. The SIP -x command line option is used to disable a feature.
If a feature is enabled then SIP will automatically generate a corresponding C preprocessor symbol for use by handwritten code. The symbol is the name of the feature prefixed by SIP_FEATURE_.
For example:
%Feature FOO_SUPPORT %If (FOO_SUPPORT) void foo(); %End
%GCClearCode code %End
Python has a cyclic garbage collector which can identify and release unneeded objects even when their reference counts are not zero. If a wrapped C structure or C++ class keeps its own reference to a Python object then, if the garbage collector is to do its job, it needs to provide some handwritten code to traverse and potentially clear those embedded references.
See the section Supporting cyclic garbage collection in Embedding and Extending the Python Interpreter for the details.
This directive is used to specify the code that clears any embedded references. (See %GCTraverseCode for specifying the code that traverses any embedded references.)
The following variables are made available to the handwritten code:
The following simplified example is taken from PyQt. The QCustomEvent class allows arbitary data to be attached to the event. In PyQt this data is always a Python object and so should be handled by the garbage collector:
%GCClearCode PyObject *obj; // Get the object. obj = reinterpret_cast<PyObject *>(sipCpp -> data()); // Clear the pointer. sipCpp -> setData(0); // Clear the reference. Py_XDECREF(obj); // Report no error. sipRes = 0; %End
%GCTraverseCode code %End
This directive is used to specify the code that traverses any embedded references for Python's cyclic garbage collector. (See %GCClearCode for a full explanation.)
The following variables are made available to the handwritten code:
The following simplified example is taken from PyQt's QCustomEvent class:
%GCTraverseCode PyObject *obj; // Get the object. obj = reinterpret_cast<PyObject *>(sipCpp -> data()); // Call the visit function if there was an object. if (obj) sipRes = sipVisit(obj, sipArg); else sipRes = 0; %End
%GetCode code %End
This directive is used after the declaration of a C++ class variable or C structure member to specify handwritten code to convert it to a Python object. It is usually used to handle types that SIP cannot deal with automatically.
The following variables are made available to the handwritten code:
For example:
struct Entity { /* * In this contrived example the C library we are wrapping actually * defines this as char buffer[100] which SIP cannot handle * automatically. */ char *buffer; %GetCode sipPy = PyString_FromStringAndSize(sipCpp -> buffer, 100); %End %SetCode char *ptr; int length; if (PyString_AsStringAndSize(sipPy, &ptr, &length) == -1) sipErr = 1; else if (length != 100) { /* * Raise an exception because the length isn't exactly right. */ PyErr_SetString(PyExc_ValueError, "an Entity.buffer must be exactly 100 bytes"); sipErr = 1; } else memcpy(sipCpp -> buffer, ptr, 100); %End }
%If (expression) specification %End
where
expression ::= [ored-qualifiers | range] ored-qualifiers ::= [qualifier | qualifier || ored-qualifiers] qualifier ::= [!] [feature | platform] range ::= [version] - [version]
This directive is used in conjunction with features (see %Feature), platforms (see %Platforms) and versions (see %Timeline) to control whether or not parts of a specification are processed or not.
A range of versions means all versions starting with the lower bound up to but excluding the upper bound. If the lower bound is omitted then it is interpreted as being before the earliest version. If the upper bound is omitted then it is interpreted as being after the latest version.
For example:
%Feature SUPPORT_FOO %Platforms {WIN32_PLATFORM POSIX_PLATFORM MACOS_PLATFORM} %Timeline {V1_0 V1_1 V2_0 V3_0} %If (!SUPPORT_FOO) // Process this if the SUPPORT_FOO feature is disabled. %End %If (POSIX_PLATFORM || MACOS_PLATFORM) // Process this if either the POSIX_PLATFORM or MACOS_PLATFORM // platforms are enabled. %End %If (V1_0 - V2_0) // Process this if either V1_0 or V1_1 is enabled. %End %If (V2_0 - ) // Process this if either V2_0 or V3_0 is enabled. %End %If ( - ) // Always process this. %End
Note that this directive is not implemented as a preprocessor. Only the following parts of a specification are affected by it:
- class
- %ConvertFromTypeCode
- %ConvertToSubClassCode
- %ConvertToTypeCode
- enum
- %ExportedHeaderCode
- functions
- %GCClearCode
- %GCTraverseCode
- %If
- %MappedType
- %MethodCode
- %ModuleCode
- %ModuleHeaderCode
- namespace
- %PostInitialisationCode
- %PreInitialisationCode
- struct
- typedef
- %TypeCode
- %TypeHeaderCode
- %UnitCode
- variables
- %VirtualCatcherCode
Also note that the only way to specify the logical and of qualifiers is to use nested %If directives.
%Import filename
This directive is used to import the specification of another module. This is needed if the current module makes use of any types defined in the imported module, e.g. as an argument to a function, or to sub-class.
If filename cannot be opened then SIP prepends filename with the name of the directory containing the current specification file (i.e. the one containing the %Import directive) and tries again. If this also fails then SIP prepends filename with each of the directories, in turn, specified by the -I command line option.
For example:
%Import qt/qtmod.sip
%Include filename
This directive is used to include contents of another file as part of the specification of the current module. It is the equivalent of the C preprocessor's #include directive and is used to structure a large module specification into manageable pieces.
%Include follows the same search process as %Import when trying to open filename.
For example:
%Include qwidget.sip
%License /license-annotations/
This directive is used to specify the contents of an optional license dictionary. The license dictionary is called __license__ and is stored in the module dictionary. The elements of the dictionary are specified using the Licensee, Signature, Timestamp and Type annotations. Only the Type annotation is compulsory.
Note that this directive isn't an attempt to impose any licensing restrictions on a module. It is simply a method for easily embedding licensing information in a module so that it is accessible to Python scripts.
For example:
%License /Type="GPL"/
template<type-list> %MappedType type { [header-code] [convert-to-code] [convert-from-code] }; %MappedType type { [header-code] [convert-to-code] [convert-from-code] };
This directive is used to define an automatic mapping between a C or C++ type and a Python type. It can be used as part of a template, or to map a specific type.
When used as part of a template type cannot itself refer to a template. Any occurrences of any of the type names (but not any * or &) in type-list will be replaced by the actual type names used when the template is instantiated. Template mapped types are instantiated automatically as required (unlike template classes which are only instantiated using typedef).
Any explicit mapped type will be used in preference to any template that maps the same type, ie. a template will not be automatically instantiated if there is an explicit mapped type.
header-code is the %TypeHeaderCode used to specify the library interface to the type being mapped.
convert-to-code is the %ConvertToTypeCode used to specify the handwritten code that converts a Python object to an instance of the mapped type.
convert-from-code is the %ConvertFromTypeCode used to specify the handwritten code that converts an instance of the mapped type to a Python object.
For example:
template<Type *> %MappedType QList { %TypeHeaderCode // Include the library interface to the type being mapped. #include <qlist.h> %End %ConvertToTypeCode // See if we are just being asked to check the type of the Python // object. if (sipIsErr == NULL) { // Check it is a list. if (!PyList_Check(sipPy)) return 0; // Now check each element of the list is of the type we expect. // The template is for a pointer type so we don't disallow None. for (int i = 0; i < PyList_GET_SIZE(sipPy); ++i) if (!sipCanConvertToInstance(PyList_GET_ITEM(sipPy, i), sipClass_Type, 0)) return 0; return 1; } // Create the instance on the heap. QList<Type *> *ql = new QList<Type *>; for (int i = 0; i < PyList_GET_SIZE(sipPy); ++i) { // Use the SIP API to convert the Python object to the // corresponding C++ instance. Note that we apply any ownership // transfer to the list itself, not the individual elements. Type *t = reinterpret_cast<Type *>(sipConvertToInstance( PyList_GET_ITEM(sipPy, i), sipClass_Type, 0, 0, 0, sipIsErr)); if (*sipIsErr) { // Tidy up. delete ql; // There is nothing on the heap. return 0; } // Add the pointer to the C++ instance. ql -> append(t); } // Return the instance on the heap. *sipCppPtr = ql; // Apply the normal transfer. return sipGetState(sipTransferObj); %End %ConvertFromTypeCode PyObject *l; // Create the Python list of the correct length. if ((l = PyList_New(sipCpp -> size())) == NULL) return NULL; // Go through each element in the C++ instance and convert it to the // corresponding Python object. for (int i = 0; i < sipCpp -> size(); ++i) { Type *t = sipCpp -> at(i); PyObject *tobj; if ((tobj = sipConvertFromInstance(t, sipClass_Type, sipTransferObj)) == NULL) { // There was an error so garbage collect the Python list. Py_DECREF(l); return NULL; } PyList_SET_ITEM(l, i, tobj); } // Return the Python list. return l; %End }
Using this we can use, for example, QList<QObject *> throughout the module's specification files (and in any module that imports this one). The generated code will automatically map this to and from a Python list of QObject instances when appropriate.
%MethodCode code %End
This directive is used as part of the specification of a global function, class method, operator, constructor or destructor to specify handwritten code that replaces the normally generated call to the function being wrapped. It is usually used to handle argument types and results that SIP cannot deal with automatically.
The specified code is embedded in-line after the function's arguments have been successfully converted from Python objects to their C or C++ equivalents. The specified code must not include any return statements.
In the context of a destructor the specified code is embedded in-line in the Python type's deallocation function. Unlike other contexts it supplements rather than replaces the normally generated code, so it must not include code to return the C structure or C++ class instance to the heap. The code is only called if ownership of the structure or class is with Python.
The specified code must also handle the Python Global Interpreter Lock (GIL). If compatibility with SIP v3.x is required then the GIL must be released immediately before the C++ call and reacquired immediately afterwards as shown in this example fragment:
Py_BEGIN_ALLOW_THREADS sipCpp -> foo(); Py_END_ALLOW_THREADS
If compatibility with SIP v3.x is not required then this is optional but should be done if the C++ function might block the current thread or take a significant amount of time to execute. (See The Python Global Interpreter Lock and the ReleaseGIL and HoldGIL annotations.)
The following variables are made available to the handwritten code:
There is a variable for each argument of the Python signature (excluding any self argument) named a0, a1, etc. The type of the variable is the same as the type defined in the specification with the following exceptions:
Note that handwritten code for destructors never has any arguments.
The handwritten code should set this to a non-zero value, and raise an appropriate Python exception, if an error is detected.
sipIsErr is not provided for destructors.
The handwritten code should set this to the result to be returned. The type of the variable is the same as the type defined in the Python signature in the specification with the following exception:
sipRes is not provided for inplace operators (e.g. += or __imul__) as their results are handled automatically, nor for class constructors.
This is only made available for non-abstract, virtual methods. It is set if self was explicitly passed as the first argument of the method rather than being bound to the method. In other words, the call was:
Klass.foo(self, ...)
rather than:
self.foo(...)
The following is a complete example:
class Klass { public: virtual int foo(SIP_PYTUPLE); %MethodCode // The C++ API takes a 2 element array of integers but passing a // two element tuple is more Pythonic. int iarr[2]; if (PyArg_ParseTuple(a0, "ii", &iarr[0], &iarr[1])) { Py_BEGIN_ALLOW_THREADS sipRes = sipSelfWasArg ? sipCpp -> Klass::foo(iarr) : sipCpp -> foo(iarr); Py_END_ALLOW_THREADS } else { // PyArg_ParseTuple() will have raised the exception. sipIsErr = 1; } %End };
As the example is a virtual method [7], note the use of sipSelfWasArg to determine exactly which implementation of foo() to call.
If a method is in the protected section of a C++ class then the call should instead be:
sipRes = sipCpp -> sipProtectVirt_foo(sipSelfWasArg, iarr);
If a method is in the protected section of a C++ class but is not virtual then the call should instead be:
sipRes = sipCpp -> sipProtect_foo(iarr);
[7] | See %VirtualCatcherCode for a description of how SIP generated code handles the reimplementation of C++ virtual methods in Python. |
%Module name [version]
This directive is used to identify that the library being wrapped is a C++ library and to define the name of the module and it's optional version number.
The name may contain periods to specify that the module is part of a Python package.
The optional version number is useful if you (or others) might create other modules that build on this module, i.e. if another module might %Import this module. Under the covers, a module exports an API that is used by modules that %Import it and the API is given a version number. A module built on that module knows the version number of the API that it is expecting. If, when the modules are imported at run-time, the version numbers do not match then a Python exception is raised. The dependent module must then be re-built using the correct specification files for the base module.
The version number should be incremented whenever a module is changed. Some changes don't affect the exported API, but it is good practice to change the version number anyway.
For example:
%Module qt 5
%ModuleCode code %End
This directive is used to specify handwritten code, typically the implementations of utility functions, that can be called by other handwritten code in the module.
For example:
%ModuleCode // Print an object on stderr for debugging purposes. void dump_object(PyObject *o) { PyObject_Print(o, stderr, 0); fprintf(stderr, "\n"); } %End
See also %ExportedHeaderCode and %ModuleHeaderCode.
%ModuleHeaderCode code %End
This directive is used to specify handwritten code, typically the declarations of utility functions, that is placed in a header file that is included by all generated code for the same module.
For example:
%ModuleHeaderCode void dump_object(PyObject *o); %End
See also %ExportedHeaderCode and %ModuleCode.
%OptionalInclude filename
This directive is identical to the %Include directive except that SIP silently continues processing if filename could not be opened.
For example:
%OptionalInclude license.sip
%Platforms {name name ...}
This directive is used to declare a set of platforms. Platforms (along with %Feature and %Timeline) are used by the %If directive to control whether or not parts of a specification are processed or ignored.
Platforms are mutually exclusive - only one platform can be enabled at a time. By default all platforms are disabled. The SIP -t command line option is used to enable a platform.
For example:
%Platforms {WIN32_PLATFORM POSIX_PLATFORM MACOS_PLATFORM} %If (WIN32_PLATFORM) void undocumented(); %End %If (POSIX_PLATFORM) void documented(); %End
%PostInitialisationCode code %End
This directive is used to specify handwritten code that is embedded in-line at the very end of the generated module initialisation code.
The following variables are made available to the handwritten code:
For example:
%PostInitialisationCode // The code will be executed when the module is first imported and // after all other initialisation has been completed. %End
%PreInitialisationCode code %End
This directive is used to specify handwritten code that is embedded in-line at the very start of the generated module initialisation code.
For example:
%PreInitialisationCode // The code will be executed when the module is first imported and // before other initialisation has been completed. %End
%RaiseCode code %End
This directive is used as part of the definition of an exception using the %Exception directive to specify handwritten code that raises a Python exception when a C++ exception has been caught. The code is embedded in-line as the body of a C++ catch () clause.
The specified code must handle the Python Global Interpreter Lock (GIL) if necessary. The GIL must be acquired before any calls to the Python API and released after the last call as shown in this example fragment:
SIP_BLOCK_THREADS PyErr_SetNone(PyErr_Exception); SIP_UNBLOCK_THREADS
Finally, the specified code must not include any return statements.
The following variable is made available to the handwritten code:
See the %Exception directive for an example.
%SetCode code %End
This directive is used after the declaration of a C++ class variable or C structure member to specify handwritten code to convert it from a Python object. It is usually used to handle types that SIP cannot deal with automatically.
The following variables are made available to the handwritten code:
See the %GetCode directive for an example.
This directive sets one or more options that controls different aspects of SIP's behaviour. In this version all the available options are provided specifically to support PyQt and so are not documented.
%Timeline {name name ...}
This directive is used to declare a set of versions released over a period of time. Versions (along with %Feature and %Platforms) are used by the %If directive to control whether or not parts of a specification are processed or ignored.
Versions are mutually exclusive - only one version can be enabled at a time. By default all versions are disabled. The SIP -t command line option is used to enable a version.
For example:
%Timeline {V1_0 V1_1 V2_0 V3_0} %If (V1_0 - V2_0) void foo(); %End %If (V2_0 -) void foo(int = 0); %End
%Timeline can be used any number of times in a module to allow multiple libraries to be wrapped in the same module.
%TypeCode code %End
This directive is used as part of the specification of a C structure or a C++ class to specify handwritten code, typically the implementations of utility functions, that can be called by other handwritten code in the structure or class.
For example:
class Klass { %TypeCode // Print an instance on stderr for debugging purposes. static void dump_klass(const Klass *k) { fprintf(stderr,"Klass %s at %p\n", k -> name(), k); } %End // The rest of the class specification. };
Because the scope of the code is normally within the generated file that implements the type, any utility functions would normally be declared static. However a naming convention should still be adopted to prevent clashes of function names within a module in case the SIP -j command line option is used.
%TypeHeaderCode code %End
This directive is used to specify handwritten code that defines the interface to a C or C++ type being wrapped, either a structure, a class, or a template. It is used within a class definition or a %MappedType directive.
Normally code will be a pre-processor #include statement.
For example:
// Wrap the Klass class. class Klass { %TypeHeaderCode #include <klass.h> %End // The rest of the class specification. };
%UnitCode code %End
This directive is used to specify handwritten code that it included at the very start of a generated compilation unit (ie. C or C++ source file). It is typically used to #include a C++ precompiled header file.
%VirtualCatcherCode code %End
For most classes there are corresponding generated derived classes that contain reimplementations of the class's virtual methods. These methods (which SIP calls catchers) determine if there is a corresponding Python reimplementation and call it if so. If there is no Python reimplementation then the method in the original class is called instead.
This directive is used to specify handwritten code that replaces the normally generated call to the Python reimplementation and the handling of any returned results. It is usually used to handle argument types and results that SIP cannot deal with automatically.
This directive can also be used in the context of a class destructor to specify handwritten code that is embedded in-line in the internal derived class's destructor.
In the context of a method the Python Global Interpreter Lock (GIL) is automatically acquired before the specified code is executed and automatically released afterwards.
In the context of a destructor the specified code must handle the GIL. The GIL must be acquired before any calls to the Python API and released after the last call as shown in this example fragment:
SIP_BLOCK_THREADS Py_DECREF(obj); SIP_UNBLOCK_THREADS
The following variables are made available to the handwritten code in the context of a method:
No variables are made available in the context of a destructor.
For example:
class Klass { public: virtual int foo(SIP_PYTUPLE) [int (int *)]; %MethodCode // The C++ API takes a 2 element array of integers but passing a // two element tuple is more Pythonic. int iarr[2]; if (PyArg_ParseTuple(a0, "ii", &iarr[0], &iarr[1])) { Py_BEGIN_ALLOW_THREADS sipRes = sipCpp -> Klass::foo(iarr); Py_END_ALLOW_THREADS } else { // PyArg_ParseTuple() will have raised the exception. sipIsErr = 1; } %End %VirtualCatcherCode // Convert the 2 element array of integers to the two element // tuple. PyObject *result; result = sipCallMethod(&sipIsErr, sipMethod, "ii", a0[0], a0[1]); if (result != NULL) { // Convert the result to the C++ type. sipParseResult(&sipIsErr, sipMethod, result, "i", &sipRes); Py_DECREF(result); } %End };
In this section we describe each of the annotations that can be used in specification files.
Annotations can either be argument annotations, class annotations, enum annotations, exception annotations, function annotations, license annotations, or variable annotations depending on the context in which they can be used.
Annotations are placed between forward slashes (/). Multiple annotations are comma separated within the slashes.
Annotations have a type and, possibly, a value. The type determines the format of the value. The name of an annotation and its value are separated by =.
Annotations can have one of the following types:
The following example shows argument and function annotations:
void exec(QWidget * /Transfer/) /ReleaseGIL, PyName=call_exec/;
Note that the current version of SIP does not complain about unknown annotations, or annotations used out of their correct context.
This boolean annotation specifies that the value of the corresponding argument (which should be either SIP_PYCALLABLE, SIP_PYDICT, SIP_PYLIST, SIP_PYSLICE, SIP_PYTUPLE or SIP_PYTYPE) may be None.
This boolean annotation specifies that the corresponding argument (which should be either char * or unsigned char *) refers to an array rather than a '\0' terminated string. There must be a corresponding argument with the ArraySize annotation specified. The annotation may only be specified once in a list of arguments.
This boolean annotation specifies that the corresponding argument (which should be either short, unsigned short, int, unsigned, long or unsigned long) refers to the size of an array. There must be a corresponding argument with the Array annotation specified. The annotation may only be specified once in a list of arguments.
Python will automatically convert between certain compatible types. For example, if a floating pointer number is expected and an integer supplied, then the integer will be converted appropriately. This can cause problems when wrapping C or C++ functions with similar signatures. For example:
// The wrapper for this function will also accept an integer argument // which Python will automatically convert to a floating point number. void foo(double); // The wrapper for this function will never get used. void foo(int);
This boolean annotation specifies that the corresponding argument (which should be either bool, int, float, double or a wrapped class) must match the type without any automatic conversions. In the context of a wrapped class the invocation of any %ConvertToTypeCode is suppressed.
The following example gets around the above problem:
// The wrapper for this function will only accept floating point numbers. void foo(double /Constrained/); // The wrapper for this function will be used for anything that Python can // convert to an integer, except for floating point numbers. void foo(int);
This boolean annotation is only ever used in conjunction with handwritten code specified with the %MethodCode directive. It causes an extra variable to be generated for the corresponding argument (which should be a wrapped C structure or C++ class instance) which is a pointer to the Python object that wraps the argument.
See the %MethodCode directive for more detail.
This boolean annotation is used to specify that the corresponding argument (which should be a pointer type) is used to pass a value to the function.
For pointers to wrapped C structures or C++ class instances, char * and unsigned char * then this annotation is assumed unless the Out annotation is specified.
For pointers to other types then this annotation must be explicitly specified if required. The argument will be dereferenced to obtain the actual value.
This boolean annotation is used to specify that the corresponding argument (which should be a pointer type) is used by the function to return a value as an element of a tuple.
For pointers to wrapped C structures or C++ class instances, char * and unsigned char * then this annotation must be explicitly specified if required.
For pointers to other types then this annotation is assumed unless the In annotation is specified.
This boolean annotation is used to specify that ownership of the corresponding argument (which should be a wrapped C structure or C++ class instance) is transferred from Python to C++. In addition, if the argument is of a class method, then it is associated with the class instance with regard to the cyclic garbage collector.
See Ownership of Objects for more detail.
This boolean annotation is used to specify that ownership of the corresponding argument (which should be a wrapped C structure or C++ class instance) is transferred back to Python from C++. In addition, any association of the argument with regard to the cyclic garbage collector with another instance is removed.
Note that this can also be used as a function annotation.
See Ownership of Objects for more detail.
This boolean annotation is only used in C++ constructors or methods. In the context of a constructor or factory method it specifies that ownership of the instance being created is transferred from Python to C++ if the corresponding argument (which should be a wrapped C structure or C++ class instance) is not None. In addition, the newly created instance is associated with the argument with regard to the cyclic garbage collector.
In the context of a non-factory method it specifies that ownership of this is transferred from Python to C++ if the corresponding argument is not None. If it is None then ownership is transferred to Python.
The annotation may be used more that once, in which case ownership is transferred to last instance that is not None.
See Ownership of Objects for more detail.
This boolean annotation is used to specify that the class has additional pure virtual methods that have not been specified and so it cannot be instantiated or sub-classed from Python.
This boolean annotation is used to specify that the class's destructor should not be called until the Python interpreter exits. It would normally only be applied to singleton classes.
When the Python interpreter exits the order in which any wrapped instances are garbage collected is unpredictable. However, the underlying C or C++ instances may need to be destroyed in a certain order. If this annotation is specified then when the wrapped instance is garbage collected the C or C++ instance is not destroyed but instead added to a list of delayed instances. When the interpreter exits then the function sipDelayedDtors is called with the list of delayed instances. sipDelayedDtors can then choose to call (or ignore) the destructors in any desired order.
The sipDelayedDtors function must be specified using the %ModuleCode directive. It's signature is as follows:
static void sipDelayedDtors(const sipDelayedDtor *dd_list);
dd_list is the linked list of delayed instances. The following fields are defined.
Note that the above applies only to C and C++ instances that are owned by Python.
This boolean annotation is used to specify that the class is defined in another module. Declarations of external classes are private to the module in which they appear.
This boolean annotation is used to suppress the automatic generation of default constructors for the class.
This name annotation specifies an alternative name for the class being wrapped which is used when it is referred to from Python. It is required when a class name is the same as a Python keyword. It may also be used to avoid name clashes with other objects (e.g. enums, exceptions, functions) that have the same name in the same C++ scope.
This name annotation specifies an alternative name for the enum or enum member being wrapped which is used when it is referred to from Python. It is required when an enum or enum member name is the same as a Python keyword. It may also be used to avoid name clashes with other objects (e.g. classes, exceptions, functions) that have the same name in the same C++ scope.
This name annotation specifies an alternative name for the exception being defined which is used when it is referred to from Python. It is required when an exception name is the same as a Python keyword. It may also be used to avoid name clashes with other objects (e.g. classes, enums, functions) that have the same name.
This optional name annotation is used with class methods to specify that the method be automatically included in all sub-classes. The value is the name of a feature (specified using the %Feature directive) which must be enabled for the method to be generated.
This boolean annotation is only used with C++ constructors. Sometimes SIP needs to create a class instance. By default it uses a constructor with no compulsory arguments if one is specified. (SIP will automatically generate a constructor with no arguments if no constructors are specified.) This annotation is used to explicitly specify which constructor to use. Zero is passed as the value of any arguments to the constructor.
This boolean annotation specifies that the value returned by the function (which should be a wrapped C structure or C++ class instance) is a newly created instance and is owned by Python.
See Ownership of Objects for more detail.
This boolean annotation specifies that the Python Global Interpreter Lock (GIL) is not released before the call to the underlying C or C++ function. See The Python Global Interpreter Lock and the ReleaseGIL annotation.
This boolean annotation specifies that the function will create a new thread.
This boolean annotation is only used with C++ constructors. In many cases SIP generates a derived class for each class being wrapped (see Generated Derived Classes). This derived class contains constructors with the same C++ signatures as the class being wrapped. Sometimes you may want to define a Python constructor that has no corresponding C++ constructor. This annotation is used to suppress the generation of the constructor in the derived class.
This boolean annotation specifies that the operator should be interpreted as a numeric operator rather than a sequence operator. Python uses the + operator for adding numbers and concatanating sequences, and the * operator for multiplying numbers and repeating sequences. SIP tries to work out which is meant by looking at other operators that have been defined for the type. If it finds either -, -=, /, /=, % or %= defined then it assumes that +, +=, * and *= should be numeric operators. Otherwise, if it finds either [], __getitem__(), __setitem__() or __delitem__() defined then it assumes that they should be sequence operators. This annotation is used to force SIP to treat the operator as numeric.
This name annotation is used to specify the name of a Python builtin that is called immediately after the call to the underlying C or C++ function or any handwritten code. The builtin is not called if an error occurred. It is primarily used to integrate with debuggers.
This name annotation is used to specify the name of a Python builtin that is called immediately after the function's arguments have been successfully parsed and before the call to the underlying C or C++ function or any handwritten code. It is primarily used to integrate with debuggers.
This name annotation specifies an alternative name for the function being wrapped which is used when it is referred to from Python. It is required when a function or method name is the same as a Python keyword. It may also be used to avoid name clashes with other objects (e.g. classes, enums, exceptions) that have the same name in the same C++ scope.
This boolean annotation specifies that the Python Global Interpreter Lock (GIL) is released before the call to the underlying C or C++ function and reacquired afterwards. It should be used for functions that might block or take a significant amount of time to execute. See The Python Global Interpreter Lock and the HoldGIL annotation.
This boolean annotation specifies that ownership of the value returned by the function (which should be a wrapped C structure or C++ class instance) is transferred back to Python from C++. Normally returned values (unless they are new references to already wrapped values) are owned by C++. In addition, any association of the returned value with regard to the cyclic garbage collector with another instance is removed.
Note that this can also be used as an argument annotation.
See Ownership of Objects for more detail.
This optional string annotation specifies the license's licensee. No restrictions are placed on the contents of the string.
See the %License directive.
This optional string annotation specifies the license's signature. No restrictions are placed on the contents of the string.
See the %License directive.
This optional string annotation specifies the license's timestamp. No restrictions are placed on the contents of the string.
See the %License directive.
This string annotation specifies the license's type. No restrictions are placed on the contents of the string.
See the %License directive.
This name annotation specifies an alternative name for the variable being wrapped which is used when it is referred to from Python. It is required when a variable name is the same as a Python keyword. It may also be used to avoid name clashes with other objects (e.g. classes, functions) that have the same name in the same C++ scope.
In this section we describe the API that can be used by handwritten code in specification files.
This is a C preprocessor symbol that defines the major number of the SIP API. Its value is a number. There is no direct relationship between this and the SIP version number.
This is a C preprocessor symbol that defines the minor number of the SIP API. Its value is a number. There is no direct relationship between this and the SIP version number.
This is a C preprocessor macro that will make sure the Python Global Interpreter Lock (GIL) is acquired. Python API calls must only be made when the GIL has been acquired. There must be a corresponding SIP_UNBLOCK_THREADS at the same lexical scope.
This is a C preprocessor macro that is defined as Py_ssize_t for Python v2.5 and later, and as int for earlier versions of Python. It makes it easier to write PEP 353 compliant handwritten code.
This is a C preprocessor macro that will restore the Python Global Interpreter Lock (GIL) to the state it was prior to the corresponding SIP_BLOCK_THREADS.
This is a C preprocessor symbol that defines the SIP version number represented as a 3 part hexadecimal number (e.g. v4.0.0 is represented as 0x040000).
This is a C preprocessor symbol that defines the SIP version number represented as a string. For development snapshots it will start with snapshot-.
This creates a Python object based on a format string and associated values in a similar way to the Python Py_BuildValue() function. If there was an error then NULL is returned and a Python exception is raised. If iserr is not NULL then the location it points to is set to a non-zero value. format is the string of format characters.
If format begins and ends with parentheses then a tuple of objects is created. If format contains more than one format character then parentheses must be specified.
In the following description the first letter is the format character, the entry in parentheses is the Python object type that the format character will create, and the entry in brackets are the types of the C/C++ values to be passed.
This calls a Python method passing a tuple of arguments based on a format string and associated values in a similar way to the Python PyObject_CallObject() function. If there was an error then NULL is returned and a Python exception is raised. If iserr is not NULL then the location it points to is set to a non-zero value. method is the Python bound method to call. format is the string of format characters (see sipBuildResult()).
This is normally called by handwritten code specified with the %VirtualCatcherCode directive with method being the supplied sipMethod.
This returns a non-zero value if a Python object can be converted to an instance of a C structure or C++ class. obj is the Python object. type is the generated type corresponding to the C/C++ type being checked. flags is any combination of the following values used to fine tune the check.
- SIP_NOT_NONE causes the check to fail if obj is None.
- SIP_NO_CONVERTORS suppresses the use of of any %ConvertToTypeCode for type.
This returns a non-zero value if a Python object can be converted to an instance of a C structure or C++ class which has been implemented as a mapped type. obj is the Python object. mt is an opaque structure returned by sipFindMappedType(). flags is any combination of the following values used to fine tune the check.
- SIP_NOT_NONE causes the check to fail if obj is None.
This function is deprecated from SIP v4.4. It is equivalent to:
sipConvertToInstance(obj, type, NULL, SIP_NO_CONVERTORS, NULL, iserr);
This converts a Python object to an instance of a C structure or C++ class assuming that a previous call to sipCanConvertToInstance() has been successful. obj is the Python object. type is the generated type corresponding to the C/C++ type returned. It may be any class in the object's class hierarchy. transferObj controls any ownership changes to obj. If it is NULL then the ownership is unchanged. If it is Py_None then ownership is transferred to Python via a call to sipTransferBack(). Otherwise ownership is transferred to C/C++ and obj associated with transferObj via a call to sipTransferTo(). flags is any combination of the following values used to fine tune the check.
- SIP_NOT_NONE causes the check to fail if obj is None.
- SIP_NO_CONVERTORS suppresses the use of of any %ConvertToTypeCode for type.
If state is not NULL then the location it points to is set to describe the state of the returned C/C++ instance and is the value returned by any %ConvertToTypeCode. The calling code must then release the value at some point to prevent a memory leak by calling sipReleaseInstance(). If there is an error then the location iserr points to is set to a non-zero value. If it was initially a non-zero value then the conversion isn't attempted in the first place. (This allows several calls to be made that share the same error flag so that it only needs to be tested once rather than after each call.)
This converts a Python object to an instance of a C structure or C++ class that is implemented as a mapped type assuming that a previous call to sipCanConvertToMappedType() has been successful. obj is the Python object. mt is the opaque structure returned by sipFindMappedType(). transferObj controls any ownership changes to obj. If it is NULL then the ownership is unchanged. If it is Py_None then ownership is transferred to Python via a call to sipTransferBack(). Otherwise ownership is transferred to C/C++ and obj associated with transferObj via a call to sipTransferTo(). flags is any combination of the following values used to fine tune the check.
- SIP_NOT_NONE causes the check to fail if obj is None.
If state is not NULL then the location it points to is set to describe the state of the returned C/C++ instance and is the value returned by any %ConvertToTypeCode. The calling code must then release the value at some point to prevent a memory leak by calling sipReleaseMappedType(). If there is an error then the location iserr points to is set to a non-zero value. If it was initially a non-zero value then the conversion isn't attempted in the first place. (This allows several calls to be made that share the same error flag so that it only needs to be tested once rather than after each call.)
This C structure is used with sipMapIntToClass() to define a mapping between integer based RTTI and generated type objects. The structure elements are as follows.
This function is deprecated from SIP v4.4. It is equivalent to:
sipCanConvertToInstance(obj, type, SIP_NOT_NONE | SIP_NO_CONVERTORS);
This converts a Python object (usually returned by a method) to C/C++ based on a format string and associated values in a similar way to the Python PyArg_ParseTuple() function. If there was an error then a negative value is returned and a Python exception is raised. If iserr is not NULL then the location it points to is set to a non-zero value. method is the Python bound method that returned the result object. format is the string of format characters.
This is normally called by handwritten code specified with the %VirtualCatcherCode directive with method being the supplied sipMethod and result being the value returned by sipCallMethod().
If format begins and ends with parentheses then result must be a Python tuple and the rest of format is applied to the tuple contents.
In the following description the first letter is the format character, the entry in parentheses is the Python object type that the format character will convert, and the entry in brackets are the types of the C/C++ values to be passed.
Convert a Python object to a C structure or a C++ class instance and return its state as described in sipConvertToInstance(). f is a combination of the following flags encoded as an ASCII character by adding 0 to the combined value:
0x01 disallows the conversion of Py_None to NULL
0x02 implements the Factory annotation
- 0x04 suppresses the return of the state of the returned C/C++
- instance. Note that the int * used to return the state is not passed if this flag is specified.
Convert a Python object to a C structure or a C++ class instance implemented as a mapped type and return its state as described in sipConvertToMappedType(). f is a combination of the following flags encoded as an ASCII character by adding 0 to the combined value:
0x01 disallows the conversion of Py_None to NULL
0x02 implements the Factory annotation
- 0x04 suppresses the return of the state of the returned C/C++
- instance. Note that the int * used to return the state is not passed if this flag is specified.
This C structure is used with sipMapStringToClass() to define a mapping between '\0' terminated string based RTTI and generated type objects. The structure elements are as follows.
This function is deprecated from SIP v4.3. If tocpp is non-zero then the equivalent call is:
sipTransferTo(obj, obj);
If tocpp is zero then the equivalent call is:
sipTransferBack(obj);
This is a C structure that represents a Python wrapped instance. It is an extension of the Python PyObject structure and so may be safely cast to PyObject. It includes a member called user which is of type PyObject *. This can be used for any purpose by handwritten code and will automatically be garbage collected at the appropriate time.
This is a C structure that represents a SIP generated type object. It is an extension of the Python PyTypeObject structure (which is itself an extension of the Python PyObject structure) and so may be safely cast to PyTypeObject (and PyObject).
These functions are deprecated from SIP v4.4.
SIP generates functions for all types being wrapped (including mapped types defined with the %MappedType directive) that convert a Python object to the C structure or C++ class instance. The name of this convertor is the name of the structure or class prefixed by sipForceConvertTo_.
SIP also generates functions for mapped types that convert a C structure or C++ class instance to a Python object. The name of this convertor is the name of the structure or class prefixed by sipConvertFrom_.
The convertor functions of all imported types are available to handwritten code.
SIP generates a type object for each C structure or C++ class being wrapped. These are sipWrapperType structures and are used extensively by the SIP API.
These objects are named with the structure or class name prefixed by sipClass_. For example, the type object for class Klass is sipClass_Klass.
The type objects of all imported classes are available to handwritten code.
SIP generates a type object for each named enum being wrapped. These are PyTypeObject structures. (Anonymous enums are wrapped as Python integers.)
These objects are named with the fully qualified enum name (i.e. including any enclosing scope) prefixed by sipEnum_. For example, the type object for enum Enum defined in class Klass is sipEnum_Klass_Enum.
The type objects of all imported named enums are available to handwritten code.
For most C++ classes being wrapped SIP generates a derived class with the same name prefixed by sip. For example, the derived class for class Klass is sipKlass.
If a C++ class doesn't have any virtual or protected methods in it or any of it's super-class hierarchy, or does not emit any Qt signals, then a derived class is not generated.
Most of the time handwritten code should ignore the derived classes. The only exception is that handwritten constructor code specified using the %MethodCode directive should call the derived class's constructor (which has the same C++ signature) rather then the wrapped class's constructor.
SIP generates a Python object for each exception defined with the %Exception_ directive.
These objects are named with the fully qualified exception name (i.e. including any enclosing scope) prefixed by sipException_. For example, the type object for enum Except defined in class Klass is sipException_Klass_Except.
The objects of all imported exceptions are available to handwritten code.
The main purpose of the SIP module is to provide functionality common to all SIP generated bindings. It is loaded automatically and most of the time you will completely ignore it. However, it does expose some functionality that can be used by applications.
If the bindings have been created with SIP's -r command line option then the generated code will produce debugging statements that trace the execution of the code. (It is particularly useful when trying to understand the operation of a C++ library's virtual function calls.)
Debugging statements are generated at the following points:
By default the trace mask is zero and all debugging statements are disabled.
This function is deprecated from SIP v4.3. If direction is non-zero then the equivalent call is:
sip.transferto(obj, None)
If direction is zero then the equivalent call is:
sip.transferback(obj)
This is the type object for the type SIP uses to represent a C/C++ void *. The type constructor takes a single argument that must either be another voidptr, None, a Python CObject, or an integer. The type has the following methods:
The purpose of the build system is to make it easy for you to write configuration scripts in Python for your own bindings. The build system takes care of the details of particular combinations of platform and compiler. It supports over 50 different platform/compiler combinations.
The build system is implemented as a pure Python module called sipconfig that contains a number of classes and functions. Using this module you can write bespoke configuration scripts (e.g. PyQt's configure.py) or use it with other Python based build systems (e.g. Distutils and SCons).
An important feature of SIP is the ability to generate bindings that are built on top of existing bindings. For example, both PyKDE and PyQwt are built on top of PyQt but all three packages are maintained by different developers. To make this easier PyQt includes its own configuration module, pyqtconfig, that contains additional classes intended to be used by the configuration scripts of bindings built on top of PyQt. The SIP build system includes facilities that do a lot of the work of creating these additional configuration modules.
This creates a configuration module (e.g. pyqtconfig) from a template file and a string.
module is the name of the configuration module file to create.
template is the name of the template file.
content is a string which replaces every occurence of the pattern @SIP_CONFIGURATION@ in the template file. The content string is usually created from a Python dictionary using sipconfig.create_content(). content may also be a dictionary, in which case sipconfig.create_content() is automatically called to convert it to a string.
macros is an optional dictionary of platform specific build macros. It is only used if sipconfig.create_content() is called automatically to convert a content dictionary to a string.
This converts a Python dictionary to a string that can be parsed by the Python interpreter and converted back to an equivalent dictionary. It is typically used to generate the content string for sipconfig.create_config_module().
dict is the Python dictionary to convert.
macros is the optional dictionary of platform specific build macros.
Returns the dictionary as a string.
This creates a platform dependent executable wrapper around a Python script.
script is the full pathname of the script.
wrapper is the pathname of the wrapper to create.
gui is non-zero if a GUI enabled version of the interpreter should be used on platforms that require it.
Returns the platform specific name of the wrapper.
This displays an error message on stderr and calls sys.exit() with a value of 1.
msg is the text of the message and should not include any newline characters.
This formats a message by inserting newline characters at appropriate places.
msg is the text of the message and should not include any newline characters.
leftmargin is the optional position of the left margin.
rightmargin is the optional position of the right margin.
This displays an information message on stdout.
msg is the text of the message and should not include any newline characters.
This parses a qmake compatible file of build system macros and converts it to a dictionary. A macro is a name/value pair. The dictionary is returned or None if any of the overrides was invalid.
filename is the name of the file to parse.
names is a list of the macro names to extract from the file.
overrides is an optional list of macro names and values that modify those found in the file. They are of the form name=value (in which case the value replaces the value found in the file) or name+=value (in which case the value is appended to the value found in the file).
properties is an optional dictionary of property name and values that are used to resolve any expressions of the form $[name] in the file.
This extracts version information for a package from a file, usually a C or C++ header file. The version information must each be specified as a #define of a numeric (hexadecimal or decimal) value and/or a string value.
filename is the name of the file to read.
description is a descriptive name of the package used in error messages.
numdefine is the optional name of the #define of the version as a number. If it is None then the numeric version is ignored.
strdefine is the optional name of the #define of the version as a string. If it is None then the string version is ignored.
Returns a tuple of the numeric and string versions. sipconfig.error() is called if either were required but could not be found.
This converts a version number to a SIP version tag. SIP uses the %Timeline directive to define the chronology of the different versions of the C/C++ library being wrapped. Typically it is not necessary to define a version tag for every version of the library, but only for those versions that affect the library's API as SIP sees it.
version is the numeric version number of the C/C++ library being wrapped. If it is negative then the latest version is assumed. (This is typically useful if a snapshot is indicated by a negative version number.)
tags is the dictionary of SIP version tags keyed by the corresponding C/C++ library version number. The tag used is the one with the smallest key (i.e. earliest version) that is greater than version.
description is a descriptive name of the C/C++ library used in error messages.
Returns the SIP version tag. sipconfig.error() is called if the C/C++ library version number did not correspond to a SIP version tag.
This converts a 3 part version number encoded as a hexadecimal value to a string.
v is the version number.
Returns a string.
This class encapsulates configuration values that can be accessed as instance objects. A sub-class may provide a dictionary of additional configuration values in its constructor the elements of which will have precedence over the super-class's values.
The following configuration values are provided:
- default_bin_dir
- The name of the directory where executables should be installed by default.
- default_mod_dir
- The name of the directory where SIP generated modules should be installed by default.
- default_sip_dir
- The name of the base directory where the .sip files for SIP generated modules should be installed by default. A sub-directory with the same name as the module should be created and its .sip files should be installed in the sub-directory. The .sip files only need to be installed if you might want to build other bindings based on them.
- platform
- The name of the platform/compiler for which the build system has been configured for.
- py_conf_inc_dir
- The name of the directory containing the pyconfig.h header file.
- py_inc_dir
- The name of the directory containing the Python.h header file.
- py_lib_dir
- The name of the directory containing the Python interpreter library.
- py_version
- The Python version as a 3 part hexadecimal number (e.g. v2.3.3 is represented as 0x020303).
- sip_bin
- The full pathname of the SIP executable.
- sip_config_args
- The command line passed to configure.py when SIP was configured.
- sip_inc_dir
- The name of the directory containing the sip.h header file.
- sip_mod_dir
- The name of the directory containing the SIP module.
- sip_version
- The SIP version as a 3 part hexadecimal number (e.g. v4.0.0 is represented as 0x040000).
- sip_version_str
- The SIP version as a string. For development snapshots it will start with snapshot-.
- universal
- The name of the MacOS/X SDK used when creating universal binaries.
Initialise the instance.
sub_cfg is an optional list of sub-class configurations. It should only be used by the __init__() method of a sub-class to append its own dictionary of configuration values before passing the list to its super-class.
This class encapsulates a Makefile. It is intended to be sub-classed to generate Makefiles for particular purposes. It handles all platform and compiler specific flags, but allows them to be adjusted to suit the requirements of a particular module or program. These are defined using a number of macros which can be accessed as instance objects.
The following instance objects are provided to help in fine tuning the generated Makefile:
- chkdir
- A string that will check for the existence of a directory.
- config
- A reference to the configuration argument that was passed to the constructor.
- console
- A reference to the console argument that was passed to the constructor.
- copy
- A string that will copy a file.
- extra_cflags
- A list of additional flags passed to the C compiler.
- extra_cxxflags
- A list of additional flags passed to the C++ compiler.
- extra_defines
- A list of additional macro names passed to the C/C++ preprocessor.
- extra_include_dirs
- A list of additional include directories passed to the C/C++ preprocessor.
- extra_lflags
- A list of additional flags passed to the linker.
- extra_lib_dirs
- A list of additional library directories passed to the linker.
- extra_libs
- A list of additional libraries passed to the linker. The names of the libraries must be in platform neutral form (i.e. without any platform specific prefixes, version numbers or extensions).
- generator
- A string that defines the platform specific style of Makefile. The only supported values are UNIX and something else that is not UNIX.
- mkdir
- A string that will create a directory.
- rm
- A string that will remove a file.
Initialise the instance.
configuration is the current configuration and is an instance of the Configuration class or a sub-class.
console is set if the target is a console (rather than GUI) target. This only affects Windows and is ignored on other platforms.
qt is set if the target uses Qt. For Qt v4 a list of Qt libraries may be specified and a simple non-zero value implies QtCore and QtGui.
opengl is set if the target uses OpenGL.
python is set if the target uses Python.h.
threaded is set if the target requires thread support. It is set automatically if the target uses Qt and Qt has thread support enabled.
warnings is set if compiler warning messages should be enabled. The default of None means that warnings are enabled for SIP v4.x and disabled for SIP v3.x.
debug is set if debugging symbols should be generated.
dir is the name of the directory where build files are read from and Makefiles are written to. The default of None means the current directory is used.
makefile is the name of the generated Makefile.
installs is a list of extra install targets. Each element is a two part list, the first of which is the source and the second is the destination. If the source is another list then it is a list of source files and the destination is a directory.
universal is the name of the SDK if universal binaries are to be created under MacOS/X.
This generates the Makefile commands that will remove any files generated during the build of the default target.
mfile is the Python file object of the Makefile.
build is the dictionary created from parsing the build file.
This is the default implementation of the Makefile macros and rules generation.
mfile is the Python file object of the Makefile.
This is the default implementation of the Makefile clean target generation.
mfile is the Python file object of the Makefile.
This is the default implementation of the Makefile default target generation.
mfile is the Python file object of the Makefile.
This is the default implementation of the Makefile install target generation.
mfile is the Python file object of the Makefile.
This generates the Makefile commands to install one or more files to a directory.
mfile is the Python file object of the Makefile.
src is the name of a single file to install or a list of a number of files to install.
dst is the name of the destination directory.
strip is set if the files should be stripped of unneeded symbols after having been installed.
This returns an optional Makefile macro as a list.
name is the name of the macro.
Returns the macro as a list.
This returns an optional Makefile macro as a string.
name is the name of the macro.
default is the optional default value of the macro.
Returns the macro as a string.
This parses a build file (created with the -b SIP command line option) and converts it to a dictionary. It can also validate an existing dictionary created through other means.
filename is the name of the build file, or is a dictionary to be validated. A valid dictionary will contain the name of the target to build (excluding any platform specific extension) keyed by target; the names of all source files keyed by sources; and, optionally, the names of all header files keyed by headers.
Returns a dictionary corresponding to the parsed build file.
This converts a library name to a platform specific form.
clib is the name of the library in cannonical form.
framework is set if the library is implemented as a MacOS framework.
Return the platform specific name.
This returns a required Makefile macro as a string.
name is the name of the macro.
Returns the macro as a string. An exception is raised if the macro does not exist or has an empty value.
This class encapsulates a Makefile to build a generic Python extension module.
Initialise the instance.
configuration - see sipconfig.Makefile.__init__().
build_file is the name of the build file. Build files are generated using the -b SIP command line option.
install_dir is the name of the directory where the module will be optionally installed.
static is set if the module should be built as a static library (see Builtin Modules and Custom Interpreters).
console - see sipconfig.Makefile.__init__().
qt - see sipconfig.Makefile.__init__().
opengl - see sipconfig.Makefile.__init__().
threaded - see sipconfig.Makefile.__init__().
warnings - see sipconfig.Makefile.__init__().
debug - see sipconfig.Makefile.__init__().
dir - see sipconfig.Makefile.__init__().
makefile - see sipconfig.Makefile.__init__().
installs - see sipconfig.Makefile.__init__().
strip is set if the module should be stripped of unneeded symbols after installation. It is ignored if either debug or static is set, or if the platform doesn't support it.
export_all is set if all of the module's symbols should be exported rather than just the module's initialisation function. Exporting all symbols increases the size of the module and slows down module load times but may avoid problems with modules that use C++ exceptions. All symbols are exported if either debug or static is set, or if the platform doesn't support it.
This returns the name of a SIP v3.x module for when it is used as a library to be linked against. An exception will be raised if it is used with SIP v4.x modules.
mname is the name of the module.
Returns the corresponding library name.
This class encapsulates a Makefile that sits above a number of other Makefiles in sub-directories.
Initialise the instance.
configuration - see sipconfig.Makefile.__init__().
subdirs is the sequence of sub-directories.
dir - see sipconfig.Makefile.__init__().
makefile - see sipconfig.Makefile.__init__().
installs - see sipconfig.Makefile.__init__().
This class encapsulates a Makefile to build an executable program.
Initialise the instance.
configuration - see sipconfig.Makefile.__init__().
build_file is the name of the optional build file. Build files are generated using the -b SIP command line option.
install_dir is the name of the directory where the executable program will be optionally installed.
console - see sipconfig.Makefile.__init__().
qt - see sipconfig.Makefile.__init__().
opengl - see sipconfig.Makefile.__init__().
python - see sipconfig.Makefile.__init__().
threaded - see sipconfig.Makefile.__init__().
warnings - see sipconfig.Makefile.__init__().
debug - see sipconfig.Makefile.__init__().
dir - see sipconfig.Makefile.__init__().
makefile - see sipconfig.Makefile.__init__().
installs - see sipconfig.Makefile.__init__().
This creates a single command line that will create an executable program from a single source file.
source is the name of the source file.
Returns a tuple of the name of the executable that will be created and the command line.
This class encapsulates a Makefile that installs a pure Python module.
Initialise the instance.
configuration - see sipconfig.Makefile.__init__().
dstdir is the name of the directory in which the module's Python code will be installed.
srcdir is the name of the directory (relative to dir) containing the module's Python code. It defaults to the same directory.
dir - see sipconfig.Makefile.__init__().
makefile - see sipconfig.Makefile.__init__().
installs - see sipconfig.Makefile.__init__().
This class encapsulates a Makefile to build a SIP generated Python extension module.
To build the example in A Simple C++ Example using distutils, it is sufficient to create a standard setup.py, listing word.sip among the files to build, and hook-up SIP into distutils:
from distutils.core import setup, Extension import sipdistutils setup( name = 'word', versione = '1.0', ext_modules=[ Extension("word", ["word.sip", "word.cpp"]), ], cmdclass = {'build_ext': sipdistutils.build_ext} )
As we can see, the above is a normal distutils setup script, with just a special line which is needed so that SIP can see and process word.sip. Then, running setup.py build will build our extension module.
Sometimes you want to create a custom Python interpreter with some modules built in to the interpreter itself rather than being dynamically loaded. To do this the module must be created as a static library and linked with a custom stub and the normal Python library.
To build the SIP module as a static library you must pass the -k command line option to configure.py. You should then build and install SIP as normal. (Note that, because the module is now a static library, you will not be able to import it.)
To build a module you have created for your own library you must modify your own configuration script to pass a non-zero value as the static argument of the __init__() method of the ModuleMakefile class (or any derived class you have created). Normally you would make this configurable using a command line option in the same way that SIP's configure.py handles it.
The next stage is to create a custom stub and a Makefile. The SIP distribution contains a directory called custom which contains example stubs and a Python script that will create a correct Makefile. Note that, if your copy of SIP was part of a standard Linux distribution, the custom directory may not be installed on your system.
The custom directory contains the following files. They are provided as examples - each needs to be modified according to your particular requirements.
- mkcustom.py is a Python script that will create a Makefile which is then used to build the custom interpreter. Comments in the file describe how it should be modified.
- custom.c is a stub for a custom interpreter on Linux/UNIX. It should also be used for a custom console interpreter on Windows (i.e. like python.exe). Comments in the file describe how it should be modified.
- customw.c is a stub for a custom GUI interpreter on Windows (i.e. like pythonw.exe). Comments in the file describe how it should be modified.
Note that this technique does not restrict how the interpreter can be used. For example, it still allows users to write their own applications that can import your builtin modules. If you want to prevent users from doing that, perhaps to protect a proprietary API, then take a look at the VendorID package.