Don't populate the typenames array for each and every created object and don't
search it when deleting every object, this is O(N) in number of classes and is
completely impractical for any big library where N can easily be several
thousands. Use destructor function which will correctly destroy the object
instead.
Also don't store each created objects in the global registry, there doesn't
seem to be any point in it.
And, in fact, don't bother with the typenames at all, just use another
pseudo-virtual function for checking whether a class inherits from a class
with the given name.
There is unfortunately one problem with the new approach: it doesn't work when
the same C++ type is wrapped under different names as this results in multiple
specializations of SWIG_derives_from<> for this type. But this is a relatively
rare situation (but which does arise in template_default2 unit test, which had
to be disabled) and could be fixed in the future by completely resolving the
type, including the default template parameters values, and checking if
SWIG_derives_from had been already specialized for it. In the meanwhile, this
regression is not a big deal compared to the advantages of the new approach.
The proxy layer, and all the extra complexity associated with it, seemed to be
only necessary in order to try to allow using the same name for the wrapped
global functions as were used for them in the original C or C++ code being
wrapped. However this could simply never work in all cases, notably it didn't
work at all when using ELF shared libraries under Unix as the functions with
the same name defined in the main program were interposed and replaced the
functions defined in the shared library, meaning that the proxy function foo()
called wrapper function _wrap_foo() which called back into proxy function
foo() itself again, resulting in guaranteed stack overflow. The only possible
way to fix this would be to use "protected" ELF visibility for the original
functions, but this is not always possible (e.g. if the sources of the
original library are not available) and not simple even when it is and,
besides, protected visibility has its own problems -- notably by making it
impossible to hook the library functions when you actually want to do it.
Besides, proxy-based approach simply couldn't work at all when using static
linking as it resulted in two copies of the function with the same name
Most importantly, however, the main task of this module is to wrap C++
classes, not C functions, and renaming them in the wrapper is not necessary at
all in this case as there is no conflict with the original names in this case.
So simply drop the idea of generating a separate proxy header and generate a
header declaring the functions declared in the wrapper instead and, also, do
not give them "_wrap_" prefix whenever possible, i.e. only do it for the
global functions.
This simplifies SWIG code itself and makes it simpler to use its output as
it's not necessary to link both with the wrapper (dynamically) and with the
proxy (statically) and it's not enough to link with the wrapper only and it
can be done in any way (i.e. either statically or dynamically).
As a side effect of this change, Swig_name_proxy() is not necessary any more
and was removed, eliminating the only difference with the master branch in any
source file other than c.cxx itself.
* char-escaping:
Add missing string_constant.i testcase
changes file update for char wrappers
C# char wrappers fixes for enum values, static const member char values and %csconst
D testing added for %dmanifestconst and char constants
Fix wrapping D constants using %dmanifestconst
Php fix for enum value of '\0'
Fix static const char member variables wrappers with %javaconst(1).
Expand char testing in enums and %constant
Java char changes file update
Java enum and static member variable escaping fix for chars
Add tests for enum values and static const member variables chars containing escape sequences
Minor documentation tweak
Conflicts:
CHANGES.current
* interfaces:
Documentation and CHANGES entry for interface feature
interface macro argument name tweaks
Change the name of the interface in %feature to be more portable
Interface feature fix for typedef types
Add limited support for %interface_impl and %shared_ptr
Multiple inheritance warning wording tweak
C# "override" fix for "extend" case
Add checks for interface name symbol clashes
interface feature test changes for the tests to pass for all languages
Rename feature_interface.i to swiginterface.i
Re-organization of the interface feature common code
Port Java interface tests to C#
Test %interface
Test %interface_impl
Use rstrip instead of regex encoder in %feature_rename
Add rstrip encoder for use in %rename.
Interface macros: %interface %interface_impl %interface_custom
Add $interfacename family of special variable expansions
Add multiple_inheritance_nspace testcase
Interface name handling improvements and special variable changes
Correct regex example comment
Properly hide unexposed naming functions in naming.c
C++ namespace testing for interface feature
interface feature - SWIG_JAVABODY_PROXY does not need to be overridden
Support namespaces and nspace with the interface feature for C#
Support namespaces and nspace with the interface feature for Java
Cosmetic test case changes
Add Java premature garbage collection prevention parameter (pgcpp) to interface feature
Create javainterfacecode and csinterfacecode typemaps
IntPtr & HandleRef absolute names used
virtual/override fix
Improve interface feature checks
Add another interface test selecting just one base as an interface
Comments added to interface feature implementation and cosmetic code changes
Add overloading tests for interface feature
Refactor multiple inheritance warnings
director generation fixes
interface feature updates for C# latest on master
interfaces branch merge fixes
Remove unnecessary interfaces for concrete classes
cosmetic code formatting changes
Correct interface name attributes that are internal
interface macro changes to support templates
Test non-virtual method in Derived classes
interface tests for a most derived class inheriting the interfaces further up the chain
Rename GetCPtr/getCPtr to SWIGInterfaceUpcast
interface feature support for const ref pointers (used by the STL)
Interface feature support for arrays
More interface feature testing for return values
interface feature support for passing by value
interface feature support for references
Multiple inheritance parameters as pointers testing
Simplify multiple_inheritance_abstract Java runtime test
Warning fixes
Rename test functions in multiple_inheritance_abstract testcase
Formatting fixes in generated code for interface feature
Cosmetic spacing changes in test case
interface feature typemap corrections to handle NULL pointers
interface test added
javadirectorin fix
interface implementation visibility
interface inheritance (2)
interface inheritance (1)
feature:interface ported to Java
propagate non-abstract "interface" base methods (3)
propagate non-abstract "interface" base methods (2)
propagate non-abstract "interface" base methods (1)
namespace support added GetCPtr now returns HandleRef "feature:interface:name" is now mandatory attribute
interfaces (1)
interfaces (1)
Conflicts:
CHANGES.current
Remove duplicate (aside from comment formatting) embed15.i.
Remove references to Python 1.5.
Tested and remarkably still works with Python 2.7, so update
documentation and comments to reflect that.
wkhtmltopdf isn't using a fixed-width font for CSS font-family:monospace.
Nor is it using one for <PRE> <CODE> or <TT> elements.
Add in some Courier fonts for it to use - note that Courier 10 Pitch is
installed on Ubuntu by default. Note these fonts need to be installed on
the system that generates the pdf documentation.
Previously the htmldoc stylesheet was kept in place and the SWIG
stylesheet was prepended to it inline in SWIGDocumentation.html.
Now the SWIG stylesheet has been amended with most of the htmldoc
stylesheet changes and completely replaced after htmldoc is run.
htmldoc does not seem to be generating pdfs properly any more (on Ubuntu
14.04). It has been replaced with wkhtmltopdf which is better as it
supports css and so the patched version of htmldoc with the simple css
support is no longer required. wkhtmldoc does have have a few problems
though which have been addressed in prior commits:
- <H1><a name="X"></a>Text</H1> style links need changing to:
<H1><a name="X">Text</a></H1>
- tabs in <pre> elements should be expanded to 8 spaces by default, but
are expanded to just one space and css expand-tab is not recognised.
The <pre> <tt> <code> elements do not always select a fixed-width font -
try installing a Courier font.
wkhtmltopdf is not expanding tabs within <pre> elements to 8 spaces as it
should. Workaround the problem by converting all tabs to an appropriate
number of spaces.
Corrects position of heading text to be as mentioned in the 4.01
transitional standard, see http://www.w3.org/TR/html4/struct/links.html#h-12.1.1.
For example, changes
<H1><a name="Introduction"></a>2 Introduction</H1>
to
<H1><a name="Introduction">2 Introduction</a></H1>
The changes will convert the old incorrect usage should an html file
using the old approach be added in the future.
Use the CCache.html docs instead of the ccache-swig man page.
The yodl2man and yodl2html tools are no longer used and so SWIG no
longer has a dependency on these packages which were required when
building from git.
Closes#286Closes#128
