resolved one if is needed.
Add examples showing the problem with SWIG_TypeQuery
and template+typedef and the old type str, and how
it works now.
git-svn-id: https://swig.svn.sourceforge.net/svnroot/swig/trunk@5704 626c5289-ae23-0410-ae9c-e8d60b6d4f22
Add the macros
%typemap_asfromcheck()
%typemap_asfrom()
that can be used to defined all the different typemaps
for types where the As/From/Check methods are provided.
All the basic type (int, char,...) typemaps are implemented
using them.
The std::string and std::complex<T> are reimplemented using
the new %typemap_asfrom/check macros too. This helps to complete
all the previously missing typemaps (consttab, varin, varout,..)
and also ilustrates how to define the As/From/Check methods
to use with the %typemap_asfrom/check macros.
As a byproduct, the C complex typemap was added, and the file
complex.i
can be used to load the complex support for either C or C++. The
original C++ std_complex.i is still there, and the corresponding C
ccomplex.i too, if they need to be included explicitly.
Also, the As/From methods are declared via %fragment, so, they
can be reused as needed, and only appear in the wrapped code if
they corresponding typemap is invoked, making the typemaps
and the entire code shorter and simpler.
Marcelo.
git-svn-id: https://swig.svn.sourceforge.net/svnroot/swig/trunk@5691 626c5289-ae23-0410-ae9c-e8d60b6d4f22
%fragment("Hello","header") "..."
%fragment("Hi","header",fragment="Hello") "..."
the latter fragment will include the first one if is invoked.
more than one fragment can be included at the time, just
keep adding fragment="f1",fragment="f2", etc.
this is used to emulate typemaps reuse, where all the
reusable typemap code is put in a fragment static method,
and then it can be included from another fragment typemap
as needed.
git-svn-id: https://swig.svn.sourceforge.net/svnroot/swig/trunk@5690 626c5289-ae23-0410-ae9c-e8d60b6d4f22
they only accept PyStrings.
now the raw pointer from calloc.i, for example, works
fine.
some of the SWIG_PyObj_(As|From)XXX routines are getting
too large, so, they also will go into the runtime library.
git-svn-id: https://swig.svn.sourceforge.net/svnroot/swig/trunk@5689 626c5289-ae23-0410-ae9c-e8d60b6d4f22
null chars, which are allowed in C++ and python, and proper
sizes:
you can use a shorter string to assign a char[6]:
x = "he\0l" -> x := {'h','e',0,'l',0, 0}
but you always get back in python the original size:
x := {'h','i','0','0',0, 0} -> "hi\0\0\0\0"
Better mapping of (char*)0, now you get a None in the python
side, and using None, you get a (char*)0, just like
all the other C++ pointers. Before (char*)0 was mapped
into python as "", or as seg. faults.
Better mapping of "", which is not None nor (char*)0.
Added more tests using primitive types with simple templates.
git-svn-id: https://swig.svn.sourceforge.net/svnroot/swig/trunk@5687 626c5289-ae23-0410-ae9c-e8d60b6d4f22
typedef char namet[5];
const namet def_namet = {'h','o', 0, 'l','a'};
ie, char arrays with null embedded characters, and other odd
char[ANY] situations.
add the corresponding test to the primitive_types.i and runme files.
Marcelo
----------------------------------------------------------------------
git-svn-id: https://swig.svn.sourceforge.net/svnroot/swig/trunk@5685 626c5289-ae23-0410-ae9c-e8d60b6d4f22
(throw typemap).
- fix the examples with exception problems and warnings
- proper and consist treatment of basic types (short, unsigned char,...)
now all are checked for range and sign. Before, this was depending of
the use of parsing or no parsing, and/or the converter method, and/or
the use of directors, etc.
git-svn-id: https://swig.svn.sourceforge.net/svnroot/swig/trunk@5659 626c5289-ae23-0410-ae9c-e8d60b6d4f22
WARN_PARSE_REDUNDANT 322
similar to the g++ -Wredundant-decls flag.
This recovers the warnings that now are not been reported by
the original code
WARN_PARSE_REDEFINED 302
Redundant example:
int foo(int);
int foo(int);
Redefined example:
int foo(int);
double foo(int);
git-svn-id: https://swig.svn.sourceforge.net/svnroot/swig/trunk@5634 626c5289-ae23-0410-ae9c-e8d60b6d4f22
Now, this doesn't generate warnings:
class A; class B*;
int foo(A*, B*);
struct A {
friend int foo(A*, B*);
};
but this generates
struct B {
friend double foo(A*, B*);
};
which is correct, since they have different return types.
See the redefined.i file for most cases where
repeated (but not redefined) declarations are
not generating warnings now, such as:
// no warning
#define REPEATED 1
#define REPEATED 1
// yes warning
#define REDEFINED 1
#define REDEFINED 2
this is following the C, C++ and preprocessor
standard behavior.
The function 'need_redefined_warn(..)' was added to
util.c, if some strange corner appears, and therefore,
the parser.y file doesn't need to be changed latter.
Also, the redefined warning format in parser.y
was changed, so now it respond to William's -Fmicrosoft
flag.
git-svn-id: https://swig.svn.sourceforge.net/svnroot/swig/trunk@5633 626c5289-ae23-0410-ae9c-e8d60b6d4f22
