aboutsummaryrefslogtreecommitdiff
path: root/CHANGES.current
diff options
context:
space:
mode:
authorWilliam S Fulton <wsf@fultondesigns.co.uk>2013-09-17 23:43:30 +0100
committerWilliam S Fulton <wsf@fultondesigns.co.uk>2013-09-18 00:40:17 +0100
commit37bccfd517dfebb4d191a144975d04c7ebfdea12 (patch)
tree59cfc2c193313631c8b1c107df6d2e6c444ab0f7 /CHANGES.current
parentb0b8d4c87e0077baccf770441768f5ad3b15dcd3 (diff)
downloadswig-37bccfd517dfebb4d191a144975d04c7ebfdea12.tar.gz
Bump version to 3.0.0
SWIG 3 is open for business - looks go crazy!
Diffstat (limited to 'CHANGES.current')
-rw-r--r--CHANGES.current136
1 files changed, 1 insertions, 135 deletions
diff --git a/CHANGES.current b/CHANGES.current
index 1727ecb2c..ddf123364 100644
--- a/CHANGES.current
+++ b/CHANGES.current
@@ -2,140 +2,6 @@ Below are the changes for the current release.
See the CHANGES file for changes in older releases.
See the RELEASENOTES file for a summary of changes in each release.
-Version 2.0.11 (15 Sep 2013)
+Version 3.0.0 (in progress)
============================
-2013-09-15: wsfulton
- [R] Fix attempt to free a non-heap object in OUTPUT typemaps for:
- unsigned short *OUTPUT
- unsigned long *OUTPUT
- signed long long *OUTPUT
- char *OUTPUT
- signed char*OUTPUT
- unsigned char*OUTPUT
-
-2013-09-12: wsfulton
- [Lua] Pull Git patch #62.
- 1) Static members and static functions inside class can be accessed as
- ModuleName.ClassName.FunctionName (MemberName respectively). Old way such as
- ModuleName.ClassName_FunctionName still works.
- 2) Same goes for enums inside classes: ModuleName.ClassName.EnumValue1 etc.
-
-2013-09-12: wsfulton
- [UTL] Infinity is now by default an acceptable value for type 'float'. This fix makes
- the handling of type 'float' and 'double' the same. The implementation requires the
- C99 isfinite() macro, or otherwise some platform dependent equivalents, to be available.
-
- Users requiring the old behaviour of not accepting infinity, can define a 'check' typemap
- wherever a float is used, such as:
-
- %typemap(check,fragment="<float.h>") float, const float & %{
- if ($1 < -FLT_MAX || $1 > FLT_MAX) {
- SWIG_exception_fail(SWIG_TypeError, "Overflow in type float");
- }
- %}
-
- *** POTENTIAL INCOMPATIBILITY ***
-
-2013-08-30: wsfulton
- [Lua] Pull Git patch #81: Include Lua error locus in SWIG error messages.
- This is standard information in Lua error messages, and makes it much
- easier to find bugs.
-
-2013-08-29: wsfulton
- Pull Git patch #75: Handle UTF-8 files with BOM at beginning of file. Was giving an
- 'Illegal token' syntax error.
-
-2013-08-29: wsfulton
- [C#] Pull Git patch #77: Allow exporting std::map using non-default comparison function.
-
-2013-08-28: wsfulton
- [Python] %implicitconv is improved for overloaded functions. Like in C++, the methods
- with the actual types are considered before trying implicit conversions. Example:
-
- %implicitconv A;
- struct A {
- A(int i);
- };
- class CCC {
- public:
- int xx(int i) { return 11; }
- int xx(const A& i) { return 22; }
- };
-
- The following python code:
-
- CCC().xx(-1)
-
- will now return 11 instead of 22 - the implicit conversion is not done.
-
-2013-08-23: olly
- [Python] Fix clang++ warning in generated wrapper code.
-
-2013-08-16: wsfulton
- [Python] %implicitconv will now accept None where the implicit conversion takes a C/C++ pointer.
- Problem highlighted by Bo Peng. Closes SF patch #230.
-
-2013-08-07: wsfulton
- [Python] SF Patch #326 from Kris Thielemans - Remove SwigPyObject_print and SwigPyObject_str and
- make the generated wrapper use the default python implementations, which will fall back to repr
- (for -builtin option).
-
- Advantages:
- - it avoids the swig user having to jump through hoops to get print to work as expected when
- redefining repr/str slots.
- - typing the name of a variable on the python prompt now prints the result of a (possibly redefined)
- repr, without the swig user having to do any extra work.
- - when redefining repr, the swig user doesn't necessarily have to redefine str as it will call the
- redefined repr
- - the behaviour is exactly the same as without the -builtin option while requiring no extra work
- by the user (aside from adding the %feature("python:slot...) statements of course)
-
- Disadvantage:
- - default str() will give different (but clearer?) output on swigged classes
-
-2013-07-30: wsfulton
- [Python, Ruby] Fix #64 #65: Missing code in std::multimap wrappers. Previously an instantiation
- of a std::map was erroneously required in addition to an instantiation of std::multimap with the
- same template parameters to prevent compilation errors for the wrappers of a std::multimap.
-
-2013-07-14: joequant
- [R] Change types file to allow for SEXP return values
-
-2013-07-05: wsfulton
- [Python] Add %pythonbegin directive which works like %pythoncode, except the specified code is
- added at the beginning of the generated .py file. This is primarily needed for importing from
- __future__ statements required to be at the very beginning of the file. Example:
-
- %pythonbegin %{
- from __future__ import print_function
- print("Loading", "Whizz", "Bang", sep=' ... ')
- %}
-
-2013-07-01: wsfulton
- [Python] Apply SF patch #340 - Uninitialized variable fix in SWIG_Python_NonDynamicSetAttr
- when using -builtin.
-
-2013-07-01: wsfulton
- [Python, Ruby, Ocaml] Apply SF patch #341 - fix a const_cast in generated code that was generating
- a <:: digraph when using the unary scope operator (::) (global scope) in a template type.
-
-2013-07-01: wsfulton
- [Python] Add SF patch #342 from Christian Delbaere to fix some director classes crashing on
- object deletion when using -builtin. Fixes SF bug #1301.
-
-2013-06-11: wsfulton
- [Python] Add SWIG_PYTHON_INTERPRETER_NO_DEBUG macro which can be defined to use the Release version
- of the Python interpreter in Debug builds of the wrappers. The Visual Studio .dsp example
- files have been modified to use this so that Debug builds will now work without having
- to install or build a Debug build of the interpreter.
-
-2013-06-07: wsfulton
- [Ruby] Git issue #52. Fix regression with missing rb_complex_new function for Ruby
- versions prior to 1.9 using std::complex wrappers if just using std::complex as an output type.
- Also fix the Complex helper functions external visibility (to static by default).
-
-2013-06-04: olly
- [PHP] Fix SWIG_ZTS_ConvertResourcePtr() not to dereference NULL
- if the type lookup fails.
-