std::wstring parameters on SWIG directors were being truncated to a single character due to marshalling inconsistencies.
This patch applies the same MarshalAs(UnmanagedType.LPWStr) attribute to director delegate methods and allows std::wstrings to be successfully received in the C# code via director calls.
Initial contributions for Linux provided in issue #1233, modified to work
on both Windows and Linux. Dual support is possible by detecting
the sizeof wchar_t which is different on each of these systems.
This exercises a bug that was found with this fairly complex mapping, it
wasn't putting newlines in the proper place. A previous commit added
the newlines, this makes sure it doesn't happen again.
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Shows how to use a go directory for passing an NULL terminated argv type
array.
This also exercises the bug in the previous commit. That's why the
call1 function isn't assigned to zero, that the only case where that bug
happens.
Signed-off-by: Corey Minyard <cminyard@mvista.com>
In some cases the \n was missing from goin directives in some places.
Without this change, you will get
} swig_r = (int)(C....
in the output in some cases, and that won't compile in Go, it will
error with:
../gensio/gensio.go:1508:3: expected ';', found swig_r
It only did this on director callbacks with godirectorin, and only if
the virtual function wasn't assigned to zero. So the following will
fail:
class callbacks {
public:
virtual bool call1(int v, const char * const *strarray);
virtual ~callbacks() = default;
};
but the following won't:
class callbacks {
public:
virtual bool call1(int v, const char * const *strarray) = 0;
virtual ~callbacks() = default;
};
Add "\n" to the proper places to fix this.
Ruby 3.2 (still in development) started warning about this, see
https://bugs.ruby-lang.org/issues/18007.
Note that the extension docs mention it in
https://ruby-doc.org/core-3.1.1/doc/extension_rdoc.html#label-C+struct+to+Ruby+object:
> The klass argument is the class for the object. The klass should
> derive from `rb_cObject`, and the allocator must be set by calling
> `rb_define_alloc_func` or `rb_undef_alloc_func`.
So in this case we call `rb_undef_alloc_func` to fix the warning.
Verified that this works in robinst/taglib-ruby#115.
Fixes#2257.
Found via `codespell -q 3 -L ans,anumber,ba,bae,chello,clos,cmo,coo,dout,fo,funktion,goin,inout,methid,nd,nin,nnumber,object,objekt,od,ois,packag,parm,parms,pres,statics,strack,struc,tempdate,te,thru,uint,upto,writen`
This appears to have been accidentally added in
d8cc75946b as part of the changes to
support C++11 rvalue references, presumably to match the existing
correct rule for unary &.
There is a non-standard GCC extension for taking the address of a label
using unary &&:
https://gcc.gnu.org/onlinedocs/gcc/Labels-as-Values.html
However this extension only works in an expression in a function body,
and SWIG doesn't parse the contents of function bodies so this grammar
rule isn't even providing any useful support for this extension.
./configure now enables C++11 and later C++ standards testing by default (when
running: 'make check').
The options to control this testing are the same:
./configure --enable-cpp11-testing
./configure --disable-cpp11-testing
But the former is now the default and the latter can be used to turn off C++11 and
later C++ standards testing.
Reduce the number of tests on Appveyor by only running 32-bit (x86) using
latest Visual Studio (2022). Drop 32-bit Cygwin and MinGW testing.
Add C# and Java testing on VS2022 and drop Java testing on older Visual Studio.
This reduces the really long run times on Appveyor, however, VS2022 is
somewhat slower than older compilers.
If a "docstring" feature is present it will still override a Doxygen comment.
If the "autodoc" feature is also present, the combined "autodoc" and "docstring"
will override the Doxygen comment. If no "docstring" is present then the
"autodoc" feature will not be generated when there is a Doxygen comment.
This way the "autodoc" feature can be specified and used to provide documentation
for 'missing' Doxygen comments.
Closes#1635
This option was undocumented. If used it quietly did nothing
unless you'd installed the SWILL library before building SWIG, but
SWILL is no longer maintained and hasn't seen a release since
2008-04-10.
It's clear nobody has used this functionality for some time as the
code to support it in SWIG isn't actually valid ISO C++:
Modules/browser.cxx: In member function ‘void Browser::show_attributes(Node*)’:
Modules/browser.cxx:57:23: warning: ISO C++ forbids converting a string constant to ‘char*’ [-Wwrite-strings]
57 | char *trunc = "";
| ^~
Modules/browser.cxx:61:21: warning: ISO C++ forbids converting a string constant to ‘char*’ [-Wwrite-strings]
61 | trunc = "...";
| ^~~~~
Closes#2251
MzScheme headers have a problem when using -Werror that the test-suite uses:
/usr/include/racket/sconfig.h:126:43: error: invalid suffix on literal; C++11 requires a space between literal and string macro [-Werror=literal-suffix]
126 | # define SCHEME_PLATFORM_LIBRARY_SUBPATH "x86_64-"SPLS_LINUX
Move HAVE_CXX11 into makefiles so that running test-suite
from top level directory or in the language's test-suite directory
is consistent. For example, running 'make check-java-test-suite'
behaves the same as 'cd Examples/test-suite/java && make check'.
Problem here seems to be one also affecting other ocaml broken tests:
1. Enum value names should be using sym:name not name (ie %rename is broken for ocaml enum items)
2. directorin typemaps are not correct
