#682 [java] bad code generation for typedef'd enums

[Steven K. Sharp]

The code generated for
typedef enum
{
blah = 1,
bleh
} baz;

is incorrect. In a call that takes a baz as an
argument results in "enum baz" in the cast instead of
just "baz".

Attached is an example that demonstrates the bug.

[Steven K. Sharp]

include and i files for enum bug.

[Steven K. Sharp]

Logged In: YES
user_id=1434579

This bug is using the trunk (pre-1.3.30?) version of SWIG.

[Olly Betts]

Setting category to Java, and assigning to William as the SWIG/Java maintainer.

Reproducible with trunk shortly after SWIG 2.0.0.

Needs to be processed with: swig -java -c++

Problematic line is now:

foo2::bar2::enumbugexample((enum foo::bar::baz const &)*arg1);

Removing "enum" allows this to compile.

[ganc]

Essentially the same problem is still present in SWIG 3.0.2, swig -csharp -c++

typedef enum { RED = 1, BLUE, GREEN } MyEnum;

There is an extra "enum" before the name on some but not all occasions, resulting in casting errors. Removing the keyword allows compilation.

[Olly Betts]

Still broken with git master with the same bad code generated. Testing with:

../preinst-swig -c++ -java enumbug2.i
g++ -fPIC -shared -I"/usr/lib/jvm/java-1.10.0-openjdk-amd64/include" -I"/usr/lib/jvm/java-1.10.0-openjdk-amd64/include/linux" enumbug2_wrap.cxx enumbug2.cpp -o libenumbug2.so

This is not the cause of the problem, but I noticed enumbug2.cpp in the supplied testcase needs a fix - change the second cout here to endl:

        std::cout << "woof is " << woof << std::cout;

[Olly Betts]

Still reproducible with current git master. Also I came up with a much simpler reproducer:

%module x
%inline %{
typedef enum { BAZ1 } baz;
void enumbugexample( baz const & ) { }
%}

Then:

$ swig -c++ -python simplified.i
$ grep 'enum baz' simplified_wrap.cxx
  enumbugexample((enum baz const &)*arg1);
static swig_type_info _swigt__p_baz = {"_p_baz", "baz *|enum baz *", 0, 0, (void*)0, 0};

The const & seems to be needed.

I notice the type info entry includes enum baz * which seems wrong too.

[Olly Betts]

Looks like https://github.com/swig/swig/issues/197 may be the same issue.

[Olly Betts]

Closing in favour of https://github.com/swig/swig/issues/3012