SWIG development mailing list. SWIG is a tool for wrapping C++ code making it available from other language like Python, Perl, Java etc

sploving | 2 May 2011 15:54

java typemap question

Hello all,
There is a C++ Function like this:
void get_labels(float64_t** p_labels, int32_t* len){
    *len=4;
    *p_labels=(float64_t*) malloc(sizeof(float64_t)* (*len));

    for (int32_t i=0; i<*len; i++)
       (*p_labels)[i]=get_label(i);
}

I write a typemap file to call it from java like this:
double[] y =new double[4]
get_labels(y); // y={1,2,3,4}
It works well. The typemap file is as the following:
%typemap(jni) (float64_t** p_labels, int32_t* len)        %{jdoubleArray%}
%typemap(jtype) (float64_t** p_labels, int32_t* len)        %{double[]%}
%typemap(jstype) (float64_t** p_labels, int32_t* len)    %{double[]%}

%typemap(in) (float64_t** p_labels, int32_t* len) {
    $1 = (float64_t**) malloc(sizeof(float64_t*));
    $2 = (int32_t*) malloc(sizeof(int32_t));
}

%typemap(argout) (float64_t** p_labels, int32_t* len) {
    float64_t* vec = *$1;
    jdouble *arr;
    int i;
    if (!$input)
        return ;
    arr = JCALL2(GetDoubleArrayElements, jenv, $input, 0);
    if (!arr)
        return;
    for (i=0; i < *$2; i++)
        arr[i] = (jdouble)vec[i];
    JCALL3(

ReleaseDoubleArrayElements, jenv, $input, arr, 0);
}

%typemap(javain) (float64_t** p_labels, int32_t* len) "$javainput".

But if I want to use it like this in Java:

double y[];
y = get_labels();// y should be {1,2,3,4}
then how to change the typemap file? I tried to use: typemap(in, numinputs=0) and typemap(argout), but failed. Could anyone give me some advice? thanks.

体验网易邮箱2G超大附件，轻松发优质大电影、大照片，提速3倍!

------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network 
management toolset available today.  Delivers lowest initial 
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd

------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network 
management toolset available today.  Delivers lowest initial 
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd

headers

Ian Lance Taylor | 2 May 2011 23:14

Re: varargs_typemap test is Python specific

William S Fulton <wsf <at> fultondesigns.co.uk> writes:

> On 30/04/11 07:08, Ian Lance Taylor wrote:
>> The new varargs_typemap test is Python specific and will fail for other
>> languages.  At least, it fails for Go, and causes a testsuite failure.
>> What's the right way to handle this?  Should varargs_typemap move from
>> examples/test-suite/common.mk to examples/test-suite/python/Makefile, or
>> should it somehow be adjusted so that it has a chance of working for
>> other languages?
>>
> When adding new test cases the goal should be to make them language
> neutral. However, this isn't possible with custom typemaps like this
> one, so then it should only be added in the language specific
> Makefile, and then prefix the test case name with python_, so let's
> rename it to python_varargs_typemap.

I made this change.

Ian

------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network 
management toolset available today.  Delivers lowest initial 
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd

headers

SourceForge.net | 6 May 2011 13:50

[ swig-Bugs-3298252 ] Invalid typemaps for raw wchar_t* strings

Bugs item #3298252, was opened at 2011-05-06 11:50
Message generated for change (Tracker Item Submitted) made by 
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3298252&group_id=1645

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: csharp
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: https://www.google.com/accounts ()
Assigned to: William Fulton (wsfulton)
Summary: Invalid typemaps for raw wchar_t* strings

Initial Comment:
Swig generates invalid code for raw wchar_t * strings. While compiling this code VisualStudio C# compiler
produces error:
 Cannot implicitly convert type 'System.IntPtr' to 'string'	

Example:
>> test.i
%include <wchar.i>
%immutable;
static const wchar_t *g_const;

>> testPINVOKE.cs
public static extern IntPtr g_const_get();

>> test.cs

public static string g_const {
    get {
      string ret = testPINVOKE.g_const_get(); // ERROR: Cannot implicitly convert type 'System.IntPtr' to 'string'	
      return ret;
    } 
  }

> D:\dev\swigwin-2.0.3\swig.exe -csharp -c++ test.i

Currently I'm using the following typemaps to workaround this problem
%typemap(imtype, inattributes="[MarshalAs(UnmanagedType.LPWStr)]", out="string",
outattributes="[return: MarshalAs(UnmanagedType.LPWStr)]" ) wchar_t * "string"
%typemap(out) wchar_t * %{ $result = (wchar_t *)SWIG_csharp_wstring_callback((const wchar_t *)$1); %}

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3298252&group_id=1645

------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network 
management toolset available today.  Delivers lowest initial 
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd

headers

William S Fulton | 7 May 2011 01:13

next release and python 3.2

Has anyone tried out python 3.2 and the test-suite? There are reports on 
swig-user that swig-2.0.3 and python-3.2 doesn't work. I think we are 
about ready for a new release, but python 3.2 needs to work before we 
push the next release out.

William

------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network 
management toolset available today.  Delivers lowest initial 
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd

headers

Stefan Zager | 8 May 2011 00:05

Re: next release and python 3.2

On Fri, 6 May 2011, William S Fulton wrote:

> Has anyone tried out python 3.2 and the test-suite? There are reports on
> swig-user that swig-2.0.3 and python-3.2 doesn't work. I think we are
> about ready for a new release, but python 3.2 needs to work before we
> push the next release out.

I have been busy for the last week, but I will start working on this.

Stefan

------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network 
management toolset available today.  Delivers lowest initial 
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd

headers

SourceForge.net | 8 May 2011 10:06

[ swig-Bugs-2380788 ] Bug in SWIG 1.3.35 with std::vector type system?

Bugs item #2380788, was opened at 2008-12-02 18:01
Message generated for change (Comment added) made by szager
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=2380788&group_id=1645

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: code generation (general)
Group: None
>Status: Closed
>Resolution: Duplicate
Priority: 5
Private: No
Submitted By: Ben Webb (benwebb)
Assigned to: Nobody/Anonymous (nobody)
Summary: Bug in SWIG 1.3.35 with std::vector type system?

Initial Comment:
I have encountered a problem with SWIG 1.3.35 and later with a fairly simple usage of std::vector. What
seems to be happening is that the SWIG runtime is getting confused about the type of the std::vector.

For example, running 'make' with the attached Makefile and test.i works (produces no output from Python)
with SWIG 1.3.34, but fails on SWIG 1.3.35, 1.3.36 and SVN r10960 with the error:
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "test.py", line 130, in insert
    def insert(self, *args): return _test.vectorf_insert(self, *args)
NotImplementedError: Wrong number of arguments for overloaded function 'vectorf_insert'.
  Possible C/C++ prototypes are:
    insert(std::vector< float > *,std::vector< float >::iterator,std::vector< float >::value_type
const &)
    insert(std::vector< float > *,std::vector< float >::iterator,std::vector< float
>::size_type,std::vector< float >::value_type const &)

Obviously this makes STL vectors rather hard for us to use in practice!

I ran into the problem with Fedora 10's package of SWIG 1.3.35 (Fedora 9's build of 1.3.33 works just fine)
but confirmed that it also fails for me with 1.3.35 built from your upstream sources. (But ideally a
backport of any fix to 1.3.35 would be useful, so that Fedora can patch their package without having to
upgrade to an SVN snapshot.)

----------------------------------------------------------------------

>Comment By: szager (szager)
Date: 2011-05-08 01:06

Message:
Duplicate of bug 2811549, now fixed.

----------------------------------------------------------------------

Comment By: William Fulton (wsfulton)
Date: 2011-02-17 15:06

Message:
Possibly related to #2811549

----------------------------------------------------------------------

Comment By: Ben Webb (benwebb)
Date: 2008-12-04 19:49

Message:
I can certainly help test such a solution, although my experience with the
SWIG internals is minimal. If you can remember anything further about the
Ruby iterator fix, that would be helpful.

----------------------------------------------------------------------

Comment By: William Fulton (wsfulton)
Date: 2008-12-04 14:28

Message:
The warnings are indicative of other problems. How about fixing the
iterators instead? I have a vague recollection that Ruby has the iterators
fixed, so perhaps the Ruby solution can be ported to Python.

----------------------------------------------------------------------

Comment By: Ben Webb (benwebb)
Date: 2008-12-04 10:49

Message:
Thanks - that's just it. Backing out SVN r10322 fixes the problem for me
both with current SVN and 1.3.35. I agree that if you can only have one
version of begin() and friends, it should probably be the non-const version
(I tried backing out the patch and then removing just the const iterators,
and that also works for me). Otherwise, it seems to me that backing out the
patch is the best solution - surely a few warnings are better than code
that fails to run?

----------------------------------------------------------------------

Comment By: Haoyu Bai (bhy)
Date: 2008-12-03 13:26

Message:
This is due to SF patch #1925702 (SVN r10322) "Remove warning 512 when
using std_vector.i":

std_container.i used to provide an interface for both the const and
non-const versions of begin(), end(), rbegin(), and rend(). This causes a
bunch of warning 512's when using std_vector.i. This patch removes the
non-const versions of those four functions, which gets rid of the
warnings.

This patch removed non-const version of begin(), so in SWIG 1.3.35 we got
a const_iterator returned, while std::vector::insert expect a non-const
iterator:

void insert( iterator loc, input_iterator start, input_iterator end );

So seems we should keep the non-const version of begin() etc., and remove
(or rename?) the const version.

----------------------------------------------------------------------

Comment By: Ben Webb (benwebb)
Date: 2008-12-02 18:02

Message:
File Added: test.i

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=2380788&group_id=1645

------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network 
management toolset available today.  Delivers lowest initial 
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd

headers

Stefan Zager | 9 May 2011 07:04

Re: next release and python 3.2

On Fri, 6 May 2011, William S Fulton wrote:

> Has anyone tried out python 3.2 and the test-suite? There are reports on
> swig-user that swig-2.0.3 and python-3.2 doesn't work. I think we are
> about ready for a new release, but python 3.2 needs to work before we
> push the next release out.

I committed a couple of minor fixes for python3.2 this weekend.  I'm now able to
run the full test-suite with python3.2, with and without -O and -builtin.

Stefan

------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network 
management toolset available today.  Delivers lowest initial 
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd

headers

SourceForge.net | 9 May 2011 09:51

[ swig-Bugs-3299307 ] Does not support repeat cascading

Bugs item #3299307, was opened at 2011-05-09 09:51
Message generated for change (Tracker Item Submitted) made by akimd
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3299307&group_id=1645

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: preprocessor
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Akim Demaille (akimd)
Assigned to: Nobody/Anonymous (nobody)
Summary: Does not support repeat cascading

Initial Comment:
Hi,

I have been toying with Boost.Preprocessor, and maybe the SWIG preprocessor is not too far from being able
to grok it.  But there is a first issue, demonstrated as follows.

Thanks in advance.

$ swig -version

SWIG Version 2.0.3

Compiled with /usr/bin/g++-4.2 [i386-apple-darwin10.7.0]

Configured options: +pcre

Please see http://www.swig.org for reporting bugs and further information

$ cat /tmp/bar.cc
#define ONE(X) TWO
#define TWO(X) X

ONE(1)(2)

$ swig -E -java /tmp/bar.cc | tail

]

]
%includefile(maininput="/tmp/bar.cc") "/tmp/bar.cc" [

TWO(2)
]

$ g++ -Wall -E /tmp/bar.cc
# 1 "/tmp/bar.cc"
# 1 "<built-in>"
# 1 "<command-line>"
# 1 "/tmp/bar.cc"

2

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3299307&group_id=1645

------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network 
management toolset available today.  Delivers lowest initial 
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd

headers

SourceForge.net | 9 May 2011 10:28

[ swig-Bugs-3299307 ] Does not support cascading macro invocations

Bugs item #3299307, was opened at 2011-05-09 09:51
Message generated for change (Settings changed) made by akimd
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3299307&group_id=1645

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: preprocessor
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Akim Demaille (akimd)
Assigned to: Nobody/Anonymous (nobody)
>Summary: Does not support cascading macro invocations

Initial Comment:
Hi,

I have been toying with Boost.Preprocessor, and maybe the SWIG preprocessor is not too far from being able
to grok it.  But there is a first issue, demonstrated as follows.

Thanks in advance.

$ swig -version

SWIG Version 2.0.3

Compiled with /usr/bin/g++-4.2 [i386-apple-darwin10.7.0]

Configured options: +pcre

Please see http://www.swig.org for reporting bugs and further information

$ cat /tmp/bar.cc
#define ONE(X) TWO
#define TWO(X) X

ONE(1)(2)

$ swig -E -java /tmp/bar.cc | tail

]

]
%includefile(maininput="/tmp/bar.cc") "/tmp/bar.cc" [

TWO(2)
]

$ g++ -Wall -E /tmp/bar.cc
# 1 "/tmp/bar.cc"
# 1 "<built-in>"
# 1 "<command-line>"
# 1 "/tmp/bar.cc"

2

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3299307&group_id=1645

------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network 
management toolset available today.  Delivers lowest initial 
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd

headers

SourceForge.net | 10 May 2011 19:54

[ swig-Patches-3300072 ] Tcl: fix for multi-import global variable problem

Patches item #3300072, was opened at 2011-05-10 13:54
Message generated for change (Tracker Item Submitted) made by delbaere
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=301645&aid=3300072&group_id=1645

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Christian Delbaere (delbaere)
Assigned to: Nobody/Anonymous (nobody)
Summary: Tcl:  fix for multi-import global variable problem

Initial Comment:
SWIG uses a global variable in the Tcl interpreter to track the first module that was loaded.  Currently, the
code that sets the variable does not explicitly use the global namespace.  This works correctly as long as
you load the modules outside of a namespace and outside of a procedure.  If you load the modules within a
namespace or procedure, or if you using Tcl packaging system, the 2nd module will overwrite the global
variable and there will be no linkage between them.

The attached patch contains a tiny fix for the problem as well as a modification to the tcl imports test to
reproduce the issue.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=301645&aid=3300072&group_id=1645

------------------------------------------------------------------------------
Achieve unprecedented app performance and reliability
What every C/C++ and Fortran developer should know.
Learn how Intel has extended the reach of its next-generation tools
to help boost performance applications - inlcuding clusters.
http://p.sf.net/sfu/intel-dev2devmay

27	May 2013
69	April 2013
70	March 2013
25	February 2013
105	January 2013
147	December 2012
27	November 2012
8	October 2012
33	September 2012
98	August 2012
127	July 2012
38	June 2012
142	May 2012
174	April 2012
232	March 2012
32	February 2012
59	January 2012
43	December 2011
60	November 2011
50	October 2011
30	September 2011
55	August 2011
30	July 2011
36	June 2011
93	May 2011
71	April 2011
101	March 2011
113	February 2011
66	January 2011
55	December 2010
35	November 2010
74	October 2010
64	September 2010
27	August 2010
60	July 2010
76	June 2010
107	May 2010
54	April 2010
41	March 2010
68	February 2010
33	January 2010
48	December 2009
106	November 2009
38	October 2009
80	September 2009
99	August 2009
137	July 2009
110	June 2009
104	May 2009
50	April 2009
89	March 2009
48	February 2009
92	January 2009
92	December 2008
53	November 2008
35	October 2008
96	September 2008
26	August 2008
101	July 2008
37	June 2008
73	May 2008
144	April 2008
125	March 2008
81	February 2008
73	January 2008
67	December 2007
47	November 2007
159	October 2007
131	September 2007
68	August 2007
39	July 2007
58	June 2007
84	May 2007
93	April 2007
83	March 2007
29	February 2007
62	January 2007
79	December 2006
227	November 2006
102	October 2006
83	September 2006
39	August 2006
43	July 2006
83	June 2006
83	May 2006
32	April 2006
129	March 2006
202	February 2006
378	January 2006
162	December 2005
63	November 2005
77	October 2005
57	September 2005
30	August 2005
38	July 2005
46	June 2005
36	May 2005
45	April 2005
57	March 2005
53	February 2005
33	January 2005
63	December 2004
151	November 2004
256	October 2004
107	September 2004
53	August 2004
38	July 2004
26	June 2004
22	May 2004
150	April 2004
148	March 2004
154	February 2004
351	January 2004
630	December 2003
477	November 2003
129	October 2003
333	September 2003
158	August 2003
109	July 2003
257	June 2003
164	May 2003
267	April 2003
519	March 2003
469	February 2003
86	January 2003
203	December 2002
470	November 2002
238	October 2002
640	September 2002
836	August 2002
759	July 2002
372	June 2002
578	May 2002
311	April 2002
292	March 2002
177	February 2002
379	January 2002
281	December 2001
266	November 2001
254	October 2001
281	September 2001
593	August 2001
337	July 2001
389	June 2001
49	May 2001
23	April 2001
36	March 2001
37	February 2001
83	January 2001
49	December 2000
13	November 2000
49	October 2000
255	September 2000
315	August 2000
180	July 2000
65	June 2000
24	May 2000
113	April 2000
97	March 2000
277	February 2000
141	January 2000

Mon	Tue	Wed	Thu	Fri	Sat	Sun

						1
2	3	4	5	6	7	8
9	10	11	12	13	14	15
16	17	18	19	20	21	22
23	24	25	26	27	28	29
30	31