gmane.comp.programming.swig.devel

[ swig-Bugs-1607576 ] std_string and std_sstream.i bug

SourceForge.net — 2012-05-18T03:04:10

Bugs item #1607576, was opened at 2006-12-02 14:11
Message generated for change (Comment added) made by olly
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=1607576&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: Gaetan Lehmann (gleh)
Assigned to: Nobody/Anonymous (nobody)
Summary: std_string and std_sstream.i bug

Initial Comment:
the following (very) simple interface:

%include std_string.i
%include std_sstream.i

produce invalid c++ code, where basic_string is not prefixed with "std::".
Including the 2 libs in reversed order works without any problem:

%include std_sstream.i
%include std_string.i

Reproduced on mac os and linux, with python and swig 1.3.31

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

Date: 2012-05

[ swig-Bugs-1670570 ] segfault with two modules,inheritance and virtual functions

SourceForge.net — 2012-05-18T02:59:24

Bugs item #1670570, was opened at 2007-02-27 23:20
Message generated for change (Settings changed) made by olly
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=1670570&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.
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Maks N. Polunin (maks_polunin)
Assigned to: Nobody/Anonymous (nobody)
Summary: segfault with two modules, inheritance and virtual functions

Initial Comment:
I have two modules
baseA contains class baseA w/ virtual function foo
and A contains child of baseA w/out foo
I enable directors for them
When trying to inherit in python from A and call foo, I get segfault
Temporary workarounds are to use only one module or to declare A::foo

just run runme.sh from attachment
To see workaround working - read notes in A.h and A.i

swig is from CVS

$ swig -ve

[ swig-Bugs-1694806 ] Python gets deleted object whenstoring it passed from c++

SourceForge.net — 2012-05-18T02:57:06

Bugs item #1694806, was opened at 2007-04-04 23:43
Message generated for change (Settings changed) made by olly
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=1694806&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.
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Maks N. Polunin (maks_polunin)
Assigned to: Nobody/Anonymous (nobody)
Summary: Python gets deleted object when storing it passed from c++

Initial Comment:
I'm not sure if it is a bug, but it's confusing anyway.
The problem appears when I create object in C++ and pass it into python callback. Python code creates reference to it and when callback is finished, py reference points to deleted object.
Is it possible someway to track if there are any links from python after calling callback and replace wrapped pointer to c++ object w/ pointer to copy of

[ swig-Bugs-3454399 ] Memory leak after exception

SourceForge.net — 2012-05-18T02:53:49

Bugs item #3454399, was opened at 2011-12-08 06:53
Message generated for change (Settings changed) made by olly
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3454399&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.
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: https://www.google.com/accounts ()
Assigned to: Nobody/Anonymous (nobody)
Summary: Memory leak after exception

Initial Comment:
If a function throws an exception, any applicable 'freearg' typemaps are not executed and leaks may occur.

%typemap(in) Data*
{
   $1 = new Data($input)
}

%typemap(freearg) Data*
{
  delete $1;
}

class Data
{
// something
};

class Foo
{
public:
  void Bar(Data* data) throw (Exception);
};


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

You can respond by visiting: 
https://sourceforge.net/tr

[ swig-Bugs-3526487 ] [python] SWIG_From(unsigned int)return a Python long

SourceForge.net — 2012-05-18T02:52:54

Bugs item #3526487, was opened at 2012-05-14 00:45
Message generated for change (Settings changed) made by olly
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3526487&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.
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Johan Hake ()

Initial Comment:
After revision 12732 by Stefan Zager SWIG_From(unsigned int) now returns a Python long instead of a Python int. This breaks a lot of running code for us, and we have had to write our own typemaps to circumvent this bug.

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

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3526487&group_id=1645

------------------------------------------------------------------------------
Live Security Virtual Conference
Ex

[ swig-Bugs-2792261 ] [lua] SWIG_check_num_args toostrict

SourceForge.net — 2012-05-18T02:47:16

Bugs item #2792261, was opened at 2009-05-15 06:34
Message generated for change (Settings changed) made by olly
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=2792261&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: lua
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Sebastian (bastisemail)
Assigned to: Mark Gossage (mgossage)

Initial Comment:
We moved from SWIG version 1.3.31 to 1.3.39 and found out, that SWIG now introduced the following sanity check:

SWIG_check_num_args

By doing so we can not write any wraper code anymore taking variable number of arguments.

For example, we wrote some typemap 



%typemap(in) (int n, double* x) (double*) {
double *y;
double m, key;
int n, k=0;
const char *text;
size_t nbytes;
int i;
/* get number of arguments */
n = lua_gettop(L)-1;
$1 =

[ swig-Bugs-3428833 ] std::vector template in PHP

SourceForge.net — 2012-05-18T02:44:28

Bugs item #3428833, was opened at 2011-10-26 10:48
Message generated for change (Comment added) made by olly
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3428833&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: php
Group: None
Priority: 5
Private: No
Submitted By: jcf (frickson)
Assigned to: Olly Betts (olly)
Summary: std::vector template in PHP

Initial Comment:
In my test.i file, I have:
    %include "std_vector.i"
    %template(vectorDbl) std::vector<double>;
    %include "test.h"

In a test.h, I have:
    #include <vector>
    typedef std::vector<double>DblVector;
    class test
    {
           ...
           DblVector  vals1, vals2;
    };

In test.php, if I have:
     $count = $obj->vals1->size();

At this point I get an error saying "Class vectorT_double_t not found".

If I change test.i from
    %template(vectorDbl)

Re: Request for a Matlab-branch

Oliver Buchtala — 2012-05-17T07:31:26

Hi Josef,

(All?) other modules behave like this.

Though, IMO it doesn't mean that your module has to.

Namespace information is available in the 'name' attribute e.g., of a 
class declaration node.
You can retrieve the namespace with

     Swig_scopename_prefix(Getattr(n, "name"));

HTH,
Oliver


On 17.05.2012 09:19, Josef Pacula wrote:


------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Re: Request for a Matlab-branch

Josef Pacula — 2012-05-17T07:19:10

Thanks!

The documentation of SWIG says that namespaces are flattened. Is there
a way to use the namespace information? Like a simple switch or a
macro that would prepend namespace names to each wrapped item?

Regards,
Josef

On 13 May 2012 22:31, Oliver Buchtala <oliver.buchtala< at >googlemail.com> wrote:

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

[ swig-Bugs-3428833 ] std::vector template in PHP

SourceForge.net — 2012-05-16T01:13:27

Bugs item #3428833, was opened at 2011-10-26 10:48
Message generated for change (Comment added) made by highcommander4
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3428833&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: php
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: jcf (frickson)
Assigned to: Olly Betts (olly)
Summary: std::vector template in PHP

Initial Comment:
In my test.i file, I have:
    %include "std_vector.i"
    %template(vectorDbl) std::vector<double>;
    %include "test.h"

In a test.h, I have:
    #include <vector>
    typedef std::vector<double>DblVector;
    class test
    {
           ...
           DblVector  vals1, vals2;
    };

In test.php, if I have:
     $count = $obj->vals1->size();

At this point I get an error saying "Class vectorT_double_t not found".

If I cha

Re: SWIG_From_unsigned_SS_int now returns a long

Johan Hake — 2012-05-15T07:32:10

So basically you suggest that:

   SWIG_From_unsigned_SS_int

should use PyInt_fromLong instead of SWIG_From_unsigned_SS_long? Such 
solution make sense to me.

Johan

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Re: SWIG_From_unsigned_SS_int now returns a long

Stefan Zager — 2012-05-15T07:23:23





On a 32-bit system, an unsigned long int can overflow a python int.

But if I recall correctly, the motivation for the change was this:
previously, the code fragments in question would return either a python
long or a python int, depending on the input value, and it was problematic
not knowing what type those code fragments would return.

Rather than reverting that change, the right thing to do is probably to
disambiguate SWIG_From_unsigned_SS_int and SWIG_From_unsigned_SS_long.

Stefan
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

[ swig-Bugs-3526748 ] SWIG generates calls to privatecopy constructor

SourceForge.net — 2012-05-15T07:05:35

Bugs item #3526748, was opened at 2012-05-14 16:51
Message generated for change (Comment added) made by wsfulton
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3526748&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
Priority: 5
Private: No
Submitted By: Jeff Laughlin (n1ywb)
Summary: SWIG generates calls to private copy constructor

Initial Comment:
SWIG version: 2.0.6

When wrapping a function which returns a C++ object by value, SWIG declares the "result" object with automatic storage class, hence allocating it on the stack. It then calls the copy constructor to make a copy of the object which is allocated on the heap. Besides being inefficient, it fails to compile if the copy cosntructor is private. In other words, SWIG fails to respect the fact that the copy constructor is private.

E.

SWIG_From_unsigned_SS_int now returns a long

William S Fulton — 2012-05-15T06:54:03

Hi Stefan

Can you talk us through the motivation for making this change? I see from
http://docs.python.org/library/stdtypes.html that a Python long is 
implemented using a C long. From this I'd say that the changes that 
changed the long typemaps to use a Python long are thus incorrect. Also 
the changes for long long seem unnecessary to always return a Python 
long as they are range checked beforehand to see if they fit into a 
Python int.

William'

On 08/05/12 12:37, Johan Hake wrote:


------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Re: request developer access

William S Fulton — 2012-05-15T06:44:23

Hi Peter

You've probably noticed that there hasn't been any active maintenance on 
Modula3 for many years and I'm sure any sort of maintenance on this 
would be welcome by Modula3 users. Do you envisage providing regular 
updates, so svn access would be useful, or is this a one off patch 
someone else could apply? In either case, could you post your patch for 
us to look at and perhaps introduce yourself with some background.

Thanks
William

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

[ swig-Bugs-3526748 ] SWIG generates calls to privatecopy constructor

SourceForge.net — 2012-05-14T23:51:49

Bugs item #3526748, was opened at 2012-05-14 16:51
Message generated for change (Tracker Item Submitted) made by n1ywb
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3526748&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: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Jeff Laughlin (n1ywb)
Assigned to: Nobody/Anonymous (nobody)
Summary: SWIG generates calls to private copy constructor

Initial Comment:
SWIG version: 2.0.6

When wrapping a function which returns a C++ object by value, SWIG declares the "result" object with automatic storage class, hence allocating it on the stack. It then calls the copy constructor to make a copy of the object which is allocated on the heap. Besides being inefficient, it fails to compile if the copy cosntructor is private. In other word

request developer access

Peter McKinna — 2012-05-14T23:38:14

Hi,

  I have made a number of changes to the modula3.cxx module for swig
and would like to incorporate them in the current source tree.

Regards Peter

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

[ swig-Bugs-3344762 ] Adding %include "std_string.i"causes g++ compile to fail

SourceForge.net — 2012-05-14T18:55:48

Bugs item #3344762, was opened at 2011-06-29 14:09
Message generated for change (Comment added) made by wsfulton
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3344762&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: tcl
Group: None
Priority: 5
Private: No
Submitted By: fedorabug (fedorabug)
Summary: Adding %include "std_string.i" causes g++ compile to fail

Initial Comment:
This is the entire SWIG file which works with g++ when compiled using "swig -tcl8 func.i"

%module base
//%include "std_string.i"
%{
float pow(float one, float two)
{
  float result = 1;
  for(int i=0;i<two;i++)
    result = result * one;
  return result;
}
%}

float pow(float one, float two);


However merely uncommenting the std_string include causes the g++ compile to fail due to two lines of code in the baseFunc_wrap.c file:

1. value

[ swig-Bugs-3489392 ] namespace + operator in defaultargument parsed uncorrectly

SourceForge.net — 2012-05-14T18:40:21

Bugs item #3489392, was opened at 2012-02-19 06:45
Message generated for change (Comment added) made by wsfulton
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3489392&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: parsing
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Martin Fiers (mfiers)
Assigned to: Marcelo Matus (marcelomatus)
Summary: namespace + operator in default argument parsed uncorrectly

Initial Comment:
I got this error while trying to compile python-meep (https://launchpad.net/python-meep) with kwargs on. The following function is wrapped incorrectly:

in meep.hpp:

namespace meep {

class  custom_src_time {
    custom_src_time(complex<double> (*func)(double t, void *), void *data,
  double st = -infinity, double et = infinity)

    }
}

In the wrapped file meep_w

[ swig-Bugs-3526487 ] SWIG_From(unsigned int) return aPython long

SourceForge.net — 2012-05-14T07:45:54

Bugs item #3526487, was opened at 2012-05-14 00:45
Message generated for change (Tracker Item Submitted) made by 
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3526487&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: Johan Hake ()
Assigned to: Nobody/Anonymous (nobody)
Summary: SWIG_From(unsigned int) return a Python long 

Initial Comment:
After revision 12732 by Stefan Zager SWIG_From(unsigned int) now returns a Python long instead of a Python int. This breaks a lot of running code for us, and we have had to write our own typemaps to circumvent this bug.

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

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=101645&aid=3526487&group_id=1645

---

Re: Request for a Matlab-branch

Oliver Buchtala — 2012-05-13T20:31:57

Hello Josef,

you might also want to have a look at my control flow analyzer module 
which I use to evolve the specification of my V8 module: 
https://github.com/oliver----/swig-v8 (branch: analyzer).
Have also a look at: 
https://github.com/oliver----/swig-v8/blob/devel/Doc/Devel/Javascript/V8_CodeGeneratorSpecification.md#control-flow-analysis

Cheers,
Oliver

On 12.05.2012 10:22, Josef Pacula wrote:

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/