Re: [Swig-user] adding init function to c structure

[William S F. <ws...@fu...>]

Without a standalone testcase and compiler options etc, it is hard to 
tell, but I've never seen gcc have problems with the generated code from 
a simple constructor type %extend.

William

Rich E wrote:
> This works very elegantly!
> 
> One thing, I get a warning during compilation that the following
> function is not a prototype:
> 
> SWIGINTERN SMS_AnalParams *new_SMS_AnalParams(){
>    SMS_AnalParams *s = (SMS_AnalParams *)malloc(sizeof(SMS_AnalParams));
>    sms_initAnalParams(s);
>    return s;
>  }
> 
> It works, so does anyone know why gcc is complaining?
> 
> 
> thanks for the help, cheers..
> 
> rich
> 
> On Wed, Feb 4, 2009 at 11:17 PM, William S Fulton
> <ws...@fu...> wrote:
>> Rich E wrote:
>>> Anyone know of a good example on how to do this?  Basically, in my C
>>> code, I have a function that initializes a structure to default
>>> values.  So, in python I need to do something like:
>>>
>>>>> analParams = SMS_AnalParams()
>>>>> sms_initAnalParams(analParams)
>>> I'm sure it is possible to make the second step happen automatically,
>>> but the various stabs I just took at it didn't work.  Anyone that has
>>> done this or knows how, would you mind giving a recommendation?
>>>
>> Use constructor syntax in a %extend block, something like:
>>
>> %extend SMS_AnalParam {
>>  SMS_AnalParam() {
>>    SMS_AnalParam *s = (SMS_AnalParam *)malloc(sizeof(SMS_AnalParam));
>>    sms_initAnalParams(s);
>>    return s;
>>  }
>> }
>>
>> It works for C code.
>>
>> William
>>
>