Converting code from c to c++ [duplicate] - c++

Question

This question already has answers here:
What is the C++ equivalent to C's designated initializers?
(4 answers)
Closed 7 years ago.
I'm trying to convert some code from C to C++
This is a project using a raspberry pi with camera module and I'd like to analyse pictures with it.
But on this piece of code (somebody else created) I get this error
231:8: error: expected primary-expression before ‘.’ token
Which is this line:
.max_stills_w = state->width,
I tried everything I could find but it keeps giving me other errors
video_port = camera->output[MMAL_CAMERA_VIDEO_PORT];
still_port = camera->output[MMAL_CAMERA_CAPTURE_PORT];
// set up the camera configuration
{
MMAL_PARAMETER_CAMERA_CONFIG_T cam_config =
{
{ MMAL_PARAMETER_CAMERA_CONFIG, sizeof(cam_config) },
.max_stills_w = state->width,
.max_stills_h = state->height,
.stills_yuv422 = 0,
.one_shot_stills = 0,
.max_preview_video_w = state->width,
.max_preview_video_h = state->height,
.num_preview_video_frames = 3,
.stills_capture_circular_buffer_height = 0,
.fast_preview_resume = 0,
.use_stc_timestamp = MMAL_PARAM_TIMESTAMP_MODE_RESET_STC
};
mmal_port_parameter_set(camera->control, &cam_config.hdr);
}
// Set the encode format on the video port

Answer 1

C++ doesn't support named initializers, only positional ones.
You'll have to read the struct definition and arrange all the initializers into declaration order.
Any members which weren't being named in the C code were being initialized to zero implicitly. You may have to make some of them explicit in order to not skip positions.
If this code is from JVCleave then it appears that the members are already in the correct order, you can just comment out the names and it will still work.

Answer 2

The named structure member initialisation in C99 is not implemented in C++. In C++ you can perform member initialisation using a constructor. E.g:
struct sConfig
{
int m_x ;
int m_y ;
sConfig( int x, int y ) : m_x(x),
m_y(y)
{}
} ;
Or:
struct sConfig
{
int m_x ;
int m_y ;
sConfig( int x, int y )
{
m_x = x ;
m_y = y ;
}
} ;
or even a combination of the two methods. Then you instantiate an object thus:
sConfig myConfig( 10, 20 ) ;
Initialising m_x and m_y to 10 and 20 respectively in this example.
You can then of course only perform the initialisation provided by the structures defined constructors; but that is not necessarily a bad thing - letting the user arbitrarily decide which members to initialise is not particularly safe or maintainable. You can of course define multiple constructors to perform different initialisations; you would normally want define a default constructor for example:
struct sConfig
{
int m_x ;
int m_y ;
sConfig( int x, int y ) : m_x(x),
m_y(y)
{}
sConfig( ) : m_x(0),
m_y(0)
{}
} ;
So that in this example:
sConfig myConfig ;
is equivalent to all of the following:
sConfig myConfig( 0, 0 ) ;
sConfig myConfig = { 0, 0 } ;
sConfig myConfig = {0} ;