mdlKISolid_bodyToElements is returning success but no elements

How is pElemDescr declared? You've passed a variable called pElemDescr into both mdlKISolid_elementToBody and mdlKISolid_bodyToElements, but one takes an MSElementDescr* for an existing element and the other takes an MSElementDescr** to allocate and return a new element. So if pElemDescr is declared as MSElementDecr* pElemDescr = nullptr; you'd do this: mdlKISolid_bodyToElements(&pElemDescr, ...).

-B

* kBody passed to mdlKISolid_elementToBody looks suspicious as well...if this is a KIBODY** then it can't be passed to mdlKISolid_bodyToElements/mdlKISolid_free that require a KIBODY*...

MSElementDescr* pElemDescr = NULL;

ProcessElementByBodies(&pElemDescr);


//Defination
void ProcessElementByBodies(MSElementDescr** pElemDescr)
{
    int iStatus = BSIERROR; 
    KIBODY* kBody = NULL;
    iStatus = getBody(m_pElemDescr, &kBody);

    if (iStatus == BSISUCCESS && NULL != kBody)
    {
        try
        {
            iStatus = BSIERROR;
            int iStatusLocal = mdlKISolid_bodyToElements(pElemDescr, kBody, FALSE, -1, &m_pElemDescr->el, m_pDgnModelRef); 
            mdlKISolid_free(kBody);
            if (BSISUCCESS == iStatusLocal)
            iStatus = BSISUCCESS;
        }
        catch (...)
        {
            iStatus = GOTEXCEPTION;
        }
    }
}

Sorry. Its a typo. 

MSElementDescr* pElemDescr = NULL; This is what there in my code.

This is the getBody() implementation : 

int getBody(MSElementDescr*  pElemDescr,KIBODY** kBody)
{	
	int	iStatus = BSIERROR;
	try
	{
		iStatus = mdlKISolid_elementToBody(kBody, pElemDescr, m_pDgnModelRef);
	}
	catch (...)
	{
		iStatus = GOTEXCEPTION;
	}
	return iStatus;
}

I hope you understood, how i use the kBody now. 

BTW, i am not facing any syntactical errors or crashes with my code. Issue is with the Microstation API itself. mdlKISolid_bodyToElements() is getting success but there are no elements in it. 

Instead of using that API, if i use mdlKISolid_satDataFromElement(), i am able to get the sat data and graphics properly. But as i said earlier, this is an undocumented API and giving crashes to other dgn files and not encouraging to use it. 

Regards

Sampath.

I wrote a simplest test code snippet without any error check. It works well for my attached test.dgn. FYI.

void kiSolidTest()
{
	MSElementDescrP inEdP=NULL, outEdP=NULL;
	KIBODY*  kBody=NULL;
	mdlAssoc_getElementDescr (&inEdP, NULL, 16904L, ACTIVEMODEL, FALSE);
	mdlKISolid_elementToBody(&kBody, inEdP, ACTIVEMODEL);
	mdlKISolid_bodyToElement(&outEdP, kBody, FALSE, -1, NULL, ACTIVEMODEL);
	mdlElmdscr_add (outEdP);
	mdlElmdscr_freeAll (&outEdP);
	mdlElmdscr_freeAll (&inEdP);
	mdlKISolid_freeBody(&kBody);
}

8712.test.dgn

Hi Yong.

Its almost similar code what i have written, but the smart surfaces are not converting to elements. 

Regards

Sampath

A couple things to try:

1) Pass TRUE to create the preferred SmartSurface representation (FALSE to create surface geometry was a V7 format legacy option so that even older versions that did not yet include a solid kernel could still render the smart surfaces/solids).

2) Use mdlKISolid_elementToBody2 instead of mdlKISolid_elementToBody and make sure you have mdlKISolid_elementToBody2 return the body to uor transform. mdlKISolid_elementToBody tries to return the body transformed to world coordinates which can be a problem since this may scale or move the body outside the solid kernel's 1 km size box making it invalid.

If using mdlKISolid_elementToBody2 corrects the issue with mdlKISolid_bodyToElement not creating any elements, we'll discuss what to do with that body-to-uor transforms to get those elements at the correct location/scale next.

NOTE: mdlKISolid_elementToBody is actually documented as being deprecated in V8i:

/*---------------------------------------------------------------------------------**//**
* @description This function converts an element to a brep which can be a solid
* (type 19 elements), a sheet (type 18 elements, shapes, spline surfaces, filled
* ellipses etc.) or a wire (curves, line strings etc). If the input element is
* a smart element, the embedded brep is returned. For all other elements, the
* kernel in which the body is created depends on the workmode. The bodies need to
* be deleted by the calling application after they are manipulated. This function is
* deprecated, use mdlKISolid_elementToBody2 instead.
* @param bodyPP OUT is the body created by the function.
* @param edP IN is the MicroStation element.
* @param modelRef IN indicates the model containing the element.
* @return SUCCESS/ERROR depending on whether operation is successful or not
* @alinkjoin usmthmdlKISolid_bodyToElement usmthmdlKISolid_bodyToElementsC
* @group "MDL Solids API"
* @bsimethod
+---------------+---------------+---------------+---------------+---------------+------*/
nativeCode StatusInt mdlKISolid_elementToBody
(
KIBODY **bodyPP, /* OUT body */
MSElementDescr *edP, /* IN element */
DgnModelRefP modelRef /* IN model ref */
);

Would you mind to upload your test DGN file here ?

Another thing to notice: I use mdlKISolid_bodyToElement, not your mdlKISolid_bodyToElements.

Text.dgn

Please find the attached dgn file. It has only one smart surface (Text "B"). 

I think i understood the problem. The API is correct, but the range of the object is more and its becoming invalid. 

Brien Bastings said:

mdlKISolid_elementToBody is actually documented as being deprecated in V8i

Yes, but it's easy to miss that note.  You could reinforce that statement by moving the function declaration to, say, mdlKISolid_deprecated.fdf.

You've done something similar with the text API, which unhesitatingly focusses the developer's attention.

Bency Gregorioz said:

I think i understood the problem. The API is correct, but the range of the object is more and its becoming invalid. 

Brien Bastings said:

If using mdlKISolid_elementToBody2 corrects the issue with mdlKISolid_bodyToElement not creating any elements, we'll discuss what to do with that body-to-uor transforms to get those elements at the correct location/scale next.

Not sure if this means using mdlKISolid_elementToBody2 worked, but just in case, to follow up on my previous comment...here's how to account for the bodyTransform when creating elements from a body in local (solid kernel) coordinates.

MSElementDescrP inEdP=NULL, outEdP=NULL;
KIBODY* kBody=NULL;
Transform bodyTransform;
	
mdlAssoc_getElementDescr(&inEdP, NULL, 16904L, ACTIVEMODEL, FALSE);
mdlKISolid_elementToBody2(&kBody, &bodyTransform, inEdP, ACTIVEMODEL, 0L, FALSE);

mdlKISolid_bodyToElement(&outEdP, kBody, TRUE, -1, NULL, ACTIVEMODEL);
mdlKISolid_beginCurrTrans(ACTIVEMODEL);
mdlElmdscr_transform(outEdP, &bodyTransform);
mdlKISolid_endCurrTrans();
mdlElmdscr_add(outEdP);

mdlElmdscr_freeAll(&outEdP);
mdlElmdscr_freeAll(&inEdP);
mdlKISolid_freeBody(&kBody);

Jon Summers said:

Yes, but it's easy to miss that note.  You could reinforce that statement by moving the function declaration to, say, mdlKISolid_deprecated.fdf.

...

Jon Summers said:

Future-proof: IElementGraphicsProcessor is available, with few changes, in CONNECT, whereas the mdlKISolid_api doesn't exist in CONNECT

Seems sufficiently reinforced?

Jon Summers said:

Yes, but it's easy to miss that note.  You could reinforce that statement by moving the function declaration to, say, mdlKISolid_deprecated.fdf.

...

Jon Summers said:

Future-proof: IElementGraphicsProcessor is available, with few changes, in CONNECT, whereas the mdlKISolid_api doesn't exist in CONNECT

Seems sufficiently reinforced?