error: conversion from 'F64 {aka double}' to non-scalar type

Guys, help me make gcc happy.  This snippet doesn't bother VS at all but gcc tells me no:

Code:

void LLFloaterObjectFunctions::onClickBlink()
{
    // move current selection based on delta from position and update z position
    for (LLObjectSelection::root_iterator iter = LLSelectMgr::getInstance()-&tgetSelection()-&troot_begin();
        iter != LLSelectMgr::getInstance()-&tgetSelection()-&troot_end(); iter++)
    {
        LLSelectNode* node = *iter;
        if (node)
        {
            LLVector3d cur_pos = node-&tgetObject()-&tgetPositionGlobal();
            LLVector3d new_pos = cur_pos.mdV[VZ] = 340282346638528859811704183484516925440.0f;
            node-&tmDuplicatePos = node-&tgetObject()-&tgetPositionGlobal();
            node-&tgetObject()-&tsetPositionGlobal(new_pos);
        }
    }
    LLSelectMgr::getInstance()-&tsendMultipleUpdate(UPD_POSITION);
}

the full error message with the indicator:
error: conversion from ‘F64 {aka double}’ to non-scalar type ‘LLVector3d’ requested
   LLVector3d new_pos = cur_pos.mdV[VZ] = 340282346638528859811704183484516925440.0f;
                                          ^

from Kinggoon Copybot Forum - All Forums http://ift.tt/2kKMuly
via IFTTT