hi all,
I write this program under the C++ not C. But I heaively used the
malloc() and free() functions in my code.
In brief what my code does is just allocate memory for NFA transition
table. It grows dynamically. When the number of transitions is a %4
then it will automatically free the previously allocated memory for the
two arrays and allocate a new array with more additional 17 bytes.
4 integer and 1 - zero terminator byte (zero terminator is necessary
because it is used in by the free() call).
here is the code ,
int TransistionTable::AddTransistion ( int stateId , int nextStateId )
{
// we are first check if adding this state will
// overflow our allocated memory.
if ( (numberOfTransistions ) %4 == 0 )
{
// then we need to allocate some memory
int * pNewIntId = (int *)malloc ( numberOfTransistions *4 + 17);
int * pNewTransistion = (int *) malloc ( numberOfTransistions * 4 + 17);
memcpy ( ( void *)pNewIntId ,( void* ) pIntId ,numberOfTransistions*4);
memcpy ( ( void *)pNewTransistion , ( void *) pTransistion , numberOfTransistions*4);
// then we just release the old memory
if ( ! numberOfTransistions == 0)
{
free (( void*) pIntId);
free (( void*) pTransistion);
}
// then we update the pointer values
pIntId = pNewIntId ;
pTransistion = pNewTransistion ;
}
// then add the new value
pIntId [numberOfTransistions] = stateId ;
pTransistion [ numberOfTransistions ] = nextStateId ;
numberOfTransistions++;
// just return 0 to say nothing goes wrong
return 0;
}This code works prefect , and I already write a unite test for this.
My Question is how should I improve performance of the memory
management algorithm , is there is a better algorithm ?
is it oky to just allocate only 4 bytes using malloc() ? Need suggestions....
please forgive me about C++ coding standards , I know some C ppl
find very difficult to read the C++ styled code.
