Am 20.10.2012 00:08, schrieb Anthony D. Baye:
> would there be any conceivable problem with changing the maximum number of blob
> components to sizeof(unsigned int)?

You probably mean INT_MAX.

> I mean, aside from type mismatch warnings pouring from every orifice when you
> compiled unless you went through and changed the types on all the counters as
> well...

That. A thorough review of the blob code would be needed to determine 
the possible maximum value.

There's at least on statement suggesting that MAX_BLOB_COMPONENTS should 
be lower than INT_MAX/7.

Other code suggests that MAX_BLOB_COMPONENTS should be lower than any of 
the following expressions:

   SIZE_MAX/(7*sizeof(double))
   SIZE_MAX/(2*sizeof(BSPHERE_TREE))
   SIZE_MAX/sizeof(Blob_Interval_Struct)

And that's just what a superficial review reveals.


> Honestly, though, 1M seems a little low...

640k should be enough for everybody...

Post a reply to this message