Blitzmax Supertips & Quirkies

BlitzMax Forums/BlitzMax Programming/Blitzmax *Supertips & Quirkies*

Tom	(Posted 2005) [#1]

This isn't something you might have come across unless working with C structs & max types, but it caused me a few hours of head scratching.

When working in C, the field data of an array of structs is sequential in memory. In Blitzmax this is not the case. Here's how you can get a 'C friendly' array of your Blitzmax type array data:

Type test
	Field a:Byte,b:Byte
	
	Function GetArray:Byte[](srcArray:test[] Var)
		Local sz:Int = SizeOf(test)
		Local dataArray:Byte[srcArray.length * sz]
		Local i:Int
		Local bPtr:Byte Ptr = Varptr(dataArray[0])
		
		For i = 0 Until srcArray.length
			MemCopy bPtr+i*sz, Varptr(SrcArray[i].a),sz
		Next
		Return dataArray
	End Function
End Type

'Create a 'test' array
Local t:test[3]
t[0] = New test
t[1] = New test
t[2] = New test
t[0].a=1
t[0].b=2
t[1].a=3
t[1].b=4
t[2].a=5
t[2].b=6


Local tPtr:Byte Ptr = Varptr(t[0].a) 'Pointer to first byte in the array
Local i
Print "In Blitzmax type arrays, the fields are not sequential in memory as you might think"
For i = 0 Until 6
	Print tPtr[i]
Next
Print

'A 'C' struct friendly array
Local fieldData:Byte[] = test.GetArray(t)
Print "Using myType.GetArray() you can get a byte[] of"
Print "your type array data fields that will be 'C friendly"
For i = 0 Until fieldData.length
	Print fieldData[i]
Next

'Typically you would pass varptr(fieldData) to your C functions