<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
I have looked in the GetElementPtr FAQ but could not find the answer to
this.<br>
<br>
I have encountered a GetElementPtr instruction that looks like this:<br>
<br>
%u.i = alloca %struct..0anon, align 8 ;
<%struct..0anon*> [#uses=2]<br>
%u = alloca %struct..0anon, align 8 ;
<%struct..0anon*> [#uses=2]<br>
%tmp = getelementptr %struct..0anon* %u, int 0, uint 0 ;
<ulong*> [#uses=1]<br>
<br>
So far so good. The bytecode for this GetElementPtr instruction is:<br>
<br>
<font face="Courier New, Courier, monospace">00000000000100001010101101101011<br>
^^ Type 3, 3 operands<br>
^^^^^^-- Opcode = 0x01a = 26 = GetElementPtr<br>
^^^^^^-------- Type = 0x02b = 43<br>
^^^^^^-------------- Operand 0 = 2 = %u<br>
^^^^^^-------------------- Operand 1 = 1 = int (because of
compaction table)<br>
^^^^^^-------------------------- Operand 2 = 0 = zero<br>
</font><br>
The bytecode has only three operands while the ll instruction clearly
has five. How do I know from the bytecode that the result type is
ulong* and not struct..0anon*?<br>
<br>
Thanks in advance,<br>
<br>
-- Robert.<br>
<br>
<br>
<br>
</body>
</html>