llvm dev - Oct 2013 - [LLVMdev] Typecasting int32ty to int64ty

I have something like this on my code :-

vector<Value*> args3;
args3.push_back(instr->getOperand(0)); //where 'instr' is store
instruction

Now when I check the Type of instr->getOperand(0).
It varies to i32 and i64, depending on the operand(0) of instruction. 
Due to which my vector<Value*> has different Type of Value*.

How can I convert i32 to i64 type before pushing it to the vector<Value*>,
so that all Value* on my vector is of same Type?





--
View this message in context:
http://llvm.1065342.n5.nabble.com/Typecasting-int32ty-to-int64ty-tp61814.html
Sent from the LLVM - Dev mailing list archive at Nabble.com.

> How can I convert i32 to i64 type before pushing it to the
vector<Value*>,
> so that all Value* on my vector is of same Type?
You have to decide whether your i32s are signed or unsigned (should
0xffffffff give an i64 of -1 or 0xffffffff?). After that you can
insert an appropriate ZExt or SExt instruction into the basic block
and push its result onto your vector.

If inserting a sext/zext into the IR isn't appropriate then your only
option is to adapt your own code to deal with both i32 and i64 Values
being in that vector.

Cheers.

Tim.