Build 32 Bit Number from Non-8bit aligned vars

[crumgary]

CCS Version: CCS PCM C Compiler, Version 3.242, 32566

Just wondering if anyone can suggest a better way to build a 32bit number from a couple vars, but these other vars are not 8bit vars.

idealy I would have done something like:

typedef struct {
int16 i13Var : 13;
int8 i5Var : 5;
int16 i9Var : 9;
}myBigVar;

but I guess you can't do that with 8bit vars in CCS

sooo.....instead maybe this:

inline int32 SomeFunction(int16 i13Var, int8 i5Var, int16 i9Var)
{

int32 myBigVar = 0x0;
int32 i13Var_t = 0;
int32 i5Var_t = 0;
int32 i9Var_t = 0;

i9Var_t = (i9Var & 0x01FF);
i5Var_t = ( ( (int32)i5Var ) << 9 );
i13Var_t = ( (int32)i13Var << 14 );

myBigVar = i13Var_t + i5Var_t + i9Var_t;

return myBigVar;
}

The above seems to work, but if I put it all on one line, it doesn't

myBigVar = (i9Var & 0x01FF) + ( ( (int32)i5Var ) << 9 ) + ( (int32)i13Var << 14 );



example of use:
int32 i32Addr = 0x00;

int16 iBlock = 0b1111111111111;
int8 iNAND = 0b11111;
int16 iByte = 0b111111111;

i32Addr = scMake32Addr(iBlock,iNAND,iByte);

---results---
i32Addr ==> 0x07FFFFF -- correct

but the one line version results in
i32Addr ==> 0x0000C0FF -- incorrect


Any ideas on why the one liner version doesn't work?
Any Better ways of doing this (I really don't like the type case to 32bit because it's adding overhead)

disclaimer: I just type up the above code, so there _might_ be mistakes..but the code I'm actually use compiles and works (except as mentioned)


Thanks
Gary
[/code]

[crumgary]

ok, I lied, the one-liner version works too!

I had the MPLAB simulator set for a different chip than I was compiling for.
You would think that after I had a simular mistake with the simulator before, I would have made sure of that setting.... doh

So, I guess I was just wondering if there is a better way to do this.

Thanks again
Gary

[crumgary]

no thoughts on this?
Thanks
Gary