| View previous topic :: View next topic |
| Author |
Message |
colin382
Joined: 03 Jun 2020
Posts: 37
Location: UK
|
| Assemble long int from shifted bytes |
 Posted: Tue Jan 10, 2023 6:10 am |
 |
|
I receive an int32 as four int8 and I need to reconstruct the original int32
e.g. Byte3 = 32, byte2 4, byte1 = 8 and byte0 =99
longInt = (Byte3<<24) gives the correct value of 536870912 but
longInt = (Byte3<<24) + (Byte2<<16) (or any other combination of shifted bytes) gives zero.
How does one add shifted bytes to form a long int?
code snip:
| Code: | PulseInt = 537135203; // test value, all bytes non-zero
// create the received bytes
Byte3 = PulseInt >> 24;
Byte2 = (PulseInt >> 16) & 0xFF;
Byte1 = (PulseInt >> 8) & 0xFF;
Byte0 = PulseInt & 0xFF;
// the individual byte values are correct
Interval = Byte3<<24; // OK
Interval = (Byte3<<24) + (Byte2<<16); // not OK
|
|
|
 |
temtronic
Joined: 01 Jul 2010
Posts: 9225
Location: Greensville,Ontario
|
|
| Posted: Tue Jan 10, 2023 6:52 am |
|
|
Why not use the make() function that CCS supplies ?
quik cut and paste from 2019 manual..
make32( )
Syntax:
i32 = MAKE32(var1, var2, var3, var4)
Parameters:
var1-4 are a 8 or 16 bit integers
var2-4 are optional
Returns:
32 bit integer
Function:
Makes a 32 bit number out of any combination of 8 and 16 bit numbers. Note that the
number of parameters may be 1 to 4. The msb is first. If the total bits provided is less
than 32 then zeros are added at the msb
Availability:
All Devices
Requires:
-----
Examples:
int32 x;
int y;
long z;
x = make32(1,2,3,4); // x is 0x01020304
y=0x12;
z=0x4321;
x = make32(y,z); // x is 0x00124321
x = make32(y,y,z); // x is 0x12124321
Example Files:
ex_freqc.c
See Also:
make() is also very fast to execute. |
|
|
colin382
Joined: 03 Jun 2020
Posts: 37
Location: UK
|
|
| Posted: Tue Jan 10, 2023 7:28 am |
|
|
Thanks temtronic, that worked perfectly.
"Why not use the make() function that CCS supplies ?"
because I didnt know it existed, nor did I expect it to exist! |
|
|
Ttelmah
Joined: 11 Mar 2010
Posts: 19504
|
|
| Posted: Tue Jan 10, 2023 7:31 am |
|
|
It's worth also saying that the same applies to your extraction operations:
| Code: |
PulseInt = 537135203; // test value, all bytes non-zero
// create the received bytes
Byte3 = make8(PulseInt,3; //Take the top byte from the value
Byte2 = make8(PulseInt,2); //take the third byte
Byte1 = make8(PulseInt,1); //second byte
Byte0 = make8(PulseInt,0); //bottom byte
|
You will find that this code is smaller and quicker than your version.
The make8 operation just extracts the single 8bit byte from a variable.
Just what you want. The make32 operation does effectively the reverse
(they are both more capable than this so make32, can for example
assemble two 16bit values).
It is also worth raising the other alternative way of doing this with a
union:
| Code: |
union {
unsigned int32 whole;
unsigned int8 Bytes[4];
} PulseInt, Interval;
PulseInt.whole = 537135203;
//Then PulseInt.Bytes[0] to [3] are directly the bytes
Byte3 = PulseInt.Bytes[3];
Byte2 = PulseInt.Bytes[2];
Byte1 = PulseInt.Bytes[1];
Byte0 = PulseInt.Bytes[0];
//as a demo of this.
//You can then reassemble the other way round.
Interval.Bytes[3]=Byte3;
Interval.Bytes[2]=Byte2;
//etc...
|
The great thing about both the union and the make accesses is the
compiler does both by simple direct byte movements. Super efficient. |
|
|
temtronic
Joined: 01 Jul 2010
Posts: 9225
Location: Greensville,Ontario
|
|
| Posted: Tue Jan 10, 2023 8:40 am |
|
|
It's kinda interesting what's in the manual......
all sorts of EASY ways to get PICs to perform....
be sure to look at the FAQ section too !!
Also worth looking at (need 2-3 cups of coffee ) are all the examples they put in, um, ah, the 'EXAMPLES' folder.....
I always have the manual open ( F11) cause I forget stuff real easy now that I'm nearing 70. |
|
|
colin382
Joined: 03 Jun 2020
Posts: 37
Location: UK
|
|
| Posted: Tue Jan 10, 2023 9:00 am |
|
|
| Thanks for the additional material. Like you I nearly always have the help file open. At 78 years old, it's even easier to forget. |
|
|
PrinceNai
Joined: 31 Oct 2016
Posts: 479
Location: Montenegro
|
|
| Posted: Tue Jan 10, 2023 12:47 pm |
|
|
| It still seems ye ol' geezers forgot more than I'll ever know :-) |
|
|
temtronic
Joined: 01 Jul 2010
Posts: 9225
Location: Greensville,Ontario
|
|
| Posted: Tue Jan 10, 2023 12:57 pm |
|
|
| I know 113355 is a magical number for hungry folks...... |
|
|
PrinceNai
Joined: 31 Oct 2016
Posts: 479
Location: Montenegro
|
|
| Posted: Tue Jan 10, 2023 1:04 pm |
|
|
| Pi? |
|
|
temtronic
Joined: 01 Jul 2010
Posts: 9225
Location: Greensville,Ontario
|
|
| Posted: Tue Jan 10, 2023 5:32 pm |
|
|
wow, good guess !!
magical number 113355
divide left 3 digits into the right 3 digits...
355/113 = Pi, good to 6 or 7 places.
It's the best integer approximation I've seen since the mid/late '50s.. |
|
|
|
FAQ 
Search
Register
Profile 
Log in to check your private messages 
Log in
