Even better imho... use the standard types in <stdint.h>:
typedef signed char int8_t;
typedef signed int int16_t;
typedef signed long int int32_t;
typedef unsigned char uint8_t;
typedef unsigned int uint16_t;
typedef unsigned long int uint32_t;

These make it explicit the size of the int you intend to use.  C tried=20
to make it so that the compiler could pick for you, to some degree=20
anyway, the size of "int" so you didn't force (or guess) what size would=20
run fastest on whatever computer your code was being compiled for.  This=20
is pretty irrelevant nowadays, and usually one cares about the size of=20
an int (for anything low-level or microcontroller-CPU-scale) more than a=20
little more speed.

I don't know if you're using floats, but XC8 gives you the choice of=20
24-bit or 32-bit floats.  24-bit pulls in less library code, runs=20
faster, and uses 1 byte less storage than 32-bit, but you'll have to=20
decide the precision vs. (everything else) tradeoff.

J


Isaac Marino Bavaresco wrote:
> The first struct is anonymous but its sub-fields are not. The second
> struct is also anonymous but its only field is anonymous also, so you
> don't have a means to reference it.
>
> You must give a name to the bit-field. Better, you could also use a
> 'char' field instead of the second struct.
>
> One more point, 'unsigned' is a 16-bit type, if you want your union to
> occupy just one byte, declare the fields as 'unsigned char'.
>
>
> Isaac
>
>
> On 18/03/2015 21:24, Josh Koffman wrote:
>> Hi all,
>>
>> I'm trying to define a status register in XC8. I'd like to be able to
>> reference individual bits, but also be able to say, clear the variable
>> as a whole.
>>
>> I currently have it defined as this:
>>
>> typedef union {
>>          struct
>>          {
>>              unsigned ENC1UP                    :1;
>>              unsigned ENC1DN                    :1;
>>              unsigned ENC2UP                    :1;
>>              unsigned ENC2DN                    :1;
>>              unsigned ENC3UP                    :1;
>>              unsigned ENC3DN                    :1;
>>              unsigned ENC4UP                    :1;
>>              unsigned ENC4DN                    :1;
>>          };
>>          struct
>>          {
>>              unsigned                            :8;
>>          };
>>
>> } EncStatus_t;
>>
>> In another part of the code, I define a variable using the typedef EncSt=
atus_t.
>>
>> In my code, I have two lines that try to write to that variable:
>>
>>     EncStatus.ENC1DN =3D 1;
>>     EncStatus =3D 0;                  // Clear encoder status register
>>
>> The first line works perfect. The second line throws an error though -
>> illegal conversion between types - int -> volatile union.
>>
>> I understand the basic problem here, but I can't seem to fix it. I've
>> spent a bunch of time looking at the header file for my processor, to
>> see how I can write to both PORTB and PORTBbits.RB0 successfully. Near
>> as I can tell, it's because both of those variables are explicitly
>> being declared at the same address, not because there's anything fancy
>> in the union.
>>
>> So...is there a way to deal with this when I don't know where the
>> variable will be assigned?
>>
>> To be clear, this isn't a work stopping problem, I'd just like to know
>> if there's a way to do this that is efficient. After all, I can always
>> just clear each bit individually.
>>
>> Thanks!
>>
>> Josh
>
--=20
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.