struct definition problem

"bit" is not a standard C type. Even if two different compilers both happen to choose that word as a keyword for an extension, they will most likely implement their extensions in different ways. If you use non-standard extensions, then you'll have to do a bit of work to port between compilers.

With Keil C, the "bit" type is a way to declare an individual variable that resides in the bit-addressable memory region of the 8051. Variables of "bit" cannot be members of a structure. As the manual says for error 150, "A union-aggregate may not contain members of type bit". (Seems pretty clear to me, if blunt.) Other limitations of the bit type are discussed in the manual in Chapter 3, "Bit Types", page 98.

If you want to access data as either a byte or bits, and want to use the 8051's bit-addressable memory to hold that byte, you probably want to use the Keil "bdata" memory type and "sbit" extensions. The "badata" memory qualifer locates variables in the bit-addressable memory. The "sbit" type is a way to declare a bit that is part of an object in bdata. So, one possible port of your code might look something like:

char bdata cValue;
sbit cValueBit0 = cValue ^ 0;
sbit cValueBit1 = cValue ^ 1;
sbit cValueBit2 = cValue ^ 2;
sbit cValueBit3 = cValue ^ 3;
sbit cValueBit4 = cValue ^ 4;
sbit cValueBit5 = cValue ^ 5;
sbit cValueBit6 = cValue ^ 6;
sbit cValueBit7 = cValue ^ 7;

"If the data you want to access doesn't reside in the bit addressable memory (or an SFR), but is just an 'ordinary' variable in data or xdata, then you'll need to use those mechanisms instead of 'bit' or 'sbit'."

OR copy the data to a bit-addressable temporary location, and work with that?

>Usually I'm using this to access each bit
>separatly in a byte, rather than using
>shifting ans masking. In my MPLAB C its
>working fine. But when using this
>declaration in KEIL I get the error:

As said by others, the "bit" directive is compiler specific.

PIC processors can manipulate bits in every register, so naturally the PIC compiler allows you to define bit variables anywhere in memory.

does anyone know a better and easy way to access the value of one bit in a byte ?

Use a bit-field?

Jon

The bit field method would look something like this:

struct structSplitByte
{
    unsigned char    b0:1;
    unsigned char    b1:1;
    unsigned char    b2:1;
    unsigned char    b3:1;
    unsigned char    b4:1;
    unsigned char    b5:1;
    unsigned char    b6:1;
    unsigned char    b7:1;
};

union unionSplit
{
    struct structSplitByte Split;
    char cValue;
};

"...is potentially difficult to port as C does not specify the order of the bit fields in memory"

and has its own difficulties in C51, as Keil's implementation can be somewhat unexpected!

Search the knowledgebase (and this forum) as there are articles on the common pitfalls!

Another reason to stick with the masks...?