I2C and PORTB outpts.... how to change output_XXXX()??

[Mazza365]

Using a 16F88 with the I2C as a slave, I have found that using the CCS built in functions for outputting on other PORTB pins (output_high() and output_low()) they use the BSF/BCF instructions when modifying the TRISB register. Because these are Read-Modify-Write instructions, it corrupts the status of the I2C pins direction reigster bits when I2C communications are happening. The only way around this is to use the method shown in example 10-1 on page 92 of the current datasheet. (Read TRISB, modify it for the pin being changed to output, sets I2C pins back to outputs, write the new value back using MOVWF). This ensures the I2C direction bits are always left as 1's.

My question is: Is it possible to easily re-define the CCS functions output_high() and output_low() so that they use this method that does not interfere with the I2C direction bits?
Will I have to write my own code and redefine the functions using something like this:
#define output_high( my_output_high(
#define output_low( my_output_low(

Or is there some other way I can get the compiler to compile using different code that does not use BCF and BSF when changing TRISB (thats if CCS have written this different code already)?

Thanks.

[PCM programmer]

They already do it for Port C. Look at the generated code for
the test program shown below. They maintain a shadow register
for TRISC that is called "TRIS_C". They do all bit operations on it
and then copy it to the physical TRISC register. This eliminates
the Read-Modify-Write problems that might occur because of the
peripherals that override the TRIS register. You could do something
similar to this.

[Mazza365]

Great. Thats roughly what I have done now. The way i've written it looks far from pretty and is specific to my application, but it works!
I've also E-Mailed CCS to see if they will add a 'No RMW on TRISB'...now waiting to see what they say.

Thankyou!