Post by thread:Getting an "not in bank 0" error that I can not ge

Hello Mathew. > Getting an "not in bank 0" error that I can not get rid of. The assembler is brainless. It has *no idea* what bank you may be in, or even should be in. It generates errors literally, just to keep you on your toes! There is a way to do it, but for the moment I'd just forget it. I mailed a message with the commands to suppress that message and another even sillier one IIRC. Just use them and carefully set up the banks the way *you* think they should be after careful consideration. And do learn to use the TRIS instruction that makes it all far easier in the first place. I'll analyse your code much more carefully after I have a nap and feel better. -- Cheers, Paul B.

Paul; I remember this issue over safe use of high and low bank naming in assembly, so I asked the supreme professor, Andrew Warren. His answer to me is below. Chris Eddy Hi, Chris. Actually, I DISCOURAGED the practice of doing the XOR in the file-register equate; here are the four methods that are commonly used, along with my reasons for using Method #4 over the others: Note: In the following examples, "XXX" is a register address on page 1 (i.e., between 0x80 and 0xFF). METHOD 1: REG EQU XXX-128 ;Generates no warnings, but .... ;also doesn't allow REG to MOVWF REG ;be viewed in MPASM watch ;windows. Equivalent to: ; REG EQU XXX & 0x7F ; and ; REG EQU XXX ^ 0x80 ;both of which have the same ;problem. METHOD 2: REG EQU XXX ;Generates an MPASM warning. .... ; MOVWF REG ; METHOD 3: REG EQU XXX ;Generates no warning, EVEN .... ;when accidentally used with MOVWF REG & 0x7F ;a page-0 register. METHOD 4: REG EQU XXX ;Generates no warning when .... ;used correctly (i.e., when MOVWF REG ^ 0x80 ;REG is on page 1), but DOES ;generate a warning when ;accidentally used with a ;page-0 register.

Chris Eddy wrote: > Paul; I remember this issue over safe use of high and low bank naming > in assembly, so I asked the supreme professor, Andrew Warren. His > answer to me is below. He didn't mention using the TRIS instruction and just doing it properly? Since you ask. The *really* neat way is to use a variable called, say, RBANK, and using MOVWF REG^RBANK RBANK is set to 0 at the start of the code, and you use two macros: BANK0 MACRO RBANK SET 0 BCF STATUS,RB0 ENDM BANK1 MACRO RBANK SET 0x80 BSF STATUS,RB0 ENDM And use them as commands to set and clear the RB bit where you think they are needed in the code. If you now get an error, it implies you didn't set it right though this can be confused by program flow. -- Cheers, Paul B.