MADS assembler, why no .DBYTE?

[Atari_Ace]

I've been reconstructing DOS 1.0 source code from DOS 2.0 and a disassembly, and running the result through MADS assembler to validate the binary is correct. In one place, the DOS 2.0 source code uses big-endian 16-bit integers, e.g.

 

.DBYTE XRENAME-1

 

which I can simulate in MADS via

 

.BYTE >(XRENAME-1),<(XRENAME-1)

 

but I was expecting it would be there like .WORD, since MAC/65 supports .DBYTE and already supports big-endian 32-bit integers.

 

The code does support big endian 32-bit (via format 'R'), so this is a curious omission.

 

BTW, MADS is overall an awesome assembler, which accepts syntax close enough to the various Atari assemblers that I can use it with only minor adjustments. Just providing some feedback to hopefully make it even more useful.

 

 

 

[Heaven/TQA]

but there is a .word?

[Atari_Ace]

.WORD is little endian

[Rybags]

Strange... the DTA directive allows type g which is double-word (32-bit) in big endian order but no provision for 16-bit words in big-endian order.

 

It might be easiest to just create a Macro which does the work for you rather than having the annoyance of doing .BYTE >(label),<(label)

[tebe]

Mad-Assembler, 32-bit:

 

.DWORD

 

dta f(xxxx)

 

dta r(xxxx) ; reverse 32bit

 

 

p.s.

Mad-Pascal, 32bit: CARDINAL, INTEGER

 

p.s. #2 ATASM instruction

 

.DBYTE <words>

Stores words in memory at the current memory address in MSB/LSB format.

Ex:

.DBYTE $1234,-1,1

will generate:

12 34 FF FF 00 01

.WORD <words>

Stores words in memory at the current memory address in native format

(LSB/MSB).

p.s. #3

probably I did not understand this .DBYTE and I created .DB (simple .BYTE)

Edited March 28, 2016 by tebe

[+Stephen]

Strange... the DTA directive allows type g which is double-word (32-bit) in big endian order but no provision for 16-bit words in big-endian order.

 

It might be easiest to just create a Macro which does the work for you rather than having the annoyance of doing .BYTE >(label),<(label)

Forgive the dumb question, but why would you do something big-endian on the 6502?

[tebe]

MADS (2.0.4 b7) + .DBYTE

 

mads_dbyte.zip

[Atari_Ace]

Thanks tebe,

 

As for why .DBYTE, in DOS.SYS, it's used to implement a jump table compactly using RTS, i.e.

 

ASL A
TAY

LDA DVDCVT,Y
PHA ;PUSH EXECUTION ADDR
LDA DVDCVT+1,Y
PHA
RTS​

 

It could have used .WORD and switched the LDA lines, but arguably this is more natural.