pdp-11 assembly standards

[Brent Hilpert]

On 2017-Jan-09, at 6:27 AM, Pete Lancashire wrote:
> wow ... the memories ... someday I've got to get a PDP-11 again :-).
> 
> had most of the opcodes memorized, for a story ....
> 
> Had a coworker who played the piano, he could enter/patch code from
> the 11/35's panel from memory so fast all you saw was a blur.
> 
> When we replacing the 11/35's with 11/34A he hated it.
> 
> -pete
> 
> On Mon, Jan 9, 2017 at 6:13 AM, Paul Koning <paulkoning at comcast.net> wrote:
>> 
>>> On Jan 9, 2017, at 12:38 AM, Don North <north at alum.mit.edu> wrote:
>>> 
>>> On 1/8/2017 9:10 PM, Brent Hilpert wrote:
>>>> OK, what was the standard (if there was one) number-base syntax for PDP-11 assembler?
>>>> 
>>>> Despite all the PDP-11 assembly info on web sites, this seems to be a buried bit of info.
>>>> One assembler doc uses a prefix of "&o", another specifies octal as default and prefix of zero for decimal (opposite of the common C-derived standard . . great).
>>>> 
>>>> Is this for example standard?:
>>>> 
>>>>             BIT #&o200, @#&o177564          ; test 2^7 bit at address octal 177564
>>>> 
>>>> (I'm just trying to make some written commentary consistent with common policy.)
>>>> 
>>>> 
>>> MACRO11 Language Manual v5.5 section 6.4
>>> 
>>> All numbers are octal radix, unless the default radix is changed via the .RADIX N directive (N can be 2, 8, 10, or 16). N blank resets the radix to octal.
>>> 
>>> So 0100, 100 would be octal 100, decimal value 64.
>>> 
>>> Any number followed by a period (decimal point) is forced to be base 10.
>>> 
>>> So 100. would be decimal 100, octal 144.
>>> 
>>> Prefix operators ^B (binary), ^O (octal), ^D (decimal), ^X (hexadecimal) force the following digits/characters to the designated radix.
>>> 
>>> So ^B101000 == ^O50 == ^D40 == ^X28 all represent the same value (decimal 40.) irrespective of the current .RADIX N setting.
>> 
>> I don't remember ^X.  Other ways to specify numeric values is with prefix ' (single quote) for a single byte value, i.e., 'x is the ASCII code for character x.  Similarly, "xy is a 16 bit value for the two-character sequence xy (little endian).  And ^Rxyz is the RAD50 coded value for the three characters xyz.
>> 
>> &o doesn't match anything I've ever seen, not even in the wildly different world of Unix.


So the answer is, by modern expectations the old standard would be ambiguous or misleading.
I guess I should just comment it.