How to eliminate the need for @@@ in PropBASIC ?

Bean

Here is some code showing how PropBASIC uses the @ @ @ operator.

'  hubVar1    HUB   LONG
'
  mov __temp,#0    ' WRLONG hubVar1,0
  wrlong __temp,__hubVar_adr
'
__hubVar_adr   LONG @ @ @ hubVar1

DAT
hubVar1   LONG  0
'

How can the use of @ @ @ be avoided ?
I am open to any ideas.

Bean

Dave Hein

In the top object "@ @ @ hubvar" is equivalent to "@ hubvar + 16". However, this only works in the top object. In the past, if I needed absolute address in lower objects and I wanted to avoid the @ @ @ operator I would adjust the addresses in memory at startup.

Dave Hein

Here's how CSPIN handles a list of strings in C, such as

static char *ops[] = {
    "++", "+=", "+", "--", "-=", "-", ">>=", ">>",
    ">-=", ">-", "><=", "><", ">|", ">=", ">", "<<=",
    "<<", "<-=", "<-", "<#=", "<#", "<>=", "<>", "<=",
    "<", "**=", "**", "*=", "*", "//=", "//", "/=",
    "/", "||", "|<", "|=", "|", " @ @ ", " @ ", "#>=",
    "#>", "^^", "^=", "^", "=>=", "=>", "=<=", "=<",
    "!", "~~", "~>=", "~>", "~", "?", "===", "==", 0};

/* INLINE SPIN
PUB start | ptr
  ptr := @ops
  repeat while long[ptr]
    long[ptr] += @ @0 - 16
    ptr += 4
*/

It produces the following Spin code

DAT
  ops long
  long @datstr + 16, @datstr + 19, @datstr + 22, @datstr + 24, @datstr + 27, @datstr + 30, @datstr + 32, @datstr + 36
  long @datstr + 39, @datstr + 43, @datstr + 46, @datstr + 50, @datstr + 53, @datstr + 56, @datstr + 59, @datstr + 61
  long @datstr + 65, @datstr + 68, @datstr + 72, @datstr + 75, @datstr + 79, @datstr + 82, @datstr + 86, @datstr + 89
  long @datstr + 92, @datstr + 94, @datstr + 98, @datstr + 101, @datstr + 104, @datstr + 106, @datstr + 110, @datstr + 113
  long @datstr + 116, @datstr + 118, @datstr + 121, @datstr + 124, @datstr + 127, @datstr + 129, @datstr + 132, @datstr + 134
  long @datstr + 138, @datstr + 141, @datstr + 144, @datstr + 147, @datstr + 149, @datstr + 153, @datstr + 156, @datstr + 160
  long @datstr + 163, @datstr + 165, @datstr + 168, @datstr + 172, @datstr + 175, @datstr + 177, @datstr + 179, @datstr + 183, 0

  datstr byte "++", 0, "+=", 0, "+", 0, "--", 0, "-=", 0, "-", 0, ">>=", 0, ">>", 0
  byte ">-=", 0, ">-", 0, "><=", 0, "><", 0, ">|", 0, ">=", 0, ">", 0, "<<=", 0
  byte "<<", 0, "<-=", 0, "<-", 0, "<#=", 0, "<#", 0, "<>=", 0, "<>", 0, "<=", 0
  byte "<", 0, "**=", 0, "**", 0, "*=", 0, "*", 0, "//=", 0, "//", 0, "/=", 0
  byte "/", 0, "||", 0, "|<", 0, "|=", 0, "|", 0, " @ @ ", 0, " @ ", 0, "#>=", 0
  byte "#>", 0, "^^", 0, "^=", 0, "^", 0, "=>=", 0, "=>", 0, "=<=", 0, "=<", 0
  byte "!", 0, "~~", 0, "~>=", 0, "~>", 0, "~", 0, "?", 0, "===", 0, "==", 0

' INLINE SPIN
PUB start | ptr
  ptr := @ops
  repeat while long[ptr]
    long[ptr] += @ @0 - 16
    ptr += 4

The start method is not needed for the top object. It is only needed for the lower objects. The start method is generated using CSPIN's inline directive. I never implemented an automatic way of generating the start method.

Bean

It was a long time ago, but I think the issue was I wasn't sure the +16 would ALWAYS be valid.
I think you are saying that it is, but only for the top object.
If that is true, I can try using the +16 and see if it works.

Thanks for the tip, I will try it.

Bean

Bean

Dave,
Can you explain this line ?

 long[ptr] += @ @0 - 16

Bean

ersmith

There are a number of compilers that support @ @ @: bstc (you already know about that one), homespun, and fastspin, at least. homespun and fastspin are open source, so neither of those is going away anytime soon. I have used fastspin to successfully compile PropBASIC output.

Dave Hein

Bean wrote: »

Dave,
Can you explain this line ?

 long[ptr] += @ @0 - 16

Bean

Cspin adds 16 for the top object, so when adjusting for a lower object I have to subtract 16 and add the lower object's offset. The "@ @ " operator is used to add the object offset to a value, so "@ @ 0" is the object offset plus zero. I could have written this expression as "@ @ (-16)" to make it more efficient.