[fasmg] struc ? and struc name

Index > Macroinstructions > [fasmg] struc ? and struc name

Author

Thread

zhak

Joined: 12 Apr 2005
Posts: 501
Location: Belarus

zhak 02 Jun 2017, 11:30

Long story short, having the below definitions, why match type more, def: is not processed for named struc defined from struct macro -- I never reach here! point from struct LINE

Code:

define @mode
@mode = 64

; ---------------------------------------------------

macro struct name*
  local n

  display 'struct ', `name,13,10

  define name
  namespace name
  virtual at 0
  n = 0
  struc (__name__) ? def&
    local m

    display '    ', `__name__,0x20, `def,13,10

    match type more, def:

      display 'here!',13,10

      __name__ def
    end match
  end struc

  macro end?.struct?!
    restruc ?
    end namespace
    sizeof.name = $
    end virtual

    struc (__name__) name def&
      display '>>', `__name__, 0x20, `name,13,10
    end struc

    purge end?.struct?
  end macro
end macro

struct POINT
  x intn
  y intn
end struct

struct LINE
  startpoint POINT
  endpoint POINT
end struct


;------------------ just helpers below -------------


element datatype
element generic:  datatype
element aligned:  datatype
element natural:  datatype

macro align? n*, x:?
  if n > 0
    emit 1: (n - 1) - ($ + n - 1) mod (n) dup (x)
  else if n <> 0
    err 'invalid argument'
  end if
end macro

macro datatypes meta, def&
  irp <name, size>, def
    element name: meta + size
    macro name args:?&
      local m, n
      m = name metadata 1
      n = m scale 0
      if n = 0
        n = @mode / 8
      end if
      if m relativeto aligned | m relativeto natural
        align n
      end if
      match [value], args
        emit n: (0 + value) dup ?
      else
        emit n: args
      end match
    end macro
    struc (__name__) name args:?&
      local m, n
      m = name metadata 1
      n = m scale 0
      if n = 0
        n = @mode / 8
      end if
      if m relativeto aligned | m relativeto natural
        align n
      end if
      label .: n
      match [value], args
        emit n: (0 + value) dup ?
      else
        emit n: args
      end match
      __name__.__size = n
      __name__.__length = $ - .
    end struc
  end irp
end macro

datatypes aligned, int8,1, int16,2, int32,4, int64,8
datatypes natural, intn,0
; ---------------------------------------------------

02 Jun 2017, 11:30

Tomasz Grysztar

Joined: 16 Jun 2003
Posts: 8353
Location: Kraków, Poland

Tomasz Grysztar 02 Jun 2017, 12:18

You have POINT expression-class symbol defined as a symbolic variable with empty text. So when you do:

Code:

match type more, POINT:

the text that is matched becomes simply ":", since the "POINT" is replaced with empty text. You can detect this by replacing your debugging display just above that MATCH with this one:

Code:

    match any, def:
        display ' -  ',`__name__,0x20,`any,13,10
    end match

To match the text without performing replacements of symbolic variables you can use the old trick from fasm 1:

Code:

    local proxy
    define proxy def:
    match type more, proxy

PS. Just in case you thought that you could use RAWMATCH to do it without proxy: I would not recommend it, because it also discards any recognition context from the text it matches. RAWMATCH is not a shortcut for the above construction, it is only useful when you're interested in the text as-is and not the values of symbols that the text may represent.

02 Jun 2017, 12:18

zhak

Joined: 12 Apr 2005
Posts: 501
Location: Belarus

zhak 02 Jun 2017, 12:42

Ah, that define thing. Thanks, Tomasz.
If I do

Code:

define name name

could there be any problems? This seems to resolve the issue

02 Jun 2017, 12:42

Tomasz Grysztar

Joined: 16 Jun 2003
Posts: 8353
Location: Kraków, Poland

Tomasz Grysztar 02 Jun 2017, 12:49

zhak wrote:

Ah, that define thing. Thanks, Tomasz.
If I do
Code:
define name name
    
could there be any problems? This seems to resolve the issue

That's a clever trick. This would create a circular reference if the symbol was used in an expression, but as this symbol is not intended to be used in an expression anyway, I think it might be even better than defining it with empty value - because when it has empty value you can put it into an expression with no actual effect, and in this case it would at least signal some error (circular reference).

02 Jun 2017, 12:49

< Last Thread | Next Thread >

Forum Rules:

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum