From: dxforth@gmail.com   
      
   On 6/05/2025 4:58 pm, Ruvim wrote:   
   > On 2025-05-06 06:55, dxf wrote:   
   >> On 6/05/2025 3:29 am, Ruvim wrote:   
   >>> On 2025-05-05 08:11, dxf wrote:   
   >>>> ...   
   >>>> However modifying it to do so gives these results:   
   >>>>   
   >>>> 1 .( a ) cr -56 throw .( b )   
   >>>>   
   >>>>     'a' is displayed and the stack is: 1   
   >>>   
   >>> This violates the behavior of `throw` specified in 9.6.1.2275,   
   >>> because if there is no a user's exception frame, the data stack must be   
   emptied.   
   >>>   
   >>> The table 9.1 in Forth-2012 (or 9.2 in Forth-94) does not affect the   
   behavior of `throw` at all.   
   >>   
   >> Except THROW special-cased codes -1 and -2 effectively reserving their   
   behaviour.   
   >> One can speculate why ANS didn't do so for -56 but to my mind QUIT when   
   implemented   
   >> should function as expected i.e. -56 THROW uncaught should perform QUIT.   
   >>   
   >   
   > You suggest to make -56 a special case in `throw`.  But what is the   
   rationale for this?   
      
   So there's no confusion as to what happen in the case of -56 THROW.  This   
   doesn't mean   
   systems must implement : QUIT -56 THROW ;   
      
   > Yes, -1 and -2 are special cases in 9.6.1.2275 (when there is no user   
   exception frame). But for those cases it *only* specifies what message shall   
   be displayed, and *nothing more*. The standard should have specified this   
   messages anyway, and the only    
   choice was which section to specify it in.   
   >   
   >   
   > --   
   > Ruvim   
      
   --- SoupGate-DOS v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)