Additions to InFile and ObjectLoader...

[Ken Cline]

Hi all,

Here's a small method that might be useful to add to the
InFile class:
 
-(int) unGetWord: (char *) aWord {
   int i;
 
   if ( !aWord ) return 0;
 
   for ( i = strlen( aWord ) - 1; i >= 0; i-- ) {
      if ( [ self unGetChar: aWord[ i ] ] == 0 )
         return 0;
   }
 
   return 1 ; 
}

I added this to my InFile class because I was combining the
setup paramters for several objects into the same file.  For
example my setup files look like:

      @Begin Producer
         agentName foo
         ...
      @End Producer

      @Begin Producer
         agentName bar
         ...
      @End Producer

      @Begin Consumer
         agentName foobar
         ...
      @End Consumer

and so on...

In my model swarm, I read the "@Begin" and then the object
type (eg "Producer", "Consumer" etc.).  Next, I move the
file pointer backwards two words so that it is pointing at
the "@Begin". Then, based on the object type, I call the
appropriate method to create a new instance and load that
new instance with the ObjectLoader.

I've also modified my objectloader so that if it encounters
an "@Begin" while loading an object, for example:

     @Begin Producer
        agentName foo
        ...
        @Begin
           methodname argument
           methodname argument
           ...
        @End
        ...
     @End Producer

then it assumes everything between the inner "@Begin" and
"@End" are method names followed by arguments.  I use the
`perform' method to call the appropriate method for the
object being loaded.

Here is my modification of the ObjectLoader:

-loadObject: anObject {
     ...
  SEL aSEL;
     ...
    if( (aString[0] == '@') && 
        (aString[1] == 'E') &&
        (aString[2] == 'n') &&
        (aString[3] == 'd') ) {
      [theFileObject skipLine] ; //Note: if this fails we don't care...
                                 //(users may or may not have a newline
                                 //character at the end of their last 
                                 //object specification (last line in file).   
      break ;
    }

    if( (aString[0] == '@') && 
        (aString[1] == 'B') &&
        (aString[2] == 'e') &&
        (aString[3] == 'g') &&
        (aString[4] == 'i') &&
        (aString[5] == 'n')    )
    {
      if(![theFileObject skipLine])
        [self _crash_: anObject] ;   

      while(1) {
        while(1){ 

          if(!([theFileObject getChar: &aChar]))
            [self _crash_: anObject] ;

          if(aChar == '#'){
            if(![theFileObject skipLine])
              [self _crash_: anObject] ;
          } else {
            if(![theFileObject unGetChar: aChar])
              [self _crash_: anObject] ;
            break ;
          }
        }

        if(![theFileObject getWord: aString])
          [self _crash_: anObject] ;
   
        if( (aString[0] == '@') && 
            (aString[1] == 'E') &&
            (aString[2] == 'n') &&
            (aString[3] == 'd') ) {
          if(![theFileObject skipLine])
            [self _crash_: anObject] ;   
          break ;
        }

        aSEL = sel_get_uid( aString ); 
        if( ![anObject respondsTo: aSEL] )
          [self _crash_: anObject] ;

        if(![theFileObject getWord: aString])
          [self _crash_: anObject] ;
       
        [ anObject perform: aSEL with: (id)aString ];

        if(![theFileObject skipLine])
          [self _crash_: anObject] ;   
      }
      continue ;
    }

    aProbe = [aProbeMap getProbeForVariable: aString] ; 
    if( !aProbe )
      [self _crash_: anObject] ;

      ...

  return self ;  
}


As you can see most of this is cut-n-paste of what was more
or less already there.  I just chose the quick way to
implement this stuff.

Well, I hope these will be useful to some of y'all.

Ken.


PS: Oh yeah, I forgot to mention that I switched from
"@begin" and "@end" to "@Begin" and "@End", just so you
know.  Maybe ObjectLoader could have some char *'s called
"beginString" and "endString" and could use strncasecmp to
compare the input to these values?

Something like:

   char * beginString;
   char * endString;
   int n, m;

   beginString = "@begin";
   endString = "@end";
   n = strlen( beginString );
   m = strlen( endString );
   
   if ( strlen(aString) >= n && 
        strncasecmp(beginString, aString, n) == 0 ) {
      ...
   }

   if ( strlen(aString) >= m && 
        strncasecmp(endString, aString, m) == 0 ) {
      ...
   }

Whatca think?

_________________________________________________________
Ken Cline                             address@hidden
SAIC                                 VOICE (410) 571-0413
Annapolis, MD                          FAX (301) 261-8427


                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.
                  ==================================