Changeset 8719 for trunk/MagicSoft/Mars/mdata

Timestamp:

08/28/07 13:03:04 (18 years ago)

Author:

tbretz

Message:

*** empty log message ***

File:

• trunk/MagicSoft/Mars/mdata/MDataPhrase.cc (modified) (1 diff)

trunk/MagicSoft/Mars/mdata/MDataPhrase.cc

@@ -321 +321 @@
         idx++;
     }
-
+    /*
+
+    // * HOW DO WE PROCESS THE FILTERS?
+    // * DO WE NEED THIS FOR MData derived classes? Is there any need for it?
+    // * MAYBE FIRST FILTERS (MF?) MUST REPLACE {name[class]} BEFORE
+    //   THE DATA PHRASSE IS CREATED?
+    // --> MFDataPhrase must have a list of MFilter. In Process first
+    //     all self created MFilter are processed (see MF). Then
+    //     they are evaluated and the result is given to the MDataPhrase.
+    //     Can this be done using MDataMember? We replace {Name[class]}
+    //     by Name.IsExpressionTrue and we need a way that MDataPhrase
+    //     gets the coresponding pointer.
+    // --> Alternatively we can create a MDataPhrase which allows
+    //     Pre/Processing
+    //
+    // We convert {Name[Class]} to Name.IsExpressionTrue. To process these
+    // data/filters we need a request from MFDataPhrase (new virtual
+    // memeber function?)
+    //
+    // {} Is alreaqdy used in ReadEnv.
+    //
+    // Enhance ReadEnv to allow Cut1.5 to be just a class.
+    //
+    // The difference between MFDataPhrase is
+    //   MFDataPhrase only knows MDataPhrase, while
+    //   MF also knows a filter-class.
+    //
+
+    p = 0;
+
+    // And now we check for other phrases or filters
+    // They are defined by a [, a pribtable character and
+    // any numer of word characters (a-zA-Z0-9_) and a closing ]
+    reg = TPRegexp("{[A-Za-z}\\w+(\\[[A-Za-z]\\w+\\])?}");
+    while (1)
+    {
+        // If some indices are already existing
+        // initialize them by a MDataValue
+        idx = CheckForVariable(phrase, idx);
+
+        // Check whether expression is found
+        if (reg.Match(phrase, mods, p, 130, &pos)==0)
+            break;
+
+        // Get expression from phrase
+        TString expr = phrase(pos[0], pos[1]-pos[0]);
+
+        // Zerlegen: {Name[Class]}
+
+        // Create a new MData object of kind
+        MData *dat = (MData*)GetNewObject(cls, MData::Class());
+        if (!dat)
+            return "";
+        dat->SetName(name);
+
+        // Add a corresponding MDataMember to our list
+        fMembers.AddLast(dat);
+
+        // Make the expression "Regular expression proofed"
+        expr.ReplaceAll("[", "\\[");
+        expr.ReplaceAll("]", "\\]");
+
+        // Find other occurances of arg by this regexp
+        // and start next search behind first match
+        p = pos[0] + Substitute(phrase, expr, idx);
+
+        // Step forward to the next argument
+        idx++;
+    }
+ */
     // Now we have to check if there are additional indices [idx]
     // This is mainly important if the rule has indices only!