directory-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
Subject svn commit: rev 20557 - in incubator/directory/snickers/trunk: ber-codec/src/java/org/apache/snickers/ber/digester/rules xdocs/ber-codec
Date Fri, 28 May 2004 23:18:41 GMT
Author: akarasulu
Date: Fri May 28 16:18:39 2004
New Revision: 20557

some docs were garbage

Modified: incubator/directory/snickers/trunk/ber-codec/src/java/org/apache/snickers/ber/digester/rules/
--- incubator/directory/snickers/trunk/ber-codec/src/java/org/apache/snickers/ber/digester/rules/
+++ incubator/directory/snickers/trunk/ber-codec/src/java/org/apache/snickers/ber/digester/rules/
Fri May 28 16:18:39 2004
@@ -18,6 +18,8 @@
 import org.apache.snickers.ber.TypeClass ;
+import org.apache.snickers.ber.digester.BERDigester ;
+import org.apache.snickers.ber.digester.AbstractRule ;
 import org.apache.commons.lang.exception.NestableRuntimeException ;

Modified: incubator/directory/snickers/trunk/xdocs/ber-codec/BERDigesterDesign.xml
--- incubator/directory/snickers/trunk/xdocs/ber-codec/BERDigesterDesign.xml	(original)
+++ incubator/directory/snickers/trunk/xdocs/ber-codec/BERDigesterDesign.xml	Fri May 28 16:18:39
@@ -232,34 +232,9 @@
           Now the question is how do we search with a wildcard in the front or
-          at the tail end of the pattern.  Perhaps the easiest of the two to
-          handle would be a wildcard in the tail end of the pattern.  A search
-          could be conducted using everything but the wildcard as if it were a
-          regular search without wildcards.  If this lands on a TagNode then
-          all the rules of that node and that nodes children have matched been
-          matched by the pattern.  The other use case, #1, is not as easy to
-          implement.  For this another special TagTree could be built, however
-          it would be assembled in reverse order using only the tails of those
-          patterns starting with a wildcard.  If for example *-4-8-9 and
-          *-6-9-1 patterns are used to register rules r1, and r2 respectively
-          then the root node would contain two child nodes one for a tag int
-          equal to 9 and another equal to 1.  These two rule pattern would
-          result in two tree branches.  Again when adding new patterns we reuse
-          what already exists before branching.  To search these trees for a
-          match the reverse nesting pattern is used to walk the tree.  Take
-          for example the pattern 1-8-9.  This would start off at the root,
-          select the child node with a tag int equal to 9, then select the
-          child under node 9 with a tag int equal to 8.  At this point we can
-          go no further.  At this point a check is performed to see if the node
-          we are stuck at is a leaf node.  If it is a leaf node then we have a
-          match for the rules at that node, if not then no rules were matched.
-          For 1-8-9 no rules are matched.  Now use 1-6-4-8-9 as the nesting
-          pattern to test.  Using the same algorithm we find ourselves stuck
-          walking the tree at node 4 which is a leaf node.  In this case the
-          rules in node 4 have been matched by the nesting pattern.  Keep in
-          mind that the search against the reverse TagTree is conducted in
-          addition to the search against the forward TagTree.  Its easy to see
-          that this use case is far more painful to implement.
+          at the tail end of the pattern.

View raw message