You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@xerces.apache.org by pe...@apache.org on 2003/11/05 19:19:09 UTC
cvs commit: xml-xerces/c/src/xercesc/framework XMLGrammarPool.hpp
peiyongz 2003/11/05 10:19:09
Modified: c/src/xercesc/framework XMLGrammarPool.hpp
Log:
Documentation update
Revision Changes Path
1.7 +23 -13 xml-xerces/c/src/xercesc/framework/XMLGrammarPool.hpp
Index: XMLGrammarPool.hpp
===================================================================
RCS file: /home/cvs/xml-xerces/c/src/xercesc/framework/XMLGrammarPool.hpp,v
retrieving revision 1.6
retrieving revision 1.7
diff -u -r1.6 -r1.7
--- XMLGrammarPool.hpp 29 Oct 2003 16:16:48 -0000 1.6
+++ XMLGrammarPool.hpp 5 Nov 2003 18:19:09 -0000 1.7
@@ -56,6 +56,9 @@
/*
* $Log$
+ * Revision 1.7 2003/11/05 18:19:09 peiyongz
+ * Documentation update
+ *
* Revision 1.6 2003/10/29 16:16:48 peiyongz
* GrammarPool' serialization/deserialization support
*
@@ -267,31 +270,30 @@
* 1. Context: Serialize/Deserialize All Grammars In One Session
*
* Since it is common that a declaration in one grammar may reference
- * to definitions in another grammar, it is required to serialize those
+ * to definitions in other grammars, it is required to serialize those
* related (or interdependent) grammars in to one persistent data store
* in one serialization session (storing), and deserialize them from the
* persistent data store in one deserialization session (loading) back
* to the grammar pool.
*
- * 2. Multiple serialization
+ * 2. Multiple serializations
*
* It is acceptable that client application requests more than one
- * gramamr serialization on a particular grammar pool, to track the
+ * grammar serialization on a particular grammar pool, to track the
* different grammars cached, or for whatever reasons that client
- * application is interested in. However it is strongly recommended that
- * the client application requests no more than one gramamr deserialization
- * and only does it when the grammar pool is empty.
- *
- * For multiple serializations, if the same file name is given, then the
- * last result will be in the file (overwriting mode), if different file
- * names are given, then we have multiple data stores for each serialization.
+ * application is interested in.
*
- * 3. Multiple deserialization
+ * 3. Multiple deserializations
*
* Request for grammar deserialization either after the grammar pool has
* its own cached grammars, or request for more than one grammar
* deserialization, may cause undesired and unpredictable consequence
- * and therefore is NOT supported.
+ * and therefore client application shall be aware that individual
+ * implementationis may NOT support this.
+ *
+ * However it is strongly recommended that the client application requests
+ * no more than one grammar deserialization even a particular implementation
+ * may allow multiple deserializations.
*
* 4. Locking
*
@@ -306,6 +308,14 @@
* deserialization, thus a grammar pool may decide if it supports
* a binary data created by a different release of Xerces.
*
+ * 6. Clean up
+ *
+ * The client application shall be aware that in the event of an exception
+ * thrown due to a corrupted data store during deserialization, implementation
+ * may not be able to clean up all resources allocated, and therefore it is
+ * client application's responsibility to clean up those unreleased resources.
+ *
+ *
*/
virtual void serializeGrammars(BinOutputStream* const) = 0;
virtual void deserializeGrammars(BinInputStream* const) = 0;
---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-cvs-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-cvs-help@xml.apache.org