From ad6f7847317b0a46224204d795d055ba775d103f Mon Sep 17 00:00:00 2001
From: Laurence Withers <lwithers@lochside.guralp.local>
Date: Mon, 8 Jan 2007 12:26:55 +0000
Subject: [PATCH] Fix handling of restart markers. csxml_reset() was clearing
 the state after encountering a restart marker, leaving us in stream level
 after <![RESTART[, which meant that the closing bit of the marker was
 interpreted as stream-level content. Also, the state machine now handles
 restart markers at high level as well as low level; this simply means that a
 (false) error is not issued in the case of perfectly well-formed Streamed XML
 containing restart markers.

---
 src/libCStreamedXML/parser.c | 29 +++++++++++++++++++++++------
 1 file changed, 23 insertions(+), 6 deletions(-)

diff --git a/src/libCStreamedXML/parser.c b/src/libCStreamedXML/parser.c
index 2e7ae84..4fa66f4 100644
--- a/src/libCStreamedXML/parser.c
+++ b/src/libCStreamedXML/parser.c
@@ -23,7 +23,9 @@ enum csxml_State {
     StateRestartMarker2, // after second ]
     StateOpen, // after <
     StateOpenBang, // after <!
-    StateOpenCdataMarker, // after <![
+    StateOpenBangMarker, // after <![
+    StateOpenCdataMarker, // after <![C
+    StateOpenRestartMarker, // after <![R
     StateCDATA,
     StateCDATA1, // first ]
     StateCDATA2, // second ]
@@ -110,7 +112,7 @@ int csxml_feedChar(struct csxml* ctx, char ch)
     int try;
 
 #ifdef DEBUG_STATE_MACHINE
-    printf("%p: Character '%c' (%02X), state %d, buffer %s\n", ctx, ch, ch, ctx->state, ctx->buffer.data);
+    printf("%p: Character '%c' (%02X), state %d, buffer `%s', restartCount %d\n", ctx, ch, ch, ctx->state, ctx->buffer.data, ctx->restartCount);
 #endif
 
 #define ERROR(_reason) do { \
@@ -141,8 +143,8 @@ int csxml_feedChar(struct csxml* ctx, char ch)
 
     if(ch == xmlRestartMarker[ctx->restartCount]) {
         if(++ctx->restartCount == 11) {
-            ctx->state = StateRestartMarker;
             csxml_reset(ctx);
+            ctx->state = StateRestartMarker;
             return 0;
         }
     } else {
@@ -343,19 +345,34 @@ int csxml_feedChar(struct csxml* ctx, char ch)
 
     case StateOpenBang:
         if(ch == '[') {
-            // restart markers handled by lower layer
-            ctx->state = StateOpenCdataMarker;
+            ctx->state = StateOpenBangMarker;
             ctx->xmlCount = 3;
-            if(!ctx->elementDepth) ERROR("CDATA sections not valid at stream level.");
         } else if(ch == '-') ctx->state = StateOpenComment;
         else ERROR("Invalid special tag.");
         break;
 
+    case StateOpenBangMarker:
+        if(ch == xmlCdataMarker[ctx->xmlCount]) {
+            ctx->state = StateOpenCdataMarker;
+            if(!ctx->elementDepth) ERROR("CDATA sections not valid at stream level.");
+        } else if(ch == xmlRestartMarker[ctx->xmlCount]) {
+            ctx->state = StateOpenRestartMarker;
+        } else {
+            ERROR("Invalid marked section.");
+        }
+        ++(ctx->xmlCount);
+        break;
+
     case StateOpenCdataMarker:
         if(ch != xmlCdataMarker[ctx->xmlCount]) ERROR("Invalid marked section.");
         if(!xmlCdataMarker[++ctx->xmlCount]) ctx->state = StateCDATA;
         break;
 
+    case StateOpenRestartMarker:
+        if(ch != xmlRestartMarker[ctx->xmlCount++]) ERROR("Invalid marked section.");
+        /* Restart markers handled by lower layer -- if(!xmlRestartMarker[++ctx->xmlCount]) { } */
+        break;
+
     case StateOpenComment:
         if(ch != '-') ERROR("Invalid special tag.");
         ctx->state = StateComment;