Bug#18201: XML: ExtractValue works even if the xml

fragment is not well-formed xml

Problem:
- ExtractValue silently returned NULL if a wrong XML value is passed.
- In some cases "unexpected END-OF-INPUT" error was not detected, and
  a non-NULL result could be returned for a bad XML value.

Fix:
- Adding warning messages, to make user aware why NULL was returned.
- Missing "unexpected END-OF-INPUT" error is reported now.


mysql-test/r/xml.result:
  - Fixing XML systax error in old test
  - Adding test cases.
mysql-test/t/xml.test:
  - Fixing XML systax error in old test
  - Adding test cases.
sql/item_xmlfunc.cc:
  Produce warning in case of XML systax error,
  instead of silentrly returning NULL.
strings/xml.c:
  - Making error messages better looking and clearer:
  It is important because now they're seen in SHOW WARNINGS
  (previously they were used only for debugging purposes).
  - Adding "unexpected END-OF-INPUT" error if after scanning
    closing tag for the root element some input is left
    (previously this error was ignored in a mistake).

1 files changed, 19 insertions, 8 deletions

diff --git a/strings/xml.c b/strings/xml.c
index 4c09e85604d..76fdd07f25e 100644
--- a/strings/xml.c
+++ b/strings/xml.c
@@ -43,7 +43,7 @@ static const char *lex2str(int lex)
 {
   switch(lex)
   {
-    case MY_XML_EOF:      return "EOF";
+    case MY_XML_EOF:      return "END-OF-INPUT";
     case MY_XML_STRING:   return "STRING";
     case MY_XML_IDENT:    return "IDENT";
     case MY_XML_CDATA:    return "CDATA";
@@ -195,8 +195,13 @@ static int my_xml_leave(MY_XML_PARSER *p, const char *str, uint slen)
   if (str && (slen != glen))
   {
     mstr(s,str,sizeof(s)-1,slen);
-    mstr(g,e+1,sizeof(g)-1,glen),
-    sprintf(p->errstr,"'</%s>' unexpected ('</%s>' wanted)",s,g);
+    if (glen)
+    {
+      mstr(g,e+1,sizeof(g)-1,glen),
+      sprintf(p->errstr,"'</%s>' unexpected ('</%s>' wanted)",s,g);
+    }
+    else
+      sprintf(p->errstr,"'</%s>' unexpected (END-OF-INPUT wanted)", s);
     return MY_XML_ERROR;
   }
   
@@ -247,7 +252,7 @@ int my_xml_parse(MY_XML_PARSER *p,const char *str, uint len)
       {
         if (MY_XML_IDENT != (lex=my_xml_scan(p,&a)))
         {
-          sprintf(p->errstr,"1: %s unexpected (ident wanted)",lex2str(lex));
+          sprintf(p->errstr,"%s unexpected (ident wanted)",lex2str(lex));
           return MY_XML_ERROR;
         }
         if (MY_XML_OK != my_xml_leave(p,a.beg,(uint) (a.end-a.beg)))
@@ -275,7 +280,7 @@ int my_xml_parse(MY_XML_PARSER *p,const char *str, uint len)
       }
       else
       {
-        sprintf(p->errstr,"3: %s unexpected (ident or '/' wanted)",
+        sprintf(p->errstr,"%s unexpected (ident or '/' wanted)",
 		lex2str(lex));
         return MY_XML_ERROR;
       }
@@ -297,7 +302,7 @@ int my_xml_parse(MY_XML_PARSER *p,const char *str, uint len)
           }
           else
           {
-            sprintf(p->errstr,"4: %s unexpected (ident or string wanted)",
+            sprintf(p->errstr,"%s unexpected (ident or string wanted)",
 		    lex2str(lex));
             return MY_XML_ERROR;
           }
@@ -325,7 +330,7 @@ gt:
       {
         if (lex != MY_XML_QUESTION)
         {
-          sprintf(p->errstr,"6: %s unexpected ('?' wanted)",lex2str(lex));
+          sprintf(p->errstr,"%s unexpected ('?' wanted)",lex2str(lex));
           return MY_XML_ERROR;
         }
         if (MY_XML_OK != my_xml_leave(p,NULL,0))
@@ -341,7 +346,7 @@ gt:
       
       if (lex != MY_XML_GT)
       {
-        sprintf(p->errstr,"5: %s unexpected ('>' wanted)",lex2str(lex));
+        sprintf(p->errstr,"%s unexpected ('>' wanted)",lex2str(lex));
         return MY_XML_ERROR;
       }
     }
@@ -359,6 +364,12 @@ gt:
       }
     }
   }
+
+  if (p->attr[0])
+  {
+    sprintf(p->errstr,"unexpected END-OF-INPUT");
+    return MY_XML_ERROR;
+  }
   return MY_XML_OK;
 }