This is the mail archive of the ecos-patches@sourceware.org mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Ezxml upgrade to v0.8.5


Hi,

This patch upgrades the services/ezxml package to upstream version
0.8.5.

This version has a number of bugfixes and new features compared to the
current 0.8.1 version. From the change log:

ezXML 0.8.5
- fixed ezxml_toxml() to not output siblings of tag being converted
- fixed a segfault when ezxml_set_attr() was used on a new root tag
- added ezxml_name() function macro
- all external functions now handle NULL ezxml_t structs without segfaulting

ezXML 0.8.4
- fixed to compile under win-doze when NOMMAP make option is set
- fixed a bug where ezxml_toxml() could segfault if tag offset is out of bounds
- ezxml_add_child() now works properly when tags are added out of order
- improved error messages now include line numbers
- fixed memory leak when entity reference is shorter than replacement text
- added ezxml_new_d(), ezxml_add_child_d(), ezxml_set_txt_d() and 
  ezxml_set_attr_d() function macros as wrappers that strdup() their arguments

ezXML 0.8.3
- fixed a UTF-16 decoding bug affecting larger unicode values
- added internal dtd processing for entity declarations and default attributes
- now correctly normalizes attribute values in compliance with the XML 1.0 spec
- added check for correct tag nesting
- ezxml_toxml() now generates canonical xml (apart from the namespace stuff)

ezXML 0.8.2
- fixed compiler warning about lvalue type casting
- ezxml_get() argument list can now be terminated by an empty string tag name
- added NOMMAP make option for systems without posix memory mapping
- added support for UTF-16
- fixed bug in ezxml_toxml() where UTF-8 sequences were being ampersand encoded
- added ezxml_new(), ezxml_add_child(), ezxml_set_txt(), ezxml_set_attr(),
  and ezxml_remove() to facilitate creating and modifying xml

Index: packages/services/ezxml/current/ChangeLog
===================================================================
RCS file: /cvs/ecos/ecos/packages/services/ezxml/current/ChangeLog,v
retrieving revision 1.1
diff -u -r1.1 ChangeLog
--- packages/services/ezxml/current/ChangeLog	31 Jan 2005 00:09:30 -0000	1.1
+++ packages/services/ezxml/current/ChangeLog	23 Jan 2006 15:33:08 -0000
@@ -1,3 +1,11 @@
+2006-01-23  Peter Korsgaard  <jacmet@sunsite.dk>
+
+	* src/ezxml.c:
+	* include/ezxml.h:
+	* doc/ezxml.txt:
+	* doc/ezxml.html:
+	* doc/license.txt: Upgrade to ezxml-0.8.5
+
 2005-01-31  Andrew Lunn  <andrew.lunn@ascom.ch>
 
 	* Import of the ezxml XML parsing library
Index: packages/services/ezxml/current/doc/ezxml.html
===================================================================
RCS file: /cvs/ecos/ecos/packages/services/ezxml/current/doc/ezxml.html,v
retrieving revision 1.1
diff -u -r1.1 ezxml.html
--- packages/services/ezxml/current/doc/ezxml.html	31 Jan 2005 00:09:31 -0000	1.1
+++ packages/services/ezxml/current/doc/ezxml.html	23 Jan 2006 15:33:08 -0000
@@ -4,20 +4,21 @@
   <head><title>ezXML</title></head>
   <body>
     <h1>ezXML - XML Parsing C Library</h1>
-    <h3>version 0.8</h3>
+    <h3>version 0.8.5</h3>
     <p>
       ezXML is a C library for parsing XML documents inspired by
       <a href="http://www.php.net/SimpleXML";>simpleXML</a> for
-      PHP. As the name implies, it's easy to use. It's ideal for parsing xml
+      PHP. As the name implies, it's easy to use. It's ideal for parsing XML
       configuration files or REST web service responses. It's also fast and
-      lightweight (11k compiled). The latest version is available here:
-      <a href="http://prdownloads.sf.net/ezxml/ezxml-0.8.tar.gz?download";
-	 >ezxml-0.8.tar.gz</a>
+      lightweight (less than 20k compiled). The latest version is available
+      here:
+      <a href="http://prdownloads.sf.net/ezxml/ezxml-0.8.5.tar.gz?download";
+	 >ezxml-0.8.5.tar.gz</a>
     </p>
 
     <b>Example Usage</b>
     <p>
-      Given the following example xml document:
+      Given the following example XML document:
     </p>
     <code>
       &lt;?xml version="1.0"?&gt;<br />
@@ -37,7 +38,7 @@
       &lt;/formula1&gt;
     </code>
     <p>
-      This code snipped prints out a list of drivers, which team they drive for,
+      This code snippet prints out a list of drivers, which team they drive for,
       and how many championship points they have:
     </p>
     <code>
@@ -58,7 +59,7 @@
     </code>
     <p>
       Alternately, the following would print out the name of the second driver
-      of the first team:
+      on the first team:
     </p>
     <code>
       ezxml_t f1 = ezxml_parse_file("formula1.xml");<br />
@@ -74,14 +75,12 @@
     <b>Known Limitations</b>
     <ul>
       <li>
-	No support for UTF-16, however UTF-8 is handled correctly. UTF-16
-	support is required for XML 1.0 conformity and will be implimented for
-	the 1.0 release.
+	ezXML is not a validating parser.
 	<br />&nbsp;
       </li>
       <li>
-	Loads the entire xml document into memory at once and does not allow for
-	documents to be passed in a chunk at a time. Large xml files can still
+	Loads the entire XML document into memory at once and does not allow for
+	documents to be passed in a chunk at a time. Large XML files can still
 	be handled though through <code>ezxml_parse_file()</code> and 
 	<code>ezxml_parse_fd()</code>, which use mmap to map the file to a
 	virtual address space and rely on the virtual memory system to page in
@@ -89,9 +88,10 @@
 	<br />&nbsp;
       </li>
       <li>
-	Ignores DTDs. Parsing of the internal DTD subset is required for XML 1.0
-	conformity and will be implimented for the 1.0 release. ezXML is not,
-	and is not likely to become, a validating parser.
+	Does not currently recognize all possible well-formedness errors. It
+	should correctly handle all well-formed XML documents and will either
+	ignore or halt XML processing on well-formedness errors. More
+	well-formedness checking will be added in subsiquent releases.
 	<br />&nbsp;
       </li>
       <li>
@@ -106,7 +106,7 @@
 	  <code>"line one\nline two"</code>, and <code>&lt;br/&gt;</code> is
 	  reported as a sub tag, but the location of <code>&lt;br/&gt;</code>
 	  within the character data is not. The function
-	  <code>ezxml_toxml()</code> will convert an ezXML structure back to xml
+	  <code>ezxml_toxml()</code> will convert an ezXML structure back to XML
 	  with sub tag locations intact.
 	</p>
       </li>
Index: packages/services/ezxml/current/doc/ezxml.txt
===================================================================
RCS file: /cvs/ecos/ecos/packages/services/ezxml/current/doc/ezxml.txt,v
retrieving revision 1.1
diff -u -r1.1 ezxml.txt
--- packages/services/ezxml/current/doc/ezxml.txt	31 Jan 2005 00:09:31 -0000	1.1
+++ packages/services/ezxml/current/doc/ezxml.txt	23 Jan 2006 15:33:08 -0000
@@ -1,15 +1,15 @@
 ezXML - XML Parsing C Library
-version 0.8
+version 0.8.5
 
 ezXML is a C library for parsing XML documents inspired by simpleXML for PHP.
-As the name implies, it's easy to use. It's ideal for parsing xml configuration
-files or REST web service responses. It's also fast and lightweight (11k
-compiled).  The latest version is available here:
-http://prdownloads.sf.net/ezxml/ezxml-0.8.tar.gz?download
+As the name implies, it's easy to use. It's ideal for parsing XML configuration
+files or REST web service responses. It's also fast and lightweight (less than
+20k compiled). The latest verions is available here:
+http://prdownloads.sf.net/ezxml/ezxml-0.8.5.tar.gz?download
 
 Example Usage
 
-Given the following example xml document:
+Given the following example XML document:
 
 <?xml version="1.0"?>
 <formula1>
@@ -25,7 +25,7 @@
   </team>
 </formula1>
 
-This code snipped prints out a list of drivers, which team they drive for,
+This code snippet prints out a list of drivers, which team they drive for,
 and how many championship points they have:
 
 ezxml_t f1 = ezxml_parse_file("formula1.xml"), team, driver;
@@ -40,7 +40,7 @@
 }
 ezxml_free(f1);
 
-Alternately, the following would print out the name of the second driver of the
+Alternately, the following would print out the name of the second driver on the
 first team:
 
 ezxml_t f1 = ezxml_parse_file("formula1.xml");
@@ -53,18 +53,18 @@
 
 Known Limitations
 
-- No support for UTF-16, however UTF-8 is handled correctly. UTF-16 support is
-  required for XML 1.0 conformity and will be implimented for the 1.0 release.
+- ezXML is not a validating parser
 
-- Loads the entire xml document into memory at once and does not allow for
-  documents to be passed in a chunk at a time. Large xml files can still be
+- Loads the entire XML document into memory at once and does not allow for
+  documents to be passed in a chunk at a time. Large XML files can still be
   handled though through ezxml_parse_file() and ezxml_parse_fd(), which use mmap
   to map the file to a virtual address space and rely on the virtual memory
   system to page in data as needed.
 
-- Ignores DTDs. Parsing of the internal DTD subset is required for XML 1.0
-  conformity and will be implimented for the 1.0 release. ezXML is not, and is
-  not likely to become, a validating parser.
+- Does not currently recognize all possible well-formedness errors. It should
+  correctly handle all well-formed XML documents and will either ignore or halt
+  XML processing on well-formedness errors. More well-formedness checking will
+  be added in subsiquent releases.
 
 - In making the character content of tags easy to access, there is no way
   provided to keep track of the location of sub tags relative to the character
@@ -76,7 +76,7 @@
   The character content of the doc tag is reported as "line one\nline two", and
   <br/> is reported as a sub tag, but the location of <br/> within the
   character data is not. The function ezxml_toxml() will convert an ezXML
-  structure back to xml with sub tag locations intact.
+  structure back to XML with sub tag locations intact.
 
 Licensing
 
Index: packages/services/ezxml/current/doc/license.txt
===================================================================
RCS file: /cvs/ecos/ecos/packages/services/ezxml/current/doc/license.txt,v
retrieving revision 1.1
diff -u -r1.1 license.txt
--- packages/services/ezxml/current/doc/license.txt	31 Jan 2005 00:09:31 -0000	1.1
+++ packages/services/ezxml/current/doc/license.txt	23 Jan 2006 15:33:08 -0000
@@ -1,4 +1,4 @@
-Copyright 2004 Aaron Voisine <aaron@voisine.org>
+Copyright 2004, 2005 Aaron Voisine <aaron@voisine.org>
 
 Permission is hereby granted, free of charge, to any person obtaining
 a copy of this software and associated documentation files (the
Index: packages/services/ezxml/current/include/ezxml.h
===================================================================
RCS file: /cvs/ecos/ecos/packages/services/ezxml/current/include/ezxml.h,v
retrieving revision 1.1
diff -u -r1.1 ezxml.h
--- packages/services/ezxml/current/include/ezxml.h	31 Jan 2005 00:09:31 -0000	1.1
+++ packages/services/ezxml/current/include/ezxml.h	23 Jan 2006 15:33:08 -0000
@@ -54,7 +54,7 @@
 
 /* ezxml.h
  *
- * Copyright 2004 Aaron Voisine <aaron@voisine.org>
+ * Copyright 2004, 2005 Aaron Voisine <aaron@voisine.org>
  *
  * Permission is hereby granted, free of charge, to any person obtaining
  * a copy of this software and associated documentation files (the
@@ -88,29 +88,25 @@
 extern "C" {
 #endif
 
-#define EZXML_BUFSIZE 1024
-
-// returns the next tag of the same name in the same section and depth or NULL
-// if not found
-#define ezxml_next(xml) xml->next
-
-// returns the tag character content or empty string if none
-#define ezxml_txt(xml) xml->txt
+#define EZXML_BUFSIZE 1024 // size of internal memory buffers
+#define EZXML_NAMEM   0x80 // name is malloced
+#define EZXML_TXTM    0x40 // txt is malloced
+#define EZXML_DUP     0x20 // attribute name and value are strduped
 
 typedef struct ezxml *ezxml_t;
 struct ezxml {
     char *name;      // tag name
     char **attr;     // tag attributes { name, value, name, value, ... NULL }
     char *txt;       // tag character content, empty string if none
-    size_t off;      // tag offset in parent tag character content
+    size_t off;      // tag offset from start of parent tag character content
     ezxml_t next;    // next tag with same name in this section at this depth
     ezxml_t sibling; // next tag with different name in same section and depth
     ezxml_t ordered; // next tag, same section and depth, in original order
     ezxml_t child;   // head of sub tag list, NULL if none
     ezxml_t parent;  // parent tag, NULL if current tag is root tag
-    short flags;     // additional information, only used internally for now
+    short flags;     // additional information
 };
-    
+
 // Given a string of xml data and its length, parses it and creates an ezxml
 // structure. For efficiency, modifies the data by adding null terminators
 // and decoding ampersand sequences. If you don't want this, copy the data and
@@ -129,20 +125,31 @@
 // stream into memory and then parses it. For xml files, use ezxml_parse_file()
 // or ezxml_parse_fd()
 ezxml_t ezxml_parse_fp(FILE *fp);
-    
-// returns the first child tag (one level deeper) with the given name or NULL if
-// not found
+
+// returns the first child tag (one level deeper) with the given name or NULL
+// if not found
 ezxml_t ezxml_child(ezxml_t xml, const char *name);
 
+// returns the next tag of the same name in the same section and depth or NULL
+// if not found
+#define ezxml_next(xml) ((xml) ? xml->next : NULL)
+
 // Returns the Nth tag with the same name in the same section at the same depth
 // or NULL if not found. An index of 0 returns the tag given.
 ezxml_t ezxml_idx(ezxml_t xml, int idx);
 
+// returns the name of the given tag
+#define ezxml_name(xml) ((xml) ? xml->name : NULL)
+
+// returns the given tag's character content or empty string if none
+#define ezxml_txt(xml) ((xml) ? xml->txt : "")
+
 // returns the value of the requested tag attribute, or NULL if not found
 const char *ezxml_attr(ezxml_t xml, const char *attr);
 
-// Traverses the ezxml sturcture to retrive a specific subtag. Takes a variable
-// length list of tag names and indexes. Final index must be -1. Example: 
+// Traverses the ezxml sturcture to retrieve a specific subtag. Takes a
+// variable length list of tag names and indexes. The argument list must be
+// terminated by either an index of -1 or an empty string tag name. Example: 
 // title = ezxml_get(library, "shelf", 0, "book", 2, "title", -1);
 // This retrieves the title of the 3rd book on the 1st shelf of library.
 // Returns NULL if not found.
@@ -162,6 +169,41 @@
 // returns parser error message or empty string if none
 const char *ezxml_error(ezxml_t xml);
 
+// returns a new empty ezxml structure with the given root tag name
+ezxml_t ezxml_new(const char *name);
+
+// wrapper for ezxml_new() that strdup()s name
+#define ezxml_new_d(name) ezxml_set_flag(ezxml_new(strdup(name)), EZXML_NAMEM)
+
+// Adds a child tag. off is the offset of the child tag relative to the start
+// of the parent tag's character content. Returns the child tag.
+ezxml_t ezxml_add_child(ezxml_t xml, const char *name, size_t off);
+
+// wrapper for ezxml_add_child() that strdup()s name
+#define ezxml_add_child_d(xml, name, off) \
+    ezxml_set_flag(ezxml_add_child(xml, strdup(name), off), EZXML_NAMEM)
+
+// sets the character content for the given tag and returns the tag
+ezxml_t ezxml_set_txt(ezxml_t xml, const char *txt);
+
+// wrapper for ezxml_set_txt() that strdup()s txt
+#define ezxml_set_txt_d(xml, txt) \
+    ezxml_set_flag(ezxml_set_txt(xml, strdup(txt)), EZXML_TXTM)
+
+// Sets the given tag attribute or adds a new attribute if not found. A value
+// of NULL will remove the specified attribute.
+void ezxml_set_attr(ezxml_t xml, const char *name, const char *value);
+
+// Wrapper for ezxml_set_attr() that strdup()s name/value. Value cannot be NULL
+#define ezxml_set_attr_d(xml, name, value) \
+    ezxml_set_attr(ezxml_set_flag(xml, EZXML_DUP), strdup(name), strdup(value))
+
+// sets a flag for the given tag and returns the tag
+ezxml_t ezxml_set_flag(ezxml_t xml, short flag);
+
+// removes a tag along with all its subtags
+void ezxml_remove(ezxml_t xml);
+
 #ifdef __cplusplus
 }
 #endif
Index: packages/services/ezxml/current/src/ezxml.c
===================================================================
RCS file: /cvs/ecos/ecos/packages/services/ezxml/current/src/ezxml.c,v
retrieving revision 1.1
diff -u -r1.1 ezxml.c
--- packages/services/ezxml/current/src/ezxml.c	31 Jan 2005 00:09:31 -0000	1.1
+++ packages/services/ezxml/current/src/ezxml.c	23 Jan 2006 15:33:09 -0000
@@ -1,8 +1,8 @@
 //==========================================================================
 //
-//      crc32.c
+//      ezxml.h
 //
-//      Gary S. Brown's 32 bit CRC
+//      Simple XML parser
 //
 //==========================================================================
 //####ECOSGPLCOPYRIGHTBEGIN####
@@ -52,10 +52,9 @@
 //
 //==========================================================================
 
-
 /* ezxml.c
  *
- * Copyright 2004 Aaron Voisine <aaron@voisine.org>
+ * Copyright 2004, 2005 Aaron Voisine <aaron@voisine.org>
  *
  * Permission is hereby granted, free of charge, to any person obtaining
  * a copy of this software and associated documentation files (the
@@ -77,6 +76,8 @@
  * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
 
+#define EZXML_NOMMAP // no mmap on eCos
+
 #include <stdlib.h>
 #include <stdio.h>
 #include <stdarg.h>
@@ -84,37 +85,40 @@
 #include <ctype.h>
 #include <unistd.h>
 #include <sys/types.h>
-#ifndef __ECOS__
+#ifndef EZXML_NOMMAP
 #include <sys/mman.h>
-#endif
+#endif // EZXML_NOMMAP
 #include <sys/stat.h>
 #include "ezxml.h"
 
 #include <pkgconf/system.h>
 
-#define EZXML_TXTM      0x80 // flag value meaning txt was malloced
-#define EZXML_WS        "\t\r\n " // whitespace
-
-// called when parser finds close of tag
-#define ezxml_close_tag(root) root->cur = root->cur->parent;
+#define EZXML_WS   "\t\r\n "  // whitespace
+#define EZXML_ERRL 128        // maximum error string length
 
 typedef struct ezxml_root *ezxml_root_t;
-struct ezxml_root {   // additional data for the root tag
-    struct ezxml xml; // is a super-struct built on top of ezxml struct
-    ezxml_t cur;      // current xml tree insertion point
-    void *m;          // original xml string
-    size_t len;       // length of allocated memory for mmap, or -1 for malloc
-    const char *err;  // error string
-    char ***pi;       // processing instructions
+struct ezxml_root {       // additional data for the root tag
+    struct ezxml xml;     // is a super-struct built on top of ezxml struct
+    ezxml_t cur;          // current xml tree insertion point
+    char *m;              // original xml string
+    size_t len;           // length of allocated memory for mmap, -1 for malloc
+    char *u;              // UTF-8 conversion of string if original was UTF-16
+    char *s;              // start of work area
+    char *e;              // end of work area
+    char **ent;           // general entities (ampersand sequences)
+    char ***attr;         // default attributes
+    char ***pi;           // processing instructions
+    short standalone;     // non-zero if <?xml standalone="yes"?>
+    char err[EZXML_ERRL]; // error string
 };
 
+char *EZXML_NIL[] = { NULL }; // empty, null terminated array of strings
+
 // returns the first child tag with the given name or NULL if not found
 ezxml_t ezxml_child(ezxml_t xml, const char *name)
 {
-    if (! xml) return NULL;
-    xml = xml->child;
+    xml = (xml) ? xml->child : NULL;
     while (xml && strcmp(name, xml->name)) xml = xml->sibling;
-
     return xml;
 }
 
@@ -129,321 +133,528 @@
 // returns the value of the requested tag attribute or NULL if not found
 const char *ezxml_attr(ezxml_t xml, const char *attr)
 {
-    int i = 0;
+    int i = 0, j = 1;
+    ezxml_root_t root = (ezxml_root_t)xml;
 
-    if (! xml) return NULL;
+    if (! xml || ! xml->attr) return NULL;
     while (xml->attr[i] && strcmp(attr, xml->attr[i])) i += 2;
-    return (xml->attr[i]) ? xml->attr[i + 1] : NULL;
+    if (xml->attr[i]) return xml->attr[i + 1]; // found attribute
+
+    while (root->xml.parent) root = (ezxml_root_t)root->xml.parent; // root tag
+    for (i = 0; root->attr[i] && strcmp(xml->name, root->attr[i][0]); i++);
+    if (! root->attr[i]) return NULL; // no matching default attributes
+    while (root->attr[i][j] && strcmp(attr, root->attr[i][j])) j += 3;
+    return (root->attr[i][j]) ? root->attr[i][j + 1] : NULL; // found default
 }
 
-// same as ezxml_get but takes an alredy initialized va_list
+// same as ezxml_get but takes an already initialized va_list
 ezxml_t ezxml_vget(ezxml_t xml, va_list ap)
 {
     char *name = va_arg(ap, char *);
-    int idx = va_arg(ap, int);
+    int idx = -1;
 
-    xml = ezxml_child(xml, name);
+    if (name && *name) {
+        idx = va_arg(ap, int);    
+        xml = ezxml_child(xml, name);
+    }
     return (idx < 0) ? xml : ezxml_vget(ezxml_idx(xml, idx), ap);
 }
 
-// Traverses the xml tree to retrive a specific subtag. Takes a variable
-// length list of tag names and indexes. Final index must be -1. Example: 
+// Traverses the xml tree to retrieve a specific subtag. Takes a variable
+// length list of tag names and indexes. The argument list must be terminated
+// by either an index of -1 or an empty string tag name. Example: 
 // title = ezxml_get(library, "shelf", 0, "book", 2, "title", -1);
 // This retrieves the title of the 3rd book on the 1st shelf of library.
 // Returns NULL if not found.
 ezxml_t ezxml_get(ezxml_t xml, ...)
 {
     va_list ap;
-    ezxml_t ret;
+    ezxml_t r;
 
     va_start(ap, xml);
-    ret = ezxml_vget(xml, ap);
+    r = ezxml_vget(xml, ap);
     va_end(ap);
-
-    return ret;
+    return r;
 }
 
-// returns a NULL terminated array of processing instructions for the given
+// returns a null terminated array of processing instructions for the given
 // target
 const char **ezxml_pi(ezxml_t xml, const char *target)
 {
-    static const char *nopi = NULL;
-    ezxml_root_t root;
+    ezxml_root_t root = (ezxml_root_t)xml;
     int i = 0;
 
-    while (xml->parent) xml = xml->parent;
-    root = (ezxml_root_t)xml;
-
-    if (! root->pi) return &nopi;
-    while (root->pi[i] && strcmp(target, root->pi[i][0])) i++;
-    return (root->pi[i]) ? (const char **)root->pi[i] + 1 : &nopi;
+    if (! root) return (const char **)EZXML_NIL;
+    while (root->xml.parent) root = (ezxml_root_t)root->xml.parent; // root tag
+    while (root->pi[i] && strcmp(target, root->pi[i][0])) i++; // find target
+    return (const char **)((root->pi[i]) ? root->pi[i] + 1 : EZXML_NIL);
 }
 
-// Converts \r or \r\n to a single \n. If decode is non-zero, decodes ampersand
-// sequences in place. Returns s.
-char *ezxml_decode(char *s, int decode)
+// set an error string and return root
+ezxml_t ezxml_err(ezxml_root_t root, char *s, const char *err, ...)
 {
-    int b;
-    char *e, *ret = s;
-    long c, d;
+    va_list ap;
+    int line = 1;
+    char *t, fmt[EZXML_ERRL];
+    
+    for (t = root->s; t < s; t++) if (*t == '\n') line++;
+    snprintf(fmt, EZXML_ERRL, "[error near line %d]: %s", line, err);
 
-    for (;;) {
-        while (*s && *s != '\r' && *s != '&') s++;
+    va_start(ap, err);
+    vsnprintf(root->err, EZXML_ERRL, fmt, ap);
+    va_end(ap);
 
-        if (! *s) return ret;
-        else if (*s == '\r') {
+    return &root->xml;
+}
+
+// Recursively decodes entity and character references and normalizes new lines
+// ent is a null terminated array of alternating entity names and values. set t
+// to '&' for general entity decoding, '%' for parameter entity decoding, 'c'
+// for cdata sections, ' ' for attribute normalization, or '*' for non-cdata
+// attribute normalization. Returns s, or if the decoded string is longer than
+// s, returns a malloced string that must be freed.
+char *ezxml_decode(char *s, char **ent, char t)
+{
+    char *e, *r = s, *m = s;
+    long b, c, d, l;
+
+    for (; *s; s++) { // normalize line endings
+        while (*s == '\r') {
             *(s++) = '\n';
-            if (*s == '\n') memmove((void *)s, (void *)(s + 1), strlen(s));
-            continue;
+            if (*s == '\n') memmove(s, (s + 1), strlen(s));
         }
-        else if (! decode) { s++; continue; }
-        else if (! strncmp(s, "&lt;", 4)) *(s++) = '<';
-        else if (! strncmp(s, "&gt;", 4)) *(s++) = '>';
-        else if (! strncmp(s, "&quot;", 6)) *(s++) = '"';
-        else if (! strncmp(s, "&apos;", 6)) *(s++) = '\'';
-        else if (! strncmp(s, "&amp;", 5)) s++;
-        else if (! strncmp(s, "&#", 2)) {
-            if (s[2] == 'x') c = strtol(s + 3, &e, 16);
-            else c = strtol(s + 2, &e, 10);
-            if (! c || *e != ';') { s++; continue; }
+    }
+    
+    for (s = r; ; ) {
+        while (*s && *s != '&' && (*s != '%' || t != '%') && !isspace(*s)) s++;
 
-            if (c < 0x80) *(s++) = (char)c; // US-ASCII subset
+        if (! *s) break;
+        else if (t != 'c' && ! strncmp(s, "&#", 2)) { // character reference
+            if (s[2] == 'x') c = strtol(s + 3, &e, 16); // base 16
+            else c = strtol(s + 2, &e, 10); // base 10
+            if (! c || *e != ';') { s++; continue; } // not a character ref
+
+            if (c < 0x80) *(s++) = c; // US-ASCII subset
             else { // multi-byte UTF-8 sequence
                 for (b = 0, d = c; d; d /= 2) b++; // number of bits in c
                 b = (b - 2) / 5; // number of bytes in payload
                 *(s++) = (0xFF << (7 - b)) | (c >> (6 * b)); // head
                 while (b) *(s++) = 0x80 | ((c >> (6 * --b)) & 0x3F); // payload
             }
+
+            memmove(s, strchr(s, ';') + 1, strlen(strchr(s, ';')));
         }
-        else { s++; continue; }
-        
-        memmove((void *)s, (void *)(strchr(s, ';') + 1), strlen(strchr(s, ';')));
-    }    
+        else if ((*s == '&' && (t == '&' || t == ' ' || t == '*')) ||
+                 (*s == '%' && t == '%')) { // entity reference
+            for (b = 0; ent[b] && strncmp(s + 1, ent[b], strlen(ent[b]));
+                 b += 2); // find entity in entity list
+
+            if (ent[b++]) { // found a match
+                if ((c = strlen(ent[b])) - 1 > (e = strchr(s, ';')) - s) {
+                    l = (d = (s - r)) + c + strlen(e); // new length
+                    r = (r == m) ? strcpy(malloc(l), r) : realloc(r, l);
+                    e = strchr((s = r + d), ';'); // fix up pointers
+                }
+
+                memmove(s + c, e + 1, strlen(e)); // shift rest of string
+                strncpy(s, ent[b], c); // copy in replacement text
+            }
+            else s++; // not a known entity
+        }
+        else if ((t == ' ' || t == '*') && isspace(*s)) *(s++) = ' ';
+        else s++; // no decoding needed
+    }
+
+    if (t == '*') { // normalize spaces for non-cdata attributes
+        for (s = r; *s; s++) {
+            if ((l = strspn(s, " "))) memmove(s, s + l, strlen(s + l) + 1);
+            while (*s && *s != ' ') s++;
+        }
+        if (--s >= r && *s == ' ') *s = '\0'; // trim any trailing space
+    }
+    return r;
 }
 
 // called when parser finds start of new tag
 void ezxml_open_tag(ezxml_root_t root, char *name, char **attr)
 {
     ezxml_t xml = root->cur;
+    
+    if (xml->name) xml = ezxml_add_child(xml, name, strlen(xml->txt));
+    else xml->name = name; // first open tag
 
-    if (xml->name) { // not root tag
-        if (xml->child) { // already have sub tags
-            xml = xml->child;
-            while (xml->ordered) xml = xml->ordered;
-            xml->ordered = (ezxml_t)malloc(sizeof(struct ezxml));
-            xml->ordered->parent = root->cur;
-            root->cur = xml->ordered;
-            xml = xml->parent->child;
-
-            while (strcmp(xml->name, name) && xml->sibling) xml = xml->sibling;
-            if (! strcmp(xml->name, name)) { // already have this tag type
-                while (xml->next) xml = xml->next;
-                xml = xml->next = root->cur;
-            }
-            else xml = xml->sibling = root->cur;
-        }
-        else { // first sub tag
-            xml->child = (ezxml_t)malloc(sizeof(struct ezxml));
-            xml->child->parent = xml;
-            root->cur = xml = xml->child;
-        }
-
-        xml->off = strlen(xml->parent->txt); // offset in parent char content
-    }
-
-    // initialize new tag
-    xml->name = name;
     xml->attr = attr;
-    xml->next = xml->child = xml->sibling = xml->ordered = NULL;
-    xml->txt = "";
-    xml->flags = 0;
+    root->cur = xml; // update tag insertion point
 }
 
 // called when parser finds character content between open and closing tag
-void ezxml_char_content(ezxml_root_t root, char *s, size_t len, short decode)
+void ezxml_char_content(ezxml_root_t root, char *s, size_t len, char t)
 {
     ezxml_t xml = root->cur;
+    char *m = s;
     size_t l;
 
-    if (! xml || ! xml->name || ! len) return;
+    if (! xml || ! xml->name || ! len) return; // sanity check
 
-    s[len] = '\0';
-    ezxml_decode(s, decode);
+    s[len] = '\0'; // null terminate text (calling functions anticipate this)
+    len = strlen(s = ezxml_decode(s, root->ent, t)) + 1;
 
-    if (! *(xml->txt)) xml->txt = s;
+    if (! *(xml->txt)) xml->txt = s; // initial character content
     else { // allocate our own memory and make a copy
-        l = strlen(xml->txt);
-        if (! (xml->flags & EZXML_TXTM)) {
-            xml->txt = strcpy((char *)malloc(l + len + 1), xml->txt);
-            xml->flags |= EZXML_TXTM;
-        }
-        else xml->txt = (char *)realloc((void *)(xml->txt), l + len + 1);
-        strcpy(xml->txt + l, s);
+        xml->txt = (xml->flags & EZXML_TXTM) // allocate some space
+                   ? realloc(xml->txt, (l = strlen(xml->txt)) + len)
+                   : strcpy(malloc((l = strlen(xml->txt)) + len), xml->txt);
+        strcpy(xml->txt + l, s); // add new char content
+        if (s != m) free(s); // free s if it was malloced by ezxml_decode()
+    }
+
+    if (xml->txt != m) ezxml_set_flag(xml, EZXML_TXTM);
+}
+
+// called when parser finds closing tag
+ezxml_t ezxml_close_tag(ezxml_root_t root, char *name, char *s)
+{
+    if (! root->cur || ! root->cur->name || strcmp(name, root->cur->name))
+        return ezxml_err(root, s, "unexpected closing tag </%s>", name);
+
+    root->cur = root->cur->parent;
+    return NULL;
+}
+
+// checks for circular entity references, returns non-zero if no circular
+// references are found, zero otherwise
+int ezxml_ent_ok(char *name, char *s, char **ent)
+{
+    int i;
+
+    for (; ; s++) {
+        while (*s && *s != '&') s++; // find next entity reference
+        if (! *s) return 1;
+        if (! strncmp(s + 1, name, strlen(name))) return 0; // circular ref.
+        for (i = 0; ent[i] && strncmp(ent[i], s + 1, strlen(ent[i])); i += 2);
+        if (ent[i] && ! ezxml_ent_ok(name, ent[i + 1], ent)) return 0;
     }
 }
 
-// called when the parser finds an xml processing instruction
+// called when the parser finds a processing instruction
 void ezxml_proc_inst(ezxml_root_t root, char *s, size_t len)
 {
     int i = 0, j = 1;
     char *target = s;
 
     s[len] = '\0'; // null terminate instruction
-    *(s += strcspn(s, EZXML_WS)) = '\0'; // null terminate target
-    s += strspn(s + 1, EZXML_WS) + 1; // skip whitespace after target
+    if (*(s += strcspn(s, EZXML_WS))) {
+        *s = '\0'; // null terminate target
+        s += strspn(s + 1, EZXML_WS) + 1; // skip whitespace after target
+    }
+
+    if (! strcmp(target, "xml")) { // <?xml ... ?>
+        if ((s = strstr(s, "standalone")) && ! strncmp(s + strspn(s + 10,
+            EZXML_WS "='\"") + 10, "yes", 3)) root->standalone = 1;
+        return;
+    }
 
-    if (! root->pi) *(root->pi = (char ***)malloc(sizeof(char**))) = NULL;
+    if (! root->pi[0]) *(root->pi = malloc(sizeof(char **))) = NULL; //first pi
 
-    while (root->pi[i] && strcmp(target, root->pi[i][0])) i++;
+    while (root->pi[i] && strcmp(target, root->pi[i][0])) i++; // find target
     if (! root->pi[i]) { // new target
-        root->pi = (char ***)realloc(root->pi, sizeof(char **) * (i + 2));
-        root->pi[i] = (char **)malloc(sizeof(char *) * 2);
+        root->pi = realloc(root->pi, sizeof(char **) * (i + 2));
+        root->pi[i] = malloc(sizeof(char *) * 3);
         root->pi[i][0] = target;
-        root->pi[i + 1] = NULL; // null terminate lists
-        root->pi[i][1] = (char *)root->pi[i + 1];
+        root->pi[i][1] = (char *)(root->pi[i + 1] = NULL); // terminate pi list
+        root->pi[i][2] = strdup(""); // empty document position list
     }
 
-    while (root->pi[i][j]) j++;
-    root->pi[i] = (char **)realloc(root->pi[i], sizeof(char *) * (j + 2));
-    root->pi[i][j] = s;
-    root->pi[i][j + 1] = NULL;
+    while (root->pi[i][j]) j++; // find end of instruction list for this target
+    root->pi[i] = realloc(root->pi[i], sizeof(char *) * (j + 3));
+    root->pi[i][j + 2] = realloc(root->pi[i][j + 1], j + 1);
+    strcpy(root->pi[i][j + 2] + j - 1, (root->xml.name) ? ">" : "<");
+    root->pi[i][j + 1] = NULL; // null terminate pi list for this target
+    root->pi[i][j] = s; // set instruction
 }
 
-// set an error string and return root
-ezxml_t ezxml_seterr(ezxml_root_t root, const char *err)
+// called when the parser finds an internal doctype subset
+short ezxml_internal_dtd(ezxml_root_t root, char *s, size_t len)
 {
-    root->err = err;
-    return (ezxml_t)root;
+    char q, *c, *t, *n = NULL, *v, **ent, **pe;
+    int i, j;
+    
+    pe = memcpy(malloc(sizeof(EZXML_NIL)), EZXML_NIL, sizeof(EZXML_NIL));
+
+    for (s[len] = '\0'; s; ) {
+        while (*s && *s != '<' && *s != '%') s++; // find next declaration
+
+        if (! *s) break;
+        else if (! strncmp(s, "<!ENTITY", 8)) { // parse entity definitions
+            c = s += strspn(s + 8, EZXML_WS) + 8; // skip white space separator
+            n = s + strspn(s, EZXML_WS "%"); // find name
+            *(s = n + strcspn(n, EZXML_WS)) = ';'; // append ; to name
+
+            v = s + strspn(s + 1, EZXML_WS) + 1; // find value
+            if ((q = *(v++)) != '"' && q != '\'') { // skip externals
+                s = strchr(s, '>');
+                continue;
+            }
+
+            for (i = 0, ent = (*c == '%') ? pe : root->ent; ent[i]; i++);
+            ent = realloc(ent, (i + 3) * sizeof(char *)); // space for next ent
+            if (*c == '%') pe = ent;
+            else root->ent = ent;
+
+            *(++s) = '\0'; // null terminate name
+            if ((s = strchr(v, q))) *(s++) = '\0'; // null terminate value
+            ent[i + 1] = ezxml_decode(v, pe, '%'); // set value
+            ent[i + 2] = NULL; // null terminate entity list
+            if (! ezxml_ent_ok(n, ent[i + 1], ent)) { // circular reference
+                if (ent[i + 1] != v) free(ent[i + 1]);
+                ezxml_err(root, v, "circular entity declaration &%s", n);
+                break;
+            }
+            else ent[i] = n; // set entity name
+        }
+        else if (! strncmp(s, "<!ATTLIST", 9)) { // parse default attributes
+            t = s + strspn(s + 9, EZXML_WS) + 9; // skip whitespace separator
+            if (! *t) { ezxml_err(root, t, "unclosed <!ATTLIST"); break; }
+            if (*(s = t + strcspn(t, EZXML_WS ">")) == '>') continue;
+            else *s = '\0'; // null terminate tag name
+            for (i = 0; root->attr[i] && strcmp(n, root->attr[i][0]); i++);
+
+            while (*(n = ++s + strspn(s, EZXML_WS)) && *n != '>') {
+                if (*(s = n + strcspn(n, EZXML_WS))) *s = '\0'; // attr name
+                else { ezxml_err(root, t, "malformed <!ATTLIST"); break; }
+
+                s += strspn(s + 1, EZXML_WS) + 1; // find next token
+                c = (strncmp(s, "CDATA", 5)) ? "*" : " "; // is it cdata?
+                if (! strncmp(s, "NOTATION", 8))
+                    s += strspn(s + 8, EZXML_WS) + 8;
+                s = (*s == '(') ? strchr(s, ')') : s + strcspn(s, EZXML_WS);
+                if (! s) { ezxml_err(root, t, "malformed <!ATTLIST"); break; }
+
+                s += strspn(s, EZXML_WS ")"); // skip white space separator
+                if (! strncmp(s, "#FIXED", 6))
+                    s += strspn(s + 6, EZXML_WS) + 6;
+                if (*s == '#') { // no default value
+                    s += strcspn(s, EZXML_WS ">") - 1;
+                    if (*c == ' ') continue; // cdata is default, nothing to do
+                    v = NULL;
+                }
+                else if ((*s == '"' || *s == '\'')  &&  // default value
+                         (s = strchr(v = s + 1, *s))) *s = '\0';
+                else { ezxml_err(root, t, "malformed <!ATTLIST"); break; }
+
+                if (! root->attr[i]) { // new tag name
+                    root->attr = (! i) ? malloc(2 * sizeof(char **))
+                                       : realloc(root->attr,
+                                                 (i + 2) * sizeof(char **));
+                    root->attr[i] = malloc(2 * sizeof(char *));
+                    root->attr[i][0] = t; // set tag name
+                    root->attr[i][1] = (char *)(root->attr[i + 1] = NULL);
+                }
+
+                for (j = 1; root->attr[i][j]; j += 3); // find end of list
+                root->attr[i] = realloc(root->attr[i],
+                                        (j + 4) * sizeof(char *));
+
+                root->attr[i][j + 3] = NULL; // null terminate list
+                root->attr[i][j + 2] = c; // is it cdata?
+                root->attr[i][j + 1] = (v) ? ezxml_decode(v, root->ent, *c)
+                                           : NULL;
+                root->attr[i][j] = n; // attribute name 
+            }
+        }
+        else if (! strncmp(s, "<!--", 4)) s = strstr(s + 4, "-->"); // comments
+        else if (! strncmp(s, "<?", 2)) { // processing instructions
+            if ((s = strstr(c = s + 2, "?>")))
+                ezxml_proc_inst(root, c, s++ - c);
+        }
+        else if (*s == '<') s = strchr(s, '>'); // skip other declarations
+        else if (*(s++) == '%' && ! root->standalone) break;
+    }
+
+    free(pe);
+    return ! *root->err;
+}
+
+// Converts a UTF-16 string to UTF-8. Returns a new string that must be freed
+// or NULL if no conversion was needed.
+char *ezxml_str2utf8(char **s, size_t *len)
+{
+    char *u;
+    size_t l = 0, sl, max = *len;
+    long c, d;
+    int b, be = (**s == '\xFE') ? 1 : (**s == '\xFF') ? 0 : -1;
+
+    if (be == -1) return NULL; // not UTF-16
+
+    u = malloc(max);
+    for (sl = 2; sl < *len - 1; sl += 2) {
+        c = (be) ? (((*s)[sl] & 0xFF) << 8) | ((*s)[sl + 1] & 0xFF)  //UTF-16BE
+                 : (((*s)[sl + 1] & 0xFF) << 8) | ((*s)[sl] & 0xFF); //UTF-16LE
+        if (c >= 0xD800 && c <= 0xDFFF && (sl += 2) < *len - 1) { // high-half
+            d = (be) ? (((*s)[sl] & 0xFF) << 8) | ((*s)[sl + 1] & 0xFF)
+                     : (((*s)[sl + 1] & 0xFF) << 8) | ((*s)[sl] & 0xFF);
+            c = (((c & 0x3FF) << 10) | (d & 0x3FF)) + 0x10000;
+        }
+
+        while (l + 6 > max) u = realloc(u, max += EZXML_BUFSIZE);
+        if (c < 0x80) u[l++] = c; // US-ASCII subset
+        else { // multi-byte UTF-8 sequence
+            for (b = 0, d = c; d; d /= 2) b++; // bits in c
+            b = (b - 2) / 5; // bytes in payload
+            u[l++] = (0xFF << (7 - b)) | (c >> (6 * b)); // head
+            while (b) u[l++] = 0x80 | ((c >> (6 * --b)) & 0x3F); // payload
+        }
+    }
+    return *s = realloc(u, *len = l);
+}
+
+// frees a tag attribute list
+void ezxml_free_attr(char **attr) {
+    int i = 0;
+    char *m;
+    
+    if (! attr || attr == EZXML_NIL) return; // nothing to free
+    while (attr[i]) i += 2; // find end of attribute list
+    m = attr[i + 1]; // list of which names and values are malloced
+    for (i = 0; m[i]; i++) {
+        if (m[i] & EZXML_NAMEM) free(attr[i * 2]);
+        if (m[i] & EZXML_TXTM) free(attr[(i * 2) + 1]);
+    }
+    free(m);
+    free(attr);
 }
 
 // parse the given xml string and return an ezxml structure
 ezxml_t ezxml_parse_str(char *s, size_t len)
 {
-    ezxml_root_t root = (ezxml_root_t)malloc(sizeof(struct ezxml_root));
-    char *d, **attr, q, e;
-    static char *noattr[] = { NULL };
-    int l;
-
-    if (! root) return NULL;
+    ezxml_root_t root = (ezxml_root_t)ezxml_new(NULL);
+    char q, e, *d, **attr, **a = NULL; // initialize a to avoid compile warning
+    int l, i, j;
+
+    root->m = s;
+    if (! len) return ezxml_err(root, s, "root tag missing");
+    root->u = ezxml_str2utf8(&s, &len); // convert utf-16 to utf-8
+    root->e = (root->s = s) + len; // record start and end of work area
     
-    // initialize root tag
-    memset((void *)root, '\0', sizeof (struct ezxml_root));
-    root->xml.attr = noattr;
-    root->cur = (ezxml_t)root;
-    root->m = (void *)s;
-    root->err = root->xml.txt = "";
-
-    if (! len) return ezxml_seterr(root, "root tag missing");
-    e = s[len - 1];
-    s[len - 1] = '\0';
+    e = s[len - 1]; // save end char
+    s[len - 1] = '\0'; // turn end char into null terminator
 
     while (*s && *s != '<') s++; // find first tag
-    if (! *s) return ezxml_seterr(root, "root tag missing");
+    if (! *s) return ezxml_err(root, s, "root tag missing");
 
-    for (;;) {
-        attr = noattr;
+    for (; ; ) {
+        attr = (char **)EZXML_NIL;
         d = ++s;
         
         if (isalpha(*s) || *s == '_' || *s == ':') { // new tag
-            if (! root->cur) return ezxml_seterr(root, "unmatched closing tag");
+            if (! root->cur)
+                return ezxml_err(root, d, "markup outside of root element");
 
             s += strcspn(s, EZXML_WS "/>");
-            if (isspace(*s)) *(s++) = '\0';
+            while (isspace(*s)) *(s++) = '\0'; // null terminate tag name
   
-            l = 0;
-            while (*s && *s != '/' && *s != '>') { // new tag attribute
-                while (isspace(*s)) s++;
- 
-                attr = (char **)((! l) ? malloc(3 * sizeof (char *)) :
-                       realloc((void *)attr, (l + 3) * sizeof (char *)));
-                attr[l] = s;
+            if (*s && *s != '/' && *s != '>') // find tag in default attr list
+                for (i = 0; (a = root->attr[i]) && strcmp(a[0], d); i++);
+
+            for (l = 0; *s && *s != '/' && *s != '>'; l += 2) { // new attrib
+                attr = (l) ? realloc(attr, (l + 4) * sizeof(char *))
+                           : malloc(4 * sizeof(char *)); // allocate space
+                attr[l + 3] = (l) ? realloc(attr[l + 1], (l / 2) + 2)
+                                  : malloc(2); // mem for list of maloced vals
+                strcpy(attr[l + 3] + (l / 2), " "); // value is not malloced
+                attr[l + 2] = NULL; // null terminate list
+                attr[l + 1] = ""; // temporary attribute value
+                attr[l] = s; // set attribute name
 
                 s += strcspn(s, EZXML_WS "=/>");
                 if (*s == '=' || isspace(*s)) { 
-                    *(s++) = '\0';
+                    *(s++) = '\0'; // null terminate tag attribute name
                     q = *(s += strspn(s, EZXML_WS "="));
                     if (q == '"' || q == '\'') { // attribute value
                         attr[l + 1] = ++s;
                         while (*s && *s != q) s++;
-                        if (*s) *(s++) = '\0';
+                        if (*s) *(s++) = '\0'; // null terminate attribute val
                         else {
-                            free(attr);
-                            return ezxml_seterr(root, (q == '"') ? 
-                                                "missing \"" : "missing '");
+                            ezxml_free_attr(attr);
+                            return ezxml_err(root, d, "missing %c", q);
                         }
-                        ezxml_decode(attr[l + 1], 1);
+
+                        for (j = 1; a && a[j] && strcmp(a[j], attr[l]); j +=3);
+                        attr[l + 1] = ezxml_decode(attr[l + 1], root->ent, (a
+                                                   && a[j]) ? *a[j + 2] : ' ');
+                        if (attr[l + 1] < d || attr[l + 1] > s)
+                            attr[l + 3][l / 2] = EZXML_TXTM; // value malloced
                     }
-                    else attr[l + 1] = "";
                 }
-                else attr[l + 1] = "";
-
-                attr[(l += 2)] = NULL;
+                while (isspace(*s)) s++;
             }
 
             if (*s == '/') { // self closing tag
                 *(s++) = '\0';
                 if ((*s && *s != '>') || (! *s && e != '>')) {
-                    if (l) free(attr);
-                    return ezxml_seterr(root, "missing >");
+                    if (l) ezxml_free_attr(attr);
+                    return ezxml_err(root, d, "missing >");
                 }
                 ezxml_open_tag(root, d, attr);
-                ezxml_close_tag(root);
+                ezxml_close_tag(root, d, s);
             }
-            else if (*s == '>' || (! *s && e == '>')) { // open tag
-                q = *s;
-                *s = '\0';
+            else if ((q = *s) == '>' || (! *s && e == '>')) { // open tag
+                *s = '\0'; // temporarily null terminate tag name
                 ezxml_open_tag(root, d, attr);
                 *s = q;
             }
             else {
-                if (l) free(attr);
-                return ezxml_seterr(root, "missing >"); 
+                if (l) ezxml_free_attr(attr);
+                return ezxml_err(root, d, "missing >"); 
             }
         }
         else if (*s == '/') { // close tag
-            if (! root->cur) return ezxml_seterr(root, "unmatched closing tag");
-            ezxml_close_tag(root);
-            while (*s && *s != '>') s++;
-            if (! *s && e != '>') return ezxml_seterr(root, "missing >");
+            s += strcspn(d = s + 1, EZXML_WS ">") + 1;
+            if (! (q = *s) && e != '>') return ezxml_err(root, d, "missing >");
+            *s = '\0'; // temporarily null terminate tag name
+            if (ezxml_close_tag(root, d, s)) return &root->xml;
+            if (isspace(*s = q)) s += strspn(s, EZXML_WS);
         }
         else if (! strncmp(s, "!--", 3)) { // comment
-            do { s = strstr(s, "--"); } while (s && *(s += 2) && *s != '>');
-            if (! s || (! *s && e != '>'))
-                return ezxml_seterr(root, "unclosed <!--");
+            if (! (s = strstr(s + 3, "--")) || (*(s += 2) != '>' && *s) ||
+                (! *s && e != '>')) return ezxml_err(root, d, "unclosed <!--");
         }
         else if (! strncmp(s, "![CDATA[", 8)) { // cdata
             if ((s = strstr(s, "]]>")))
-                ezxml_char_content(root, d + 8, (s += 2) - d - 10, 0);
-            else return ezxml_seterr(root, "unclosed <![CDATA[");
+                ezxml_char_content(root, d + 8, (s += 2) - d - 10, 'c');
+            else return ezxml_err(root, d, "unclosed <![CDATA[");
         }
-        else if (! strncmp(s, "!DOCTYPE", 8)) { // skip <!DOCTYPE declarations
+        else if (! strncmp(s, "!DOCTYPE", 8)) { // dtd
             for (l = 0; *s && ((! l && *s != '>') || (l && (*s != ']' || 
-                 s[strspn(s + 1, EZXML_WS) + 1] != '>')));
+                 *(s + strspn(s + 1, EZXML_WS) + 1) != '>')));
                  l = (*s == '[') ? 1 : l) s += strcspn(s + 1, "[]>") + 1;
-            if (! *s && e != '>') 
-                return ezxml_seterr(root, "unclosed <!DOCTYPE");
+            if (! *s && e != '>')
+                return ezxml_err(root, d, "unclosed <!DOCTYPE");
+            d = (l) ? strchr(d, '[') + 1 : d;
+            if (l && ! ezxml_internal_dtd(root, d, s++ - d)) return &root->xml;
         }
         else if (*s == '?') { // <?...?> processing instructions
             do { s = strchr(s, '?'); } while (s && *(++s) && *s != '>');
             if (! s || (! *s && e != '>')) 
-                return ezxml_seterr(root, "unclosed <?");
+                return ezxml_err(root, d, "unclosed <?");
             else ezxml_proc_inst(root, d + 1, s - d - 2);
         }
-        else return ezxml_seterr(root, "syntax error");
+        else return ezxml_err(root, d, "unexpected <");
         
         if (! s || ! *s) break;
         *s = '\0';
         d = ++s;
         if (*s && *s != '<') { // tag character content
             while (*s && *s != '<') s++;
-            if (*s) ezxml_char_content(root, d, s - d, 1);
+            if (*s) ezxml_char_content(root, d, s - d, '&');
             else break;
         }
         else if (! *s) break;
     }
-    
-    return (root->cur) ? ezxml_seterr(root, (root->cur->name) ? "unclosed tag" :
-                                      "root tag missing") : (ezxml_t)root;
+
+    if (! root->cur) return &root->xml;
+    else if (! root->cur->name) return ezxml_err(root, d, "root tag missing");
+    else return ezxml_err(root, d, "unclosed tag <%s>", root->cur->name);
 }
 
 #ifdef CYGPKG_IO_FILEIO
@@ -452,22 +663,20 @@
 // or ezxml_parse_fd()
 ezxml_t ezxml_parse_fp(FILE *fp)
 {
-    ezxml_root_t ret;
-    size_t l, len = 0, ps = (size_t)sysconf(_SC_PAGESIZE);
-    void *s;
+    ezxml_root_t root;
+    size_t l, len = 0;
+    char *s;
 
-    if (! (s = (char *)malloc(ps))) return NULL;
+    if (! (s = malloc(EZXML_BUFSIZE))) return NULL;
     do {
-        len += (l = fread((void *)((int)s + len), 1, ps, fp));
-        if (l == ps) s = (char *)realloc((void *)s, len + ps);
-    } while (s && l == ps);
+        len += (l = fread((s + len), 1, EZXML_BUFSIZE, fp));
+        if (l == EZXML_BUFSIZE) s = realloc(s, len + EZXML_BUFSIZE);
+    } while (s && l == EZXML_BUFSIZE);
 
     if (! s) return NULL;
-
-    ret = (ezxml_root_t)ezxml_parse_str(s, len);
-    ret->len = -1; // so we know to free s in ezxml_free()
-
-    return (ezxml_t)ret;
+    root = (ezxml_root_t)ezxml_parse_str(s, len);
+    root->len = -1; // so we know to free s in ezxml_free()
+    return &root->xml;
 }
 
 // A wrapper for ezxml_parse_str() that accepts a file descriptor. First
@@ -475,50 +684,48 @@
 // Returns NULL on failure.
 ezxml_t ezxml_parse_fd(int fd)
 {
-    ezxml_root_t ret;
-    struct stat stat;
-    size_t len, ps = (size_t)sysconf(_SC_PAGESIZE);
+    ezxml_root_t root;
+    struct stat st;
+    size_t l;
     void *m;
 
     if (fd < 0) return NULL;
+    fstat(fd, &st);
 
-    fstat(fd, &stat);
-    len = (stat.st_size + ps - 1) & ~(ps - 1); // round up to next page boundry
-#ifdef __ECOS__
-	if (0) {
-#else 
-    if ((m = mmap(NULL, len, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0)) !=
+#ifndef EZXML_NOMMAP
+    l = (st.st_size + sysconf(_SC_PAGESIZE) - 1) & ~(sysconf(_SC_PAGESIZE) -1);
+    if ((m = mmap(NULL, l, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0)) !=
         MAP_FAILED) {
-        madvise(m, len, MADV_SEQUENTIAL); // optimize for sequential access
-        ret = (ezxml_root_t)ezxml_parse_str((char *)m, stat.st_size);
-        madvise(m, len, MADV_NORMAL);     // put it back to normal
-        ret->len = len;
-#endif
+        madvise(m, l, MADV_SEQUENTIAL); // optimize for sequential access
+        root = (ezxml_root_t)ezxml_parse_str(m, st.st_size);
+        madvise(m, root->len = l, MADV_NORMAL); // put it back to normal
     }
     else { // mmap failed, read file into memory
-        if (! (m = malloc(stat.st_size))) return NULL;
-        len = read(fd, m, stat.st_size);
-        ret = (ezxml_root_t)ezxml_parse_str((char *)m, len);
-        ret->len = -1;
+#endif // EZXML_NOMMAP
+        l = read(fd, m = malloc(st.st_size), st.st_size);
+        root = (ezxml_root_t)ezxml_parse_str(m, l);
+        root->len = -1; // so we know to free s in ezxml_free()
+#ifndef EZXML_NOMMAP
     }
-
-    return (ezxml_t)ret;
+#endif // EZXML_NOMMAP
+    return &root->xml;
 }
 
 // a wrapper for ezxml_parse_fd that accepts a file name
 ezxml_t ezxml_parse_file(const char *file)
 {
     int fd = open(file, O_RDONLY, 0);
-    ezxml_t ret = ezxml_parse_fd(fd);
+    ezxml_t xml = ezxml_parse_fd(fd);
     
     if (fd >= 0) close(fd);
-    return ret;
+    return xml;
 }
-#endif
-// Encodes ampersand sequences appending the results to dst, reallocating dst
-// if it's length excedes max. Returns *dst.
+#endif // CYGPKG_IO_FILEIO
+
+// Encodes ampersand sequences appending the results to *dst, reallocating *dst
+// if length excedes max. a is non-zero for attribute encoding. Returns *dst
 char *ezxml_ampencode(const char *s, size_t len, char **dst, size_t *dlen,
-                      size_t *max)
+                      size_t *max, short a)
 {
     const char *e;
     
@@ -530,130 +737,327 @@
         case '&': *dlen += sprintf(*dst + *dlen, "&amp;"); break;
         case '<': *dlen += sprintf(*dst + *dlen, "&lt;"); break;
         case '>': *dlen += sprintf(*dst + *dlen, "&gt;"); break;
-        case '"': *dlen += sprintf(*dst + *dlen, "&quot;"); break;
-        default:
-            if (*s >= ' ' || *s == '\n' || *s == '\t') (*dst)[(*dlen)++] = *s;
-            else *dlen += sprintf(*dst + *dlen, "&#%02d;", *s);
+        case '"': *dlen += sprintf(*dst + *dlen, (a) ? "&quot;" : "\""); break;
+        case '\n': *dlen += sprintf(*dst + *dlen, (a) ? "&#xA;" : "\n"); break;
+        case '\t': *dlen += sprintf(*dst + *dlen, (a) ? "&#x9;" : "\t"); break;
+        case '\r': *dlen += sprintf(*dst + *dlen, "&#xD;"); break;
+        default: (*dst)[(*dlen)++] = *s;
         }
     }
-    
     return *dst;
 }
 
-// Recursively converts each tag to xml appending it to s. Reallocates s if it's
-// length excedes max. start is the location of the previous tag in the parent
-// tag's character content. Returns *s.
+// Recursively converts each tag to xml appending it to *s. Reallocates *s if
+// its length excedes max. start is the location of the previous tag in the
+// parent tag's character content. Returns *s.
 char *ezxml_toxml_r(ezxml_t xml, char **s, size_t *len, size_t *max,
-                    size_t start)
+                    size_t start, char ***attr)
 {
-    int i;
+    int i, j;
     char *txt = (xml->parent) ? xml->parent->txt : "";
+    size_t off = 0;
 
     // parent character content up to this tag
-    ezxml_ampencode(txt + start, xml->off - start, s, len, max);
+    *s = ezxml_ampencode(txt + start, xml->off - start, s, len, max, 0);
 
     while (*len + strlen(xml->name) + 4 > *max) // reallocate s
         *s = realloc(*s, *max += EZXML_BUFSIZE);
 
     *len += sprintf(*s + *len, "<%s", xml->name); // open tag
     for (i = 0; xml->attr[i]; i += 2) { // tag attributes
+        if (ezxml_attr(xml, xml->attr[i]) != xml->attr[i + 1]) continue;
         while (*len + strlen(xml->attr[i]) + 7 > *max) // reallocate s
             *s = realloc(*s, *max += EZXML_BUFSIZE);
 
         *len += sprintf(*s + *len, " %s=\"", xml->attr[i]);
-        ezxml_ampencode(xml->attr[i + 1], -1, s, len, max);
+        ezxml_ampencode(xml->attr[i + 1], -1, s, len, max, 1);
         *len += sprintf(*s + *len, "\"");
     }
 
-    if (xml->child || *(xml->txt)) { // tag content
-        *len += sprintf(*s + *len, ">");
-        if (xml->child) ezxml_toxml_r(xml->child, s, len, max, 0);
-        else ezxml_ampencode(xml->txt, -1, s, len, max); // char content
-
-        while (*len + strlen(xml->name) + 4 > *max) // reallocate s
+    for (i = 0; attr[i] && strcmp(attr[i][0], xml->name); i++);
+    for (j = 1; attr[i] && attr[i][j]; j += 3) { // default attributes
+        if (! attr[i][j + 1] || ezxml_attr(xml, attr[i][j]) != attr[i][j + 1])
+            continue; // skip duplicates and non-values
+        while (*len + strlen(attr[i][j]) + 7 > *max) // reallocate s
             *s = realloc(*s, *max += EZXML_BUFSIZE);
 
-        *len += sprintf(*s + *len, "</%s>", xml->name); // close tag
+        *len += sprintf(*s + *len, " %s=\"", attr[i][j]);
+        ezxml_ampencode(attr[i][j + 1], -1, s, len, max, 1);
+        *len += sprintf(*s + *len, "\"");
     }
-    else *len += sprintf(*s + *len, "/>"); // self closing tag
+    *len += sprintf(*s + *len, ">");
+
+    *s = (xml->child) ? ezxml_toxml_r(xml->child, s, len, max, 0, attr) //child
+                      : ezxml_ampencode(xml->txt, -1, s, len, max, 0);  //data
+    
+    while (*len + strlen(xml->name) + 4 > *max) // reallocate s
+        *s = realloc(*s, *max += EZXML_BUFSIZE);
 
-    if (xml->ordered) return ezxml_toxml_r(xml->ordered, s, len, max, xml->off);
-    return ezxml_ampencode(txt + xml->off, -1, s, len, max);
+    *len += sprintf(*s + *len, "</%s>", xml->name); // close tag
+
+    while (txt[off] && off < xml->off) off++; // make sure off is within bounds
+    return (xml->ordered) ? ezxml_toxml_r(xml->ordered, s, len, max, off, attr)
+                          : ezxml_ampencode(txt + off, -1, s, len, max, 0);
 }
 
 // converts an ezxml structure back to xml, returning it as a string that must
 // be freed
 char *ezxml_toxml(ezxml_t xml)
 {
-    ezxml_t p = xml->parent;
+    ezxml_t p = (xml) ? xml->parent : NULL, o = (xml) ? xml->ordered : NULL;
+    ezxml_root_t root = (ezxml_root_t)xml;
     size_t len = 0, max = EZXML_BUFSIZE;
-    char *s = strcpy((char *)malloc(max), "");
+    char *s = strcpy(malloc(max), ""), *t, *n;
+    int i, j, k;
 
     if (! xml || ! xml->name) return realloc(s, len + 1);
+    while (root->xml.parent) root = (ezxml_root_t)root->xml.parent; // root tag
 
-    xml->parent = NULL;
-    ezxml_toxml_r(xml, &s, &len, &max, 0);
+    for (i = 0; ! p && root->pi[i]; i++) { // pre-root processing instructions
+        for (k = 2; root->pi[i][k - 1]; k++);
+        for (j = 1; (n = root->pi[i][j]); j++) {
+            if (root->pi[i][k][j - 1] == '>') continue; // not pre-root
+            while (len + strlen(t = root->pi[i][0]) + strlen(n) + 7 > max)
+                s = realloc(s, max += EZXML_BUFSIZE);
+            len += sprintf(s + len, "<?%s%s%s?>\n", t, *n ? " " : "", n);
+        }
+    }
+
+    xml->parent = xml->ordered = NULL;
+    s = ezxml_toxml_r(xml, &s, &len, &max, 0, root->attr);
     xml->parent = p;
+    xml->ordered = o;
 
+    for (i = 0; ! p && root->pi[i]; i++) { // post-root processing instructions
+        for (k = 2; root->pi[i][k - 1]; k++);
+        for (j = 1; (n = root->pi[i][j]); j++) {
+            if (root->pi[i][k][j - 1] == '<') continue; // not post-root
+            while (len + strlen(t = root->pi[i][0]) + strlen(n) + 7 > max)
+                s = realloc(s, max += EZXML_BUFSIZE);
+            len += sprintf(s + len, "\n<?%s%s%s?>", t, *n ? " " : "", n);
+        }
+    }
     return realloc(s, len + 1);
 }
 
 // free the memory allocated for the ezxml structure
 void ezxml_free(ezxml_t xml)
 {
-    void *m = NULL; // assigned to null to avoid compiler warning
-    size_t len = 0;
-    int i;
+    ezxml_root_t root = (ezxml_root_t)xml;
+    int i, j;
+    char **a, *s;
 
     if (! xml) return;
+    ezxml_free(xml->child);
+    ezxml_free(xml->ordered);
 
-    if (! xml->parent) {
-        m = ((ezxml_root_t)xml)->m;
-        len = ((ezxml_root_t)xml)->len;
-
-        if (((ezxml_root_t)xml)->pi) {
-            for (i = 0; ((ezxml_root_t)xml)->pi[i]; i++)
-                free(((ezxml_root_t)xml)->pi[i]);
-            free(((ezxml_root_t)xml)->pi);
+    if (! xml->parent) { // free root tag allocations
+        for (i = 10; root->ent[i]; i += 2) // 0 - 9 are default entites (<>&"')
+            if ((s = root->ent[i + 1]) < root->s || s > root->e) free(s);
+        free(root->ent); // free list of general entities
+
+        for (i = 0; (a = root->attr[i]); i++) {
+            for (j = 1; a[j++]; j += 2) // free malloced attribute values
+                if (a[j] && (a[j] < root->s || a[j] > root->e)) free(a[j]);
+            free(a);
         }
-    }
+        if (root->attr[0]) free(root->attr); // free default attribute list
 
-    ezxml_free(xml->child);
-    ezxml_free(xml->ordered);
+        for (i = 0; root->pi[i]; i++) {
+            for (j = 1; root->pi[i][j]; j++);
+            free(root->pi[i][j + 1]);
+            free(root->pi[i]);
+        }            
+        if (root->pi[0]) free(root->pi); // free processing instructions
+
+        if (root->len == -1) free(root->m); // malloced xml data
+#ifndef EZXML_NOMMAP
+        else if (root->len) munmap(root->m, root->len); // mem mapped xml data
+#endif // EZXML_NOMMAP
+        if (root->u) free(root->u); // utf8 conversion
+    }
 
-    if (xml->attr[0]) free((void *)xml->attr);
-    if ((xml->flags & EZXML_TXTM)) free((void *)xml->txt);
-    free((void *)xml);
-
-    if (len == -1) free(m);
-#ifndef __ECOS__
-    else if (len) munmap(m, len);
-#endif
+    ezxml_free_attr(xml->attr); // tag attributes
+    if ((xml->flags & EZXML_TXTM)) free(xml->txt); // character content
+    if ((xml->flags & EZXML_NAMEM)) free(xml->name); // tag name
+    free(xml);
 }
 
 // return parser error message or empty string if none
 const char *ezxml_error(ezxml_t xml)
 {
-    while (xml->parent) xml = xml->parent;
-    return ((ezxml_root_t)xml)->err;
+    while (xml && xml->parent) xml = xml->parent; // find root tag
+    return (xml) ? ((ezxml_root_t)xml)->err : "";
 }
 
-#ifdef EZXMLTEST // test harness
+// returns a new empty ezxml structure with the given root tag name
+ezxml_t ezxml_new(const char *name)
+{
+    static char *ent[] = { "lt;", "&#60;", "gt;", "&#62;", "quot;", "&#34;",
+                           "apos;", "&#39;", "amp;", "&#38;", NULL };
+    ezxml_root_t root = (ezxml_root_t)memset(malloc(sizeof(struct ezxml_root)), 
+                                             '\0', sizeof(struct ezxml_root));
+    root->xml.name = (char *)name;
+    root->cur = &root->xml;
+    strcpy(root->err, root->xml.txt = "");
+    root->ent = memcpy(malloc(sizeof(ent)), ent, sizeof(ent));
+    root->attr = root->pi = (char ***)(root->xml.attr = EZXML_NIL);
+    return &root->xml;
+}
+
+// Adds a child tag. off is the offset of the child tag relative to the start
+// of the parent tag's character content. returns the child tag
+ezxml_t ezxml_add_child(ezxml_t xml, const char *name, size_t off)
+{
+    ezxml_t cur, head, child;
 
+    if (! xml) return NULL;
+    child = (ezxml_t)memset(malloc(sizeof(struct ezxml)), '\0',
+                            sizeof(struct ezxml));
+    child->name = (char *)name;
+    child->attr = EZXML_NIL;
+    child->off = off;
+    child->parent = xml;
+    child->txt = "";
+
+    if ((head = xml->child)) { // already have sub tags
+        if (head->off <= off) { // not first subtag
+            for (cur = head; cur->ordered && cur->ordered->off <= off;
+                 cur = cur->ordered);
+            child->ordered = cur->ordered;
+            cur->ordered = child;
+        }
+        else { // first subtag
+            child->ordered = head;
+            xml->child = child;
+        }
+
+        for (cur = head; cur->sibling && strcmp(cur->name, name);
+             cur = cur->sibling); // find tag type
+        if (! strcmp(cur->name, name) && cur->off <= off) { //not first of type
+            while (cur->next && cur->next->off <= off) cur = cur->next;
+            child->next = cur->next;
+            cur->next = child;
+        }
+        else { // first tag of this type
+            if (cur->off > off) child->next = cur; // not only tag of this type
+            for (cur = head; cur->sibling && cur->sibling->off <= off;
+                 cur = cur->sibling);
+            child->sibling = cur->sibling;
+            cur->sibling = child;
+        }
+    }
+    else xml->child = child; // only sub tag
+    
+    return child;
+}
+
+// sets the character content for the given tag and returns the tag
+ezxml_t ezxml_set_txt(ezxml_t xml, const char *txt)
+{
+    if (! xml) return NULL;
+    if (xml->flags & EZXML_TXTM) free(xml->txt); // existing txt was malloced
+    xml->flags &= ~EZXML_TXTM;
+    xml->txt = (char *)txt;
+    return xml;
+}
+
+// Sets the given tag attribute or adds a new attribute if not found. A value
+// of NULL will remove the specified attribute.
+void ezxml_set_attr(ezxml_t xml, const char *name, const char *value)
+{
+    int l = 0, c;
+
+    if (! xml) return;
+    while (xml->attr[l] && strcmp(xml->attr[l], name)) l += 2;
+    if (! xml->attr[l]) { // not found, add as new attribute
+        if (! value) return; // nothing to do
+        if (xml->attr == EZXML_NIL) { // first attribute
+            xml->attr = malloc(4 * sizeof(char *));
+            xml->attr[1] = strdup(""); // empty list of malloced names/vals
+        }
+        else xml->attr = realloc(xml->attr, (l + 4) * sizeof(char *));
+
+        xml->attr[l] = (char *)name; // set attribute name
+        xml->attr[l + 2] = NULL; // null terminate attribute list
+        xml->attr[l + 3] = realloc(xml->attr[l + 1],
+                                   (c = strlen(xml->attr[l + 1])) + 2);
+        strcpy(xml->attr[l + 3] + c, " "); // set name/value as not malloced
+        if (xml->flags & EZXML_DUP) xml->attr[l + 3][c] = EZXML_NAMEM;
+    }
+    else if (xml->flags & EZXML_DUP) free((char *)name); // name was strduped
+
+    for (c = l; xml->attr[c]; c += 2); // find end of attribute list
+    if (xml->attr[c + 1][l / 2] & EZXML_TXTM) free(xml->attr[l + 1]); //old val
+    if (xml->flags & EZXML_DUP) xml->attr[c + 1][l / 2] |= EZXML_TXTM;
+    else xml->attr[c + 1][l / 2] &= ~EZXML_TXTM;
+
+    if (value) xml->attr[l + 1] = (char *)value; // set attribute value
+    else { // remove attribute
+        if (xml->attr[c + 1][l / 2] & EZXML_NAMEM) free(xml->attr[l]);
+        memmove(xml->attr + l, xml->attr + l + 2, (c - l + 2) * sizeof(char*));
+        xml->attr = realloc(xml->attr, (c + 2) * sizeof(char *));
+        memmove(xml->attr[c + 1] + (l / 2), xml->attr[c + 1] + (l / 2) + 1,
+                (c / 2) - (l / 2)); // fix list of which name/vals are malloced
+    }
+    xml->flags &= ~EZXML_DUP; // clear strdup() flag
+}
+
+// sets a flag for the given tag and returns the tag
+ezxml_t ezxml_set_flag(ezxml_t xml, short flag)
+{
+    if (xml) xml->flags |= flag;
+    return xml;
+}
+
+// removes a tag along with all its subtags
+void ezxml_remove(ezxml_t xml)
+{
+    ezxml_t cur;
+
+    if (! xml) return; // nothing to do
+    if (xml->next) xml->next->sibling = xml->sibling; // patch sibling list
+
+    if (xml->parent) { // not root tag
+        cur = xml->parent->child; // find head of subtag list
+        if (cur == xml) xml->parent->child = xml->ordered; // first subtag
+        else { // not first subtag
+            while (cur->ordered != xml) cur = cur->ordered;
+            cur->ordered = cur->ordered->ordered; // patch ordered list
+
+            cur = xml->parent->child; // go back to head of subtag list
+            if (strcmp(cur->name, xml->name)) { // not in first sibling list
+                while (strcmp(cur->sibling->name, xml->name))
+                    cur = cur->sibling;
+                if (cur->sibling == xml) { // first of a sibling list
+                    cur->sibling = (xml->next) ? xml->next
+                                               : cur->sibling->sibling;
+                }
+                else cur = cur->sibling; // not first of a sibling list
+            }
+
+            while (cur->next && cur->next != xml) cur = cur->next;
+            if (cur->next) cur->next = cur->next->next; // patch next list
+        }        
+    }
+    xml->ordered = NULL; // prevent ezxml_free() from clobbering ordered list
+    ezxml_free(xml);
+}
+
+#ifdef EZXML_TEST // test harness
 int main(int argc, char **argv)
 {
     ezxml_t xml;
     char *s;
+    int i;
 
     if (argc != 2) return fprintf(stderr, "usage: %s xmlfile\n", argv[0]);
 
     xml = ezxml_parse_file(argv[1]);
-    printf("%s", (s = ezxml_toxml(xml)));
-    fprintf(stderr, "%s", ezxml_error(xml));
+    printf("%s\n", (s = ezxml_toxml(xml)));
     free(s);
+    i = fprintf(stderr, "%s", ezxml_error(xml));
     ezxml_free(xml);
-
-    return 0;
+    return (i) ? 1 : 0;
 }
-
-#endif // EZXMLTEST
+#endif // EZXML_TEST
-- 
Bye, Peter Korsgaard

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]