miniXML parsing C API

Question

I'm trying parse the follow XML file:

<root>Root
    <pai>Pai_1
        <filho>Pai1,Filho1</filho>
        <filho>Pai1,Filho2</filho>
    </pai>
    <pai>Pai_2
        <filho>Pai2,Filho1</filho>
        <filho>Pai2,Filho2</filho>
    </pai>
</root>

I'm using the follow C code:

//... open file
xml_tree = mxmlLoadFile(NULL, fp, MXML_TEXT_CALLBACK);

node = xml_tree;
printf("%s\n", mxmlGetText(node, NULL));
// here the return is: Root
// I expected: Root, OK

node = xml_tree->child;
printf("%s\n", mxmlGetText(node, NULL));
// here the return is: Root
// I expected: Pai_1, not OK

node = mxmlGetFirstChild(xml_tree);
printf("%s\n", mxmlGetText(node, NULL));
// here the return is: Root
// I expected: Pai_1, not OK

node = mxmlFindElement(xml_tree, xml_tree, "pai", NULL, NULL, MXML_DESCEND);
printf("%s\n", mxmlGetText(node, NULL));
// here the return is: Pai_1
// I expected: Pai_1, OK

node = mxmlGetNextSibling(node);
printf("%s\n", mxmlGetText(node, NULL));
// here the return is: (NULL)
// I expected: Pai_2, not OK

How can I access the child of the root? Where my concept of access is wrong?

Thank you.

EDIT after @RutgersMike response

I expand your while loop to try understand the concept of minixml:

root = mxmlLoadFile(NULL,fp,MXML_TEXT_CALLBACK);
node = root;

printf("------- Root\n");
fprintf(stdout,"Element = %s\n",mxmlGetElement(node));
fprintf(stdout,"  Value = %s\n",mxmlGetText(node,0));
printf("\n");

printf("------- First child of Root\n");
node = mxmlGetFirstChild(node);
fprintf(stdout,"Element = %s\n",mxmlGetElement(node));
fprintf(stdout,"  Value = %s\n",mxmlGetText(node,0));
printf("\n");

printf("------- Sibling 1 of First child of Root\n");
node = mxmlGetNextSibling(node);
fprintf(stdout,"Element = %s\n",mxmlGetElement(node));
fprintf(stdout,"  Value = %s\n",mxmlGetText(node,0));
printf("\n");

printf("------- Sibling 2 of First child of Root\n");
node = mxmlGetNextSibling(node);
fprintf(stdout,"Element = %s\n",mxmlGetElement(node));
fprintf(stdout,"  Value = %s\n",mxmlGetText(node,0));
printf("\n");

printf("------- Sibling 3 of First child of Root\n");
node = mxmlGetNextSibling(node);
fprintf(stdout,"Element = %s\n",mxmlGetElement(node));
fprintf(stdout,"  Value = %s\n",mxmlGetText(node,0));
printf("\n");

printf("------- Sibling 4 of First child of Root\n");
node = mxmlGetNextSibling(node);
fprintf(stdout,"Element = %s\n",mxmlGetElement(node));
fprintf(stdout,"  Value = %s\n",mxmlGetText(node,0));
printf("\n");

The result this was:

------- Root
Element = root
  Value = Root

------- First child of Root
Element = (null)
  Value = Root

------- Sibling 1 of First child of Root
Element = (null)
  Value = 

------- Sibling 2 of First child of Root
Element = pai
  Value = Pai_1

------- Sibling 3 of First child of Root
Element = (null)
  Value = 

------- Sibling 4 of First child of Root
Element = pai
  Value = Pai_2

I think this concept of navigation between child and parent a little strange. Why there are (null) values between sibling?

I am considering go back to ezxml.

Thank you

score 7 · Answer 1 · edited Jul 15 '12 at 09:14

Just starting playing with min-xml too and though I've been extremely frustrated by lack of good examples, I've borrowed and enhanced to give a decent but not perfect example of reading an XML file and seeing all the parts. It shows tag name, attributes, and text values between tags. Not sure how to identify ending tag. Make sure the standard xml tag is at top of xml file. Include stdio, stdlib, string .h files.

#include "mxml.h"

int main (int argc, char **argv ) {

         FILE *fp  = NULL;

         int k = 0;

         mxml_node_t * tree = NULL;
         mxml_node_t * node  = NULL;

         if (argc < 2){
             perror("Argument Required XML File ");
             exit(1);
         }
         fp = fopen (argv[1], "r");
         if (fp ){
             tree = mxmlLoadFile (NULL , fp , MXML_OPAQUE_CALLBACK);
         }else {
             perror("Could Not Open the File Provided");
             exit(1);
         }
         if (tree){
                 for (node = mxmlFindElement(tree, tree,NULL,NULL, NULL,MXML_DESCEND);
                         node != NULL;
                         node=mxmlWalkNext (node, NULL, MXML_DESCEND)
                         //node = mxmlFindElement(node, tree, NULL,NULL,NULL,MXML_DESCEND)
                 ){
                         if (node->type  == MXML_ELEMENT) {
                             printf("MXML_ELEMENT Node <%s>:%d \n", node->value.element.name, node->value.element.num_attrs);
                             for (k = 0; k < node->value.element.num_attrs; k++){
                                 if (node->value.element.attrs ){
                                     printf ("Attribute Name :: %s \n", node->value.element.attrs[k].name);
                                     printf ("Attribute Value:: %s \n", node->value.element.attrs[k].value);
                                 }
                                 //if (!strncmp(node->value.element.name , "display-name", 12 )){
                                 //    printf(" String %s \n", (char*) node->child->value.text.string);
                                 //}
                             }
                         }
                         else if (node->type == MXML_REAL){
                             printf("MXML_REAL Node is %s \n", node->value.element.name);
                         }
                         else if(node->type == MXML_OPAQUE){
                             printf("MXML_OPAQUE Node is %s \n", node->value.element.name);
                         }
                         else if(node->type == MXML_INTEGER){
                             printf("MXML_INTEGER Node is %s \n", node->value.element.name);
                         }
                         else if(node->type == MXML_TEXT){
                             printf("MXML_TEXT Node is %s \n", node->value.element.name);
                         }
                         else if(node->type == MXML_IGNORE){
                             printf("MXML_IGNORE Node is %s \n", node->value.element.name);
                         }
                         else if(node->type == MXML_CUSTOM){
                             printf("MXML_IGNORE Node is %s \n", node->value.element.name);
                         }
                         else {
                             printf("Type Default Node is %s \n", node->value.element.name);
                         }
                 }
         }
         if (tree){
            mxmlDelete(tree);
         }
         if (fp){
            fclose(fp);
         }
         return 0;
}

Thank you for adding this! I appreciate it immensely! – Lesley Gushurst May 10 '16 at 20:47 — Lesley Gushurst, May 10 '16 at 20:47

score 6 · Accepted Answer · edited Jul 15 '12 at 09:16

It looks like you want to use the iteration functions described here (http://www.minixml.org/mxml.html#3_7) to get the child nodes.

EDIT: I wrote this to iterate down through the first child node and it works fine, used mxmlGetFirstChild and mxmlGetNextSibling:

<!-- language: c -->
mxml_node_t* node = mxmlLoadFile(NULL,f,MXML_TEXT_CALLBACK);
while ( node != NULL )
{
   switch ( mxmlGetType(node) )
   {
   case MXML_ELEMENT:
   {
      fprintf(stdout,"Element = %s\n",mxmlGetElement(node));
   }
   break;
   case MXML_TEXT:
   {
      fprintf(stdout,"  Value = %s\n",mxmlGetText(node,0));
   }
   break;
   default:
   {
   }
   break;
   }
   mxml_node_t* next = mxmlGetFirstChild(node);
   if ( next != NULL )
   {
      node = next;
   }
   else
   {
      next = mxmlGetNextSibling(node);
      if ( next != NULL )
      {
         node = next;
      }
      else
      {
         node = next;
         fprintf(stdout,"Done\n");
      }
   }
}

Produces output:

Element = root
Value = Root
Value = 
Element = pai
Value = Pai_1
Value = 
Element = filho
Value = Pai1,Filho1

I presume you can use one of the getParent functions to iterate back up, or save off the last node before diving down into a child, using a stack of node pointers, if you want to iterate through the whole file. Note that I only handle / print data for two node types - you'll want to experiment to see what the other node types contain, if you need that info as well.

** MORE EDIT after your edit **

I suggested just the other day that someone else try libxml2 (http://xmlsoft.org/examples/index.html#xmlReader) - check that link out. The xmlReader example shows usage. It's incredibly easy to create a reader and iterate through the nodes - when you hit each node just check it's type to be sure it's one you care about (usually ELEMENT, ATTRIBUTE, TEXT, and END_ELEMENT), and then pull out either the name or the value. I like it much better than mxml.

Did it not compile or did it not produce the expected result? What did it produce? — rutgersmike, Mar 21 '12 at 16:51
I do this sequence: `node = xml_tree;` `node = mxmlGetFirstChild(xml_tree);` `node = mxmlWalkNext(node, xml_tree, MXML_DESCEND);` `printf("> %s\n", mxmlGetElement(node));` `printf("> %s\n", mxmlGetText(node, NULL));` The result: (null) — Ricardo Crudo, Mar 21 '12 at 17:09

miniXML parsing C API

2 Answers2