Joke: What's the difference between HTML and XML? HTML requires 100,000 lines of code to satisfy one W3C Standard and XML requires 100,000 W3C Standards to write one line of code.
Sure, that's an exaggeration but there are far more W3C standards for XML than there ever were for HTML.
Today, I'd like to write about some actual C++ source code. This site is for software engineers, after all, even if the focus is on business, so let's start off with something that will make us all feel comfortable. Like, actual C++ source code.
XML started simple. Tags, attribute-value pairs, a little text. Looks like well-organized HTML with funny tag names. Easy.
Then, many people asked: "How do I get XML into my program?" So, we got XML parsers, both DOM and SAX.
And, then: "How do I spot sloppy or invalid XML?" DTDs, XML Schema, well-formedness and validation.
Others chimed in: "How can I convert my XML tags into somebody else's XML tags?" XSL and XSLT.
Oh, my. Now, years later, we have oodles of W3C Standards for XML. XPath, XPointer, XML Query and Web Services and all the rest.
It'd be a lot of work to conform to these standards if we didn't have help in the form of libraries. The non-profit Apache Foundation provides Xerces which implements many, if not all, the XML standards in Java, C++, Perl and COM. Sun Microsystems has their own enterprise-ready XML ecosystem in Java. And Microsoft, as well, in .NET. So, it's all taken care of for us, right?
Wrong.
While we don't have to read the XML standards themselves, we've got to learn the APIs to these XML libraries that implement the standards. These APIs simplify and abstract XML but that doesn't mean that these APIs themselves are simple. A small bit of simple code to do a simple XML operation can be easier than these APIs for libraries that implement all these XML standards. It may be unorthodox to forsake standards but maybe it's worthwhile in some cases to get useful attributes like simplicity and ubiquity.
Let's rewind for a sec. Expat, the original XML library, written by the other Jim Clark (not the Jim Clark) started as a relatively simple XML parser. A dozen or so files for the most common XML operation: to read an XML file and put it in a convenient data structure in memory. It has grown a bit: the expat 1.2 distribution contains 42 files of various sorts. 42 files? That's not bad.
Why grow? Why not shrink? Why not one file, instead?
Forget XML standards. Think the C standard. What does most software want to do with XML? Read it in from a file, process it and write it out. Every ancient C compiler could read files in, process them and write them out.
Scroll down to the code below. If you copy it into a C++ program, it will work. It works on Visual C++ 6.0 and g++ without a makefile or #ifdef's. (At the bottom of this page, you can download the C version that uses malloc() and free() which runs on Visual C++ 6.0, cc, gcc, g++, Turbo C 2.0 for DOS and Pacific C 7.51 for DOS. Heck, anything, really.)
One file, one XML parser, one C struct, two C functions, three example C functions. 500 lines. Runs everywhere. Simple.
The simplest XML reader in the world.
/*
Copyright (c) 2003-2004 Daniel W. Howard
Permission is hereby granted, free of charge, to any
person obtaining a copy of this software and associated
documentation files (the "Software"), to deal in the
Software without restriction, including without
limitation the rights to use, copy, modify, merge,
publish, distribute, sublicense, and/or sell copies of the
Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice
shall be included in all copies or substantial portions of
the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY
KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE
WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR
PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS
OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. IN
NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE
FOR ANY SPECIAL, INDIRECT, INCIDENTAL, OR CONSEQUENTIAL
DAMAGES, OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF
USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT,
NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN
CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*/
#include <stdio.h>
#include <string.h>
/* basicxmlnode: simple xml node memory representation */
struct basicxmlnode
{
char * tag; /* always non-NULL */
char * text; /* body + all whitespace, always non-NULL */
char * * attrs; /* array of strings, NULL marks end */
char * * values; /* array of strings, NULL marks end */
struct basicxmlnode * * children; /* similar */
int * childreni; /* children positions in text */
};
/* deletebasicxmlnode: frees all memory for xml tree */
void deletebasicxmlnode( struct basicxmlnode * node )
{
int ii; if (!node) return;
for (ii=0; node->children[ii]; ++ii)
deletebasicxmlnode(node->children[ii]);
delete node->tag; /*free(node->tag);*/
delete node->text; /*free(node->text);*/
for (ii=0; node->attrs[ii]; ++ii) {
delete [] node->attrs[ii]; /*free(node->attrs[ii]);*/
delete [] node->values[ii]; /*free(node->values[ii]);*/
}
delete [] node->attrs; /*free(node->attrs);*/
delete [] node->values; /*free(node->values);*/
delete [] node->children; /*free(node->children);*/
delete [] node->childreni; /*free(node->childreni);*/
delete node; /*free(node);*/
}
/* readbasicxmlnode: reads simple XML file */
struct basicxmlnode * readbasicxmlnode( FILE * fpi )
{
int bufinc=500, textinc=50, childinc=50; /* tuning */
char * buf=0; int nbuf=0;
char * ibuf=0, * buftmp, * wspos, * ps, * pe, * pe2;
int phase=0, ch, err=0, ii=0, ia, natts, * nodeitmp;
int itex=0, ntex=textinc, ichi=0, nchi=childinc, square=0;
typedef /*struct*/ basicxmlnode nodetype;
nodetype * node, * child, * * nodetmp;
size_t so_c = sizeof(char), so_cp = sizeof(char *);
size_t so_i = sizeof(int);
size_t so_n = sizeof(struct basicxmlnode);
size_t so_np = sizeof(struct basicxmlnode *);
if (!err && !fpi) err=1; /* bad arg */
/* get chars between < and >; strip most whitespace */
while (1) {
int isws=0, issl=0, islt=0, isgt=0, iseq=0, isqu=0;
int isqs=0, isex=0, isda=0, issi=0, isun=0, ispe=0;
int islb=0, isrb=0, isn1=0, isn2=0;
if ((ibuf-buf+5) > nbuf) { /* realloc buf */
nbuf += bufinc;
buftmp = new char[nbuf]; /*(char*)malloc(so_c*nbuf);*/
if (!buftmp) {
err=2; break; /* out of memory */
}
if (buf) {
ii = ibuf-buf;
strncpy(buftmp, buf, ii);
delete [] buf; /*free(buf);*/
}
buf = buftmp;
ibuf = buf + ii;
}
if (phase != 100) ch = fgetc(fpi);
if (err || feof(fpi) || (phase == 100)) break;
/* analyze char */ /* /<>="?!-'12*/
if ((ch == ' ') || ((ch > 8) && (ch < 14))) isws = 1;
if (ch == '/') issl = 1; if (ch == '<') islt = 1;
if (ch == '>') isgt = 1; if (ch == '=') iseq = 1;
if (ch == '"') isqu = 1; if (ch == '?') isqs = 1;
if (ch == '!') isex = 1; if (ch == '-') isda = 1;
if (ch == '\'') issi = 1; if (ch == '_') isun = 1;
if (ch == '[') islb = 1; if (ch == ']') isrb = 1;
if (ch == '.') ispe = 1;
if ((ch >= 'a') && (ch <= 'z')) isn1 = 1;
if ((ch >= 'A') && (ch <= 'Z')) isn1 = 1;
if (isun) isn1 = 1;
if (isn1 || isda || ispe) isn2 = 1;
if ((ch >= '0') && (ch <= '9')) isn2 = 1;
/* err=3, bad xml file (parsing error) */
switch (phase) {
case 0: /* eat whitespace before < */
if (isws) { } /* */
else if (islt) { ++phase; *ibuf=ch; ++ibuf; } /*<*/
else { err=3; } /**/ /*/>="?!-'t1*/
break;
case 1: /* handle the first char after < */
if (issl) { phase=10; *ibuf=ch; ++ibuf; } /*</*/
else if (isqs) { phase=14; --ibuf; } /*<?*/
else if (isex) { phase=15; --ibuf; } /*<!*/
else if (isn1) { ++phase; *ibuf=ch; ++ibuf; } /*<t*/
else { err=3; } /*<*/ /* <>="-'1*/
break;
case 2: /* store tag name and a single space after it */
if (isws) { ++phase; *ibuf = ' '; ++ibuf; } /*<t */
else if (issl) { phase=13; strncpy(ibuf," /",2); ibuf+=2; } /*<t/*/
else if (isgt) { phase=100; strncpy(ibuf," >",2); ibuf+=2; } /*<t>*/
else if (isn2) { *ibuf=ch; ++ibuf; } /*<t*/ /*-1*/
else { err=3; } /*<t*/ /*<="?!'*/
break;
case 3: /* eat whitespace after tag name */
if (isws) { } /*<tag */
else if (issl) { phase=13; *ibuf=ch; ++ibuf; } /*<tag /*/
else if (isgt) { phase=100; *ibuf=ch; ++ibuf; } /*<tag >*/
else if (isn1) { ++phase; *ibuf=ch; ++ibuf; } /*<tag t*/
else { err=3; } /*<tag */ /*<="?!-'1*/
break;
case 4: /* store attribute name */
if (isws) { ++phase; } /*<tag a */
else if (iseq) { phase=6; *ibuf=ch; ++ibuf; } /*<tag a=*/
else if (isn2) { *ibuf=ch; ++ibuf; } /*<tag a*/ /*-t1*/
else { err=3; } /*<tag a*/ /*/<>"?!'t*/
break;
case 5: /* eat whitespace after attribute name */
if (isws) { } /*<tag attr */
else if (iseq) { ++phase; *ibuf=ch; ++ibuf; } /*<tag attr =*/
else { err=3; } /*<tag attr */ /*/<>"?!-'t1*/
break;
case 6: /* eat whitespace after = */
if (isws) { } /*<tag attr= */
else if (isqu) { ++phase; *ibuf=ch; ++ibuf; } /*<tag attr="*/
else if (issi) { phase=8; *ibuf=ch; ++ibuf; } /*<tag attr='*/
else { err=3; } /*<tag attr=*/ /*/<>=?!-t1*/
break;
case 7: /* store the attribute value ("value") */
if (isws) { *ibuf=' '; ++ibuf; } /*tag attr=" */
else if (islt) { err=3; } /*<tag attr="<*/
else if (isqu) { phase=9; *ibuf=ch; ++ibuf; } /*<tag attr=""*/
else { *ibuf=ch; ++ibuf; } /*tag attr="*/ /*/>=?!-'t1*/
break;
case 8: /* store the attribute value ('value') */
if (isws) { *ibuf=' '; ++ibuf; } /*<tag attr=' */
else if (islt) { err=3; } /*<tag attr='<*/
else if (issi) { ++phase; *ibuf=ch; ++ibuf; } /*<tag attr=''*/
else { *ibuf=ch; ++ibuf; } /*tag attr='*/ /*/>="?!-t1*/
break;
case 9: /* eat whitespace after attribute value */
if (isws) { phase=3; *ibuf=' '; ++ibuf; } /*<tag attr="value" */
else if (issl) { phase=13; strncpy(ibuf," /",2); ibuf+=2; } /*..."/*/
else if (isgt) { phase=100; strncpy(ibuf," >",2); ibuf+=2; } /*...">*/
else { err=3; } /*tag attr="value"*/ /*<="?!-'t1*/
break;
case 10: /* handle the first char after </ */
if (isn1) { ++phase; *ibuf=ch; ++ibuf; } /*</t*/
else { err=3; } /*</*/ /* /<>="?!-'1*/
break;
case 11: /* store tag name and a single space after it */
if (isws) { ++phase; *ibuf=' '; ++ibuf; } /*</t */
else if (isgt) { phase=100; strncpy(ibuf," >",2); ibuf+=2; } /*</t>*/
else if (isn2) { *ibuf=ch; ++ibuf; } /*</t*/ /*-t*/
else { err=3; } /*</t*/ /*/<="?!'*/
break;
case 12: /* eat whitespace before > */
if (isws) { } /*</tag */
else if (isgt) { phase=100; *ibuf=ch; ++ibuf; } /*</tag >*/
else { err=3; } /*</tag */ /*/<="?!-'t1*/
break;
case 13: /* ensure final > */
if (isgt) { phase=100; *ibuf=ch; ++ibuf; } /*<tag />*/
else { err=3; } /*<tag /*/ /* /<="?!-'t1*/
break;
case 14: /* skip tag (<? tag) */
if (isgt) { phase=0; } /*<?>*/
else { } /*<?*/ /* /<="?!-'t1*/
break;
case 15: /* find first - (<! tag) */
if (isgt) { phase=0; } /*<!>*/
else if (isda) { ++phase; } /*<!-*/
else if (islb) { ++square; phase=20; } /*<![*/
else { phase=14; } /*<!*/ /* /<="?!'t1*/
break;
case 16: /* find second - (<!- tag) */
if (isgt) { phase=0; } /*<!->*/
else if (isda) { ++phase; } /*<!--*/
else { phase=14; } /*<!-*/ /* /<="?!'t1*/
break;
case 17: /* skip comment */
if (isda) { ++phase; } /*-*/
else { } /**/ /* /<>="?!'t1*/
break;
case 18: /* find second - in --> */
if (isda) { ++phase; } /*--*/
else { phase=17; } /*-*/ /* /<>="?!'t1*/
break;
case 19: /* find > in --> */
if (isgt) { phase=0; } /*-->*/
else { err=3; } /*--*/ /* /<="?!-'t1*/
break;
case 20: /* find ] in <![ */
if (islb) { ++square; }
if (isrb) { if (--square==0) ++phase; }
else { } /**/ /* /<>="?!'t1*/
break;
case 21: /* find > in (<![CDATA[]] tag) */
if (isgt) { phase=0; } /*<?>*/
else { } /*<?*/ /* /<="?!-'t1*/
break;
};
if (err) break;
}
if (!err && (phase != 100)) {
err=4; /* incomplete xml file */
}
if (!err) *ibuf = 0;
/* buf is now in one of these formats: */
/* <tag > or </tag > or <tag /> or */
/* <tag attr="value" ... attr="value" > or */
/* <tag attr="value" ... attr="value" /> */
/* allocate memory for a node */
node = new nodetype; /*(nodetype*)malloc(so_n);*/
if (node) {
node->tag = node->text = 0;
node->attrs = node->values = 0;
node->children = 0;
node->childreni = 0;
}
if (!err) {
if (!node) {
err=5; /* out of memory */
} else {
memset(node, 0, sizeof(node));
} }
/* put tag name into node */
if (!err) {
wspos = strchr(buf, ' ');
node->tag = new char[wspos-buf]; /*(char*)malloc(so_c*(wspos-buf));*/
if (!node->tag) err=6; /* out of memory */
}
if (!err) {
strncpy(node->tag, buf+1, wspos-buf-1);
node->tag[wspos-buf-1] = 0;
if (buf[1] == '/') { /* end tag (not err) */
delete [] buf; /*free(buf);*/
return node;
}
/* estimate # of attrs and values (maybe too much) */
pe = strstr(wspos, "\" ");
pe2 = strstr(wspos, "' ");
for (natts=0; pe || pe2; ++natts) {
if (!pe || (pe2 && (pe2 < pe))) pe = pe2;
pe2 = strstr(pe+1, "' "); pe = strstr(pe+1, "\" ");
}
node->attrs = new char *[natts+1]; /*(char**)malloc(so_cp*(natts+1));*/
node->values = new char *[natts+1]; /*(char**)malloc(so_cp*(natts+1));*/
if (!node->attrs||!node->values) {
if (node->attrs) delete [] node->attrs; /*free(node->attrs);*/
if (node->values) delete [] node->values; /*free(node->values);*/
node->attrs=node->values=0; err=7; /* out of memory */
} }
/* put attrs and values into node (and recount #) */
if (!err) { ps = wspos+1;
pe = strstr(wspos, "=\""); pe2 = strstr(wspos, "='");
}
for (natts=0, ia=0; !err && (pe||pe2); ++ia, ++natts) {
if (!pe || (pe2 && (pe2 < pe))) pe = pe2;
node->attrs[ia] = new char[pe-ps+1]; /*(char*)malloc(so_c*(pe-ps+1));*/
if (!node->attrs[ia]) {
node->values[ia]=0; err=8; break; /* out of memory */
}
strncpy(node->attrs[ia], ps, pe-ps);
node->attrs[ia][pe-ps] = 0;
node->attrs[ia+1] = 0;
ps = pe + 2;
if (pe == pe2) pe = strstr(ps, "' ");
else pe = strstr(ps, "\" ");
node->values[ia] = new char[pe-ps+1]; /*(char*)malloc(so_c*(pe-ps+1));*/
if (!node->values[ia]) { err=9; break; } /* out of memory */
strncpy(node->values[ia], ps, pe-ps);
node->values[ia][pe-ps] = 0;
node->values[ia+1] = 0;
ps = pe + 2;
pe2 = strstr(ps, "='"); pe = strstr(ps, "=\"");
}
if (!err) {
node->attrs[natts] = node->values[natts] = 0;
/* no children for self-contained tag, save memory */
if (buf[strlen(buf)-2] == '/') {
ntex = 1; nchi = 1;
}
/* no body or children yet */
node->text = new char[ntex]; /*(char*)malloc(so_c*ntex);*/
node->children = new nodetype *[nchi]; /*(nodetype**)malloc(so_np*nchi);*/
node->childreni = new int [nchi]; /*(int*)malloc(so_i*nchi);*/
if (!node->text || !node->children || !node->childreni) {
err=10; /* out of memory */
} else {
node->text[0] = 0; node->children[0] = 0; node->childreni[0] = 0;
} }
if (!err) {
if (ntex == 1) { /* end tag, no children (not err) */
delete [] buf; /*free(buf);*/
return node;
} }
/* read child tags */
for (ichi=0; !err; ++ichi) {
/* add more text to node (between child tags) */
for (ch=fgetc(fpi); ch != '<'; ch=fgetc(fpi)) {
if (feof(fpi)) {
err=11; /* incomplete xml file */
break;
}
if ((itex+2) == ntex) { /* realloc text */
ntex += textinc;
buftmp = new char[ntex]; /*(char*)malloc(so_c*ntex);*/
if (!buftmp) { err=12; break; } /* out of memory */
strncpy(buftmp, node->text, itex);
delete [] node->text; /*free(node->text);*/
node->text = buftmp;
}
node->text[itex++] = ch;
}
if (err) continue;
node->text[itex] = 0;
ungetc(ch, fpi); /* oops, put back next tag's "<" */
child = readbasicxmlnode(fpi);
if (child) {
if (child->tag[0] == '/') {
if (strcmp(child->tag+1, node->tag)) {
err=13; /* end tag mismatch */
}
delete [] child->tag; /*free(child->tag);*/
delete child; /*free(child);*/
child=0;
} } else if (!err) {
err=14; /* bad child */
}
if (err || !child) break;
node->children[ichi] = child;
node->childreni[ichi] = itex;
if ((ichi+2) == nchi) { /* realloc child tags array */
nchi += childinc;
nodetmp = new nodetype *[nchi]; /*(nodetype**)malloc(so_np*nchi);*/
if (!nodetmp) { err=15; continue; } /* out of memory */
nodeitmp = new int [nchi]; /*(int*)malloc(so_i*nchi);*/
if (!nodeitmp) {
delete [] nodetmp; /*free(nodetmp);*/
nodetmp=0; err=16; continue; /* out of memory */
}
for (ii=0; ii <= ichi; ++ii) { /* copy */
nodetmp[ii] = node->children[ii];
nodeitmp[ii] = node->childreni[ii];
}
delete [] node->children; /*free(node->children);*/
delete [] node->childreni; /*free(node->childreni);*/
node->children = nodetmp;
node->childreni = nodeitmp;
} }
if (node) {
if (node->children) {
node->children[ichi] = 0;
node->childreni[ichi] = 0;
if (err) { /* delete complete node */
deletebasicxmlnode(node); node = 0;
} } else { /* delete partial node */
if (node->tag) delete [] node->tag; /*free(node->tag);*/
if (node->text) delete [] node->text; /*free(node->text);*/
if (node->attrs) {
for (ii=0; node->attrs[ii]; ++ii) {
delete [] node->attrs[ii]; /*free(node->attrs[ii]);*/
} }
if (node->values) {
for (ii=0; node->values[ii]; ++ii) {
delete [] node->values[ii]; /*free(node->values[ii]);*/
} }
if (node->attrs) delete node->attrs; /*free(node->attrs);*/
if (node->values) delete node->values; /*free(node->values);*/
delete node; /*free(node);*/
node = 0;
} }
if (buf) delete [] buf; /*free(buf);*/
return node; /* whew! return complete node */
}
/* printbasicxmlnode: prints to console */
void printbasicxmlnode( struct basicxmlnode * node )
{
int ii, ch, it=0, haschildren=0; if (!node) return;
if (node->children[0]) haschildren = 1;
printf("<%s", node->tag); /* print tag */
for (ii=0; node->attrs[ii]; ++ii) { /* print attrs */
printf(" %s=", node->attrs[ii]);
ch = strstr(node->values[ii], "\"")? '\'': '"';
printf("%c%s%c", ch, node->values[ii], ch);
}
if (!haschildren && !node->text[0]) printf(" /");
printf(">");
for (ii=0; node->children[ii]; ++ii) { /* print children */
while (it < node->childreni[ii]) /* print text */
printf("%c", node->text[it++]);
printbasicxmlnode(node->children[ii]);
}
printf("%s", node->text + it); /* print remaining text */
if (haschildren || node->text[0]) {
printf("</%s>", node->tag); /* print end tag */
}
}
/* printbasicxmlnodetagnames: visits each node once */
void printbasicxmlnodetagnames( struct basicxmlnode * node )
{
int sp=0, n;
int stack[100];
struct basicxmlnode * root = node;
while (1) {
printf("%s\n", node->tag);
/* find the next node */
stack[sp++] = 0;
while (1) {
node = root; /* find the parent */
for (n=0; n < (sp-1); ++n) {
node = node->children[stack[n]];
}
for (n=0; node->children[n]; ++n) ;
if (stack[sp-1] != n) break;
if (--sp == 0) break;
++stack[sp-1];
}
if (sp == 0) break;
node = node->children[stack[sp-1]];
}
}
/* main: read an xml file and print it */
int main()
{
struct basicxmlnode * root = NULL;
FILE * fp = fopen("file.xml", "rt");
if (!fp) { printf("open failed\n"); return 1; }
root = readbasicxmlnode(fp);
fclose(fp);
if (!root) { printf("read failed\n"); return 2; }
printbasicxmlnode(root);
deletebasicxmlnode(root);
return 0;
}
Whew!
(You can also download the C++ version, the C version and the Visual C++ 6.0 project version.)
E-mail Dan Howard about this article
