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 + + * 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 * 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 @@ ezXML

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 + 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: - ezxml-0.8.tar.gz + lightweight (less than 20k compiled). The latest version is available + here: + ezxml-0.8.5.tar.gz

Example Usage

- Given the following example xml document: + Given the following example XML document:

<?xml version="1.0"?>
@@ -37,7 +38,7 @@ </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:

@@ -58,7 +59,7 @@

Alternately, the following would print out the name of the second driver - of the first team: + on the first team:

ezxml_t f1 = ezxml_parse_file("formula1.xml");
@@ -74,14 +75,12 @@ 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 + 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 @@ -89,9 +88,10 @@
     
  • - 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.
     
  • @@ -106,7 +106,7 @@ "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 + ezxml_toxml() will convert an ezXML structure back to XML with sub tag locations intact.

    • 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: @@ -25,7 +25,7 @@ -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
    is reported as a sub tag, but the location of
    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 +Copyright 2004, 2005 Aaron Voisine 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 + * Copyright 2004, 2005 Aaron Voisine * * 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 + * Copyright 2004, 2005 Aaron Voisine * * 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 #include #include @@ -84,37 +85,40 @@ #include #include #include -#ifndef __ECOS__ +#ifndef EZXML_NOMMAP #include -#endif +#endif // EZXML_NOMMAP #include #include "ezxml.h" #include -#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 + 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, "<", 4)) *(s++) = '<'; - else if (! strncmp(s, ">", 4)) *(s++) = '>'; - else if (! strncmp(s, """, 6)) *(s++) = '"'; - else if (! strncmp(s, "'", 6)) *(s++) = '\''; - else if (! strncmp(s, "&", 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 ", 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")) { // + 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, "'); + 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, "")) == '>') 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 ") - 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 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, ""); // comments + else if (! strncmp(s, ""))) + 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