#include "libxslt/libxslt.h"
#include "libexslt/exslt.h"
+#include <stdio.h>
#ifdef HAVE_STRING_H
#include <string.h>
#endif
#ifdef HAVE_SYS_TIME_H
#include <sys/time.h>
#endif
+#ifdef HAVE_TIME_H
+#include <time.h>
+#endif
#ifdef HAVE_SYS_STAT_H
#include <sys/stat.h>
#endif
#include <libxml/debugXML.h>
#include <libxml/HTMLtree.h>
#include <libxml/xmlIO.h>
-#ifdef LIBXML_DOCB_ENABLED
-#include <libxml/DOCBparser.h>
-#endif
#ifdef LIBXML_XINCLUDE_ENABLED
#include <libxml/xinclude.h>
#endif
#ifdef LIBXML_CATALOG_ENABLED
#include <libxml/catalog.h>
#endif
+#include <libxml/parser.h>
#include <libxml/parserInternals.h>
+#include <libxml/uri.h>
#include <libxslt/xslt.h>
#include <libxslt/xsltInternals.h>
#include <libxslt/transform.h>
#include <libxslt/xsltutils.h>
#include <libxslt/extensions.h>
+#include <libxslt/security.h>
#include <libexslt/exsltconfig.h>
#if defined(WIN32) && !defined (__CYGWIN__)
-#ifdef _MSC_VER
+#if defined(_MSC_VER) || defined(__MINGW32__)
#include <winsock2.h>
-#pragma comment(lib, "ws2_32.lib")
#define gettimeofday(p1,p2)
#define HAVE_TIME_H
#include <time.h>
#define HAVE_STDARG_H
#include <stdarg.h>
+#define snprintf _snprintf
#endif /* _MS_VER */
#else /* WIN32 */
#if defined(HAVE_SYS_TIME_H)
#endif
#endif /* WIN32 */
+#ifdef HAVE_SYS_TIMEB_H
+#include <sys/timeb.h>
+#endif
+
#ifndef HAVE_STAT
# ifdef HAVE__STAT
/* MS C library seems to define stat and _stat. The definition
# endif
#endif
-xmlParserInputPtr xmlNoNetExternalEntityLoader(const char *URL,
- const char *ID,
- xmlParserCtxtPtr ctxt);
-
static int debug = 0;
static int repeat = 0;
static int timing = 0;
+static int dumpextensions = 0;
static int novalid = 0;
static int noout = 0;
-#ifdef LIBXML_DOCB_ENABLED
-static int docbook = 0;
-#endif
#ifdef LIBXML_HTML_ENABLED
static int html = 0;
#endif
+static int load_trace = 0;
#ifdef LIBXML_XINCLUDE_ENABLED
static int xinclude = 0;
#endif
static int profile = 0;
#define MAX_PARAMETERS 64
+#define MAX_PATHS 64
+#if LIBXML_VERSION >= 20600
+static int options = XSLT_PARSE_OPTIONS;
+#endif
static const char *params[MAX_PARAMETERS + 1];
static int nbparams = 0;
static xmlChar *strparams[MAX_PARAMETERS + 1];
static int nbstrparams = 0;
-static const char *output = NULL;
+static xmlChar *paths[MAX_PATHS + 1];
+static int nbpaths = 0;
+static char *output = NULL;
static int errorno = 0;
+static const char *writesubtree = NULL;
+
+/*
+ * Entity loading control and customization.
+ */
+static
+void parsePath(const xmlChar *path) {
+ const xmlChar *cur;
+
+ if (path == NULL)
+ return;
+ while (*path != 0) {
+ if (nbpaths >= MAX_PATHS) {
+ fprintf(stderr, "MAX_PATHS reached: too many paths\n");
+ return;
+ }
+ cur = path;
+ while ((*cur == ' ') || (*cur == ':'))
+ cur++;
+ path = cur;
+ while ((*cur != 0) && (*cur != ' ') && (*cur != ':'))
+ cur++;
+ if (cur != path) {
+ paths[nbpaths] = xmlStrndup(path, cur - path);
+ if (paths[nbpaths] != NULL)
+ nbpaths++;
+ path = cur;
+ }
+ }
+}
+
+xmlExternalEntityLoader defaultEntityLoader = NULL;
+
+static xmlParserInputPtr
+xsltprocExternalEntityLoader(const char *URL, const char *ID,
+ xmlParserCtxtPtr ctxt) {
+ xmlParserInputPtr ret;
+ warningSAXFunc warning = NULL;
+
+ int i;
+ const char *lastsegment = URL;
+ const char *iter = URL;
+
+ if (nbpaths > 0) {
+ while (*iter != 0) {
+ if (*iter == '/')
+ lastsegment = iter + 1;
+ iter++;
+ }
+ }
+
+ if ((ctxt != NULL) && (ctxt->sax != NULL)) {
+ warning = ctxt->sax->warning;
+ ctxt->sax->warning = NULL;
+ }
+ if (defaultEntityLoader != NULL) {
+ ret = defaultEntityLoader(URL, ID, ctxt);
+ if (ret != NULL) {
+ if (warning != NULL)
+ ctxt->sax->warning = warning;
+ if (load_trace) {
+ fprintf \
+ (stderr,
+ "Loaded URL=\"%s\" ID=\"%s\"\n",
+ URL ? URL : "(null)",
+ ID ? ID : "(null)");
+ }
+ return(ret);
+ }
+ }
+ for (i = 0;i < nbpaths;i++) {
+ xmlChar *newURL;
+
+ newURL = xmlStrdup((const xmlChar *) paths[i]);
+ newURL = xmlStrcat(newURL, (const xmlChar *) "/");
+ newURL = xmlStrcat(newURL, (const xmlChar *) lastsegment);
+ if (newURL != NULL) {
+ ret = defaultEntityLoader((const char *)newURL, ID, ctxt);
+ if (ret != NULL) {
+ if (warning != NULL)
+ ctxt->sax->warning = warning;
+ if (load_trace) {
+ fprintf \
+ (stderr,
+ "Loaded URL=\"%s\" ID=\"%s\"\n",
+ newURL,
+ ID ? ID : "(null)");
+ }
+ xmlFree(newURL);
+ return(ret);
+ }
+ xmlFree(newURL);
+ }
+ }
+ if (warning != NULL) {
+ ctxt->sax->warning = warning;
+ if (URL != NULL)
+ warning(ctxt, "failed to load external entity \"%s\"\n", URL);
+ else if (ID != NULL)
+ warning(ctxt, "failed to load external entity \"%s\"\n", ID);
+ }
+ return(NULL);
+}
/*
* Internal timing routines to remove the necessity to have unix-specific
* function calls
*/
+#ifndef HAVE_GETTIMEOFDAY
+#ifdef HAVE_SYS_TIMEB_H
+#ifdef HAVE_SYS_TIME_H
+#ifdef HAVE_FTIME
+
+int
+my_gettimeofday(struct timeval *tvp, void *tzp)
+{
+ struct timeb timebuffer;
+
+ ftime(&timebuffer);
+ if (tvp) {
+ tvp->tv_sec = timebuffer.time;
+ tvp->tv_usec = timebuffer.millitm * 1000L;
+ }
+ return (0);
+}
+#define HAVE_GETTIMEOFDAY 1
+#define gettimeofday my_gettimeofday
+
+#endif /* HAVE_FTIME */
+#endif /* HAVE_SYS_TIME_H */
+#endif /* HAVE_SYS_TIMEB_H */
+#endif /* !HAVE_GETTIMEOFDAY */
#if defined(HAVE_GETTIMEOFDAY)
-static struct timeval begin, end;
+static struct timeval begin, endtime;
/*
* startTimer: call where you want to start timing
*/
long msec;
va_list ap;
- gettimeofday(&end, NULL);
- msec = end.tv_sec - begin.tv_sec;
+ gettimeofday(&endtime, NULL);
+ msec = endtime.tv_sec - begin.tv_sec;
msec *= 1000;
- msec += (end.tv_usec - begin.tv_usec) / 1000;
+ msec += (endtime.tv_usec - begin.tv_usec) / 1000;
#ifndef HAVE_STDARG_H
#error "endTimer required stdarg functions"
* This is obviously less accurate, but there's little we can do about
* that.
*/
+#ifndef CLOCKS_PER_SEC
+#define CLOCKS_PER_SEC 100
+#endif
-clock_t begin, end;
+clock_t begin, endtime;
static void startTimer(void)
{
begin=clock();
long msec;
va_list ap;
- end=clock();
- msec = ((end-begin) * 1000) / CLOCKS_PER_SEC;
+ endtime=clock();
+ msec = ((endtime-begin) * 1000) / CLOCKS_PER_SEC;
#ifndef HAVE_STDARG_H
#error "endTimer required stdarg functions"
}
#endif
+/*
+ * xsltSubtreeCheck:
+ *
+ * allow writes only on a subtree specified on the command line
+ */
+static int
+xsltSubtreeCheck(xsltSecurityPrefsPtr sec ATTRIBUTE_UNUSED,
+ xsltTransformContextPtr ctxt ATTRIBUTE_UNUSED,
+ const char *value ATTRIBUTE_UNUSED) {
+ int len, ret;
+
+ if (writesubtree == NULL)
+ return(0);
+ if (value == NULL)
+ return(-1);
+
+ len = xmlStrlen(BAD_CAST writesubtree);
+ ret = xmlStrncmp(BAD_CAST writesubtree, BAD_CAST value, len);
+ if (ret == 0)
+ return(1);
+ return(0);
+}
+
static void
xsltProcess(xmlDocPtr doc, xsltStylesheetPtr cur, const char *filename) {
xmlDocPtr res;
xsltTransformContextPtr ctxt;
+
#ifdef LIBXML_XINCLUDE_ENABLED
if (xinclude) {
if (timing)
startTimer();
+#if LIBXML_VERSION >= 20603
+ xmlXIncludeProcessFlags(doc, XSLT_PARSE_OPTIONS);
+#else
xmlXIncludeProcess(doc);
+#endif
if (timing) {
endTimer("XInclude processing %s", filename);
}
res = xsltApplyStylesheet(cur, doc, params);
xmlFreeDoc(res);
xmlFreeDoc(doc);
+#if LIBXML_VERSION >= 20600
#ifdef LIBXML_HTML_ENABLED
if (html)
- doc = htmlParseFile(filename, NULL);
+ doc = htmlReadFile(filename, NULL, options);
else
#endif
-#ifdef LIBXML_DOCB_ENABLED
- if (docbook)
- doc = docbParseFile(filename, NULL);
+ doc = xmlReadFile(filename, NULL, options);
+#else
+#ifdef LIBXML_HTML_ENABLED
+ if (html)
+ doc = htmlParseFile(filename, NULL);
else
#endif
doc = xmlParseFile(filename);
+#endif
}
}
ctxt = xsltNewTransformContext(cur, doc);
if (ctxt == NULL)
return;
+ xsltSetCtxtParseOptions(ctxt, options);
if (profile) {
res = xsltApplyStylesheetUser(cur, doc, params, NULL,
stderr, ctxt);
}
if (ctxt->state == XSLT_STATE_ERROR)
errorno = 9;
+ if (ctxt->state == XSLT_STATE_STOPPED)
+ errorno = 10;
xsltFreeTransformContext(ctxt);
if (timing) {
if (repeat)
xmlFreeDoc(res);
} else {
- xsltRunStylesheet(cur, doc, params, output, NULL, NULL);
+
+ ctxt = xsltNewTransformContext(cur, doc);
+ if (ctxt == NULL)
+ return;
+ if (profile) {
+ xsltRunStylesheetUser(cur, doc, params, output,
+ NULL, NULL, stderr, ctxt);
+ } else {
+ xsltRunStylesheetUser(cur, doc, params, output,
+ NULL, NULL, NULL, ctxt);
+ }
+ if (ctxt->state == XSLT_STATE_ERROR)
+ errorno = 9;
+ xsltFreeTransformContext(ctxt);
if (timing)
endTimer("Running stylesheet and saving result");
xmlFreeDoc(doc);
printf("\t--output file or -o file: save to a given file\n");
printf("\t--timing: display the time used\n");
printf("\t--repeat: run the transformation 20 times\n");
+#ifdef LIBXML_DEBUG_ENABLED
printf("\t--debug: dump the tree of the result instead\n");
+#endif
+ printf("\t--dumpextensions: dump the registered extension elements and functions to stdout\n");
printf("\t--novalid skip the Dtd loading phase\n");
printf("\t--noout: do not dump the result\n");
printf("\t--maxdepth val : increase the maximum depth\n");
+#if LIBXML_VERSION >= 20600
+ printf("\t--maxparserdepth val : increase the maximum parser depth\n");
+#endif
#ifdef LIBXML_HTML_ENABLED
printf("\t--html: the input document is(are) an HTML file(s)\n");
#endif
-#ifdef LIBXML_DOCB_ENABLED
- printf("\t--docbook: the input document is SGML docbook\n");
-#endif
printf("\t--param name value : pass a (parameter,value) pair\n");
- printf("\t value is an XPath expression.\n");
+ printf("\t value is an UTF8 XPath expression.\n");
printf("\t string values must be quoted like \"'string'\"\n or");
printf("\t use stringparam to avoid it\n");
- printf("\t--stringparam name value : pass a (parameter,string value) pair\n");
- printf("\t--nonet refuse to fetch DTDs or entities over network\n");
+ printf("\t--stringparam name value : pass a (parameter, UTF8 string value) pair\n");
+ printf("\t--path 'paths': provide a set of paths for resources\n");
+ printf("\t--nonet : refuse to fetch DTDs or entities over network\n");
+ printf("\t--nowrite : refuse to write to any file or resource\n");
+ printf("\t--nomkdir : refuse to create directories\n");
+ printf("\t--writesubtree path : allow file write only with the path subtree\n");
#ifdef LIBXML_CATALOG_ENABLED
printf("\t--catalogs : use SGML catalogs from $SGML_CATALOG_FILES\n");
printf("\t otherwise XML Catalogs starting from \n");
#ifdef LIBXML_XINCLUDE_ENABLED
printf("\t--xinclude : do XInclude processing on document intput\n");
#endif
+ printf("\t--load-trace : print trace of all external entites loaded\n");
printf("\t--profile or --norman : dump profiling informations \n");
+ printf("\nProject libxslt home page: http://xmlsoft.org/XSLT/\n");
+ printf("To report bugs and get help: http://xmlsoft.org/XSLT/bugs.html\n");
}
int
int i;
xsltStylesheetPtr cur = NULL;
xmlDocPtr doc, style;
+ xsltSecurityPrefsPtr sec = NULL;
if (argc <= 1) {
usage(argv[0]);
LIBXML_TEST_VERSION
- xmlLineNumbersDefault(1);
+ sec = xsltNewSecurityPrefs();
+ xsltSetDefaultSecurityPrefs(sec);
+ defaultEntityLoader = xmlGetExternalEntityLoader();
+ xmlSetExternalEntityLoader(xsltprocExternalEntityLoader);
for (i = 1; i < argc; i++) {
if (!strcmp(argv[i], "-"))
(!strcmp(argv[i], "-output")) ||
(!strcmp(argv[i], "--output"))) {
i++;
- output = argv[i];
+#if defined(WIN32) || defined (__CYGWIN__)
+ output = xmlCanonicPath(argv[i]);
+ if (output == NULL)
+#endif
+ output = (char *) xmlStrdup((xmlChar *) argv[i]);
} else if ((!strcmp(argv[i], "-V")) ||
(!strcmp(argv[i], "-version")) ||
(!strcmp(argv[i], "--version"))) {
} else if ((!strcmp(argv[i], "-noout")) ||
(!strcmp(argv[i], "--noout"))) {
noout++;
-#ifdef LIBXML_DOCB_ENABLED
- } else if ((!strcmp(argv[i], "-docbook")) ||
- (!strcmp(argv[i], "--docbook"))) {
- docbook++;
-#endif
#ifdef LIBXML_HTML_ENABLED
} else if ((!strcmp(argv[i], "-html")) ||
(!strcmp(argv[i], "--html"))) {
profile++;
} else if ((!strcmp(argv[i], "-nonet")) ||
(!strcmp(argv[i], "--nonet"))) {
- xmlSetExternalEntityLoader(xmlNoNetExternalEntityLoader);
+ defaultEntityLoader = xmlNoNetExternalEntityLoader;
+ } else if ((!strcmp(argv[i], "-nowrite")) ||
+ (!strcmp(argv[i], "--nowrite"))) {
+ xsltSetSecurityPrefs(sec, XSLT_SECPREF_WRITE_FILE,
+ xsltSecurityForbid);
+ xsltSetSecurityPrefs(sec, XSLT_SECPREF_CREATE_DIRECTORY,
+ xsltSecurityForbid);
+ xsltSetSecurityPrefs(sec, XSLT_SECPREF_WRITE_NETWORK,
+ xsltSecurityForbid);
+ } else if ((!strcmp(argv[i], "-nomkdir")) ||
+ (!strcmp(argv[i], "--nomkdir"))) {
+ xsltSetSecurityPrefs(sec, XSLT_SECPREF_CREATE_DIRECTORY,
+ xsltSecurityForbid);
+ } else if ((!strcmp(argv[i], "-writesubtree")) ||
+ (!strcmp(argv[i], "--writesubtree"))) {
+ i++;
+ writesubtree = argv[i];
+ xsltSetSecurityPrefs(sec, XSLT_SECPREF_WRITE_FILE,
+ xsltSubtreeCheck);
+ } else if ((!strcmp(argv[i], "-path")) ||
+ (!strcmp(argv[i], "--path"))) {
+ i++;
+ parsePath(BAD_CAST argv[i]);
#ifdef LIBXML_CATALOG_ENABLED
} else if ((!strcmp(argv[i], "-catalogs")) ||
(!strcmp(argv[i], "--catalogs"))) {
xinclude++;
xsltSetXIncludeDefault(1);
#endif
+ } else if ((!strcmp(argv[i], "-load-trace")) ||
+ (!strcmp(argv[i], "--load-trace"))) {
+ load_trace++;
} else if ((!strcmp(argv[i], "-param")) ||
(!strcmp(argv[i], "--param"))) {
i++;
(!strcmp(argv[i], "--stringparam"))) {
const xmlChar *string;
xmlChar *value;
- int len;
i++;
params[nbparams++] = argv[i++];
string = (const xmlChar *) argv[i];
- len = xmlStrlen(string);
if (xmlStrchr(string, '"')) {
if (xmlStrchr(string, '\'')) {
fprintf(stderr,
if (value > 0)
xsltMaxDepth = value;
}
- } else {
+#if LIBXML_VERSION >= 20600
+ } else if ((!strcmp(argv[i], "-maxparserdepth")) ||
+ (!strcmp(argv[i], "--maxparserdepth"))) {
+ int value;
+
+ i++;
+ if (sscanf(argv[i], "%d", &value) == 1) {
+ if (value > 0)
+ xmlParserMaxDepth = value;
+ }
+#endif
+ } else if ((!strcmp(argv[i],"-dumpextensions"))||
+ (!strcmp(argv[i],"--dumpextensions"))) {
+ dumpextensions++;
+
+ } else {
fprintf(stderr, "Unknown option %s\n", argv[i]);
usage(argv[0]);
return (3);
}
params[nbparams] = NULL;
+#if LIBXML_VERSION < 20600
+ /*
+ * The old parser interfaces uses the global variables
+ */
if (novalid == 0)
xmlLoadExtDtdDefaultValue = XML_DETECT_IDS | XML_COMPLETE_ATTRS;
else
xmlLoadExtDtdDefaultValue = 0;
-
-
- /*
- * Replace entities with their content.
- */
xmlSubstituteEntitiesDefault(1);
+ xmlLineNumbersDefault(1);
+#else
+ if (novalid != 0)
+ options = XML_PARSE_NOENT | XML_PARSE_NOCDATA;
+#endif
/*
* Register the EXSLT extensions and the test module
exsltRegisterAll();
xsltRegisterTestModule();
+ if (dumpextensions)
+ xsltDebugDumpExtensions(NULL);
+
for (i = 1; i < argc; i++) {
if ((!strcmp(argv[i], "-maxdepth")) ||
(!strcmp(argv[i], "--maxdepth"))) {
i++;
continue;
+#if LIBXML_VERSION >= 20600
+ } else if ((!strcmp(argv[i], "-maxparserdepth")) ||
+ (!strcmp(argv[i], "--maxparserdepth"))) {
+ i++;
+ continue;
+#endif
} else if ((!strcmp(argv[i], "-o")) ||
(!strcmp(argv[i], "-output")) ||
(!strcmp(argv[i], "--output"))) {
i++;
continue;
+ } else if ((!strcmp(argv[i], "-writesubtree")) ||
+ (!strcmp(argv[i], "--writesubtree"))) {
+ i++;
+ continue;
+ } else if ((!strcmp(argv[i], "-path")) ||
+ (!strcmp(argv[i], "--path"))) {
+ i++;
+ continue;
}
if ((!strcmp(argv[i], "-param")) || (!strcmp(argv[i], "--param"))) {
i += 2;
continue;
}
+ if ((!strcmp(argv[i], "-stringparam")) ||
+ (!strcmp(argv[i], "--stringparam"))) {
+ i += 2;
+ continue;
+ }
if ((argv[i][0] != '-') || (strcmp(argv[i], "-") == 0)) {
if (timing)
startTimer();
+#if LIBXML_VERSION >= 20600
+ style = xmlReadFile((const char *) argv[i], NULL, options);
+#else
style = xmlParseFile((const char *) argv[i]);
+#endif
if (timing)
endTimer("Parsing stylesheet %s", argv[i]);
if (style == NULL) {
/* it is an embedded stylesheet */
xsltProcess(style, cur, argv[i]);
xsltFreeStylesheet(cur);
+ cur = NULL;
goto done;
}
cur = xsltParseStylesheetDoc(style);
if (cur != NULL) {
- if (cur->indent == 1)
- xmlIndentTreeOutput = 1;
- else
- xmlIndentTreeOutput = 0;
+ if (cur->errors != 0) {
+ errorno = 5;
+ goto done;
+ }
i++;
} else {
xmlFreeDoc(style);
}
}
+#if LIBXML_VERSION < 20600
/*
+ * The old parser interfaces uses the global variables
* disable CDATA from being built in the document tree
*/
xmlDefaultSAXHandlerInit();
xmlDefaultSAXHandler.cdataBlock = NULL;
+#endif
if ((cur != NULL) && (cur->errors == 0)) {
for (; i < argc; i++) {
doc = NULL;
if (timing)
startTimer();
+#if LIBXML_VERSION >= 20600
#ifdef LIBXML_HTML_ENABLED
if (html)
- doc = htmlParseFile(argv[i], NULL);
+ doc = htmlReadFile(argv[i], NULL, options);
else
#endif
-#ifdef LIBXML_DOCB_ENABLED
- if (docbook)
- doc = docbParseFile(argv[i], NULL);
+ doc = xmlReadFile(argv[i], NULL, options);
+#else
+#ifdef LIBXML_HTML_ENABLED
+ if (html)
+ doc = htmlParseFile(argv[i], NULL);
else
#endif
doc = xmlParseFile(argv[i]);
+#endif
if (doc == NULL) {
fprintf(stderr, "unable to parse %s\n", argv[i]);
errorno = 6;
xsltProcess(doc, cur, argv[i]);
}
}
+done:
if (cur != NULL)
xsltFreeStylesheet(cur);
for (i = 0;i < nbstrparams;i++)
xmlFree(strparams[i]);
-done:
+ if (output != NULL)
+ xmlFree(output);
+ xsltFreeSecurityPrefs(sec);
xsltCleanupGlobals();
xmlCleanupParser();
xmlMemoryDump();