xmllint.c

libxml,在UNIX/LINUX下非常重要的一个库,为XML相关应用提供方便.目前上载的是最新版本,若要取得最新版本,请参考里面的readme.

		endTimer("Parsing and validating");
	    else
#endif
#ifdef LIBXML_VALID_ENABLED
	    if (valid)
		endTimer("Parsing and validating");
	    else
#endif
	    endTimer("Parsing");
	}

#ifdef LIBXML_VALID_ENABLED
	if (valid) {
	    if (xmlTextReaderIsValid(reader) != 1) {
		xmlGenericError(xmlGenericErrorContext,
			"Document %s does not validate\n", filename);
		progresult = XMLLINT_ERR_VALID;
	    }
	}
#endif /* LIBXML_VALID_ENABLED */
#ifdef LIBXML_SCHEMAS_ENABLED
	if (relaxng != NULL) {
	    if (xmlTextReaderIsValid(reader) != 1) {
		fprintf(stderr, "%s fails to validate\n", filename);
		progresult = XMLLINT_ERR_VALID;
	    } else {
		fprintf(stderr, "%s validates\n", filename);
	    }
	}
#endif
	/*
	 * Done, cleanup and status
	 */
	xmlFreeTextReader(reader);
	if (ret != 0) {
	    fprintf(stderr, "%s : failed to parse\n", filename);
	    progresult = XMLLINT_ERR_UNCLASS;
	}
    } else {
	fprintf(stderr, "Unable to open %s\n", filename);
	progresult = XMLLINT_ERR_UNCLASS;
    }
#ifdef LIBXML_PATTERN_ENABLED
    if (patstream != NULL) {
	xmlFreeStreamCtxt(patstream);
	patstream = NULL;
    }
#endif
#ifdef HAVE_SYS_MMAN_H
    if (memory) {
        xmlFreeParserInputBuffer(input);
	munmap((char *) base, info.st_size);
	close(fd);
    }
#endif
}

static void walkDoc(xmlDocPtr doc) {
    xmlTextReaderPtr reader;
    int ret;

#ifdef LIBXML_PATTERN_ENABLED
    xmlNodePtr root;
    const xmlChar *namespaces[22];
    int i;
    xmlNsPtr ns;

    root = xmlDocGetRootElement(doc);
    for (ns = root->nsDef, i = 0;ns != NULL && i < 20;ns=ns->next) {
        namespaces[i++] = ns->href;
        namespaces[i++] = ns->prefix;
    }
    namespaces[i++] = NULL;
    namespaces[i++] = NULL;

    if (pattern != NULL) {
        patternc = xmlPatterncompile((const xmlChar *) pattern, doc->dict,
	                             0, &namespaces[0]);
	if (patternc == NULL) {
	    xmlGenericError(xmlGenericErrorContext,
		    "Pattern %s failed to compile\n", pattern);
            progresult = XMLLINT_ERR_SCHEMAPAT;
	    pattern = NULL;
	}
    }
    if (patternc != NULL) {
        patstream = xmlPatternGetStreamCtxt(patternc);
	if (patstream != NULL) {
	    ret = xmlStreamPush(patstream, NULL, NULL);
	    if (ret < 0) {
		fprintf(stderr, "xmlStreamPush() failure\n");
		xmlFreeStreamCtxt(patstream);
		patstream = NULL;
            }
	}
    }
#endif /* LIBXML_PATTERN_ENABLED */
    reader = xmlReaderWalker(doc);
    if (reader != NULL) {
	if ((timing) && (!repeat)) {
	    startTimer();
	}
	ret = xmlTextReaderRead(reader);
	while (ret == 1) {
	    if ((debug)
#ifdef LIBXML_PATTERN_ENABLED
	        || (patternc)
#endif
	       )
		processNode(reader);
	    ret = xmlTextReaderRead(reader);
	}
	if ((timing) && (!repeat)) {
	    endTimer("walking through the doc");
	}
	xmlFreeTextReader(reader);
	if (ret != 0) {
	    fprintf(stderr, "failed to walk through the doc\n");
	    progresult = XMLLINT_ERR_UNCLASS;
	}
    } else {
	fprintf(stderr, "Failed to crate a reader from the document\n");
	progresult = XMLLINT_ERR_UNCLASS;
    }
#ifdef LIBXML_PATTERN_ENABLED
    if (patstream != NULL) {
	xmlFreeStreamCtxt(patstream);
	patstream = NULL;
    }
#endif
}
#endif /* LIBXML_READER_ENABLED */

/************************************************************************
 * 									*
 * 			Tree Test processing				*
 * 									*
 ************************************************************************/
static void parseAndPrintFile(char *filename, xmlParserCtxtPtr rectxt) {
    xmlDocPtr doc = NULL;
#ifdef LIBXML_TREE_ENABLED
    xmlDocPtr tmp;
#endif /* LIBXML_TREE_ENABLED */

    if ((timing) && (!repeat))
	startTimer();
    

#ifdef LIBXML_TREE_ENABLED
    if (filename == NULL) {
	if (generate) {
	    xmlNodePtr n;

	    doc = xmlNewDoc(BAD_CAST "1.0");
	    n = xmlNewDocNode(doc, NULL, BAD_CAST "info", NULL);
	    xmlNodeSetContent(n, BAD_CAST "abc");
	    xmlDocSetRootElement(doc, n);
	}
    }
#endif /* LIBXML_TREE_ENABLED */
#ifdef LIBXML_HTML_ENABLED
#ifdef LIBXML_PUSH_ENABLED
    else if ((html) && (push)) {
        FILE *f;

#if defined(_WIN32) || defined (__DJGPP__) && !defined (__CYGWIN__)
	f = fopen(filename, "rb");
#else
	f = fopen(filename, "r");
#endif
        if (f != NULL) {
            int res, size = 3;
            char chars[4096];
            htmlParserCtxtPtr ctxt;

            /* if (repeat) */
                size = 4096;
            res = fread(chars, 1, 4, f);
            if (res > 0) {
                ctxt = htmlCreatePushParserCtxt(NULL, NULL,
                            chars, res, filename, XML_CHAR_ENCODING_NONE);
                while ((res = fread(chars, 1, size, f)) > 0) {
                    htmlParseChunk(ctxt, chars, res, 0);
                }
                htmlParseChunk(ctxt, chars, 0, 1);
                doc = ctxt->myDoc;
                htmlFreeParserCtxt(ctxt);
            }
            fclose(f);
        }
    }
#endif /* LIBXML_PUSH_ENABLED */
    else if (html) {
	doc = htmlReadFile(filename, NULL, options);
    }
#endif /* LIBXML_HTML_ENABLED */
    else {
#ifdef LIBXML_PUSH_ENABLED
	/*
	 * build an XML tree from a string;
	 */
	if (push) {
	    FILE *f;

	    /* '-' Usually means stdin -<[e-mail]> */
	    if ((filename[0] == '-') && (filename[1] == 0)) {
	      f = stdin;
	    } else {
#if defined(_WIN32) || defined (__DJGPP__) && !defined (__CYGWIN__)
		f = fopen(filename, "rb");
#else
		f = fopen(filename, "r");
#endif
	    }
	    if (f != NULL) {
		int ret;
	        int res, size = 1024;
	        char chars[1024];
                xmlParserCtxtPtr ctxt;

		/* if (repeat) size = 1024; */
		res = fread(chars, 1, 4, f);
		if (res > 0) {
		    ctxt = xmlCreatePushParserCtxt(NULL, NULL,
		                chars, res, filename);
		    xmlCtxtUseOptions(ctxt, options);
		    while ((res = fread(chars, 1, size, f)) > 0) {
			xmlParseChunk(ctxt, chars, res, 0);
		    }
		    xmlParseChunk(ctxt, chars, 0, 1);
		    doc = ctxt->myDoc;
		    ret = ctxt->wellFormed;
		    xmlFreeParserCtxt(ctxt);
		    if (!ret) {
			xmlFreeDoc(doc);
			doc = NULL;
		    }
	        }
	    }
	} else
#endif /* LIBXML_PUSH_ENABLED */
        if (testIO) {
	    if ((filename[0] == '-') && (filename[1] == 0)) {
	        doc = xmlReadFd(0, NULL, NULL, options);
	    } else {
	        FILE *f;

#if defined(_WIN32) || defined (__DJGPP__) && !defined (__CYGWIN__)
		f = fopen(filename, "rb");
#else
		f = fopen(filename, "r");
#endif
		if (f != NULL) {
		    if (rectxt == NULL)
			doc = xmlReadIO((xmlInputReadCallback) myRead,
					(xmlInputCloseCallback) myClose, f,
					filename, NULL, options);
		    else
			doc = xmlCtxtReadIO(rectxt,
			                (xmlInputReadCallback) myRead,
					(xmlInputCloseCallback) myClose, f,
					filename, NULL, options);
		} else
		    doc = NULL;
	    }
	} else if (htmlout) {
	    xmlParserCtxtPtr ctxt;

	    if (rectxt == NULL)
		ctxt = xmlNewParserCtxt();
	    else
	        ctxt = rectxt;
	    if (ctxt == NULL) {	      
	        doc = NULL;
	    } else {
	        ctxt->sax->error = xmlHTMLError;
	        ctxt->sax->warning = xmlHTMLWarning;
	        ctxt->vctxt.error = xmlHTMLValidityError;
	        ctxt->vctxt.warning = xmlHTMLValidityWarning;

		doc = xmlCtxtReadFile(ctxt, filename, NULL, options);

		if (rectxt == NULL)
		    xmlFreeParserCtxt(ctxt);
	    }
#ifdef HAVE_SYS_MMAN_H
	} else if (memory) {
	    int fd;
	    struct stat info;
	    const char *base;
	    if (stat(filename, &info) < 0) 
		return;
	    if ((fd = open(filename, O_RDONLY)) < 0)
		return;
	    base = mmap(NULL, info.st_size, PROT_READ, MAP_SHARED, fd, 0) ;
	    if (base == (void *) MAP_FAILED)
	        return;

	    if (rectxt == NULL)
		doc = xmlReadMemory((char *) base, info.st_size,
		                    filename, NULL, options);
	    else
		doc = xmlCtxtReadMemory(rectxt, (char *) base, info.st_size,
			                filename, NULL, options);
	        
	    munmap((char *) base, info.st_size);
#endif
#ifdef LIBXML_VALID_ENABLED
	} else if (valid) {
	    xmlParserCtxtPtr ctxt = NULL;

	    if (rectxt == NULL)
		ctxt = xmlNewParserCtxt();
	    else
	        ctxt = rectxt;
	    if (ctxt == NULL) {	      
	        doc = NULL;
	    } else {
		doc = xmlCtxtReadFile(ctxt, filename, NULL, options);

		if (ctxt->valid == 0)
		    progresult = XMLLINT_ERR_RDFILE;
		if (rectxt == NULL)
		    xmlFreeParserCtxt(ctxt);
	    }
#endif /* LIBXML_VALID_ENABLED */
	} else {
	    if (rectxt != NULL)
	        doc = xmlCtxtReadFile(rectxt, filename, NULL, options);
	    else
		doc = xmlReadFile(filename, NULL, options);
	}
    }

    /*
     * If we don't have a document we might as well give up.  Do we
     * want an error message here?  <[e-mail]> */
    if (doc == NULL) {
	progresult = XMLLINT_ERR_UNCLASS;
	return;
    }

    if ((timing) && (!repeat)) {
	endTimer("Parsing");
    }

    /*
     * Remove DOCTYPE nodes
     */
    if (dropdtd) {
	xmlDtdPtr dtd;

	dtd = xmlGetIntSubset(doc);
	if (dtd != NULL) {
	    xmlUnlinkNode((xmlNodePtr)dtd);
	    xmlFreeDtd(dtd);
	}
    }

#ifdef LIBXML_XINCLUDE_ENABLED
    if (xinclude) {
	if ((timing) && (!repeat)) {
	    startTimer();
	}
	if (xmlXIncludeProcessFlags(doc, options) < 0)
	    progresult = XMLLINT_ERR_UNCLASS;
	if ((timing) && (!repeat)) {
	    endTimer("Xinclude processing");
	}
    }
#endif

#ifdef LIBXML_DEBUG_ENABLED
#ifdef LIBXML_XPATH_ENABLED
    /*
     * shell interaction
     */
    if (shell)  
        xmlShell(doc, filename, xmlShellReadline, stdout);
#endif
#endif

#ifdef LIBXML_TREE_ENABLED
    /*
     * test intermediate copy if needed.
     */
    if (copy) {
        tmp = doc;
	if (timing) {
	    startTimer();
	}
	doc = xmlCopyDoc(doc, 1);
	if (timing) {
	    endTimer("Copying");
	}
	if (timing) {
	    startTimer();
	}
	xmlFreeDoc(tmp);
	if (timing) {
	    endTimer("Freeing original");
	}
    }
#endif /* LIBXML_TREE_ENABLED */

#ifdef LIBXML_VALID_ENABLED
    if ((insert) && (!html)) {
        const xmlChar* list[256];
	int nb, i;
	xmlNodePtr node;

	if (doc->children != NULL) {
	    node = doc->children;
	    while ((node != NULL) && (node->last == NULL)) node = node->next;
	    if (node != NULL) {
		nb = xmlValidGetValidElements(node->last, NULL, list, 256);
		if (nb < 0) {
		    fprintf(stderr, "could not get valid list of elements\n");
		} else if (nb == 0) {
		    fprintf(stderr, "No element can be inserted under root\n");
		} else {
		    fprintf(stderr, "%d element types can be inserted under root:\n",
		           nb);
		    for (i = 0;i < nb;i++) {
			 fprintf(stderr, "%s\n", (char *) list[i]);
		    }
		}
	    }
	}    
    }else
#endif /* LIBXML_VALID_ENABLED */
#ifdef LIBXML_READER_ENABLED
    if (walker) {
        walkDoc(doc);
    }
#endif /* LIBXML_READER_ENABLED */
#ifdef LIBXML_OUTPUT_ENABLED
    if (noout == 0) {
        int ret;

	/*
	 * print it.
	 */
#ifdef LIBXML_DEBUG_ENABLED
	if (!debug) {
#endif
	    if ((timing) && (!repeat)) {
		startTimer();
	    }
#ifdef LIBXML_HTML_ENABLED
            if ((html) && (!xmlout)) {
		if (compress) {
		    htmlSaveFile(output ? output : "-", doc);
		}
		else if (encoding != NULL) {
		    if ( format ) {
			htmlSaveFileFormat(output ? output : "-", doc, encoding, 1);
		    }
		    else {
			htmlSaveFileFormat(output ? output : "-", doc, encoding, 0);
		    }
		}
		else if (format) {
		    htmlSaveFileFormat(output ? output : "-", doc, NULL, 1);
		}
		else {
		    FILE *out;
		    if (output == NULL)
			out = stdout;
		    else {
			out = fopen(output,"wb");
		    }
		    if (out != NULL) {
			if (htmlDocDump(out, doc) < 0)
			    progresult = XMLLINT_ERR_OUT;

			if (output != NULL)
			    fclose(out);
		    } else {
			fprintf(stderr, "failed to open %s\n", output);
			progresult = XMLLINT_ERR_OUT;
		    }
		}
		if ((timing) && (!repeat)) {
		    endTimer("Saving");
		}
	    } else
#endif
#ifdef LIBXML_C14N_ENABLED
            if (canonical) {