remote html works, some function aliases work

Author: Kyle Maxwell

TODO

@@ -10,19 +10,21 @@
 # - p/br support needs multicase handling
 # - reorganize project (at least tests, makefile.am src?!)
 # - unfilter inside of nested explicit scope
+# - functions inside magic groups?!?!
+- shortcut functions	(need c hash table)
+- fix position()
 - memory leaks
-- windows, linux builds
-- flags?! 
-		^ - force group-before
-		$ - force group-after
-		! - don't group?!
-		? - optional field
-- fix not()/set-difference
 - debugging options
-- fix position()
-- shortcut functions
+- check linux build
+- python
+- fix not()/set-difference
 - complain if empty / runtime no-match errors
 - CSS equations
 - saxon compatibility?!
-- functions inside magic groups?!?!
-- XML input converter?!
+- XML input converter?!
+- check windows build
+- flags?! 
+		^ - force group-before
+		$ - force group-after
+		! - don't group?!
+		? - optional field

configure.ac

@@ -11,11 +11,17 @@ AC_PROG_CC_C99
 AM_PROG_LEX
 AC_PROG_YACC
 AC_PROG_LIBTOOL
-# AC_CHECK_HEADER(argp.h, , AC_MSG_ERROR([could not find argp.h]))
-# AC_CHECK_HEADER(json/json.h, , AC_MSG_ERROR([could not find json.h]))
+AC_CHECK_HEADER(pcre.h, , AC_MSG_ERROR([could not find pcre.h]))
+AC_CHECK_HEADER(argp.h, , AC_MSG_ERROR([could not find argp.h]))
+AC_CHECK_HEADER(json/json.h, , AC_MSG_ERROR([could not find json.h]))
 
-# AC_CHECK_HEADER(/usr/include/libxml2/libxml/HTMLtree.h, , AC_MSG_ERROR([could not find libxml2/HTMLtree.h]))
-# AC_CHECK_HEADER(libexslt/exslt.h, , AC_MSG_ERROR([could not find exslt.h]))
+AC_PATH_PROG(XML2_CONFIG, xml2-config, , [$PATH])
+if test x$XML2_CONFIG = x ; then
+  AC_MSG_ERROR([libxml2 not present or not configured])
+else
+  XML2_CFLAGS="`$XML2_CONFIG --cflags`"
+  XML2_LIBS="`$XML2_CONFIG --libs`"
+fi
 
 AC_PATH_PROG(XSLT_CONFIG, xslt-config, , [$PATH])
 if test x$XSLT_CONFIG = x ; then
@@ -25,16 +31,16 @@ else
   XSLT_LIBS="`$XSLT_CONFIG --libs`"
 fi
 
-CPPFLAGS="$CPPFLAGS $XSLT_CFLAGS"
-LIBS="$LIBS $XSLT_LIBS"
+CPPFLAGS="$CPPFLAGS $XML2_CFLAGS $XSLT_CFLAGS"
+LIBS="$LIBS $XML2_LIBS $XSLT_LIBS"
 
-AC_CHECK_LIB(pcre, pcre_compile)
-AC_CHECK_LIB(argp, argp_parse)
+AC_CHECK_LIB(pcre, pcre_compile, , AC_MSG_ERROR([could not find pcre]))
+AC_CHECK_LIB(argp, argp_parse, , AC_MSG_ERROR([could not find argp]))
 AC_CHECK_LIB(json, json_object_new_string, , AC_MSG_ERROR([could not find the json library]))
 
-AC_CHECK_LIB(xslt, xsltApplyStylesheet)
-AC_CHECK_LIB(xml, htmlParseFile)
-AC_CHECK_LIB(exslt, exsltRegisterAll)
+AC_CHECK_LIB(xslt, xsltApplyStylesheet, , AC_MSG_ERROR([could not find libxslt]))
+AC_CHECK_LIB(xml2, htmlParseFile, , AC_MSG_ERROR([could not find libxml2 with html parsing]))
+AC_CHECK_LIB(exslt, exsltRegisterAll, , AC_MSG_ERROR([could not find libexslt]))
 
 AC_CONFIG_FILES([Makefile])
 AC_OUTPUT

dexter.c

@@ -80,6 +80,7 @@ dexPtr dex_compile(char* dex_str, char* incl) {
   if(!dex_exslt_registered) {
     exsltRegisterAll();
 		dex_register_all();
+		init_xpath_alias();
 		exslt_org_regular_expressions_init();
     dex_exslt_registered = true;
   }

functions.c

@@ -287,8 +287,7 @@ xsltLoadHtmlDocument(xsltTransformContextPtr ctxt, const xmlChar *URI) {
 	ret = ret->next;
     }
 
-    doc = xsltHtmlDocLoader(URI, ctxt->dict, ctxt->parserOptions,
-                               (void *) ctxt, XSLT_LOAD_DOCUMENT);
+    doc = htmlReadFile(URI, NULL, ctxt->parserOptions | HTML_PARSE_RECOVER);
 
     if (doc == NULL)
 	return(NULL);
@@ -316,71 +315,4 @@ xsltLoadHtmlDocument(xsltTransformContextPtr ctxt, const xmlChar *URI) {
 
     ret = xsltNewDocument(ctxt, doc);
     return(ret);
-}
-
-/**
- * xsltDocDefaultLoaderFunc:
- * @URI: the URI of the document to load
- * @dict: the dictionary to use when parsing that document
- * @options: parsing options, a set of xmlParserOption
- * @ctxt: the context, either a stylesheet or a transformation context
- * @type: the xsltLoadType indicating the kind of loading required
- *
- * Default function to load document not provided by the compilation or
- * transformation API themselve, for example when an xsl:import,
- * xsl:include is found at compilation time or when a document()
- * call is made at runtime.
- *
- * Returns the pointer to the document (which will be modified and
- * freed by the engine later), or NULL in case of error.
- */
-static xmlDocPtr
-xsltHtmlDocLoader(const xmlChar * URI, xmlDictPtr dict, int options,
-                         void *ctxt ATTRIBUTE_UNUSED,
-			 xsltLoadType type ATTRIBUTE_UNUSED)
-{
-    xmlParserCtxtPtr pctxt;
-    xmlParserInputPtr inputStream;
-    xmlDocPtr doc;
-
-
-
-    pctxt = htmlNewParserCtxt();
-    if (pctxt == NULL)
-        return(NULL);
-    if ((dict != NULL) && (pctxt->dict != NULL)) {
-        xmlDictFree(pctxt->dict);
-	pctxt->dict = NULL;
-    }
-    if (dict != NULL) {
-	pctxt->dict = dict;
-	xmlDictReference(pctxt->dict);
-#ifdef WITH_XSLT_DEBUG
-	xsltGenericDebug(xsltGenericDebugContext,
-                     "Reusing dictionary for document\n");
-#endif
-    }
-    // htmlCtxtUseOptions(pctxt, options | HTML_PARSE_RECOVER | HTML_PARSE_NOERROR );
-    inputStream = xmlLoadExternalEntity((const char *) URI, NULL, pctxt);
-    if (inputStream == NULL) {
-        xmlFreeParserCtxt(pctxt);
-	return(NULL);
-    }
-    inputPush(pctxt, inputStream);
-    if (pctxt->directory == NULL)
-        pctxt->directory = xmlParserGetDirectory((const char *) URI);
-
-		htmlDoRead(pctxt, NULL, NULL, options | HTML_PARSE_RECOVER, 1);
-
-    if (pctxt->wellFormed) {
-        doc = pctxt->myDoc;
-    }
-    else {
-        doc = NULL;
-        xmlFreeDoc(pctxt->myDoc);
-        pctxt->myDoc = NULL;
-    }
-    xmlFreeParserCtxt(pctxt);
-
-    return(doc);
-}
+}

functions.h

@@ -1,4 +1,8 @@
+#ifndef DEX_FUNCTIONS_H_INCLUDED
+#define DEX_FUNCTIONS_H_INCLUDED
+
 #include <libxml/xpath.h>
+#include <libxml/HTMLparser.h>
 #include <libxslt/xslt.h>
 #include <libxslt/xsltInternals.h>
 #include <libxslt/transform.h>
@@ -8,6 +12,6 @@ void dex_register_all();
 
 static void xsltHtmlDocumentFunction(xmlXPathParserContextPtr, int);
 static void xsltHtmlDocumentFunctionLoadDocument(xmlXPathParserContextPtr, xmlChar*);
-static xsltDocumentPtr	xsltLoadHtmlDocument(xsltTransformContextPtr, const xmlChar *);
-static xmlDocPtr xsltHtmlDocLoader(const xmlChar *, xmlDictPtr, int, void *, xsltLoadType type);
+xsltDocumentPtr	xsltLoadHtmlDocument(xsltTransformContextPtr, const xmlChar *);
 
+#endif

parser.y

@@ -4,6 +4,7 @@
 #include <stdlib.h>
 #include <string.h>
 #include "kstring.h"
+#include <libxml/hash.h>
 
 #ifndef PARSER_Y_H_INCLUDED
 #define PARSER_Y_H_INCLUDED
@@ -19,6 +20,11 @@ void prepare_parse(char*);
 void cleanup_parse(void);
 void start_debugging(void);
 
+static xmlHashTablePtr alias_hash;
+
+char* xpath_alias(char*);
+void init_xpath_alias();
+
 int yyparse(void);
 char* myparse(char*);
 void answer(char*);
@@ -223,7 +229,7 @@ PrimaryExpr
 	;
 
 FunctionCall
-  : FunctionName LPAREN Arguments RPAREN		{ $$ = astrcat4($1, $2, $3, $4); }
+  : FunctionName LPAREN Arguments RPAREN		{ $$ = astrcat4(xpath_alias($1), $2, $3, $4); }
 	| FunctionName LPAREN Arguments						{ yyerror("unclosed parenthesis"); }
 	| FunctionName LPAREN Arguments RPAREN RPAREN { yyerror("too many parenthesis"); }
 	;
@@ -548,6 +554,16 @@ OptS
 
 %%
 
+char* xpath_alias(char* key) {
+	char* value = (char*) xmlHashLookup(alias_hash, key);
+	return value == NULL ? key : value;
+}
+
+void init_xpath_alias() {
+	alias_hash = xmlHashCreate(100);
+	xmlHashAddEntry(alias_hash, "html", "dex:html-document");
+}
+
 char* myparse(char* string){
 	// start_debugging();
   prepare_parse(string);

test/remote.dex

@@ -0,0 +1,4 @@
+{
+	"title": "html('http://www.google.com/')//title"
+}
+

test/remote.html

@@ -0,0 +1,3 @@
+<html>
+	<body>Doesn't matter</body>
+</html>

test/remote.json

@@ -0,0 +1 @@
+{ "title": "Google" }

xml2json.c

@@ -1,6 +1,7 @@
 #include <libxml/parser.h>
 #include <json/json.h>
 #include "xml2json.h"
+#include <string.h>
 
 /**
  * Handles a simplified xml