X-Git-Url: https://code.citadel.org/?a=blobdiff_plain;f=webcit%2Fhtml2html.c;h=00a2adcbb9301260c93462cac4515b6ea11dd3b2;hb=efb97b068c75a55d164acee228ed1199b66ef08c;hp=6eb490a602faf891c2a2fbb1fa5584da65157768;hpb=5f6d11980b48033e589189f1e9ca597dee456a46;p=citadel.git

diff --git a/webcit/html2html.c b/webcit/html2html.c
index 6eb490a60..00a2adcbb 100644
--- a/webcit/html2html.c
+++ b/webcit/html2html.c
@@ -1,37 +1,76 @@
 /*
  * $Id$
- *
- * Output an HTML message, modifying it slightly to make sure it plays nice
+ */
+/**
+ * \defgroup HTML2HTML Output an HTML message, modifying it slightly to make sure it plays nice
  * with the rest of our web framework.
- *
+ * \ingroup WebcitHttpServer
  */
-
-#include <ctype.h>
-#include <stdlib.h>
-#include <unistd.h>
-#include <stdio.h>
-#include <fcntl.h>
-#include <signal.h>
-#include <sys/types.h>
-#include <sys/wait.h>
-#include <sys/socket.h>
-#include <limits.h>
-#include <netinet/in.h>
-#include <netdb.h>
-#include <string.h>
-#include <pwd.h>
-#include <errno.h>
-#include <stdarg.h>
-#include <pthread.h>
-#include <signal.h>
+/*@{*/
 #include "webcit.h"
 #include "vcard.h"
 #include "webserver.h"
 
 
-/*
+/**
+ * \brief	Strip surrounding single or double quotes from a string.
+ *
+ * \param s	String to be stripped.
+ */
+void stripquotes(char *s)
+{
+	int len;
+
+	if (!s) return;
+
+	len = strlen(s);
+	if (len < 2) return;
+
+	if ( ( (s[0] == '\"') && (s[len-1] == '\"') ) || ( (s[0] == '\'') && (s[len-1] == '\'') ) ) {
+		s[len-1] = 0;
+		strcpy(s, &s[1]);
+	}
+}
+
+
+/**
+ * \brief Check to see if a META tag has overridden the declared MIME character set.
+ *
+ * \param charset		Character set name (left unchanged if we don't do anything)
+ * \param meta_http_equiv	Content of the "http-equiv" portion of the META tag
+ * \param meta_content		Content of the "content" portion of the META tag
  */
-void output_html(void) {
+void extract_charset_from_meta(char *charset, char *meta_http_equiv, char *meta_content)
+{
+	char *ptr;
+	char buf[64];
+
+	if (!charset) return;
+	if (!meta_http_equiv) return;
+	if (!meta_content) return;
+
+
+	if (strcasecmp(meta_http_equiv, "Content-type")) return;
+
+	ptr = strchr(meta_content, ';');
+	if (!ptr) return;
+
+	safestrncpy(buf, ++ptr, sizeof buf);
+	striplt(buf);
+	if (!strncasecmp(buf, "charset=", 8)) {
+		strcpy(charset, &buf[8]);
+	}
+}
+
+
+
+/**
+ * \brief Sanitize and enhance an HTML message for display.
+ *        Also convert weird character sets to UTF-8 if necessary.
+ *
+ * \param supplied_charset the input charset as declared in the MIME headers
+ */
+void output_html(char *supplied_charset, int treat_as_wiki) {
 	char buf[SIZ];
 	char *msg;
 	char *ptr;
@@ -47,19 +86,30 @@ void output_html(void) {
 	int alevel = 0;
 	int i;
 	int linklen;
+	char charset[128];
+#ifdef HAVE_ICONV
+	iconv_t ic = (iconv_t)(-1) ;
+	char *ibuf;                   /**< Buffer of characters to be converted */
+	char *obuf;                   /**< Buffer for converted characters      */
+	size_t ibuflen;               /**< Length of input buffer               */
+	size_t obuflen;               /**< Length of output buffer              */
+	char *osav;                   /**< Saved pointer to output buffer       */
+#endif
 
+	safestrncpy(charset, supplied_charset, sizeof charset);
 	msg = strdup("");
-	sprintf(new_window, "<A TARGET=\"%s\" HREF=", TARGET);
+	sprintf(new_window, "<a target=\"%s\" href=", TARGET);
 
-	while (serv_gets(buf), strcmp(buf, "000")) {
+	while (serv_getln(buf, sizeof buf), strcmp(buf, "000")) {
 		line_length = strlen(buf);
 		buffer_length = content_length + line_length + 2;
 		msg = realloc(msg, buffer_length);
 		if (msg == NULL) {
-			wprintf("<B>realloc() error!  "
-				"couldn't get %d bytes: %s</B><BR><BR>\n",
+			wprintf("<b>");
+			wprintf(_("realloc() error! couldn't get %d bytes: %s"),
 				buffer_length + 1,
 				strerror(errno));
+			wprintf("</b><br /><br />\n");
 			return;
 		}
 		strcpy(&msg[content_length], buf);
@@ -68,19 +118,66 @@ void output_html(void) {
 		content_length += 1;
 	}
 
+	/** Do a first pass to isolate the message body */
 	ptr = msg;
 	msgstart = msg;
 	msgend = &msg[content_length];
 
 	while (ptr < msgend) {
 
-		/* Advance to next tag */
+		/** Advance to next tag */
 		ptr = strchr(ptr, '<');
 		if ((ptr == NULL) || (ptr >= msgend)) break;
 		++ptr;
 		if ((ptr == NULL) || (ptr >= msgend)) break;
 
-		/* Any of these tags cause everything up to and including
+		/**
+		 *  Look for META tags.  Some messages (particularly in
+		 *  Asian locales) illegally declare a message's character
+		 *  set in the HTML instead of in the MIME headers.  This
+		 *  is wrong but we have to work around it anyway.
+		 */
+		if (!strncasecmp(ptr, "META", 4)) {
+
+			char *meta_start;
+			char *meta_end;
+			int meta_length;
+			char *meta;
+			char *meta_http_equiv;
+			char *meta_content;
+			char *spaceptr;
+
+			meta_start = &ptr[4];
+			meta_end = strchr(ptr, '>');
+			if ((meta_end != NULL) && (meta_end <= msgend)) {
+				meta_length = meta_end - meta_start + 1;
+				meta = malloc(meta_length + 1);
+				safestrncpy(meta, meta_start, meta_length);
+				meta[meta_length] = 0;
+				striplt(meta);
+				if (!strncasecmp(meta, "HTTP-EQUIV=", 11)) {
+					meta_http_equiv = strdup(&meta[11]);
+					spaceptr = strchr(meta_http_equiv, ' ');
+					if (spaceptr != NULL) {
+						*spaceptr = 0;
+						meta_content = strdup(++spaceptr);
+						if (!strncasecmp(meta_content, "content=", 8)) {
+							strcpy(meta_content, &meta_content[8]);
+							stripquotes(meta_http_equiv);
+							stripquotes(meta_content);
+							extract_charset_from_meta(charset,
+								meta_http_equiv, meta_content);
+						}
+						free(meta_content);
+					}
+					free(meta_http_equiv);
+				}
+				free(meta);
+			}
+		}
+
+		/**
+		 * Any of these tags cause everything up to and including
 		 * the tag to be removed.
 		 */	
 		if ( (!strncasecmp(ptr, "HTML", 4))
@@ -94,7 +191,8 @@ void output_html(void) {
 			msgstart = ptr;
 		}
 
-		/* Any of these tags cause everything including and following
+		/**
+		 * Any of these tags cause everything including and following
 		 * the tag to be removed.
 		 */
 		if ( (!strncasecmp(ptr, "/HTML", 5))
@@ -107,42 +205,98 @@ void output_html(void) {
 
 		++ptr;
 	}
+	if (msgstart > msg) {
+		strcpy(msg, msgstart);
+	}
+
+	/** Convert foreign character sets to UTF-8 if necessary. */
+#ifdef HAVE_ICONV
+	if ( (strcasecmp(charset, "us-ascii"))
+	   && (strcasecmp(charset, "UTF-8"))
+	   && (strcasecmp(charset, ""))
+	) {
+		lprintf(9, "Converting %s to UTF-8\n", charset);
+		ic = iconv_open("UTF-8", charset);
+		if (ic == (iconv_t)(-1) ) {
+			lprintf(5, "%s:%d iconv_open() failed: %s\n",
+				__FILE__, __LINE__, strerror(errno));
+		}
+	}
+	if (ic != (iconv_t)(-1) ) {
+		ibuf = msg;
+		ibuflen = content_length;
+		obuflen = content_length + (content_length / 2) ;
+		obuf = (char *) malloc(obuflen);
+		osav = obuf;
+		iconv(ic, &ibuf, &ibuflen, &obuf, &obuflen);
+		content_length = content_length + (content_length / 2) - obuflen;
+		osav[content_length] = 0;
+		free(msg);
+		msg = osav;
+		iconv_close(ic);
+	}
+#endif
+
+	/** FIXME At this point, shigerugo's messages are still clean.
+	 *        Figure out what is mangling them below.
+	 */
 
+	/** Now go through the message, parsing tags as necessary. */
 	converted_msg = malloc(content_length);
 	strcpy(converted_msg, "");
 	ptr = msgstart;
 	while (ptr < msgend) {
-		/* Change mailto: links to WebCit mail, by replacing the
+		/**
+		 * Change mailto: links to WebCit mail, by replacing the
 		 * link with one that points back to our mail room.  Due to
 		 * the way we parse URL's, it'll even handle mailto: links
 		 * that have "?subject=" in them.
 		 */
-		if (!strncasecmp(ptr, "<A HREF=\"mailto:", 16)) {
+		if (!strncasecmp(ptr, "<a href=\"mailto:", 16)) {
 			content_length += 64;
 			converted_msg = realloc(converted_msg, content_length);
 			sprintf(&converted_msg[output_length],
-				"<A HREF=\"/display_enter"
+				"<a href=\"display_enter"
 				"?force_room=_MAIL_&recp=");
 			output_length += 47;
 			ptr = &ptr[16];
 			++alevel;
 		}
-		/* Make links open in a separate window */
-		else if (!strncasecmp(ptr, "<A HREF=", 8)) {
-			content_length += 64;
-			converted_msg = realloc(converted_msg, content_length);
-			sprintf(&converted_msg[output_length], new_window);
-			output_length += strlen(new_window);
-			ptr = &ptr[8];
+		/** Make external links open in a separate window */
+		else if (!strncasecmp(ptr, "<a href=", 8)) {
 			++alevel;
+			if ( ((strchr(ptr, ':') < strchr(ptr, '/')))
+			     &&  ((strchr(ptr, '/') < strchr(ptr, '>'))) 
+			     ) {
+				/* open external links to new window */
+				content_length += 64;
+				converted_msg = realloc(converted_msg, content_length);
+				sprintf(&converted_msg[output_length], new_window);
+				output_length += strlen(new_window);
+				ptr = &ptr[8];
+			}
+			else if ( (treat_as_wiki) && (strncasecmp(ptr, "<a href=\"wiki?", 14)) ) {
+				lprintf(9, "converting wiki link\n");
+				content_length += 64;
+				converted_msg = realloc(converted_msg, content_length);
+				sprintf(&converted_msg[output_length], "<a href=\"wiki?page=");
+				output_length += 19;
+				ptr = &ptr[9];
+			}
+			else {
+				sprintf(&converted_msg[output_length], "<a href=");
+				output_length += 8;
+				ptr = &ptr[8];
+			}
 		}
-		/* Turn anything that looks like a URL into a real link, as long
+		/**
+		 * Turn anything that looks like a URL into a real link, as long
 		 * as it's not inside a tag already
 		 */
 		else if ( (brak == 0) && (alevel == 0)
 		     && (!strncasecmp(ptr, "http://", 7))) {
 				linklen = 0;
-				/* Find the end of the link */
+				/** Find the end of the link */
 				for (i=0; i<=strlen(ptr); ++i) {
 					if ((ptr[i]==0)
 					   ||(isspace(ptr[i]))
@@ -179,7 +333,7 @@ void output_html(void) {
 				}
 		}
 		else {
-			/*
+			/**
 			 * We need to know when we're inside a tag,
 			 * so we don't turn things that look like URL's into
 			 * links, when they're already links - or image sources.
@@ -192,14 +346,15 @@ void output_html(void) {
 		}
 	}
 
-	/* Output our big pile of markup */
+	/** Output our big pile of markup */
 	client_write(converted_msg, output_length);
 
-	/* A little trailing vertical whitespace... */
-	wprintf("<BR><BR>\n");
+	/** A little trailing vertical whitespace... */
+	wprintf("<br /><br />\n");
 
-	/* Now give back the memory */
+	/** Now give back the memory */
 	free(converted_msg);
 	free(msg);
 }
 
+/*@}*/