4 * Output an HTML message, modifying it slightly to make sure it plays nice
5 * with the rest of our web framework.
11 #include "webserver.h"
15 * Sanitize and enhance an HTML message for display.
16 * Also convert weird character sets to UTF-8 if necessary.
18 void output_html(char *charset) {
25 int buffer_length = 1;
27 int content_length = 0;
28 int output_length = 0;
35 iconv_t ic = (iconv_t)(-1) ;
36 char *ibuf; /* Buffer of characters to be converted */
37 char *obuf; /* Buffer for converted characters */
38 size_t ibuflen; /* Length of input buffer */
39 size_t obuflen; /* Length of output buffer */
40 char *osav; /* Saved pointer to output buffer */
44 sprintf(new_window, "<A TARGET=\"%s\" HREF=", TARGET);
46 while (serv_getln(buf, sizeof buf), strcmp(buf, "000")) {
47 line_length = strlen(buf);
48 buffer_length = content_length + line_length + 2;
49 msg = realloc(msg, buffer_length);
52 wprintf(_("realloc() error! couldn't get %d bytes: %s"),
55 wprintf("</B><br /><br />\n");
58 strcpy(&msg[content_length], buf);
59 content_length += line_length;
60 strcpy(&msg[content_length], "\n");
65 if ( (strcasecmp(charset, "us-ascii"))
66 && (strcasecmp(charset, "UTF-8")) ) {
67 ic = iconv_open("UTF-8", charset);
68 if (ic == (iconv_t)(-1) ) {
69 lprintf(5, "iconv_open() failed: %s\n", strerror(errno));
72 if (ic != (iconv_t)(-1) ) {
74 ibuflen = content_length;
75 obuflen = content_length + (content_length / 2) ;
76 obuf = (char *) malloc(obuflen);
78 iconv(ic, &ibuf, &ibuflen, &obuf, &obuflen);
79 content_length = content_length + (content_length / 2) - obuflen;
80 osav[content_length] = 0;
89 msgend = &msg[content_length];
91 while (ptr < msgend) {
93 /* Advance to next tag */
94 ptr = strchr(ptr, '<');
95 if ((ptr == NULL) || (ptr >= msgend)) break;
97 if ((ptr == NULL) || (ptr >= msgend)) break;
99 /* Any of these tags cause everything up to and including
100 * the tag to be removed.
102 if ( (!strncasecmp(ptr, "HTML", 4))
103 ||(!strncasecmp(ptr, "HEAD", 4))
104 ||(!strncasecmp(ptr, "/HEAD", 5))
105 ||(!strncasecmp(ptr, "BODY", 4)) ) {
106 ptr = strchr(ptr, '>');
107 if ((ptr == NULL) || (ptr >= msgend)) break;
109 if ((ptr == NULL) || (ptr >= msgend)) break;
113 /* Any of these tags cause everything including and following
114 * the tag to be removed.
116 if ( (!strncasecmp(ptr, "/HTML", 5))
117 ||(!strncasecmp(ptr, "/BODY", 5)) ) {
127 converted_msg = malloc(content_length);
128 strcpy(converted_msg, "");
130 while (ptr < msgend) {
131 /* Change mailto: links to WebCit mail, by replacing the
132 * link with one that points back to our mail room. Due to
133 * the way we parse URL's, it'll even handle mailto: links
134 * that have "?subject=" in them.
136 if (!strncasecmp(ptr, "<A HREF=\"mailto:", 16)) {
137 content_length += 64;
138 converted_msg = realloc(converted_msg, content_length);
139 sprintf(&converted_msg[output_length],
140 "<A HREF=\"/display_enter"
141 "?force_room=_MAIL_&recp=");
146 /* Make links open in a separate window */
147 else if (!strncasecmp(ptr, "<A HREF=", 8)) {
148 content_length += 64;
149 converted_msg = realloc(converted_msg, content_length);
150 sprintf(&converted_msg[output_length], new_window);
151 output_length += strlen(new_window);
155 /* Turn anything that looks like a URL into a real link, as long
156 * as it's not inside a tag already
158 else if ( (brak == 0) && (alevel == 0)
159 && (!strncasecmp(ptr, "http://", 7))) {
161 /* Find the end of the link */
162 for (i=0; i<=strlen(ptr); ++i) {
174 if (linklen > 0) break;
177 content_length += (32 + linklen);
178 converted_msg = realloc(converted_msg, content_length);
179 sprintf(&converted_msg[output_length], new_window);
180 output_length += strlen(new_window);
181 converted_msg[output_length] = '\"';
182 converted_msg[++output_length] = 0;
183 for (i=0; i<linklen; ++i) {
184 converted_msg[output_length] = ptr[i];
185 converted_msg[++output_length] = 0;
187 sprintf(&converted_msg[output_length], "\">");
189 for (i=0; i<linklen; ++i) {
190 converted_msg[output_length] = *ptr++;
191 converted_msg[++output_length] = 0;
193 sprintf(&converted_msg[output_length], "</A>");
199 * We need to know when we're inside a tag,
200 * so we don't turn things that look like URL's into
201 * links, when they're already links - or image sources.
203 if (*ptr == '<') ++brak;
204 if (*ptr == '>') --brak;
205 if (!strncasecmp(ptr, "</A>", 3)) --alevel;
206 converted_msg[output_length] = *ptr++;
207 converted_msg[++output_length] = 0;
211 /* Output our big pile of markup */
212 client_write(converted_msg, output_length);
214 /* A little trailing vertical whitespace... */
215 wprintf("<br /><br />\n");
217 /* Now give back the memory */