When converting <A> tags, don't omit the change to

[citadel.git] / webcit / html2html.c

diff --git a/webcit/html2html.c b/webcit/html2html.c
index abd5904e846a1577fa6c2af0c5a2d3fa3e72eed6..72356c700bf539566f5baee2cd92b3c74e7372a9 100644 (file)
--- a/webcit/html2html.c
+++ b/webcit/html2html.c
@@ -59,6 +59,18 @@ void extract_charset_from_meta(char *charset, char *meta_http_equiv, char *meta_
        striplt(buf);
        if (!strncasecmp(buf, "charset=", 8)) {
                strcpy(charset, &buf[8]);
+
+               /*
+                * The brain-damaged webmail program in Microsoft Exchange declares
+                * a charset of "unicode" when they really mean "UTF-8".  GNU iconv
+                * treats "unicode" as an alias for "UTF-16" so we have to manually
+                * fix this here, otherwise messages generated in Exchange webmail
+                * show up as a big pile of weird characters.
+                */
+               if (!strcasecmp(charset, "unicode")) {
+                       strcpy(charset, "UTF-8");
+               }
+
        }
 }
 
@@ -255,8 +267,10 @@ void output_html(char *supplied_charset, int treat_as_wiki) {
        converted_alloc = content_length + 8192;
        converted_msg = malloc(converted_alloc);
        if (converted_msg == NULL) {
-               abort();                        /* FIXME */
+               wprintf("Error %d: %s<br />%s:%s", errno, strerror(errno), __FILE__, __LINE__);
+               goto BAIL;
        }
+
        strcpy(converted_msg, "");
        ptr = msg;
        msgend = strchr(msg, 0);
@@ -372,17 +386,17 @@ void output_html(char *supplied_charset, int treat_as_wiki) {
                                }
                }
                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.
-                        */
-                       if (*ptr == '<') ++brak;
-                       if (*ptr == '>') --brak;
-                       if (!strncasecmp(ptr, "</A>", 3)) --alevel;
                        converted_msg[output_length] = *ptr++;
                        converted_msg[++output_length] = 0;
                }
+               /**
+                * 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.
+                */
+               if (*ptr == '<') ++brak;
+               if (*ptr == '>') --brak;
+               if (!strncasecmp(ptr, "</A>", 3)) --alevel;
        }
 
        /**     uncomment these two lines to override conversion        */
@@ -392,7 +406,7 @@ void output_html(char *supplied_charset, int treat_as_wiki) {
        /** Output our big pile of markup */
        client_write(converted_msg, output_length);
 
-       /** A little trailing vertical whitespace... */
+BAIL:  /** A little trailing vertical whitespace... */
        wprintf("<br /><br />\n");
 
        /** Now give back the memory */