2012-05-16 [paul] 3.8.0cvs43
[claws.git] / src / matcher.c
index 844ff32734834e55556dc6a6e93240b4eb215900..734a825033ff1dabaa500edd95dfd078520d2c6a 100644 (file)
@@ -1,6 +1,6 @@
 /*
  * Sylpheed -- a GTK+ based, lightweight, and fast e-mail client
- * Copyright (C) 2002-2004 by the Claws Mail Team and Hiroyuki Yamamoto
+ * Copyright (C) 2002-2011 by the Claws Mail Team and Hiroyuki Yamamoto
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -85,6 +85,10 @@ static const MatchParser matchparser_tab[] = {
        {MATCHCRITERIA_NOT_WATCH_THREAD, "~watch_thread"},
        {MATCHCRITERIA_SPAM, "spam"},
        {MATCHCRITERIA_NOT_SPAM, "~spam"},
+       {MATCHCRITERIA_HAS_ATTACHMENT, "has_attachment"},
+       {MATCHCRITERIA_HAS_NO_ATTACHMENT, "~has_attachment"},
+       {MATCHCRITERIA_SIGNED, "signed"},
+       {MATCHCRITERIA_NOT_SIGNED, "~signed"},
 
        /* msginfo headers */
        {MATCHCRITERIA_SUBJECT, "subject"},
@@ -351,13 +355,13 @@ static gboolean match_with_addresses_in_addressbook
        gboolean found = FALSE;
        gchar *path = NULL;
 
-       g_return_val_if_fail(address_list != NULL, FALSE);
+       cm_return_val_if_fail(address_list != NULL, FALSE);
 
        debug_print("match_with_addresses_in_addressbook(%d, %s)\n",
                                g_slist_length(address_list), folderpath?folderpath:"(null)");
 
        if (folderpath == NULL ||
-               strcasecmp(folderpath, _("Any")) == 0 ||
+               strcasecmp(folderpath, "Any") == 0 ||
                *folderpath == '\0')
                path = NULL;
        else
@@ -463,6 +467,7 @@ static gboolean matcherprop_string_match(MatcherProp *prop, const gchar *str,
        switch (prop->matchtype) {
        case MATCHTYPE_REGEXPCASE:
        case MATCHTYPE_REGEXP:
+#ifndef G_OS_WIN32
                if (!prop->preg && (prop->error == 0)) {
                        prop->preg = g_new0(regex_t, 1);
                        /* if regexp then don't use the escaped string */
@@ -505,7 +510,7 @@ static gboolean matcherprop_string_match(MatcherProp *prop, const gchar *str,
                        g_free(stripped);
                }
                break;
-                       
+#endif                 
        case MATCHTYPE_MATCHCASE:
        case MATCHTYPE_MATCH:
                ret = (strstr(str1, down_expr) != NULL);
@@ -594,48 +599,19 @@ const gchar *debug_context)
        return FALSE;
 }
 
-/* FIXME body search is a hack. */
-static gboolean matcherprop_string_decode_match(MatcherProp *prop, const gchar *str,
-                                                                                               const gchar *debug_context)
+static gboolean matcherprop_header_line_match(MatcherProp *prop, const gchar *hdr,
+                                             const gchar *str, const gchar *debug_context)
 {
-       gchar *utf = NULL;
-       gchar tmp[BUFFSIZE];
+       gchar *line = NULL;
        gboolean res = FALSE;
 
-       if (str == NULL)
+       if (hdr == NULL || str == NULL)
                return FALSE;
 
-       /* we try to decode QP first, because it's faster than base64 */
-       qp_decode_const(tmp, BUFFSIZE-1, str);
-       if (!g_utf8_validate(tmp, -1, NULL)) {
-               utf = conv_codeset_strdup
-                       (tmp, conv_get_locale_charset_str_no_utf8(),
-                        CS_INTERNAL);
-               res = matcherprop_string_match(prop, utf, debug_context);
-               g_free(utf);
-       } else {
-               res = matcherprop_string_match(prop, tmp, debug_context);
-       }
-       
-       if (res == FALSE && (strchr(prop->expr, '=') || strchr(prop->expr, '_')
-                           || strchr(str, '=') || strchr(str, '_'))) {
-               /* if searching for something with an equal char, maybe 
-                * we should try to match the non-decoded string. 
-                * In case it was not qp-encoded. */
-               if (!g_utf8_validate(str, -1, NULL)) {
-                       utf = conv_codeset_strdup
-                               (str, conv_get_locale_charset_str_no_utf8(),
-                                CS_INTERNAL);
-                       res = matcherprop_string_match(prop, utf, debug_context);
-                       g_free(utf);
-               } else {
-                       res = matcherprop_string_match(prop, str, debug_context);
-               }
-       }
-
-       /* FIXME base64 decoding is too slow, especially since text can 
-        * easily be handled as base64. Don't even try now. */
-
+       line = g_strdup_printf("%s %s", hdr, str);
+       res = matcherprop_string_match(prop, line, debug_context);
+       g_free(line);
+       
        return res;
 }
 
@@ -684,13 +660,21 @@ static gboolean matcherprop_match_test(const MatcherProp *prop,
 #endif
 
        file = procmsg_get_message_file(info);
-       if (file == NULL)
+       if (file == NULL) {
+#ifdef USE_PTHREAD
+               g_free(td);
+#endif
                return FALSE;
+       }
        g_free(file);           
 
        cmd = matching_build_command(prop->expr, info);
-       if (cmd == NULL)
+       if (cmd == NULL) {
+#ifdef USE_PTHREAD
+               g_free(td);
+#endif 
                return FALSE;
+}
 
 #ifdef USE_PTHREAD
        /* debug output */
@@ -797,6 +781,14 @@ gboolean matcherprop_match(MatcherProp *prop,
                return MSG_IS_SPAM(info->flags);
        case MATCHCRITERIA_NOT_SPAM:
                return !MSG_IS_SPAM(info->flags);
+       case MATCHCRITERIA_HAS_ATTACHMENT:
+               return MSG_IS_WITH_ATTACHMENT(info->flags);
+       case MATCHCRITERIA_HAS_NO_ATTACHMENT:
+               return !MSG_IS_WITH_ATTACHMENT(info->flags);
+       case MATCHCRITERIA_SIGNED:
+               return MSG_IS_SIGNED(info->flags);
+       case MATCHCRITERIA_NOT_SIGNED:
+               return !MSG_IS_SIGNED(info->flags);
        case MATCHCRITERIA_COLORLABEL:
        {
                gint color = MSG_GET_COLORLABEL_VALUE(info->flags);
@@ -933,7 +925,7 @@ gboolean matcherprop_match(MatcherProp *prop,
 
                t = time(NULL);
                age = ((t - info->date_t) / (60 * 60 * 24));
-               ret = (age > prop->value);
+               ret = (age >= prop->value);
 
                /* debug output */
                if (debug_filtering_session
@@ -1033,76 +1025,76 @@ gboolean matcherprop_match(MatcherProp *prop,
        }
        case MATCHCRITERIA_SIZE_GREATER:
        {
-               /* FIXME: info->size is an off_t */
-               gboolean ret = (info->size > (off_t) prop->value);
+               /* FIXME: info->size is a goffset */
+               gboolean ret = (info->size > (goffset) prop->value);
 
                /* debug output */
                if (debug_filtering_session
                                && prefs_common.filtering_debug_level >= FILTERING_DEBUG_LEVEL_HIGH) {
                        if (ret) {
                                log_print(LOG_DEBUG_FILTERING,
-                                               "message size [ %ld ] is greater than [ %d ]\n",
-                                               info->size, prop->value);
+                                               "message size is greater than [ %d ]\n",
+                                               prop->value);
                        } else {
                                log_print(LOG_DEBUG_FILTERING,
-                                               "message size [ %ld ] is not greater than [ %d ]\n",
-                                               info->size, prop->value);
+                                               "message size is not greater than [ %d ]\n",
+                                               prop->value);
                        }
                }
                return ret;
        }
        case MATCHCRITERIA_SIZE_SMALLER:
        {
-               /* FIXME: info->size is an off_t */
-               gboolean ret = (info->size < (off_t) prop->value);
+               /* FIXME: info->size is a goffset */
+               gboolean ret = (info->size < (goffset) prop->value);
 
                /* debug output */
                if (debug_filtering_session
                                && prefs_common.filtering_debug_level >= FILTERING_DEBUG_LEVEL_HIGH) {
                        if (ret) {
                                log_print(LOG_DEBUG_FILTERING,
-                                               "message size [ %ld ] is smaller than [ %d ]\n",
-                                               info->size, prop->value);
+                                               "message size is smaller than [ %d ]\n",
+                                               prop->value);
                        } else {
                                log_print(LOG_DEBUG_FILTERING,
-                                               "message size [ %ld ] is not smaller than [ %d ]\n",
-                                               info->size, prop->value);
+                                               "message size is not smaller than [ %d ]\n",
+                                               prop->value);
                        }
                }
                return ret;
        }
        case MATCHCRITERIA_SIZE_EQUAL:
        {
-               /* FIXME: info->size is an off_t */
-               gboolean ret = (info->size == (off_t) prop->value);
+               /* FIXME: info->size is a goffset */
+               gboolean ret = (info->size == (goffset) prop->value);
 
                /* debug output */
                if (debug_filtering_session
                                && prefs_common.filtering_debug_level >= FILTERING_DEBUG_LEVEL_HIGH) {
                        if (ret) {
                                log_print(LOG_DEBUG_FILTERING,
-                                               "message size [ %ld ] is equal to [ %d ]\n",
-                                               info->size, prop->value);
+                                               "message size is equal to [ %d ]\n",
+                                               prop->value);
                        } else {
                                log_print(LOG_DEBUG_FILTERING,
-                                               "message size [ %ld ] is not equal to [ %d ]\n",
-                                               info->size, prop->value);
+                                               "message size is not equal to [ %d ]\n",
+                                               prop->value);
                        }
                }
                return ret;
        }
        case MATCHCRITERIA_PARTIAL:
        {
-               /* FIXME: info->size is an off_t */
-               gboolean ret = (info->total_size != 0 && info->size != (off_t)info->total_size);
+               /* FIXME: info->size is a goffset */
+               gboolean ret = (info->total_size != 0 && info->size != (goffset)info->total_size);
 
                /* debug output */
                if (debug_filtering_session
                                && prefs_common.filtering_debug_level >= FILTERING_DEBUG_LEVEL_HIGH) {
                        if (ret) {
                                log_print(LOG_DEBUG_FILTERING,
-                                               "message is partially downloaded, size [ %ld ] is less than total size [ %d ])\n",
-                                               info->size, info->total_size);
+                                               "message is partially downloaded, size is less than total size [ %d ])\n",
+                                               info->total_size);
                        } else {
                                log_print(LOG_DEBUG_FILTERING,
                                                "message is not partially downloaded\n");
@@ -1112,8 +1104,8 @@ gboolean matcherprop_match(MatcherProp *prop,
        }
        case MATCHCRITERIA_NOT_PARTIAL:
        {
-               /* FIXME: info->size is an off_t */
-               gboolean ret = (info->total_size == 0 || info->size == (off_t)info->total_size);
+               /* FIXME: info->size is a goffset */
+               gboolean ret = (info->total_size == 0 || info->size == (goffset)info->total_size);
 
                /* debug output */
                if (debug_filtering_session
@@ -1123,8 +1115,8 @@ gboolean matcherprop_match(MatcherProp *prop,
                                                "message is not partially downloaded\n");
                        } else {
                                log_print(LOG_DEBUG_FILTERING,
-                                               "message is partially downloaded, size [ %ld ] is less than total size [ %d ])\n",
-                                               info->size, info->total_size);
+                                               "message is partially downloaded, size is less than total size [ %d ])\n",
+                                               info->total_size);
                        }
                }
                return ret;
@@ -1205,10 +1197,11 @@ void matcherlist_free(MatcherList *cond)
 {
        GSList *l;
 
-       g_return_if_fail(cond);
+       cm_return_if_fail(cond);
        for (l = cond->matchers ; l != NULL ; l = g_slist_next(l)) {
                matcherprop_free((MatcherProp *) l->data);
        }
+       g_slist_free(cond->matchers);
        g_free(cond);
 }
 
@@ -1259,13 +1252,23 @@ static gboolean matcherprop_match_one_header(MatcherProp *matcher,
                }
                break;
        case MATCHCRITERIA_HEADERS_PART:
-               return matcherprop_string_match(matcher, buf, _("header line"));
-       case MATCHCRITERIA_NOT_HEADERS_PART:
-               return !matcherprop_string_match(matcher, buf, _("headers line"));
        case MATCHCRITERIA_MESSAGE:
-               return matcherprop_string_decode_match(matcher, buf, _("message line"));
+               header = procheader_parse_header(buf);
+               if (!header)
+                       return FALSE;
+               result = matcherprop_header_line_match(matcher, 
+                              header->name, header->body, _("header line"));
+               procheader_header_free(header);
+               return result;
+       case MATCHCRITERIA_NOT_HEADERS_PART:
        case MATCHCRITERIA_NOT_MESSAGE:
-               return !matcherprop_string_decode_match(matcher, buf, _("message line"));
+               header = procheader_parse_header(buf);
+               if (!header)
+                       return FALSE;
+               result = !matcherprop_header_line_match(matcher, 
+                              header->name, header->body, _("header line"));
+               procheader_header_free(header);
+               return result;
        case MATCHCRITERIA_FOUND_IN_ADDRESSBOOK:
        case MATCHCRITERIA_NOT_FOUND_IN_ADDRESSBOOK:
                {
@@ -1274,9 +1277,9 @@ static gboolean matcherprop_match_one_header(MatcherProp *matcher,
                        gboolean found = FALSE;
 
                        /* how many address headers are me trying to mach? */
-                       if (strcasecmp(matcher->header, _("Any")) == 0)
+                       if (strcasecmp(matcher->header, "Any") == 0)
                                match = MATCH_ANY;
-                       else if (strcasecmp(matcher->header, Q_("Filtering Matcher Menu|All")) == 0)
+                       else if (strcasecmp(matcher->header, "All") == 0)
                                        match = MATCH_ALL;
 
                        if (match == MATCH_ONE) {
@@ -1407,9 +1410,9 @@ static gboolean matcherlist_match_headers(MatcherList *matchers, FILE *fp)
                                         procheader_headername_equal(header->name, "Reply-To") ||
                                         procheader_headername_equal(header->name, "Sender"))) {
 
-                                       if (strcasecmp(matcher->header, _("Any")) == 0)
+                                       if (strcasecmp(matcher->header, "Any") == 0)
                                                match = MATCH_ANY;
-                                       else if (strcasecmp(matcher->header, Q_("Filtering Matcher Menu|All")) == 0)
+                                       else if (strcasecmp(matcher->header, "All") == 0)
                                                match = MATCH_ALL;
                                        else
                                                match = MATCH_ONE;
@@ -1469,28 +1472,6 @@ static gboolean matcherprop_criteria_body(const MatcherProp *matcher)
        }
 }
 
-/*!
- *\brief       Check if a (line) string matches the criteria
- *             described by a matcher structure
- *
- *\param       matcher Matcher structure
- *\param       line String
- *
- *\return      gboolean TRUE if string matches criteria
- */
-static gboolean matcherprop_match_line(MatcherProp *matcher, const gchar *line)
-{
-       switch (matcher->criteria) {
-       case MATCHCRITERIA_BODY_PART:
-       case MATCHCRITERIA_MESSAGE:
-               return matcherprop_string_decode_match(matcher, line, _("body line"));
-       case MATCHCRITERIA_NOT_BODY_PART:
-       case MATCHCRITERIA_NOT_MESSAGE:
-               return !matcherprop_string_decode_match(matcher, line, _("body line"));
-       }
-       return FALSE;
-}
-
 /*!
  *\brief       Check if a line in a message file's body matches
  *             the criteria
@@ -1500,45 +1481,94 @@ static gboolean matcherprop_match_line(MatcherProp *matcher, const gchar *line)
  *
  *\return      gboolean TRUE if succesful match
  */
-static gboolean matcherlist_match_body(MatcherList *matchers, FILE *fp)
+static gboolean matcherlist_match_body(MatcherList *matchers, gboolean body_only, MsgInfo *info)
 {
        GSList *l;
+       MimeInfo *mimeinfo = NULL;
+       MimeInfo *partinfo = NULL;
        gchar buf[BUFFSIZE];
-       
-       while (fgets(buf, sizeof(buf), fp) != NULL) {
-               for (l = matchers->matchers ; l != NULL ; l = g_slist_next(l)) {
-                       MatcherProp *matcher = (MatcherProp *) l->data;
-                       
-                       if (matcher->done) 
-                               continue;
+       gboolean first_text_found = FALSE;
+       FILE *outfp = NULL;
 
-                       /* if the criteria is ~body_part or ~message, ZERO lines
-                        * must NOT match for the rule to match. */
-                       if (matcher->criteria == MATCHCRITERIA_NOT_BODY_PART ||
-                           matcher->criteria == MATCHCRITERIA_NOT_MESSAGE) {
-                               if (matcherprop_match_line(matcher, buf)) {
-                                       matcher->result = TRUE;
-                               } else {
-                                       matcher->result = FALSE;
-                                       matcher->done = TRUE;
-                               }
-                       /* else, just one line has to match */
-                       } else if (matcherprop_criteria_body(matcher) ||
-                                  matcherprop_criteria_message(matcher)) {
-                               if (matcherprop_match_line(matcher, buf)) {
-                                       matcher->result = TRUE;
-                                       matcher->done = TRUE;
+       cm_return_val_if_fail(info != NULL, FALSE);
+
+       mimeinfo = procmime_scan_message(info);
+
+       /* Skip headers */
+       partinfo = procmime_mimeinfo_next(mimeinfo);
+
+       for (; partinfo != NULL; partinfo = procmime_mimeinfo_next(partinfo)) {
+
+               if (partinfo->type != MIMETYPE_TEXT && body_only)
+                       continue;
+
+               if (partinfo->type == MIMETYPE_TEXT) {
+                       first_text_found = TRUE;
+                       outfp = procmime_get_text_content(partinfo);
+               } else
+                       outfp = procmime_get_binary_content(partinfo);
+
+               if (!outfp) {
+                       procmime_mimeinfo_free_all(mimeinfo);
+                       return FALSE;
+               }
+
+               while (fgets(buf, sizeof(buf), outfp) != NULL) {
+                       strretchomp(buf);
+
+                       for (l = matchers->matchers ; l != NULL ; l = g_slist_next(l)) {
+                               MatcherProp *matcher = (MatcherProp *) l->data;
+
+                               if (matcher->done) 
+                                       continue;
+
+                               /* Don't scan non-text parts when looking in body, only
+                                * when looking in whole message
+                                */
+                               if (partinfo && partinfo->type != MIMETYPE_TEXT &&
+                               (matcher->criteria == MATCHCRITERIA_NOT_BODY_PART ||
+                               matcher->criteria == MATCHCRITERIA_BODY_PART))
+                                       continue;
+
+                               /* if the criteria is ~body_part or ~message, ZERO lines
+                                * must match for the rule to match.
+                                */
+                               if (matcher->criteria == MATCHCRITERIA_NOT_BODY_PART ||
+                                   matcher->criteria == MATCHCRITERIA_NOT_MESSAGE) {
+                                       if (matcherprop_string_match(matcher, buf, 
+                                                               _("body line"))) {
+                                               matcher->result = FALSE;
+                                               matcher->done = TRUE;
+                                       } else
+                                               matcher->result = TRUE;
+                               /* else, just one line has to match */
+                               } else if (matcherprop_criteria_body(matcher) ||
+                                          matcherprop_criteria_message(matcher)) {
+                                       if (matcherprop_string_match(matcher, buf,
+                                                               _("body line"))) {
+                                               matcher->result = TRUE;
+                                               matcher->done = TRUE;
+                                       }
                                }
-                       }
 
-                       /* if the matchers are OR'ed and the rule matched,
-                        * no need to check the others. */
-                       if (matcher->result && matcher->done) {
-                               if (!matchers->bool_and)
-                                       return TRUE;
+                               /* if the matchers are OR'ed and the rule matched,
+                                * no need to check the others. */
+                               if (matcher->result && matcher->done) {
+                                       if (!matchers->bool_and) {
+                                               procmime_mimeinfo_free_all(mimeinfo);
+                                               fclose(outfp);
+                                               return TRUE;
+                                       }
+                               }
                        }
                }
+               fclose(outfp);
+
+               if (body_only && first_text_found)
+                       break;
        }
+       procmime_mimeinfo_free_all(mimeinfo);
+
        return FALSE;
 }
 
@@ -1556,6 +1586,7 @@ static gboolean matcherlist_match_file(MatcherList *matchers, MsgInfo *info,
 {
        gboolean read_headers;
        gboolean read_body;
+       gboolean body_only;
        GSList *l;
        FILE *fp;
        gchar *file;
@@ -1564,6 +1595,7 @@ static gboolean matcherlist_match_file(MatcherList *matchers, MsgInfo *info,
 
        read_headers = FALSE;
        read_body = FALSE;
+       body_only = TRUE;
        for (l = matchers->matchers ; l != NULL ; l = g_slist_next(l)) {
                MatcherProp *matcher = (MatcherProp *) l->data;
 
@@ -1574,6 +1606,7 @@ static gboolean matcherlist_match_file(MatcherList *matchers, MsgInfo *info,
                if (matcherprop_criteria_message(matcher)) {
                        read_headers = TRUE;
                        read_body = TRUE;
+                       body_only = FALSE;
                }
                matcher->result = FALSE;
                matcher->done = FALSE;
@@ -1603,7 +1636,7 @@ static gboolean matcherlist_match_file(MatcherList *matchers, MsgInfo *info,
 
        /* read the body */
        if (read_body) {
-               matcherlist_match_body(matchers, fp);
+               matcherlist_match_body(matchers, body_only, info);
        }
        
        for (l = matchers->matchers; l != NULL; l = g_slist_next(l)) {
@@ -1684,6 +1717,10 @@ gboolean matcherlist_match(MatcherList *matchers, MsgInfo *info)
                case MATCHCRITERIA_NOT_LOCKED:
                case MATCHCRITERIA_SPAM:
                case MATCHCRITERIA_NOT_SPAM:
+               case MATCHCRITERIA_HAS_ATTACHMENT:
+               case MATCHCRITERIA_HAS_NO_ATTACHMENT:
+               case MATCHCRITERIA_SIGNED:
+               case MATCHCRITERIA_NOT_SIGNED:
                case MATCHCRITERIA_COLORLABEL:
                case MATCHCRITERIA_NOT_COLORLABEL:
                case MATCHCRITERIA_IGNORE_THREAD:
@@ -1879,6 +1916,10 @@ gchar *matcherprop_to_string(MatcherProp *matcher)
        case MATCHCRITERIA_NOT_LOCKED:
        case MATCHCRITERIA_SPAM:
        case MATCHCRITERIA_NOT_SPAM:
+       case MATCHCRITERIA_HAS_ATTACHMENT:
+       case MATCHCRITERIA_HAS_NO_ATTACHMENT:
+       case MATCHCRITERIA_SIGNED:
+       case MATCHCRITERIA_NOT_SIGNED:
        case MATCHCRITERIA_PARTIAL:
        case MATCHCRITERIA_NOT_PARTIAL:
        case MATCHCRITERIA_IGNORE_THREAD:
@@ -2001,9 +2042,9 @@ static void add_str_default(gchar ** dest,
 
 /* matching_build_command() - preferably cmd should be unescaped */
 /*!
- *\brief       Build the command line to execute
+ *\brief       Build the command-line to execute
  *
- *\param       cmd String with command line specifiers
+ *\param       cmd String with command-line specifiers
  *\param       info Message info to use for command
  *
  *\return      gchar * Newly allocated string
@@ -2370,8 +2411,17 @@ static void matcher_add_rulenames(const gchar *rcpath)
        FILE *dst = g_fopen(newpath, "wb");
        gchar buf[BUFFSIZE];
        int r;
+       if (src == NULL) {
+               perror("fopen");
+               if (dst)
+                       fclose(dst);
+               g_free(newpath);
+               return;
+       }
        if (dst == NULL) {
                perror("fopen");
+               if (src)
+                       fclose(src);
                g_free(newpath);
                return;
        }