[Feature] Enchance text_part:get_content method

This method now supports otional `type` attribute which could be
following:

- `content` (default): utf8 content with HTML tags stripped and newlines preserved
- `content_oneline`: utf8 content with HTML tags and newlines stripped
- `raw`: raw content, not mime decoded nor utf8 converted
- `raw_parsed`: raw content, mime decoded, not utf8 converted
- `raw_utf`: raw content, mime decoded, utf8 converted (but with HTML tags and newlines)

diff --git a/src/libmime/message.c b/src/libmime/message.c
index a4f3be5ca765b902c032bfe77cc7f6b9453eff8a..40769037b5c801deb57a29e2e4bf17cb3fcf4bbf 100644 (file)
--- a/src/libmime/message.c
+++ b/src/libmime/message.c
@@ -497,6 +497,7 @@ rspamd_message_process_text_part (struct rspamd_task *task,
                                &text_part->exceptions,
                                task->urls,
                                task->emails);
+               text_part->utf_raw_content = part_content;
 
                if (text_part->content->len == 0) {
                        text_part->flags |= RSPAMD_MIME_TEXT_PART_FLAG_EMPTY;
@@ -526,6 +527,7 @@ rspamd_message_process_text_part (struct rspamd_task *task,
 
                text_part->content = rspamd_mime_text_part_maybe_convert (task,
                                text_part);
+               text_part->utf_raw_content = text_part->content;
 
                if (text_part->content != NULL) {
                        /*

diff --git a/src/libmime/message.h b/src/libmime/message.h
index 8c0f919eab679b49d5bd31a3f9d2e88ae4d6779c..15fcfcccc1ffadcb801a6bfc61bda97ecdec8f89 100644 (file)
--- a/src/libmime/message.h
+++ b/src/libmime/message.h
@@ -82,6 +82,7 @@ struct rspamd_mime_text_part {
        rspamd_ftok_t raw;
        rspamd_ftok_t parsed;
        GByteArray *content;
+       GByteArray *utf_raw_content;
        GByteArray *stripped_content;
        GPtrArray *newlines;    /**< positions of newlines in text                                      */
        struct html_content *html;

diff --git a/src/lua/lua_mimepart.c b/src/lua/lua_mimepart.c
index 6d17c3a661e03c08c0a762dafc86cb37e89eaa96..1ff3dbd582064a2ffc3b5dbfeb18d2d43403e65d 100644 (file)
--- a/src/lua/lua_mimepart.c
+++ b/src/lua/lua_mimepart.c
@@ -46,8 +46,13 @@ end
  */
 LUA_FUNCTION_DEF (textpart, is_utf);
 /***
- * @method text_part:get_content()
- * Get the text of the part (html tags stripped)
+ * @method text_part:get_content([type])
+ * Get the text of the part (html tags stripped). Optional `type` defines type of content to get:
+ * - `content` (default): utf8 content with HTML tags stripped and newlines preserved
+ * - `content_oneline`: utf8 content with HTML tags and newlines stripped
+ * - `raw`: raw content, not mime decoded nor utf8 converted
+ * - `raw_parsed`: raw content, mime decoded, not utf8 converted
+ * - `raw_utf`: raw content, mime decoded, utf8 converted (but with HTML tags and newlines)
  * @return {text} `UTF8` encoded content of the part (zero-copy if not converted to a lua string)
  */
 LUA_FUNCTION_DEF (textpart, get_content);
@@ -354,16 +359,51 @@ lua_textpart_get_content (lua_State * L)
 {
        struct rspamd_mime_text_part *part = lua_check_textpart (L);
        struct rspamd_lua_text *t;
+       gsize len;
+       const gchar *start, *type = NULL;
 
        if (part == NULL || IS_PART_EMPTY (part)) {
                lua_pushnil (L);
                return 1;
        }
 
+       if (lua_type (L, 2) == LUA_TSTRING) {
+               type = lua_tostring (L, 2);
+       }
+
        t = lua_newuserdata (L, sizeof (*t));
        rspamd_lua_setclass (L, "rspamd{text}", -1);
-       t->start = part->content->data;
-       t->len = part->content->len;
+
+       if (!type) {
+               start = part->content->data;
+               len = part->content->len;
+       }
+       else if (strcmp (type, "content") == 0) {
+               start = part->content->data;
+               len = part->content->len;
+       }
+       else if (strcmp (type, "content_oneline") == 0) {
+               start = part->stripped_content->data;
+               len = part->stripped_content->len;
+       }
+       else if (strcmp (type, "raw_parsed") == 0) {
+               start = part->parsed.begin;
+               len = part->parsed.len;
+       }
+       else if (strcmp (type, "raw_utf") == 0) {
+               start = part->utf_raw_content->data;
+               len = part->utf_raw_content->len;
+       }
+       else if (strcmp (type, "raw") == 0) {
+               start = part->raw.begin;
+               len = part->raw.len;
+       }
+       else {
+               return luaL_error (L, "invalid content type: %s", type);
+       }
+
+       t->start = start;
+       t->len = len;
        t->flags = 0;
 
        return 1;