commit 2249d42: [Fix] Adjust length of the fuzzy checks for short text parts

[Vsevolod Stakhov]

Author: Vsevolod Stakhov
Date: 2022-08-02 21:12:38 +0100
URL: https://github.com/rspamd/rspamd/commit/2249d428bd8f27d8e952233c4a15a75f69b12316

[Fix] Adjust length of the fuzzy checks for short text parts
Reported by: @citrin

---
 lualib/lua_fuzzy.lua   | 6 +++++-
 src/lua/lua_mimepart.c | 6 +-----
 2 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/lualib/lua_fuzzy.lua b/lualib/lua_fuzzy.lua
index 1f82752f9..ce4ac55ee 100644
--- a/lualib/lua_fuzzy.lua
+++ b/lualib/lua_fuzzy.lua
@@ -126,7 +126,11 @@ local function check_length(task, part, rule)
     local adjusted_bytes = bytes
 
     if part:is_text() then
-      bytes = part:get_text():get_length()
+      -- Fuzzy plugin uses stripped utf content to get an exact hash, that
+      -- corresponds to `get_content_oneline()`
+      -- However, in the case of empty parts this method returns `nil`, so extra
+      -- sanity check is required.
+      bytes = #(part:get_text():get_content_oneline() or '')
       if rule.text_multiplier then
         adjusted_bytes = bytes * rule.text_multiplier
       end
diff --git a/src/lua/lua_mimepart.c b/src/lua/lua_mimepart.c
index ae959c34e..459a78e69 100644
--- a/src/lua/lua_mimepart.c
+++ b/src/lua/lua_mimepart.c
@@ -803,11 +803,7 @@ lua_textpart_get_content_oneline (lua_State * L)
 		return 1;
 	}
 
-	t = lua_newuserdata (L, sizeof (*t));
-	rspamd_lua_setclass (L, "rspamd{text}", -1);
-	t->start = part->utf_stripped_content->data;
-	t->len = part->utf_stripped_content->len;
-	t->flags = 0;
+	lua_new_text(L, part->utf_stripped_content->data, part->utf_stripped_content->len, FALSE);
 
 	return 1;
 }