From: Vsevolod Stakhov Date: Fri, 12 Mar 2021 21:13:04 +0000 (+0000) Subject: [Minor] Treat absent Content-Type more wisely X-Git-Tag: 3.0~588 X-Git-Url: https://source.dussan.org/?a=commitdiff_plain;h=3377d26289d64f828169c6dfbf4c71b3f5cdea09;p=rspamd.git [Minor] Treat absent Content-Type more wisely --- diff --git a/src/libmime/mime_parser.c b/src/libmime/mime_parser.c index 0d6659867..a89b76b3a 100644 --- a/src/libmime/mime_parser.c +++ b/src/libmime/mime_parser.c @@ -25,6 +25,7 @@ #include "contrib/uthash/utlist.h" #include #include +#include "contrib/fastutf8/fastutf8.h" struct rspamd_mime_parser_lib_ctx { struct rspamd_multipattern *mp_boundary; @@ -610,8 +611,20 @@ rspamd_mime_parse_normal_part (struct rspamd_task *task, * In theory, it is very unsafe to process it as a text part * as we unlikely get some sane result */ - part->ct->flags &= ~RSPAMD_CONTENT_TYPE_TEXT; - part->ct->flags |= RSPAMD_CONTENT_TYPE_BROKEN; + + /* + * On the other hand, there is an evidence that some + * emails actually rely on that. + * So we apply an expensive hack here: + * if there are no 8bit characters -OR- the content is valid + * UTF8, we can still imply Content-Type == text/plain + */ + + if (rspamd_str_has_8bit (part->raw_data.begin, part->raw_data.len) && + !rspamd_fast_utf8_validate (part->raw_data.begin, part->raw_data.len)) { + part->ct->flags &= ~RSPAMD_CONTENT_TYPE_TEXT; + part->ct->flags |= RSPAMD_CONTENT_TYPE_BROKEN; + } } }