diff options
| author | Mikhail Galanin <mgalanin@mimecast.com> | 2018-09-18 10:34:37 +0100 |
|---|---|---|
| committer | Mikhail Galanin <mgalanin@mimecast.com> | 2018-09-18 10:34:37 +0100 |
| commit | 2d6d215df32c129b15502efdf5ec3c21c3d167ff (patch) | |
| tree | 8795a09e92950f6bc41271947ae1381e0b817bce /src/libmime | |
| parent | 5b571379f3a2493f53eabd331da42ddcdcce4921 (diff) | |
| download | rspamd-2d6d215df32c129b15502efdf5ec3c21c3d167ff.tar.gz rspamd-2d6d215df32c129b15502efdf5ec3c21c3d167ff.zip | |
[Minor] Insert spaces only if text part is not suspected to be a link
sa_body requires newlines to be replaces with spaces.
However, some email clients highlight URL in triangular braces (<>) even if it's
broken by a newline. Let's just remove \n if current text part potentially
can be such URL.
Diffstat (limited to 'src/libmime')
| -rw-r--r-- | src/libmime/message.c | 16 |
1 files changed, 14 insertions, 2 deletions
diff --git a/src/libmime/message.c b/src/libmime/message.c index f0a235a4c..39768e0e0 100644 --- a/src/libmime/message.c +++ b/src/libmime/message.c @@ -244,6 +244,8 @@ rspamd_strip_newlines_parse (const gchar *begin, const gchar *pe, const gchar *p = begin, *c = begin; gchar last_c = '\0'; gboolean crlf_added = FALSE; + gboolean url_open_bracket = FALSE; + enum { normal_char, seen_cr, @@ -285,6 +287,8 @@ rspamd_strip_newlines_parse (const gchar *begin, const gchar *pe, break; } + url_open_bracket = FALSE; + p ++; } else if (G_UNLIKELY (*p == '\n')) { @@ -300,7 +304,7 @@ rspamd_strip_newlines_parse (const gchar *begin, const gchar *pe, c = p + 1; - if (IS_PART_HTML (part) || g_ascii_ispunct (last_c)) { + if (IS_PART_HTML (part) || !url_open_bracket) { g_byte_array_append (part->utf_stripped_content, (const guint8 *)" ", 1); g_ptr_array_add (part->newlines, @@ -315,7 +319,7 @@ rspamd_strip_newlines_parse (const gchar *begin, const gchar *pe, case seen_cr: /* \r\n */ if (!crlf_added) { - if (IS_PART_HTML (part) || g_ascii_ispunct (last_c)) { + if (IS_PART_HTML (part) || !url_open_bracket) { g_byte_array_append (part->utf_stripped_content, (const guint8 *) " ", 1); crlf_added = TRUE; @@ -345,10 +349,18 @@ rspamd_strip_newlines_parse (const gchar *begin, const gchar *pe, c = p + 1; break; } + url_open_bracket = FALSE; p ++; } else { + if ((*p) == '<') { + url_open_bracket = TRUE; + } + else if ((*p) == '>') { + url_open_bracket = FALSE; + } + switch (state) { case normal_char: if (G_UNLIKELY (*p) == ' ') { |