From 4b1c580bb9bc03d656e59d702c72c3e793a1bbe0 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?louiz=E2=80=99?= <louiz@louiz.org>
Date: Wed, 15 Jun 2016 12:19:19 +0200
Subject: cut messages at 512 bytes, taking into account the UTF-8 codepoints

ref #3067
---
 louloulibs/utils/string.cpp | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

(limited to 'louloulibs/utils')

diff --git a/louloulibs/utils/string.cpp b/louloulibs/utils/string.cpp
index 7ed1aa3..2447f84 100644
--- a/louloulibs/utils/string.cpp
+++ b/louloulibs/utils/string.cpp
@@ -1,4 +1,5 @@
 #include <utils/string.hpp>
+#include <utils/encoding.hpp>
 
 bool to_bool(const std::string& val)
 {
@@ -11,8 +12,17 @@ std::vector<std::string> cut(const std::string& val, const std::size_t size)
   std::string::size_type pos = 0;
   while (pos < val.size())
     {
-      res.emplace_back(val.substr(pos, size));
-      pos += size;
+      // Get the number of chars, <= size, that contain only whole
+      // UTF-8 codepoints.
+      std::size_t s = 0;
+      auto codepoint_size = utils::get_next_codepoint_size(val[pos + s]);
+      while (s + codepoint_size <= size)
+        {
+          s += codepoint_size;
+          codepoint_size = utils::get_next_codepoint_size(val[pos + s]);
+        }
+      res.emplace_back(val.substr(pos, s));
+      pos += s;
     }
   return res;
 }
-- 
cgit v1.2.3