lib: Add str_truncate_utf8().

It works similar to str_truncate(), but it makes sure the truncated string
remains valid UTF-8.

diff --git a/src/lib/str.c b/src/lib/str.c
index 0f6d3b84888de3c58a9d9d539f67328627dfab59..2ec597023db833a14eb9948bc5885b4ab024f03c 100644 (file)
--- a/src/lib/str.c
+++ b/src/lib/str.c
@@ -3,6 +3,7 @@
 #include "lib.h"
 #include "buffer.h"
 #include "printf-format-fix.h"
+#include "unichar.h"
 #include "str.h"
 
 #include <stdio.h>
@@ -146,3 +147,12 @@ void str_vprintfa(string_t *str, const char *fmt, va_list args)
        /* drop the unused data, including terminating NUL */
        buffer_set_used_size(str, pos + ret);
 }
+
+void str_truncate_utf8(string_t *str, size_t len)
+{
+       size_t size = str_len(str);
+
+       if (size <= len)
+               return;
+       str_truncate(str, uni_utf8_data_truncate(str_data(str), size, len));
+}

diff --git a/src/lib/str.h b/src/lib/str.h
index 16ccc9314c5aa32f048e97a8b1ec7e6f989e781d..e29b9b0cad69d563a29336419c0d45ed99c823fa 100644 (file)
--- a/src/lib/str.h
+++ b/src/lib/str.h
@@ -78,4 +78,10 @@ static inline void str_truncate(string_t *str, size_t len)
                buffer_set_used_size(str, len);
 }
 
+/* Truncate the string to specified length, but also make sure the truncation
+   doesn't happen in the middle of an UTF-8 character sequence. In that case,
+   the string will end up being up to a few bytes smaller than len. If it's
+   already smaller to begin with, do nothing. */
+void str_truncate_utf8(string_t *str, size_t len);
+
 #endif

diff --git a/src/lib/test-str.c b/src/lib/test-str.c
index 79ce5fbc454b89bd741d90267b185aa138559710..e20e7f98f9eee92ba3f08451901dc7a2ec6673f2 100644 (file)
--- a/src/lib/test-str.c
+++ b/src/lib/test-str.c
@@ -1,6 +1,7 @@
 /* Copyright (c) 2012-2018 Dovecot authors, see the included COPYING file */
 
 #include "test-lib.h"
+#include "unichar.h"
 #include "str.h"
 
 static void test_str_append(void)
@@ -127,6 +128,48 @@ static void test_str_truncate(void)
        test_end();
 }
 
+static void test_str_truncate_utf8(void)
+{
+       string_t *str = t_str_new(8);
+       int i;
+
+       test_begin("str_truncate_utf8()");
+       str_append(str, "123456");
+       for (i = 100; i >= 6; i--) {
+               str_truncate_utf8(str, i);
+               test_assert_idx(str_len(str) == 6, i);
+       }
+       for (; i >= 0; i--) {
+               str_truncate_utf8(str, i);
+               test_assert_idx(str_len(str) == (unsigned int)i, i);
+       }
+
+       str_append(str, "\xE4\xB8\x80\xE4\xBa\x8C\xE4\xB8\x89"
+                       "\xE5\x9b\x9b\xE4\xBa\x94\xE5\x85\xAD");
+       for (i = 100; i >= 18; i--) {
+               str_truncate_utf8(str, i);
+               test_assert_idx(str_len(str) == 18, i);
+       }
+       for (; i >= 0; i--) {
+               str_truncate_utf8(str, i);
+               test_assert_idx(str_len(str) % 3 == 0, i);
+               test_assert_idx((str_len(str) / 3) == ((unsigned int)i / 3), i);
+       }
+
+       str_append(str, "\xE4\xB8\x80""1""\xE4\xBa\x8C""2""\xE4\xB8\x89""3"
+                       "\xE5\x9b\x9b""4""\xE4\xBa\x94""5""\xE5\x85\xAD""6");
+       for (i = 100; i >= 24; i--) {
+               str_truncate_utf8(str, i);
+               test_assert_idx(str_len(str) == 24, i);
+       }
+       for (; i >= 0; i--) {
+               str_truncate_utf8(str, i);
+               test_assert_idx(uni_utf8_data_is_valid(str_data(str),
+                                                      str_len(str)), i);
+       }
+       test_end();
+}
+
 void test_str(void)
 {
        test_str_append();
@@ -135,4 +178,5 @@ void test_str(void)
        test_str_delete();
        test_str_append_max();
        test_str_truncate();
+       test_str_truncate_utf8();
 }