src/test/test-utf8.c

   1 /* SPDX-License-Identifier: LGPL-2.1+ */
   2 /***
   3   This file is part of systemd.
   4
   5   Copyright 2013 Dave Reisner
   6 ***/
   7
   8 #include "alloc-util.h"
   9 #include "string-util.h"
  10 #include "utf8.h"
  11 #include "util.h"
  12
  13 static void test_utf8_is_printable(void) {
  14         assert_se(utf8_is_printable("ascii is valid\tunicode", 22));
  15         assert_se(utf8_is_printable("\342\204\242", 3));
  16         assert_se(!utf8_is_printable("\341\204", 2));
  17         assert_se(utf8_is_printable("ąę", 4));
  18 }
  19
  20 static void test_utf8_is_valid(void) {
  21         assert_se(utf8_is_valid("ascii is valid unicode"));
  22         assert_se(utf8_is_valid("\342\204\242"));
  23         assert_se(!utf8_is_valid("\341\204"));
  24 }
  25
  26 static void test_ascii_is_valid(void) {
  27         assert_se(ascii_is_valid("alsdjf\t\vbarr\nba z"));
  28         assert_se(!ascii_is_valid("\342\204\242"));
  29         assert_se(!ascii_is_valid("\341\204"));
  30 }
  31
  32 static void test_utf8_encoded_valid_unichar(void) {
  33         assert_se(utf8_encoded_valid_unichar("\342\204\242") == 3);
  34         assert_se(utf8_encoded_valid_unichar("\302\256") == 2);
  35         assert_se(utf8_encoded_valid_unichar("a") == 1);
  36         assert_se(utf8_encoded_valid_unichar("\341\204") < 0);
  37         assert_se(utf8_encoded_valid_unichar("\341\204\341\204") < 0);
  38 }
  39
  40 static void test_utf8_escaping(void) {
  41         _cleanup_free_ char *p1, *p2, *p3;
  42
  43         p1 = utf8_escape_invalid("goo goo goo");
  44         puts(p1);
  45         assert_se(utf8_is_valid(p1));
  46
  47         p2 = utf8_escape_invalid("\341\204\341\204");
  48         puts(p2);
  49         assert_se(utf8_is_valid(p2));
  50
  51         p3 = utf8_escape_invalid("\341\204");
  52         puts(p3);
  53         assert_se(utf8_is_valid(p3));
  54 }
  55
  56 static void test_utf8_escaping_printable(void) {
  57         _cleanup_free_ char *p1, *p2, *p3, *p4, *p5, *p6;
  58
  59         p1 = utf8_escape_non_printable("goo goo goo");
  60         puts(p1);
  61         assert_se(utf8_is_valid(p1));
  62
  63         p2 = utf8_escape_non_printable("\341\204\341\204");
  64         puts(p2);
  65         assert_se(utf8_is_valid(p2));
  66
  67         p3 = utf8_escape_non_printable("\341\204");
  68         puts(p3);
  69         assert_se(utf8_is_valid(p3));
  70
  71         p4 = utf8_escape_non_printable("ąę\n가너도루\n1234\n\341\204\341\204\n\001 \019\20\a");
  72         puts(p4);
  73         assert_se(utf8_is_valid(p4));
  74
  75         p5 = utf8_escape_non_printable("\001 \019\20\a");
  76         puts(p5);
  77         assert_se(utf8_is_valid(p5));
  78
  79         p6 = utf8_escape_non_printable("\xef\xbf\x30\x13");
  80         puts(p6);
  81         assert_se(utf8_is_valid(p6));
  82 }
  83
  84 static void test_utf16_to_utf8(void) {
  85         char *a = NULL;
  86         const uint16_t utf16[] = { htole16('a'), htole16(0xd800), htole16('b'), htole16(0xdc00), htole16('c'), htole16(0xd801), htole16(0xdc37) };
  87         const char utf8[] = { 'a', 'b', 'c', 0xf0, 0x90, 0x90, 0xb7, 0 };
  88
  89         a = utf16_to_utf8(utf16, 14);
  90         assert_se(a);
  91         assert_se(streq(a, utf8));
  92
  93         free(a);
  94 }
  95
  96 static void test_utf8_n_codepoints(void) {
  97         assert_se(utf8_n_codepoints("abc") == 3);
  98         assert_se(utf8_n_codepoints("zażółcić gęślą jaźń") == 19);
  99         assert_se(utf8_n_codepoints("串") == 1);
 100         assert_se(utf8_n_codepoints("") == 0);
 101         assert_se(utf8_n_codepoints("…👊🔪💐…") == 5);
 102         assert_se(utf8_n_codepoints("\xF1") == (size_t) -1);
 103 }
 104
 105 int main(int argc, char *argv[]) {
 106         test_utf8_is_valid();
 107         test_utf8_is_printable();
 108         test_ascii_is_valid();
 109         test_utf8_encoded_valid_unichar();
 110         test_utf8_escaping();
 111         test_utf8_escaping_printable();
 112         test_utf16_to_utf8();
 113         test_utf8_n_codepoints();
 114
 115         return 0;
 116 }