1 //===- unittests/Support/LocaleTest.cpp - Locale.h tests ------------------===// 2 // 3 // The LLVM Compiler Infrastructure 4 // 5 // This file is distributed under the University of Illinois Open Source 6 // License. See LICENSE.TXT for details. 7 // 8 //===----------------------------------------------------------------------===// 9 10 #include "llvm/Support/Locale.h" 11 #include "gtest/gtest.h" 12 13 namespace llvm { 14 namespace sys { 15 namespace locale { 16 namespace { 17 18 // FIXME: WIN32 implementation is incorrect. We should consider using the one 19 // from LocaleGeneric.inc for WIN32. 20 #ifndef _WIN32 21 TEST(Locale, columnWidth) { 22 // FIXME: This test fails with MacOSX implementation of columnWidth. 23 #ifndef __APPLE__ 24 EXPECT_EQ(0, columnWidth("")); 25 EXPECT_EQ(1, columnWidth(" ")); 26 EXPECT_EQ(1, columnWidth("a")); 27 EXPECT_EQ(1, columnWidth("~")); 28 29 EXPECT_EQ(6, columnWidth("abcdef")); 30 31 EXPECT_EQ(-1, columnWidth("\x01")); 32 EXPECT_EQ(-1, columnWidth("aaaaaaaaaa\x01")); 33 EXPECT_EQ(-1, columnWidth("\342\200\213")); // 200B ZERO WIDTH SPACE 34 35 EXPECT_EQ(0, columnWidth("\314\200")); // 0300 COMBINING GRAVE ACCENT 36 EXPECT_EQ(1, columnWidth("\340\270\201")); // 0E01 THAI CHARACTER KO KAI 37 EXPECT_EQ(2, columnWidth("\344\270\200")); // CJK UNIFIED IDEOGRAPH-4E00 38 39 EXPECT_EQ(4, columnWidth("\344\270\200\344\270\200")); 40 EXPECT_EQ(3, columnWidth("q\344\270\200")); 41 EXPECT_EQ(3, columnWidth("\314\200\340\270\201\344\270\200")); 42 43 // Invalid UTF-8 strings, columnWidth should error out. 44 EXPECT_EQ(-2, columnWidth("\344")); 45 EXPECT_EQ(-2, columnWidth("\344\270")); 46 EXPECT_EQ(-2, columnWidth("\344\270\033")); 47 EXPECT_EQ(-2, columnWidth("\344\270\300")); 48 EXPECT_EQ(-2, columnWidth("\377\366\355")); 49 50 EXPECT_EQ(-2, columnWidth("qwer\344")); 51 EXPECT_EQ(-2, columnWidth("qwer\344\270")); 52 EXPECT_EQ(-2, columnWidth("qwer\344\270\033")); 53 EXPECT_EQ(-2, columnWidth("qwer\344\270\300")); 54 EXPECT_EQ(-2, columnWidth("qwer\377\366\355")); 55 56 // UTF-8 sequences longer than 4 bytes correspond to unallocated Unicode 57 // characters. 58 EXPECT_EQ(-2, columnWidth("\370\200\200\200\200")); // U+200000 59 EXPECT_EQ(-2, columnWidth("\374\200\200\200\200\200")); // U+4000000 60 #endif // __APPLE__ 61 } 62 63 TEST(Locale, isPrint) { 64 EXPECT_EQ(false, isPrint(0)); // <control-0000>-<control-001F> 65 EXPECT_EQ(false, isPrint(0x01)); 66 EXPECT_EQ(false, isPrint(0x1F)); 67 EXPECT_EQ(true, isPrint(' ')); 68 EXPECT_EQ(true, isPrint('A')); 69 EXPECT_EQ(true, isPrint('~')); 70 EXPECT_EQ(false, isPrint(0x7F)); // <control-007F>..<control-009F> 71 EXPECT_EQ(false, isPrint(0x90)); 72 EXPECT_EQ(false, isPrint(0x9F)); 73 74 EXPECT_EQ(true, isPrint(0xAC)); 75 // FIXME: Figure out if we want to treat SOFT HYPHEN as printable character. 76 #ifndef __APPLE__ 77 EXPECT_EQ(false, isPrint(0xAD)); // SOFT HYPHEN 78 #endif // __APPLE__ 79 EXPECT_EQ(true, isPrint(0xAE)); 80 81 // MacOS implementation doesn't think it's printable. 82 #ifndef __APPLE__ 83 EXPECT_EQ(true, isPrint(0x0377)); // GREEK SMALL LETTER PAMPHYLIAN DIGAMMA 84 #endif // __APPLE__ 85 EXPECT_EQ(false, isPrint(0x0378)); // <reserved-0378>..<reserved-0379> 86 87 EXPECT_EQ(false, isPrint(0x0600)); // ARABIC NUMBER SIGN 88 89 EXPECT_EQ(false, isPrint(0x1FFFF)); // <reserved-1F774>..<noncharacter-1FFFF> 90 EXPECT_EQ(true, isPrint(0x20000)); // CJK UNIFIED IDEOGRAPH-20000 91 92 EXPECT_EQ(false, isPrint(0x10FFFF)); // noncharacter 93 } 94 95 #endif // _WIN32 96 97 } // namespace 98 } // namespace locale 99 } // namespace sys 100 } // namespace llvm 101
