pike.git/ src/ testsuite.in

2005-04-02

2005-04-02 11:43:03 by Martin Stjernholm <mast@lysator.liu.se>

• dee8943255b4fb97a831cdb66c1b0cd15f497d12 (63 lines) (+56/-7) ^{[ Show | Annotate ]}
  Branch: 7.9

Made string_to_utf8 and utf8_to_string unicode 3.1 compliant by not allowing
chars outside the valid ranges and by not decoding non-shortest forms.

Rev: src/builtin_functions.c:1.594
Rev: src/testsuite.in:1.753

1:   START_MARKER - test_true([["$Id: testsuite.in,v 1.752 2005/02/18 13:54:04 grubba Exp $"]]); + test_true([["$Id: testsuite.in,v 1.753 2005/04/02 11:43:03 mast Exp $"]]);      // This triggered a bug only if run sufficiently early.   test_compile_any([[#pike 7.2]])
10030:   test_eq(string_to_utf8("foo"), "foo")   test_eq(string_to_utf8("blä"), "bl\303\244")   test_eq(string_to_utf8("\77077"), "\347\270\277") - test_eq(string_to_utf8("\7077077"), "\367\207\270\277") - test_eq(string_to_utf8("\77077077"), "\370\277\207\270\277") - test_eq(string_to_utf8("\7077077077"), "\374\270\277\207\270\277") + test_eq(string_to_utf8("\7077077", 1), "\367\207\270\277") + test_eq(string_to_utf8("\77077077", 1), "\370\277\207\270\277") + test_eq(string_to_utf8("\7077077077", 1), "\374\270\277\207\270\277")      // 077077077077 has 33 bits unsigned. The escape sequence parser used   // to silently truncate too long char values, but not anymore.
10042:   test_eq(string_to_utf8("\37077077077", 1), "\376\203\270\277\207\270\277")   test_eq(utf8_to_string("\376\203\270\277\207\270\277", 1), "\37077077077")    - test_eq(utf8_to_string("\374\270\277\207\270\277"), "\7077077077") - test_eq(utf8_to_string("\370\277\207\270\277"), "\77077077") - test_eq(utf8_to_string("\367\207\270\277"), "\7077077") + test_eq(utf8_to_string("\374\270\277\207\270\277", 1), "\7077077077") + test_eq(utf8_to_string("\370\277\207\270\277", 1), "\77077077") + test_eq(utf8_to_string("\367\207\270\277", 1), "\7077077")   test_eq(utf8_to_string("\347\270\277"), "\77077")   test_eq(utf8_to_string("bl\303\244"), "blä")   test_eq(utf8_to_string("foo"), "foo")
10066:   test_eval_error(return utf8_to_string("\347\270a"));   test_eval_error(return utf8_to_string("\303a"));    + // Invalid ranges + test_eq(string_to_utf8 ("\ud7ff"), "\u00ed\u009f\u00bf") + test_eval_error(return string_to_utf8 ("\ud800")) + test_eq(string_to_utf8 ("\ud800", 1), "\u00ed\u00a0\u0080") + test_eq(string_to_utf8 ("\udfff", 1), "\u00ed\u00bf\u00bf") + test_eval_error(return string_to_utf8 ("\udfff")) + test_eq(string_to_utf8 ("\ue000"), "\u00ee\u0080\u0080") + test_eq(string_to_utf8 ("\U0010ffff"), "\u00f4\u008f\u00bf\u00bf") + test_eval_error(return string_to_utf8 ("\U00110000")) + test_eq(string_to_utf8 ("\U00110000", 1), "\u00f4\u0090\u0080\u0080") +  + test_eq(utf8_to_string ("\u00ed\u009f\u00bf"), "\ud7ff") + test_eval_error(return utf8_to_string ("\u00ed\u00a0\u0080")) + test_eq(utf8_to_string ("\u00ed\u00a0\u0080", 1), "\ud800") + test_eq(utf8_to_string ("\u00ed\u00bf\u00bf", 1), "\udfff") + test_eval_error(return utf8_to_string ("\u00ed\u00bf\u00bf")) + test_eq(utf8_to_string ("\u00ee\u0080\u0080"), "\ue000") + test_eq(utf8_to_string ("\u00f4\u008f\u00bf\u00bf"), "\U0010ffff") + test_eval_error(return utf8_to_string ("\u00f4\u0090\u0080\u0080")) + test_eq(utf8_to_string ("\u00f4\u0090\u0080\u0080", 1), "\U00110000") + test_eval_error(return utf8_to_string ("\u00f8\u0088\u0080\u0080\u0080")) + test_eval_error(return utf8_to_string ("\u00fc\u0084\u0080\u0080\u0080\u0080")) + test_eval_error(return utf8_to_string ("\u00fe\u0082\u0080\u0080\u0080\u0080\u0080")) + test_eq(utf8_to_string ("\u00fe\u0083\u00bf\u00bf\u00bf\u00bf\u00bf", 1), "\Uffffffff") + test_eval_error(return utf8_to_string ("\u00fe\u0084\u0080\u0080\u0080\u0080\u0080", 1)) + test_eval_error(return utf8_to_string ("\u00ff")) + test_eval_error(return utf8_to_string ("\u00ff", 1)) +  + // Non-shortest forms + test_eval_error(return utf8_to_string ("\u00c0\u0080")) + test_eval_error(return utf8_to_string ("\u00c1\u00bf")) + test_eq(utf8_to_string ("\u00c2\u0080"), "\u0080") + test_eval_error(return utf8_to_string ("\u00e0\u0080\u0080")) + test_eval_error(return utf8_to_string ("\u00e0\u009f\u00bf")) + test_eq(utf8_to_string ("\u00e0\u00a0\u0080"), "\u0800")) + test_eval_error(return utf8_to_string ("\u00f0\u0080\u0080\u0080")) + test_eval_error(return utf8_to_string ("\u00f0\u008f\u00bf\u00bf")) + test_eq(utf8_to_string ("\u00f0\u0090\u0080\u0080"), "\U00010000") + test_eval_error(return utf8_to_string ("\u00f8\u0080\u0080\u0080\u0080", 1)) + test_eval_error(return utf8_to_string ("\u00f8\u0087\u00bf\u00bf\u00bf", 1)) + test_eq(utf8_to_string ("\u00f8\u0088\u0080\u0080\u0080", 1), "\U00200000") + test_eval_error(return utf8_to_string ("\u00fc\u0080\u0080\u0080\u0080\u0080", 1)) + test_eval_error(return utf8_to_string ("\u00fc\u0083\u00bf\u00bf\u00bf\u00bf", 1)) + test_eq(utf8_to_string ("\u00fc\u0084\u0080\u0080\u0080\u0080", 1), "\U04000000") + test_eval_error(return utf8_to_string ("\u00fe\u0080\u0080\u0080\u0080\u0080\u0080", 1)) + test_eval_error(return utf8_to_string ("\u00fe\u0081\u00bf\u00bf\u00bf\u00bf\u00bf", 1)) + test_eq(utf8_to_string ("\u00fe\u0082\u0080\u0080\u0080\u0080\u0080", 1), "\U80000000") +    // - stringp   // Tested in foop