sl@0: # 2002 May 24
sl@0: #
sl@0: # The author disclaims copyright to this source code.  In place of
sl@0: # a legal notice, here is a blessing:
sl@0: #
sl@0: #    May you do good and not evil.
sl@0: #    May you find forgiveness for yourself and forgive others.
sl@0: #    May you share freely, never taking more than you give.
sl@0: #
sl@0: #***********************************************************************
sl@0: # This file implements regression tests for SQLite library.  The focus of
sl@0: # this file is testing the SQLite routines used for converting between the
sl@0: # various suported unicode encodings (UTF-8, UTF-16, UTF-16le and
sl@0: # UTF-16be).
sl@0: #
sl@0: # $Id: enc.test,v 1.7 2007/05/23 16:23:09 danielk1977 Exp $
sl@0: 
sl@0: set testdir [file dirname $argv0]
sl@0: source $testdir/tester.tcl
sl@0: 
sl@0: # Skip this test if the build does not support multiple encodings.
sl@0: #
sl@0: ifcapable {!utf16} {
sl@0:   finish_test
sl@0:   return
sl@0: }
sl@0: 
sl@0: proc do_bincmp_test {testname got expect} {
sl@0:   binary scan $expect \c* expectvals
sl@0:   binary scan $got \c* gotvals
sl@0:   do_test $testname [list set dummy $gotvals] $expectvals
sl@0: }
sl@0: 
sl@0: # $utf16 is a UTF-16 encoded string. Swap each pair of bytes around
sl@0: # to change the byte-order of the string.
sl@0: proc swap_byte_order {utf16} {
sl@0:   binary scan $utf16 \c* ints
sl@0: 
sl@0:   foreach {a b} $ints {
sl@0:     lappend ints2 $b
sl@0:     lappend ints2 $a
sl@0:   }
sl@0: 
sl@0:   return [binary format \c* $ints2]
sl@0: }
sl@0: 
sl@0: #
sl@0: # Test that the SQLite routines for converting between UTF encodings
sl@0: # produce the same results as their TCL counterparts.
sl@0: #
sl@0: # $testname is the prefix to be used for the test names.
sl@0: # $str is a string to use for testing (encoded in UTF-8, as normal for TCL).
sl@0: #
sl@0: # The test procedure is:
sl@0: # 1. Convert the string from UTF-8 to UTF-16le and check that the TCL and
sl@0: #    SQLite routines produce the same results.
sl@0: #
sl@0: # 2. Convert the string from UTF-8 to UTF-16be and check that the TCL and
sl@0: #    SQLite routines produce the same results.
sl@0: #
sl@0: # 3. Use the SQLite routines to convert the native machine order UTF-16
sl@0: #    representation back to the original UTF-8. Check that the result
sl@0: #    matches the original representation.
sl@0: #
sl@0: # 4. Add a byte-order mark to each of the UTF-16 representations and
sl@0: #    check that the SQLite routines can convert them back to UTF-8.  For
sl@0: #    byte-order mark info, refer to section 3.10 of the unicode standard.
sl@0: #
sl@0: # 5. Take the byte-order marked UTF-16 strings from step 4 and ensure
sl@0: #    that SQLite can convert them both to native byte order UTF-16 
sl@0: #    strings, sans BOM.
sl@0: #
sl@0: # Coverage:
sl@0: #
sl@0: # sqlite_utf8to16be (step 2)
sl@0: # sqlite_utf8to16le (step 1)
sl@0: # sqlite_utf16to8 (steps 3, 4)
sl@0: # sqlite_utf16to16le (step 5)
sl@0: # sqlite_utf16to16be (step 5)
sl@0: #
sl@0: proc test_conversion {testname str} {
sl@0:  
sl@0:   # Step 1.
sl@0:   set utf16le_sqlite3 [test_translate $str UTF8 UTF16LE]
sl@0:   set utf16le_tcl [encoding convertto unicode $str]
sl@0:   append utf16le_tcl "\x00\x00"
sl@0:   if { $::tcl_platform(byteOrder)!="littleEndian" } {
sl@0:     set utf16le_tcl [swap_byte_order $utf16le_tcl]
sl@0:   }
sl@0:   do_bincmp_test $testname.1 $utf16le_sqlite3 $utf16le_tcl
sl@0:   set utf16le $utf16le_tcl
sl@0: 
sl@0:   # Step 2.
sl@0:   set utf16be_sqlite3 [test_translate $str UTF8 UTF16BE]
sl@0:   set utf16be_tcl [encoding convertto unicode $str]
sl@0:   append utf16be_tcl "\x00\x00"
sl@0:   if { $::tcl_platform(byteOrder)=="littleEndian" } {
sl@0:     set utf16be_tcl [swap_byte_order $utf16be_tcl]
sl@0:   }
sl@0:   do_bincmp_test $testname.2 $utf16be_sqlite3 $utf16be_tcl
sl@0:   set utf16be $utf16be_tcl
sl@0:  
sl@0:   # Step 3.
sl@0:   if { $::tcl_platform(byteOrder)=="littleEndian" } {
sl@0:     set utf16 $utf16le
sl@0:   } else {
sl@0:     set utf16 $utf16be
sl@0:   }
sl@0:   set utf8_sqlite3 [test_translate $utf16 UTF16 UTF8]
sl@0:   do_bincmp_test $testname.3 $utf8_sqlite3 [binarize $str]
sl@0: 
sl@0:   # Step 4 (little endian).
sl@0:   append utf16le_bom "\xFF\xFE" $utf16le
sl@0:   set utf8_sqlite3 [test_translate $utf16le_bom UTF16 UTF8 1]
sl@0:   do_bincmp_test $testname.4.le $utf8_sqlite3 [binarize $str]
sl@0: 
sl@0:   # Step 4 (big endian).
sl@0:   append utf16be_bom "\xFE\xFF" $utf16be
sl@0:   set utf8_sqlite3 [test_translate $utf16be_bom UTF16 UTF8]
sl@0:   do_bincmp_test $testname.4.be $utf8_sqlite3 [binarize $str]
sl@0: 
sl@0:   # Step 5 (little endian to little endian).
sl@0:   set utf16_sqlite3 [test_translate $utf16le_bom UTF16LE UTF16LE]
sl@0:   do_bincmp_test $testname.5.le.le $utf16_sqlite3 $utf16le
sl@0: 
sl@0:   # Step 5 (big endian to big endian).
sl@0:   set utf16_sqlite3 [test_translate $utf16be_bom UTF16 UTF16BE]
sl@0:   do_bincmp_test $testname.5.be.be $utf16_sqlite3 $utf16be
sl@0: 
sl@0:   # Step 5 (big endian to little endian).
sl@0:   set utf16_sqlite3 [test_translate $utf16be_bom UTF16 UTF16LE]
sl@0:   do_bincmp_test $testname.5.be.le $utf16_sqlite3 $utf16le
sl@0: 
sl@0:   # Step 5 (little endian to big endian).
sl@0:   set utf16_sqlite3 [test_translate $utf16le_bom UTF16 UTF16BE]
sl@0:   do_bincmp_test $testname.5.le.be $utf16_sqlite3 $utf16be
sl@0: }
sl@0: 
sl@0: translate_selftest
sl@0: 
sl@0: test_conversion enc-1 "hello world"
sl@0: test_conversion enc-2 "sqlite"
sl@0: test_conversion enc-3 ""
sl@0: test_conversion enc-X "\u0100"
sl@0: test_conversion enc-4 "\u1234"
sl@0: test_conversion enc-5 "\u4321abc"
sl@0: test_conversion enc-6 "\u4321\u1234"
sl@0: test_conversion enc-7 [string repeat "abcde\u00EF\u00EE\uFFFCabc" 100]
sl@0: test_conversion enc-8 [string repeat "\u007E\u007F\u0080\u0081" 100]
sl@0: test_conversion enc-9 [string repeat "\u07FE\u07FF\u0800\u0801\uFFF0" 100]
sl@0: test_conversion enc-10 [string repeat "\uE000" 100]
sl@0: 
sl@0: proc test_collate {enc zLeft zRight} {
sl@0:   return [string compare $zLeft $zRight]
sl@0: }
sl@0: add_test_collate $::DB 0 0 1
sl@0: do_test enc-11.1 {
sl@0:   execsql {
sl@0:     CREATE TABLE ab(a COLLATE test_collate, b);
sl@0:     INSERT INTO ab VALUES(CAST (X'C388' AS TEXT), X'888800');
sl@0:     INSERT INTO ab VALUES(CAST (X'C0808080808080808080808080808080808080808080808080808080808080808080808080808080808080808080808080808080808388' AS TEXT), X'888800');
sl@0:     CREATE INDEX ab_i ON ab(a, b);
sl@0:   }
sl@0: } {}
sl@0: do_test enc-11.2 {
sl@0:   set cp200 "\u00C8"
sl@0:   execsql {
sl@0:     SELECT count(*) FROM ab WHERE a = $::cp200;
sl@0:   }
sl@0: } {2}
sl@0: 
sl@0: finish_test