1# 2002 May 24 2# 3# The author disclaims copyright to this source code. In place of 4# a legal notice, here is a blessing: 5# 6# May you do good and not evil. 7# May you find forgiveness for yourself and forgive others. 8# May you share freely, never taking more than you give. 9# 10#*********************************************************************** 11# This file implements regression tests for SQLite library. The focus of 12# this file is testing the SQLite routines used for converting between the 13# various suported unicode encodings (UTF-8, UTF-16, UTF-16le and 14# UTF-16be). 15# 16# $Id: enc2.test,v 1.22 2005/02/26 17:31:28 drh Exp $ 17 18set testdir [file dirname $argv0] 19source $testdir/tester.tcl 20 21# If UTF16 support is disabled, ignore the tests in this file 22# 23ifcapable {!utf16} { 24 finish_test 25 return 26} 27 28# The rough organisation of tests in this file is: 29# 30# enc2.1.*: Simple tests with a UTF-8 db. 31# enc2.2.*: Simple tests with a UTF-16LE db. 32# enc2.3.*: Simple tests with a UTF-16BE db. 33# enc2.4.*: Test that attached databases must have the same text encoding 34# as the main database. 35# enc2.5.*: Test the behaviour of the library when a collation sequence is 36# not available for the most desirable text encoding. 37# enc2.6.*: Similar test for user functions. 38# enc2.7.*: Test that the VerifyCookie opcode protects against assuming the 39# wrong text encoding for the database. 40# enc2.8.*: Test sqlite3_complete16() 41# 42 43db close 44 45# Return the UTF-8 representation of the supplied UTF-16 string $str. 46proc utf8 {str} { 47 # If $str ends in two 0x00 0x00 bytes, knock these off before 48 # converting to UTF-8 using TCL. 49 binary scan $str \c* vals 50 if {[lindex $vals end]==0 && [lindex $vals end-1]==0} { 51 set str [binary format \c* [lrange $vals 0 end-2]] 52 } 53 54 set r [encoding convertfrom unicode $str] 55 return $r 56} 57 58# 59# This proc contains all the tests in this file. It is run 60# three times. Each time the file 'test.db' contains a database 61# with the following contents: 62set dbcontents { 63 CREATE TABLE t1(a PRIMARY KEY, b, c); 64 INSERT INTO t1 VALUES('one', 'I', 1); 65} 66# This proc tests that we can open and manipulate the test.db 67# database, and that it is possible to retreive values in 68# various text encodings. 69# 70proc run_test_script {t enc} { 71 72# Open the database and pull out a (the) row. 73do_test $t.1 { 74 set DB [sqlite3 db test.db] 75 execsql {SELECT * FROM t1} 76} {one I 1} 77 78# Insert some data 79do_test $t.2 { 80 execsql {INSERT INTO t1 VALUES('two', 'II', 2);} 81 execsql {SELECT * FROM t1} 82} {one I 1 two II 2} 83 84# Insert some data 85do_test $t.3 { 86 execsql { 87 INSERT INTO t1 VALUES('three','III',3); 88 INSERT INTO t1 VALUES('four','IV',4); 89 INSERT INTO t1 VALUES('five','V',5); 90 } 91 execsql {SELECT * FROM t1} 92} {one I 1 two II 2 three III 3 four IV 4 five V 5} 93 94# Use the index 95do_test $t.4 { 96 execsql { 97 SELECT * FROM t1 WHERE a = 'one'; 98 } 99} {one I 1} 100do_test $t.5 { 101 execsql { 102 SELECT * FROM t1 WHERE a = 'four'; 103 } 104} {four IV 4} 105ifcapable subquery { 106 do_test $t.6 { 107 execsql { 108 SELECT * FROM t1 WHERE a IN ('one', 'two'); 109 } 110 } {one I 1 two II 2} 111} 112 113# Now check that we can retrieve data in both UTF-16 and UTF-8 114do_test $t.7 { 115 set STMT [sqlite3_prepare $DB "SELECT a FROM t1 WHERE c>3;" -1 TAIL] 116 sqlite3_step $STMT 117 sqlite3_column_text $STMT 0 118} {four} 119 120do_test $t.8 { 121 sqlite3_step $STMT 122 utf8 [sqlite3_column_text16 $STMT 0] 123} {five} 124 125do_test $t.9 { 126 sqlite3_finalize $STMT 127} SQLITE_OK 128 129ifcapable vacuum { 130 execsql VACUUM 131} 132 133do_test $t.10 { 134 db eval {PRAGMA encoding} 135} $enc 136 137} 138 139# The three unicode encodings understood by SQLite. 140set encodings [list UTF-8 UTF-16le UTF-16be] 141 142set sqlite_os_trace 0 143set i 1 144foreach enc $encodings { 145 file delete -force test.db 146 sqlite3 db test.db 147 db eval "PRAGMA encoding = \"$enc\"" 148 execsql $dbcontents 149 db close 150 run_test_script enc2-$i $enc 151 db close 152 incr i 153} 154 155# Test that it is an error to try to attach a database with a different 156# encoding to the main database. 157do_test enc2-4.1 { 158 file delete -force test.db 159 sqlite3 db test.db 160 db eval "PRAGMA encoding = 'UTF-8'" 161 db eval "CREATE TABLE abc(a, b, c);" 162} {} 163do_test enc2-4.2 { 164 file delete -force test2.db 165 sqlite3 db2 test2.db 166 db2 eval "PRAGMA encoding = 'UTF-16'" 167 db2 eval "CREATE TABLE abc(a, b, c);" 168} {} 169do_test enc2-4.3 { 170 catchsql { 171 ATTACH 'test2.db' as aux; 172 } 173} {1 {attached databases must use the same text encoding as main database}} 174 175db2 close 176db close 177 178# The following tests - enc2-5.* - test that SQLite selects the correct 179# collation sequence when more than one is available. 180 181set ::values [list one two three four five] 182set ::test_collate_enc INVALID 183proc test_collate {enc lhs rhs} { 184 set ::test_collate_enc $enc 185 set l [lsearch -exact $::values $lhs] 186 set r [lsearch -exact $::values $rhs] 187 set res [expr $l - $r] 188 return $res 189} 190 191file delete -force test.db 192set DB [sqlite3 db test.db] 193do_test enc2-5.0 { 194 execsql { 195 CREATE TABLE t5(a); 196 INSERT INTO t5 VALUES('one'); 197 INSERT INTO t5 VALUES('two'); 198 INSERT INTO t5 VALUES('five'); 199 INSERT INTO t5 VALUES('three'); 200 INSERT INTO t5 VALUES('four'); 201 } 202} {} 203do_test enc2-5.1 { 204 add_test_collate $DB 1 1 1 205 set res [execsql {SELECT * FROM t5 ORDER BY 1 COLLATE test_collate}] 206 lappend res $::test_collate_enc 207} {one two three four five UTF-8} 208do_test enc2-5.2 { 209 add_test_collate $DB 0 1 0 210 set res [execsql {SELECT * FROM t5 ORDER BY 1 COLLATE test_collate}] 211 lappend res $::test_collate_enc 212} {one two three four five UTF-16LE} 213do_test enc2-5.3 { 214 add_test_collate $DB 0 0 1 215 set res [execsql {SELECT * FROM t5 ORDER BY 1 COLLATE test_collate}] 216 lappend res $::test_collate_enc 217} {one two three four five UTF-16BE} 218 219db close 220file delete -force test.db 221set DB [sqlite3 db test.db] 222execsql {pragma encoding = 'UTF-16LE'} 223do_test enc2-5.4 { 224 execsql { 225 CREATE TABLE t5(a); 226 INSERT INTO t5 VALUES('one'); 227 INSERT INTO t5 VALUES('two'); 228 INSERT INTO t5 VALUES('five'); 229 INSERT INTO t5 VALUES('three'); 230 INSERT INTO t5 VALUES('four'); 231 } 232} {} 233do_test enc2-5.5 { 234 add_test_collate $DB 1 1 1 235 set res [execsql {SELECT * FROM t5 ORDER BY 1 COLLATE test_collate}] 236 lappend res $::test_collate_enc 237} {one two three four five UTF-16LE} 238do_test enc2-5.6 { 239 add_test_collate $DB 1 0 1 240 set res [execsql {SELECT * FROM t5 ORDER BY 1 COLLATE test_collate}] 241 lappend res $::test_collate_enc 242} {one two three four five UTF-16BE} 243do_test enc2-5.7 { 244 add_test_collate $DB 1 0 0 245 set res [execsql {SELECT * FROM t5 ORDER BY 1 COLLATE test_collate}] 246 lappend res $::test_collate_enc 247} {one two three four five UTF-8} 248 249db close 250file delete -force test.db 251set DB [sqlite3 db test.db] 252execsql {pragma encoding = 'UTF-16BE'} 253do_test enc2-5.8 { 254 execsql { 255 CREATE TABLE t5(a); 256 INSERT INTO t5 VALUES('one'); 257 INSERT INTO t5 VALUES('two'); 258 INSERT INTO t5 VALUES('five'); 259 INSERT INTO t5 VALUES('three'); 260 INSERT INTO t5 VALUES('four'); 261 } 262} {} 263do_test enc2-5.9 { 264 add_test_collate $DB 1 1 1 265 set res [execsql {SELECT * FROM t5 ORDER BY 1 COLLATE test_collate}] 266 lappend res $::test_collate_enc 267} {one two three four five UTF-16BE} 268do_test enc2-5.10 { 269 add_test_collate $DB 1 1 0 270 set res [execsql {SELECT * FROM t5 ORDER BY 1 COLLATE test_collate}] 271 lappend res $::test_collate_enc 272} {one two three four five UTF-16LE} 273do_test enc2-5.11 { 274 add_test_collate $DB 1 0 0 275 set res [execsql {SELECT * FROM t5 ORDER BY 1 COLLATE test_collate}] 276 lappend res $::test_collate_enc 277} {one two three four five UTF-8} 278 279# Also test that a UTF-16 collation factory works. 280do_test enc2-5-12 { 281 add_test_collate $DB 0 0 0 282 catchsql { 283 SELECT * FROM t5 ORDER BY 1 COLLATE test_collate 284 } 285} {1 {no such collation sequence: test_collate}} 286do_test enc2-5.13 { 287 add_test_collate_needed $DB 288 set res [execsql {SELECT * FROM t5 ORDER BY 1 COLLATE test_collate}] 289 lappend res $::test_collate_enc 290} {one two three four five UTF-16BE} 291 292db close 293file delete -force test.db 294 295# The following tests - enc2-6.* - test that SQLite selects the correct 296# user function when more than one is available. 297 298proc test_function {enc arg} { 299 return "$enc $arg" 300} 301 302file delete -force test.db 303set DB [sqlite3 db test.db] 304execsql {pragma encoding = 'UTF-8'} 305do_test enc2-6.0 { 306 execsql { 307 CREATE TABLE t5(a); 308 INSERT INTO t5 VALUES('one'); 309 } 310} {} 311do_test enc2-6.1 { 312 add_test_function $DB 1 1 1 313 execsql { 314 SELECT test_function('sqlite') 315 } 316} {{UTF-8 sqlite}} 317db close 318set DB [sqlite3 db test.db] 319do_test enc2-6.2 { 320 add_test_function $DB 0 1 0 321 execsql { 322 SELECT test_function('sqlite') 323 } 324} {{UTF-16LE sqlite}} 325db close 326set DB [sqlite3 db test.db] 327do_test enc2-6.3 { 328 add_test_function $DB 0 0 1 329 execsql { 330 SELECT test_function('sqlite') 331 } 332} {{UTF-16BE sqlite}} 333 334db close 335file delete -force test.db 336set DB [sqlite3 db test.db] 337execsql {pragma encoding = 'UTF-16LE'} 338do_test enc2-6.3 { 339 execsql { 340 CREATE TABLE t5(a); 341 INSERT INTO t5 VALUES('sqlite'); 342 } 343} {} 344do_test enc2-6.4 { 345 add_test_function $DB 1 1 1 346 execsql { 347 SELECT test_function('sqlite') 348 } 349} {{UTF-16LE sqlite}} 350db close 351set DB [sqlite3 db test.db] 352do_test enc2-6.5 { 353 add_test_function $DB 0 1 0 354 execsql { 355 SELECT test_function('sqlite') 356 } 357} {{UTF-16LE sqlite}} 358db close 359set DB [sqlite3 db test.db] 360do_test enc2-6.6 { 361 add_test_function $DB 0 0 1 362 execsql { 363 SELECT test_function('sqlite') 364 } 365} {{UTF-16BE sqlite}} 366 367db close 368file delete -force test.db 369set DB [sqlite3 db test.db] 370execsql {pragma encoding = 'UTF-16BE'} 371do_test enc2-6.7 { 372 execsql { 373 CREATE TABLE t5(a); 374 INSERT INTO t5 VALUES('sqlite'); 375 } 376} {} 377do_test enc2-6.8 { 378 add_test_function $DB 1 1 1 379 execsql { 380 SELECT test_function('sqlite') 381 } 382} {{UTF-16BE sqlite}} 383db close 384set DB [sqlite3 db test.db] 385do_test enc2-6.9 { 386 add_test_function $DB 0 1 0 387 execsql { 388 SELECT test_function('sqlite') 389 } 390} {{UTF-16LE sqlite}} 391db close 392set DB [sqlite3 db test.db] 393do_test enc2-6.10 { 394 add_test_function $DB 0 0 1 395 execsql { 396 SELECT test_function('sqlite') 397 } 398} {{UTF-16BE sqlite}} 399 400 401db close 402file delete -force test.db 403 404# The following tests - enc2-7.* - function as follows: 405# 406# 1: Open an empty database file assuming UTF-16 encoding. 407# 2: Open the same database with a different handle assuming UTF-8. Create 408# a table using this handle. 409# 3: Read the sqlite_master table from the first handle. 410# 4: Ensure the first handle recognises the database encoding is UTF-8. 411# 412do_test enc2-7.1 { 413 sqlite3 db test.db 414 execsql { 415 PRAGMA encoding = 'UTF-16'; 416 SELECT * FROM sqlite_master; 417 } 418} {} 419do_test enc2-7.2 { 420 set enc [execsql { 421 PRAGMA encoding; 422 }] 423 string range $enc 0 end-2 ;# Chop off the "le" or "be" 424} {UTF-16} 425do_test enc2-7.3 { 426 sqlite3 db2 test.db 427 execsql { 428 PRAGMA encoding = 'UTF-8'; 429 CREATE TABLE abc(a, b, c); 430 } db2 431} {} 432do_test enc2-7.4 { 433 execsql { 434 SELECT * FROM sqlite_master; 435 } 436} "table abc abc [expr $AUTOVACUUM?3:2] {CREATE TABLE abc(a, b, c)}" 437do_test enc2-7.5 { 438 execsql { 439 PRAGMA encoding; 440 } 441} {UTF-8} 442 443db close 444db2 close 445 446proc utf16 {utf8} { 447 set utf16 [encoding convertto unicode $utf8] 448 append utf16 "\x00\x00" 449 return $utf16 450} 451ifcapable {complete} { 452 do_test enc2-8.1 { 453 sqlite3_complete16 [utf16 "SELECT * FROM t1;"] 454 } {1} 455 do_test enc2-8.2 { 456 sqlite3_complete16 [utf16 "SELECT * FROM"] 457 } {0} 458} 459 460finish_test 461