Additional tests for malformed UTF-8. (CVS 4011)

FossilOrigin-Name: 448d3ef670dce6d27c7d7b1be58088d45f8b6274

manifest

@@ -1,5 +1,5 @@
-C Make\ssure\spParse->rc\sgets\sset\swhenever\ssqlite3ErrorMsg()\sis\scalled.\s\sThis\sis\nadded\sinsurance\sthat\sparsing\swill\sstop\squickly\safter\san\serror.\s\sThis\schange\ndid\smake\sthe\sparser\sstop\sfaster\sin\ssome\scases,\swhich\srequired\ssome\srevisions\nto\stests.\s(CVS\s4010)
+C Additional\stests\sfor\smalformed\sUTF-8.\s(CVS\s4011)
-D 2007-05-15T16:51:37
+D 2007-05-15T18:35:21
 F Makefile.in 87b200ad9970907f76df734d29dff3d294c10935
 F Makefile.linux-gcc 2d8574d1ba75f129aba2019f0b959db380a90935
 F README 9c4e2d6706bdcc3efdd773ce752a8cdab4f90028
@@ -164,6 +164,7 @@ F test/autovacuum_crash.test 05a63b8805b20cfba7ace82856ce4ccdda075a31
 F test/autovacuum_ioerr.test c46a76869cb6eddbbb40b419b2b6c4c001766b1f
 F test/autovacuum_ioerr2.test dc189f323cf0546289b5a9bbda60bcb1fe52bd4b
 F test/avtrans.test b77740800de0d2cdcf394ef94d2d61225a1941d8
+F test/badutf.test e5cd3fd946b76fb0d05024156618ed905298f44a
 F test/between.test 16b1776c6323faadb097a52d673e8e3d8be7d070
 F test/bigfile.test ebc9ce9216e08bead63734ab816d0f27858f3b80
 F test/bigrow.test f0aeb7573dcb8caaafea76454be3ade29b7fc747
@@ -247,7 +248,7 @@ F test/fts2k.test 222d0b3bc8667753f18406aaea9906a6098ea016
 F test/fts2l.test 4c53c89ce3919003765ff4fd8d98ecf724d97dd3
 F test/fts2m.test 4b30142ead6f3ed076e880a2a464064c5ad58c51
 F test/fts2n.test a70357e72742681eaebfdbe9007b87ff3b771638
-F test/func.test bf30bac1c5ce10448ab739994268cf18f8b3fa30
+F test/func.test 605989453d1b42cec1d05c17aa232dc98e3e04e6
 F test/fuzz.test 3d382b2d3c6b909a4f08b46dc6ca6a149eeb3a81
 F test/fuzz2.test ea38692ce2da99ad79fe0be5eb1a452c1c4d37bb
 F test/hook.test 7e7645fd9a033f79cce8fdff151e32715e7ec50a
@@ -491,7 +492,7 @@ F www/tclsqlite.tcl bb0d1357328a42b1993d78573e587c6dcbc964b9
 F www/vdbe.tcl 87a31ace769f20d3627a64fa1fade7fed47b90d0
 F www/version3.tcl 890248cf7b70e60c383b0e84d77d5132b3ead42b
 F www/whentouse.tcl fc46eae081251c3c181bd79c5faef8195d7991a5
-P 9b91502ef0ea9c0bfb8362c4b70314dc8449eb9e
+P f84d9dab110c4415d9b772f8043397640162b6b2
-R 05f932c99108ea2fd9e2cf63bbb5a075
+R 53fa1bf671cabc8be454726443bd9326
 U drh
-Z 9a27c4d77fa449acc8811ad8c3119491
+Z 08fe10d1e019c102afe7db67bcfd77dd

manifest.uuid

@@ -1 +1 @@
-f84d9dab110c4415d9b772f8043397640162b6b2
+448d3ef670dce6d27c7d7b1be58088d45f8b6274

test/badutf.test

@@ -0,0 +1,139 @@
+# 2007 May 15
+#
+# The author disclaims copyright to this source code.  In place of
+# a legal notice, here is a blessing:
+#
+#    May you do good and not evil.
+#    May you find forgiveness for yourself and forgive others.
+#    May you share freely, never taking more than you give.
+#
+#***********************************************************************
+# This file implements regression tests for SQLite library. 
+#
+# This file checks to make sure SQLite is able to gracefully
+# handle malformed UTF-8.
+#
+# $Id: badutf.test,v 1.1 2007/05/15 18:35:21 drh Exp $
+
+set testdir [file dirname $argv0]
+source $testdir/tester.tcl
+
+do_test badutf-1.1 {
+  db eval {PRAGMA encoding=UTF8}
+  sqlite3_exec db {SELECT hex('%80') AS x}
+} {0 {x 80}}
+do_test badutf-1.2 {
+  sqlite3_exec db {SELECT hex('%81') AS x}
+} {0 {x 81}}
+do_test badutf-1.3 {
+  sqlite3_exec db {SELECT hex('%bf') AS x}
+} {0 {x BF}}
+do_test badutf-1.4 {
+  sqlite3_exec db {SELECT hex('%c0') AS x}
+} {0 {x C0}}
+do_test badutf-1.5 {
+  sqlite3_exec db {SELECT hex('%e0') AS x}
+} {0 {x E0}}
+do_test badutf-1.6 {
+  sqlite3_exec db {SELECT hex('%f0') AS x}
+} {0 {x F0}}
+do_test badutf-1.7 {
+  sqlite3_exec db {SELECT hex('%ff') AS x}
+} {0 {x FF}}
+
+do_test badutf-1.10 {
+  sqlite3 db2 {}
+  db2 eval {PRAGMA encoding=UTF16be}
+  sqlite3_exec db2 {SELECT hex('%80') AS x}
+} {0 {x 0080}}
+do_test badutf-1.11 {
+  sqlite3_exec db2 {SELECT hex('%81') AS x}
+} {0 {x 0081}}
+do_test badutf-1.12 {
+  sqlite3_exec db2 {SELECT hex('%bf') AS x}
+} {0 {x 00BF}}
+do_test badutf-1.13 {
+  sqlite3_exec db2 {SELECT hex('%c0') AS x}
+} {0 {x FFFD}}
+do_test badutf-1.14 {
+  sqlite3_exec db2 {SELECT hex('%c1') AS x}
+} {0 {x FFFD}}
+do_test badutf-1.15 {
+  sqlite3_exec db2 {SELECT hex('%c0%bf') AS x}
+} {0 {x FFFD}}
+do_test badutf-1.16 {
+  sqlite3_exec db2 {SELECT hex('%c1%bf') AS x}
+} {0 {x FFFD}}
+do_test badutf-1.17 {
+  sqlite3_exec db2 {SELECT hex('%c3%bf') AS x}
+} {0 {x 00FF}}
+do_test badutf-1.18 {
+  sqlite3_exec db2 {SELECT hex('%e0') AS x}
+} {0 {x FFFD}}
+do_test badutf-1.19 {
+  sqlite3_exec db2 {SELECT hex('%f0') AS x}
+} {0 {x FFFD}}
+do_test badutf-1.20 {
+  sqlite3_exec db2 {SELECT hex('%ff') AS x}
+} {0 {x FFFD}}
+
+
+do_test badutf-2.1 {
+  sqlite3_exec db {SELECT '%80'=CAST(x'80' AS text) AS x}
+} {0 {x 1}}
+do_test badutf-2.2 {
+  sqlite3_exec db {SELECT CAST('%80' AS blob)=x'80' AS x}
+} {0 {x 1}}
+
+do_test badutf-3.1 {
+  sqlite3_exec db {SELECT length('%80') AS x}
+} {0 {x 1}}
+do_test badutf-3.2 {
+  sqlite3_exec db {SELECT length('%61%62%63') AS x}
+} {0 {x 3}}
+do_test badutf-3.3 {
+  sqlite3_exec db {SELECT length('%7f%80%81') AS x}
+} {0 {x 3}}
+do_test badutf-3.4 {
+  sqlite3_exec db {SELECT length('%61%c0') AS x}
+} {0 {x 2}}
+do_test badutf-3.5 {
+  sqlite3_exec db {SELECT length('%61%c0%80%80%80%80%80%80%80%80%80%80') AS x}
+} {0 {x 2}}
+do_test badutf-3.6 {
+  sqlite3_exec db {SELECT length('%c0%80%80%80%80%80%80%80%80%80%80') AS x}
+} {0 {x 1}}
+do_test badutf-3.7 {
+  sqlite3_exec db {SELECT length('%80%80%80%80%80%80%80%80%80%80') AS x}
+} {0 {x 10}}
+do_test badutf-3.8 {
+  sqlite3_exec db {SELECT length('%80%80%80%80%80%f0%80%80%80%80') AS x}
+} {0 {x 6}}
+do_test badutf-3.9 {
+  sqlite3_exec db {SELECT length('%80%80%80%80%80%f0%80%80%80%ff') AS x}
+} {0 {x 7}}
+
+do_test badutf-4.1 {
+  sqlite3_exec db {SELECT hex(trim('%80%80%80%f0%80%80%80%ff','%80%ff')) AS x}
+} {0 {x F0}}
+do_test badutf-4.2 {
+  sqlite3_exec db {SELECT hex(ltrim('%80%80%80%f0%80%80%80%ff','%80%ff')) AS x}
+} {0 {x F0808080FF}}
+do_test badutf-4.3 {
+  sqlite3_exec db {SELECT hex(rtrim('%80%80%80%f0%80%80%80%ff','%80%ff')) AS x}
+} {0 {x 808080F0}}
+do_test badutf-4.4 {
+  sqlite3_exec db {SELECT hex(trim('%80%80%80%f0%80%80%80%ff','%ff%80')) AS x}
+} {0 {x 808080F0808080FF}}
+do_test badutf-4.5 {
+  sqlite3_exec db {SELECT hex(trim('%ff%80%80%f0%80%80%80%ff','%ff%80')) AS x}
+} {0 {x 80F0808080FF}}
+do_test badutf-4.6 {
+  sqlite3_exec db {SELECT hex(trim('%ff%80%f0%80%80%80%ff','%ff%80')) AS x}
+} {0 {x F0808080FF}}
+do_test badutf-4.7 {
+  sqlite3_exec db {SELECT hex(trim('%ff%80%f0%80%80%80%ff','%ff%80%80')) AS x}
+} {0 {x FF80F0808080FF}}
+
+db2 close
+finish_test

test/func.test

@@ -11,7 +11,7 @@
 # This file implements regression tests for SQLite library.  The
 # focus of this file is testing built-in functions.
 #
-# $Id: func.test,v 1.66 2007/05/11 01:44:52 drh Exp $
+# $Id: func.test,v 1.67 2007/05/15 18:35:21 drh Exp $
 
 set testdir [file dirname $argv0]
 source $testdir/tester.tcl
@@ -327,16 +327,30 @@ do_test func-9.5 {
 do_test func-9.10 {
   execsql {SELECT hex(x'00112233445566778899aAbBcCdDeEfF')}
 } {00112233445566778899AABBCCDDEEFF}
-do_test func-9.11 {
+set encoding [db one {PRAGMA encoding}]
+if {$encoding=="UTF-16le"} {
+  do_test func-9.11-utf16le {
     execsql {SELECT hex(replace('abcdefg','ef','12'))}
-} {61626364313267}
+  } {6100620063006400310032006700}
-do_test func-9.12 {
+  do_test func-9.12-utf16le {
     execsql {SELECT hex(replace('abcdefg','','12'))}
-} {{}}
+  } {{}}
-breakpoint
+  breakpoint
-do_test func-9.13 {
+  do_test func-9.13-utf16le {
     execsql {SELECT hex(replace('aabcdefg','a','aaa'))}
-} {616161616161626364656667}
+  } {610061006100610061006100620063006400650066006700}
+} elseif {$encoding=="UTF-8"} {
+  do_test func-9.11-utf8 {
+    execsql {SELECT hex(replace('abcdefg','ef','12'))}
+  } {61626364313267}
+  do_test func-9.12-utf8 {
+    execsql {SELECT hex(replace('abcdefg','','12'))}
+  } {{}}
+  breakpoint
+  do_test func-9.13-utf8 {
+    execsql {SELECT hex(replace('aabcdefg','a','aaa'))}
+  } {616161616161626364656667}
+}
   
 # Use the "sqlite_register_test_function" TCL command which is part of
 # the text fixture in order to verify correct operation of some of