* tests/string.test:

	* generic/tclVar.c (Tcl_SetVar2Ex):
	* generic/tclStringObj.c (Tcl_AppendObjToObj):
	* generic/tclCmdMZ.c (Tcl_StringObjCmd): optimized the string
	index, string length, string range, and append command in cases
	where the object's internal rep is a bytearray.  Objects with
	other internal reps are converted to have the new unicode internal
	rep.

	* unix/Makefile.in:
	* win/Makefile.in:
	* win/Makefile.vc:
	* tests/unicode.test:
	* generic/tclInt.h:
	* generic/tclObj.c:
	* generic/tclUnicodeObj.c: added a new object type to store the
	unicode representation of a string.

	* generic/tclTestObj.c: added the objtype option to the testobj
	command.  This option returns the name of the type of internal rep
	an object has.

2 files changed, 234 insertions, 1 deletions

diff --git a/tests/string.test b/tests/string.test
index 01ad4bf..235dba8 100644
--- a/tests/string.test
+++ b/tests/string.test
@@ -11,12 +11,17 @@
 # See the file "license.terms" for information on usage and redistribution
 # of this file, and for a DISCLAIMER OF ALL WARRANTIES.
 #
-# RCS: @(#) $Id: string.test,v 1.11 1999/06/03 18:43:30 stanton Exp $
+# RCS: @(#) $Id: string.test,v 1.12 1999/06/08 02:59:28 hershey Exp $
 
 if {[lsearch [namespace children] ::tcltest] == -1} {
     source [file join [pwd] [file dirname [info script]] defs.tcl]
 }
 
+# Some tests require the testobj command
+
+set ::tcltest::testConfig(testobj) \
+	[expr {[info commands testobj] != {}}]
+
 test string-1.1 {error conditions} {
     list [catch {string gorp a b} msg] $msg
 } {1 {bad option "gorp": must be bytelength, compare, equal, first, index, is, last, length, map, match, range, repeat, replace, tolower, toupper, totitle, trim, trimleft, trimright, wordend, or wordstart}}
@@ -226,6 +231,18 @@ test string-5.11 {string index, unicode} {
 test string-5.12 {string index, unicode over char length, under byte length} {
     string index \334\374\334\374 6
 } {}
+test string-5.13 {string index, bytearray object} {
+    string index [binary format a5 fuz] 0
+} f
+test string-5.14 {string index, bytearray object} {
+    string index [binary format I* {0x50515253 0x52}] 3
+} S
+test string-5.15 {string index, bytearray object} {
+    set b [binary format I* {0x50515253 0x52}]
+    set i1 [string index $b end-6]
+    set i2 [string index $b 1]
+    string compare $i1 $i2
+} 0
 
 test string-6.1 {string is, too few args} {
     list [catch {string is} msg] $msg
@@ -585,6 +602,12 @@ test string-9.4 {string length} {
 test string-9.5 {string length, unicode} {
     string le "abcd\u7266"
 } 5
+test string-9.6 {string length, bytearray object} {
+    string length [binary format a5 foo]
+} 5
+test string-9.7 {string length, bytearray object} {
+    string length [binary format I* {0x50515253 0x52}]
+} 8
 
 test string-10.1 {string map, too few args} {
     list [catch {string map} msg] $msg
@@ -798,6 +821,12 @@ test string-12.17 {string range, unicode} {
 test string-12.18 {string range, unicode} {
     string range ab\u7266cdefghijklmnop 2 3
 } \u7266c
+test string-12.19 {string range, bytearray object} {
+    set b [binary format I* {0x50515253 0x52}]
+    set r1 [string range $b 1 end-1]
+    set r2 [string range $b 1 6]
+    string compare $r1 $r2
+} 0
 
 test string-13.1 {string repeat} {
     list [catch {string repeat} msg] $msg
diff --git a/tests/unicode.test b/tests/unicode.test
new file mode 100644
index 0000000..6ee91c8
--- /dev/null
+++ b/tests/unicode.test
@@ -0,0 +1,204 @@
+# This file tests the tclUnicode.c file. 
+#
+# This file contains a collection of tests for one or more of the Tcl
+# built-in commands.  Sourcing this file into Tcl runs the tests and
+# generates output for errors.  No output means no errors were found.
+#
+# Copyright (c) 1999 by Scriptics Corporation.
+#
+# See the file "license.terms" for information on usage and redistribution
+# of this file, and for a DISCLAIMER OF ALL WARRANTIES.
+#
+# RCS: @(#) $Id: unicode.test,v 1.2 1999/06/08 02:59:30 hershey Exp $
+
+if {[lsearch [namespace children] ::tcltest] == -1} {
+    source [file join [pwd] [file dirname [info script]] defs.tcl]
+}
+
+# Some tests require the testobj command
+
+set ::tcltest::testConfig(testobj) \
+	[expr {[info commands testobj] != {}}]
+
+test unicode-1.1 {TclGetUniCharFromObj with byte-size chars} {
+    string index "abcdefghi" 0
+} "a"
+test unicode-1.2 {TclGetUniCharFromObj with byte-size chars} {
+    string index "abcdefghi" 3
+} "d"
+test unicode-1.3 {TclGetUniCharFromObj with byte-size chars} {
+    string index "abcdefghi" end
+} "i"
+test unicode-1.4 {TclGetUniCharFromObj with mixed width chars} {
+    string index "ïa¿b®c®¿dï" 0
+} "ï"
+test unicode-1.5 {TclGetUniCharFromObj} {
+    string index "ïa¿b®c®¿dï" 4
+} "®"
+test unicode-1.6 {TclGetUniCharFromObj} {
+    string index "ïa¿b®cï¿d®" end
+} "®"
+
+test unicode-2.1 {TclGetUnicodeLengthFromObj with byte-size chars} {
+    string length "" 
+} 0
+test unicode-2.2 {TclGetUnicodeLengthFromObj with byte-size chars} {
+    string length "a"
+} 1
+test unicode-2.3 {TclGetUnicodeLengthFromObj with byte-size chars} {
+    string length "abcdef"
+} 6
+test unicode-2.4 {TclGetUnicodeLengthFromObj with mixed width chars} {
+    string length "®" 
+} 1
+test unicode-2.5 {TclGetUnicodeLengthFromObj with mixed width chars} {
+    string length "￮￮" 
+} 6
+test unicode-2.6 {TclGetUnicodeLengthFromObj with mixed width chars} {
+    string length "ïa¿b®cï¿d®" 
+} 10
+
+test unicode-3.1 {TclGetRangeFromObj with all byte-size chars} {testobj} {
+    set x "abcdef"
+    list [testobj objtype $x] [set y [string range $x 1 end-1]] \
+	    [testobj objtype $x] [testobj objtype $y]
+} {none bcde unicode none}
+
+test unicode-3.2 {TclGetRangeFromObj with some mixed width chars} {testobj} {
+    set x "abcïïdef"
+    list [testobj objtype $x] [set y [string range $x 1 end-1]] \
+	    [testobj objtype $x] [testobj objtype $y]
+} {none bcïïde unicode unicode}
+
+test unicode-4.1 {UpdateStringOfUnicode} {testobj} {
+    set x 2345
+    list [string index $x end] [testobj objtype $x] [incr x] \
+	    [testobj objtype $x]
+} {5 unicode 2346 int}
+
+test unicode-5.1 {SetUnicodeFromAny called with non-unicode obj} {testobj} {
+    set x 2345
+    list [incr x] [testobj objtype $x] [string index $x end] \
+	    [testobj objtype $x]
+} {2346 int 6 unicode}
+
+test unicode-5.2 {SetUnicodeFromAny called with unicode obj} {testobj} {
+    set x "abcdef"
+    list [string length $x] [testobj objtype $x] \
+	    [string length $x] [testobj objtype $x]
+} {6 unicode 6 unicode}
+
+test unicode-6.1 {DupUnicodeInternalRep, mixed width chars} {testobj} {
+    set x abcï¿®ghi
+    string length $x
+    set y $x
+    list [testobj objtype $x] [testobj objtype $y] [append x "®¿ï"] \
+	    [set y] [testobj objtype $x] [testobj objtype $y]
+} {unicode unicode abcï¿®ghi®¿ï abcï¿®ghi unicode unicode}
+
+test unicode-6.2 {DupUnicodeInternalRep, mixed width chars} {testobj} {
+    set x abcï¿®ghi
+    set y $x
+    string length $x
+    list [testobj objtype $x] [testobj objtype $y] [append x "®¿ï"] \
+	    [set y] [testobj objtype $x] [testobj objtype $y]
+} {unicode unicode abcï¿®ghi®¿ï abcï¿®ghi unicode unicode}
+
+test unicode-6.3 {DupUnicodeInternalRep, all byte-size chars} {testobj} {
+    set x abcdefghi
+    string length $x
+    set y $x
+    list [testobj objtype $x] [testobj objtype $y] [append x jkl] \
+	    [set y] [testobj objtype $x] [testobj objtype $y]
+} {unicode unicode abcdefghijkl abcdefghi unicode unicode}
+
+test unicode-6.4 {DupUnicodeInternalRep, all byte-size chars} {testobj} {
+    set x abcdefghi
+    set y $x
+    string length $x
+    list [testobj objtype $x] [testobj objtype $y] [append x jkl] \
+	    [set y] [testobj objtype $x] [testobj objtype $y]
+} {unicode unicode abcdefghijkl abcdefghi unicode unicode}
+
+test unicode-7.1 {TclAppendObjToUnicodeObj, mixed src & dest} {testobj} {
+    set x abcï¿®ghi
+    set y ®¿ï
+    string length $x
+    list [testobj objtype $x] [testobj objtype $y] [append x $y] \
+	    [set y] [testobj objtype $x] [testobj objtype $y]
+} {unicode none abcï¿®ghi®¿ï ®¿ï unicode unicode}
+
+test unicode-7.2 {TclAppendObjToUnicodeObj, mixed src & dest} {testobj} {
+    set x abcï¿®ghi
+    string length $x
+    list [testobj objtype $x] [append x $x] [testobj objtype $x] \
+	    [append x $x] [testobj objtype $x]
+} {unicode abcï¿®ghiabcï¿®ghi unicode\
+abcï¿®ghiabcï¿®ghiabcï¿®ghiabcï¿®ghi\
+unicode}
+
+test unicode-7.3 {TclAppendObjToUnicodeObj, mixed src & 1-byte dest} {testobj} {
+    set x abcdefghi
+    set y ®¿ï
+    string length $x
+    list [testobj objtype $x] [testobj objtype $y] [append x $y] \
+	    [set y] [testobj objtype $x] [testobj objtype $y]
+} {unicode none abcdefghi®¿ï ®¿ï string unicode}
+
+test unicode-7.4 {TclAppendObjToUnicodeObj, 1-byte src & dest} {testobj} {
+    set x abcdefghi
+    set y jkl
+    string length $x
+    list [testobj objtype $x] [testobj objtype $y] [append x $y] \
+	    [set y] [testobj objtype $x] [testobj objtype $y]
+} {unicode none abcdefghijkl jkl unicode unicode}
+
+test unicode-7.5 {TclAppendObjToUnicodeObj, 1-byte src & dest} {testobj} {
+    set x abcdefghi
+    string length $x
+    list [testobj objtype $x] [append x $x] [testobj objtype $x] \
+	    [append x $x] [testobj objtype $x]
+} {unicode abcdefghiabcdefghi unicode abcdefghiabcdefghiabcdefghiabcdefghi\
+unicode}
+
+test unicode-7.6 {TclAppendObjToUnicodeObj, 1-byte src & mixed dest} {testobj} {
+    set x abcï¿®ghi
+    set y jkl
+    string length $x
+    list [testobj objtype $x] [testobj objtype $y] [append x $y] \
+	    [set y] [testobj objtype $x] [testobj objtype $y]
+} {unicode none abcï¿®ghijkl jkl unicode unicode}
+
+test unicode-7.7 {TclAppendObjToUnicodeObj, integer src & dest} {testobj} {
+    set x [expr {4 * 5}]
+    set y [expr {4 + 5}]
+    list [testobj objtype $x] [testobj objtype $y] [append x $y] \
+	    [testobj objtype $x] [append x $y] [testobj objtype $x] \
+	    [testobj objtype $y]
+} {int int 209 string 2099 string int}
+
+test unicode-7.8 {TclAppendObjToUnicodeObj, integer src & dest} {testobj} {
+    set x [expr {4 * 5}]
+    list [testobj objtype $x] [append x $x] [testobj objtype $x] \
+	    [append x $x] [testobj objtype $x]
+} {int 2020 string 20202020 unicode}
+
+test unicode-7.9 {TclAppendObjToUnicodeObj, integer src & 1-byte dest} {testobj} {
+    set x abcdefghi
+    set y [expr {4 + 5}]
+    string length $x
+    list [testobj objtype $x] [testobj objtype $y] [append x $y] \
+	    [set y] [testobj objtype $x] [testobj objtype $y]
+} {unicode int abcdefghi9 9 string int}
+
+test unicode-7.10 {TclAppendObjToUnicodeObj, integer src & mixed dest} {testobj} {
+    set x abcï¿®ghi
+    set y [expr {4 + 5}]
+    string length $x
+    list [testobj objtype $x] [testobj objtype $y] [append x $y] \
+	    [set y] [testobj objtype $x] [testobj objtype $y]
+} {unicode int abcï¿®ghi9 9 unicode int}
+
+# cleanup
+::tcltest::cleanupTests
+return