diff options
author | Steve Bennett <steveb@workware.net.au> | 2010-10-20 16:01:17 +1000 |
---|---|---|
committer | Steve Bennett <steveb@workware.net.au> | 2010-11-17 07:57:37 +1000 |
commit | 9f6ad73686d6dc1fc8628be60a0d42a6ee20817c (patch) | |
tree | 455e400d7d49937b5814d824ff40461aee93b8ff /tests/stringmatch.test | |
parent | abac7fb5ee7d37150951b9618ba6a0ee57d98085 (diff) | |
download | jimtcl-9f6ad73686d6dc1fc8628be60a0d42a6ee20817c.zip jimtcl-9f6ad73686d6dc1fc8628be60a0d42a6ee20817c.tar.gz jimtcl-9f6ad73686d6dc1fc8628be60a0d42a6ee20817c.tar.bz2 |
Add UTF-8 support to Jim
Signed-off-by: Steve Bennett <steveb@workware.net.au>
Diffstat (limited to 'tests/stringmatch.test')
-rw-r--r-- | tests/stringmatch.test | 228 |
1 files changed, 228 insertions, 0 deletions
diff --git a/tests/stringmatch.test b/tests/stringmatch.test new file mode 100644 index 0000000..b964ed4 --- /dev/null +++ b/tests/stringmatch.test @@ -0,0 +1,228 @@ +# This file is a Tcl script to test the code in the file tclUtil.c. +# This file is organized in the standard fashion for Tcl tests. +# +# Copyright (c) 1995-1998 Sun Microsystems, Inc. +# Copyright (c) 1998-1999 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: util.test,v 1.7.2.1 2001/07/16 23:14:13 hobbs Exp $ + +source testing.tcl + +test stringmatch-5.1 {Tcl_StringMatch} { + string match ab*c abc +} 1 +test stringmatch-5.2 {Tcl_StringMatch} { + string match ab**c abc +} 1 +test stringmatch-5.3 {Tcl_StringMatch} { + string match ab* abcdef +} 1 +test stringmatch-5.4 {Tcl_StringMatch} { + string match *c abc +} 1 +test stringmatch-5.5 {Tcl_StringMatch} { + string match *3*6*9 0123456789 +} 1 +test stringmatch-5.6 {Tcl_StringMatch} { + string match *3*6*9 01234567890 +} 0 +test stringmatch-5.7 {Tcl_StringMatch: UTF-8} { + string match *u \u4e4fu +} 1 +test stringmatch-5.8 {Tcl_StringMatch} { + string match a?c abc +} 1 +ifutf8 { + test stringmatch-5.9 {Tcl_StringMatch: UTF-8} { + # skip one character in string + + string match a?c a\u4e4fc + } 1 +} +test stringmatch-5.10 {Tcl_StringMatch} { + string match a??c abc +} 0 +test stringmatch-5.11 {Tcl_StringMatch} { + string match ?1??4???8? 0123456789 +} 1 +test stringmatch-5.12 {Tcl_StringMatch} { + string match {[abc]bc} abc +} 1 +ifutf8 { + test stringmatch-5.13 {Tcl_StringMatch: UTF-8} { + # string += Tcl_UtfToUniChar(string, &ch); + + string match "\[\u4e4fxy\]bc" "\u4e4fbc" + } 1 +} +test stringmatch-5.14 {Tcl_StringMatch} { + # if ((*pattern == ']') || (*pattern == '\0')) + # badly formed pattern + + string match {[]} {[]} +} 0 +test stringmatch-5.15 {Tcl_StringMatch} { + # if ((*pattern == ']') || (*pattern == '\0')) + # badly formed pattern + + string match {[} {[} +} 0 +test stringmatch-5.16 {Tcl_StringMatch} { + string match {a[abc]c} abc +} 1 +ifutf8 { + test stringmatch-5.17 {Tcl_StringMatch: UTF-8} { + # pattern += Tcl_UtfToUniChar(pattern, &endChar); + # get 1 UTF-8 character + + string match "a\[a\u4e4fc]c" "a\u4e4fc" + } 1 +} +test stringmatch-5.18 {Tcl_StringMatch: UTF-8} { + # pattern += Tcl_UtfToUniChar(pattern, &endChar); + # proper advance: wrong answer would match on UTF trail byte of \u4e4f + + string match {a[a\u4e4fc]c} a\u008fc +} 0 +test stringmatch-5.19 {Tcl_StringMatch: UTF-8} { + # pattern += Tcl_UtfToUniChar(pattern, &endChar); + # proper advance. + + string match {a[a\u4e4fc]c} "acc" +} 1 +test stringmatch-5.20 {Tcl_StringMatch} { + string match {a[xyz]c} abc +} 0 +test stringmatch-5.21 {Tcl_StringMatch} { + string match {12[2-7]45} 12345 +} 1 +test stringmatch-5.22 {Tcl_StringMatch: UTF-8 range} { + string match "\[\u4e00-\u4e4f]" "0" +} 0 +ifutf8 { + test stringmatch-5.23 {Tcl_StringMatch: UTF-8 range} { + string match "\[\u4e00-\u4e4f]" "\u4e33" + } 1 + test stringmatch-5.24 {Tcl_StringMatch: UTF-8 range} { + string match "\[\u4e00-\u4e4f]" "\uff08" + } 0 +} +test stringmatch-5.25 {Tcl_StringMatch} { + string match {12[ab2-4cd]45} 12345 +} 1 +test stringmatch-5.26 {Tcl_StringMatch} { + string match {12[ab2-4cd]45} 12b45 +} 1 +test stringmatch-5.27 {Tcl_StringMatch} { + string match {12[ab2-4cd]45} 12d45 +} 1 +test stringmatch-5.28 {Tcl_StringMatch} { + string match {12[ab2-4cd]45} 12145 +} 0 +test stringmatch-5.29 {Tcl_StringMatch} { + string match {12[ab2-4cd]45} 12545 +} 0 +test stringmatch-5.30 {Tcl_StringMatch: forwards range} { + string match {[k-w]} "z" +} 0 +test stringmatch-5.31 {Tcl_StringMatch: forwards range} { + string match {[k-w]} "w" +} 1 +test stringmatch-5.32 {Tcl_StringMatch: forwards range} { + string match {[k-w]} "r" +} 1 +test stringmatch-5.33 {Tcl_StringMatch: forwards range} { + string match {[k-w]} "k" +} 1 +test stringmatch-5.34 {Tcl_StringMatch: forwards range} { + string match {[k-w]} "a" +} 0 +test stringmatch-5.35 {Tcl_StringMatch: reverse range} { + string match {[w-k]} "z" +} 0 +test stringmatch-5.36 {Tcl_StringMatch: reverse range} { + string match {[w-k]} "w" +} 1 +test stringmatch-5.37 {Tcl_StringMatch: reverse range} { + string match {[w-k]} "r" +} 1 +test stringmatch-5.38 {Tcl_StringMatch: reverse range} { + string match {[w-k]} "k" +} 1 +test stringmatch-5.39 {Tcl_StringMatch: reverse range} { + string match {[w-k]} "a" +} 0 +test stringmatch-5.40 {Tcl_StringMatch: skip correct number of ']'} { + string match {[A-]x} Ax +} 0 +test stringmatch-5.41 {Tcl_StringMatch: skip correct number of ']'} { + string match {[A-]]x} Ax +} 1 +test stringmatch-5.42 {Tcl_StringMatch: skip correct number of ']'} { + string match {[A-]]x} \ue1x +} 0 +ifutf8 { + test stringmatch-5.43 {Tcl_StringMatch: skip correct number of ']'} { + string match \[A-]\ue1]x \ue1x + } 1 +} +test stringmatch-5.44 {Tcl_StringMatch: skip correct number of ']'} { + string match {[A-]h]x} hx +} 1 +test stringmatch-5.45 {Tcl_StringMatch} { + # if (*pattern == '\0') + # badly formed pattern, still treats as a set + + string match {[a} a +} 1 +test stringmatch-5.46 {Tcl_StringMatch} { + string match {a\*b} a*b +} 1 +test stringmatch-5.47 {Tcl_StringMatch} { + string match {a\*b} ab +} 0 +test stringmatch-5.48 {Tcl_StringMatch} { + string match {a\*\?\[\]\\\x} "a*?\[\]\\x" +} 1 +test stringmatch-5.49 {Tcl_StringMatch} { + string match ** "" +} 1 +test stringmatch-5.50 {Tcl_StringMatch} { + string match *. "" +} 0 +test stringmatch-5.51 {Tcl_StringMatch} { + string match "" "" +} 1 + +# 'string match' doesn't support ^, which is different +# from 'scan' + +test stringmatch-6.1 {bracket in charset} { + string match {a[]b]c} {a]c} +} 0 + +test stringmatch-6.2 {bracket in charset} { + string match {a[]b]c} {abc} +} 0 + +test stringmatch-6.3 {charset with ^} { + string match {a[^]b]c} {axc} +} 0 + +test stringmatch-6.4 {charset with ^} { + string match {a[^]b]c} {a]c} +} 0 + +test stringmatch-6.5 {charset with ^} { + string match {a[^bc]d} {axd} +} 0 + +test stringmatch-6.6 {charset with ^} { + string match {a[\]]c} {a]c} +} 0 + + +testreport |