organicmaps/icu
13
Fork 0
mirror of https://github.com/unicode-org/icu.git synced 2025-04-07 22:44:49 +00:00
Code Issues 1 Projects Releases Wiki Activity Actions 2

ICU-8046 CLDR 1.9 integration, fix he,fi brkitr

Browse source 
X-SVN-Rev: 28914
This commit is contained in:
Peter Edberg 2010-10-26 07:38:20 +00:00
parent f27448c127
commit 33120b6943
6 changed files with 25 additions and 15 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
icu4c/source/data/brkitr/fi.txt
View file

@ -9,6 +9,6 @@
fi{
Version{"2.2.89.3"}
boundaries{
sentence:process(dependency){"line_fi.brk"}
line:process(dependency){"line_fi.brk"}
}
}

2
icu4c/source/data/brkitr/he.txt
View file

@ -9,6 +9,6 @@
he{
Version{"2.2.88.46"}
boundaries{
sentence:process(dependency){"line_he.brk"}
line:process(dependency){"line_he.brk"}
}
}

4
icu4c/source/data/brkitr/line_fi.txt
View file

@ -326,12 +326,12 @@ $QUcm $LB18NonBreaks $CM*; # Don't let a combining mark go onto $CR, $BK, etc
$LB20NonBreaks = [$LB18NonBreaks - $CB];
# LB 20.09 added rule for Finnish tailoring
$SP / ($HY | $HH) $AL;
# LB 21 x (BA | HY | NS)
# BB x
#
$LB20NonBreaks $CM* ($BAcm | $HHcm | $HYcm | $NScm) / $AL;
$LB20NonBreaks $CM* ($BAcm | $HHcm | $HYcm | $NScm);
($HY | $HH) $AL;
$BBcm [^$CB]; # $BB x
$BBcm $LB20NonBreaks $CM*;

2
icu4c/source/data/xml/brkitr/fi.xml
View file

@ -17,7 +17,7 @@
<special xmlns:icu="http://www.icu-project.org/">
<icu:breakIteratorData>
<icu:boundaries>
<icu:sentence icu:dependency="line_fi.brk"/>
<icu:line icu:dependency="line_fi.brk"/>
</icu:boundaries>
</icu:breakIteratorData>
</special>

2
icu4c/source/data/xml/brkitr/he.xml
View file

@ -17,7 +17,7 @@
<special xmlns:icu="http://www.icu-project.org/">
<icu:breakIteratorData>
<icu:boundaries>
<icu:sentence icu:dependency="line_he.brk"/>
<icu:line icu:dependency="line_he.brk"/>
</icu:boundaries>
</icu:breakIteratorData>
</special>

28
icu4c/source/test/cintltst/cbiapts.c
View file

@ -713,6 +713,14 @@ static void TestBreakIteratorUText(void) {
* Test break iterator tailorings from CLDR data.
*/
/* Thai/Lao grapheme break tailoring */
static const UChar thTest[] = { 0x0020, 0x0E40, 0x0E01, 0x0020,
0x0E01, 0x0E30, 0x0020, 0x0E01, 0x0E33, 0x0020, 0 };
static const int32_t thTestOffs_enFwd[] = { 1, 3, 4, 6, 7, 9, 10 };
static const int32_t thTestOffs_thFwd[] = { 1, 2, 3, 4, 5, 6, 7, 9, 10 };
static const int32_t thTestOffs_enRev[] = { 9, 7, 6, 4, 3, 1, 0 };
static const int32_t thTestOffs_thRev[] = { 9, 7, 6, 5, 4, 3, 2, 1, 0 };
/* Hebrew line break tailoring, for cldrbug 3028 */
static const UChar heTest[] = { 0x0020, 0x002D, 0x0031, 0x0032, 0x0020,
0x0061, 0x002D, 0x006B, 0x0020,
@ -725,11 +733,11 @@ static const int32_t heTestOffs_enRev[] = { 22, 19, 17, 14, 12, 9, 7, 5, 1,
static const int32_t heTestOffs_heRev[] = { 19, 14, 12, 9, 7, 5, 1, 0 };
/* Finnish line break tailoring, for cldrbug 3029 */
static const UChar fiTest[] = { 0x0020, 0x002D, 0x0031, 0x0032, 0x0020,
0x0061, 0x002D, 0x006B, 0x0020,
0x0061, 0x0300, 0x2010, 0x006B, 0x0020,
0x0061, 0x0020, 0x002D, 0x006B, 0x0020,
0x0061, 0x0300, 0x0020, 0x2010, 0x006B, 0x0020, 0 };
static const UChar fiTest[] = { /* 00 */ 0x0020, 0x002D, 0x0031, 0x0032, 0x0020,
/* 05 */ 0x0061, 0x002D, 0x006B, 0x0020,
/* 09 */ 0x0061, 0x0300, 0x2010, 0x006B, 0x0020,
/* 14 */ 0x0061, 0x0020, 0x002D, 0x006B, 0x0020,
/* 19 */ 0x0061, 0x0300, 0x0020, 0x2010, 0x006B, 0x0020, 0 };
static const int32_t fiTestOffs_enFwd[] = { 1, 5, 7, 9, 12, 14, 16, 17, 19, 22, 23, 25 };
static const int32_t fiTestOffs_fiFwd[] = { 1, 5, 7, 9, 12, 14, 16, 19, 22, 25 };
static const int32_t fiTestOffs_enRev[] = { 23, 22, 19, 17, 16, 14, 12, 9, 7, 5, 1, 0 };
@ -745,10 +753,12 @@ typedef struct {
} RBBITailoringTest;
static const RBBITailoringTest tailoringTests[] = {
{ "en", UBRK_LINE, heTest, heTestOffs_enFwd, heTestOffs_enRev, sizeof(heTestOffs_enFwd)/ sizeof(heTestOffs_enFwd[0]) },
/* { "he", UBRK_LINE, heTest, heTestOffs_heFwd, heTestOffs_heRev, sizeof(heTestOffs_heFwd)/ sizeof(heTestOffs_heFwd[0]) }, not working yet */
{ "en", UBRK_LINE, fiTest, fiTestOffs_enFwd, fiTestOffs_enRev, sizeof(fiTestOffs_enFwd)/ sizeof(fiTestOffs_enFwd[0]) },
/* { "fi", UBRK_LINE, fiTest, fiTestOffs_fiFwd, fiTestOffs_fiRev, sizeof(fiTestOffs_fiFwd)/ sizeof(fiTestOffs_fiFwd[0]) }, not working yet */
{ "en", UBRK_CHARACTER, thTest, thTestOffs_enFwd, thTestOffs_enRev, sizeof(thTestOffs_enFwd)/sizeof(thTestOffs_enFwd[0]) },
{ "th", UBRK_CHARACTER, thTest, thTestOffs_thFwd, thTestOffs_thRev, sizeof(thTestOffs_thFwd)/sizeof(thTestOffs_thFwd[0]) },
{ "en", UBRK_LINE, heTest, heTestOffs_enFwd, heTestOffs_enRev, sizeof(heTestOffs_enFwd)/sizeof(heTestOffs_enFwd[0]) },
{ "he", UBRK_LINE, heTest, heTestOffs_heFwd, heTestOffs_heRev, sizeof(heTestOffs_heFwd)/sizeof(heTestOffs_heFwd[0]) },
{ "en", UBRK_LINE, fiTest, fiTestOffs_enFwd, fiTestOffs_enRev, sizeof(fiTestOffs_enFwd)/sizeof(fiTestOffs_enFwd[0]) },
{ "fi", UBRK_LINE, fiTest, fiTestOffs_fiFwd, fiTestOffs_fiRev, sizeof(fiTestOffs_fiFwd)/sizeof(fiTestOffs_fiFwd[0]) },
{ NULL, 0, NULL, NULL, NULL, 0 },
};