ICU-22986 GL takes CM

unicode-org · Dec 10, 2024 · 23d9a3e · 23d9a3e
1 parent 515d0a7
commit 23d9a3e
Show file tree

Hide file tree

Showing 4 changed files with 10 additions and 2 deletions.
diff --git a/icu4c/source/data/brkitr/rules/line.txt b/icu4c/source/data/brkitr/rules/line.txt
@@ -297,7 +297,7 @@ $LB20NonBreaks = [$LB18NonBreaks - $CB];
 #             and then to default UAX #14 behaviour (UTC-179-C32).
 #
 ^($HY | $HH) $CM* $ALPlus;
-$GL ($HY | $HH) $CM* $ALPlus; 
+$GL $CM* ($HY | $HH) $CM* $ALPlus; 
 # Non-breaking CB from LB8a:
 $CB $CM* $ZWJ ($HY | $HH) $CM* $ALPlus;
 # Non-breaking SP from LB14:

diff --git a/icu4c/source/test/testdata/break_rules/line.txt b/icu4c/source/test/testdata/break_rules/line.txt
@@ -176,7 +176,7 @@ LB11.2:      SP WJ;
 LB11.3:      WJ CM* [^CM];
 
 # Needs to apply before LB12, because the new monkeys are not greedy.
-LB20a.2:   GL (HY | HH) CM* AL;
+LB20a.2:   GL CM* (HY | HH) CM* AL;
 LB12:      GL CM* [^CM];
 
 LB12a:       [^SP BA HY] CM* GL;

diff --git a/icu4c/source/test/testdata/rbbitst.txt b/icu4c/source/test/testdata/rbbitst.txt
@@ -2214,3 +2214,7 @@ Bangkok)•</data>
 <data>•« Complex »« chaining » •</data>
 <data>•« .618 »•</data>  # Interaction with the ICU tailoring to break before such numbers.
 
+# A hyphen following non-breaking space that carries an intervening combining
+# mark is treated as word-initial; by LB20a it has no break opportunity after
+# it.  A bug in ICU 76 incorrectly handled that case (ICU-22986).
+<data>• ̄-k•</data>
diff --git a/icu4j/main/core/src/test/resources/com/ibm/icu/dev/test/rbbi/rbbitst.txt b/icu4j/main/core/src/test/resources/com/ibm/icu/dev/test/rbbi/rbbitst.txt
@@ -2214,3 +2214,7 @@ Bangkok)•</data>
 <data>•« Complex »« chaining » •</data>
 <data>•« .618 »•</data>  # Interaction with the ICU tailoring to break before such numbers.
 
+# A hyphen following non-breaking space that carries an intervening combining
+# mark is treated as word-initial; by LB20a it has no break opportunity after
+# it.  A bug in ICU 76 incorrectly handled that case (ICU-22986).
+<data>• ̄-k•</data>