ICU-22707 UTC-179-C32 Upstream and improve the old Finnish tailoring …

…LB20a from CLDR-3029 and ICU-8151
unicode-org · Jun 21, 2024 · 0e71e57 · 0e71e57
1 parent c466f45
commit 0e71e57
Show file tree

Hide file tree

Showing 3 changed files with 426 additions and 8 deletions.
diff --git a/icu4c/source/data/brkitr/rules/line.txt b/icu4c/source/data/brkitr/rules/line.txt
@@ -311,6 +311,18 @@ $LB20NonBreaks = [$LB18NonBreaks - $CB];
 #             Note: this is not default UAX-14 behaviour. See issue ICU-8151.
 #
 ^($HY | $HH) $CM* $ALPlus;
+$GL ($HY | $HH) $CM* $ALPlus; 
+# Non-breaking CB from LB8a:
+$CB $CM* $ZWJ ($HY | $HH) $CM* $ALPlus;
+# Non-breaking SP from LB14:
+$OP $CM* $SP+ ($HY | $HH) $CM* $ALPlus; 
+# Non-breaking SP from LB15a:
+($OP $CM* $SP+ | [$OP $QU $GL] $CM*) ([\p{Pi} & $QU] $CM* $SP*)+ $SP ($HY | $HH) $CM* $ALPlus;
+^([\p{Pi} & $QU] $CM* $SP*)+ $SP ($HY | $HH) $CM* $ALPlus;
+# Non-breaking SP from LB15a following LB15b:
+$LB8NonBreaks [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ $SP ($HY | $HH) $CM* $ALPlus;
+$CAN_CM $CM*  [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ $SP ($HY | $HH) $CM* $ALPlus;
+^$CM+  [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ $SP ($HY | $HH) $CM* $ALPlus;
 
 # LB 21        x   (BA | HY | NS)
 #           BB x

diff --git a/icu4c/source/test/intltest/rbbitst.cpp b/icu4c/source/test/intltest/rbbitst.cpp
@@ -3320,14 +3320,36 @@ int32_t RBBILineMonkey::next(int32_t startPos) {
             break;
         }
 
-        //           Don't break between Hyphens and letters if a break precedes the hyphen.
-        //           Formerly this was a Finnish tailoring.
-        //           Moved to root in ICU 63. This is an ICU customization, not in UAX-14.
-        //           ^($HY | $HH) $AL;
-        if (fAL->contains(thisChar) && (fHY->contains(prevChar) || fHH->contains(prevChar)) &&
-                prevPosX2 == -1) {
-            setAppliedRule(pos, "LB 20.09");
-            continue;
+        // Don't break between Hyphens and letters if a break or a space precedes the hyphen.
+        // Formerly this was a Finnish tailoring.
+        // (sot | BK | CR | LF | NL | SP | ZW | CB | GL) ( HY | [\u2010] ) × AL
+        if (fAL->contains(thisChar) && (fHY->contains(prevChar) || fHH->contains(prevChar))) {
+            // sot ( HY | [\u2010] ) × AL.
+            if (prevPos == 0) {
+                setAppliedRule(pos, "LB 20a");
+                continue;
+            }
+            // prevPosX2 is -1 if there was a break; but the UAX #14 rules can
+            // look through breaks.
+            int breakObliviousPrevPosX2 = fText->moveIndex32(prevPos, -1);
+            if (fBK->contains(fText->char32At(breakObliviousPrevPosX2)) ||
+                fCR->contains(fText->char32At(breakObliviousPrevPosX2)) ||
+                fLF->contains(fText->char32At(breakObliviousPrevPosX2)) ||
+                fNL->contains(fText->char32At(breakObliviousPrevPosX2)) ||
+                fSP->contains(fText->char32At(breakObliviousPrevPosX2)) ||
+                fGL->contains(fText->char32At(breakObliviousPrevPosX2)) ||
+                fZW->contains(fText->char32At(breakObliviousPrevPosX2))) {
+                setAppliedRule(pos, "LB 20a");
+                continue;
+            }
+            while (breakObliviousPrevPosX2 > 0 &&
+                    fCM->contains(fText->char32At(breakObliviousPrevPosX2))) {
+                breakObliviousPrevPosX2 = fText->moveIndex32(breakObliviousPrevPosX2, -1);
+            }
+            if (fCB->contains(fText->char32At(breakObliviousPrevPosX2))) {
+                setAppliedRule(pos, "LB 20a");
+                continue;
+            }
         }
 
         if (fBA->contains(thisChar) ||