Apache Solr open-source search software
LUCENE-9775: Hunspell: make FORCEUCASE work when the first compound word is inherently title-case (#2375)
Peter Gromov902cb93db227
diff --git a/lucene/analysis/common/src/java/org/apache/lucene/analysis/hunspell/CheckCompoundPattern.java b/lucene/analysis/common/src/java/org/apache/lucene/analysis/hunspell/CheckCompoundPattern.javaindex 5b4c504c818..b0617a7fbba 100644--- a/lucene/analysis/common/src/java/org/apache/lucene/analysis/hunspell/CheckCompoundPattern.java+++ b/lucene/analysis/common/src/java/org/apache/lucene/analysis/hunspell/CheckCompoundPattern.java@@ -94,7 +94,7 @@ class CheckCompoundPattern {CharsRef expandReplacement(CharsRef word, int breakPos) {if (replacement != null && charsMatch(word, breakPos, replacement)) {return new CharsRef(- word.subSequence(0, breakPos)+ new String(word.chars, 0, word.offset + breakPos)+ endChars+ beginChars+ word.subSequence(breakPos + replacement.length(), word.length));diff --git a/lucene/analysis/common/src/java/org/apache/lucene/analysis/hunspell/Hunspell.java b/lucene/analysis/common/src/java/org/apache/lucene/analysis/hunspell/Hunspell.javaindex e98200a5287..81254d8d8cb 100644--- a/lucene/analysis/common/src/java/org/apache/lucene/analysis/hunspell/Hunspell.java+++ b/lucene/analysis/common/src/java/org/apache/lucene/analysis/hunspell/Hunspell.java@@ -171,6 +171,10 @@ public class Hunspell {private boolean checkCompounds(CharsRef word, WordCase originalCase, CompoundPart prev) {if (prev != null && prev.index > dictionary.compoundMax - 2) return false;+ if (prev == null && word.offset != 0) {+ // we check the word's beginning for FORCEUCASE and expect to find it at 0+ throw new IllegalArgumentException();+ }int limit = word.length - dictionary.compoundMin + 1;for (int breakPos = dictionary.compoundMin; breakPos < limit; breakPos++) {@@ -231,7 +235,7 @@ public class Hunspell {if (lastRoot != null&& !dictionary.hasFlag(lastRoot.entryId, dictionary.forbiddenword)&& !(dictionary.checkCompoundDup && prev.root.equals(lastRoot))- && !hasForceUCaseProblem(lastRoot, originalCase)+ && !hasForceUCaseProblem(lastRoot, originalCase, word.chars)&& prev.mayCompound(lastRoot, remainingLength, originalCase)) {return true;}@@ -240,8 +244,9 @@ public class Hunspell {return checkCompounds(tail, originalCase, prev);}- private boolean hasForceUCaseProblem(Root<?> root, WordCase originalCase) {+ private boolean hasForceUCaseProblem(Root<?> root, WordCase originalCase, char[] wordChars) {if (originalCase == WordCase.TITLE || originalCase == WordCase.UPPER) return false;+ if (originalCase == null && Character.isUpperCase(wordChars[0])) return false;return dictionary.hasFlag(root.entryId, dictionary.forceUCase);}diff --git a/lucene/analysis/common/src/test/org/apache/lucene/analysis/hunspell/forceucase.dic b/lucene/analysis/common/src/test/org/apache/lucene/analysis/hunspell/forceucase.dicindex 82fd93b3097..c30fbc414f6 100644--- a/lucene/analysis/common/src/test/org/apache/lucene/analysis/hunspell/forceucase.dic+++ b/lucene/analysis/common/src/test/org/apache/lucene/analysis/hunspell/forceucase.dic@@ -2,3 +2,4 @@foo/Cbar/Cbaz/CA+Upper/C\ No newline at end of filediff --git a/lucene/analysis/common/src/test/org/apache/lucene/analysis/hunspell/forceucase.good b/lucene/analysis/common/src/test/org/apache/lucene/analysis/hunspell/forceucase.goodindex 37ecf495731..6da2cdfd647 100644--- a/lucene/analysis/common/src/test/org/apache/lucene/analysis/hunspell/forceucase.good+++ b/lucene/analysis/common/src/test/org/apache/lucene/analysis/hunspell/forceucase.good@@ -1,3 +1,4 @@+Foobazfoobarbaz@@ -5,3 +6,4 @@ foobarFoobazfoobazbarFoobarbaz+Upperbaz