Fix a bug with the Greek question mark.

Bug: 17398284
Change-Id: If49593e2c959935a183a4cd9d988407cc1160599

5 files changed, 67 insertions, 5 deletions

diff --git a/java/res/values-el/donottranslate-config-spacing-and-punctuations.xml b/java/res/values-el/donottranslate-config-spacing-and-punctuations.xml
new file mode 100644
index 000000000..117c075a6
--- /dev/null
+++ b/java/res/values-el/donottranslate-config-spacing-and-punctuations.xml
@@ -0,0 +1,23 @@
+<?xml version="1.0" encoding="utf-8"?>
+<!--
+/*
+**
+** Copyright 2014, The Android Open Source Project
+**
+** Licensed under the Apache License, Version 2.0 (the "License");
+** you may not use this file except in compliance with the License.
+** You may obtain a copy of the License at
+**
+**     http://www.apache.org/licenses/LICENSE-2.0
+**
+** Unless required by applicable law or agreed to in writing, software
+** distributed under the License is distributed on an "AS IS" BASIS,
+** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+** See the License for the specific language governing permissions and
+** limitations under the License.
+*/
+-->
+<resources xmlns:xliff="urn:oasis:names:tc:xliff:document:1.2">
+    <!-- Symbols that terminate sentences and require capitalization on the next char -->
+    <string name="symbols_sentence_terminators">.;!?</string>
+</resources>
diff --git a/java/res/values/donottranslate-config-spacing-and-punctuations.xml b/java/res/values/donottranslate-config-spacing-and-punctuations.xml
index 2faf578d2..06144d021 100644
--- a/java/res/values/donottranslate-config-spacing-and-punctuations.xml
+++ b/java/res/values/donottranslate-config-spacing-and-punctuations.xml
@@ -33,9 +33,14 @@
     <string name="symbols_word_separators">"&#x0009;&#x0020;&#x000A;&#x00A0;"()[]{}*&amp;&lt;&gt;+=|.,;:!?/_\"</string>
     <!-- Word connectors -->
     <string name="symbols_word_connectors">\'-</string>
-    <!-- The sentence separator code point, for capitalization -->
+    <!-- The sentence separator code point, for capitalization and auto-insertion -->
     <!-- U+002E: "." FULL STOP   ; 2Eh = 46d -->
     <integer name="sentence_separator">46</integer>
+    <!-- The abbreviation marker code point -->
+    <!-- U+002E: "." FULL STOP   ; 2Eh = 46d -->
+    <integer name="abbreviation_marker">46</integer>
+    <!-- Symbols that terminate sentences and require capitalization on the next char -->
+    <string name="symbols_sentence_terminators">.?!</string>
     <!-- Whether this language uses spaces between words -->
     <bool name="current_language_has_spaces">true</bool>
 </resources>
diff --git a/java/src/com/android/inputmethod/latin/settings/SpacingAndPunctuations.java b/java/src/com/android/inputmethod/latin/settings/SpacingAndPunctuations.java
index b8d2a2248..8f2e9de61 100644
--- a/java/src/com/android/inputmethod/latin/settings/SpacingAndPunctuations.java
+++ b/java/src/com/android/inputmethod/latin/settings/SpacingAndPunctuations.java
@@ -35,6 +35,8 @@ public final class SpacingAndPunctuations {
     public final int[] mSortedWordSeparators;
     public final PunctuationSuggestions mSuggestPuncList;
     private final int mSentenceSeparator;
+    private final int mAbbreviationMarker;
+    private final int[] mSortedSentenceTerminators;
     public final String mSentenceSeparatorAndSpace;
     public final boolean mCurrentLanguageHasSpaces;
     public final boolean mUsesAmericanTypography;
@@ -54,7 +56,10 @@ public final class SpacingAndPunctuations {
                 res.getString(R.string.symbols_word_connectors));
         mSortedWordSeparators = StringUtils.toSortedCodePointArray(
                 res.getString(R.string.symbols_word_separators));
+        mSortedSentenceTerminators = StringUtils.toSortedCodePointArray(
+                res.getString(R.string.symbols_sentence_terminators));
         mSentenceSeparator = res.getInteger(R.integer.sentence_separator);
+        mAbbreviationMarker = res.getInteger(R.integer.abbreviation_marker);
         mSentenceSeparatorAndSpace = new String(new int[] {
                 mSentenceSeparator, Constants.CODE_SPACE }, 0, 2);
         mCurrentLanguageHasSpaces = res.getBoolean(R.bool.current_language_has_spaces);
@@ -92,6 +97,14 @@ public final class SpacingAndPunctuations {
         return Arrays.binarySearch(mSortedSymbolsClusteringTogether, code) >= 0;
     }
 
+    public boolean isSentenceTerminator(final int code) {
+        return Arrays.binarySearch(mSortedSentenceTerminators, code) >= 0;
+    }
+
+    public boolean isAbbreviationMarker(final int code) {
+        return code == mAbbreviationMarker;
+    }
+
     public boolean isSentenceSeparator(final int code) {
         return code == mSentenceSeparator;
     }
diff --git a/java/src/com/android/inputmethod/latin/utils/CapsModeUtils.java b/java/src/com/android/inputmethod/latin/utils/CapsModeUtils.java
index 92bd02c54..02f1c5f00 100644
--- a/java/src/com/android/inputmethod/latin/utils/CapsModeUtils.java
+++ b/java/src/com/android/inputmethod/latin/utils/CapsModeUtils.java
@@ -213,13 +213,22 @@ public final class CapsModeUtils {
         char c = cs.charAt(--j);
 
         // We found the next interesting chunk of text ; next we need to determine if it's the
-        // end of a sentence. If we have a question mark or an exclamation mark, it's the end of
-        // a sentence. If it's neither, the only remaining case is the period so we get the opposite
-        // case out of the way.
-        if (c == Constants.CODE_QUESTION_MARK || c == Constants.CODE_EXCLAMATION_MARK) {
+        // end of a sentence. If we have a sentence terminator (typically a question mark or an
+        // exclamation mark), then it's the end of a sentence; however, we treat the abbreviation
+        // marker specially because usually is the same char as the sentence separator (the
+        // period in most languages) and in this case we need to apply a heuristic to determine
+        // in which of these senses it's used.
+        if (spacingAndPunctuations.isSentenceTerminator(c)
+                && !spacingAndPunctuations.isAbbreviationMarker(c)) {
             return (TextUtils.CAP_MODE_CHARACTERS | TextUtils.CAP_MODE_WORDS
                     | TextUtils.CAP_MODE_SENTENCES) & reqModes;
         }
+        // If we reach here, we know we have whitespace before the cursor and before that there
+        // is something that either does not terminate the sentence, or a symbol preceded by the
+        // start of the text, or it's the sentence separator AND it happens to be the same code
+        // point as the abbreviation marker.
+        // If it's a symbol or something that does not terminate the sentence, then we need to
+        // return caps for MODE_CHARACTERS and MODE_WORDS, but not for MODE_SENTENCES.
         if (!spacingAndPunctuations.isSentenceSeparator(c) || j <= 0) {
             return (TextUtils.CAP_MODE_CHARACTERS | TextUtils.CAP_MODE_WORDS) & reqModes;
         }
diff --git a/tests/src/com/android/inputmethod/latin/utils/CapsModeUtilsTests.java b/tests/src/com/android/inputmethod/latin/utils/CapsModeUtilsTests.java
index 5d3664765..4646a823d 100644
--- a/tests/src/com/android/inputmethod/latin/utils/CapsModeUtilsTests.java
+++ b/tests/src/com/android/inputmethod/latin/utils/CapsModeUtilsTests.java
@@ -136,5 +136,17 @@ public class CapsModeUtilsTests extends AndroidTestCase {
         allPathsForCaps("Word; ", c | w, sp, false);
         allPathsForCaps("Word;", c | w, sp, true);
         allPathsForCaps("Word;", c, sp, false);
+        // Test for sentence terminators in Greek
+        sp = job.runInLocale(res, LocaleUtils.constructLocaleFromString("el"));
+        allPathsForCaps("Word? ", c | w | s, sp, false);
+        allPathsForCaps("Word?", c | w | s, sp, true);
+        allPathsForCaps("Word?", c, sp, false);
+        allPathsForCaps("Word! ", c | w | s, sp, false);
+        allPathsForCaps("Word!", c | w | s, sp, true);
+        allPathsForCaps("Word!", c, sp, false);
+        // In Greek ";" is the question mark and it terminates the sentence
+        allPathsForCaps("Word; ", c | w | s, sp, false);
+        allPathsForCaps("Word;", c | w | s, sp, true);
+        allPathsForCaps("Word;", c, sp, false);
     }
 }