include text context in researchLogger

when logging LatinIME.onUpdateSelection(), now include the current word and
preceding word.  no escaping of the word is performed; this is temporary
until the output format is cleaned up.

also fix EditingUtils.getWordRangeAtCursor to support supplementary
UTF-16 characters.

Bug: 6188932
Change-Id: If4612a2627537d5d8bb2f9585a3ad1b4e56c2e26

3 files changed, 54 insertions, 13 deletions

diff --git a/java/src/com/android/inputmethod/latin/EditingUtils.java b/java/src/com/android/inputmethod/latin/EditingUtils.java
index 93106ac27..be4cb7787 100644
--- a/java/src/com/android/inputmethod/latin/EditingUtils.java
+++ b/java/src/com/android/inputmethod/latin/EditingUtils.java
@@ -54,7 +54,7 @@ public class EditingUtils {
      */
     public static String getWordAtCursor(InputConnection connection, String separators) {
         // getWordRangeAtCursor returns null if the connection is null
-        Range r = getWordRangeAtCursor(connection, separators);
+        Range r = getWordRangeAtCursor(connection, separators, 0);
         return (r == null) ? null : r.mWord;
     }
 
@@ -84,7 +84,17 @@ public class EditingUtils {
         }
     }
 
-    private static Range getWordRangeAtCursor(InputConnection connection, String sep) {
+    /**
+     * Returns the text surrounding the cursor.
+     *
+     * @param connection the InputConnection to the TextView
+     * @param sep a string of characters that split words.
+     * @param additionalPrecedingWordsCount the number of words before the current word that should
+     *   be included in the returned range
+     * @return a range containing the text surrounding the cursor
+     */
+    public static Range getWordRangeAtCursor(InputConnection connection, String sep,
+            int additionalPrecedingWordsCount) {
         if (connection == null || sep == null) {
             return null;
         }
@@ -94,14 +104,40 @@ public class EditingUtils {
             return null;
         }
 
-        // Find first word separator before the cursor
+        // Going backward, alternate skipping non-separators and separators until enough words
+        // have been read.
         int start = before.length();
-        while (start > 0 && !isWhitespace(before.charAt(start - 1), sep)) start--;
+        boolean isStoppingAtWhitespace = true;  // toggles to indicate what to stop at
+        while (true) { // see comments below for why this is guaranteed to halt
+            while (start > 0) {
+                final int codePoint = Character.codePointBefore(before, start);
+                if (isStoppingAtWhitespace == isSeparator(codePoint, sep)) {
+                    break;  // inner loop
+                }
+                --start;
+                if (Character.isSupplementaryCodePoint(codePoint)) {
+                    --start;
+                }
+            }
+            // isStoppingAtWhitespace is true every other time through the loop,
+            // so additionalPrecedingWordsCount is guaranteed to become < 0, which
+            // guarantees outer loop termination
+            if (isStoppingAtWhitespace && (--additionalPrecedingWordsCount < 0)) {
+                break;  // outer loop
+            }
+            isStoppingAtWhitespace = !isStoppingAtWhitespace;
+        }
 
         // Find last word separator after the cursor
         int end = -1;
-        while (++end < after.length() && !isWhitespace(after.charAt(end), sep)) {
-            // Nothing to do here.
+        while (++end < after.length()) {
+            final int codePoint = Character.codePointAt(after, end);
+            if (isSeparator(codePoint, sep)) {
+                break;
+            }
+            if (Character.isSupplementaryCodePoint(codePoint)) {
+                ++end;
+            }
         }
 
         int cursor = getCursorPosition(connection);
@@ -114,8 +150,8 @@ public class EditingUtils {
         return null;
     }
 
-    private static boolean isWhitespace(int code, String whitespace) {
-        return whitespace.contains(String.valueOf((char) code));
+    private static boolean isSeparator(int code, String sep) {
+        return sep.indexOf(code) != -1;
     }
 
     private static final Pattern spaceRegex = Pattern.compile("\\s+");
diff --git a/java/src/com/android/inputmethod/latin/LatinIME.java b/java/src/com/android/inputmethod/latin/LatinIME.java
index e8211aecf..b7385ed21 100644
--- a/java/src/com/android/inputmethod/latin/LatinIME.java
+++ b/java/src/com/android/inputmethod/latin/LatinIME.java
@@ -755,7 +755,7 @@ public class LatinIME extends InputMethodService implements KeyboardActionListen
             ResearchLogger.latinIME_onUpdateSelection(mLastSelectionStart, mLastSelectionEnd,
                     oldSelStart, oldSelEnd, newSelStart, newSelEnd, composingSpanStart,
                     composingSpanEnd, mExpectingUpdateSelection,
-                    expectingUpdateSelectionFromLogger);
+                    expectingUpdateSelectionFromLogger, getCurrentInputConnection());
             if (expectingUpdateSelectionFromLogger) {
                 return;
             }
diff --git a/java/src/com/android/inputmethod/latin/ResearchLogger.java b/java/src/com/android/inputmethod/latin/ResearchLogger.java
index 92a9633de..a46ed03af 100644
--- a/java/src/com/android/inputmethod/latin/ResearchLogger.java
+++ b/java/src/com/android/inputmethod/latin/ResearchLogger.java
@@ -36,6 +36,7 @@ import com.android.inputmethod.keyboard.Key;
 import com.android.inputmethod.keyboard.KeyDetector;
 import com.android.inputmethod.keyboard.Keyboard;
 import com.android.inputmethod.keyboard.internal.KeyboardState;
+import com.android.inputmethod.latin.EditingUtils.Range;
 import com.android.inputmethod.latin.define.ProductionFlag;
 
 import java.io.BufferedWriter;
@@ -64,6 +65,7 @@ public class ResearchLogger implements SharedPreferences.OnSharedPreferenceChang
     private static final String PREF_USABILITY_STUDY_MODE = "usability_study_mode";
     private static final String PREF_RESEARCH_LOGGER_UUID_STRING = "pref_research_logger_uuid";
     private static final boolean DEBUG = false;
+    private static final String WHITESPACE_SEPARATORS = " \t\n\r";
 
     private static final ResearchLogger sInstance = new ResearchLogger(new LogFileManager());
     private static final int MAX_INPUTVIEW_LENGTH_TO_CAPTURE = 8192; // must be >=1
@@ -558,9 +560,10 @@ public class ResearchLogger implements SharedPreferences.OnSharedPreferenceChang
                     logUnstructured("LatinIME_onWindowHidden", "<no text>");
                 } else {
                     if (charSequence.length() > MAX_INPUTVIEW_LENGTH_TO_CAPTURE) {
-                        // do not cut in the middle of a supplementary character
                         int length = MAX_INPUTVIEW_LENGTH_TO_CAPTURE;
-                        if (!Character.isLetter(charSequence.charAt(length))) {
+                        // do not cut in the middle of a supplementary character
+                        final char c = charSequence.charAt(length-1);
+                        if (Character.isHighSurrogate(c)) {
                             length--;
                         }
                         final CharSequence truncatedCharSequence = charSequence.subSequence(0,
@@ -614,7 +617,7 @@ public class ResearchLogger implements SharedPreferences.OnSharedPreferenceChang
             final int lastSelectionEnd, final int oldSelStart, final int oldSelEnd,
             final int newSelStart, final int newSelEnd, final int composingSpanStart,
             final int composingSpanEnd, final boolean expectingUpdateSelection,
-            final boolean expectingUpdateSelectionFromLogger) {
+            final boolean expectingUpdateSelectionFromLogger, final InputConnection connection) {
         if (UnsLogGroup.LATINIME_ONUPDATESELECTION_ENABLED) {
             final String s = "onUpdateSelection: oss=" + oldSelStart
                     + ", ose=" + oldSelEnd
@@ -625,7 +628,9 @@ public class ResearchLogger implements SharedPreferences.OnSharedPreferenceChang
                     + ", cs=" + composingSpanStart
                     + ", ce=" + composingSpanEnd
                     + ", eus=" + expectingUpdateSelection
-                    + ", eusfl=" + expectingUpdateSelectionFromLogger;
+                    + ", eusfl=" + expectingUpdateSelectionFromLogger
+                    + ", context=\"" + EditingUtils.getWordRangeAtCursor(connection,
+                            WHITESPACE_SEPARATORS, 1).mWord + "\"";
             logUnstructured("LatinIME_onUpdateSelection", s);
         }
     }