Read multi-byte char group counts

Change-Id: Idc62382f1c814e9bd1466c9f7dda1fcc8ba4137d

2 files changed, 6 insertions, 3 deletions

diff --git a/native/src/binary_format.h b/native/src/binary_format.h
index 9944fa2bd..1d74998f6 100644
--- a/native/src/binary_format.h
+++ b/native/src/binary_format.h
@@ -61,7 +61,9 @@ inline int BinaryFormat::detectFormat(const uint8_t* const dict) {
 }
 
 inline int BinaryFormat::getGroupCountAndForwardPointer(const uint8_t* const dict, int* pos) {
-    return dict[(*pos)++];
+    const int msb = dict[(*pos)++];
+    if (msb < 0x80) return msb;
+    return ((msb & 0x7F) << 8) | dict[(*pos)++];
 }
 
 inline uint8_t BinaryFormat::getFlagsAndForwardPointer(const uint8_t* const dict, int* pos) {
diff --git a/native/src/unigram_dictionary.cpp b/native/src/unigram_dictionary.cpp
index cd73fe3f8..ca7f0be0c 100644
--- a/native/src/unigram_dictionary.cpp
+++ b/native/src/unigram_dictionary.cpp
@@ -507,9 +507,10 @@ int UnigramDictionary::getMostFrequentWordLikeInner(const uint16_t * const inWor
     int maxFreq = -1;
     const uint8_t* const root = DICT_ROOT;
 
-    mStackChildCount[0] = root[0];
+    int startPos = 0;
+    mStackChildCount[0] = BinaryFormat::getGroupCountAndForwardPointer(root, &startPos);
     mStackInputIndex[0] = 0;
-    mStackSiblingPos[0] = 1;
+    mStackSiblingPos[0] = startPos;
     while (depth >= 0) {
         const int charGroupCount = mStackChildCount[depth];
         int pos = mStackSiblingPos[depth];