[parsing] Fix past the end position for streaming streams.

Also, as this is hard to track down, always DCHECK position after ReadBlock().

Change-Id: Ie32c3a311dd8df91f651b6d82ccacc7c95e6fde0
Reviewed-on: https://chromium-review.googlesource.com/528196
Commit-Queue: Marja Hölttä <marja@chromium.org>
Reviewed-by: Marja Hölttä <marja@chromium.org>
Reviewed-by: Daniel Vogelheim <vogelheim@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45811}

AUTHORS

@@ -125,6 +125,7 @@ Victor Costan <costan@gmail.com>
 Vlad Burlik <vladbph@gmail.com>
 Vladimir Krivosheev <develar@gmail.com>
 Vladimir Shutoff <vovan@shutoff.ru>
+Wiktor Garbacz <wiktor.garbacz@gmail.com>
 Yu Yin <xwafish@gmail.com>
 Zac Hansen <xaxxon@gmail.com>
 Zhongping Wang <kewpie.w.zp@gmail.com>

src/parsing/scanner-character-streams.cc

@@ -598,6 +598,7 @@ bool TwoByteExternalStreamingStream::ReadBlock() {
 
   // Out of data? Return 0.
   if (chunks_[chunk_no].byte_length == 0) {
+    buffer_pos_ = position;
     buffer_cursor_ = buffer_start_;
     buffer_end_ = buffer_start_;
     return false;
@@ -700,6 +701,7 @@ bool TwoByteExternalBufferedStream::ReadBlock() {
 
   // Out of data? Return 0.
   if (chunks_[chunk_no].byte_length == 0) {
+    buffer_pos_ = position;
     buffer_cursor_ = buffer_start_;
     buffer_end_ = buffer_start_;
     return false;

src/parsing/scanner.h

@@ -43,7 +43,7 @@ class Utf16CharacterStream {
   inline uc32 Advance() {
     if (V8_LIKELY(buffer_cursor_ < buffer_end_)) {
       return static_cast<uc32>(*(buffer_cursor_++));
-    } else if (ReadBlock()) {
+    } else if (ReadBlockChecked()) {
       return static_cast<uc32>(*(buffer_cursor_++));
     } else {
       // Note: currently the following increment is necessary to avoid a
@@ -102,6 +102,21 @@ class Utf16CharacterStream {
         buffer_pos_(buffer_pos) {}
   Utf16CharacterStream() : Utf16CharacterStream(nullptr, nullptr, nullptr, 0) {}
 
+  bool ReadBlockChecked() {
+    size_t position = pos();
+    USE(position);
+    bool success = ReadBlock();
+
+    // Post-conditions: 1, We should always be at the right position.
+    //                  2, Cursor should be inside the buffer.
+    //                  3, We should have more characters available iff success.
+    DCHECK_EQ(pos(), position);
+    DCHECK_LE(buffer_cursor_, buffer_end_);
+    DCHECK_LE(buffer_start_, buffer_cursor_);
+    DCHECK_EQ(success, buffer_cursor_ < buffer_end_);
+    return success;
+  }
+
   void ReadBlockAt(size_t new_pos) {
     // The callers of this method (Back/Back2/Seek) should handle the easy
     // case (seeking within the current buffer), and we should only get here
@@ -113,14 +128,8 @@ class Utf16CharacterStream {
     // Change pos() to point to new_pos.
     buffer_pos_ = new_pos;
     buffer_cursor_ = buffer_start_;
-    bool success = ReadBlock();
-    USE(success);
-
-    // Post-conditions: 1, on success, we should be at the right position.
-    //                  2, success == we should have more characters available.
-    DCHECK_IMPLIES(success, pos() == new_pos);
-    DCHECK_EQ(success, buffer_cursor_ < buffer_end_);
-    DCHECK_EQ(success, buffer_start_ < buffer_end_);
+    DCHECK_EQ(pos(), new_pos);
+    ReadBlockChecked();
   }
 
   // Read more data, and update buffer_*_ to point to it.

test/cctest/parsing/test-scanner-streams.cc

@@ -254,6 +254,9 @@ void TestCharacterStream(const char* reference, i::Utf16CharacterStream* stream,
     CHECK_EQU(reference[i], stream->Advance());
   }
   CHECK_EQU(end, stream->pos());
+  CHECK_EQU(i::Utf16CharacterStream::kEndOfInput, stream->Advance());
+  CHECK_EQU(end + 1, stream->pos());
+  stream->Back();
 
   // Pushback, re-read, pushback again.
   while (i > end / 4) {