Add logs to check glyphs retrieved from HarfBuzz while shaping complex text 66/168866/2
authorRichard Huang <r.huang@samsung.com>
Wed, 31 Jan 2018 15:53:32 +0000 (15:53 +0000)
committerHeeyong Song <heeyong.song@samsung.com>
Thu, 1 Feb 2018 00:17:43 +0000 (09:17 +0900)
This is needed to get more information for investigating the extra space issue.

Change-Id: Id630367f2af1c8948cf5783f281cc49ad88edb8a

text/dali/internal/text-abstraction/shaping-impl.cpp

index 24c542c..4d5fd6e 100644 (file)
 #include <ft2build.h>
 #include <iostream>
 
+namespace
+{
+
+const static uint8_t U2 = 2u;
+const static uint8_t U3 = 3u;
+const static uint8_t U4 = 4u;
+
+uint32_t GetNumberOfUtf8Bytes( const uint32_t* const utf32, uint32_t numberOfCharacters )
+{
+  uint32_t numberOfBytes = 0u;
+
+  const uint32_t* begin = utf32;
+  const uint32_t* end = utf32 + numberOfCharacters;
+
+  for( ; begin < end; ++begin )
+  {
+    const uint32_t code = *begin;
+
+    if( code < 0x80u )
+    {
+      ++numberOfBytes;
+    }
+    else if( code < 0x800u )
+    {
+      numberOfBytes += U2;
+    }
+    else if( code < 0x10000u )
+    {
+      numberOfBytes += U3;
+    }
+    else if( code < 0x200000u )
+    {
+      numberOfBytes += U4;
+    }
+  }
+
+  return numberOfBytes;
+}
+
+uint32_t Utf32ToUtf8( const uint32_t* const utf32, uint32_t numberOfCharacters, uint8_t* utf8 )
+{
+  const uint32_t* begin = utf32;
+  const uint32_t* end = utf32 + numberOfCharacters;
+
+  uint8_t* utf8Begin = utf8;
+
+  for( ; begin < end; ++begin )
+  {
+    const uint32_t code = *begin;
+
+    if( code < 0x80u )
+    {
+      *utf8++ = code;
+    }
+    else if( code < 0x800u )
+    {
+      *utf8++ = static_cast<uint8_t>(   code >> 6u )           | 0xc0u; // lead byte for 2 byte sequence
+      *utf8++ = static_cast<uint8_t>(   code          & 0x3f ) | 0x80u; // continuation byte
+    }
+    else if( code < 0x10000u )
+    {
+      *utf8++ = static_cast<uint8_t>(   code >> 12u )          | 0xe0u; // lead byte for 2 byte sequence
+      *utf8++ = static_cast<uint8_t>( ( code >> 6u )  & 0x3f ) | 0x80u; // continuation byte
+      *utf8++ = static_cast<uint8_t>(   code          & 0x3f ) | 0x80u; // continuation byte
+    }
+    else if( code < 0x200000u )
+    {
+      *utf8++ = static_cast<uint8_t>(   code >> 18u )          | 0xf0u; // lead byte for 2 byte sequence
+      *utf8++ = static_cast<uint8_t>( ( code >> 12u ) & 0x3f ) | 0x80u; // continuation byte
+      *utf8++ = static_cast<uint8_t>( ( code >> 6u )  & 0x3f ) | 0x80u; // continuation byte
+      *utf8++ = static_cast<uint8_t>(   code          & 0x3f ) | 0x80u; // continuation byte
+    }
+  }
+
+  return utf8 - utf8Begin;
+}
+
+void Utf32ToUtf8( const uint32_t* const utf32, uint32_t numberOfCharacters, std::string& utf8 )
+{
+  utf8.clear();
+
+  uint32_t numberOfBytes = GetNumberOfUtf8Bytes( &utf32[0], numberOfCharacters );
+  utf8.resize( numberOfBytes );
+
+  // This is a bit horrible but std::string returns a (signed) char*
+  Utf32ToUtf8( utf32, numberOfCharacters, reinterpret_cast<uint8_t*>(&utf8[0]) );
+}
+
+}
+
 namespace Dali
 {
 
@@ -208,6 +298,11 @@ struct Shaping::Plugin
 
     hb_shape( harfBuzzFont, harfBuzzBuffer, NULL, 0u );
 
+    std::string currentText;
+    Utf32ToUtf8( text, numberOfCharacters, currentText );
+
+    DALI_LOG_RELEASE_INFO( "Shape: currentText: %s, font: %s, pointSize: %d\n", currentText.c_str(), fontDescription.path.c_str(), static_cast<int>( fontClient.GetPointSize( fontId ) * FROM_266 ) );
+
     /* Get glyph data */
     unsigned int glyphCount;
     hb_glyph_info_t* glyphInfo = hb_buffer_get_glyph_infos( harfBuzzBuffer, &glyphCount );
@@ -266,6 +361,8 @@ struct Shaping::Plugin
         mOffset.PushBack( floor( glyphPositions[i].x_offset * FROM_266 ) );
         mOffset.PushBack( floor( glyphPositions[i].y_offset * FROM_266 ) );
 
+        DALI_LOG_RELEASE_INFO( "glyphIndex: %u, glyph.advance: %f, glyph.xBearing: %f\n", glyphInfo[i].codepoint, floor( glyphPositions[i].x_advance * FROM_266 ), floor( glyphPositions[i].x_offset * FROM_266 ) );
+
         ++i;
       }
     }