commit 21ae2b7297e005576afeb9f0230d1f69b3abc857
author npm <npm@chromium.org> Tue Jan 17 15:15:38 2017 -0800
committer Commit bot <commit-bot@chromium.org> Tue Jan 17 15:15:38 2017 -0800
tree 803869c2e20d0b38f9fc0f6f32d41a78a8b560db
parent a715f524e66a1fd8b974f6e2af0db3c23b8c5190

Stop using Encoding array to get unicode from charcode in Type 3 fonts

We should not fill out the m_Encoding for Type3 fonts. This way, we
stop getting garbage characters from text extraction. Guessing that
unicode == charcode (in the absence of ToUnicode) is our best bet.

BUG=pdfium:642

Review-Url: https://codereview.chromium.org/2643543002

core/fpdfapi/font/cpdf_type3font.cpp

diff --git a/core/fpdfapi/font/cpdf_type3font.cpp b/core/fpdfapi/font/cpdf_type3font.cpp
index 8a0a115..9542945 100644
--- a/core/fpdfapi/font/cpdf_type3font.cpp
+++ b/core/fpdfapi/font/cpdf_type3font.cpp
@@ -72,18 +72,8 @@
   }
   m_pCharProcs = m_pFontDict->GetDictFor("CharProcs");
   CPDF_Object* pEncoding = m_pFontDict->GetDirectObjectFor("Encoding");
-  if (pEncoding) {
+  if (pEncoding)
     LoadPDFEncoding(pEncoding, m_BaseEncoding, &m_CharNames, false, false);
-    if (!m_CharNames.empty()) {
-      for (int i = 0; i < 256; i++) {
-        m_Encoding.m_Unicodes[i] =
-            PDF_UnicodeFromAdobeName(m_CharNames[i].c_str());
-        if (m_Encoding.m_Unicodes[i] == 0) {
-          m_Encoding.m_Unicodes[i] = i;
-        }
-      }
-    }
-  }
   return true;
 }