From f765805983d157e9a21d7202ba49148f24d226f3 Mon Sep 17 00:00:00 2001
From: Tom Sepez <tsepez@chromium.org>
Date: Wed, 8 Aug 2018 22:20:29 +0000
Subject: Remove unused codepage from WideString::GetWideString().

Consolidate code into sole caller, which always retrieves a string
from the "default ANSI code page". Add unit test for this code path.
Fix sign extension bug giving bogus 0xffffff80 codepoint uncovered
by unit test on non-windows. Note windows result is different because
the system conversion routine does actually work.

Change-Id: I1c1246b203ee66e9ff0b6fccf97594788bcd0fca
Reviewed-on: https://pdfium-review.googlesource.com/39730
Reviewed-by: Lei Zhang <thestig@chromium.org>
Commit-Queue: Tom Sepez <tsepez@chromium.org>
---
 core/fxcrt/fx_system.cpp           |  6 ++--
 core/fxcrt/widestring.cpp          | 59 ++++++++++----------------------------
 core/fxcrt/widestring.h            |  3 --
 core/fxcrt/widestring_unittest.cpp | 21 ++++++++++++++
 4 files changed, 39 insertions(+), 50 deletions(-)

diff --git a/core/fxcrt/fx_system.cpp b/core/fxcrt/fx_system.cpp
index 27cbd65f3c..061ca6d82d 100644
--- a/core/fxcrt/fx_system.cpp
+++ b/core/fxcrt/fx_system.cpp
@@ -203,6 +203,7 @@ int FXSYS_WideCharToMultiByte(uint32_t codepage,
   }
   return len;
 }
+
 int FXSYS_MultiByteToWideChar(uint32_t codepage,
                               uint32_t dwFlags,
                               const char* bstr,
@@ -211,9 +212,8 @@ int FXSYS_MultiByteToWideChar(uint32_t codepage,
                               int buflen) {
   int wlen = 0;
   for (int i = 0; i < blen; i++) {
-    if (buf && wlen < buflen) {
-      buf[wlen] = bstr[i];
-    }
+    if (buf && wlen < buflen)
+      buf[wlen] = reinterpret_cast<const uint8_t*>(bstr)[i];
     wlen++;
   }
   return wlen;
diff --git a/core/fxcrt/widestring.cpp b/core/fxcrt/widestring.cpp
index 7dd1c30eb0..97073f170d 100644
--- a/core/fxcrt/widestring.cpp
+++ b/core/fxcrt/widestring.cpp
@@ -277,43 +277,6 @@ Optional<WideString> TryVSWPrintf(size_t size,
   return {str};
 }
 
-#ifndef NDEBUG
-bool IsValidWideCodePage(uint16_t codepage) {
-  switch (codepage) {
-    case FX_CODEPAGE_DefANSI:
-    case FX_CODEPAGE_ShiftJIS:
-    case FX_CODEPAGE_ChineseSimplified:
-    case FX_CODEPAGE_Hangul:
-    case FX_CODEPAGE_ChineseTraditional:
-      return true;
-    default:
-      return false;
-  }
-}
-#endif
-
-WideString GetWideString(uint16_t codepage, const ByteStringView& bstr) {
-#ifndef NDEBUG
-  ASSERT(IsValidWideCodePage(codepage));
-#endif
-
-  int src_len = bstr.GetLength();
-  int dest_len = FXSYS_MultiByteToWideChar(
-      codepage, 0, bstr.unterminated_c_str(), src_len, nullptr, 0);
-  if (!dest_len)
-    return WideString();
-
-  WideString wstr;
-  {
-    // Span's lifetime must end before ReleaseBuffer() below.
-    pdfium::span<wchar_t> dest_buf = wstr.GetBuffer(dest_len);
-    FXSYS_MultiByteToWideChar(codepage, 0, bstr.unterminated_c_str(), src_len,
-                              dest_buf.data(), dest_len);
-  }
-  wstr.ReleaseBuffer(dest_len);
-  return wstr;
-}
-
 }  // namespace
 
 namespace fxcrt {
@@ -902,14 +865,22 @@ size_t WideString::Replace(const WideStringView& pOld,
 }
 
 // static
-WideString WideString::FromLocal(const ByteStringView& str) {
-  return FromCodePage(str, 0);
-}
+WideString WideString::FromLocal(const ByteStringView& bstr) {
+  int src_len = bstr.GetLength();
+  int dest_len = FXSYS_MultiByteToWideChar(
+      FX_CODEPAGE_DefANSI, 0, bstr.unterminated_c_str(), src_len, nullptr, 0);
+  if (!dest_len)
+    return WideString();
 
-// static
-WideString WideString::FromCodePage(const ByteStringView& str,
-                                    uint16_t codepage) {
-  return GetWideString(codepage, str);
+  WideString wstr;
+  {
+    // Span's lifetime must end before ReleaseBuffer() below.
+    pdfium::span<wchar_t> dest_buf = wstr.GetBuffer(dest_len);
+    FXSYS_MultiByteToWideChar(FX_CODEPAGE_DefANSI, 0, bstr.unterminated_c_str(),
+                              src_len, dest_buf.data(), dest_len);
+  }
+  wstr.ReleaseBuffer(dest_len);
+  return wstr;
 }
 
 // static
diff --git a/core/fxcrt/widestring.h b/core/fxcrt/widestring.h
index dc5dd23428..c51c74506b 100644
--- a/core/fxcrt/widestring.h
+++ b/core/fxcrt/widestring.h
@@ -61,9 +61,6 @@ class WideString {
   ~WideString();
 
   static WideString FromLocal(const ByteStringView& str) WARN_UNUSED_RESULT;
-  static WideString FromCodePage(const ByteStringView& str,
-                                 uint16_t codepage) WARN_UNUSED_RESULT;
-
   static WideString FromUTF8(const ByteStringView& str) WARN_UNUSED_RESULT;
   static WideString FromUTF16LE(const unsigned short* str,
                                 size_t len) WARN_UNUSED_RESULT;
diff --git a/core/fxcrt/widestring_unittest.cpp b/core/fxcrt/widestring_unittest.cpp
index 9d38aa45e9..160dc929db 100644
--- a/core/fxcrt/widestring_unittest.cpp
+++ b/core/fxcrt/widestring_unittest.cpp
@@ -1025,6 +1025,27 @@ TEST(WideString, ToDefANSI) {
                          .ToDefANSI());
 }
 
+TEST(WideString, FromLocal) {
+  EXPECT_EQ(L"", WideString::FromLocal(ByteStringView()));
+#if _FX_PLATFORM_ == _FX_PLATFORM_WINDOWS_
+  const wchar_t* kResult =
+      L"x"
+      L"\u20ac"
+      L"\u00ff"
+      L"y";
+#else
+  const wchar_t* kResult =
+      L"x"
+      L"\u0080"
+      L"\u00ff"
+      L"y";
+#endif
+  EXPECT_EQ(kResult, WideString::FromLocal("x"
+                                           "\x80"
+                                           "\xff"
+                                           "y"));
+}
+
 TEST(WideStringView, FromVector) {
   std::vector<WideStringView::UnsignedType> null_vec;
   WideStringView null_string(null_vec);
-- 
cgit v1.2.3