Noissue - Add tracing function 'ToHexLiteral()'

src/plugin/PluginDebug.cpp

 /*
  * This file is part of Adblock Plus <https://adblockplus.org/>,
  * Copyright (C) 2006-2015 Eyeo GmbH
  *
  * Adblock Plus is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 3 as
  * published by the Free Software Foundation.
  *
  * Adblock Plus is distributed in the hope that it will be useful,
  * but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ skipping 238 matching lines @@
     srcTrunc = srcTrunc.Left(67) + L"..." + srcTrunc.Right(30);
   }
 
   CString blocking;
   blocking.Format(L"Ignored  %-12s  %s  %s", ToCString(type), domain.empty()? L"-" : ToCString(domain), srcTrunc);
 
   DebugResultLegacy(blocking);
 }
 
 #endif // ENABLE_DEBUG_RESULT_IGNORED
+
+namespace
+{
+  /*
+   * To convert a pointer to a hexadecimal number, we need an integral type that has the same size as that of the pointer.
+   */
+#if defined(_WIN64)
+  typedef long long voidIntegral;

[sergei, 2015/12/07 11:04:02]

what about uint64_t and uint32_t instead of `long long` and `long`?

[Eric, 2015/12/07 13:06:57]

Done.

Do not, however, ask me to remove the 'static_assert' statements. Anything
that's got behavior dependent upon a 'reinterpret_cast' needs sanity checking.

+  static_assert(sizeof(void*)==sizeof(voidIntegral),"WIN64: sizeof(long long) is not the same as sizeof(void*)");
+#elif defined(_WIN32)
+  typedef long voidIntegral;
+  static_assert(sizeof(void*)==sizeof(voidIntegral),"WIN32: sizeof(long) is not the same as sizeof(void*)");
+#else
+#error Must compile with either _WIN32 or _WIN64
+#endif
+}
+
+std::wstring ToHexLiteral(void const* p)
+{
+  std::wstringstream ss;
+  ss << L"0x";
+  ss.width(sizeof(p) * 2);
+  ss.fill(L'0');
+  ss << std::hex << reinterpret_cast<voidIntegral>(p);
+  return ss.str();

[sergei, 2015/12/07 11:04:02]

Do we really need it so complicated? It seems `basic_ostream& operator<<( const
void* value );` does the same thing. The following should be enough, shouldn't
it?
{
  std::wstringstream ss;
  ss << p;
  return ss.str();
}

[Oleksandr, 2015/12/07 12:00:50]

Alternatively there's also StringCbPrintf (aka sprintf) with "%p". Anyway, I
think any of these can be added when we need actual tracing. 

[Eric, 2015/12/07 13:06:57]

I doesn't do the same thing. The formatting is different.

And, in addition, it's implementation-defined behavior. Insofar as I know,
there's no standard for stream formatting of 'void *', nor does Microsoft
provide documentation for it. Since you have previously expressed discomfort
with relying upon undocumented behavior (see 'IDispatch'), I would think you
would accept the use of standard, documented C++ to obtain reliable behavior.

[Eric, 2015/12/07 13:06:57]

It doesn't save code to use a C-style function that requires explicit buffer
management.
We already need actual tracing. I have code ready to post for review that's
depends upon other reviews that have taken interminably long because of endless
dithering and niggling about very small things. For example, this one
https://codereview.adblockplus.org/29323561/ which is almost four months old
now. That tracing code is still not ready to post because we haven't managed to
finish that review yet.

[Eric, 2015/12/15 16:31:16]

Actual tracing in new review: https://codereview.adblockplus.org/29332677/

+}