Marius Bancila’s Blog

Visual Studio 11 brings many new things for native development, including support for new features from C++11 (unfortunately not all), or ability to write Metro apps with C++/CX including modeling the UI with XAML. In this post I will talk a bit about three favorite features that I noticed immediately after trying VS11 from Windows 8 Developer Preview.

Use of namespaces
Finally, I see namespaces promoted in native code. Yes, it’s C++/CX and they were probably forced to use namespaces for a consistent experience from the various languages that target the Windows Runtime, but it’s a very nice change to the default templates for C++ projects where everything is put in the global namespace. I can only hope they will improve that in this version or the next for standard C++ applications (whether Win32 console apps or MFC apps).

namespace Sample
{
    public ref class MainPage
    {
        public:
            MainPage();
            ~MainPage();
    };
}

UPDATE: looks like I wasn’t clear enough, I’m not saying namespaces is a new C++ feature (duh), I’m saying Visual Studio templates for C++ don’t promote that. Create a Win32 project, an MFC project, and ATL project, there are no namespaces. You’d have to code everything manually, but if you do it, you mess the wizards. So, what I’m saying is that I hope we can see namespaces promoted for other project and item templates too.

Partial classes
I already wrote about partial classes, but I want to reiterate this feature. Partial classes gives you the ability to define a class in several files. This is great because developers and code generator tools such as designers can edit different parts of the same class without interfering one with the other. This feature made it possible for supporting XAML user interfaces in C++/CX Metro applications.

I’m actually wondering why isn’t this already part of standard C++ and I can only wish that the next version (which hopefully will not take another decade to conclude) will include this feature.

Better Syntax Highlighting
Below is a comparison for the same piece of code highlighted by Visual Studio 2010 on the left, and Visual Studio.vNext (11) on the right.

There is hardly any formatting in VS10. However, on the other hand, the highlighting in VS11 is beautiful. User defined types (including library types) are displayed with another color than the built-in types (such as int), including in their definition. STL types (string, vector, etc.) are finally identified as types and displayed with the appropriate color. Also the name of parameters is displayed in italics which makes them easily identifiable. There are other things about the highlighting, but these striking changes.

Not long ago I ran into a COM interop problem that was a bit tricky to fix. So I’m sharing the problem and the solution here in case others encounter the same problem.

I had this native in-proc COM server that initially was built only for x86. It was used in a native MFC application as well as a C# Windows Forms application, where it was added as a COM reference. Both worked nicely. But then I needed to port the MFC app to the x64 platform, so I had to do the same with the in-proc COM server. They both worked correctly, but the managed app, which had to be available also both as 64-bit and 32-bit (even on 64-bit machines), was broken. Eventually I traced the problem to some COM method calls which were a bit atypical, because the arguments to the methods were not COM “native” types, but custom structures.

These structures looked like this:

[uuid(6F13C84D-0E01-48cd-BFD4-F7071A32B49F)] struct S
{
      long a;
      BSTR b;
      long c;
      BSTR d;
      long e;
      BSTR f;
      BSTR g;
      BSTR h;
      BSTR i;
      long j;
      BSTR k;
      long l;
      BSTR m;
      long n;
};

and the COM method signature:

[id(42)] HRESULT GetListOfStructs(SAFEARRAY(struct S)* arrRes);

When you add a COM reference to a .NET assembly, Visual Studio runs tlbimp.exe to generate a .NET assembly with equivalent definitions to the type definitions found in the COM type library. This interop assembly is used to access the COM server. It contains Runtime Callable Wrappers that bridge the two worlds together. By default, the interop assembly, generated in the project’s output folder, is called Interop.<comname>Lib.dll. For instance if the COM server is called NativeCOMServer, the interop assembly is called Interop.NativeCOMServerLib.dll.

The wrapper that was generated in the interop assembly had the following signature for the aforementioned method:

[DispId(42)]
void GetListOfStructs(ref Array arrRes);

and therefore used like this:

System.Array result = null;
obj.GetListOfStructs(ref result);

The call was performed correctly, the native code was executing, but as soon as it was returning an access violation exception was occurring. Thanks to Hans Passant I figured the problem was rooted in the way Visual Studio generates the interop assembly. The generated RCW didn’t know how to handle the custom structure correctly. Probably different padding on the two sides generated the access violation.

The trick was to generate the interop assembly directly, as a custom build step in the COM server project, and include it as an assembly reference in the managed project. Here are the commands for the custom build (must to make sure you have the correct path to the 32-bit and the 64-bit version of tlbimp.exe):

\TlbImp.exe $(TargetPath) /out:$(TargetDir)\NativeCOMLib.Interop.dll /primary  /keyfile:mykey.snk /machine:x86
\TlbImp.exe $(TargetPath) /out:$(TargetDir)\NativeCOMLib.Interop.dll /primary  /keyfile:mykey.snk /machine:x64

The result was a wrapper with methods like this:

[DispId(42)]
void GetListOfStructs(ref S[] arrRes);

which of course had to be called like this:

S[] result = null;
obj.GetListOfStructs(ref result);

To include either one or the other in the C# project I had to manually change the project file, since Visual Studio does not know conditional references, a feature available in MSBuild. I didn’t build the managed app for Any CPU because it had to be able to run as 32-bit on 64-bit machines, so I had two configurations, one for x86 and one fore x64.

    <Reference Condition=" '$(Platform)' == 'x86' " Include="NativeCOMServer.Interop,
               Version=1.0.0.0, Culture=neutral, PublicKeyToken=f5b9312191a42d52, processorArchitecture=x86">
      <SpecificVersion>False</SpecificVersion>
      <HintPath>..\Bin\NativeCOMServer.Interop.dll</HintPath>
    </Reference>
    <Reference Condition=" '$(Platform)' == 'x64' " Include="NativeCOMServer.Interop,
               Version=1.0.0.0, Culture=neutral, PublicKeyToken=f5b9312191a42d52, processorArchitecture=x64">
      <SpecificVersion>False</SpecificVersion>
      <HintPath>..\Bin64\NativeCOMServer.Interop.dll</HintPath>
    </Reference>

But this time the wrappers were able to bridge the call and everything worked smoothly.

The lesson learned is that when you have COM custom structures you should not rely on the way Visual Studio generates the interop assembly. You should build the interop explicitly (maybe as a custom build step, like I did) and include it as an assembly reference to your managed project.

Partial classes are finally available to C++. Sort of. It’s not part of the new C++11 standard, it’s part of the C++/CX language developed by Microsoft for targeting WinRT on Windows 8.

Partial classes mean that you can define a class spanned across several files. Why is this great? Because it allows developers and automatic code generator tools (such as designers) to edit parts of the same class without interfering one with another. WinRT allows C++ developers to write UI in XAML. This could not have been possible without the support for partial classes.

Partial classes:

• are available only for ref classes; native classes are not supported
• are introduced with the partial keyword in all definitions but one

Here is an example:

// foo.private.h
#pragma once

partial ref class foo // <- here the partial keyword is used
{
private:
   int _id;
   Platform::String^ _name;
};

// foo.public.h
#pragma once
#include "foo.private.h"

ref class foo // <- partial keyword is not used here
{
public:
   int GetId();
   Platform::String^ GetName();
};

// foo.cpp
#include "pch.h"
#include "foo.public.h"

int foo::GetId() {return _id;}
Platform::String^ foo::GetName {return _name;}

What happens when you add a new page to a C++ Metro style application? The wizard generates three files: a XAML file and a header and cpp file as the code behind. Let's say the page is called MainPage. In this case the three files are MainPage.xaml (code below is a dummy example), MainPage.xaml.h and MainPage.xaml.cpp.

<UserControl x:Class="DemoApp.MainPage"
    xmlns="http://schemas.microsoft.com/winfx/2006/xaml/presentation"
    xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml"
    xmlns:d="http://schemas.microsoft.com/expression/blend/2008"
    xmlns:mc="http://schemas.openxmlformats.org/markup-compatibility/2006"
    mc:Ignorable="d"
    d:DesignHeight="768" d:DesignWidth="1366">

    <StackPanel Name="firstPanel">
        <Button Name="firstButon" />
    </StackPanel>

</UserControl>

//
// MainPage.xaml.h
// Declaration of the MainPage.xaml class.
//

#pragma once

#include "pch.h"
#include "MainPage.g.h"

namespace DemoApp
{
   public ref class MainPage
   {
      public:
         MainPage();
         ~MainPage();
   };
}

//
// MainPage.xaml.cpp
// Implementation of the MainPage.xaml class.
//

#include "pch.h"
#include "MainPage.xaml.h"

using namespace Windows::UI::Xaml;
using namespace Windows::UI::Xaml::Controls;
using namespace Windows::UI::Xaml::Data;
using namespace DemoApp;

MainPage::MainPage()
{
   InitializeComponent();
}

MainPage::~MainPage()
{
}

You can notice that the objects firstPanel and firstButton are not defined in the header for MainPage and second, MainPage.xaml.h includes MainPage.g.h. So what is this? This is a designer generated file, which together with MainPage.g.cpp completes the definition of the MainPage ref class. These files are not generated until you start a build. After you do that you can find them in the output folder (Debug or Release for instance). This is how they look:

#pragma once
//------------------------------------------------------------------------------
//     This code was generated by a tool.
//
//     Changes to this file may cause incorrect behavior and will be lost if
//     the code is regenerated.
//------------------------------------------------------------------------------

namespace Windows {
    namespace UI {
        namespace Xaml {
            namespace Controls {
                ref class StackPanel;
                ref class Button;
            }
        }
    }
}

namespace DemoApp
{
    partial ref class MainPage : public Windows::UI::Xaml::Controls::UserControl,
                                                     public Windows::UI::Xaml::Markup::IComponentConnector
    {
    public:
        void InitializeComponent();
        void Connect(int connectionId, Platform::Object^ pTarget);

    private:
        Windows::UI::Xaml::Controls::StackPanel^ firstPanel;
        Windows::UI::Xaml::Controls::Button^ firstButon;
    };
}

//------------------------------------------------------------------------------
//     This code was generated by a tool.
//
//     Changes to this file may cause incorrect behavior and will be lost if
//     the code is regenerated.
//------------------------------------------------------------------------------
#include "pch.h"

#include "MainPage.xaml.h"

using namespace Windows::Foundation;
using namespace Windows::UI::Xaml;
using namespace Windows::UI::Xaml::Controls;
using namespace Windows::UI::Xaml::Markup;
using namespace MyDemoApplication1;

void MainPage::InitializeComponent()
{
    // Call LoadComponent on ms-resource://DemoApp/Files/MainPage.xaml
    Windows::UI::Xaml::Application::LoadComponent(this, ref new Windows::Foundation::Uri("ms-resource://DemoApp/Files/MainPage.xaml"));

    // Get the StackPanel named 'firstPanel'
    firstPanel = safe_cast<Windows::UI::Xaml::Controls::StackPanel^>(static_cast<IFrameworkElement^>(this)->FindName("firstPanel"));

    // Get the Button named 'firstButon'
    firstButon = safe_cast<Windows::UI::Xaml::Controls::Button^>(static_cast<IFrameworkElement^>(this)->FindName("firstButon"));

}

void MainPage::Connect(int connectionId, Platform::Object^ pTarget)
{
}

The following image illustrates the grouping of these files:

MainPage.xaml is a XAML files, which can be edited both by the designer or manually by the developer (basically with the designer is still the developer that models the UI). The other files are C++/CX files. MainPage.xaml.h and MainPage.xaml.cpp are the files the developer writes, while MainPage.g.h and MainPage.g.cpp are edited by the designer. Do not modify the code in these files, because you could either mess up the designer, or all your changes would get lost when the file is regenerated.

At the beginning of this year, Microsoft announced a “C++ renaissance”. Quoting from the description of a Channel 9 video with Craig Symonds and Mohsen Agsen:

C++ is currently undergoing a renaissance. This means that, by definition, the language, compilers and compositional tooling are evolving and coalescing into a state that maximizes native developer efficiency, productivity, and creativity across hardware and software domains.

Everybody agrees that Microsoft made C++ a sort of second class citizen in the past years, while the company invested a lot in the .NET framework. Many developers have switched from native development to managed (.NET) simply because it offers a more productive environment. And the postponing of the ISO standard committee in releasing the new C++0x standard only made things worse.

However, with the completion of the new C++ standard this year, Microsoft, apparently, plans to change that, and make C++ again appealing to developers. They already made C++0x features available in the VS2010 C++ compiler and are working on implementing most of the rest for Visual Studio vNext. They are also investing in tools (now labeled Application Lifecycle Management), and for instance are bringing intellisence to C++/CLI. One of the most important areas of development is parallelism, where they are developing the PPL and Agents libraries and now the C++ AMP that they just announced. And also recently the Kinect for Windows SDK beta that provides Kinect capabilities to developers who build applications with C++ (and other laguanges). And in the mean time they hired Erich Gamma in the Visual Studio team.

But this is not enough in my opinion. Improvements in language and tools are an important part, but not everything. It is equally necessary for Microsoft to evangelize it, using any necessary means. Unless they can spread the word, the work might pass unnoticed. To be honest, I was very reluctant about this part, half an year ago, when they announced the “renaissance”. However, looking back at what they done I’d say they are on the right track. Of course, there is still a lot of work to match the “advertising” effort put into .NET. But right now C++ is getting more attention at conferences such as PDC or TechEd, or their publishing assets, such as Channel 9, MSDN or their team blogs. So I tried to assemble a collection of videos, blogs, books and code samples related to C++ or native development that they published since the announcement of the renaissance. So far it looks good, in my opinion.

Channel 9
E2E: Herb Sutter and Erik Meijer – Perspectives on C++
Craig Symonds and Mohsen Agsen: C++ Renaissance
Windows 7 Taskbar Integration for MFC Applications
Tony Goodhew: VC++ Developer Communication – Questions and Answers
Talkin’ C++ with Kate Gregory
MVP Summit 2011: Meet C++ MVPs Angel, PJ, Tom and Sheng
Talkin’ C++ with Alon, Marius, Bruno, and Jim
Talkin’ C++ with Boris Jabes: C++ Intellisense, Game Development, and Boris Faces His Demons
Application Restart and Recovery on Windows 7 in Native Code
Parallel Programming for C++ Developers: Tasks and Continuations, Part 1 of 2
Parallel Programming for C++ Developers: Tasks and Continuations, Part 2 of 2
Conversation with Herb Sutter: Perspectives on Modern C++(0x/11)
First Look: New ALM Tools for VC++ Developers
Modern Native C++ Development for Maximum Productivity
Mohsen Agsen – C++ Today and Tomorrow
Herb Sutter: C++ Questions and Answers
Herb Sutter – Heterogeneous Computing and C++ AMP
Daniel Moth: Blazing-fast code using GPUs and more, with C++ AMP
C9 Lectures: Stephan T Lavavej – Advanced STL, 1 of n
C9 Lectures: Stephan T Lavavej – Advanced STL, 2 of n
C9 Lectures: Stephan T Lavavej – Advanced STL, 3 of n
C9 Lectures: Stephan T Lavavej – Advanced STL, 4 of n
C9 Lectures: Stephan T Lavavej – Advanced STL, 5 of n

Visual C++ Team Blog
Grr… My VC++ Project Is Building Slower in VS2010. What Do I Do Now? (A Step by Step Guide)
C++/CLI IntelliSense in Visual Studio vNext
Exception Boundaries: Working With Multiple Error Handling Mechanisms
Troubleshooting Tips for IntelliSense Slowness
Build Related Improvement in VS2010 SP1
Converting An MFC Ribbon To Designer Format
Enforcing Correct Concurrent Access of Class Data

Parallel Programming in Native Code Blog
Sorting in PPL
How to pick your parallel sort?
The Concurrency Runtime and Visual C++ 2010: Lambda Expressions
The Concurrency Runtime and Visual C++ 2010: Automatic Type Deduction
The Concurrency Runtime and Visual C++ 2010: The decltype Type Specifier
The Concurrency Runtime and Visual C++ 2010: Rvalue References
The Concurrency Runtime and Visual C++ 2010: Transporting Exceptions between Threads
Building Responsive GUI Applications with PPL Tasks

MSDN Magazine
Writing a Debugging Tools for Windows Extension
Writing a Debugging Tools for Windows Extension, Part 2: Output
Writing a Debugging Tools for Windows Extension, Part 3: Clients and Callbacks
Agile C++ Development and Testing with Visual Studio and TFS

Books & Publications
Parallel Programming with Microsoft Visual C++
The Visual C++ Weekly

Code & Samples
Code samples for the Concurrency Runtime and Parallel Pattern Library in Visual Studio 2010
Bing Maps Trip Optimizer
Hilo: Developing C++ Applications for Windows 7
All-in-One Code Framework

Microsoft has recently announced at the AMD Fusion Developer Summit the introduction of a new technology called C++ Accelerated Massive Parallelism or shortly C++ AMP that helps C++ developers use the GPU for parallel programming. The new technology will be part of Visual C++, integrated with full support (edit, build, debug, profile) in the next version of Visual Studio. It is built in modern C++ on top of DirectX. It will provide an STL-like library, part of the Concurrency namespace, delivered in a amp.h header.

S.Somasegar (Senior Vice President of the Developer Division) said:

By building on the Windows DirectX platform, our implementation of C++ AMP allows you to target hardware from all the major hardware vendors. We expect that it will be part of the next Visual C++ compiler and fully integrated in the next release of Visual Studio experience.

You can read more about it here:

• Targeting Heterogeneity with C++ AMP and PPL
• C++ Accelerated Massive Parallelism
• Introducing C++ Accelerated Massive Parallelism (C++ AMP)

Looking forward for the first CTP and samples.

UPDATE
Herb Sutter introduces C++ AMP at the AMD Fusion Developer Summit 11

Daniel Moth digs deeper into C++ AMP with code samples and more

Herb Sutter

How long is your password? How long it will take a 100,000,000 GPU cores running at what, a million attempts per second to crack your password just by brute force? That where almost a kid can write that. An that is just a tiny example of how game changing this is.

When you run your (unmanaged/C++) application in debugger, you see at the end a report of memory leaks (if any are detected). Something like this:

Detected memory leaks!
Dumping objects ->
f:\dd\vctools\vc7libs\ship\atlmfc\src\mfc\strcore.cpp(141) : {381} normal block at 0x001FFC30, 54 bytes long.
Data: < x > 0C 00 B9 78 12 00 00 00 12 00 00 00 01 00 00 00
d:\marius\vc++\memoryleakstest\memoryleakstestdlg.cpp(163) : {380} normal block at 0x001FFBF0, 4 bytes long.
Data: <@ > 40 FC 1F 00
f:\dd\vctools\vc7libs\ship\atlmfc\src\mfc\thrdcore.cpp(306) : {374} client block at 0x001FFA38, subtype c0, 68 bytes long.
a CWinThread object at $001FFA38, 68 bytes long
Object dump complete.

Some of them can be fixed immediately, because when you double click on them Visual Studio will take you to the line where the allocation was made. Some of them are harder to spot, because Visual Studio is not able to do the same. Question is how do you find the source of those allocations? Luckily, there is this global variable called _crtBreakAlloc, that can be used to force the debugger to stop the execution when a certain block is allocated.

In order to use that, you should follow several steps:

• First, you have to find a reproducible sequence that produces the same memory allocation number. When memory blocks are allocated, they are identified with a number, called allocation number. This number is reported by Visual Studio in brackets when it lists the memory leaks (e.g. {381}).
• Second, you have to put a breakpoint somewhere in your program to force a stop at the beginning of the execution. That means you could use function main(), your CWinApp derived class constructor, function InitInstance() or other, depending on your application type and how early in the execution of the program the block that leaks is allocated. The smaller the allocation number is, the earlier in the execution the allocation occurs.
• Run your program in debugger.
• When the program stops (at the first breakpoint) open the Watch window and add the following expression: {,,msvcr90d.dll}_crtBreakAlloc in the Name column. In the Value column (which by default should have the value -1) write the allocation number. Take notice that msvcr90d.dll (which is the DLL that contains the C++ runtime library) is specific to Visual Studio 2008 (and is the debug version). If you use another version of Visual Studio, you have to use the appropriate DLL.
• Continue debugging.
• When the block identified by the allocation number set in the Watch window is allocated, the debugger will stop the execution and jump to a line from dbgheap.c.
  
  In order to see the line in your code that triggered the allocation, open the Call Stack window and find, from top down, the first function from your own code.
  
  That will lead you to the source of the memory leak.

To read more about this topic see:
How to: Set Breakpoints on a Memory Allocation Number
How to use _crtBreakAlloc to debug a memory allocation

I was doing some development in Visual Studio 2010 Beta 2 and I had to add some references to my projects. When I opened the Add Reference dialog I realized something was wrong: it was working very fast. Since I’m using Visual Studio 2008 for every day development I am used to wait tens of seconds before the dialog loads all the references and only after that I can select what I want. But in Visual Studio 2010 it popped up instantly and all the tabs were browse able at the same speed. This was not normal. Usually new versions are slower that older ones (and I suspect Visual Studio 2010 has such features), but Add Reference dialog works great.

Then I browsed the web I came across this post from Scott Guthrie who explain what has changed:

• default active tab when the Add Reference dialog is opened is now Projects, and not .NET
• .NET and COM tabs load asynchronously in worker threads, populating the lists as references are discovered, without blocking the UI thread, which means you can browse through the references as soon as you open the tab

There are only two things that I can say: first is that I’m impressed. I now get instantly what it used to take maybe half a minute. Second is that I’m puzzled that it took so many years to implement that. Anyway, good work.

I have modified my VSBuildStatus add-in to work with Visual Studio 2005. Basically I had to remove the LINQ stuff to break the dependency on the version 3.5 of the .NET framework.

The new version (1.0.3) that supports Visual Studio 2005 is available here. Let me know if you have any problems with it. Especially in VS 2005.

This is the record number of error I ever got in a project: 50917.

After compiling that VC# project (with T4), Visual Studio 2008 was a little bit overwhelmed. It started to report out of memory errors and failed to build after fixing the error. A restart was, of course, the cure.