Windows 7 rocks

Ok I know this blog is not about Windows but I thought I’d share the love.

To me it looks like as is someone has combed through the entire user experience and fixed most of the stuff that bothered me.

The taskbar change alone are good enough for the upgrade. E.g. the taskbar feels very natural when set in vertical position (it didn’t work well in this configuration before). The icons in tray and the status bar can be dragged around and it actually shows download progress in the IE taskbar icon itself (See the green progress in the screenshot below). I can even close an app by hovering over it’s icon on the taskbar (no need to restore it and then travel across the entire screen and hit close button). The taskbar is very accessible through keyboard, hit Win+T and taskbar is in focus and then shift through the icons using the arrow keys.

The greatest shocker was when I connected my laptop to my TV and it instantly popped up the following, it then correctly picked up my TV’s resolution (1080p) and matched screen resolution to that. Almost felt magical…

I could go on and on (like the awesome window management features) but you can head over to Tim’s blog and read more details there.

Back to basics: Why use garbage collection

This is a part of a series of post on GC. See the index here

Garbage Collection does exactly what it’s more fancier name “Automatic dynamic memory management” suggests. As I discussed in my last post on Memory Allocation dynamic memory is hard to manage and GC attempts to do that automatically and relieves the coder from the hard task.

GC basically attempts to take care of two basic scenarios remove garbage and avoid dangling pointers. They are very inter-related but are different scenarios

Garbage Collect

Consider a typical object reference graph. In the graph every rectangle is an object and the arrows denote object references. So A is an object which references B. For a program to be able to use an object it should be in one of these reference chains. The reference chain start from what is called Root and are typically references held in registers, on stack as local variable or global variables.

Let’s assume that due to some operation, A relinquishes the reference to B and the graph becomes something like this…

 

Now B and hence C is not reachable from any valid root in the program and hence have become Garbage (un-reachable). The programmer must ensure to follow all reference and free (de-allocate them). One of the duty of a GC system is to automate this process by tracking down (using various algorithms) such objects and reclaim the memory used by them automatically. So in a GC system when the reference is broken it will figure out that B and hence C is not reachable and will de-allocate them.

Hanging/dangling reference

Let’s consider another object graph which is similar to the one above but in addition to A, another object A’ also has a reference to B (or in other words B is shared between them)

Even here after some operation object A doesn’t need reference to B. The programmer does what he thinks is right and de-allocates B and C

However, A’ still has a references to B and hence that reference is now hanging or in more specific term pointing to invalid memory and would typically return unpredictable result when accessed. The key here is the unpredictable behavior. It is not necessary that program will crash. Unless the memory location in B is re-used it will seem to have valid data and de-references from A’ will work fine. So the failure will come up in un-expected ways and totally un-related places in the program and will make locating the root cause extremely hard.

GC helps by automatically taking care of both of the above scenarios and ensuring that the system doesn’t land up in either of them.

How important is GC

If GC is so cool then why doesn’t all systems use GC. There are multiple reasons but with newer technology most of them are going away. One of the primary reasons of not using GC is the performance overhead. This makes GC a less lucrative deal for real-time systems, device drivers and even gaming platforms. However, there are examples where GC is used even for these systems, e.g. .NET Compact Framework (the team I work for) is used with full GC support very successfully in XNA games on Xbox.

However, on some systems like functional languages which relies a lot on closures and deferred execution of those, it makes execution flow very un-predictable and hence GC becomes almost mandatory.

Back To Basics: Memory allocation, a walk down the history

This is a part of a series of post on GC. See the index here

Most languages we deal with today support a whole bunch of different memory allocation options. We have an option to use static allocation, stack-allocation and heap allocation. However, same was not true in the pre-historic era, life was simple then and the earlier languages just supported static allocation, then slowly stack and then heap allocation came by and today we have come to expect automatic memory management from most new languages.

However, that didn’t really mean that there was no advantage is the simplistic approach taken by these early languages.

Static allocation

This is the simplest allocation option and was supported by initial languages like Fortran (early 50’s). All memory was allocated (reserved) during compile time and variable names were bound to the target memory and frozen. This had some serious limitation like variable memory couldn’t be supported. In case memory usage might vary the developer simply allocated the largest size he anticipated and used from it.

This also meant that size of all Data Structures were known at runtime (e.g. fixed length arrays, no link-lists). Since name binding happened statically, simple things we have come to expect over time like recursion was not possible because the second instantiation of a method would mean that same variables are accessed and since names are bound to addresses statically they would over-write data.

Even though this seems to be pretty restrictive this had some advantages. The most obvious is robustness and predictable behavior. Since there are no runtime memory allocation, there is no possibility of running out of memory (so no OutOfMemoryException :) ). The execution was also very fast because there is no memory to manage and no stack/heap to tear down. Moreover the compiler could emit direct memory accesses instruction without going through any indirection (as symbol->address mapping doesn’t change).

Stack Allocation

The first languages to bring in stack allocation was Algol in late 50’s. This worked past some limitations of static allocation by supporting memory frames that got pushed onto the system stack as procedures were called. What this meant was that based on the input parameters of a method it could push a different sized frame and hence had variable memory allocation (atleast for method local variables). However, return value had to be of fixed size (fixed by compile time) because the caller would need to know how much was left behind by callee on the stack.

Recursion was possible because every invocation would have it’s exclusive frame, and hence when a method returned the frame would unwind (pop) and memory allocated by it would be no longer available to the callee.

Obviously with this came reduced robustness because based on control flow the program could get out of stack and the additional stack management overhead reduced speed (though not significantly).

Even though it wasn’t much but it still was a huge improvement…

Heap Allocation

Funny to call it so but this is the state of the art in memory allocation :) and has remained so for a long time (with some improvements).

With heap data could be allocated in chunks of variable size and later de-allocated. There is no ordering requirement in-between the allocation and de-allocation unlike the stack frames where callee memory gets deallocated before the caller. So now it was possible to allocate and pass variable sized memory around (e.g. from Callee to Caller). It was easier to manage memory better, e.g. by using growable non-contiguous lists over arrays, model data closer to their real life existence like a tree.

With the heap allocation life seemed simpler at first but actually became harder. Robustness nose-dived. Memory allocated dynamically had to be managed carefully because if allocated memory is not de-allocated after it’s use is over, it becomes garbage and un-available (called memory leak) and slowly runs out of space. At the end the dreaded out of memory scenario comes up where there is no more memory left to be allocated to the requestor and the program fails to continue. Heap allocation is also a costly operation and can have severe perf implications.

Even with all the tooling support it still remains hard to locate memory leaks and constitutes one of the harder classes of bugs to be tackled.

Garbage Collection

LISP in 1959 started a new technique called Automatic Memory Management known by the more popular term Garbage Collection (GC). This tries to address the memory leak issue by needing explicit memory allocation but removing the need to de-allocate memory. Instead it provides language facilities that actually hunt down the un-used memory and automatically free them without user-code intervention.

As you might guess that even though GC increases robustness and reduces coding overhead significantly it also has performance implications. This makes using GC in hard in a whole bunch of scenarios. E.g. you typically do not expect to see a GC being used for a real-time system or a system driver (but there are exceptions to this as well); but you do expect to see it in a web-development platform that manages huge amount of memory and can sometimes live with a bit of response delay when the GC is actually running.

Putting it all together

Most modern programming languages support all three forms of storage. E.g. C++ or C# supports all 3.

Even though all of them might be available at the same time, the choice of usage is key to ensure robustness and meet performance goals of a program. Unfortunately a lot of developers take a casual approach to it and later land up into trouble when the memory foot-print become larger and harder to handle.

Consider the following pseudo-code

void foo()
{
    MyType m1;                                   // Option-1 stack allocation
    MyType* m2 = AllocateOnHeap(sizeof(MyType)); // Option-2 heap allocation
    ...
    if (SomeCond)
       foo();     // recursive call    
    ...
    DeAllocate(m);
}

-

Here foo is called in recursion. If MyType is of significant size and we use option-1 of using stack allocation then the depth of recursion supported by the method will be significantly reduced because each function call will need additional stack space to accommodate MyType. However, if we use heap (Option-2) then the stack needed will be significantly lower (increase in depth of recursion) but since heap allocation is much slower the speed will suffer. Based on the size of MyType, the need of the program the allocation would have to be carefully chosen.

Based on target domain different languages have chosen

When the spell checkers attack

Today Boing Boing had a post about the Cupertino effect in which spell-checkers erroneously change valid words. I had worked on Adobe Acrobat and Adobe FrameMaker's spell checking feature and do happen to appreciate the headache with getting them right.

I face Cupertino effect most often while typing in Indian names because most word-processors don't recognize them. For my wife's name Somtapa the offered spelling is Stomata followed by Sumatra. So I started calling her Stomata until it got onto her nerves and something very similar to the photo above happened to me

Is interlocked increment followed by comparison thread safe?

Sorry about the blog title, my imagination failed me :(.

In our internal alias someone asked the question "Is the following thread safe"

if(Interlocked.Increment(ref someInt) == CONSTANT_VAL)
{
    doSomeStuff();
}

My instant reaction was no because even though the increment is done in a thread safe way using System.Threading.Interlocked class, the comparison that follows is not safe.

My reasoning was that the "if" expression can be broken down to the following operations

1. Fetch of someInt
   
2. Increment operation
   
3. Write back of someInt
   
4. Comparison

The first 3 are done inside the Increment method and it provides concurrency protection and hence cannot be interleaved by another instruction.

So if two threads are running in parallel (one marked in red and the other in green) I assumed that the following interleaving is possible

1. someInt is 4 and CONSTANT_VAL is 5
   
2. Fetch of someint      -> someInt ==4
   
3. Increment operation   -> someInt ==5
   
4. Write back of someint -> someInt ==5
   
5. Fetch of someint      -> someInt == 5
   
6. Increment operation   -> someInt == 6
   
7. Write back of someint -> someInt == 6
   
8. comparison            -> compare 6 & CONSTANT_VAL
   
9. comparison            -> compare 6 & CONSTANT_VAL

This means that the comparison of both thread will fail.

However, someone responded back that I was wrong as the return value is being used and not the written back value. This made me do some more investigation.

If I see the JITted code then it looks like

if (Interlocked.Increment(ref someInt) == CONSTANT_VAL)
00000024  lea         ecx,ds:[002B9314h] 
0000002a  call        796F1221 
0000002f  mov         esi,eax 
00000031  cmp         esi,5

The call (at 0x000002a) is to the native code inside CLR which in turn calls Win32 api (InterlockedIncrement).

However, the last 2 lines are the interesting ones. Register EAX contains the return value and comparison is happening against that and CONSTANT_VAL. So even if the second thread had already changed the value of someInt it doesn’t have any effect as the return of the first increment is being used and not the safeInt value in memory. So first comparison (step 8 above) will actually compare CONSTANT_VAL against 5 and succeed.

C/C++ Compile Time Asserts

The Problem

Run time asserts are fairly commonly used in C++. As the MSDN documentation for assert states

"(assert) Evaluates an expression and, when the result is false, prints a diagnostic message and aborts the program."

There is another type of asserts which can be used to catch code issues right at the time of compilation. These are called static or compile-time asserts. These asserts can be used to do compile time validations and are very effectively used in the .NET Compact Framework code base.

E.g. you have two types Foo and Bar and your code assumes (may be for a reinterpret_cast) that they are of the same size. Now being in separate places there is always a possibility that someone modifies one without changing the other and that results in some weird bugs. How do you express this assumption in code? Obviously you can do a run-time check like

assert(sizeof(foo) == sizeof(bar));

If that code is not hit during running this assert will not get fired. This might be caught in testing later. However, if you notice carefully all of the information is available during compilation (both the type and the sizeof is resolved while compilation). So we should be able to do compile time validation, with-something to the effect

COMPILE_ASSERT(sizeof(int) == sizeof(char));

This should be tested during compilation and hence whether the code is run or not the assert should fail.

The Solution

There are many ways to get this done. I will discuss two quick ways

Array creation

You can create a MACRO expression as follows

#define COMPILE_ASSERT(x) extern int __dummy[(int)x]

This macro works as follows

// compiles fine 
COMPILE_ASSERT(sizeof(int) == sizeof(unsigned)); 
// error C2466: cannot allocate an array of constant size 0
COMPILE_ASSERT(sizeof(int) == sizeof(char)); 

The first expression gets expanded to int __dummy[1] and compiles fine, but the later expands to int __dummy[0] and fails to compile.

The advantage of this approach is that it works for both C and C++, however, the failure message is very confusing and doesn't indicate what the failure is for. It is left to the developer to visit the line of compilation failure to see that it's a COMPILE_ASSERT.

sizeof on incomplete type

This approach works using explicit-specialization of template types and the fact that sizeof of incomplete types fail to compile.

Consider the following

namespace static_assert
{
    template <bool> struct STATIC_ASSERT_FAILURE;
    template <> struct STATIC_ASSERT_FAILURE<true> { enum { value = 1 }; };
}

Here we defined the generic type STATIC_ASSERT_FAILURE. As you see the type is incomplete (no member definition). However we do provide a explicit-specialization of that type for the value true. However, the same for false is not provided. This means that sizeof(STATIC_ASSERT_FAILURE<true>) is valid but sizeof(STATIC_ASSERT_FAILURE<false>) is not. This can be used to create a compile time assert as follows

namespace static_assert
{
    template <bool> struct STATIC_ASSERT_FAILURE;
    template <> struct STATIC_ASSERT_FAILURE<true> { enum { value = 1 }; };

    template<int x> struct static_assert_test{};
}

#define COMPILE_ASSERT(x) \
    typedef ::static_assert::static_assert_test<\
        sizeof(::static_assert::STATIC_ASSERT_FAILURE< (bool)( x ) >)>\
            _static_assert_typedef_

Here the error we get is as follows

// compiles fine
COMPILE_ASSERT(sizeof(int) == sizeof(unsigned)); 
// error C2027: use of undefined type 'static_assert::STATIC_ASSERT_FAILURE<__formal>
COMPILE_ASSERT(sizeof(int) == sizeof(char)); 

So the advantage is that the STATIC_ASSERT_FAILURE is called out right at the point of failure and is more obvious to figure out

The macro expansion is as follows

1. typedef static_assert_test< sizeof(STATIC_ASSERT_FAILURE<false>) >  _static_assert_typedef_
   
2. typedef static_assert_test< sizeof(incomplete type) >  _static_assert_typedef_

Similarly for true the type is not incomplete and the expansion is

1. typedef static_assert_test< sizeof(STATIC_ASSERT_FAILURE<true>) >  _static_assert_typedef_
   
2. typedef static_assert_test< sizeof(valid type with one enum member) >  _static_assert_typedef_
   
3. typedef static_assert_test< 1 > _static_assert_typedef_

Put it all together

All together the following source gives a good working point to create static or compile time assert that works for both C and C++

#ifdef __cplusplus

#define JOIN( X, Y ) JOIN2(X,Y)
#define JOIN2( X, Y ) X##Y

namespace static_assert
{
    template <bool> struct STATIC_ASSERT_FAILURE;
    template <> struct STATIC_ASSERT_FAILURE<true> { enum { value = 1 }; };

    template<int x> struct static_assert_test{};
}

#define COMPILE_ASSERT(x) \
    typedef ::static_assert::static_assert_test<\
        sizeof(::static_assert::STATIC_ASSERT_FAILURE< (bool)( x ) >)>\
            JOIN(_static_assert_typedef, __LINE__)

#else // __cplusplus

#define COMPILE_ASSERT(x) extern int __dummy[(int)x]

#endif // __cplusplus

#define VERIFY_EXPLICIT_CAST(from, to) COMPILE_ASSERT(sizeof(from) == sizeof(to)) 

#endif // _COMPILE_ASSERT_H_

The only extra part is the JOIN macros. They just ensure that the typedef is using new names each time and doesn't give type already exists errors.

More Reading

1. Boost static_assert has an even better implementation that takes cares of much more scenarios and compiler quirks. Head over to the source at http://www.boost.org/doc/libs/1_36_0/boost/static_assert.hpp
   
2. Modern C++ Design by the famed Andrei Alexandrescu

Taking my job more seriously

I generally take my work very seriously. However, the following from SICP was still my favorite quote

"I think that it's extraordinarily important that we in computer science keep fun in computing. When it started out, it was an awful lot of fun. Of course, the paying customers got shafted every now and then, and after a while we began to take their complaints seriously. We began to feel as if we really were responsible for the successful, error-free perfect use of these machines. I don't think we are. I think we're responsible for stretching them, setting them off in new directions, and keeping fun in the house. I hope the field of computer science never loses its sense of fun. Above all, I hope we don't become missionaries. Don't feel as if you're Bible salesmen. The world has too many of those already. What you know about computing other people will learn. Don't feel as if the key to successful computing is only in your hands. What's in your hands, I think and hope, is intelligence: the ability to see the machine as more than when you were first led up to it, that you can make it more.''

However, couple of days back I was lying in a cold room with a doctor probing my Thyroid gland with ultrasound beams. She went on explaining the details of the tumor that I had, pointing to the screen. In my mind I was thinking about folks who have to code for these things and how careful they need to be.

The first thing I did when I came out of the room is search for whether .NET Compact Framework gets used for these kinds of equipment. Of the hundreds of pages I hit the one titled "Control System for Lung Ventilation Equipment with Windows CE , Microsoft .Net Compact Framework and Visual Studio Team System" struck me. Mostly because it had two products I personally coded for (NETCF and VSTS). There were even more life saving equipment listed on the search page.

It was a very humbling experience. I made a little vow, I'll be more careful when I code from tomorrow.

Silverlight on Nokia S60 devices

In many of my blog posts (e.g.here and here) I refer to .NET Compact Framework and Symbian OS (S60) and obviously folks keep asking me via comments (or assume) that we are porting .NETCF on S60 devices. So I thought it's time to clarify :)

The short answer is that we are not porting .NETCF to S60 devices, but we are porting the Silverlight on S60 devices. This was jointly announced by Microsoft and Nokia, read Nokia's press release here.

We are in very early stage of development and it is very hard to tell how much will be in it (e.g. will it be SL v1.0 or v2.0). However, we are working hard and as and when more details emerge I will share it out from this blog. So keep reading.

Who halted the code

Today a bed time story involving un-initialized variable access and a weird coincidence.

Couple of weeks back I was kind of baffled with a weird issue I was facing with the .NET Compact Framework Jitter. We were making the Jitter work for Symbian OS (S60) emulator. The S60 emulator is a Windows process and hence we needed to bring up the i386 Jitter in our code base. After running some JITted code the application would stop executing with the message "Execution halted" in the stack trace.

Now obviously the user didn't halt anything and the program was a simple hello world.

After some debugging I found an interesting thing. For the debug build the Symbian C compiler memsets un-initialized variables to 0xcccc. This is fairly standard and is used to catch un-initialized variable access.

However, our Jit code had a bug such that in some scenario it was not emitting function call/returns correctly. So instead of returning we were executing arbitrary memory. But instead of some sort of access violation (or some thing equally bad) the execution was actually halting.

The reason was we started executing un-initialized data. Now this data is 0xcccc. The S60 debugger uses interrupt 3 (INT 3) for inserting breakpoints and the following line reveals the rest of the story

For i386 the instruction used for inserting breakpoint and the pattern used for un-initialized data matched and the debugger thought that it hit a breakpoint.

Hex signatures

Frequently in code we need to add bit patterns or magic numbers. Since hex numbers have the alphabets A-F folks become creative and create all kinds of interesting words. Typical being 0xDEAD, 0xBEEF, 0xFEED or various combination of these like 0xFEEDBEEF. However, someone actually used the following in a sample

0xbed1babe

No naming the guilty :)

Back To Basics: Finding your stack usage

Handling stack overflow is a critical requirement for most Virtual Machines. For .NET Compact framework (NETCF) it is more important because it runs on embedded devices with very little memory and hence little stack space.

The .NETCF Execution Engine (EE) needs to detect stack overflows and report it via StackOverflowException. To do this it uses some logic which we internally refer to as StackCheck (that will be covered in a later post). The algorithm needs fair prediction of stack usage for system APIs (e.g. Win32 APIs or for Symbian OS, S60 APIs).

Each time we target a new platform we do some measurements in addition to referring to specs :) to find it's stack characteristics. As we are currently making the NETCF work on Symbian OS we are doing these measurements again. So I thought of sharing how we are going about measuring stack usage using simple watermarking.

The technique

Step:1

On method Entry store the two current and max stack values. This typically available via system APIs which in case of Symbian is available over the TThreadStackInfo class (iBase, iLimit and other members).

                       +---------------+
                       |               |
                       |               |
                       |               |
current stack ------>  +---------------+
                       |               |
                       |   Available   |
                       |    Stack      |
                       |               |
                       |               |
                       |               |
                       |               |
Stack limit ---------> +---------------+
                       |               |
                       |               |
                       .               .
                       .               .
                       .               . 

Step :2

Get a pointer on to the current stack pointer. How to get the pointer will vary based on the target platform. Options include system APIs, de-referencing stack pointer register (e.g. ESP register on x86) or simply creating a local variable on the stack and getting it's pointer.

Then Memset the whole region from current stack to the total available with some known pattern, e.g. 0xDEAD (a larger signature is a better approach to ensure there is no accidental match)

                       +---------------+
                       |               |
                       |               |
                       |               |
current stack ------>  +---------------+
                       |     DEAD      |
                       |     DEAD      |
                       |     DEAD      |
                       |     DEAD      |
                       |     DEAD      |
                       |     DEAD      |
                       |     DEAD      |
                       |     DEAD      |
Stack limit ---------> +---------------+
                       |               |
                       |               |
                       .               .
                       .               .
                       .               . 

Step: 3

Make an OS or whatever call you want to measure.

                       +---------------+
                       |               |
                       |               |
                       |               |
current stack ------>  +---------------+
                       |     1231      | --+
                       |     1231      |   |
                       |     D433      |   +--> Stack usage
                       |     D324      |   |
                       |     3453      | --+
                       |     DEAD      |
                       |     DEAD      |
                       |     DEAD      |
Stack limit ---------> +---------------+
                       |               |
                       |               |
                       .               .
                       .               .
                       .               . 

Step 4:

When the call returns the stack will get modified. Iterate through the memory starting from the current stack pointer looking for the first occurrence of the pattern you’ve set in Step:2. This is the water mark and is the point till which the stack got used by the OS call. Subtract the water mark from the original entry point saved in Step 1 and you have the stack usage.

Tail call optimization

I had posted about tail call and how .NET handles it before. However there was some email exchanges and confusion on internal DLs around it and so I thought I'd try to elaborate more on how .NET goes about handling tail calls

Let’s try to break this down a bit.

a) A piece of C# code contains tail recursion .

static void CountDown(int num)
{ 
    Console.WriteLine("{0} ", num); 
    if (num > 0) 
        CountDown(--num); 
} 

-

b) The C# compiler does not optimize this and generates a normal recursive call in IL

 IL_000c:  ...
 IL_000d:  ...
 IL_000e:  sub
 IL_000f:  dup
 IL_0010:  starg.s num
 IL_0012:  call void TailRecursion.Program::CountDown(int32)

-

c) The Jitter on seeing this call doesn’t optimize it in any way and just inserts a normal call instruction for the platform it targets (call for x86)

However, if any compiler is smart enough to add the tail. recursive IL statement then things change.

a) Scheme code

(define (CountDown n)
    (if (= n 0)
         n
      (CountDown (- n 1))))

-

b) Hypothetical IronScheme compiler will generate (note for scheme it has to do the optimization)

 IL_000c:  ...
 IL_000d:  ...
 IL_000e:  sub
  tail.
 IL_0023:  call void TailRecursion.Program::CountDown(int32)
 IL_0029:  ret

-

c) Based on which JIT you are using and various other scenarios the JIT now may honour the tail. IL instruction and optimize this with a jmp when generating machine instruction

Please note the may in the very last point and refer to here for some instances where the optimization still might not take place…

A* Pathfinding algorithm animation screen saver

I'm trying to learn WPF and IMO it is not a simple task. Previously whenever I upgraded my UI technology knowledge it had been easy as it build on a lot of pre-existing concepts. However, WPF is more of a disruptive change.

I decided to write some simple fun application in the effort to learn WPF. Since I cannot write a BabySmash application as Scott Hanselman has already done a awesome job out of it, I decided to code up a simulation of A* path finding algorithm and push it out as a screen saver. The final product looks as follows.

What it does is that it shows a source, a destination with blockages (wall, mountain?) in between and then uses A* algorithm to find a path in between them. This is the same kind of algorithm that is used in say games like Age of the Empires as workers move around collecting wood and other resources. The algorithm is documented here.

Features

1. It supports plugging in your own scenes with help of the awesome screen/board designer I blogged about
2. Comes with a WPF and a console client to show the animation
3. Algorithm can be played with easily to change or tweak it.
4. Shows full animation including start, end, obstacle, closed cells, current path being probed and the final path
5. Multi-screen support. Each screen shows a different board being solved.

Limitations:

Obviously this was more of a quick dirty hobby project and there remains a ton to work on. E.g.

1. Screen saver preview doesn't work.
2. Setting dialog is a sham.
3. The boards do not flip for vertically aligned screens
4. The XAML data binding is primitive and needs some work
5. The path can choose diagonally across cell even if it seems to cross over diagonal wall of obstacle.
6. Mouse move alone doesn't shut down the screen saver. You need to hit a
7. Many more :)

Download:

Download the final binaries from here. Unzip to a folder, right click on the *.scr and choose Install

Download sources (including VS 2008 sln) from here.

Enjoy.

How Many Types are loaded for Hello World

Consider the following super simple C# code

namespace SmartDeviceProject1
{
    class Program
    {
        static void Main(string[] args)
        {
            System.Console.WriteLine("Hello");
        }
    }
}

Can you guess how many managed Type gets loaded to run this? I was doing some profiling the .NET Compact Framework loader (for entirely unrelated reason) and was surprised by the list that got dumped. 87 types**, never could've guessed that...

1. System.Object
   
2. System.ValueType
   
3. System.Type
   
4. System.Reflection.MemberInfo
   
5. System.Delegate
   
6. System.MarshalByRefObject
   
7. System.SystemException
   
8. System.Exception
   
9. System.Attribute
   
10. System.Collections.Hashtable
    
11. System.Enum
    
12. System.Reflection.BindingFlags
    
13. System.MulticastDelegate
    
14. System.Reflection.MemberFilter
    
15. System.Reflection.Binder
    
16. System.Reflection.TypeAttributes
    
17. System.DateTime
    
18. System.Collections.IDictionary
    
19. System.UnhandledExceptionEventHandler
    
20. System.AppDomainManager
    
21. System.Version
    
22. System.Decimal
    
23. System.Runtime.InteropServices.ComInterfaceType
    
24. System.Collections.ICollection
    
25. System.Collections.IEqualityComparer
    
26. System.AppDomainManagerInitializationOptions
    
27. System.ArithmeticException
    
28. System.ArgumentException
    
29. System.MissingMemberException
    
30. System.MemberAccessException
    
31. System.AppDomainSetup
    
32. System.Reflection.AssemblyName
    
33. System.Globalization.CultureInfo
    
34. System.Reflection.Assembly
    
35. System.Configuration.Assemblies.AssemblyHashAlgorithm
    
36. System.Configuration.Assemblies.AssemblyVersionCompatibility
    
37. System.Reflection.AssemblyNameFlags
    
38. System.Globalization.CultureTableRecord
    
39. System.Globalization.CompareInfo
    
40. System.Globalization.TextInfo
    
41. System.Globalization.NumberFormatInfo
    
42. System.Globalization.DateTimeFormatInfo
    
43. System.Globalization.Calendar
    
44. System.Globalization.BaseInfoTable
    
45. System.Globalization.CultureTable
    
46. System.Globalization.CultureTableData
    
47. System.Globalization.NumberStyles
    
48. System.Globalization.DateTimeStyles
    
49. System.Globalization.DateTimeFormatFlags
    
50. System.Globalization.TokenHashValue
    
51. System.Globalization.CultureTableHeader
    
52. System.Globalization.CultureNameOffsetItem
    
53. System.Globalization.RegionNameOffsetItem
    
54. System.Globalization.IDOffsetItem
    
55. System.TokenType
    
56. System.Char
    
57. System.IO.TextReader
    
58. System.IO.TextWriter
    
59. System.IFormatProvider
    
60. System.Console
    
61. System.RuntimeTypeHandle
    
62. System.NotSupportedException
    
63. System.Globalization.EndianessHeader
    
64. System.Reflection.Missing
    
65. System.RuntimeType
    
66. System.Threading.StackCrawlMark
    
67. System.Globalization.CultureTableItem
    
68. System.Int32
    
69. System.Security.CodeAccessSecurityEngine
    
70. System.AppDomain
    
71. System.LocalDataStoreMgr
    
72. System.Threading.ExecutionContext
    
73. System.LocalDataStore
    
74. System.Collections.ArrayList
    
75. System.Threading.SynchronizationContext
    
76. System.Runtime.Remoting.Messaging.LogicalCallContext
    
77. System.Runtime.Remoting.Messaging.IllogicalCallContext
    
78. System.Threading.Thread
    
79. System.Collections.Generic.Dictionary`2
    
80. System.Runtime.Remoting.Messaging.CallContextRemotingData
    
81. System.RuApplication starting
    
82. System.Collections.Generic.IEqualityComparer`1
    
83. Runtime.Remoting.Messaging.CallContextSecurityData
    
84. System.Array
    
85. System.RuntimeFieldHandle
    
86. System.Globalization.CultureTableRecord[]
    
87. System.Text.StringBuilder

**This is for the compact framework CLR. Your mileage will vary if you run the same on the desktop CLR.

Designer for my path finding boards

I'm writing a small application (or rather a screen saver) that animates and demonstrates A* search algorithm. The idea is simple. On screen you see a start and end point and some random obstacles in between them. Then you see animation on how A* algorithm is used to navigate around the board to find a path (close to the shortest possible) between the two.

All of this is fairly standard. However, I got hit by a simple issue. I wanted to have the ability to design this board visually and also let my potential million users do the same. Obviously I don't have time to code up a designer and hence choose the all time manager's favorite technique. Re-use :)

So the final solution I took was to use Microsoft Office Excel as the WYSIWYG editor. I created a xlsx file with the following conditional formatting which colors cells based on the value of the cells.

So in this case

1. w = Wall marking the end of the table
2. b = blocks/bricks/obstacle
3. s = start
4. e = end

Using this I can easily design the board. The designer in action looks as follows

Since the excel sheet has conditional formatting the user just types in s, e, b, w in the cells and they all light up visually. At the end he/she just saves the file using File => Save As and uses CSV format. This saves the file shown above as

w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w
w,,,,,,,,,,,,,,,,,,,,,,,b,,,,,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,b,b,b,,,,,,,,b,,,,,,,,,,,,,,,,,,w
w,,,,s,,,,,,,,b,b,b,b,b,b,,,,,,b,,,,,,,,,,,,,,,,,,w
w,,,,,,,,,,b,b,b,b,,b,b,b,b,,,,,b,,,,,,,,,,,,,,,,,,w
w,,,,,,,,,b,b,b,,,,,,b,b,b,,,,b,,,,,,,,,,,,,,,,,,w
w,,,,,,,,,b,b,,,,,,,,b,b,,,,b,,,,,,,,,,,,,,,,,,w
w,,,,,,,,,b,b,,,,,,,,b,b,,,,b,,,,,,,,,,,,,,,,,,w
w,,,,,,,,,b,b,,,,,,,b,b,,,,,b,,,,,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,,,,b,b,,,,,,b,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,,,b,b,,,,,,,b,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,,b,b,,,,,,,,b,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,b,b,,,,,,,,,b,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,b,b,,,,,,,,,b,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,b,b,,,,,,,,,b,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,b,b,,,,,,,,,b,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,b,b,,,,,,,,,b,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,b,b,,,,,,,,,b,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,,,,,,,,,,,b,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,b,b,,,,,,,,,b,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,,,,,,,,,,,b,,,,b,,,,,,,,,,,e,,,w
w,,,,,,,,,,,,,,,,,,,,,,,b,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,,,,,,,,,,,b,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,,,,,,,,,,,,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,,,,,,,,,,,,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,,,,,,,,,,,,,,,b,,,,,,,,,,,,,,w
w,,,,,,,,,,,,,,,,,,,,,,,,,,,b,,,,,,,,,,,,,,w
w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w,w

As you see the format is simply a row per line and every column separated by a comma. My simulator reads this file and renders using whatever renderer is configured (console or WPF).

More about the A* simulator to come soon...

It's is easy to claim that the world won't get destroyed today

The clock is ticking and the Large Hadron Collider in Cern is going to get switched on today (9/10/2008). Even though there are speculations, CERN is claiming it's perfectly safe and the world won't end. But it's easy to claim that, who'll be around to prove them wrong in case they are :)

It's one of the coolest device at an operating temperature of < -270°C. But I'd get really angry if there's any disturbance to my Birthday celebrations!!

Team Foundation Server tool dump workspace details

I juggle around with a lot of workspaces. The reason is .NET Compact Framework is consumed in a whole bunch of things like Windows Mobile, Xbox, Zune, Nokia and most of them are on different source branches. On top of this active feature work happens in feature branches and there are other service branches to frequently peek into.

So the net effect is I need to sometime go to different folders and see what it's source control workspace mapping is or I need to dump details of workspaces on other computers and see if I can use that workspace as a template to create a new workspace.

The thing I hate here is to run cd to that folder and do a tf workspace to bring up the GUI and then dismiss the UI. I don't like GUI when things can be done equally well on the console. So I quickly hacked up a tool that dumps workspace details onto the console, something as below

d:\>workspace.exe NETCF\F01
Name:      ABHINAB2_S60
Server:    TfsServer01
Owner:     Abhinaba
Computer:  ABHINAB2

Comment:

Working folders:

Active   $/NetCF/F01/tools     d:\NETCF\F01\tools
Active   $/NetCF/F01/source    d:\NETCF\F01\source

-

This tool can be used either to see the mapping of a folder or given a workspace name and owner dump it details. It uses the exact same format as show in the tf workspace UI.

Source

The source and a build binary is linked below. It is a single source file (Program.cs) and you should be able to pull it into any version of Visual Studio and even build it from the command line using CSC

1. Source
   
2. Binary

Using the tool

Build it or use the pre-built binary and edit the .config file to point it to your TFS server and you are all set to go.

Obsession with desktop continues

This is my new office setup. I have been pushing around a lot of code lately and felt I needed more real-estate to effectively do what I'm doing. So I hooked up another monitor. All  3 are standard HP LP1965 19" monitors.

However, since none of my video cards support 3 monitors and I have a whole bunch of computers to hook up I had to do some complex wiring around :). If I number the screens 1,2,3 from left to right this is how it works out

• 1 and 2 is connected to my main dev box (machine1)
• 2 and 3 is connected to the machine I use for emails and browsing (machine 2)
• 2 actually goes through a 4-port KVM switch using which it can circle through machine1, machine2 and two other less used machines
• 1 also goes through a 2 port KVM switch and connects a xbox and machine 1
• I don't use the laptop at work directly, I connect to it using remote desktop.

Sweet :)

It's not as complex as it sounds. For my normal flow I just use the KVM to switch between machine-1 and machine-2. I rarely need to switch between the xbox and machine1.

PS: Don't try to make much sense of the Vintage books on the shelve, I plan to keep them for some more time before handing them off to some historian.

Back to Basic: Using a System.Threading.Interlocked is a great idea

I just saw some code which actually takes a lock to do a simple set operation. This is bad because taking locks are expensive and there is an easy alternative. The System.Threading.Interlocked class and its members can get the same job done and much faster.

I wrote the following two methods which increments a variable a million times. The first method does that by taking a lock and the other uses the Interlocked.Increment API.

static int IncrementorLock()
{
    int val = 0;
    object lockObject = new object();
    for (int i = 0; i < 1000000; ++i)
    {
        lock (lockObject)
        {
            val++;
        }
    }

    return val;
}

static int IncrementorInterlocked()
{
    int val = 0;
    for (int i = 0; i < 1000000; ++i)
    {
        System.Threading.Interlocked.Increment(ref val);
    }

    return val;
}

-

I then used the Visual Studio Team System Profiler (instrumented profiling) and got the following performance data.

Function Name	Elapsed Inclusive Time	Application Inclusive Time
IncrementorInterlocked()	1,363.45	134.43
IncrementorLock()	4,374.23	388.69

Even though this is a micro benchmark and uses a completely skewed scenario, it still drives the simple point that using the interlocked class is way faster.

The reason the interlocked version is faster is because instead of using .NET locks it directly calls Win32 apis like InterlockedIncrement which ensures atomic updation to variables at the OS scheduler level.

These Interlocked APIs complete depend on the support of the underlying OS. This is the reason you'd see that all the APIs are not available uniformly across all versions of .NET (e.g. there is no uniform coverage over WinCE and Xbox). While implementing these for the Nokia platform (S60) we are hitting some scenarios where there is no corresponding OS API.

String equality

akutz has one of the most detailed post on string interning and equality comparison performance metrics I have ever seen. Head over to the post here

I loved his conclusion which is the crux of the whole story.

"In conclusion, the String class’s static Equals method is the most efficient way to compare two string literals and the String class’s instance Equals method is the most efficient way to compare two runtime strings. The kicker is that there must be 10^5 (100,000) string comparisons taking place before one method starts becoming more efficient than the next. Until then, use whichever method floats your boat."

Kismat Konnection

<sorry for the total off-topic />

I had a real hard time giving a title to this post.

The post was originally supposed to be about the fire I saw yesterday in a movie theatre Talkie Town (map). However it is not really about the fire, it is more about the people and how they reacted to it.

First things first, the fire.

Folks who know me, know that I go to a movie theatre once in couple of years, I prefer movies visiting me rather than the other way around. In the 4 years I'm in Hyderabad, this was my 3rd visit to a movie. So I can say it was a momentous occasion when me and my wife impromptu decided to see a movie in the 2 screen theatre in Miyapur called Talkie Town. The decision was heavily biased on the fact that my father-in-law was at home looking after our daughter.

Batman  (or rather The Dark Knight) lost to my wife and I was forced to see this bollywood flick Kismat Konnection.

This is where all the fun started. Just after the interval (1.5 hours through, and yes bollywood movies are that long) people saw some light on the theatre ceiling. Soon the light spread and there was a small hole in the ceiling and we could actually see flames. Whatever was above the sound proofed ceiling had caught fire and the heat actually caused the material of the ceiling to burn.

Then the most amazing thing happened, 50% people didn't care. They were gleefully looking up and not even caring to leave the theatre. I'm not talking about some false fire alarm, or smell of smoke, it was a real fire with flames and a burned hole in the ceiling!! The fire alarm didn't sound and no one from the theatre authority seemed to be around. I left and called some folks and they simply went to the terrace. All the noise now prompted some more people to leave but even at that time there were others happily watching as if the whole thing was a part of the movie.

Then even more amazing thing happened, the folks from the theatre came and said that the fire has been doused and re-started the show. The hall was smoky and the AC was off. Me and another guy (yea there were some more sane people around) caught hold of an official and he simply said he has checked and he can guarantee that everything is safe. I asked him that since he couldn't ensure that there is no fire in the first place how can he guarantee against a recurrence, he simply said I can get a refund of my ticket. Which I did and left the hall.

On the way back I had a revelation. From childhood I had seen that people around me had little care about safety in general, but this is the first time I figured out that people didn't even care about their own safety. How can a movie be worth taking the risk of sitting in a fire hazard zone and that too with small children on their lap is something I will never figure out.

A lot of things including the traffic situation and the weird jay walking I see around suddenly makes more sense to me.

Writing exception handlers as separate methods may prove to be a good idea

Let us consider a scenario where you catch some exception and in the exception handler do some costly operation. You can write that code in either of the following ways

Method-1 : Separate method call

public class Program
{
    public static void Main(string[] args)
    {
        try
        {
            using (DataStore ds = new DataStore())
            {
                // ...
            }
        }
        catch (Exception ex)
        {
            ErrorReporter(ex);
        }
    }

    private static void ErrorReporter(Exception ex)
    {
        string path = System.IO.Path.GetTempFileName();
        ErrorDumper ed = new ErrorDumper(path, ex);
        ed.WriteError();

        XmlDocument xmlDoc = new XmlDocument();
        xmlDoc.Load(path);
        RemoteErrorReporter er = new RemoteErrorReporter(xmlDoc);
        er.ReportError();
    }
}

-

Method-2 : Inline

public static void Main(string[] args)
{
    try
    {
        using (DataStore ds = new DataStore())
        {
            // ...
        }
    }
    catch (Exception ex)
    {
        string path = System.IO.Path.GetTempFileName();
        ErrorDumper ed = new ErrorDumper(path, ex);
        ed.WriteError();

        XmlDocument xmlDoc = new XmlDocument();
        xmlDoc.Load(path);
        RemoteErrorReporter er = new RemoteErrorReporter(xmlDoc);
        er.ReportError();
    }
}

-

The simple difference is that in the first case the exception handler is written as a separate method and in the second case it is placed directly inline inside the handler itself.

The question is which is better in terms of performance?

In case you do have significant code and type reference in the handler and you expect the exception to be thrown rarely in an application execution then the Method-1 is going to be more performant.

The reason is that just before executing a method the whole method gets jitted. The jitted code contains stubs to the other method's it will call but it doesn't do a recursive jitting. This means when Main gets called it gets jitted but the method ErrorReporter is still not jitted. So in case the exception is never fired all the code inside ErrorReporter never gets Jitted. This might prove to be significant saving in terms of time and space if the handling code is complex and refers to type not already referenced.

However, if the code is inline then the moment Main gets jitted all the code inside the catch block gets jitted. This is expensive not only because it leads to Jitting of code that is never executed but also because all types referenced in the catch block is also resolved resulting in loading a bunch of dlls after searching though the disk. In our example above System.Xml.dll and the other dll containing remote error reporting gets loaded even though they will never be used. Since disk access, loading assemblies and type resolution are slow, the simple change can prove to give some saving.

Microsoft RoundTable

Our conference rooms have been fitted with this really weird looking device (click to enlarge).

I had no clue what the thing was. Fortunately it's box was still placed in the room along with the manual. It's called the Microsoft RoundTable and it is actually a 360-degree camera (with 5 cameras and 6 microphones). It comes with bundled software that let's all participant be visible to the other side in a live meeting at real time. It shows the white board and the software is intelligent enough to focus on and track the active speaker (using microphone and face recognition) and much much more (lot of MS Research stuff has gone into it). The video below gives you some idea and head on to this post for some review and inside view of the device.

Simply put it's AWSOME

Do namespace using directives affect Assembly Loading?

The simple answer is no, the inquisitive reader can read on :)

Close to 2 year back I had posted about the two styles of coding using directives as follows

Style 1

namespace MyNameSpace
{
    using System;
    using System.Collections.Generic;
    using System.Text;
    // ...
}

-

Style 2

using System;
using System.Collections.Generic;
using System.Text;

namespace MyNameSpace
{
    // ...
}

-

and outlined the benefits of the first style (using directives inside the namespace). This post is not to re-iterate them.

This post to figure out if either of the styles have any bearing in the loading order of assemblies. Obviously at the first look it clearly indicates that is shouldn't, but this has caused some back and forth discussions over the web.

Scot Hanselman posted about a statement on the Microsoft Style cop blog which states

"When using directives are declared outside of a namespace, the .Net Framework will load all assemblies referenced by these using statements at the same time that the referencing assembly is loaded.

However, placing the using statements within a namespace element allows the framework to lazy load the referenced assemblies at runtime. In some cases, if the referencing code is not actually executed, the framework can avoid having to load one or more of the referenced assemblies completely. This follows general best practice rule about lazy loading for performance.

Note, this is subject to change as the .Net Framework evolves, and there are subtle differences between the various versions of the framework."

This just doesn't sound right because using directives have no bearing to assembly loading.

Hanselman did a simple experiment with the following code

using System;  
using System.Xml;  
  
namespace Microsoft.Sample  
{  
   public class Program  
   {  
      public static void Main(string[] args)  
      {  
         Guid g = Guid.NewGuid();  
         Console.WriteLine("Before XML usage");  
         Console.ReadLine();  
         Foo();  
         Console.WriteLine("After XML usage");  
         Console.ReadLine();  
      }  
  
      public static void Foo()  
      {  
         XmlDocument x = new XmlDocument();  
      }  
   }  
}  

-

and then he watched the loading time using process explorer and then he moved the using inside the namespace and did the same. Both loaded the System.Xml.dll after he hit enter on the console clearly indicating that for both the cases they got lazy loaded.

Let me try to give a step by step rundown of how the whole type look up of XmlDocument happens in .NETCF which in turn would throw light on whether using directives have bearing on assembly loading.

1. When Main method is Jitted and ran the System.Xml.dll is not yet loaded
   
2. When method Foo is called the execution engine (referred to as EE) tries to JIT the method. As documented the Jitter only JITs methods that are to be executed.
   
3. The Jitter tries to see if the method Foo is managed (could be native as well due to mixed mode support) and then tries to see if it's already Jitted (by a previous call), since it's not it goes ahead with jitting it
   
4. The jitter validates a bunch of stuff like whether the class on which the method Foo is being called (in this case Microsoft.Sample.Program) is valid, been initialized, stack requirements, etc...
   
5. Then it tries to resolve the local variables of the method. It waits to resolve the local variable type reference till this point so that it is able to save time and memory by not Jitting/loading types that are referenced by methods that are never executed
   
6. Then it tries to resolve the type of the variable which in this case if System.Xml.XmlDocument.
   
7. It sees if it's already in the cache, that is if that type is already loaded
   
8. Since it's not the case it tries to search for the reference based on the type reference information
   
9. This information contains the full type reference including the assembly name, which in this case is System.Xml.dll and also version information,strong name information, etc...
   
10. All of the above information along with other information like the executing application's path is passed to the assembly loader to load the assembly
    
11. The usual assembly search sequence is used to look for the assembly and then it is loaded and the type reference subsequently gets resolved

If you see the above steps there is in no way a dependency of assembly loading on using directive. Hence at least on .NETCF whether you put the using outside or inside the namespace you'd get the referenced assemblies loaded exactly at the time of first reference of a type from that assembly (the step #5 above is the key).

Auto generating Code Review Email for TFS

We use a small command line tool called crmail to auto-generate code review email from shelveset. I find the whole process very helpful and thought I'd share the process and the tool (which has some really cool features).

Features

1. Automatic generation of the email from the shelveset details
2. Hyperlinks are put to TFS webaccess so that you can review code from machines without any tools installed, even without source enlistment. Yes it's true!!! The only thing you need is your office's intranet access
3. You can even use a Windows mobile phone :) and even some non MS browsers. Ok I guess I have sold this enuf
4. This is how the email looks like with all the details pointed out
   
5. Effectively you can see the file diff, history, blame (annotate), shelveset details, associated bugs, everything from your browser and best thing is that all of these takes one click each.
   This is how the fill diff looks in the browser
   

Pre-reqs

1. Team System Web Access (TSWA) 2008 power tool installed on your TFS server
2. Outlook installed on the machine on which the email is generated
3. Enlistment and TFS client installed on the machine on which the email is generated
4. For reviewers there is no pre-req other than a browser and email reader.

Dev process

1. The developer creates a shelveset after he is done with his changes. He ensures he fills up all the details including the reviewers email address ; separated
2. He runs the tool with a simple command
   crmail shelvesetname
3. Email gets generated and opened he fills in additional information and fires send
4. Done!!

Reviewers

Ok they just click on the email links. Since mostly these are managers what more do you expect out of them? Real devs will stick with firing up tfpt command line :)

Configuring the tool

1. Download the binaries from here
2. Unzip. Open the crmail.exe.config file and modify the values in it to point to your tfsserver and your code review distribution list (if you do not have one then make it empty)
3. Checkin to some tools folder in your source control so that everyone in your team has access to it

Support

Self help is the best help :), download the sources from here and enjoy. Buck Hodges post on the direct link URLs would help you in case you want to modify the sources to do more.

How does the .NET CF handle null reference

What happens when we have code as bellow

class B
{
    public virtual void Virt(){
        Console.WriteLine("Base::Virt");
    }
}

class Program
{
    static void Main(string[] args){
        B b = null;
        b.Virt(); // throws System.NullReferenceException
    }
}

Obviously we have a null reference exception being thrown. If you see the IL the call looks like

    L_0000: nop 
    L_0001: ldnull 
    L_0002: stloc.0 
    L_0003: ldloc.0 
    L_0004: callvirt instance void ConsoleApplication1.B::Virt()
    L_0009: nop 
    L_000a: ret 

 

So in effect you'd expect the jitter to generate the following kind of code (in processor instruction)

if (b == null)
   throw new NullReferenceException
else
   b->Virt() // actually call safely using the this pointer

 

However, generating null checks for every call is going to lead to code bloat. So to work around this on some platforms (e.g. .NETCF on WinCE 6.0 and above) it uses the following approach

1. Hook up native access violation exception (WinCE 6.0 supports this) to a method in the execution engine (EE)
   
2. Do not generate any null checking and directly generate calls through references
   
3. In case the reference is null then a native AV (access violation is raised as invalid 0 address is accessed) and the hook method is called
   
4. At this point the EE checks to see if the source of the access violation (native code) is inside Jitted code block. If yes it creates the managed NullRefenceException and propagates it up the call chain.
   
5. If it's outside then obviously it's either CLR itself or some other native component is crashing and it has nothing to do about it..

C# generates virtual calls to non-virtual methods as well

Sometime back I had posted about a case where non-virtual calls are used for virtual methods and promised posting about the reverse scenario. This issue of C# generating callvirt IL instruction even for non-virtual method calls keeps coming back on C# discussion DLs every couple of months. So here it goes :)

Consider the following code

class B
{
    public virtual void Virt(){
        Console.WriteLine("Base::Virt");
    }

    public void Stat(){
        Console.WriteLine("Base::Stat");
    }
}

class D : B
{
    public override void Virt(){
        Console.WriteLine("Derived::Virt");
    }
}

class Program
{
    static void Main(string[] args)
    {
        D d = new D();
        d.Stat(); // should emit the call IL instruction
        d.Virt(); // should emit the callvirt IL instruction
    }
}

The basic scenario is that a base class defines a virtual method and a non-virtual method. A call is made to base using a derived class pointer. The expectation is that the call to the virtual method (B.Virt) will be through the intermediate language (IL) callvirt instruction and that to the non-virtual method (B.Stat) through call IL instruction.

However, this is not true and callvirt is used for both. If we open the disassembly for the Main method using reflector or ILDASM this is what we see

    L_0000: nop 
    L_0001: newobj instance void ConsoleApplication1.D::.ctor()
    L_0006: stloc.0 
    L_0007: ldloc.0 
    L_0008: callvirt instance void ConsoleApplication1.B::Stat()
    L_000d: nop 
    L_000e: ldloc.0 
    L_000f: callvirt instance void ConsoleApplication1.B::Virt()
    L_0014: nop 
    L_0015: ret 

Question is why? There are two reasons that have been brought forward by the CLR team

1. API change.
   The reason is that .NET team wanted a change in an method (API) from non-virtual to virtual to be non-breaking. So in effect since the call is anyway generated as callvirt a caller need not be recompiled in case the callee changes to be a virtual method.
   
   
2. Null checking
   If a call is generated and the method body doesn't access any instance variable then it is possible to even call on null objects successfully. This is currently possible in C++, see a post I made on this here.
   

   With callvirt there's a forced access to this pointer and hence the object on which the method is being called is automatically checked for null.

callvirt does come with additional performance cost but measurement showed that there's no significant performance difference between call with null check vs callvirt. Moreover, since the Jitter has full metadata of the callee, while jitting the callvirt it can generate processor instructions to do static call if it figures out that the callee is indeed non-virtual.

However, the compiler does try to optimize situations where it knows for sure that the target object cannot be null. E.g. for the expression i.ToString(); where i is an int call is used to call the ToString method because Int32 is value type (cannot be null) and sealed.