LEAVE and TRAP:Exception Handling in Symbian C++

 

Most of us which are beggining in Symbian programming have dificulties on to adapt yourself with the exception handling schema used in the Symbian C++ dialect. Personally, for me was very bore to know how to use the CleanupStack and the ELeave constant!!! Let me try to show you what I learned...

For starting, some definitions: an exception differs from an error, because an error, or a bug, you MUST (or NEED) to correct it to put the application properly running. We would say that an exception is an error, but an exception appears at runtime. The main difference is that an error, once it had discovered, gives to you the opportunity to correct it, and an exception not, because it depends on environment factors (such memory limitations...)

With the Symbian C++ you can't use the famous try/catch keywords for exception handling, very well knowed such by Java developers, as by C++ programmers... (try/catch was removed by performance issues, and to downsize the generated code)

Thinking on it, the Symbian creators implemented an alternative model for exception handling such as the encountered in the modern languages, but with a minor impact in the performance and "object code" size, using only operator overloading, the creation of some classes and functions, and the definition of some exclusive macros for exception handling. In a such way, we could assume that the Symbian LEAVE is like throws, and the TRAP is the catch.

A simplified schema for Symbian exception handling is to think that every function you call, which would throws an exception, is putted inside a TRAP. TRAP is our try/catch, and is defined through a macro "define" keyword. For example:

TInt error;
TRAP(error, fooFunctionL());
if (error!=KErrNone)
{
   // threats exceptions throwed by "fooFunctionL()"
}

In the Symbian developer day-by-day, it's unusual to reference TRAP calls in a Symbian programm, because Symbian Active Scheduler (an execution daemon thread) already call TRAP macros in many execution points in the application. This turn its use a rare occurrence. So, let's pass to the next strucuture, which is more frequently used. I'm talking about CleanupStack. CleanupStack class, like a generic stack struct, defines 2 fundamental methods, PushL and Pop. The CleanupStack main intent is to address every allocated object that may be "orphaned" when an exception occurs. CleanupStack maintains a stack where its forming elements stores pointers to the objects in the heap. It's very importante on that cases where an object may lose its reference during object instantiation and memory space allocation. Suppose a function that can throw an exception (nominatedely, every function with the character "L" sufix can throw an exception, or a LEAVE):

void CTest::testFunctionL()
{
   CSomeClass* obj1 = new (ELeave) CComeClass;
   CComeClass* obj2 = new (ELeave) CComeClass;
   ConstructL();
}

There is an misuse in the function above: all 2 lines in the function can throw an exception (leave). So, what happens when the first variable (obj1) is succesfully allocated, but the second one (obj2) couldn't be instantiated because a memory leack occurs? The result is that a LEAVE is throwed on the second line, the second object wasn't created, but the first one will be orphaned, because the TRAP in the Active Scheduler (or in the user's program) will clean all automatic variables (automatic variables are that declared inside a function)

To fix the "testFunctionL()" method:

void CTest::testFunctionL()
{
   CSomeClass* obj1 = new (ELeave) CSomeClass;
   CleanupStack::PushL(obj1);
   CSomeClass* obj2 = new (ELeave) CSomeClass;
   CleanupStack::PushL(obj2);
   ConstructL();
}

With the CleanupStack's PushL inclusion, when a Leave occurs, the associated TRAP will call a PopAndDestroy (this function do the same thing the Pop do, but call the reference's destructor too, and clear space occupied by it) over every object declared and initialized inside this function. You can remove his object's pointer from the stack, calling the CleanupStack::Pop() function. In that case, you call Pop wherever you are certified that the object won't throw an exception (or a Leave), because once it was removed from CleanupStack, a TRAP wouldn't clear its reference from memory.

The last tip: to specify when an instantiation instruction (new operator) could Leave, you can use the overloaded new operator, passing the ELeave constant parameter. For example:

CUser* usr = new (ELeave) CUser(name, age);

Using operator overloading, the Symbian designers redefined the new operator to receive a constant, whiuch will say that it can throw an exception. A similar code in C++ ANSI could use the following strategy for threating an excpetion in the object initialization:

CUser* user;
if ( (user = new CUser(name, age)) == NULL )
{
   // manage the inapropriated initialization
}

Perceive the differences from that solution, to that used on Symbian...

  • 0
    点赞
  • 0
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
在信号处理领域,DOA(Direction of Arrival)估计是一项关键技术,主要用于确定多个信号源到达接收阵列的方向。本文将详细探讨三种ESPRIT(Estimation of Signal Parameters via Rotational Invariance Techniques)算法在DOA估计中的实现,以及它们在MATLAB环境中的具体应用。 ESPRIT算法是由Paul Kailath等人于1986年提出的,其核心思想是利用阵列数据的旋转不变性来估计信号源的角度。这种算法相比传统的 MUSIC(Multiple Signal Classification)算法具有较低的计算复杂度,且无需进行特征值分解,因此在实际应用中颇具优势。 1. 普通ESPRIT算法 普通ESPRIT算法分为两个主要步骤:构造等效旋转不变系统和估计角度。通过空间平移(如延时)构建两个子阵列,使得它们之间的关系具有旋转不变性。然后,通过对子阵列数据进行最小二乘拟合,可以得到信号源的角频率估计,进一步转换为DOA估计。 2. 常规ESPRIT算法实现 在描述中提到的`common_esprit_method1.m`和`common_esprit_method2.m`是两种不同的普通ESPRIT算法实现。它们可能在实现细节上略有差异,比如选择子阵列的方式、参数估计的策略等。MATLAB代码通常会包含预处理步骤(如数据归一化)、子阵列构造、旋转不变性矩阵的建立、最小二乘估计等部分。通过运行这两个文件,可以比较它们在估计精度和计算效率上的异同。 3. TLS_ESPRIT算法 TLS(Total Least Squares)ESPRIT是对普通ESPRIT的优化,它考虑了数据噪声的影响,提高了估计的稳健性。在TLS_ESPRIT算法中,不假设数据噪声是高斯白噪声,而是采用总最小二乘准则来拟合数据。这使得算法在噪声环境下表现更优。`TLS_esprit.m`文件应该包含了TLS_ESPRIT算法的完整实现,包括TLS估计的步骤和旋转不变性矩阵的改进处理。 在实际应用中,选择合适的ESPRIT变体取决于系统条件,例如噪声水平、信号质量以及计算资源。通过MATLAB实现,研究者和工程师可以方便地比较不同算法的效果,并根据需要进行调整和优化。同时,这些代码也为教学和学习DOA估计提供了一个直观的平台,有助于深入理解ESPRIT算法的工作原理。
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值