我们如何才能轻松找到代码中的哪一行导致运行时异常？(How can we easily find what line in the code causes a runtime exception?)

 考虑以下分析器： 
 
public void AnalyzeNode(SyntaxNode node, SemanticModel semanticModel, Action<Diagnostic> addDiagnostic, CancellationToken cancellationToken)
{
    var throwStatement = node as ThrowStatementSyntax;

    var isObjectCreationExpression = throwStatement.Expression is ObjectCreationExpressionSyntax;
    var obj = throwStatement.Expression as ObjectCreationExpressionSyntax;

    var isCorrectTypeOfExpression = (obj.Type as IdentifierNameSyntax).Identifier.Text == typeof(ArgumentException).Name;
}
 
 使用SyntaxKind.ThrowStatement作为兴趣种类。 
 
 如果抛出的异常没有在new Exception()形式中声明，则obj应该为null ，而是表示为throw e ，其中e是先前声明的异常。 
 
 当在之后调用obj.Type时，这反过来将抛出NullReferenceException 。 
 
 有问题的例子： 
 
static void Method1()
{
    throw new ArgumentException();
}

static void Method2(ArgumentException e)
{
    throw e;
}
 
 第一次throw会很好地通过分析器，但第二次throw会导致obj为null因为它不是ObjectCreationExpressionSyntax类型。 
 
 在沙箱Visual Studio环境中，这将显示为信息消息： 
 
 
 
 用户诊断分析器'FormattingFixes.EmptyArgumentException.ArgumentExceptionAnalyzer'抛出了一个异常，消息'对象引用未设置为对象的实例。'。 
 
 
 在这个简短的样本中，很容易分辨问题所在，但在一个不那么人为的例子中，它将更难被发现。 行和列都是1 ，没有帮助。 
 
 在“传统”编程中，您的环境会自动显示抛出运行时异常的位置以及该时间点的值。 我可以在我的代码中的某个地方放置一个随机断点，每次命中时查看所有值，并尝试从那里推断出它，但是一旦节点数远远高于这些值，这就不能很好地扩展。 
 
 我们如何才能轻松找到代码中的哪一行导致运行时异常？ 

Consider the following analyzer:
 
public void AnalyzeNode(SyntaxNode node, SemanticModel semanticModel, Action<Diagnostic> addDiagnostic, CancellationToken cancellationToken)
{
    var throwStatement = node as ThrowStatementSyntax;

    var isObjectCreationExpression = throwStatement.Expression is ObjectCreationExpressionSyntax;
    var obj = throwStatement.Expression as ObjectCreationExpressionSyntax;

    var isCorrectTypeOfExpression = (obj.Type as IdentifierNameSyntax).Identifier.Text == typeof(ArgumentException).Name;
}
 
With SyntaxKind.ThrowStatement as the Kind of Interest.
 
obj should be null if the thrown exception has not been declared right there in the form new Exception() but is rather presented as throw e where e is a previously declared exception.
 
This in turn will throw a NullReferenceException when obj.Type is called right after.
 
Example in question:
 
static void Method1()
{
    throw new ArgumentException();
}

static void Method2(ArgumentException e)
{
    throw e;
}
 
The first throw will pass the analyzer just fine, but the second one will cause obj to be null since it is not of type ObjectCreationExpressionSyntax.
 
In the sandbox Visual Studio environment this will show up as an information message:
 
 
 
The User Diagnostic Analyzer 'FormattingFixes.EmptyArgumentException.ArgumentExceptionAnalyzer' threw an exception with message 'Object reference not set to an instance of an object.'.
 
 
In this short sample it's easy to tell where the problem is, but in a less contrived example it will be harder to spot. The line and column are both 1, not helpful. 
 
In "traditional" programming, your environment automatically shows you where the runtime exception was thrown and what the values are at that point in time. I can put a random breakpoint somewhere in my code, look at all values every time it's hit and try to deduce it from there, but this doesn't scale well once the amount of nodes goes much higher than these 2.
 
How can we easily find what line in the code causes a runtime exception?

原文：https://stackoverflow.com/questions/23093709