我想收集尽可能多的关于。net / clr中API版本控制的信息,特别是API更改如何破坏或不破坏客户端应用程序。首先,让我们定义一些术语:

API change - a change in the publicly visible definition of a type, including any of its public members. This includes changing type and member names, changing base type of a type, adding/removing interfaces from list of implemented interfaces of a type, adding/removing members (including overloads), changing member visibility, renaming method and type parameters, adding default values for method parameters, adding/removing attributes on types and members, and adding/removing generic type parameters on types and members (did I miss anything?). This does not include any changes in member bodies, or any changes to private members (i.e. we do not take into account Reflection).

二进制级中断——一种API更改,导致针对旧版本API编译的客户端程序集可能无法装入新版本。例如:改变方法签名,即使它允许以与以前相同的方式被调用(即:void返回类型/参数默认值重载)。

源代码级中断——API更改会导致针对旧版本API编写的现有代码可能无法使用新版本进行编译。但是,已经编译的客户机程序集与以前一样工作。例如:添加一个新的重载,可能导致之前明确的方法调用出现歧义。

源代码级安静的语义变化——API的变化会导致编写的现有代码针对旧版本的API悄悄地改变其语义,例如通过调用不同的方法。然而,代码应该继续编译,没有警告/错误,以前编译的程序集应该像以前一样工作。示例:在现有类上实现一个新接口,这会导致在重载解析过程中选择不同的重载。

The ultimate goal is to catalogize as many breaking and quiet semantics API changes as possible, and describe exact effect of breakage, and which languages are and are not affected by it. To expand on the latter: while some changes affect all languages universally (e.g. adding a new member to an interface will break implementations of that interface in any language), some require very specific language semantics to enter into play to get a break. This most typically involves method overloading, and, in general, anything having to do with implicit type conversions. There doesn't seem to be any way to define the "least common denominator" here even for CLS-conformant languages (i.e. those conforming at least to rules of "CLS consumer" as defined in CLI spec) - though I'll appreciate if someone corrects me as being wrong here - so this will have to go language by language. Those of most interest are naturally the ones that come with .NET out of the box: C#, VB and F#; but others, such as IronPython, IronRuby, Delphi Prism etc are also relevant. The more of a corner case it is, the more interesting it will be - things like removing members are pretty self-evident, but subtle interactions between e.g. method overloading, optional/default parameters, lambda type inference, and conversion operators can be very surprising at times.

这里有几个例子:

添加新的方法重载

Kind:源级中断

受影响的语言:c#, VB, f#

更改前的API:

public class Foo
{
    public void Bar(IEnumerable x);
}

更改后的API:

public class Foo
{
    public void Bar(IEnumerable x);
    public void Bar(ICloneable x);
}

样例客户端代码在更改前工作,更改后失效:

new Foo().Bar(new int[0]);

添加新的隐式转换运算符重载

Kind:源级中断。

受影响的语言:c#, VB

不受影响语言:f#

更改前的API:

public class Foo
{
    public static implicit operator int ();
}

更改后的API:

public class Foo
{
    public static implicit operator int ();
    public static implicit operator float ();
}

样例客户端代码在更改前工作,更改后失效:

void Bar(int x);
void Bar(float x);
Bar(new Foo());

注意:f#并没有被破坏,因为它对重载操作符没有任何语言级别的支持,既不是显式的也不是隐式的——两者都必须作为op_Explicit和op_Implicit方法直接调用。

添加新的实例方法

Kind:源级安静语义更改。

受影响的语言:c#, VB

不受影响语言:f#

更改前的API:

public class Foo
{
}

更改后的API:

public class Foo
{
    public void Bar();
}

客户端代码示例:

public static class FooExtensions
{
    public void Bar(this Foo foo);
}

new Foo().Bar();

注意:f#并没有被破坏,因为它没有对ExtensionMethodAttribute的语言级支持,并且需要将CLS扩展方法作为静态方法调用。


当前回答

Visual Studio扩展NDepend在API Breaking Changes类别中提供了几个规则来检测二进制级别的中断。只有定义了NDepend基线,这些规则才会执行。

API Breaking Changes: Types: This rule warns if a type publicly visible in the baseline, is not publicly visible anymore or if it has been removed. Clients code using such type will be broken. API Breaking Changes: Methods: This rule warns if a method publicly visible in the baseline, is not publicly visible anymore or if it has been removed. Clients code using such method will be broken. Note that if a method signature gets changed the old method version is seen as removed and the new method version is seen as added, so a breaking change will be detected on the old method version. API Breaking Changes: Fields: This rule warns if a field publicly visible in the baseline, is not publicly visible anymore or if it has been removed. Clients code using such field will be broken. API Breaking Changes: Interfaces and Abstract Classes: This rule warns if a publicly visible interface or abstract class has been changed and contains new abstract methods or if some abstract methods have been removed. Clients code that implement such interface or derive from such abstract class will be broken. Broken serializable types: This rule warns about breaking changes in types tagged with SerializableAttribute. To do so, this rule searches for serializable type with serializable instance fields added or removed since the baseline. Notice that it doesn't take account of fields tagged with NonSerializedAttribute. Avoid changing enumerations Flags status: This rule matches enumeration types that used to be tagged with FlagsAttribute in the baseline, and not anymore. It also matches the opposite, enumeration types that are now tagged with FlagsAttribute, and were not tagged in the baseline. Being tagged with FlagsAttribute is a strong property for an enumeration. Not so much in terms of behavior (only the enum.ToString() method behavior changes when an enumeration is tagged with FlagsAttribute) but in terms of meaning: is the enumeration a range of values or a range of flags?

还有3个代码查询建议让用户浏览新的公共API元素:

新的公开可见类型。 新的公开可见方法。 新的公开可见字段。

其他回答

将显式接口实现转换为隐式接口实现。

打破:来源

受影响语言:所有

将显式接口实现重构为隐式接口实现在如何破坏API方面更为微妙。从表面上看,这似乎是相对安全的,然而,当与继承结合在一起时,它可能会导致问题。

更改前的API:

public class Foo : IEnumerable
{
    IEnumerator IEnumerable.GetEnumerator() { yield return "Foo"; }
}

API变更后:

public class Foo : IEnumerable
{
    public IEnumerator GetEnumerator() { yield return "Foo"; }
}

样例客户端代码在更改前工作,更改后被破坏:

class Bar : Foo, IEnumerable
{
    IEnumerator IEnumerable.GetEnumerator() // silently hides base instance
    { yield return "Bar"; }
}

foreach( var x in new Bar() )
    Console.WriteLine(x);    // originally output "Bar", now outputs "Foo"

当我发现它时,这一点非常不明显,特别是考虑到接口的相同情况。这根本不是一个休息,但我决定把它包括在内,这是足够令人惊讶的:

将类成员重构为基类

Kind:不休息!

受影响的语言:无(即没有损坏)

更改前的API:

class Foo
{
    public virtual void Bar() {}
    public virtual void Baz() {}
}

更改后的API:

class FooBase
{
    public virtual void Bar() {}
}

class Foo : FooBase
{
    public virtual void Baz() {}
}

在整个更改过程中保持工作的示例代码(即使我预计它会中断):

// C++/CLI
ref class Derived : Foo
{
   public virtual void Baz() {{

   // Explicit override    
   public virtual void BarOverride() = Foo::Bar {}
};

注:

C++/CLI is the only .NET language that has a construct analogous to explicit interface implementation for virtual base class members - "explicit override". I fully expected that to result in the same kind of breakage as when moving interface members to a base interface (since IL generated for explicit override is the same as for explicit implementation). To my surprise, this is not the case - even though generated IL still specifies that BarOverride overrides Foo::Bar rather than FooBase::Bar, assembly loader is smart enough to substitute one for another correctly without any complaints - apparently, the fact that Foo is a class is what makes the difference. Go figure...

添加具有默认值的参数。

中断类型:二进制级别的中断

即使调用源代码不需要更改,它仍然需要重新编译(就像添加常规参数一样)。

这是因为c#将参数的默认值直接编译到调用程序集中。这意味着如果您不重新编译,您将得到一个MissingMethodException,因为旧程序集试图调用带有较少参数的方法。

更改前的API

public void Foo(int a) { }

更改后的API

public void Foo(int a, string b = null) { }

之后被破坏的示例客户端代码

Foo(5);

客户端代码需要在字节码级别上重新编译为Foo(5, null)。被调用的程序集将只包含Foo(int, string),而不是Foo(int)。这是因为默认参数值纯粹是一种语言特性,. net运行时对它们一无所知。(这也解释了为什么在c#中默认值必须是编译时常量)。

具有可空类型参数的重载方法

类型:源级中断

受影响的语言:c#, VB

更改前的API:

public class Foo
{
    public void Bar(string param);
}

更改后的API:

public class Foo
{
    public void Bar(string param);
    public void Bar(int? param);
}

样例客户端代码在更改前工作,更改后失效:

new Foo().Bar(null);

例外:以下方法或属性之间的调用是模糊的。

添加重载方法以终止默认参数的使用

中断类型:源级安静语义更改

由于编译器将缺少默认参数值的方法调用转换为调用端具有默认值的显式调用,因此提供了与现有编译代码的兼容性;将为之前编译的所有代码找到具有正确签名的方法。

另一方面,不使用可选参数的调用现在被编译为对缺少可选参数的新方法的调用。这一切仍然正常工作,但是如果被调用的代码位于另一个程序集中,那么调用它的新编译代码现在依赖于该程序集中的新版本。部署调用重构代码的程序集而不部署重构代码所在的程序集将导致“方法未找到”异常。

更改前的API

  public int MyMethod(int mandatoryParameter, int optionalParameter = 0)
  {
     return mandatoryParameter + optionalParameter;
  }    

更改后的API

  public int MyMethod(int mandatoryParameter, int optionalParameter)
  {
     return mandatoryParameter + optionalParameter;
  }

  public int MyMethod(int mandatoryParameter)
  {
     return MyMethod(mandatoryParameter, 0);
  }

仍然可以工作的示例代码

  public int CodeNotDependentToNewVersion()
  {
     return MyMethod(5, 6); 
  }

在编译时依赖于新版本的示例代码

  public int CodeDependentToNewVersion()
  {
     return MyMethod(5); 
  }