| |
| |
| Preface | |
| |
| |
| |
| CLR/Framework Gotchas | |
| |
| |
| Type alias size doesn't match what you're familiar with | |
| |
| |
| struct and class differ in behavior | |
| |
| |
| Returning value types from a method/property is risky | |
| |
| |
| You can't force calls to your value-type constructors | |
| |
| |
| String concatenation is expensive | |
| |
| |
| Exceptions may go unhandled | |
| |
| |
| Uninitialized event handlers aren't treated gracefully | |
| |
| |
| Division operation isn't consistent between types | |
| |
| |
| Typeless ArrayList isn't type-safe | |
| |
| |
| Type.GetType() may not locate all types | |
| |
| |
| Public key reported by sn.exe is inconsistent | |
| |
| |
| |
| Visual Studio and Compiler Gotchas | |
| |
| |
| Compiler warnings may not be benign | |
| |
| |
| Ordering of catch processing isn't consist across languages | |
| |
| |
| Type.GetType() might fail at run-time | |
| |
| |
| rethrow isn't consistent | |
| |
| |
| Default of Option Strict (off) isn't good | |
| |
| |
| Versioning may lead to Serialization headaches | |
| |
| |
| Creating Web apps can be painful | |
| |
| |
| Naming XML documentation for IntelliSense support isn't intuitive | |
| |
| |
| |
| Language and API Gotchas | |
| |
| |
| Singleton isn't guaranteed process-wide | |
| |
| |
| Default performance of Data. ReadXML is poor | |
| |
| |
| enum lacks type-safety | |
| |
| |
| Copy Constructor hampers exensibility | |
| |
| |
| Clone() has limitations | |
| |
| |
| Access to static/Shared members isn't enforced consistently | |
| |
| |
| Details of exception may be hidden | |
| |
| |
| Object initialization sequence isn't consistent | |
| |
| |
| Polymorphism kicks in prematurely | |
| |
| |
| Unit testing private methods is tricky | |
| |
| |
| |
| Language Interoperability Gotchas | |
| |
| |
| Common Language Specification Compliance isn't the default | |
| |
| |
| Optional parameters break interoperability | |
| |
| |
| Mixing case between class members breaks interoperability | |
| |
| |
| Name collision with keywords breaks interoperability | |
| |
| |
| Defining an array isn't consistent | |
| |
| |
| |
| Garbage Collection Gotchas | |
| |
| |
| Writing Finalize() is rarely a good idea | |
| |
| |
| Releasing managed resources in Finalize() can wreak havoc | |
| |
| |
| Rules to invoke base.Finalize() aren't consistent | |
| |
| |
| Depending on Finalize() can tie up critical resources | |
| |
| |
| Using Finalize() on disposed objects is costly | |
| |
| |
| Implementing IDisposable isn't enough | |
| |
| |
| Using the Dispose Design Pattern doesn't guarantee cleanup | |
| |
| |
| |
| Inheritance and Polymorphism Gotchas | |
| |
| |
| Runtime Type Identification can hurt extensibility | |
| |
| |
| Using new/shadows causes "hideous hiding" | |
| |
| |
| Compilers are lenient toward forgotten override/overrides | |
| |
| |
| Compilers lean toward hiding virtual methods | |
| |
| |
| Exception handling can break polymorphism | |
| |
| |
| Signature mismatches can lead to method hiding | |
| |
| |
| |
| Multithreading Gotchas | |
| |
| |
| The Thread class supports some dangerous methods/properties | |
| |
| |
| Foreground threads may prevent a program from terminating | |
| |
| |
| Background threads don't terminate gracefully | |
| |
| |
| Interrupt () kicks in only when a thread is blocked | |
| |
| |
| ThreadAbortException-a hot potato | |
| |
| |
| Environment.Exit() brings down the CLR | |
| |
| |
| ResetAbort() may lead to surprises | |
| |
| |
| Abort() takes time to clean up | |
| |
| |
| Calling Type.GetType() may not return what you expect | |
| |
| |
| Locking on globally visible objects is too sweeping | |
| |
| |
| Threads from the thread pool are scarce | |
| |
| |
| Threads invoked using delegates behave like background threads | |
| |
| |
| Passing parameters to threads is tricky | |
| |
| |
| Exceptions thrown from threads in the pool are lost | |
| |
| |
| Accessing WinForm controls from arbitrary threads is dangerous | |
| |
| |
| Web-service proxy may fail when used for multiple asynchronous calls | |
| |
| |
| Raising events lacks thread-safety | |
| |
| |
| |
| COM-Interop and Enterprise Services Gotchas | |
| |
| |
| Release of COM object is confusing | |
| |
| |
| Using interface pointers after calling ReleaseComObject() will fail | |
| |
| |
| Cross-apartment calls are expensive | |
| |
| |
| Default apartment of main thread is inconsistent across languages | |
| |
| |
| STAThread attribute may have no effect on your methods | |
| |
| |
| Spattering access to COM components makes code hard to maintain | |
| |
| |
| Auto-generating GUID for your classes leads to versioning woes | |
| |
| |
| All but one of the ClassInterface options are ineffective | |
| |
| |
| Simply turning the switch for COM interop is dangerous | |
| |
| |
| ServicedComponents implemented inconsistently on XP and 2003 | |
| |
| |
| AutoComplete comes with undesirable side effects | |
| |
| |
| Appendix | |
| |
| |
| Index | |