CIL: Difference between revisions

Most of the time, developers will not be programming directly in CIL but use a higher level language such as [[C Sharp|C#]], [[C++/CLI]], [[Visual Basic .NET]] and [[J#]]. However, CIL has it'sits own assembly language which can be assembled by ILAsm (<code>ilasm.exe</code>). To display the byte code in assembly language, the CIL disassembler ILDAsm (<code>ildasm.exe</code>) is used. However, some specialized tools such as .NET Reflector can represent the byte code as any of the supported higher level languages, such as C#, including the classes, methods and other metadata.

CIL byte code can either be considered safe or unsafe. By default, all C# is safe, which means that the execution system knows where all the data is, all the time. Therefore it is not possible to use pointers in safe code. Thanks to this property, the most common security and stability problems are eliminated, such as buffer overflows and dangling pointers. A [[Garbage collection|garbage collector]] is employed to use the memory as efficient as possible.

Although pointers are not possible in safe code, CIL supports unsafe code which does have support for pointers. Unsafe code is contained in blocks of code which is marked 'unsafe', and in C# it uses the same syntax as C and C++. Objects can be ''pinned'' so that the [[Garbage collection|garbage collector]] doesn't move the data around and pointers remain valid. Using unsafe code is perfectly legal, although discouraged in normal application programming, and may be used to write kernel-level methods.