Go to the first, previous, next, last section, table of contents.


Camlidl generates stub code for interfacing Caml with C (as described in chapter "Interfacing with C" of the Objective Caml reference manual (http://caml.inria.fr/ocaml/htmlman/index.html)) from an IDL description of the C functions to be made available in Caml. Thus, Camlidl automates the most tedious task in interfacing C libraries with Caml programs. It can also be used to interface Caml programs with other languages, as long as those languages have a well-defined C interface.

In addition, Camlidl provides basic support for COM interfaces and components. It supports both using COM components (usually written in C++ or C) from Caml programs, and packaging Caml objects as COM components that can then be used from C++ or C.

What is IDL?

IDL stands for Interface Description Language. This is a generic term for a family of small languages that have been developed to provide type specifications for libraries written in C and C++. Those languages resembles C declarations (as found in C header files), with extra annotations to provide more precise types for the arguments and results of the functions.

The particular IDL used by Camlidl is inspired by Microsoft's IDL, which itself is an extension of the IDL found in DCE (The Open Group's Distributed Common Environment). The initial motivation for those IDLs was to automate the generation of stub code for remote procedure calls and network objects, where the arguments to the function are marshaled at the calling site, then sent across the network or through interprocess communications to a server process, which unmarshals the arguments, compute the function application, marshal the results, sends them back to the calling site, where they are unmarshaled and returned to the caller. IDLs were also found to be very useful for inter-language communications, since the same type information that guides the generation of marshaling stubs can be used to generate stubs to convert between the data representations of several languages.

What is COM?

COM is Microsoft's Common Object Model. It provides a set of programming conventions as well as system support for packaging C++ objects as executable components that can be used in other programs, either by dynamic linking of the component inside the program, or through interprocess or internetwork communications between the program and a remote server. COM components implement one or several interfaces, (similar to Caml object types or Java interfaces) identified by unique 128-bit interface identifiers (IIDs). COM specifies a standard protocol for reference counting of components, and for asking a component which interfaces it implements.

While the full range of COM services and third-party components is available only on Microsoft's Windows operating systems, the basic COM conventions can also be used on Unix and other operating systems to exchange objects between Caml and C or C++. Of particular interest is the encapsulation of Caml objects as COM components, which can then be used inside larger C or C++ applications; those applications do not need to know anything about Caml: they just call the component methods as if they were C++ methods or C functions, without knowing that they are actually implemented in Caml.

For more information about COM, see for instance Inside COM by Dale Rogerson (Microsoft Press), or the Microsoft developer Web site (http://msdn.microsoft.com).

Go to the first, previous, next, last section, table of contents.