C (programming language)

Many later languages have borrowed directly or indirectly from C, including C++, C#, Unix's C shell, D, Go, Java, JavaScript (including transpilers), Julia, Limbo, LPC, Objective-C, Perl, PHP, Python, Ruby, Rust, Swift, Verilog and SystemVerilog (hardware description languages).[5] These languages have drawn many of their control structures and other basic features from C. Most of them (Python being a dramatic exception) also express highly similar syntax to C, and they tend to combine the recognizable expression and statement syntax of C with underlying type systems, data models, and semantics that can be radically different.

The origin of C is closely tied to the development of the Unix operating system, originally implemented in assembly language on a PDP-7 by Dennis Ritchie and Ken Thompson, incorporating several ideas from colleagues. Eventually, they decided to port the operating system to a PDP-11. The original PDP-11 version of Unix was also developed in assembly language.[7]

Thompson desired a programming language to make utilities for the new platform. At first, he tried to make a Fortran compiler, but soon gave up the idea. Instead, he created a cut-down version of the recently developed BCPL systems programming language. The official description of BCPL was not available at the time,[12] and Thompson modified the syntax to be less wordy, and similar to a simplified ALGOL known as SMALGOL.[13] The result was what Thompson called B.[7] He described B as "BCPL semantics with a lot of SMALGOL syntax".[13] Like BCPL, B had a bootstrapping compiler to facilitate porting to new machines.[13] However, few utilities were ultimately written in B because it was too slow, and could not take advantage of PDP-11 features such as byte addressability.

In 1971, Ritchie started to improve B, to utilise the features of the more-powerful PDP-11. A significant addition was a character type. He called this New B.[13] Thompson started to use NB to write the Unix kernel, and his requirements shaped the direction of the language development.[13][14] Through to 1972, richer types were added to the NB language: NB had arrays of int and char. Pointers, the ability to generate pointers to other types, arrays of all types, and types to be returned from functions were all also added. Arrays within expressions became pointers. A new compiler was written, and the language was renamed C.[7]

The C compiler and some utilities made with it were included in Version 2 Unix, which is also known as Research Unix.[15]

At Version 4 Unix, released in November 1973, the Unix kernel was extensively re-implemented in C.[7] By this time, the C language had acquired some powerful features such as struct types.

The preprocessor was introduced around 1973 at the urging of Alan Snyder and also in recognition of the usefulness of the file-inclusion mechanisms available in BCPL and PL/I. Its original version provided only included files and simple string replacements: #include and #define of parameterless macros. Soon after that, it was extended, mostly by Mike Lesk and then by John Reiser, to incorporate macros with arguments and conditional compilation.[7]

Unix was one of the first operating system kernels implemented in a language other than assembly. Earlier instances include the Multics system (which was written in PL/I) and Master Control Program (MCP) for the Burroughs B5000 (which was written in ALGOL) in 1961. In around 1977, Ritchie and Stephen C. Johnson made further changes to the language to facilitate portability of the Unix operating system. Johnson's Portable C Compiler served as the basis for several implementations of C on new platforms.[14]

The cover of the book The C Programming Language, first edition, by Brian Kernighan and Dennis Ritchie

In 1978, Brian Kernighan and Dennis Ritchie published the first edition of The C Programming Language.[16] This book, known to C programmers as K&R, served for many years as an informal specification of the language. The version of C that it describes is commonly referred to as "K&R C". As this was released in 1978, it is also referred to as C78.[17] The second edition of the book[18] covers the later ANSI C standard, described below.

K&R introduced several language features:

• Standard I/O library
• long int data type
• unsigned int data type
• Compound assignment operators of the form =op (such as =-) were changed to the form op= (that is, -=) to remove the semantic ambiguity created by constructs such as i=-10, which had been interpreted as i =- 10 (decrement i by 10) instead of the possibly intended i = -10 (let i be −10).

Even after the publication of the 1989 ANSI standard, for many years K&R C was still considered the "lowest common denominator" to which C programmers restricted themselves when maximum portability was desired, since many older compilers were still in use, and because carefully written K&R C code can be legal Standard C as well.

In early versions of C, only functions that return types other than int must be declared if used before the function definition; functions used without prior declaration were presumed to return type int.

For example:

long some_function(); /* This is a function declaration, so the compiler can know the name and return type of this function. */
/* int */ other_function(); /* Another function declaration.  There is an implicit 'int' type here since we're talking about early version of C.  It's commented out here to show where it could go in later variants. */

/* int */ calling_function() /* this is a function definition, including the body of the code following in the { curly brackets } the return type is 'int', but this is implicit so no need to state 'int' when using this early version of C */
{
    long test1;
    register /* int */ test2; /* again, note the 'int' is not required here, and shown as */
                              /* a comment just to illustrate where it would be required in later variants of C. */
                              /* The 'register' keyword indicates to the compiler that this variable should */
                              /* ideally be stored in a register as opposed to within the stack frame. */
    test1 = some_function();
    if (test1 > 1)
          test2 = 0;
    else
          test2 = other_function();
    return test2;
}

The int type specifiers which are commented out could be omitted in K&R C, but are required in later standards.

Since K&R function declarations did not include any information about function arguments, function parameter type checks were not performed, although some compilers would issue a warning message if a local function was called with the wrong number of arguments, or if multiple calls to an external function used different numbers or types of arguments. Separate tools such as Unix's lint utility were developed that (among other things) could check for consistency of function use across multiple source files.

In the years following the publication of K&R C, several features were added to the language, supported by compilers from AT&T (in particular PCC[19]) and some other vendors. These included:

• void functions (i.e., functions with no return value)
• functions returning struct or union types (previously only a single pointer, integer or float could be returned)
• assignment for struct data types
• enumerated types (previously, preprocessor definitions for integer fixed values were used, e.g. #define GREEN 3)

The large number of extensions and lack of agreement on a standard library, together with the language popularity and the fact that not even the Unix compilers precisely implemented the K&R specification, led to the necessity of standardization.

During the late 1970s and 1980s, versions of C were implemented for a wide variety of mainframe computers, minicomputers, and microcomputers, including the IBM PC, as its popularity began to increase significantly.

In 1983, the American National Standards Institute (ANSI) formed a committee, X3J11, to establish a standard specification of C. X3J11 based the C standard on the Unix implementation; however, the non-portable portion of the Unix C library was handed off to the IEEE working group 1003 to become the basis for the 1988 POSIX standard. In 1989, the C standard was ratified as ANSI X3.159-1989 "Programming Language C". This version of the language is often referred to as ANSI C, Standard C, or sometimes C89.

In 1990, the ANSI C standard (with formatting changes) was adopted by the International Organization for Standardization (ISO) as ISO/IEC 9899:1990, which is sometimes called C90. Therefore, the terms "C89" and "C90" refer to the same programming language.

ANSI, like other national standards bodies, no longer develops the C standard independently, but defers to the international C standard, maintained by the working group ISO/IEC JTC1/SC22/WG14. National adoption of an update to the international standard typically occurs within a year of ISO publication.

One of the aims of the C standardization process was to produce a superset of K&R C, incorporating many of the subsequently introduced unofficial features. The standards committee also included several additional features such as function prototypes (borrowed from C++), void pointers, support for international character sets and locales, and preprocessor enhancements. Although the syntax for parameter declarations was augmented to include the style used in C++, the K&R interface continued to be permitted, for compatibility with existing source code.

C89 is supported by current C compilers, and most modern C code is based on it. Any program written only in Standard C and without any hardware-dependent assumptions will run correctly on any platform with a conforming C implementation, within its resource limits. Without such precautions, programs may compile only on a certain platform or with a particular compiler, due, for example, to the use of non-standard libraries, such as GUI libraries, or to a reliance on compiler- or platform-specific attributes such as the exact size of data types and byte endianness.

In cases where code must be compilable by either standard-conforming or K&R C-based compilers, the __STDC__ macro can be used to split the code into Standard and K&R sections to prevent the use on a K&R C-based compiler of features available only in Standard C.

After the ANSI/ISO standardization process, the C language specification remained relatively static for several years. In 1995, Normative Amendment 1 to the 1990 C standard (ISO/IEC 9899/AMD1:1995, known informally as C95) was published, to correct some details and to add more extensive support for international character sets.[20]

The C standard was further revised in the late 1990s, leading to the publication of ISO/IEC 9899:1999 in 1999, which is commonly referred to as "C99". It has since been amended three times by Technical Corrigenda.[21]

C99 introduced several new features, including inline functions, several new data types (including long long int and a complex type to represent complex numbers), variable-length arrays and flexible array members, improved support for IEEE 754 floating point, support for variadic macros (macros of variable arity), and support for one-line comments beginning with //, as in BCPL or C++. Many of these had already been implemented as extensions in several C compilers.

C99 is for the most part backward compatible with C90, but is stricter in some ways; in particular, a declaration that lacks a type specifier no longer has int implicitly assumed. A standard macro __STDC_VERSION__ is defined with value 199901L to indicate that C99 support is available. GCC, Solaris Studio, and other C compilers now support many or all of the new features of C99. The C compiler in Microsoft Visual C++, however, implements the C89 standard and those parts of C99 that are required for compatibility with C++11.[22]

In addition, support for Unicode identifiers (variable / function names) in the form of escaped characters (e.g. \U0001f431) is now required. Support for raw Unicode names is optional.

In 2007, work began on another revision of the C standard, informally called "C1X" until its official publication of ISO/IEC 9899:2011 on 2011-12-08. The C standards committee adopted guidelines to limit the adoption of new features that had not been tested by existing implementations.

The C11 standard adds numerous new features to C and the library, including type generic macros, anonymous structures, improved Unicode support, atomic operations, multi-threading, and bounds-checked functions. It also makes some portions of the existing C99 library optional, and improves compatibility with C++. The standard macro __STDC_VERSION__ is defined as 201112L to indicate that C11 support is available.

Published in June 2018 as ISO/IEC 9899:2018, C17 is the current standard for the C programming language. It introduces no new language features, only technical corrections, and clarifications to defects in C11. The standard macro __STDC_VERSION__ is defined as 201710L.

C2x is an informal name for the next (after C17) major C language standard revision. It is expected to be voted on in 2023 and would therefore be called C23.[23]

Historically, embedded C programming requires nonstandard extensions to the C language in order to support exotic features such as fixed-point arithmetic, multiple distinct memory banks, and basic I/O operations.

In 2008, the C Standards Committee published a technical report extending the C language[24] to address these issues by providing a common standard for all implementations to adhere to. It includes a number of features not available in normal C, such as fixed-point arithmetic, named address spaces, and basic I/O hardware addressing.

The basic C source character set includes the following characters:

• Lowercase and uppercase letters of ISO Basic Latin Alphabet: a–z A–Z
• Decimal digits: 0–9
• Graphic characters: ! " # % & ' ( ) * + , - . / : ; < = > ? [ \ ] ^ _ { | } ~
• Whitespace characters: space, horizontal tab, vertical tab, form feed, newline

Newline indicates the end of a text line; it need not correspond to an actual single character, although for convenience C treats it as one.

Additional multi-byte encoded characters may be used in string literals, but they are not entirely portable. The latest C standard (C11) allows multi-national Unicode characters to be embedded portably within C source text by using \uXXXX or \UXXXXXXXX encoding (where the X denotes a hexadecimal character), although this feature is not yet widely implemented.

The basic C execution character set contains the same characters, along with representations for alert, backspace, and carriage return. Run-time support for extended character sets has increased with each revision of the C standard.

C89 has 32 reserved words, also known as keywords, which are the words that cannot be used for any purposes other than those for which they are predefined:

C99 reserved five more words:

C11 reserved seven more words:[28]

Most of the recently reserved words begin with an underscore followed by a capital letter, because identifiers of that form were previously reserved by the C standard for use only by implementations. Since existing program source code should not have been using these identifiers, it would not be affected when C implementations started supporting these extensions to the programming language. Some standard headers do define more convenient synonyms for underscored identifiers. The language previously included a reserved word called entry, but this was seldom implemented, and has now been removed as a reserved word.[29]

C supports a rich set of operators, which are symbols used within an expression to specify the manipulations to be performed while evaluating that expression. C has operators for:

• arithmetic: +, -, *, /, %
• assignment: =
• augmented assignment: +=, -=, *=, /=, %=, &=, |=, ^=, <<=, >>=
• bitwise logic: ~, &, |, ^
• bitwise shifts: <<, >>
• boolean logic: !, &&, ||
• conditional evaluation: ? :
• equality testing: ==, !=
• calling functions: ( )
• increment and decrement: ++, --
• member selection: ., ->
• object size: sizeof
• order relations: <, <=, >, >=
• reference and dereference: &, *, [ ]
• sequencing: ,
• subexpression grouping: ( )
• type conversion: (typename)

C uses the operator = (used in mathematics to express equality) to indicate assignment, following the precedent of Fortran and PL/I, but unlike ALGOL and its derivatives. C uses the operator == to test for equality. The similarity between these two operators (assignment and equality) may result in the accidental use of one in place of the other, and in many cases, the mistake does not produce an error message (although some compilers produce warnings). For example, the conditional expression if (a == b + 1) might mistakenly be written as if (a = b + 1), which will be evaluated as true if a is not zero after the assignment.[30]

The C operator precedence is not always intuitive. For example, the operator == binds more tightly than (is executed prior to) the operators & (bitwise AND) and | (bitwise OR) in expressions such as x & 1 == 0, which must be written as (x & 1) == 0 if that is the coder's intent.[31]

C supports the use of pointers, a type of reference that records the address or location of an object or function in memory. Pointers can be dereferenced to access data stored at the address pointed to, or to invoke a pointed-to function. Pointers can be manipulated using assignment or pointer arithmetic. The run-time representation of a pointer value is typically a raw memory address (perhaps augmented by an offset-within-word field), but since a pointer's type includes the type of the thing pointed to, expressions including pointers can be type-checked at compile time. Pointer arithmetic is automatically scaled by the size of the pointed-to data type.

Pointers are used for many purposes in C. Text strings are commonly manipulated using pointers into arrays of characters. Dynamic memory allocation is performed using pointers; the result of a malloc is usually cast to the data type of the data to be stored. Many data types, such as trees, are commonly implemented as dynamically allocated struct objects linked together using pointers. Pointers to other pointers are often used in multi-dimensional arrays and arrays of struct objects. Pointers to functions (function pointers) are useful for passing functions as arguments to higher-order functions (such as qsort or bsearch), in dispatch tables, or as callbacks to event handlers .[33]

A null pointer value explicitly points to no valid location. Dereferencing a null pointer value is undefined, often resulting in a segmentation fault. Null pointer values are useful for indicating special cases such as no "next" pointer in the final node of a linked list, or as an error indication from functions returning pointers. In appropriate contexts in source code, such as for assigning to a pointer variable, a null pointer constant can be written as 0, with or without explicit casting to a pointer type, or as the NULL macro defined by several standard headers. In conditional contexts, null pointer values evaluate to false, while all other pointer values evaluate to true.

Void pointers (void *) point to objects of unspecified type, and can therefore be used as "generic" data pointers. Since the size and type of the pointed-to object is not known, void pointers cannot be dereferenced, nor is pointer arithmetic on them allowed, although they can easily be (and in many contexts implicitly are) converted to and from any other object pointer type.[33]

Careless use of pointers is potentially dangerous. Because they are typically unchecked, a pointer variable can be made to point to any arbitrary location, which can cause undesirable effects. Although properly used pointers point to safe places, they can be made to point to unsafe places by using invalid pointer arithmetic; the objects they point to may continue to be used after deallocation (dangling pointers); they may be used without having been initialized (wild pointers); or they may be directly assigned an unsafe value using a cast, union, or through another corrupt pointer. In general, C is permissive in allowing manipulation of and conversion between pointer types, although compilers typically provide options for various levels of checking. Some other programming languages address these problems by using more restrictive reference types.

Array types in C are traditionally of a fixed, static size specified at compile time. The more recent C99 standard also allows a form of variable-length arrays. However, it is also possible to allocate a block of memory (of arbitrary size) at run-time, using the standard library's malloc function, and treat it as an array.

Since arrays are always accessed (in effect) via pointers, array accesses are typically not checked against the underlying array size, although some compilers may provide bounds checking as an option.[36][37] Array bounds violations are therefore possible and can lead to various repercussions, including illegal memory accesses, corruption of data, buffer overruns, and run-time exceptions.

C does not have a special provision for declaring multi-dimensional arrays, but rather relies on recursion within the type system to declare arrays of arrays, which effectively accomplishes the same thing. The index values of the resulting "multi-dimensional array" can be thought of as increasing in row-major order. Multi-dimensional arrays are commonly used in numerical algorithms (mainly from applied linear algebra) to store matrices. The structure of the C array is well suited to this particular task. However, in early versions of C the bounds of the array must be known fixed values or else explicitly passed to any subroutine that requires them, and dynamically sized arrays of arrays cannot be accessed using double indexing. (A workaround for this was to allocate the array with an additional "row vector" of pointers to the columns.) C99 introduced "variable-length arrays" which address this issue.

The following example using modern C (C99 or later) shows allocation of a two-dimensional array on the heap and the use of multi-dimensional array indexing for accesses (which can use bounds-checking on many C compilers):

int func(int N, int M)
{
  float (*p)[N][M] = malloc(sizeof *p);
  if (!p)
    return -1;
  for (int i = 0; i < N; i++)
    for (int j = 0; j < M; j++)
      (*p)[i][j] = i + j;
  print_array(N, M, p);
  free(p);
  return 1;
}

And here is a similar implementation using C99's Auto VLA feature:

int func(int N, int M)
{
  // Caution: checks should be made to ensure N*M*sizeof(float) does NOT exceed limitations for auto VLAs and is within available size of stack.
  float p[N][M]; // auto VLA is held on the stack, and sized when the function is invoked
  for (int i = 0; i < N; i++)
    for (int j = 0; j < M; j++)
      p[i][j] = i + j;
  // no need to free(p) since it will disappear when the function exits, along with the rest of the stack frame
  return 1;
}

The subscript notation x[i] (where x designates a pointer) is syntactic sugar for *(x+i).[38] Taking advantage of the compiler's knowledge of the pointer type, the address that x + i points to is not the base address (pointed to by x) incremented by i bytes, but rather is defined to be the base address incremented by i multiplied by the size of an element that x points to. Thus, x[i] designates the i+1th element of the array.

Furthermore, in most expression contexts (a notable exception is as operand of sizeof), an expression of array type is automatically converted to a pointer to the array's first element. This implies that an array is never copied as a whole when named as an argument to a function, but rather only the address of its first element is passed. Therefore, although function calls in C use pass-by-value semantics, arrays are in effect passed by reference.

The total size of an array x can be determined by applying sizeof to an expression of array type. The size of an element can be determined by applying the operator sizeof to any dereferenced element of an array A, as in n = sizeof A[0]. Thus, the number of elements in a declared array A can be determined as sizeof A / sizeof A[0]. Note, that if only a pointer to the first element is available as it is often the case in C code because of the automatic conversion described above, the information about the full type of the array and its length are lost.

File input and output (I/O) is not part of the C language itself but instead is handled by libraries (such as the C standard library) and their associated header files (e.g. stdio.h). File handling is generally implemented through high-level I/O which works through streams. A stream is from this perspective a data flow that is independent of devices, while a file is a concrete device. The high-level I/O is done through the association of a stream to a file. In the C standard library, a buffer (a memory area or queue) is temporarily used to store data before it is sent to the final destination. This reduces the time spent waiting for slower devices, for example a hard drive or solid state drive. Low-level I/O functions are not part of the standard C library but are generally part of "bare metal" programming (programming that's independent of any operating system such as most embedded programming). With few exceptions, implementations include low-level I/O.

The C Programming Language

C is widely used for systems programming in implementing operating systems and embedded system applications.[40] This is for several reasons:

• The code generated after compilation doesn't demand many system features, and can be invoked from some boot code in a straightforward manner – it's simple to execute.
• The C language statements and expressions typically map well on to sequences of instructions for the target processor, and consequently there is a low run-time demand on system resources – it's fast to execute.
• With its rich set of operators, the C language can utilise many of the features of target CPUs. Where a particular CPU has more esoteric instructions, a language variant can be constructed with perhaps intrinsic functions to exploit those instructions – it can use practically all the target CPU's features.
• The language makes it easy to overlay structures onto blocks of binary data, allowing the data to be comprehended, navigated and modified – it can write data structures, even file systems.
• The language supports a rich set of operators, including bit manipulation, for integer arithmetic and logic, and perhaps different sizes of floating point numbers – it can process appropriately-structured data effectively.
• C is a fairly small language, with only a handful of statements, and without too many features that generate extensive target code – it is comprehensible.
• C has direct control over memory allocation and deallocation, which gives reasonable efficiency and predictable timing to memory-handling operations, without any concerns for sporadic stop-the-world garbage collection events – it has predictable performance.
• Platform hardware can be accessed with pointers and type punning, so system-specific features (e.g. Control/Status Registers, I/O registers) can be configured and used with code written in C – it interacts well with the platform it's running on.
• Depending on the linker and environment, C code can also call libraries written in assembly language, and may be called from assembly language – it interoperates well with other lower-level code.
• C and its calling conventions and linker structures are commonly used in conjunction with other high-level languages, with calls both to C and from C supported – it interoperates well with other high-level code.
• C has a very mature and broad ecosystem, including libraries, frameworks, open source compilers, debuggers and utilities, and is the de facto standard. It's likely the drivers already exist in C, or that there is a similar CPU architecture as a back-end of a C compiler, so there is reduced incentive to choose another language.

Historically, C was sometimes used for web development using the Common Gateway Interface (CGI) as a "gateway" for information between the web application, the server, and the browser.[41] C may have been chosen over interpreted languages because of its speed, stability, and near-universal availability.[42] It is no longer common practice for web development to be done in C,[43] and many other web development tools exist.

A consequence of C's wide availability and efficiency is that compilers, libraries and interpreters of other programming languages are often implemented in C.[44] For example, the reference implementations of Python,[45] Perl,[46] Ruby,[47] and PHP[48] are written in C.

C enables programmers to create efficient implementations of algorithms and data structures, because the layer of abstraction from hardware is thin, and its overhead is low, an important criterion for computationally intensive programs. For example, the GNU Multiple Precision Arithmetic Library, the GNU Scientific Library, Mathematica, and MATLAB are completely or partially written in C. Many languages support calling library functions in C, for example, the Python-based framework NumPy uses C for the high-performance and hardware-interacting aspects.

C is sometimes used as an intermediate language by implementations of other languages. This approach may be used for portability or convenience; by using C as an intermediate language, additional machine-specific code generators are not necessary. C has some features, such as line-number preprocessor directives and optional superfluous commas at the end of initializer lists, that support compilation of generated code. However, some of C's shortcomings have prompted the development of other C-based languages specifically designed for use as intermediate languages, such as C--. Also, contemporary major compilers GCC and LLVM both feature an intermediate representation that is not C, and those compilers support front ends for many languages including C.

C has also been widely used to implement end-user applications. However, such applications can also be written in newer, higher-level languages.