tail head cat sleep
QR code linking to this page

Manual Pages  — GETCAP


cgetent, cgetset, cgetmatch, cgetcap, cgetnum, cgetstr, cgetustr, cgetfirst, cgetnext, cgetclose – capability database access routines



Standard C Library (libc, -lc)


#include <stdlib.h>

cgetent(char **buf, char **db_array, const char *name);

cgetset(const char *ent);

cgetmatch(const char *buf, const char *name);

char *
cgetcap(char *buf, const char *cap, int type);

cgetnum(char *buf, const char *cap, long *num);

cgetstr(char *buf, const char *cap, char **str);

cgetustr(char *buf, const char *cap, char **str);

cgetfirst(char **buf, char **db_array);

cgetnext(char **buf, char **db_array);



The cgetent() function extracts the capability name from the database specified by the NULL terminated file array db_array and returns a pointer to a malloc(3)amp;'d copy of it in buf. The cgetent() function will first look for files ending in .db (see cap_mkdb(1)) before accessing the ASCII file. The buf argument must be retained through all subsequent calls to cgetmatch(), cgetcap(), cgetnum(), cgetstr(), and cgetustr(), but may then be free(3)amp;'d. On success 0 is returned, 1 if the returned record contains an unresolved tc expansion, -1 if the requested record could not be found, -2 if a system error was encountered (could not open/read a file, etc.) also setting errno, and -3 if a potential reference loop is detected (see tc= comments below).

The cgetset() function enables the addition of a character buffer containing a single capability record entry to the capability database. Conceptually, the entry is added as the first ``file'' in the database, and is therefore searched first on the call to cgetent(). The entry is passed in ent. If ent is NULL, the current entry is removed from the database. A call to cgetset() must precede the database traversal. It must be called before the cgetent() call. If a sequential access is being performed (see below), it must be called before the first sequential access call (cgetfirst() or cgetnext() ), or be directly preceded by a cgetclose() call. On success 0 is returned and -1 on failure.

The cgetmatch() function will return 0 if name is one of the names of the capability record buf, -1 if not.

The cgetcap() function searches the capability record buf for the capability cap with type type. A type is specified using any single character. If a colon (`:') is used, an untyped capability will be searched for (see below for explanation of types). A pointer to the value of cap in buf is returned on success, NULL if the requested capability could not be found. The end of the capability value is signaled by a `:' or ASCII NUL (see below for capability database syntax).

The cgetnum() function retrieves the value of the numeric capability cap from the capability record pointed to by buf. The numeric value is returned in the long pointed to by num. 0 is returned on success, -1 if the requested numeric capability could not be found.

The cgetstr() function retrieves the value of the string capability cap from the capability record pointed to by buf. A pointer to a decoded, NUL terminated, malloc(3)amp;'d copy of the string is returned in the char * pointed to by str. The number of characters in the decoded string not including the trailing NUL is returned on success, -1 if the requested string capability could not be found, -2 if a system error was encountered (storage allocation failure).

The cgetustr() function is identical to cgetstr() except that it does not expand special characters, but rather returns each character of the capability string literally.

The cgetfirst() and cgetnext() functions comprise a function group that provides for sequential access of the NULL pointer terminated array of file names, db_array. The cgetfirst() function returns the first record in the database and resets the access to the first record. The cgetnext() function returns the next record in the database with respect to the record returned by the previous cgetfirst() or cgetnext() call. If there is no such previous call, the first record in the database is returned. Each record is returned in a malloc(3)amp;'d copy pointed to by buf. Tc expansion is done (see tc= comments below). Upon completion of the database 0 is returned, 1 is returned upon successful return of record with possibly more remaining (we have not reached the end of the database yet), 2 is returned if the record contains an unresolved tc expansion, -1 is returned if a system error occurred, and -2 is returned if a potential reference loop is detected (see tc= comments below). Upon completion of database (0 return) the database is closed.

The cgetclose() function closes the sequential access and frees any memory and file descriptors being used. Note that it does not erase the buffer pushed by a call to cgetset().


Capability databases are normally ASCII and may be edited with standard text editors. Blank lines and lines beginning with a `#' are comments and are ignored. Lines ending with a `\' indicate that the next line is a continuation of the current line; the `\' and following newline are ignored. Long lines are usually continued onto several physical lines by ending each line except the last with a `\'.

Capability databases consist of a series of records, one per logical line. Each record contains a variable number of `:'-separated fields (capabilities). Empty fields consisting entirely of white space characters (spaces and tabs) are ignored.

The first capability of each record specifies its names, separated by `|' characters. These names are used to reference records in the database. By convention, the last name is usually a comment and is not intended as a lookup tag. For example, the vt100 record from the termcap(5) database begins:

    d0|vt100|vt100-am|vt100am|dec vt100:

giving four names that can be used to access the record.

The remaining non-empty capabilities describe a set of (name, value) bindings, consisting of a names optionally followed by a typed value:

name typeless [boolean] capability name is present [true]
name amp;Tvalue
  capability ( name, amp;T) has value value
name@ no capability name exists
name Tamp;@ capability ( name, T) does not exist

Names consist of one or more characters. Names may contain any character except `:', but it is usually best to restrict them to the printable characters and avoid use of graphics like `#', `=', `%', `@', etc. Types are single characters used to separate capability names from their associated typed values. Types may be any character except a `:'. Typically, graphics like `#', `=', `%', etc.amp; are used. Values may be any number of characters and may contain any character except `:'.


Capability records describe a set of (name, value) bindings. Names may have multiple values bound to them. Different values for a name are distinguished by their types. The cgetcap() function will return a pointer to a value of a name given the capability name and the type of the value.

The types `#' and `=' are conventionally used to denote numeric and string typed values, but no restriction on those types is enforced. The functions cgetnum() and cgetstr() can be used to implement the traditional syntax and semantics of `#' and `='. Typeless capabilities are typically used to denote boolean objects with presence or absence indicating truth and false values respectively. This interpretation is conveniently represented by:

    (getcap(buf, name, ':') != NULL)

A special capability, tc= name, is used to indicate that the record specified by name should be substituted for the tc capability. Tc capabilities may interpolate records which also contain tc capabilities and more than one tc capability may be used in a record. A tc expansion scope (i.e., where the argument is searched for) contains the file in which the tc is declared and all subsequent files in the file array.

When a database is searched for a capability record, the first matching record in the search is returned. When a record is scanned for a capability, the first matching capability is returned; the capability :nameT@: will hide any following definition of a value of type T for name; and the capability :name@: will prevent any following values of name from being seen.

These features combined with tc capabilities can be used to generate variations of other databases and records by either adding new capabilities, overriding definitions with new definitions, or hiding following definitions via `@' capabilities.


example|an example of binding multiple values to names:\

The capability foo has two values bound to it (bar of type `%' and blah of type `^') and any other value bindings are hidden. The capability abc also has two values bound but only a value of type `$' is prevented from being defined in the capability record more.

        new|new_record|a modification of "old":\
        old|old_record|an old database record:\

The records are extracted by calling cgetent() with file1 preceding file2. In the capability record new in file1, fript=bar overrides the definition of fript=foo interpolated from the capability record old in file2, who-cares@ prevents the definition of any who-cares definitions in old from being seen, glork#200 is inherited from old, and blah and anything defined by the record extensions is added to those definitions in old. Note that the position of the fript=bar and who-cares@ definitions before tc=old is important here. If they were after, the definitions in old would take precedence.


Two types are predefined by cgetnum() and cgetstr():

nameamp;# number
  numeric capability name has value number
name= string
  string capability name has value string
nameamp;#@ the numeric capability name does not exist
nameamp;=@ the string capability name does not exist

Numeric capability values may be given in one of three numeric bases. If the number starts with either ‘0x’ or ‘0X’ it is interpreted as a hexadecimal number (both upper and lower case a-f may be used to denote the extended hexadecimal digits). Otherwise, if the number starts with a ‘0’ it is interpreted as an octal number. Otherwise the number is interpreted as a decimal number.

String capability values may contain any character. Non-printable ASCII codes, new lines, and colons may be conveniently represented by the use of escape sequences:
^X('X' & 037)control-X
\b, \B(ASCII 010)backspace
\t, \T(ASCII 011)tab
\n, \N(ASCII 012)line feed (newline)
\f, \F(ASCII 014)form feed
\r, \R(ASCII 015)carriage return
\e, \E(ASCII 027)escape
\c, \C(:)colon
\\(\)back slash
\nnn(ASCII octal nnn)

A `\' may be followed by up to three octal digits directly specifies the numeric code for a character. The use of ASCII NULs, while easily encoded, causes all sorts of problems and must be used with care since NULs are typically used to denote the end of strings; many applications use `\200' to represent a NUL.


The cgetent(), cgetset(), cgetmatch(), cgetnum(), cgetstr(), cgetustr(), cgetfirst(), and cgetnext() functions return a value greater than or equal to 0 on success and a value less than 0 on failure. The cgetcap() function returns a character pointer on success and a NULL on failure.

The cgetent(), and cgetset() functions may fail and set errno for any of the errors specified for the library functions: fopen(3), fclose(3), open(2), and close(2).

The cgetent(), cgetset(), cgetstr(), and cgetustr() functions may fail and set errno as follows:
  No memory to allocate.


cap_mkdb(1), malloc(3)


Colons (`:') cannot be used in names, types, or values.

There are no checks for tc= name loops in cgetent().

The buffer added to the database by a call to cgetset() is not unique to the database but is rather prepended to any database used.

GETCAP (3) March 22, 2002

tail head cat sleep
QR code linking to this page

Please direct any comments about this manual page service to Ben Bullock. Privacy policy.

Like a classics radio station whose play list spans decades, Unix simultaneously exhibits its mixed and dated heritage. There's Clash-era graphics interfaces; Beatles-era two-letter command names; and systems programs (for example, ps) whose terse and obscure output was designed for slow teletypes; Bing Crosby-era command editing (# and @ are still the default line editing commands), and Scott Joplin-era core dumps.
— The Unix Haters' handbook