OpenBSD manual page server

Manual Page Search Parameters

GLOB(3) Library Functions Manual GLOB(3)

glob, globfreegenerate pathnames matching a pattern

#include <glob.h>

int
glob(const char *pattern, int flags, const int (*errfunc)(const char *, int), glob_t *pglob);

void
globfree(glob_t *pglob);

The () function is a pathname generator that implements the rules for file name pattern matching used by the shell.

The include file <glob.h> defines the structure type glob_t, which contains at least the following fields:

typedef struct {
	int gl_pathc;	 /* count of total paths so far */
	int gl_matchc;	 /* count of paths matching pattern */
	int gl_offs;	 /* reserved at beginning of gl_pathv */
	int gl_flags;	 /* returned flags */
	char **gl_pathv; /* list of paths matching pattern */
} glob_t;

The argument pattern is a pointer to a pathname pattern to be expanded. () matches all accessible pathnames against the pattern and creates a list of the pathnames that match. In order to have access to a pathname, glob() requires search permission on every component of a path except the last and read permission on each directory of any filename component of pattern that contains any of the special characters ‘*’, ‘?’, or ‘[’.

The number of matched pathnames is stored in the gl_pathc field, and a pointer to a list of pointers to pathnames in the gl_pathv field. The first pointer after the last pathname is NULL. If the pattern does not match any pathnames, the returned number of matched paths is set to zero.

It is the caller's responsibility to create the structure pointed to by pglob. The () function allocates other space as needed, including the memory pointed to by gl_pathv.

The argument flags is used to modify the behavior of (). The value of flags is the bitwise inclusive OR of any of the following values defined in <glob.h>:

Append pathnames generated to the ones from a previous call (or calls) to glob(). The value of gl_pathc will be the total matches found by this call and the previous call(s). The pathnames are appended to, not merged with the pathnames returned by the previous call(s). Between calls, the caller must not change the setting of the GLOB_DOOFFS flag, nor change the value of gl_offs when GLOB_DOOFFS is set, nor (obviously) call globfree() for pglob.
Make use of the gl_offs field. If this flag is set, gl_offs is used to specify how many null pointers to prepend to the beginning of the gl_pathv field. In other words, gl_pathv will point to gl_offs null pointers, followed by gl_pathc pathname pointers, followed by a null pointer.
Causes glob() to return when it encounters a directory that it cannot open or read. Ordinarily, glob() continues to find matches.
Each pathname that is a directory that matches pattern has a slash appended.
If pattern does not match any pathname, then glob() returns a list consisting of only pattern, with the number of total pathnames set to 1, and the number of matched pathnames set to 0.
Normally, every occurrence of a backslash (‘\’) followed by a character in pattern is replaced by that character. This is done to negate any special meaning for the character. If the GLOB_NOESCAPE flag is set, a backslash character is treated as an ordinary character.
By default, the pathnames are sorted in ascending ASCII order; this flag prevents that sorting (speeding up glob()).

The following values may also be included in flags, however, they are non-standard extensions to IEEE Std 1003.2 (“POSIX.2”).

The following additional fields in the pglob structure have been initialized with alternate functions for () to use to open, read, and close directories and to get stat information on names found in those directories:
	void *(*gl_opendir)(const char *);
	struct dirent *(*gl_readdir)(void *);
	void (*gl_closedir)(void *);
	int (*gl_lstat)(const char *, struct stat *);
	int (*gl_stat)(const char *, struct stat *);

This extension is provided to allow programs such as restore(8) to provide globbing from directories stored on tape.

Pre-process the pattern string to expand ‘{pat,pat,...}’ strings like csh(1). The pattern ‘{}’ is left unexpanded for historical reasons. (csh(1) does the same thing to ease typing of find(1) patterns.)
Retain a copy of the stat(2) information retrieved for matching paths in the gl_statv array:
struct stat **gl_statv;

This option may be used to avoid lstat(2) lookups in cases where they are expensive.

Set by the () function if the pattern included globbing characters. See the description of the usage of the gl_matchc structure member for more details.
Is the same as GLOB_NOCHECK but it only appends the pattern if it does not contain any of the special characters ‘*’, ‘?’, or ‘[’. GLOB_NOMAGIC is provided to simplify implementing the historic csh(1) globbing behavior and should probably not be used anywhere else.
This option has no effect and is included for backwards compatibility with older sources.
Expand patterns that start with ‘~’ to user name home directories.
Limit the amount of memory used to store matched strings to 64K, the number of stat(2) calls to 2048, and the number of readdir(3) calls to 16K. This option should be set for programs that can be coerced to a denial of service attack via patterns that expand to a very large number of matches, such as a long string of ‘*/../*/..’.

If, during the search, a directory is encountered that cannot be opened or read and errfunc is non-null, () calls (*errfunc)(path, errno). This may be unintuitive: a pattern like “*/Makefile” will try to stat(2) “foo/Makefile” even if “foo” is not a directory, resulting in a call to errfunc. The error routine can suppress this action by testing for ENOENT and ENOTDIR; however, the GLOB_ERR flag will still cause an immediate return when this happens.

If errfunc returns non-zero, () stops the scan and returns GLOB_ABORTED after setting gl_pathc and gl_pathv to reflect any paths already matched. This also happens if an error is encountered and GLOB_ERR is set in flags, regardless of the return value of errfunc, if called. If GLOB_ERR is not set and either errfunc is NULL or errfunc returns zero, the error is ignored.

The () function frees any space associated with pglob from a previous call(s) to glob().

On successful completion, glob() returns zero. In addition the fields of pglob contain the values described below:

gl_pathc
Contains the total number of matched pathnames so far. This includes other matches from previous invocations of glob() if GLOB_APPEND was specified.
gl_matchc
Contains the number of matched pathnames in the current invocation of glob().
gl_flags
Contains a copy of the flags parameter with the bit GLOB_MAGCHAR set if pattern contained any of the special characters ‘*’, ‘?’, or ‘[’, cleared if not.
gl_pathv
Contains a pointer to a null-terminated list of matched pathnames. However, if gl_pathc is zero, the contents of gl_pathv are undefined.
gl_statv
If the GLOB_KEEPSTAT flag was set, gl_statv contains a pointer to a null-terminated list of matched stat(2) objects corresponding to the paths in gl_pathc.

If glob() terminates due to an error, it sets errno and returns one of the following non-zero constants, which are defined in the include file <glob.h>:

An attempt to allocate memory failed, or if errno was 0 GLOB_LIMIT was specified in the flags and ARG_MAX or more patterns were matched.
The scan was stopped because an error was encountered and either GLOB_ERR was set, or (*errfunc)() returned non-zero.
The pattern did not match a pathname and GLOB_NOCHECK was not set.
The requested function is not supported by this version of glob().

The arguments pglob->gl_pathc and pglob->gl_pathv are still set as specified above.

A rough equivalent of ‘ls -l *.c *.h’ can be obtained with the following code:

glob_t g;

g.gl_offs = 2;
glob("*.c", GLOB_DOOFFS, NULL, &g);
glob("*.h", GLOB_DOOFFS | GLOB_APPEND, NULL, &g);
g.gl_pathv[0] = "ls";
g.gl_pathv[1] = "-l";
execvp("ls", g.gl_pathv);

The glob() function may fail and set errno for any of the errors specified for the library routines stat(2), closedir(3), opendir(3), readdir(3), malloc(3), and free(3).

sh(1), fnmatch(3), regex(3), glob(7)

The glob() function is expected to conform to IEEE Std 1003.2 (“POSIX.2”) and X/Open Portability Guide Issue 4, Version 2 (“XPG4.2”). Note, however, that the flags GLOB_ALTDIRFUNC, GLOB_BRACE, GLOB_KEEPSTAT, GLOB_MAGCHAR, GLOB_NOMAGIC, GLOB_QUOTE, GLOB_TILDE, and GLOB_LIMIT and the fields gl_matchc, gl_statv and gl_flags should not be used by applications striving for strict standards conformance.

The glob() and globfree() functions first appeared in 4.4BSD.

Patterns longer than PATH_MAX may cause unchecked errors.

January 21, 2014 OpenBSD-6.0