Section (3) glob

Linux manual pages Section 3  


glob, globfree — find pathnames matching a pattern, free memory from glob()


#include <glob.h>
int glob( const char *pattern,
  int flags,
  int (*errfunc)(const char *epath, int eerrno),
  glob_t *pglob);
void globfree( glob_t *pglob);


The glob() function searches for all the pathnames matching pattern according to the rules used by the shell (see glob(7)). No tilde expansion or parameter substitution is done; if you want these, use wordexp(3).

The globfree() function frees the dynamically allocated storage from an earlier call to glob().

The results of a glob() call are stored in the structure pointed to by pglob. This structure is of type glob_t (declared in <glob.h> and includes the following elements defined by POSIX.2 (more may be present as an extension):

typedef struct {
  size_t   gl_pathc;
/* Count of paths matched so far  */
  char ** gl_pathv;
/* List of matched pathnames.  */
  size_t   gl_offs;
/* Slots to reserve in gl_pathv.  */
} glob_t;

Results are stored in dynamically allocated storage.

The argument flags is made up of the bitwise OR of zero or more the following symbolic constants, which modify the behavior of glob():


Return upon a read error (because a directory does not have read permission, for example). By default, glob() attempts carry on despite errors, reading all of the directories that it can.


Append a slash to each path which corresponds to a directory.


Don_zsingle_quotesz_t sort the returned pathnames. The only reason to do this is to save processing time. By default, the returned pathnames are sorted.


Reserve pglob−>gl_offs slots at the beginning of the list of strings in pglob−>pathv. The reserved slots contain null pointers.


If no pattern matches, return the original pattern. By default, glob() returns GLOB_NOMATCH if there are no matches.


Append the results of this call to the vector of results returned by a previous call to glob(). Do not set this flag on the first invocation of glob().


Don_zsingle_quotesz_t allow backslash (_zsingle_quotesz_\_zsingle_quotesz_) to be used as an escape character. Normally, a backslash can be used to quote the following character, providing a mechanism to turn off the special meaning metacharacters.

flags may also include any of the following, which are GNU extensions and not defined by POSIX.2:


Allow a leading period to be matched by metacharacters. By default, metacharacters can_zsingle_quotesz_t match a leading period.


Use alternative functions pglob−>gl_closedir, pglob−>gl_readdir, pglob−>gl_opendir, pglob−>gl_lstat, and pglob−>gl_stat for filesystem access instead of the normal library functions.


Expand csh(1) style brace expressions of the form {a,b}. Brace expressions can be nested. Thus, for example, specifying the pattern {foo/{,cat,dog},bar} would return the same results as four separate glob() calls using the strings: foo/, foo/cat, foo/dog, and bar.


If the pattern contains no metacharacters, then it should be returned as the sole matching word, even if there is no file with that name.


Carry out tilde expansion. If a tilde (_zsingle_quotesz_~_zsingle_quotesz_) is the only character in the pattern, or an initial tilde is followed immediately by a slash (_zsingle_quotesz_/_zsingle_quotesz_), then the home directory of the caller is substituted for the tilde. If an initial tilde is followed by a username (e.g., ~andrea/bin), then the tilde and username are substituted by the home directory of that user. If the username is invalid, or the home directory cannot be determined, then no substitution is performed.


This provides behavior similar to that of GLOB_TILDE. The difference is that if the username is invalid, or the home directory cannot be determined, then instead of using the pattern itself as the name, glob() returns GLOB_NOMATCH to indicate an error.


This is a hint to glob() that the caller is interested only in directories that match the pattern. If the implementation can easily determine file-type information, then nondirectory files are not returned to the caller. However, the caller must still check that returned files are directories. (The purpose of this flag is merely to optimize performance when the caller is interested only in directories.)

If errfunc is not NULL, it will be called in case of an error with the arguments epath, a pointer to the path which failed, and eerrno, the value of errno as returned from one of the calls to opendir(3), readdir(3), or stat(2). If errfunc returns nonzero, or if GLOB_ERR is set, glob() will terminate after the call to errfunc.

Upon successful return, pglob−>gl_pathc contains the number of matched pathnames and pglob−>gl_pathv contains a pointer to the list of pointers to matched pathnames. The list of pointers is terminated by a null pointer.

It is possible to call glob() several times. In that case, the GLOB_APPEND flag has to be set in flags on the second and later invocations.

As a GNU extension, pglob−>gl_flags is set to the flags specified, ored with GLOB_MAGCHAR if any metacharacters were found.


On successful completion, glob() returns zero. Other possible returns are:


for running out of memory,


for a read error, and


for no found matches.


For an explanation of the terms used in this section, see attributes(7).

Interface Attribute Value
glob() Thread safety

MT-Unsafe race:utent env

sig:ALRM timer locale

globfree() Thread safety MT-Safe

In the above table, utent in race:utent signifies that if any of the functions setutent(3), getutent(3), or endutent(3) are used in parallel in different threads of a program, then data races could occur. glob() calls those functions, so we use race:utent to remind users.


POSIX.1-2001, POSIX.1-2008, POSIX.2.


The structure elements gl_pathc and gl_offs are declared as size_t in glibc 2.1, as they should be according to POSIX.2, but are declared as int in glibc 2.0.


The glob() function may fail due to failure of underlying function calls, such as malloc(3) or opendir(3). These will store their error code in errno.


One example of use is the following code, which simulates typing

ls −l *.c ../*.c

in the shell:

glob_t globbuf;

globbuf.gl_offs = 2;
glob(*.c, GLOB_DOOFFS, NULL, &globbuf);
glob(../*.c, GLOB_DOOFFS | GLOB_APPEND, NULL, &globbuf);
globbuf.gl_pathv[0] = ls;
globbuf.gl_pathv[1] = −l;
execvp(ls, &globbuf.gl_pathv[0]);


ls(1), sh(1), stat(2), exec(3), fnmatch(3), malloc(3), opendir(3), readdir(3), wordexp(3), glob(7)


This page is part of release 5.04 of the Linux man-pages project. A description of the project, information about reporting bugs, and the latest version of this page, can be found at−pages/.

  Copyright (c) 1993 by Thomas Koenig (

Permission is granted to make and distribute verbatim copies of this
manual provided the copyright notice and this permission notice are
preserved on all copies.

Permission is granted to copy and distribute modified versions of this
manual under the conditions for verbatim copying, provided that the
entire resulting derived work is distributed under the terms of a
permission notice identical to this one.

Since the Linux kernel and libraries are constantly changing, this
manual page may be incorrect or out-of-date.  The author(s) assume no
responsibility for errors or omissions, or for damages resulting from
the use of the information contained herein.  The author(s) may not
have taken the same level of care in the production of this manual,
which is licensed free of charge, as they might when working

Formatted or processed versions of this manual, if unaccompanied by
the source, must acknowledge the copyright and authors of this work.

Modified Wed Jul 28 11:12:17 1993 by Rik Faith (
Modified Mon May 13 23:08:50 1996 by Martin Schulze (
Modified 11 May 1998 by Joseph S. Myers (
Modified 990912 by aeb
2007-10-10 mtk
    Added description of GLOB_TILDE_NOMATCH
    Expanded the description of various flags
    Various wording fixes.

Section (7) glob

Linux manual pages Section 7  


glob — globbing pathnames


Long ago, in UNIX V6, there was a program /etc/glob that would expand wildcard patterns. Soon afterward this became a shell built-in.

These days there is also a library routine glob(3) that will perform this function for a user program.

The rules are as follows (POSIX.2, 3.13).

Wildcard matching

A string is a wildcard pattern if it contains one of the characters _zsingle_quotesz_?_zsingle_quotesz_, _zsingle_quotesz_*_zsingle_quotesz_ or _zsingle_quotesz_[_zsingle_quotesz_. Globbing is the operation that expands a wildcard pattern into the list of pathnames matching the pattern. Matching is defined by:

A _zsingle_quotesz_?_zsingle_quotesz_ (not between brackets) matches any single character.

A _zsingle_quotesz_*_zsingle_quotesz_ (not between brackets) matches any string, including the empty string.

Character classes

An expression [...] where the first character after the leading _zsingle_quotesz_[_zsingle_quotesz_ is not an _zsingle_quotesz_!_zsingle_quotesz_ matches a single character, namely any of the characters enclosed by the brackets. The string enclosed by the brackets cannot be empty; therefore _zsingle_quotesz_]_zsingle_quotesz_ can be allowed between the brackets, provided that it is the first character. (Thus, [][!] matches the three characters _zsingle_quotesz_[_zsingle_quotesz_, _zsingle_quotesz_]_zsingle_quotesz_ and _zsingle_quotesz_!_zsingle_quotesz_.)


There is one special convention: two characters separated by _zsingle_quotesz_−_zsingle_quotesz_ denote a range. (Thus, [A−Fa−f0−9] is equivalent to [ABCDEFabcdef0123456789].) One may include _zsingle_quotesz_−_zsingle_quotesz_ in its literal meaning by making it the first or last character between the brackets. (Thus, []−] matches just the two characters _zsingle_quotesz_]_zsingle_quotesz_ and _zsingle_quotesz_−_zsingle_quotesz_, and [−−0] matches the three characters _zsingle_quotesz_−_zsingle_quotesz_, _zsingle_quotesz_._zsingle_quotesz_, _zsingle_quotesz_0_zsingle_quotesz_, since _zsingle_quotesz_/_zsingle_quotesz_ cannot be matched.)


An expression [!...] matches a single character, namely any character that is not matched by the expression obtained by removing the first _zsingle_quotesz_!_zsingle_quotesz_ from it. (Thus, [!]a−] matches any single character except _zsingle_quotesz_]_zsingle_quotesz_, _zsingle_quotesz_a_zsingle_quotesz_ and _zsingle_quotesz_−_zsingle_quotesz_.)

One can remove the special meaning of _zsingle_quotesz_?_zsingle_quotesz_, _zsingle_quotesz_*_zsingle_quotesz_ and _zsingle_quotesz_[_zsingle_quotesz_ by preceding them by a backslash, or, in case this is part of a shell command line, enclosing them in quotes. Between brackets these characters stand for themselves. Thus, [[?*] matches the four characters _zsingle_quotesz_[_zsingle_quotesz_, _zsingle_quotesz_?_zsingle_quotesz_, _zsingle_quotesz_*_zsingle_quotesz_ and _zsingle_quotesz_\_zsingle_quotesz_.


Globbing is applied on each of the components of a pathname separately. A _zsingle_quotesz_/_zsingle_quotesz_ in a pathname cannot be matched by a _zsingle_quotesz_?_zsingle_quotesz_ or _zsingle_quotesz_*_zsingle_quotesz_ wildcard, or by a range like [.−0]. A range containing an explicit _zsingle_quotesz_/_zsingle_quotesz_ character is syntactically incorrect. (POSIX requires that syntactically incorrect patterns are left unchanged.)

If a filename starts with a _zsingle_quotesz_._zsingle_quotesz_, this character must be matched explicitly. (Thus, rm * will not remove .profile, and tar c * will not archive all your files; tar c . is better.)

Empty lists

The nice and simple rule given above: expand a wildcard pattern into the list of matching pathnames was the original UNIX definition. It allowed one to have patterns that expand into an empty list, as in

    xv −wait 0 *.gif *.jpg

where perhaps no *.gif files are present (and this is not an error). However, POSIX requires that a wildcard pattern is left unchanged when it is syntactically incorrect, or the list of matching pathnames is empty. With bash one can force the classical behavior using this command:

shopt −s nullglob

(Similar problems occur elsewhere. For example, where old scripts have

    rm `find . −name *~`

new scripts require

    rm −f nosuchfile `find . −name *~`

to avoid error messages from rm called with an empty argument list.)


Regular expressions

Note that wildcard patterns are not regular expressions, although they are a bit similar. First of all, they match filenames, rather than text, and secondly, the conventions are not the same: for example, in a regular expression _zsingle_quotesz_*_zsingle_quotesz_ means zero or more copies of the preceding thing.

Now that regular expressions have bracket expressions where the negation is indicated by a _zsingle_quotesz_^_zsingle_quotesz_, POSIX has declared the effect of a wildcard pattern [^...] to be undefined.

Character classes and internationalization

Of course ranges were originally meant to be ASCII ranges, so that [ −%] stands for [ !#$%] and [a−z] stands for any lowercase letter. Some UNIX implementations generalized this so that a range X−Y stands for the set of characters with code between the codes for X and for Y. However, this requires the user to know the character coding in use on the local system, and moreover, is not convenient if the collating sequence for the local alphabet differs from the ordering of the character codes. Therefore, POSIX extended the bracket notation greatly, both for wildcard patterns and for regular expressions. In the above we saw three types of items that can occur in a bracket expression: namely (i) the negation, (ii) explicit single characters, and (iii) ranges. POSIX specifies ranges in an internationally more useful way and adds three more types:

(iii) Ranges X−Y comprise all characters that fall between X and Y (inclusive) in the current collating sequence as defined by the LC_COLLATE category in the current locale.

(iv) Named character classes, like

[:alnum:]  [:alpha:]  [:blank:]  [:cntrl:]
[:digit:]  [:graph:]  [:lower:]  [:print:]
[:punct:]  [:space:]  [:upper:]  [:xdigit:]

so that one can say [[:lower:]] instead of [a−z], and have things work in Denmark, too, where there are three letters past _zsingle_quotesz_z_zsingle_quotesz_ in the alphabet. These character classes are defined by the LC_CTYPE category in the current locale.

(v) Collating symbols, like [.ch.] or [.a-acute.], where the string between [. and .] is a collating element defined for the current locale. Note that this may be a multicharacter element.

(vi) Equivalence class expressions, like [=a=], where the string between [= and =] is any collating element from its equivalence class, as defined for the current locale. For example, [[=a=]] might be equivalent to [aáàäâ], that is, to [a[.a-acute.][.a-grave.][.a-umlaut.][.a-circumflex.]].


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


This page is part of release 4.16 of the Linux man-pages project. A description of the project, information about reporting bugs, and the latest version of this page, can be found at−pages/.

  Copyright (c) 1998 Andries Brouwer

This is free documentation; you can redistribute it and/or
modify it under the terms of the GNU General Public License as
published by the Free Software Foundation; either version 2 of
the License, or (at your option) any later version.

The GNU General Public License_zsingle_quotesz_s references to object code
and executables are to be interpreted as the output of any
document formatting or typesetting system, including
intermediate and printed output.

This manual is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
GNU General Public License for more details.

You should have received a copy of the GNU General Public
License along with this manual; if not, see

2003-08-24 fix for / by John Kristoff + joey