GLOB(7) Linux Programmer's Manual GLOB(7)
glob - Globbing pathnames
Long ago, in Unix V6, there was a program /etc/glob that would expand
wildcard patterns. Soon afterwards 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).
A string is a wildcard pattern if it contains one of the characters
'?', '*' or '['. Globbing is the operation that expands a wildcard
pattern into the list of pathnames matching the pattern. Matching is
A '?' (not between brackets) matches any single character.
A '*' (not between brackets) matches any string, including the empty
An expression "[...]" where the first character after the leading '['
is not an '!' matches a single character, namely any of the characters
enclosed by the brackets. The string enclosed by the brackets cannot
be empty; therefore ']' can be allowed between the brackets, provided
that it is the first character. (Thus, "[!]" matches the three char-
acters '[', ']' and '!'.)
There is one special convention: two characters separated by '-' denote
a range. (Thus, "[A-Fa-f0-9]" is equivalent to "[ABCDE-
Fabcdef0123456789]".) One may include '-' in its literal meaning by
making it the first or last character between the brackets. (Thus,
"-]" matches just the two characters ']' and '-', and "[--0]" matches
the three characters '-', '.', '0', since '/' cannot be matched.)
An expression "[!...]" matches a single character, namely any character
that is not matched by the expression obtained by removing the first
'!' from it. (Thus, "[!]a-]" matches any single character except ']',
'a' and '-'.)
One can remove the special meaning of '?', '*' and '[' 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 '[', '?', '*'
Globbing is applied on each of the components of a pathname separately.
A '/' in a pathname cannot be matched by a '?' or '*' wildcard, or by a
range like "[.-0]". A range cannot contain an explicit '/' character;
this would lead to a syntax error.
If a filename starts with a '.', this character must be matched explic-
itly. (Thus, rm * will not remove .profile, and tar c * will not ar-
chive all your files; tar c . is better.)
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 by setting
(Similar problems occur elsewhere. E.g., 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.)
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 '*' means zero or more copies of the preceding
Now that regular expressions have bracket expressions where the nega-
tion is indicated by a '^', 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 cod-
ing 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_COL-
LATE 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 'z' 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 multi-character 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 "[a]" (warning: Latin-1 here), that is, to "[a[.a-
sh(1), fnmatch(3), glob(3), locale(7), regex(7)
This page is part of release 3.05 of the Linux man-pages project. A
description of the project, and information about reporting bugs, can
be found at http://www.kernel.org/doc/man-pages/.
Linux 2003-08-24 GLOB(7)