File::BSDGlob - Perl extension for BSD glob routine


  use File::BSDGlob ':glob';
  @list = glob('*.[ch]');
  $homedir = glob('~gnat', GLOB_TILDE | GLOB_ERR);
  if (GLOB_ERROR) {
    # an error occurred reading $homedir

  ## override the core glob (even with -T)
  use File::BSDGlob 'globally';
  my @sources = <*.{c,h,y}>


File::BSDGlob implements the FreeBSD glob(3) routine, which is a superset of the POSIX glob() (described in IEEE Std 1003.2 "POSIX.2"). The glob() routine takes a mandatory pattern argument, and an optional flags argument, and returns a list of filenames matching the pattern, with interpretation of the pattern modified by the flags variable. The POSIX defined flags are:


Force glob() to return an error when it encounters a directory it cannot open or read. Ordinarily glob() continues to find matches.


Each pathname that is a directory that matches the pattern has a slash appended.


If the pattern does not match any pathname, then glob() returns a list consisting of only the pattern. If GLOB_QUOTE is set, its effect is present in the pattern returned.


By default, the pathnames are sorted in ascending ASCII order; this flag prevents that sorting (speeding up glob()).

The FreeBSD extensions to the POSIX standard are the following flags:


Pre-process the string to expand {pat,pat,...} strings like csh(1). The pattern '{}' is left unexpanded for historical reasons (and csh(1) does the same thing to ease typing of find(1) patterns).


Same as GLOB_NOCHECK but it only returns the pattern if it does not contain any of the special characters "*", "?" or "[". NOMAGIC is provided to simplify implementing the historic csh(1) globbing behaviour and should probably not be used anywhere else.


Use the backslash ('\') character for quoting: every occurrence of a backslash followed by a character in the pattern is replaced by that character, avoiding any special interpretation of the character.


Expand patterns that start with '~' to user name home directories.


For convenience, GLOB_CSH is a synonym for GLOB_BRACE | GLOB_NOMAGIC | GLOB_QUOTE | GLOB_TILDE.

The POSIX provided GLOB_APPEND, GLOB_DOOFFS, and the FreeBSD extensions GLOB_ALTDIRFUNC, and GLOB_MAGCHAR flags have not been implemented in the Perl version because they involve more complex interaction with the underlying C structures.


glob() returns a list of matching paths, possibly zero length. If an error occurred, &File::BSDGlob::GLOB_ERROR will be non-zero and $! will be set. &File::BSDGlob::GLOB_ERROR is guaranteed to be zero if no error occurred, or one of the following values otherwise:


An attempt to allocate memory failed.


The glob was stopped because an error was encountered.

In the case where glob() has found some matching paths, but is interrupted by an error, glob() will return a list of filenames and set &File::BSDGlob::ERROR.

Note that glob() deviates from POSIX and FreeBSD glob(3) behaviour by not considering ENOENT and ENOTDIR as errors - glob() will continue processing despite those errors, unless the GLOB_ERR flag is set.

Be aware that all filenames returned from File::BSDGlob are tainted.


  • If you want to use multiple patterns, e.g. glob "a* b*", you should probably throw them in a set as in glob "{a*,b*}. This is because the argument to glob isn't subjected to parsing by the C shell. Remember that you can use a backslash to escape things.

  • Win32 users should use the real slash. If you really want to use backslashes, consider using Sarathy's File::DosGlob, which comes with the standard Perl distribution.


The Perl interface was written by Nathan Torkington (, and is released under the artistic license. Further modifications were made by Greg Bacon <>. The C glob code has the following copyright:

  Copyright (c) 1989, 1993 The Regents of the University of California.
  All rights reserved.  This code is derived from software contributed
  to Berkeley by Guido van Rossum.

For redistribution of the C glob code, read the copyright notice in the file bsd_glob.c, which is part of the File::BSDGlob source distribution.

2 POD Errors

The following errors were encountered while parsing the POD:

Around line 200:

Unterminated C<...> sequence

Around line 280:

You forgot a '=back' before '=head1'