Seth Woolley's Man Viewer

File::DosGlob(3) - File::DosGlob - DOS like globbing and then some - man 3 File::DosGlob

([section] manual, -k keyword, -K [section] search, -f whatis)
man plain no title

File::DosGlob(3)       Perl Programmers Reference Guide       File::DosGlob(3)



NAME
       File::DosGlob - DOS like globbing and then some

SYNOPSIS
           require 5.004;

           # override CORE::glob in(1,8) current package
           use File::DosGlob 'glob(1,3,7,n)';

           # override CORE::glob in(1,8) ALL packages (use with extreme caution!)
           use File::DosGlob 'GLOBAL_glob';

           @perlfiles = glob(1,3,7,n)  "..\\pe?l/*.p?";
           print <..\\pe?l/*.p?>;

           # from the command line (overrides only in(1,8) main::)
           > perl -MFile::DosGlob=glob(1,3,7,n) -e "print <../pe*/*p?>"

DESCRIPTION
       A module that implements DOS-like globbing with a few enhancements.  It
       is largely compatible with perlglob.exe (the M$ setargv.obj version(1,3,5)) in(1,8)
       all but one respect--it understands wildcards in(1,8) directory components.

       For example, "<..\\l*b\\file(1,n)/*glob.p?"> will work as expected (in(1,8) that
       it will find something like '..\lib\File/DosGlob.pm' alright).  Note
       that all path components are case-insensitive, and that backslashes and
       forward slashes are both accepted, and preserved.  You may have to dou-
       ble the backslashes if(3,n) you are putting them in(1,8) literally, due to dou-
       ble-quotish parsing of the pattern by perl.

       Spaces in(1,8) the argument delimit distinct patterns, so "glob(1,3,7,n)('*.exe
       *.dll')" globs all filenames that end in(1,8) ".exe" or ".dll".  If you want
       to put in(1,8) literal spaces in(1,8) the glob(1,3,7,n) pattern, you can escape them with
       either double quotes, or backslashes.  e.g. "glob(1,3,7,n)('c:/"Program
       Files"/*/*.dll')", or "glob(1,3,7,n)('c:/Program\ Files/*/*.dll')".  The argu-
       ment is tokenized using "Text::ParseWords::parse_line()", so see
       Text::ParseWords for details of the quoting rules used.

       Extending it to csh patterns is left as an exercise to the reader.

NOTES
          Mac OS (Classic) users(1,5) should note a few differences. The specifi-
           cation of pathnames in(1,8) glob(1,3,7,n) patterns adheres to the usual Mac OS
           conventions: The path separator is a colon ':', not a slash '/' or
           backslash '\'. A full path always begins with a volume name. A rel-
           ative pathname on Mac OS must always begin with a ':', except when
           specifying a file(1,n) or directory name in(1,8) the current working direc-
           tory, where the leading colon is optional. If specifying a volume
           name only, a trailing ':' is required. Due to these rules, a glob(1,3,7,n)
           like <*:> will find all mounted volumes, while a glob(1,3,7,n) like <*> or
           <:*> will find all files and directories in(1,8) the current directory.

           Note that updirs in(1,8) the glob(1,3,7,n) pattern are resolved before the match-
           ing begins, i.e. a pattern like "*HD:t?p::a*" will be matched as
           "*HD:a*". Note also, that a single trailing ':' in(1,8) the pattern is
           ignored (unless it's a volume name pattern like "*HD:"), i.e. a
           glob(1,3,7,n) like <:*:> will find both directories and files (and not, as
           one might expect, only directories).

           The metachars '*', '?' and the escape char '\' are valid characters
           in(1,8) volume, directory and file(1,n) names on Mac OS. Hence, if(3,n) you want
           to match a '*', '?' or '\' literally, you have to escape these
           characters. Due to perl's quoting rules, things may get a bit com-
           plicated, when you want to match a string(3,n) like '\*' literally, or
           when you want to match '\' literally, but treat the immediately
           following character '*' as metachar. So, here's a rule of thumb
           (applies to both single- and double-quoted strings): escape each
           '*' or '?' or '\' with a backslash, if(3,n) you want to treat them lit-
           erally, and then double each backslash and your are done. E.g.

           - Match '\*' literally

              escape both '\' and '*'  : '\\\*'
              double the backslashes   : '\\\\\\*'

           (Internally, the glob(1,3,7,n) routine sees a '\\\*', which means that both
           '\' and '*' are escaped.)

           - Match '\' literally, treat '*' as metachar

              escape '\' but not '*'   : '\\*'
              double the backslashes   : '\\\\*'

           (Internally, the glob(1,3,7,n) routine sees a '\\*', which means that '\' is
           escaped and '*' is not.)

           Note that you also have to quote literal spaces in(1,8) the glob(1,3,7,n) pat-
           tern, as described above.

EXPORTS (by request only)
       glob(1,3,7,n)()

BUGS
       Should probably be built into the core, and needs to stop pandering to
       DOS habits.  Needs a dose of optimizium too.

AUTHOR
       Gurusamy Sarathy <gsar@activestate.com>

HISTORY
          Support for globally overriding glob(1,3,7,n)() (GSAR 3-JUN-98)

          Scalar context, independent iterator context fixes (GSAR 15-SEP-97)

          A few dir-vs-file optimizations result in(1,8) glob(1,3,7,n) importation being 10
           times faster than using perlglob.exe, and using perlglob.bat is
           only twice as slow as perlglob.exe (GSAR 28-MAY-97)

          Several cleanups prompted by lack of compatible perlglob.exe under
           Borland (GSAR 27-MAY-97)

          Initial version(1,3,5) (GSAR 20-FEB-97)

SEE ALSO
       perl

       perlglob.bat

       Text::ParseWords



perl v5.8.5                       2001-09-21                  File::DosGlob(3)

References for this manual (incoming links)