NAME

junk - find junk files, and apply an operation to each


SYNOPSIS

junk [-c command] [-f filterlist] [-n namelist] [pathname ...]


DESCRIPTION

Junk descends the directory hierarchy for each pathname argument
seeking files that satisfy an appropriate definition of junk, and
applying command to each file found.

If command is not specified, `echo' is used.  Application of command
to a file name means that the filename is tacked on to the end of
command to obtain a shell command which is executed.

The filterlist argument can be used to narrow the files considered
according to any additional expression in the style of FIND(1).  If it
is absent it defaults to "".

The namelist argument can be used to specify a list of filenames which
will be considered to represent junk.  Normal shell file matching
characters (e.g. *) may be used if quoted.  If this argument is absent
the names will be determined from the setting of environment variables
or will default to a suitably large list of patterns.  This is
described below.

If no -n option is present, then the namelist is taken from the
environment variable JUNK_NAMES if it is set.  If it is not, then the
names are accumulated from the environment variables specified by the
environment variable JUNK_TYPES, which should be some subset of "PRINT
EMACS LATEX BIBTEX TROFF EXEC MISC".  For each TYPE, the names are taken
from the environment variable TYPE_JUNK if it is set; otherwise a
rather inclusive set of names is used.  In the case of LATEX_JUNK,
there is one extra level.  If LATEX_JUNK is not set, the latex junk
names are accumulated from the environment variables LATEX_OUT_JUNK,
LATEX_IO_JUNK, and LATEX_INDEX_JUNK according to the setting of the
environment variable LATEX_JUNK_TYPES.

The default values assumed for unset variables are given below.  You
can set any subset of these variables in your environment.

  JUNK_TYPES = "PRINT EMACS LATEX BIBTEX TROFF EXEC MISC"

  PRINT_JUNK = "*.dvi .*.dvi *.imp .*.imp"

  EMACS_JUNK = "*~ .*~ #* *.bak .*.bak"

  TEX_JUNK = "*.log .*.log"

  LATEX_JUNK_TYPES = "OUT IO INDEX"

  LATEX_OUT_JUNK = "*.log .*.log"

  LATEX_IO_JUNK = "*.aux .*.aux *.toc .*.toc *.lof .*.lof *.lot .*.lot .*.idx *.glo .*.glo"

  LATEX_INDEX_JUNK = "*.bbl .*.bbl"

  BIBTEX_JUNK = "*.blg .*.blg"

  TROFF_JUNK = "*.t .*.t"

  EXEC_JUNK = "core a.out *.o .*.o"

  MISC_JUNK = ".tmp"

You may wish to make PRINT_JUNK include PostScript files;

  PRINT_JUNK = "*.dvi .*.dvi *.ps .*.ps *.PS .*.PS *.imp .*.imp"


EXAMPLES

Here are some examples of how you might use junk.  (Some of the
examples use another function, total, which I have created.  See
total.doc if you want information on other ways of using total.)  You
don't have to look at all of these examples, if you start losing
interest as they get more complicated.

1.  See the names of all junk files in the current directory (and
subdirectories recursively):

	junk

2.  See the names of all junk files in the directories foo and bar:

	junk foo bar

3.  See long ls and sizes in kilobytes (rounded up) of junk files (in cd):

	junk -c 'ls -ls'

4.  Remove junk files not accessed for more than 1 day ( BE CAREFUL !!!! ):

	junk -c rm -f '-atime +1'

5.  See total number of kilobytes (rounded up per file) in junk files:
    (Note that hard links lead to double counting):

	junk -c 'ls -s' | total '$1'

6.  See long ls, and get the number of files and total bytes:

	junk -c 'ls -l' | total -p 1 files '$4' bytes

7.  Get the total number of bytes without multiple counting of hard links:

	junk -c 'ls -l' | total '$4/$2' bytes

8.  Get every user's number of bytes of junk and the sytem total and
sort it (if you are root):

	cd /u
        echo >/tmp/junktotals
	foreach user (*)
  	  junk -c 'ls -l' $user | total '$4/$2' "bytes $user" | grep "Total"\
            >>/tmp/junktotals
	end
	cat /tmp/junktotals | total -p '$1' | grep "Total" | sort -nr
	rm /tmp/junktotals

9.  Remove files starting with a digit or ending with a digit:

	junk -c rm -n '[0-9]* *[0-9] .*[0-9]'


SEE ALSO

find(1)