# $Id: README,v 3.3 1995/08/27 16:26:37 kb207252 Exp $

Here are two programs I have found (recently) to be quite useful.
One of them expands the C-Shell curly braces notation for file name
building, the other compresses a list of file names into one of
these expressions.  I called these programs `curly' and `uncurly'.

There is am example usage of {un,}curly in the comments in the code,
copied below.  I have used them to compress a dictionary (one initial
letter at a time BTW), and to compress lists of files on tape.

One might pipe the output of a find to uncurly and compress to store
a list of filenames as:

	$ find . -type f -print | uncurly | gzip > /tmp/files.u.gz

then (later on) we need that list of files again...

	$ gzcat /tmp/files.u.gz | curly | xargs do_something

this yields substantial compression over just gzip (or compress) alone,
which it shouldn't.  (We know something quite special about the output
of find on a `normal' find output that gives us an advantage here.)

I always name the output from uncurly as files.u.  Here is some sample
compression data (for 1962 handy filenames in my home directory):

-rw-r--r--   1 ksb dspd       48335 Apr 19 09:59 /tmp/files	  #100.0%
-rw-r--r--   1 ksb dspd       15462 Apr 19 10:01 /tmp/files.Z	  # 31.9%
-rw-r--r--   1 ksb dspd        9092 Apr 19 10:00 /tmp/files.gz	  # 18.8%
-rw-r--r--   1 ksb dspd       16648 Apr 19 10:00 /tmp/files.u	  # 34.4%
-rw-r--r--   1 ksb dspd        9240 Apr 19 10:01 /tmp/files.u.Z	  # 19.1%
-rw-r--r--   1 ksb dspd        7074 Apr 19 10:00 /tmp/files.u.gz  # 14.6%

I would like to be mailed any bug reports so I can fix my own copy.

Known bugs:  doesn't handle files with `{' or `}' in them well.

--
Enjoy!
kayessbee (Kevin S Braunsdorf, ksbrauns@fedex.com)