University of California at Berkeley
           Department of Electrical Engineering & Computer Sciences
                     Instructional & Electronics Support Group

/share/b/pub/compress.help
							      December 1991


		Avoiding Deleted Files when Using "Compress"
		~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


The UNIX "compress" and "uncompress" commands can have an unwelcome side 
effect that erases your file without saving it.   Here is a technique to
avoid that.

If there is not enough disk space or disk quota to allow for the output
file, the commands will write a size 0 output file and then delete the 
original file as usual.   There must be enough disk space for both the 
input and output files in order for the commands to work properly.

One solution is to make a link to the original file and then compress or 
uncompress the link.  That results in the link being deleted at the end
but the original file being left alone.  

For example:

  % ln -s ORIGINAL TEMP
  % compress TEMP
  % mv TEMP.Z ORIGINAL.Z

In the example, the result is that a file called ORIGINAL was compressed 
into a file called ORIGINAL.Z, with both files remaining.  First, a symbolic 
link called TEMP (or anything you want) was created, then it was compressed.  
The compress command created an output file called TEMP.Z, then deleted the 
TEMP link.  This left the leave the ORIGINAL file intact, even if the TEMP.Z 
file was empty due to disk space limitations.  Then, to follow the typical 
naming convention, the compressed file was renamed to ORIGINAL.Z .

Note that you can use extra disk space in the /tmp directory, temporarily,
when you are compressing or uncompressing files.

For more information about the above commands, see these on-line manual 
pages: "man compress",  "man ln", "man mv".


                                        - Instructional Support Group
					  378/386/384 Cory, 333 Soda
					  inst@eecs.berkeley.edu