Table of Contents

File, Network and Revision Control Utilities


File Utilities


Batch processing

find

xargs

        # Find all hidden files in the cs370 dir (files that begin with 
        # period ".*") and then use xargs to run "ls -l" on all
        # the find results.
        $ find ./cs370 -type f -name '.*' | xargs ls -l

        # Same output as 
        # $ find ./cs370 -type f -name '.*' -exec ls -l {} ';'



        # List all users logged in and run finger on each userid
        $ w -h | awk '{ print $1 }' | sort | uniq | xargs finger
        #
        # w -h               - list logged in users and associated info, excluding header (-h)
        # awk '{ print $1 }' - print only userid from w output
        # sort | uniq        - reduce duplicates
        # xargs finger       - run finger on list of users from uniq

Archiving and compression

tar

        c - create archive
        t - view existing archive
        v - operate verbosely
        x - extract archive
        f - create the specified tar file ( or use "-" to send tar'ed
            files to stdout ) 
        # Create tar archive (cs370.tar) of your ~/cs370 directory in /tmp:
        # Dashes ("-") are optional for tar options.
        $ tar cvf /tmp/cs370.tar ~/cs370  # "tar -cvf ..." does the same thing

        # Change to /tmp
        $ cd /tmp

        # "tar tvf cs370.tar" shows the contents of cs370.tar.
        # "tar xvf cs370.tar" extracts the contents of cs370.tar to
        # the CURRENT directory. (BE CAREFUL.)

gzip/bzip2/xz

        $ cd /tmp

        # Copy nano config files to /tmp
        $ cp /usr/share/nano/*.nanorc  /tmp

        $ gzip *.nanorc    # will result in all .nanorc files in current dir
                           # being compressed and given the .nanorc.gz extension


        $ gunzip *.gz      # will uncompress the .nanorc.gz files and
                           # leave files w/ .nanorc extensions


        $ bzip2 *.nanorc
        $ bunzip2 *.bz2    # same as above, but with bzip2

        
        $ xz *.nanorc
        $ unxz *.xz        # same as above, but with xz
        # Copy large wordlist to /tmp
        $ cp /usr/share/dict/words  /tmp
        
        # Compress wordlist to a separate compressed files:
        $ gzip -c words > words.gz
        $ bzip2 -c words > words.bz2
        $ xz -c words > words.xz
        # Compare size of compressed file formats.
        # Also try zip:
        $ zip words.zip words
        # View nano documentation
        $ cd /usr/share/doc/nano
        $ ls
        
        # View a compressed file (NEWS.gz)
        $ gunzip -c NEWS.gz | less
        
        or, more simply,
        
        $ zless NEWS.gz
        
        # Cat a compressed file (NEWS.gz)
        $ zcat NEWS.gz:
        # Tar your ~/cs370 dir to tar's stdout (-) and xz it,
        # redirecting the result to cs370.tar.xz:
        $ tar cvf -  ~/cs370 | xz -c > /tmp/cs370.tar.xz

        # Change to /tmp
        $ cd /tmp

        # Do the reverse to view contents of cs370.tar.xz:
        $ unxz -c cs370.tar.xz | tar tvf -
        $ tar cvJf /tmp/cs370.tar.xz ~/cs370 
        # GNU tar's "J" option forces use of xz to compress the tar archive
        #        if "z"        uses          gzip
        #        if "j"        uses          bzip2

        # to view tar.xz    
        $ tar tvJf cs370.tar.xz
        
        # to extract tar.xz 
        $ cd /tmp; tar xvJf cs370.tar.xz


Network Utilities


telnet/ftp

        # Check if vnc service running on plato (port 5900)
        # Any response means the service is running
        telnet plato 5900

ssh/scp

        #
        # Remote login:
        #
        
        # Login remotely to rockhopper.
        # Authenticate using either a password or encrypted
        # key exchange:
        $ ssh <your_userid>@rockhopper
        
        # See verbose output of a ssh login process
        $ ssh -v <your_userid>@rockhopper


        #
        # Remote command execution:
        #

        # Run the 'uptime' command on csselin01:
        $ ssh csselin01 'uptime'

        # See logins on rockhopper
        $ ssh rockhopper 'finger'

        # See jchung logins on rockhopper
        $ ssh rockhopper 'finger | grep -i chung'
        
        # Same thing, but stdout from rockhopper piped to local grep
        $ ssh rockhopper 'finger' | grep -i chung


        # Tar your ~/cs370 dir locally, pipe to gzip on rockhopper to
        # create rockhopper:/tmp/$USER-cs370.tar.gz:
        $ tar cvf -  ~/cs370 | ssh rockhopper "gzip -c > /tmp/$USER-cs370.tar.gz"
        # Create and transfer /tmp/cs370.tar.xz to your home dir on the plato server:
        $ tar  cJf  /tmp/cs370.tar.xz  ~/cs370
        $ scp  /tmp/cs370.tar.xz  plato:~


        # Transfer ~/cs370.tar.xz from plato to local /tmp:
        $ scp  plato:~/cs370.tar.xz  /tmp

rsync

        # Transfer entire ~/cs370 dir to a remote machine:/tmp
        # rsync command options (similar to cp options)
        #   -a archive (recursively copy dirs and preserve all file/dir attributes)
        #   -u update  (only transfer files that are newer than destination)
        #   -v verbose
        # 
        # In this rsync command, the source is ~/cs370 and the destination is localhost:/tmp.
        $ rsync -auv  ~/cs370  localhost:/tmp
        
        # Run it again.
        # Since -u (update) is being used, nothing gets transferred because the source
        # and destination are both up-to-date.
        $ rsync -auv  ~/cs370  localhost:/tmp
        
        # Update timestamp of ~/cs370/examples dir with 'touch',
        # and run rsync again.
        $ touch ~/cs370/examples
        $ rsync -auv  ~/cs370  localhost:/tmp

wget

        $ wget "http://en.wikipedia.org/wiki/regular_expressions"
              
        # Saves article to file //regular_expressions//.

The "-" (STDOUT) convention



Diff/Patch


diff - find differences between two files

mkdir -p ~/cs370/examples/revcontrol
cd ~/cs370/examples/revcontrol
wget -q http://bit.ly/2zZgGiV -O diffpatch.tar.xz # download diffpatch.tar.xz
tar xvJf diffpatch.tar.xz                         # extract the diffpatch directory
ls
cd diffpatch
        SYNOPSIS
               diff [options] from-file to-file
# diff input file #1
# saved as seuss1
$ cat seuss1
If a packet hits a pocket on a socket on a port,
and the bus is interrupted at a very last resort,  
and the access of the memory makes your floppy disk abort,
then the socket packet pocket has an error to report.




# diff input file #2
# saved as seuss2
$ cat seuss2
If a pocket hits a rocket on a socket on a port,
and the bus is interrupted at a very last resort,  
and the access of the memory makes your floppy abort,
then the socket packet pocket has an error to report.


       


# Use diff to show differences between seuss1 and seuss2:
$ diff seuss1 seuss2
1c1
< If a packet hits a pocket on a socket on a port,
---
> If a pocket hits a rocket on a socket on a port,
3c3
< and the access of the memory makes your floppy disk abort,
---
> and the access of the memory makes your floppy abort,



# diff input file #3
# saved as seuss3
$ cat seuss3
If a pocket hits a rocket on a socket on a port,
and the bus is interrupted at a very last resort,  
and the access of the memory makes your floppy abort,
then the socket packet pocket has an error to report.
       
If your cursor finds a menu item followed by a dash,
and the double-clicking icon puts your window in the trash,
and your data is corrupted cause the index doesn't hash,
then your situation's hopeless and your system's gonna crash!




       
# Use diff to show differences between seuss2 and seuss3:
$ diff seuss2 seuss3
4a5,9
> 
> If your cursor finds a menu item followed by a dash,
> and the double-clicking icon puts your window in the trash,
> and your data is corrupted cause the index doesn't hash,
> then your situation's hopeless and your system's gonna crash!





# diff input file #4
# saved as seuss4
$ cat seuss4
If a packet hits a pocket on a socket on a port,
and the access of the memory makes your floppy disk abort,
and the bus is interrupted at a very last resort,  
then the socket packet pocket has an error to report.






# Use diff to show differences between seuss3 and seuss4:
$ diff seuss3 seuss4
1c1,2
< If a pocket hits a rocket on a socket on a port,
---
> If a packet hits a pocket on a socket on a port,
> and the access of the memory makes your floppy disk abort,
3d3
< and the access of the memory makes your floppy abort,
5,9d4
< 
< If your cursor finds a menu item followed by a dash,
< and the double-clicking icon puts your window in the trash,
< and your data is corrupted cause the index doesn't hash,
< then your situation's hopeless and your system's gonna crash!

patch - apply a diff file to an original

        SYNOPSIS
               patch [options] [originalfile [patchfile]]
# Using diff and patch to merge changes
$ diff seuss3 seuss4 > diff34        # Generate diff file


$ cp seuss3 seuss3.orig              # Backup original seuss3



$ patch --verbose seuss3 diff34      # Apply diff34 to seuss3
Hmm...  Looks like a normal diff to me...
Patching file seuss3 using Plan A...
Hunk #1 succeeded at 1.
Hunk #2 succeeded at 4.
Hunk #3 succeeded at 5.
done






$ cat seuss3         # seuss3 is now the same as seuss4
If a packet hits a pocket on a socket on a port,
and the access of the memory makes your floppy disk abort,
and the bus is interrupted at a very last resort,
then the socket packet pocket has an error to report.

Revision Control Utilities



Lab Activities


1. Find all files that contain a string or regular expression

In ~/.bashrc, define a shell function called searchfiles which uses find to list all files that contain the string (or regular expression) that you pass in as the first function parameter, $1. Note that we don't want to search file names but file contents for a string, and then list the files that match.

Answer:

# function searchfiles which uses find to list all files that 
# contain the string (or regular expression)
# that you pass in as the first function parameter, $1.
function searchfiles
{
   find . -type f |                 # list all files recursively starting in . (current dir)
   xargs grep -li "$1" 2> /dev/null # using xargs, make grep list files (-l) in which a match is found
	
   # Can also use command substitution, if not too many find results:
   # grep -li "$1" $(find . -type f) 2> /dev/null
   # 
   # If using GNU grep (most UNIX systems), can use just grep recursively (-r):
   # grep -rli "$1" 2> /dev/null
}

2. Change to directory based on find result

In ~/.bashrc, define a shell function called findcd that changes to a directory based on a find result. If what you're searching for matches a filename, then change to the directory where that file resides. If what you're searching for matches a directory name, then change to that directory.

Example usage:

# Change to a dir named randomwall or to a dir that contains a file called randomwall
findcd randomwall 

# Change to a dir named examples  
findcd examples

# Change to a dir that contains a file called roster
findcd roster

Answer:

# function findcd - changes to a directory based on the first hit from a find
function findcd
{
   # "head -n1" chooses first find result;
   # use head instead of tail here, else may have 
   # to wait for find to print many search results;
   # if find finds nothing, $findresult is ""
   findresult=$(find . -iname "*$1*" | head -n1)

   # If $findresult is a file, can't cd to it,
   # so have to trim $findresult to a directory
   if [ -f "$findresult" ]; then
      filename=$(basename "$findresult") # see man basename
      # delete $filename from end (\$) of $findresult
      findresult=$(echo "$findresult" | sed "s/$filename\$//")
   fi

   cd "$findresult" # If $findresult is "", nothing happens.
}

3. Find maximum directory depth

Within your home directory, find the maximum depth of a directory. Your results should include the directory's name.

Note: You'll need to use the find command's -printf option. See man find.

Answer:

# Starting in current directory (.), find directories (-type d), 
# print the depth of each directory found (-printf "%d "),
# print the path of each directory found (-print),
# do a descending, numeric sort (-rn), show only the first result (head -n1)
find . -type d -printf '%d ' -print | sort -rn | head -n1

or

find . -type d -printf '%d ' -exec ls -ld {} ';' | sort -rn | head -n1

4. Create and use a git repository with gitlab

(NOTE: Counts toward your participation grade.)

# Using cp
cp  -av  ~/cs370  ~/cs370-$(date +%m%d%y)
# or using rsync
rsync -av  ~/cs370  ~/cs370-$(date +%m%d%y)