Bit 0 is "execute", bit 1 is "write", and bit 2 is "read". "1" means "permission granted", "0" means "refused". The permission of the file in picture is 711 (octal number).

       r = read
       w = write
       x = execute

       u = user
       g = group
       o = others

       +   permission granted
       -   permission taken away

Exercise : If we give "psin.c" read permissions to "group" and "others", what is the command ?

Ans : chmod go+r psin.c

Exercise : If we take away all permissions from "group", "other" for the file "psin.c", what is the command ?

Ans : chmod go-rwx psin.c

Exercise : If we do not want "psin.c" to be modified by us accidentally, what is the command?

Ans : chmod u-w psin.c

Exercise : What does the command "chmod 700 psin.c" mean?

Ans : User has all permissions, but "group, others" have none.

Exercise : What does the command "chmod 644 psin.c" mean?

Ans : User has "read, write" permissions, but "group, others" have "read" only.

Exercise : What does the command "chmod 711 psin.c" mean?

Ans : User has all permissions, but "group, others" have "execute" only.

Exercise : What does the command "chmod 711 *c" mean?

Ans : For all files in the current directory whose names end with "c", user has all permissions, but "group, others" have "execute" only.

(Note : "permissions" may be used for directories too. "Read, write" has the usual menaing, but "execute" means "read" only.)

ar [-options] libraryname objectfile1 objectfile2 ....

Where "options" may be

-r	Replace object files with those provided. It is also used to insert object files. Note that library file name must be in the form "lib .... .a", i.e. it begins with "lib" and have extension ".a". If this "libraryname" does not exist, it will be created. If an object file already exists in the library, say "objectfile1", it will be replaced with those supplied.
-v	Verbose. You should supply this, so that what "ar" does will be printed on the screen.
-d	Delete. "objectfile1 objectfile2 objectfile3 .... " will be deleted.
-x	Extract. "objectfile1 objectfile2 objectfile3 ...." will be extracted from "libraryname". Notice that they are still in "libraryname". Only copies are extracted from the "libraryname".
-s	Index. It commands "ar" to create a "index file". "Index file" are important, when subroutines call one another in a complicated ways.
-t	Table of Content. List all the object files in the "libraryname".

Examples

("ar" will create the library "/home/tom/lib/libtest.a" first, if it doesn't exist) Replace/insert the binary files "quad.o, root.o, equation.o"
Create the "index table" and shows everything "ar" does, i.e. verbose mode.

List all the object files in the library "/home/tom/lib/libtest.a" (table of content).

Suppose we have modified the subroutine "quad.c", and have compiled it with "gcc -Wall -c quad.c", and we are to insert this newer version into the library.
"Index table" is to be updated too.

Extract copies of "root.o equation.o" from the library. They are still in the library, only copies are made.

The two object files "root.o equation.o" will be deleted from the library.

How to use the libary

Now that we have created a library. Suppose the program "psin.c" needs the subroutine "quad.o", then we will have to use the command

gcc -Wall -o sintable psin.c -L/home/tom/lib/ -lm -ltest

"-o sintable" means the executable binary file is "sintable", and not "a.out".

"-Ldirname" = "-L/home/tom/lib/" means the library in stored in the directory "/home/tom/lib/". We must tell "gcc" where we store the library file. "gcc" by default will search "/usr/lib/".

"-lm -ltest" means we use two libraries "libm.a" and "libtest.a". The former is a system library and is stored in "/usr/lib/", the latter is our library stored in "/home/tom/lib/"

Exercise : Suppose "psin.c" needs "quad.c" and "root.c", and we do not want to use "library", what is the command ?

Ans :

gcc -Wall -o sintable psin.c quad.c root.c -lm

Exercise : Suppose the source file is "prog1.c", and we want the executable binary file to have name "pro". "prog1.c" uses 2 libraries "libp1.a" under /home/john/lib and the usual Maths library, "libm.a". What is the command ?

Ans :

gcc -Wall -o pro prog1.c -L/home/john/lib/ -lm -lp1

Header Files stored not in "/usr/include"

(You may not understand this section now, but when you have learned some C, you will.)

The standard header files are usually stored in directory "/usr/include/". For users, it is inadvisable to store header files there.

We may make a directory in our directory, e.g. "/home/tom/include/", and put all our non-standard header files there. Now to inform gcc that we store header files there, we will have to use, e.g.

gcc -Wall psin.c -I/home/tom/include/ -L/home/tom/lib/ -lm -lp1

gcc, by default, will search "/usr/include/".

You just type "info" and press "Enter".

"info" will display a list of topics available. You have to use the following commands to navigate,

spacebar	Pressing "spacebar" will scroll down one page.
<-	This is the key at the top-right corner, not the cursor movement key. It will scroll up one page. These may be discouraging for us who are accustomed to Microsoft's mouse and scrollbars, and we have to get used to it gradually.
m	[menu] You may select the topics you want. Type "m", and you type in the "topic name" at the bottom row. "Tab" key is used for automatic completion, e.g. if the topic is "make", and if there is only one manual item beginning with "ma", then you type "ma" and press "Tab", then it will be completed as "make". Then you press "Enter". "ctrl + c" means the usual "stop the program", and may be useful when "info" get mixed up.
p	[previous] "p" will bring you to the previous "node, or page". "info" is a crude form of information system before the invention of WWW, Internet.
u	[Up]. The "info" system is organized hierarchically. "u" means moving one level up the hierarchy.
n	[Next]. Display the next page.
b	[Beginning]. Move back to the beginning of the document. When we have scrolled down many pages with "spacebar", either we may use "<-" to scroll up page by page, or simply "b", to to go the beginning.
q	[Quit]. Leave the "info" command.

gzip psin.c root.c quad.c equation.c

gzip *c

gzip prog{1,2,3}.c

gunzip prog1.c.gz prog2.c.gz

cd fina [Tab]

tar [-options -f] tarfilename file1 file2 .... dir1 dir2 ....

-f	"-f tarfilename" specifies the backup file name.
-c	Create backup file. Put all "file1 file2 file3 .... dir1 dir2 dir3 ...." into one file "tarfilename".
-x	[Extract] Restore back the files "file1 file2 file3 .... dir1 dir2 dir3 ...." and put them under the current directory, creating the directories if necessary.
-t	[table of content] List all files in "tarfilename" but without doing anything to the file "tarfilename". Just list the table of content.
-v	[verbose] This option should be used with most commands. It means "print out everything the command does".
-z	[zip or unzip] If we are creating a backup file, this option will "compress" the backup file. Hence "tarfilename" we use should have ".gz" at the end to remind us of this fact. (The command will compress even if we do not put .gz at the end of filename. But then we do not the file has been compressed.) e.g. tar -cvf /tmp/mybackup /home/ gzip /tmp/mybackup is the same as tar -cvzf /tmp/mybackup.gz /home/ Here everything under "/home/' will be put into file "/tmp/mybackup.gz" compressed. If we are retoring from a backup file which has been compressed, then this option will automatically decompress the file. e.g. tar -xvzf /tmp/mybackup.gz will restore all files in "mybackup.gz" and put them under the current directory, decompressing the file "mybackup.gz" automatically.

Examples

(Options are : create, verbose, compression, filename)
Create archive file from files "psin.c, root.c, quad.c, equation.c" and ALL files in the directory "/home/tom/lib/". Compress the resulting file, and put it in "/tmp/abc.gz". Later, we may copy this file onto floppy, or another hard drive. Details will be explained later.

(Options are : table of content, compression, filename)
List all files in this "archived" file "/tmp/abc.gz"

(Options are : extraction, verbose, (de)compression, filename)
Restores all files in "/tmp/abc.gz", decompressing it first, and put all files under the current directory.

Exercise : How are you to put files "rep1.txt, rep2.txt, rep3.txt" and all files under "/home/john/dir1/", into one compressed file "/tmp/reportbak.gz"?

Ans :

tar -cvzf /tmp/reportbak.gz rep{1,2,3}.txt /home/john/dir1/

OR

tar -cvzf /tmp/reportbak.gz rep1.txt rep2.txt rep3.txt /home/john/dir1/ 

Exercise : Suppose you are currently in directory "/tmp/workdir/" and you type the command

tar -xvzf /tmp/reportbak.gz

What will happen?

Ans : You will find "rep1.txt, rep2.txt, rep3.txt" have been restored in "/tmp/workdir/", also there is a new directory "/tmp/workdir/home/john/dir1/" and all files inside there are also restored too.

Exercise : How are you to inspect the files in "/tmp/reportbak.gz" without retoring them?

Ans :

tar -tzf /tmp/reportbak.gz | more

This will "kill" the process whose "process ID" is specified, e.g.

kill -9 104

This "kill" command is often used in the following situation : a running program suddenly "hangs", giving no response. Then we may log in as "root", i.e. "super-user" on another screen (or terminal), with "Ctrl + Alt + F2" (F1 - F6 usually may be used), then "ps -a" to see all processes, then "kill -9 [pid]" to terminate that job.

cat /mnt/file1.txt

We may put more than one commands on one line. The commands have to be separated by ";" e.g.

cd /tmp ; ls *c ; cd

e.g. If "sintable" takes a long time to run, then we may

sintable &

To see all the jobs in the background, we issue the command

jobs

fg n%

gcc -Wall psin.c -lm && a.out

Command file

How to prevent accidental deletion of files

The "alias" and "unalias" command

If we want to cancel it, we type

unalias rm

The "find" command

Format of "find" command,

Examples :

find /home/tom/ -name "*.c" -print

Find within the directory "/home/tom/" all files whose name end with ".c" and print the names on the screen. (Note : all sub-directories inside /home/tom/ will be searched too).

find . -size +3k -print

Find within current directory (note : "." means current directory, ".." means parent directory) all files whose size are greater than 3k bytes.

find . -size -3k -print

Find within current directory all files whose size are less than 3k bytes.

find . -size 3k -print

Find within current directory all files whose size are roughly 3k bytes.

find / -mtime 3 -print

Find from the topmost directory (and all directories under /) all files whose "modification time" is 3, i.e. the files have been modified in the last 3 days.

find / -atime 4 -print

Find from the topmost directory (and all directories under /) all files whose "access time" is 4, i.e. the files have been accessed in the last 4 days.

find /home/john/ -type d -name "pl*" -print

Find within the directory "/home/john/", all directories (-type d) whose name start with "pl".

find /home/john/ -type f -name "pl*" -print

Find within the directory "/home/john/", all files (-type f) whose name start with "pl".

Note : "-exec command {} \;" means "apply command to those found".
"{}" stands for those found, and "\;" signifies the end.

find /home/tom/ -type f -name "*bak" -size +5000k -print -exec rm {} \;

Find within the directory "/home/tom/", all files (-type f) whose name end with "bak" and whose size is greatter than 5000k, and then remove them.

For more details about "find", please use "man find" command to read the online manual about "find".

Recurrsive copy or move

cp -rv /home/tom/ /home/test/
mv -rv /home/tom/ /home/john/
rm -rf /home/tom/

Pascal, Fortran

There is a "pascal to C" program in LINUX, called "p2c"

The first command will change a file whose extension is ".pas" to a C program, e.g. "myprog.pas" will produce "myprog.c". Then we have to compile this C program with gcc. Note that we need a library "libp2c.a".

There is also a "Fortran to C" program in LINUX, called "f2c"

The first command will change a file whose extension is ".f" to a C program, e.g. "myprog.f" will produce "myprog.c". Then we have to compile this C program with gcc. Note that we need a library "libf2c.a". But usually there is a "f77", or "fort" that combines these 2 steps into one.

For more details, you may read the online manual, e.g. "man f2c", "man p2c".

More about Backup

(You may omit this)

If everything is under "/home/", then we just back up this directory.

For incremental backup, we may write a shell script, e.g.

      tarfile = /tmp/backup.gz
      tarlist = /tmp/backuplist
      echo "Which directory to back up ? "
      read dirbk
      echo "Backup files accessed during the last m days. m ?"
      read days
      cd $dirbk
      find $dirbk -type f -mtime -$days -print > $tarlist
      tar cvzf $tarfile `cat $tarlist`

There is also a "cpio" command to do backup, e.g.

      ls | cpio -ov > /tmp/backup.cpio
           (This  "lists" all files in the current directory, 
            and "pipe" the names to the program "cpio" for backup.  
            "-o -v" options mean "outwards, verbose". 
            And the output file is /tmp/backup.cpio)

      cpio -i < /tmp/backup.cpio
           (This command will restore the backup files.  
            Option "-i" means "inwards".)

      find . -mtime 3 -print | cpio -ov > /tmp/backup.cpio
           (Find all files within the current directory which have been 
            modified in the last 3 days and "pipe" them to "cpio" for 
            "outward" backup.)

      cpio -it < /tmp/backup.cpio
           (List the "table of contents" of /tmp/backup.cpio.  
            But restoration of files will not be done.
            Options = "inwards, table of content".)

      cpio -id < /tmp/backup.cpio
           (Restores the files. Options = "inwards, (restore)directory, if
            they have been removed".)

For more details, please read the online manual "man cpio"