cd /home/robert/dir1

cd dir1

Change up one level, change to "parent" directory.

If you do not know "absolute path name", "relative pathname", you should read my Qbasic webpages.

vdir prog*

all files whose name starts with "prog" will be listed.

vdir *.c

all files whose extension is ".c" will be listed.

vdir ab*pl

all files whose name starts with "ab" and ends with "pl" will be listed.

cat filename | more

cp * /home/tom/dir1/

mv prog{1,2,3}.c /home/john/dir2/

mv *c /tmp/abc/

rm abc*

shutdown -h now

du *
(List the size of all files)

du *.c
(List the size of all files whose name ends with ".c" in Kbytes).

du ab*cd
(List the size of all files whose name begins with "ab" and ends with "cd")

du record1 filea
(List the size of the two files, "record1" and "filea")

file *
(Show the type of all files in current directory)

file chap{1,2,3}
(Show the type of files of "chap1, chap2, chap3". Note the use of curly brackets and comma.)

head prog.c
(Type out 10 lines of "prog.c")

head -20 prog.c
(This types 20 lines, "-20" is "option". All "options" begin with "-" in UNIX.)

tail prog.c
(Type out last 10 lines of "prog.c")

tail -15 prog.c
(This types last 15 lines.)

wc *
(Shows the "word count" of all files in current directory.)

wc abc{1,2,3}.c
(Shows the word count of "abc1.c abc2.c abc3.c")

man reset (Shows how to use the command "reset".)

To quit, we type "q". (ctrl + C to quit will not work here).

apropos compiler

/home/robert/dir1/subdir2/

y	yes, replace
n	no. skip this one and continue
q	quit
.	Replace this one, and then quit
^	Go back to previous match
!	Global replacement without further asking.

The default binary file from gcc is "a.out". We may change it with "-o binaryfilename" option.
e.g. we wish the binary file name from "psin.c" to be "sintable", and not "a.out", then

gcc -Wall -o sintable psin.c -lm and to execute the program, type sintable

Suppose we are to write a database program. The main program is in file "dbmain.c", and it needs five subroutines stored separately in files "add.c", "delete.c", "change.c", "create.c", "report.c".

The final binary file should have name, "dbmain".

For large programs, we should form "good programming practice" of writing each subroutine separately, and test them individually, before linking them into the main program.

Exercise :What files will be produced by the following commands,

gcc -Wall -c add.c
gcc -Wall -c delete.c
gcc -Wall -c change.c
gcc -Wall -c create.c
gcc -Wall -c report.c
gcc -Wall -c dbmain.c

Ans : add.o, delete.o, change.o, create.o report.o dbmain.o

Exercise : How to link them all into one program. Suppose we have to use library "libreport.a", and we wish the final binary file to be named "dbmain"?

Ans :

  gcc -Wall -o dbmain dbmain.o add.o delete.o change.o  create.o report.o -lreport

Exercise : How to check that the above command has been successfully run? and how to run the program, should there be no error?

Ans : To check the result of the last command, type "echo $?". If it shows "0", that means "o.k." Any value other than "0" signals error.
To run the program, just type "dbmain"

Exercise :Suppose after several days, we make some modification to the subroutines "report.c" and "change.c", what should we re-compile?

Ans :

       gcc -Wall -c report.c
       gcc -Wall -c change.c
       gcc -Wall -o dbmain dbmain.o add.o delete.o change.o create.o report.o -lreport

It can be seen that for each change we make, we have to re-type many commands (also, we may make mistakes in re-typing these commands. Or worse, we may forget the commands, or the filenames, or the subroutines, after some period of time). There is a UNIX command (or utility) to automate all these processes, called "make". Firstly, we have to use "emacs" to key in the following file, and use "makefile" for its filename.

 dbmain :  dbmain.o add.o delete.o change.o create.o report.o 
     gcc -Wall -o dbmain dbmain.o add.o delete.o change.o create.o report.o -lreport

 add.o : add.c
     gcc -Wall -c add.c

 delete.o : delete.c
     gcc -Wall -c delete.c

 change.o : change.c
     gcc -Wall -c change.c

 create.o : create.c
     gcc -Wall -c create.c

 report.o : report.c
     gcc -Wall -c report.c

 dbmain.o : dbmain.c  
     gcc -Wall -c dbmain.c

Notice that there is "indentation" before "gcc". You must press "Tab" to create this indentation, and NOT pressing spacebar several times. The utility program will fail if you do this.

The Utility "make"

When you just type in the command "make", the utility will search the file "makefile" (it will search only this file, hence you must use this file name.) and perform functions as shown

Notice that there many be many commands after the "target : dependencies" e.g.

target1 : dep1 dep2 dep3 .... command1 command2 command3 .... target2 : dep4 dep5 dep6 .... command4 command5 .... ...........

Sometimes, we wish to check what commands "make" will execute, without actually executing the commands, (this is useful in debugging the "makefile"), we may use

make -n

i.e. we use an "-n" option.
"make" will list all commands that it will execute, but without actually executing them.

Also, if the subroutines are too many to be put in one line in the "makefile", we should write, e.g.

          gcc -Wall -o dbmain  dbmain.o sub1.o sub2.o sub3.o sub4.o \
          sub5.o sub6.o sub7.o sub8.o sub9.o sub10.o sub11.o \
          sub12.o sub13.o

Note that "target" may be a non-existent filename. In that case, it will check the dependencies, and finally execute the commands. e.g. your "makefile" may be

dummy : gcc -Wall psin.c -lm

Here "dummy" does not exist as a file, hence "make" will always execute

gcc -Wall psin.c -lm

If you want the executable binary to have name "sintable", you should use

sintable : psin.c gcc -Wall -o sintable psin.c -lm

for the "makefile".

Reason why there may be more than one buffer is :

Sometimes we wish to "copy" from one file, and "paste" to another. Hence inside "emacs", we will have to open more than one files. This is done by "ctrl + x -> ctrl + f".

You can jump from one file to another also with this "ctrl + x -> ctrl + f", just type in the filename of the file you want to jump to.

After a while, you may form the custom of openning many files at once, and you may forget the filenames!
"ctrl + x -> ctrl + b" will show all the files you have opened. You may jump to that window, put the cursor at the filename you want to jump to, and press "Enter".

How to select several lines like that we are accustomed to do in Window ?

First move cursor to the starting position, and press

Ctrl + spacebar

Then move cursor to the ending position, usually the line below the last line to be selected, and press

Alt + w

Ctrl + w

Then move to the place where we want to "paste". It may be in the same file, or in another file (that file should have already been opened with "Ctrl + x -> Ctrl + f") and press

Ctrl + y

Ctrl + y has other uses, e.g. We delete some lines somewhere (ctrl + k = kill = delete a line), and want to put the deleted lines at other places. Then move the cursor there, and press "Ctrl + y".