2.3 Working with bash shell
bash has several features that make the command line usage easier and more effective. You do not have to always type the whole command line character by character. When you are typing the command, you can go backwards and forward by using the leftwards and rightwards arrow keys. You can delete the character left to the cursor with backspace key. Pressing Ctrl-d or the Delete key deletes the character under the cursor. In addition to these basic command line editing functions you can use the keyboard commands listed in the table below.
Table 2.4: Short-cuts for editing command line
Move one the cursor character forward in the command line
Move one the cursor character backwards in the command line
Move the cursor to the beginning of the command line
Move the cursor to the end of the command line
Cuts all the characters rightwards from the cursor
Paste the characters, cut with Ctrl-k, to the command line
Show the previous command
In bash the executed commands are stored to users home directory into a file called .bash_history. To see the full list of recently executed commands give command history. In command line, you can browse the list of previous commands with the upwards arrow key and the downwards arrow key. Alternatively you can use Ctrl-p and Ctrl-n. In cases, where you need to give similar commands several times, it is often handy to get one of the previous commands to the command line with the arrow key. Then you can edit just the modifications needed to the old command and execute the modified command by pressing Return key.
It's possible to search through the command history. Press Ctrl-r then type what you want to search for. You do not need to type a complete search, and the search results will update with each character pressed. To move between the search results use the upwards arrow key and the downwards arrow key. To run the command press the Return key. If you wish to edit the search result before running the command use the leftwards arrow key, then edit the command.
If you press Tab key the bash shell tries to complete the command or argument you are writing. The completion is done as far as possible using the lists of available commands and files. Using auto-completion is very recommended as it saves the user from typing all command or argument characters and also takes care that command don't get mistyped. For example, let's say we are in a directory where we have two files: final_research_report_old.pdf and final_research_report_new.pdf.
To open the latter of the files with evince program we would need to type command:
This command would require you to type 36 characters. However, by using Tab completion you need to type only 6 characters. First type
evinand then press the Tab key. Normally at CSC, evince is the only available program or command that starts with letters evin so when the tab completion is executed, it knows to complete the rest of the command:
Then to define the file name, you can type just the first letter of the file name
When you now press Tab, the completion process checks what files, starting with f, are available. In this case there is two of them and as the beginning of the two file names is the same string, the command can now be completed to:
Now you just need to type one n to the end of file name to distinguish the file from the old version,
When the Tab key is now pressed again there is only one option that matches the beginning of the argument that has been typed and thus the command is completed and ready to be executed:
In linux, graphical interfaces and commands that are not interactive once they have started, can be executed as background processes. When the command is executed as a background process the command shell does not wait until the command is finished. Instead it remains active and allows user to submit new commands, while the background command gets executed. However, note that in the computing clusters of CSC, the heavy computing tasks should not be executed as background processes, but they should be submitted to the batch queue system.
In normal interactive usage you can launch the command to be executed in the background by adding & character to the end of the command. For example command:
eog image1.jpg &
would open the Eye of GNOME image viewing program (requires X-term or FreeNX connection) to the background so that the command shell could still be used even though the eog program is still running. A background process can be changed to normal, foreground process with command fg. When a command and program is running interactively, i.e. the command shell is waiting that the execution finishes, you can terminate the execution by pressing Ctrl-c. Another possibility is to halt the program by pressing Ctrl-z. When command (or process) is halted, it can be continued with fg command or changed to be executed as a background process with command bg.
A command that is still running can be terminated with kill command. To be able to use kill, you need to know the process identification number (PID) of the command you want to terminate. You can check your active processes, meaning the commands that you are currently executing, with command ps. For example:
kkayttaj@hippu1:~>ps PID TTY TIME CMD 385 pts/12 00:00:00 tcsh 2001 pts/12 00:00:00 eog 2003 pts/12 00:00:00 gconfd-2 2203 pts/12 00:00:00 ps
By default, the ps command shows only those processes that have been launched from the command shell you are currently using. To see all your processes in the server you have logged in, give command:
ps -f username
Once you have identified the correct process number you can kill the process with command:
For example command:
Would kill the eog program, listed in the output of the previous ps example. You can kill only processes that are owned by your account. Sometimes, when want to kill a process that is malfunctioning, the normal kill command may not able to terminate the process. In those cases you can try to terminate the process by adding option -9 to the kill command:
kill -9 process_ID
|Previous chapter||One level up||Next chapter|