You’ll need:
What you do:
For UNIX, Mac and Linux:
For Windows 10:
For Windows w/o ssh client, use PuTTY
Command:
ssh [UWO_username]@compute.gaul.csd.uwo.ca
From here, you can enter your password and access GAUL
Notes:
who
whoami
pwd
ls
cd [name_of_directory]
cat [name_of_text_file]
mkdir [new_directory_name]
mv [file_name] [new_directory_path]
cp [name_of_file] [path_to_file]
rm [name_of_file]
man [command]
sftp
% sftp user_name@compute.gail.csd.uwo.ca
password: [enter your password]
// use lls, lpwd and lcd to navigate through the home file system
// once you get to the file you want:
get [name_of_remote_file] //to download from GAUL to home
put [name_of_local_file] //to upload from home to GAUL
quit //make sure you do this when you're done!!
Editors come in 2 types:
Modal:
Modeless:
Starting, you do
vi [name_of_file] //new or existing
Basic commands:
Colon command mode:
Starting
% emacs -nw [name_of_file] //new or existing
Commands:
To cut and paste:
Commands:
Recall that each Unix process has three streams opened upon start; input, output and diagnostic/error messages
Input is represented as stdin and has a file descriptor of 0, output is stdout and 1, and error is stderr and 2
For your terminal, stdin is the keyboard, while stdout and stderr is the display
This is all well and good, but what if you wanted to change where this outputted? What about input?
Luckily, this is quite simple
< to redirect standard output (or 0<)
to redirect standard input (or 1>)
2> to redirect standard error
For example, if we want to output the following command to a text file…
% whoami
…we can modify it like so
% whoami > file1
Just like normal file usage, file1 is created if it’s not there
However, depending on your bash setup, Unix can sometimes get mad at you if file1 already exists
So what can you do? You can force the overwrite (or clobber) the file
% whoami >| file1
You can also append the output to file1
% whoami >> file1
You can also use this method to separate output and error messages
% whoami > file1 2> file_error
You can also direct them to the same place like so
% whoami &> file1
A more general way goes as follows
% whoami > file1 2>&1 //stdout goes to file1 and stderr goes to where stdout is
For redirecting input, there’s a very good use case with the command tr
% tr string1 string2 text
This command will do the following
Usually this outputs to the terminal, but we can change this with our redirections
For example, if we want to do a translation on our file1 and put it into file2, we can do so like this
% tr aeio 4310 < file1 > file2
Using this type of command, we can also do encoding, somewhat like a cypher
We can also put in regular expressions (which we will cover in a bit)
% tr a-z A-Z < file1 > file2
For an application of error redirection, let’s say you don’t want to display error messages for whatever reason
Recalling /dev/null, the directory that makes files disappear, we can set out error output to this directory to discard them
% tr a-z A-Z < file1 > file2 2> /dev/null
This tr command is an example of what we call a filter
Filters are simply programs that
Although these filters can take user input, they don’t need it thanks to our redirection capabilities
The command tr is an example of a filter command, as well as others, such as:
grep pattern
This reads stdin and writes any lines containing the regex pattern in the argument
% grep a-z < file1 > file2
wc
Counts the # of chars/words/lines on stdin and writes the result to stdout
sort
Sorts input lines in alphabetical order and writes to stdout
Something we can also do is use take the output of a command and use it as the input for another
% command1 | command2 | command3 | ... | command_n
You can also pipe any error messages that you get
% command1 |& command2
If we run into a scenario where we want to pipe to a file AND our terminal, we can use the tee command
% command1 | tee file1
Recall that Unix is a multi-tasking operating system and not all tasks are controlled by you, either by other users or in the background
When you run a task (ls, vi, etc.) it executes in the foreground of your shell, meaning that it has control over your screen and keyboard
But what if you want a task in the background? This can be easily done with the & symbol
% command &
This is especially useful for heavy tasks that you don’t want to take over your computer, leaving you to do other things while you wait
Any output, once finished, will immediately output to your screen, which can be a bit confusing
When input is required, the process prints a message and stops
Upon completion, a message is printed
Explicitly calling background jobs aren’t needed as much for windowing systems (eg. Windows) since you can just go to another window and run the command, but for terminal-based systems (Unix) this is very useful
You also (fortunately) have a lot of control over your jobs in Unix
While you can terminate a job in the foreground with Ctrl-C, you can also temporarily pause a job with Ctrl-Z
Doing this, the job can be either taken into the foreground, taken into the background or killed
To see which jobs are running/stopped, you can check using the jobs command
% jobs
This will pull up a list of jobs that have either stopped or are still running
If you wish to move a job into the background you can do the following
% bg %n //brings the nth job into the background
You can also kill jobs using the kill command, much like Ctrl-C
% kill %n
% kill -KILL %n //use this if the above command doesn't work
This kill command can also have the same functionality as Ctrl-Z with the following switch
% kill -STOP %n
Jobs are simply a special type of Unix process that are started by Unix shells, so if you want to see every current process, you must use the ps command
% ps
This can take many options depending on which ps command you use
-l
Gives a long list of what’s going on
-u login_id
Shows info about login_id’s processes
You can use the following command to see more options
% man ps
If you want to kill any process, you can do the following
% kill process_id
The ulimit command is also quite useful, setting and reporting file-size writing limits on files written by the shell or child processes
While any user can decrease the limit, you need certain privileges to increase it
This also comes with some switches you can use
-a
Prints all limits
-n
Max number of open file descriptors
-u
Max number of processes available
Sometimes we want to search for certain patterns, which is perfectly done by a regular expressions, special strings that define a search pattern for matching text
These are used in many Unix utilities (grep, vi, etc.) as well as outside of Unix in other programming languages, which then interpret these strings
These strings can be regular text…
% grep unix file //matches only unix in file
…or they can be special text
% grep '[uU]nix' file //matches unix and Unix
% grep [uU]nix file //filename expansion
With special modifications, you can match a wide array of strings
.
Matches any single character
Ex. a.b matches axb and a$b but not axxb
-
Matches zero of more occurrences of the previous single character pattern
Ex. a*b matches b, ab, and aab, but not axb
*
Matches one or more occurrences of the previous single character pattern
Ex. a+b matches ab and aab, but not b or axb
?
Matches zero or one occurrences of the previous single character pattern
Ex. a\?b matches b and ab, but not aab or axb
You can represent sets of characters with […] (ex. [wxyz] matches w, x, y or z)
This can also match ranges of characters (ex. [a-z] matches anything from a to z)
It is also worth noting that wildcards lose any meaning once they’re in these brackets
You can also match specific parts of lines or words using regular expressions
Matches the beginning of a line
ex. ^TITLE
$
Matches the end of a line
ex. FINI$
<
Match beginning of a word
>
Match the end of a word
You can match the opposite of a set with ^
ex. [^aeiou] matches any non-vowel
You can also escape the special meaning of character with \
ex. * will only match asterisks
You can refer to the most recent match with (…)
Recall the remembered portion with \n, where n is 1-9
ex. ‘([a-z])\1’ matches any pair of duplicate letters
X{m,n}
Matches m – n repeats of the one character regular expression X
X{m}
Matches exactly m repeats of the one character regular expression X
X{m,}
Matches at least m repeats of the one character regular expression X
-c
Count the number of matches
-i
Ignore case distinction
-v
Select non-matching lines
If you were to use a utility like egrep or grep -E, you can make use of an extended set of wildcards and other mechanisms
-
Works like before, except you don’t need a backslash
| r1 | r2 |
Acts as a logical or operator between two regular expressions
(…)
Allows for single character wildcards to apply to an entire encased regular expression
Some character classes are predefined, including [:lower:] (a-z), [:upper:] (A-Z), [:alpha:] (A-Za-z), [:digit:] (0-9) and [:alnum:] (A-Za-z0-9)