Redirections

When you run a command. there are 3 file descriptors open by default. 0 which referes to the standard input of that command. 1 refers to standard output. And 2 refers to standard error.

Before the command runs. you can use a special notation to redirect those file descriptors, duplicate them, close them, or make them refer to other files.

Input Redirection

The general format is as follows:

[n]<word

This will cause the file whose name results from the expansion of word to be opened for reading on file descriptor n. Or standard input (file descriptor 0) if n is not specified.

Example

cat <file.txt 

# Or
cat 0<file.txt 

# Or
cat 3<file.txt

Output Redirection

The general format is as follows:

[n]>word

This will cause the file whose name results from the expansion of word to be opened for writing on file descriptor n. Or standard output (file descriptor 1) if n is not specified. If the file does not exist it is created. If it does exist it is truncated to zero size.

Example

echo "Hello World" >file.txt 

# Or
echo "Hello World" 1>file.txt

# Or
echo "Hello World" 3>file.txt

Appending Redirected Output

The general format for appending output is as follows:

[n]>>word

Redirection of output in this fashion causes the file whose name results from the expansion of word to be opened for appending on file descriptor n, or the standard output (file descriptor 1) if n is not specified. If the file does not exist it is created.

Example

echo "Hello World" >>file.txt 

# Or
echo "Hello World" 1>>file.txt

# Or
echo "Hello World" 3>>file.txt

Redirecting Standard Output and Standard Error

This construct allows both the standard output (file descriptor 1) and the standard error output (file descriptor 2) to be redirected to the file whose name is the expansion of word.

The format is:

&>word

This is semantically equivalent to:

>word 2>&1

Example

echo "Hello World" &>file.txt 

# Or
echo "Hello World" 1>file.txt 2>&2

Appending Standard Output and Standard Error

This construct allows both the standard output (file descriptor 1) and the standard error output (file descriptor 2) to be appended to the file whose name is the expansion of word.

The format of appending all output is:

&>>word

This is semantically equivalent to:

>>word 2>&1

Example

echo "Hello World" &>>file.txt 

# Or
echo "Hello World" 1>>file.txt 2>&2

Here String

The format is simple as:

[n]<<<word

This notation causes the string resulted from the expansion of word to be provided as input at the file descriptor n. Or at the standard input (file descriptor 0) if n is not specified.

Example

cat <<<"hello World"

# Or
cat 0<<<"Hello World"

# Or
cat 3<<<"Hello World"

Duplicating file descriptors

Duplication means to make a file descriptor refers to the same file that another file descriptor refers to. For example. Duplicating file descriptor x on y means to make the file descriptor x refers to the same file that file descriptor y refers to.

The format is:

[n]<&word

This format causes the file descriptor resulted from the expansion of word to be duplicated on file descriptor n. Or on standard input (file descriptor 0) if n is not specified.

It is possible to use this other format as well:

[n]>&word

The only difference is that when you omit n. The default is standard output (file descriptor 1).

Example

In this example, we will duplicate the file descriptor 1 on 2.

echo foobar 1>file.txt 2>&1

The file descriptor 1 is made to refer to the file file.txt. then we duplicated the file descriptor 1 on file descriptor 2. This means that when the command echo writes to file descriptor 2. It will write to the file file.txt. In other words. the file descriptor 2 is now a copy of the file descriptor 1.

Closing file descriptors

The syntax for closing file descriptors is

[n]<&-

Notice the - at the end. This notation causes the file descriptor n to be closed. or standard input (file descriptor 0) if n is not specified.

It is possible to use this other format as well:

[n]>&-

The only difference is that when you omit n. The default is standard output (file descriptor 1).

Example

echo foobar 1>&-

In this example, we closed the file descriptor 1. when the command echo writes to file descriptor 1. It will fail because the file descriptor is closed.

Moving file descriptors

This notation is a combination of moving and closing file descriptors. In simple words, moving file descriptor x to y means to duplicate file descriptor y to x. then close the file descriptor y.

The format is:

[x]<&y-

This means, duplicate the file descriptor y to x. then close y. if x is not specified, 0 is the default. An alternative syntax is available:

[x]>&y-

The only difference is that when you omit x. The default is standard output (file descriptor 1).

Example

In this example, we will move the file descriptor 2 to 1.

echo foobar 1>&2-

Now, the file descriptor 1 is made a copy of the file descriptor 2. and the file descriptor 2 was closed. This means. when the command echo writes to file descriptor 1. It will write to the same file to which file descriptor 2 was pointing to. and it will fail to write to file descriptor 2 because it was closed.

Open file for reading and writing

The format is:

[n]<>word

This notation causes the file whose name is the expansion of word to be opened for both reading and writing on file descriptor n, or on file descriptor 0 if n is not specified. If the file does not exist, it is created.

Internals of file descriptor management

It is necessary to make something clear about how file descriptors are managed in bunster. bash users usually struggle when they see a bunster script using a redirection like this:

echo foobar 9999999>file.txt >&9999999

This code will fail in bash because 9999999 is not a valid file descriptor. This is logical because bash relies on the kernel to manage file descriptors. Which -- the kernel -- has a strict rules about the format of file descriptors. Also, when you run this code in bash:

cmd 5>file.txt

Bash literally opens the file descriptor 5. As a result. the command cmd will inherit that file descriptor and can read/write to it.

Bunster is built differently. The following code is going to work totally fine in bunster :

echo foobar 9999999>file.txt  >&9999999

That's because file descriptors are managed by the bunster runtime. And are not real file descriptors. They're just an alias of a file handler. This means that in the above example. The command echo will not inherit the file descriptor 9999999 because that file descriptor is not open in reality.

Why should I be concerned about file descriptor managment ?

you shouldn't ! there is nothing to worry about regarding the managment of file descriptors. Because the behavior of your script is totally compatible between bash and bunster. You don't have to change anything in your bash scripts to work in bunster. And vice-versa. The only reason why we decided to mention all these information in the documentation is to clarify that while the behavior is similar. The internals are different. and so don't worry when you run ls /proc/self/fd and you see a different output in bunster than in bash.