Class: IO

Returns a new IO object (a stream) for the given integer file descriptor fd and mode string. opt may be used to specify parts of mode in a more readable fashion. See also .sysopen and .for_fd.

.new is called by various ::File and IO opening methods such as .open, Kernel.open, and File.open.

Open Mode

When mode is an integer it must be combination of the modes defined in ::File::Constants (File::RDONLY, File::WRONLY|File::CREAT). See the open(2) man page for more information.

When mode is a string it must be in one of the following forms:

fmode
fmode ":" ext_enc
fmode ":" ext_enc ":" int_enc
fmode ":" "BOM|UTF-*"

fmode is an IO open mode string, ext_enc is the external encoding for the IO and int_enc is the internal encoding.

IO Open Mode

Ruby allows the following open modes:

"r"  Read-only, starts at beginning of file  (default mode).

"r+" Read-write, starts at beginning of file.

"w"  Write-only, truncates existing file
     to zero length or creates a new file for writing.

"w+" Read-write, truncates existing file to zero length
     or creates a new file for reading and writing.

"a"  Write-only, each write call appends data at end of file.
     Creates a new file for writing if file does not exist.

"a+" Read-write, each write call appends data at end of file.
     Creates a new file for reading and writing if file does
     not exist.

The following modes must be used separately, and along with one or more of the modes seen above.

"b"  Binary file mode
     Suppresses EOL <-> CRLF conversion on Windows. And
     sets external encoding to ASCII-8BIT unless explicitly
     specified.

"t"  Text file mode

The exclusive access mode (“x”) can be used together with “w” to ensure the file is created. Errno::EEXIST is raised when it already exists. It may not be supported with all kinds of streams (e.g. pipes).

When the open mode of original IO is read only, the mode cannot be changed to be writable. Similarly, the open mode cannot be changed from write only to readable.

When such a change is attempted the error is raised in different locations according to the platform.

IO Encoding

When ext_enc is specified, strings read will be tagged by the encoding when reading, and strings output will be converted to the specified encoding when writing.

When ext_enc and int_enc are specified read strings will be converted from ext_enc to int_enc upon input, and written strings will be converted from int_enc to ext_enc upon output. See Encoding for further details of transcoding on input and output.

If “BOM|UTF-8”, “BOM|UTF-16LE” or “BOM|UTF16-BE” are used, Ruby checks for a Unicode BOM in the input document to help determine the encoding. For UTF-16 encodings the file open mode must be binary. When present, the BOM is stripped and the external encoding from the BOM is used. When the BOM is missing the given Unicode encoding is used as ext_enc. (The BOM-set encoding option is case insensitive, so “bom|utf-8” is also valid.)

Options

opt can be used instead of mode for improved readability. The following keys are supported:

:mode

Same as mode parameter

:flags

Specifies file open flags as integer. If mode parameter is given, this parameter will be bitwise-ORed.

:external_encoding

External encoding for the IO.

:internal_encoding

Internal encoding for the IO. “-” is a synonym for the default internal encoding.

If the value is nil no conversion occurs.

:encoding

Specifies external and internal encodings as “extern:intern”.

:textmode

If the value is truth value, same as “t” in argument mode.

:binmode

If the value is truth value, same as “b” in argument mode.

:autoclose

If the value is false, the fd will be kept open after this IO instance gets finalized.

Also, opt can have same keys in String#encode for controlling conversion between the external encoding and the internal encoding.

Example 1

fd = IO.sysopen("/dev/tty", "w")
a = IO.new(fd,"w")
$stderr.puts "Hello"
a.puts "World"

Produces:

Hello
World

Example 2

require 'fcntl'

fd = STDERR.fcntl(Fcntl::F_DUPFD)
io = IO.new(fd, mode: 'w:UTF-16LE', cr_newline: true)
io.puts "Hello, World!"

fd = STDERR.fcntl(Fcntl::F_DUPFD)
io = IO.new(fd, mode: 'w', cr_newline: true,
            external_encoding: Encoding::UTF_16LE)
io.puts "Hello, World!"

Both of above print “Hello, World!” in UTF-16LE to standard error output with converting EOL generated by #puts to CR.

Same as .write except opening the file in binary mode and ASCII-8BIT encoding ("wb:ASCII-8BIT").

copy_stream copies src to dst. src and dst is either a filename or an IO-like object. IO-like object for src should have #readpartial or #read method. IO-like object for dst should have #write method. (Specialized mechanisms, such as sendfile system call, may be used on appropriate situation.)

This method returns the number of bytes copied.

If optional arguments are not given, the start position of the copy is the beginning of the filename or the current file offset of the IO. The end position of the copy is the end of file.

If copy_length is given, No more than copy_length bytes are copied.

If src_offset is given, it specifies the start position of the copy.

When src_offset is specified and src is an IO, copy_stream doesn’t move the current file offset.

Synonym for .new.

Executes the block for every line in the named I/O port, where lines are separated by sep.

If no block is given, an enumerator is returned instead.

IO.foreach("testfile") {|x| print "GOT ", x }

produces:

GOT This is line one
GOT This is line two
GOT This is line three
GOT And so on...

If the last argument is a hash, it’s the keyword argument to open. See .readlines for details about getline_args. And see also .read for details about open_args.

Alias for File.open. With no associated block, open is a synonym for .new. If the optional code block is given, it will be passed io as an argument, and the IO object will automatically be closed when the block terminates. In this instance, open returns the value of the block.

See .new for a description of the fd, mode and opt parameters.

IO.pipe(…) {|read_io, write_io| … }

Creates a pair of pipe endpoints (connected to each other) and returns them as a two-element array of IO objects: [ read_io, write_io ].

If a block is given, the block is called and returns the value of the block. read_io and write_io are sent to the block as arguments. If read_io and write_io are not closed when the block exits, they are closed. i.e. closing read_io and/or write_io doesn’t cause an error.

Not available on all platforms.

If an encoding (encoding name or encoding object) is specified as an optional argument, read string from pipe is tagged with the encoding specified. If the argument is a colon separated two encoding names “A:B”, the read string is converted from encoding A (external encoding) to encoding B (internal encoding), then tagged with B. If two optional arguments are specified, those must be encoding objects or encoding names, and the first one is the external encoding, and the second one is the internal encoding. If the external encoding and the internal encoding is specified, optional hash argument specify the conversion option.

In the example below, the two processes close the ends of the pipe that they are not using. This is not just a cosmetic nicety. The read end of a pipe will not generate an end of file condition if there are any writers with the pipe still open. In the case of the parent process, the rd.read will never return if it does not first issue a wr.close.

rd, wr = IO.pipe

if fork
  wr.close
  puts "Parent got: <#{rd.read}>"
  rd.close
  Process.wait
else
  rd.close
  puts "Sending message to parent"
  wr.write "Hi Dad"
  wr.close
end

produces:

Sending message to parent
Parent got: <Hi Dad>

Runs the specified command as a subprocess; the subprocess’s standard input and output will be connected to the returned IO object.

The PID of the started process can be obtained by #pid method.

cmd is a string or an array as follows.

cmd:
  "-"                                      : fork
  commandline                              : command line string which is passed to a shell
  [env, cmdname, arg1, ..., opts]          : command name and zero or more arguments (no shell)
  [env, [cmdname, argv0], arg1, ..., opts] : command name, argv[0] and zero or more arguments (no shell)
(env and opts are optional.)

If cmd is a ::String “-”, then a new instance of Ruby is started as the subprocess.

If cmd is an ::Array of ::String, then it will be used as the subprocess’s argv bypassing a shell. The array can contain a hash at first for environments and a hash at last for options similar to #spawn.

The default mode for the new file object is “r”, but mode may be set to any of the modes listed in the description for class IO. The last argument opt qualifies mode.

# set IO encoding
IO.popen("nkf -e filename", :external_encoding=>"EUC-JP") {|nkf_io|
  euc_jp_string = nkf_io.read
}

# merge standard output and standard error using
# spawn option.  See the document of Kernel.spawn.
IO.popen(["ls", "/", :err=>[:child, :out]]) {|ls_io|
  ls_result_with_error = ls_io.read
}

# spawn options can be mixed with IO options
IO.popen(["ls", "/"], :err=>[:child, :out]) {|ls_io|
  ls_result_with_error = ls_io.read
}

Raises exceptions which .pipe and Kernel.spawn raise.

If a block is given, Ruby will run the command as a child connected to Ruby with a pipe. Ruby’s end of the pipe will be passed as a parameter to the block. At the end of block, Ruby closes the pipe and sets $?. In this case popen returns the value of the block.

If a block is given with a cmd of “-”, the block will be run in two separate processes: once in the parent, and once in a child. The parent process will be passed the pipe object as a parameter to the block, the child version of the block will be passed nil, and the child’s standard in and standard out will be connected to the parent through the pipe. Not available on all platforms.

f = IO.popen("uname")
p f.readlines
f.close
puts "Parent is #{Process.pid}"
IO.popen("date") {|f| puts f.gets }
IO.popen("-") {|f| $stderr.puts "#{Process.pid} is here, f is #{f.inspect}"}
p $?
IO.popen(%w"sed -e s|^|<foo>| -e s&$&;zot;&", "r+") {|f|
  f.puts "bar"; f.close_write; puts f.gets
}

produces:

["Linux\n"]
Parent is 21346
Thu Jan 15 22:41:19 JST 2009
21346 is here, f is #<IO:fd 3>
21352 is here, f is nil
#<Process::Status: pid 21352 exit 0>
<foo>bar;zot;

Opens the file, optionally seeks to the given offset, then returns length bytes (defaulting to the rest of the file). #read ensures the file is closed before returning.

If name starts with a pipe character ("|"), a subprocess is created in the same way as Kernel.open, and its output is returned.

Options

The options hash accepts the following keys:

:encoding

string or encoding

Specifies the encoding of the read string. :encoding will be ignored if length is specified. See Encoding.aliases for possible encodings.

:mode

string or integer

Specifies the mode argument for open(). It must start with an “r”, otherwise it will cause an error. See IO.new for the list of possible modes.

:open_args

array

Specifies arguments for open() as an array. This key can not be used in combination with either :encoding or :mode.

Examples:

IO.read("testfile")              #=> "This is line one\nThis is line two\nThis is line three\nAnd so on...\n"
IO.read("testfile", 20)          #=> "This is line one\nThi"
IO.read("testfile", 20, 10)      #=> "ne one\nThis is line "
IO.read("binfile", mode: "rb")   #=> "\xF7\x00\x00\x0E\x12"

Reads the entire file specified by name as individual lines, and returns those lines in an array. Lines are separated by sep.

a = IO.readlines("testfile")
a[0]   #=> "This is line one\n"

b = IO.readlines("testfile", chomp: true)
b[0]   #=> "This is line one"

If the last argument is a hash, it’s the keyword argument to open.

Options for getline

The options hash accepts the following keys:

:chomp

When the optional chomp keyword argument has a true value, \n, \r, and \r\n will be removed from the end of each line.

See also .read for details about open_args.

Alias for Kernel.select. Calls select(2) system call. It monitors given arrays of IO objects, waits until one or more of IO objects are ready for reading, are ready for writing, and have pending exceptions respectively, and returns an array that contains arrays of those IO objects. It will return nil if optional timeout value is given and no IO object is ready in timeout seconds.

select peeks the buffer of IO objects for testing readability. If the IO buffer is not empty, select immediately notifies readability. This “peek” only happens for IO objects. It does not happen for IO-like objects such as OpenSSL::SSL::SSLSocket.

The best way to use select is invoking it after nonblocking methods such as #read_nonblock, #write_nonblock, etc. The methods raise an exception which is extended by ::IO::WaitReadable or ::IO::WaitWritable. The modules notify how the caller should wait with select. If ::IO::WaitReadable is raised, the caller should wait for reading. If ::IO::WaitWritable is raised, the caller should wait for writing.

So, blocking read (#readpartial) can be emulated using #read_nonblock and select as follows:

begin
  result = io_like.read_nonblock(maxlen)
rescue IO::WaitReadable
  IO.select([io_like])
  retry
rescue IO::WaitWritable
  IO.select(nil, [io_like])
  retry
end

Especially, the combination of nonblocking methods and select is preferred for IO like objects such as OpenSSL::SSL::SSLSocket. It has #to_io method to return underlying IO object. select calls #to_io to obtain the file descriptor to wait.

This means that readability notified by select doesn’t mean readability from OpenSSL::SSL::SSLSocket object.

The most likely situation is that OpenSSL::SSL::SSLSocket buffers some data. select doesn’t see the buffer. So select can block when OpenSSL::SSL::SSLSocket#readpartial doesn’t block.

However, several more complicated situations exist.

SSL is a protocol which is sequence of records. The record consists of multiple bytes. So, the remote side of SSL sends a partial record, select notifies readability but OpenSSL::SSL::SSLSocket cannot decrypt a byte and OpenSSL::SSL::SSLSocket#readpartial will block.

Also, the remote side can request SSL renegotiation which forces the local SSL engine to write some data. This means OpenSSL::SSL::SSLSocket#readpartial may invoke #write system call and it can block. In such a situation, OpenSSL::SSL::SSLSocket#read_nonblock raises ::IO::WaitWritable instead of blocking. So, the caller should wait for ready for writability as above example.

The combination of nonblocking methods and select is also useful for streams such as tty, pipe socket socket when multiple processes read from a stream.

Finally, Linux kernel developers don’t guarantee that readability of select(2) means readability of following read(2) even for a single process. See select(2) manual on GNU/Linux system.

Invoking select before #readpartial works well as usual. However it is not the best way to use select.

The writability notified by select(2) doesn’t show how many bytes are writable. #write method blocks until given whole string is written. So, IO#write(two or more bytes) can block after writability is notified by select. #write_nonblock is required to avoid the blocking.

Blocking write (#write) can be emulated using #write_nonblock and select as follows: ::IO::WaitReadable should also be rescued for SSL renegotiation in OpenSSL::SSL::SSLSocket.

while 0 < string.bytesize
  begin
    written = io_like.write_nonblock(string)
  rescue IO::WaitReadable
    IO.select([io_like])
    retry
  rescue IO::WaitWritable
    IO.select(nil, [io_like])
    retry
  end
  string = string.byteslice(written..-1)
end

Parameters

read_array

an array of IO objects that wait until ready for read

write_array

an array of IO objects that wait until ready for write

error_array

an array of IO objects that wait for exceptions

timeout

a numeric value in second

Example

rp, wp = IO.pipe
mesg = "ping "
100.times {
  # IO.select follows IO#read.  Not the best way to use IO.select.
  rs, ws, = IO.select([rp], [wp])
  if r = rs[0]
    ret = r.read(5)
    print ret
    case ret
    when /ping/
      mesg = "pong\n"
    when /pong/
      mesg = "ping "
    end
  end
  if w = ws[0]
    w.write(mesg)
  end
}

produces:

ping pong
ping pong
ping pong
(snipped)
ping

Opens the given path, returning the underlying file descriptor as a ::Integer.

IO.sysopen("testfile")   #=> 3

Try to convert obj into an IO, using to_io method. Returns converted IO or nil if obj cannot be converted for any reason.

IO.try_convert(STDOUT)     #=> STDOUT
IO.try_convert("STDOUT")   #=> nil

require 'zlib'
f = open("/tmp/zz.gz")       #=> #<File:/tmp/zz.gz>
z = Zlib::GzipReader.open(f) #=> #<Zlib::GzipReader:0x81d8744>
IO.try_convert(z)            #=> #<File:/tmp/zz.gz>

Opens the file, optionally seeks to the given offset, writes string, then returns the length written. #write ensures the file is closed before returning. If offset is not given in write mode, the file is truncated. Otherwise, it is not truncated.

IO.write("testfile", "0123456789", 20)  #=> 10
# File could contain:  "This is line one\nThi0123456789two\nThis is line three\nAnd so on...\n"
IO.write("testfile", "0123456789")      #=> 10
# File would now read: "0123456789"

If the last argument is a hash, it specifies options for the internal open(). It accepts the following keys:

:encoding

string or encoding

Specifies the encoding of the read string. See Encoding.aliases for possible encodings.

:mode

string or integer

Specifies the mode argument for open(). It must start with “w”, “a”, or “r+”, otherwise it will cause an error. See IO.new for the list of possible modes.

:perm

integer

Specifies the perm argument for open().

:open_args

array

Specifies arguments for open() as an array. This key can not be used in combination with other keys.

Returns true if the underlying file descriptor of ios will be closed automatically at its finalization, otherwise false.

Puts ios into binary mode. Once a stream is in binary mode, it cannot be reset to nonbinary mode.

• newline conversion disabled

• encoding conversion disabled

• content is treated as ASCII-8BIT

Returns true if ios is binmode.

Sets a close-on-exec flag.

f = open("/dev/null")
f.close_on_exec = true
system("cat", "/proc/self/fd/#{f.fileno}") # cat: /proc/self/fd/3: No such file or directory
f.closed?                #=> false

Ruby sets close-on-exec flags of all file descriptors by default since Ruby 2.0.0. So you don’t need to set by yourself. Also, unsetting a close-on-exec flag can cause file descriptor leak if another thread use fork() and exec() (via system() method for example). If you really needs file descriptor inheritance to child process, use spawn()‘s argument such as fd=>fd.

Returns true if ios will be closed on exec.

f = open("/dev/null")
f.close_on_exec?                 #=> false
f.close_on_exec = true
f.close_on_exec?                 #=> true
f.close_on_exec = false
f.close_on_exec?                 #=> false

Returns true if ios is completely closed (for duplex streams, both reader and writer), false otherwise.

f = File.new("testfile")
f.close         #=> nil
f.closed?       #=> true
f = IO.popen("/bin/sh","r+")
f.close_write   #=> nil
f.closed?       #=> false
f.close_read    #=> nil
f.closed?       #=> true

Returns true if ios is at end of file that means there are no more data to read. The stream must be opened for reading or an ::IOError will be raised.

f = File.new("testfile")
dummy = f.readlines
f.eof   #=> true

If ios is a stream such as pipe or socket, #eof? blocks until the other end sends some data or closes it.

r, w = IO.pipe
Thread.new { sleep 1; w.close }
r.eof?  #=> true after 1 second blocking

r, w = IO.pipe
Thread.new { sleep 1; w.puts "a" }
r.eof?  #=> false after 1 second blocking

r, w = IO.pipe
r.eof?  # blocks forever

Note that #eof? reads data to the input byte buffer. So #sysread may not behave as you intend with #eof?, unless you call #rewind first (which is not available for some streams).

Alias for #eof.

Alias for #tty?.

Returns the current line number in ios. The stream must be opened for reading. #lineno counts the number of times #gets is called rather than the number of newlines encountered. The two values will differ if #gets is called with a separator other than newline.

Methods that use $/ like #each, #lines and #readline will also increment #lineno.

See also the $. variable.

f = File.new("testfile")
f.lineno   #=> 0
f.gets     #=> "This is line one\n"
f.lineno   #=> 1
f.gets     #=> "This is line two\n"
f.lineno   #=> 2

Manually sets the current line number to the given value. $. is updated only on the next read.

f = File.new("testfile")
f.gets                     #=> "This is line one\n"
$.                         #=> 1
f.lineno = 1000
f.lineno                   #=> 1000
$.                         #=> 1         # lineno of last read
f.gets                     #=> "This is line two\n"
$.                         #=> 1001      # lineno of last read

Returns the current offset (in bytes) of ios.

f = File.new("testfile")
f.pos    #=> 0
f.gets   #=> "This is line one\n"
f.pos    #=> 17

Seeks to the given position (in bytes) in ios. It is not guaranteed that seeking to the right position when ios is textmode.

f = File.new("testfile")
f.pos = 17
f.gets   #=> "This is line two\n"

Returns the current “sync mode” of ios. When sync mode is true, all output is immediately flushed to the underlying operating system and is not buffered by Ruby internally. See also #fsync.

f = File.new("testfile")
f.sync   #=> false

Sets the “sync mode” to true or false. When sync mode is true, all output is immediately flushed to the underlying operating system and is not buffered internally. Returns the new state. See also #fsync.

f = File.new("testfile")
f.sync = true

Alias for #pos.

Returns true if ios is associated with a terminal device (tty), false otherwise.

File.new("testfile").isatty   #=> false
File.new("/dev/tty").isatty   #=> true

Announce an intention to access data from the current file in a specific pattern. On platforms that do not support the posix_fadvise(2) system call, this method is a no-op.

advice is one of the following symbols:

:normal

No advice to give; the default assumption for an open file.

:sequential

The data will be accessed sequentially with lower offsets read before higher ones.

:random

The data will be accessed in random order.

:willneed

The data will be accessed in the near future.

:dontneed

The data will not be accessed in the near future.

:noreuse

The data will only be accessed once.

The semantics of a piece of advice are platform-dependent. See man 2 posix_fadvise for details.

“data” means the region of the current file that begins at offset and extends for len bytes. If len is 0, the region ends at the last byte of the file. By default, both offset and len are 0, meaning that the advice applies to the entire file.

If an error occurs, one of the following exceptions will be raised:

IOError

The IO stream is closed.

Errno::EBADF

The file descriptor of the current file is invalid.

Errno::EINVAL

An invalid value for advice was given.

Errno::ESPIPE

The file descriptor of the current file refers to a FIFO or pipe. (Linux raises Errno::EINVAL in this case).

TypeError

Either advice was not a Symbol, or one of the other arguments was not an Integer.

RangeError

One of the arguments given was too big/small.

This list is not exhaustive; other Errno

exceptions are also possible.

Closes ios and flushes any pending writes to the operating system. The stream is unavailable for any further data operations; an ::IOError is raised if such an attempt is made. I/O streams are automatically closed when they are claimed by the garbage collector.

If ios is opened by .popen, #close sets $?.

Calling this method on closed IO object is just ignored since Ruby 2.3.

Closes the read end of a duplex I/O stream (i.e., one that contains both a read and a write stream, such as a pipe). Will raise an ::IOError if the stream is not duplexed.

f = IO.popen("/bin/sh","r+")
f.close_read
f.readlines

produces:

prog.rb:3:in `readlines': not opened for reading (IOError)
 from prog.rb:3

Calling this method on closed IO object is just ignored since Ruby 2.3.

Closes the write end of a duplex I/O stream (i.e., one that contains both a read and a write stream, such as a pipe). Will raise an ::IOError if the stream is not duplexed.

f = IO.popen("/bin/sh","r+")
f.close_write
f.print "nowhere"

produces:

prog.rb:3:in `write': not opened for writing (IOError)
 from prog.rb:3:in `print'
 from prog.rb:3

Calling this method on closed IO object is just ignored since Ruby 2.3.

ios.each_line(sep=$/ [, getline_args]) {|line| block } -> ios

ios.each_line(limit [, getline_args])      {|line| block } -> ios
ios.each_line(sep, limit [, getline_args]) {|line| block } -> ios
ios.each_line(...)                        -> an_enumerator

Executes the block for every line in ios, where lines are separated by sep. ios must be opened for reading or an ::IOError will be raised.

If no block is given, an enumerator is returned instead.

f = File.new("testfile")
f.each {|line| puts "#{f.lineno}: #{line}" }

produces:

1: This is line one
2: This is line two
3: This is line three
4: And so on...

See .readlines for details about getline_args.

Calls the given block once for each byte (0..255) in ios, passing the byte as an argument. The stream must be opened for reading or an ::IOError will be raised.

If no block is given, an enumerator is returned instead.

f = File.new("testfile")
checksum = 0
f.each_byte {|x| checksum ^= x }   #=> #<File:testfile>
checksum                           #=> 12

Calls the given block once for each character in ios, passing the character as an argument. The stream must be opened for reading or an ::IOError will be raised.

If no block is given, an enumerator is returned instead.

f = File.new("testfile")
f.each_char {|c| print c, ' ' }   #=> #<File:testfile>

Passes the ::Integer ordinal of each character in ios, passing the codepoint as an argument. The stream must be opened for reading or an ::IOError will be raised.

If no block is given, an enumerator is returned instead.

Alias for #each.

Returns the ::Encoding object that represents the encoding of the file. If io is in write mode and no encoding is specified, returns nil.

Provides a mechanism for issuing low-level commands to control or query file-oriented I/O streams. Arguments and results are platform dependent. If arg is a number, its value is passed directly. If it is a string, it is interpreted as a binary sequence of bytes (Array#pack might be a useful way to build this string). On Unix platforms, see fcntl(2) for details. Not implemented on all platforms.

Immediately writes all buffered data in ios to disk.

If the underlying operating system does not support fdatasync(2), #fsync is called instead (which might raise a ::NotImplementedError).

Alias for #to_i.

Flushes any buffered data within ios to the underlying operating system (note that this is Ruby internal buffering only; the OS may buffer the data as well).

$stdout.print "no newline"
$stdout.flush

produces:

no newline

Immediately writes all buffered data in ios to disk. Note that #fsync differs from using #sync=. The latter ensures that data is flushed from Ruby’s buffers, but does not guarantee that the underlying operating system actually writes it to disk.

::NotImplementedError is raised if the underlying operating system does not support fsync(2).

Gets the next 8-bit byte (0..255) from ios. Returns nil if called at end of file.

f = File.new("testfile")
f.getbyte   #=> 84
f.getbyte   #=> 104

Reads a one-character string from ios. Returns nil if called at end of file.

f = File.new("testfile")
f.getc   #=> "h"
f.getc   #=> "e"

Reads the next “line” from the I/O stream; lines are separated by sep. A separator of nil reads the entire contents, and a zero-length separator reads the input a paragraph at a time (two successive newlines in the input separate paragraphs). The stream must be opened for reading or an ::IOError will be raised. The line read in will be returned and also assigned to $_. Returns nil if called at end of file. If the first argument is an integer, or optional second argument is given, the returning string would not be longer than the given value in bytes.

File.new("testfile").gets   #=> "This is line one\n"
$_                          #=> "This is line one\n"

File.new("testfile").gets(4)#=> "This"

If IO contains multibyte characters byte then gets(1) returns character entirely:

# Russian characters take 2 bytes
File.write("testfile", "\u{442 435 441 442}")
File.open("testfile") {|f|f.gets(1)} #=> "\u0442"
File.open("testfile") {|f|f.gets(2)} #=> "\u0442"
File.open("testfile") {|f|f.gets(3)} #=> "\u0442\u0435"
File.open("testfile") {|f|f.gets(4)} #=> "\u0442\u0435"

Return a string describing this IO object.

Returns the ::Encoding of the internal string if conversion is specified. Otherwise returns nil.

Provides a mechanism for issuing low-level commands to control or query I/O devices. Arguments and results are platform dependent. If arg is a number, its value is passed directly. If it is a string, it is interpreted as a binary sequence of bytes. On Unix platforms, see ioctl(2) for details. Not implemented on all platforms.

Returns the process ID of a child process associated with ios. This will be set by .popen.

pipe = IO.popen("-")
if pipe
  $stderr.puts "In parent, child pid is #{pipe.pid}"
else
  $stderr.puts "In child, pid is #{$$}"
end

produces:

In child, pid is 26209
In parent, child pid is 26209

Reads maxlen bytes from ios using the pread system call and returns them as a string without modifying the underlying descriptor offset. This is advantageous compared to combining #seek and #read in that it is atomic, allowing multiple threads/process to share the same IO object for reading the file at various locations. This bypasses any userspace buffering of the IO layer. If the optional outbuf argument is present, it must reference a ::String, which will receive the data. Raises SystemCallError on error, ::EOFError at end of file and ::NotImplementedError if platform does not implement the system call.

File.write("testfile", "This is line one\nThis is line two\n")
File.open("testfile") do |f|
  p f.read           # => "This is line one\nThis is line two\n"
  p f.pread(12, 0)   # => "This is line"
  p f.pread(9, 8)    # => "line one\n"
end

Writes the given object(s) to ios. Returns nil.

The stream must be opened for writing. Each given object that isn’t a string will be converted by calling its to_s method. When called without arguments, prints the contents of $_.

If the output field separator ($,) is not nil, it is inserted between objects. If the output record separator ($\) is not nil, it is appended to the output.

$stdout.print("This is ", 100, " percent.\n")

produces:

This is 100 percent.

Formats and writes to ios, converting parameters under control of the format string. See Kernel.sprintf for details.

If obj is ::Numeric, write the character whose code is the least-significant byte of obj. If obj is ::String, write the first character of obj to ios. Otherwise, raise ::TypeError.

$stdout.putc "A"
$stdout.putc 65

produces:

AA

Writes the given object(s) to ios. Writes a newline after any that do not already end with a newline sequence. Returns nil.

The stream must be opened for writing. If called with an array argument, writes each element on a new line. Each given object that isn’t a string or array will be converted by calling its to_s method. If called without arguments, outputs a single newline.

$stdout.puts("this", "is", ["a", "test"])

produces:

this
is
a
test

Note that puts always uses newlines and is not affected by the output record separator ($\).

Writes the given string to ios at offset using pwrite() system call. This is advantageous to combining #seek and #write in that it is atomic, allowing multiple threads/process to share the same IO object for reading the file at various locations. This bypasses any userspace buffering of the IO layer. Returns the number of bytes written. Raises SystemCallError on error and ::NotImplementedError if platform does not implement the system call.

File.open("out", "w") do |f|
  f.pwrite("ABCDEF", 3)   #=> 6
end

File.read("out")          #=> "\u0000\u0000\u0000ABCDEF"

Reads length bytes from the I/O stream.

length must be a non-negative integer or nil.

If length is a positive integer, read tries to read length bytes without any conversion (binary mode). It returns nil if an EOF is encountered before anything can be read. Fewer than length bytes are returned if an EOF is encountered during the read. In the case of an integer length, the resulting string is always in ASCII-8BIT encoding.

If length is omitted or is nil, it reads until EOF and the encoding conversion is applied, if applicable. A string is returned even if EOF is encountered before any data is read.

If length is zero, it returns an empty string ("").

If the optional outbuf argument is present, it must reference a ::String, which will receive the data. The outbuf will contain only the received data after the method call even if it is not empty at the beginning.

When this method is called at end of file, it returns nil or "", depending on length: read, read(nil), and read(0) return "", read(positive_integer) returns nil.

f = File.new("testfile")
f.read(16)   #=> "This is line one"

# read whole file
open("file") do |f|
  data = f.read   # This returns a string even if the file is empty.
  # ...
end

# iterate over fixed length records
open("fixed-record-file") do |f|
  while record = f.read(256)
    # ...
  end
end

# iterate over variable length records,
# each record is prefixed by its 32-bit length
open("variable-record-file") do |f|
  while len = f.read(4)
    len = len.unpack("N")[0]   # 32-bit length
    record = f.read(len)       # This returns a string even if len is 0.
  end
end

Note that this method behaves like the fread() function in C. This means it retries to invoke read(2) system calls to read data with the specified length (or until EOF). This behavior is preserved even if ios is in non-blocking mode. (This method is non-blocking flag insensitive as other methods.) If you need the behavior like a single read(2) system call, consider #readpartial, #read_nonblock, and #sysread.

Reads at most maxlen bytes from ios using the read(2) system call after O_NONBLOCK is set for the underlying file descriptor.

If the optional outbuf argument is present, it must reference a ::String, which will receive the data. The outbuf will contain only the received data after the method call even if it is not empty at the beginning.

read_nonblock just calls the read(2) system call. It causes all errors the read(2) system call causes: Errno::EWOULDBLOCK, Errno::EINTR, etc. The caller should care such errors.

If the exception is Errno::EWOULDBLOCK or Errno::EAGAIN, it is extended by ::IO::WaitReadable. So ::IO::WaitReadable can be used to rescue the exceptions for retrying read_nonblock.

read_nonblock causes ::EOFError on EOF.

On some platforms, such as Windows, non-blocking mode is not supported on IO objects other than sockets. In such cases, Errno::EBADF will be raised.

If the read byte buffer is not empty, read_nonblock reads from the buffer like readpartial. In this case, the read(2) system call is not called.

When read_nonblock raises an exception kind of ::IO::WaitReadable, read_nonblock should not be called until io is readable for avoiding busy loop. This can be done as follows.

# emulates blocking read (readpartial).
begin
  result = io.read_nonblock(maxlen)
rescue IO::WaitReadable
  IO.select([io])
  retry
end

Although read_nonblock doesn’t raise ::IO::WaitWritable. OpenSSL::Buffering#read_nonblock can raise ::IO::WaitWritable. If IO and SSL should be used polymorphically, ::IO::WaitWritable should be rescued too. See the document of OpenSSL::Buffering#read_nonblock for sample code.

Note that this method is identical to readpartial except the non-blocking flag is set.

By specifying a keyword argument exception to false, you can indicate that read_nonblock should not raise an ::IO::WaitReadable exception, but return the symbol :wait_readable instead. At EOF, it will return nil instead of raising ::EOFError.

Reads a byte as with #getbyte, but raises an ::EOFError on end of file.

Reads a one-character string from ios. Raises an ::EOFError on end of file.

f = File.new("testfile")
f.readchar   #=> "h"
f.readchar   #=> "e"

Reads a line as with #gets, but raises an ::EOFError on end of file.

Reads all of the lines in ios, and returns them in an array. Lines are separated by the optional sep. If sep is nil, the rest of the stream is returned as a single record. If the first argument is an integer, or an optional second argument is given, the returning string would not be longer than the given value in bytes. The stream must be opened for reading or an ::IOError will be raised.

f = File.new("testfile")
f.readlines[0]   #=> "This is line one\n"

f = File.new("testfile", chomp: true)
f.readlines[0]   #=> "This is line one"

See .readlines for details about getline_args.

Reads at most maxlen bytes from the I/O stream. It blocks only if ios has no data immediately available. It doesn’t block if some data available.

If the optional outbuf argument is present, it must reference a ::String, which will receive the data. The outbuf will contain only the received data after the method call even if it is not empty at the beginning.

It raises ::EOFError on end of file.

readpartial is designed for streams such as pipe, socket, tty, etc. It blocks only when no data immediately available. This means that it blocks only when following all conditions hold.

• the byte buffer in the IO object is empty.

• the content of the stream is empty.

• the stream is not reached to EOF.

When readpartial blocks, it waits data or EOF on the stream. If some data is reached, readpartial returns with the data. If EOF is reached, readpartial raises ::EOFError.

When readpartial doesn’t blocks, it returns or raises immediately. If the byte buffer is not empty, it returns the data in the buffer. Otherwise if the stream has some content, it returns the data in the stream. Otherwise if the stream is reached to EOF, it raises ::EOFError.

r, w = IO.pipe           #               buffer          pipe content
w << "abc"               #               ""              "abc".
r.readpartial(4096)      #=> "abc"       ""              ""
r.readpartial(4096)      # blocks because buffer and pipe is empty.

r, w = IO.pipe           #               buffer          pipe content
w << "abc"               #               ""              "abc"
w.close                  #               ""              "abc" EOF
r.readpartial(4096)      #=> "abc"       ""              EOF
r.readpartial(4096)      # raises EOFError

r, w = IO.pipe           #               buffer          pipe content
w << "abc\ndef\n"        #               ""              "abc\ndef\n"
r.gets                   #=> "abc\n"     "def\n"         ""
w << "ghi\n"             #               "def\n"         "ghi\n"
r.readpartial(4096)      #=> "def\n"     ""              "ghi\n"
r.readpartial(4096)      #=> "ghi\n"     ""              ""

Note that readpartial behaves similar to sysread. The differences are:

• If the byte buffer is not empty, read from the byte buffer instead of “sysread for buffered IO (IOError)”.

• It doesn’t cause Errno::EWOULDBLOCK and Errno::EINTR. When readpartial meets EWOULDBLOCK and EINTR by read system call, readpartial retry the system call.

The latter means that readpartial is nonblocking-flag insensitive. It blocks on the situation #sysread causes Errno::EWOULDBLOCK as if the fd is blocking mode.

Reassociates ios with the I/O stream given in other_IO or to a new stream opened on path. This may dynamically change the actual class of this stream. The mode and opt parameters accept the same values as .open.

f1 = File.new("testfile")
f2 = File.new("testfile")
f2.readlines[0]   #=> "This is line one\n"
f2.reopen(f1)     #=> #<File:testfile>
f2.readlines[0]   #=> "This is line one\n"

Positions ios to the beginning of input, resetting #lineno to zero.

f = File.new("testfile")
f.readline   #=> "This is line one\n"
f.rewind     #=> 0
f.lineno     #=> 0
f.readline   #=> "This is line one\n"

Note that it cannot be used with streams such as pipes, ttys, and sockets.

Seeks to a given offset anInteger in the stream according to the value of whence:

:CUR or IO::SEEK_CUR  | Seeks to _amount_ plus current position
----------------------+--------------------------------------------------
:END or IO::SEEK_END  | Seeks to _amount_ plus end of stream (you
                      | probably want a negative value for _amount_)
----------------------+--------------------------------------------------
:SET or IO::SEEK_SET  | Seeks to the absolute location given by _amount_

Example:

f = File.new("testfile")
f.seek(-13, IO::SEEK_END)   #=> 0
f.readline                  #=> "And so on...\n"

If single argument is specified, read string from io is tagged with the encoding specified. If encoding is a colon separated two encoding names “A:B”, the read string is converted from encoding A (external encoding) to encoding B (internal encoding), then tagged with B. If two arguments are specified, those must be encoding objects or encoding names, and the first one is the external encoding, and the second one is the internal encoding. If the external encoding and the internal encoding is specified, optional hash argument specify the conversion option.

Checks if ios starts with a BOM, and then consumes it and sets the external encoding. Returns the result encoding if found, or nil. If ios is not binmode or its encoding has been set already, an exception will be raised.

File.write("bom.txt", "\u{FEFF}abc")
ios = File.open("bom.txt", "rb")
ios.set_encoding_by_bom    #=>  #<Encoding:UTF-8>

File.write("nobom.txt", "abc")
ios = File.open("nobom.txt", "rb")
ios.set_encoding_by_bom    #=>  nil

Returns status information for ios as an object of type ::File::Stat.

f = File.new("testfile")
s = f.stat
"%o" % s.mode   #=> "100644"
s.blksize       #=> 4096
s.atime         #=> Wed Apr 09 08:53:54 CDT 2003

Reads maxlen bytes from ios using a low-level read and returns them as a string. Do not mix with other methods that read from ios or you may get unpredictable results.

If the optional outbuf argument is present, it must reference a ::String, which will receive the data. The outbuf will contain only the received data after the method call even if it is not empty at the beginning.

Raises SystemCallError on error and ::EOFError at end of file.

f = File.new("testfile")
f.sysread(16)   #=> "This is line one"

Seeks to a given offset in the stream according to the value of whence (see #seek for values of whence). Returns the new offset into the file.

f = File.new("testfile")
f.sysseek(-13, IO::SEEK_END)   #=> 53
f.sysread(10)                  #=> "And so on."

Writes the given string to ios using a low-level write. Returns the number of bytes written. Do not mix with other methods that write to ios or you may get unpredictable results. Raises SystemCallError on error.

f = File.new("out", "w")
f.syswrite("ABCDEF")   #=> 6

Returns an integer representing the numeric file descriptor for ios.

$stdin.fileno    #=> 0
$stdout.fileno   #=> 1

Returns ios.

Pushes back bytes (passed as a parameter) onto ios, such that a subsequent buffered read will return it. Only one byte may be pushed back before a subsequent read operation (that is, you will be able to read only the last of several bytes that have been pushed back). Has no effect with unbuffered reads (such as #sysread).

f = File.new("testfile")   #=> #<File:testfile>
b = f.getbyte              #=> 0x38
f.ungetbyte(b)             #=> nil
f.getbyte                  #=> 0x38

Pushes back one character (passed as a parameter) onto ios, such that a subsequent buffered character read will return it. Only one character may be pushed back before a subsequent read operation (that is, you will be able to read only the last of several characters that have been pushed back). Has no effect with unbuffered reads (such as #sysread).

f = File.new("testfile")   #=> #<File:testfile>
c = f.getc                 #=> "8"
f.ungetc(c)                #=> nil
f.getc                     #=> "8"

Writes the given strings to ios. The stream must be opened for writing. Arguments that are not a string will be converted to a string using to_s. Returns the number of bytes written in total.

count = $stdout.write("This is", " a test\n")
puts "That was #{count} bytes of data"

produces:

This is a test
That was 15 bytes of data

Writes the given string to ios using the write(2) system call after O_NONBLOCK is set for the underlying file descriptor.

It returns the number of bytes written.

write_nonblock just calls the write(2) system call. It causes all errors the write(2) system call causes: Errno::EWOULDBLOCK, Errno::EINTR, etc. The result may also be smaller than string.length (partial write). The caller should care such errors and partial write.

If the exception is Errno::EWOULDBLOCK or Errno::EAGAIN, it is extended by ::IO::WaitWritable. So ::IO::WaitWritable can be used to rescue the exceptions for retrying write_nonblock.

# Creates a pipe.
r, w = IO.pipe

# write_nonblock writes only 65536 bytes and return 65536.
# (The pipe size is 65536 bytes on this environment.)
s = "a" * 100000
p w.write_nonblock(s)     #=> 65536

# write_nonblock cannot write a byte and raise EWOULDBLOCK (EAGAIN).
p w.write_nonblock("b")   # Resource temporarily unavailable (Errno::EAGAIN)

If the write buffer is not empty, it is flushed at first.

When write_nonblock raises an exception kind of ::IO::WaitWritable, write_nonblock should not be called until io is writable for avoiding busy loop. This can be done as follows.

begin
  result = io.write_nonblock(string)
rescue IO::WaitWritable, Errno::EINTR
  IO.select(nil, [io])
  retry
end

Note that this doesn’t guarantee to write all data in string. The length written is reported as result and it should be checked later.

On some platforms such as Windows, write_nonblock is not supported according to the kind of the IO object. In such cases, write_nonblock raises Errno::EBADF.

By specifying a keyword argument exception to false, you can indicate that write_nonblock should not raise an ::IO::WaitWritable exception, but return the symbol :wait_writable instead.