path: root/doc/sed-in.texi

\input texinfo  @c -*-texinfo-*-
@c
@c -- Stuff that needs adding: ----------------------------------------------
@c (nothing!)
@c --------------------------------------------------------------------------
@c Check for consistency: regexps in @code, text that they match in @samp.
@c 
@c Tips:
@c    @command for command
@c    @samp for command fragments: @samp{cat -s}
@c    @code for sed commands and flags
@c    Use ``quote'' not `quote' or "quote".
@c
@c %**start of header
@setfilename sed.info
@settitle sed, a stream editor
@c %**end of header

@c @smallbook

@include version.texi

@c Combine indices.
@syncodeindex ky cp
@syncodeindex pg cp
@syncodeindex tp cp

@defcodeindex op
@syncodeindex op fn

@include config.texi

@copying
This file documents version @value{VERSION} of
@value{SSED}, a stream editor.

Copyright @copyright{} 1998, 1999, 2001, 2002, 2003, 2004 Free
Software Foundation, Inc.

This document is released under the terms of the @acronym{GNU} Free
Documentation License as published by the Free Software Foundation;
either version 1.1, or (at your option) any later version.

You should have received a copy of the @acronym{GNU} Free Documentation
License along with @value{SSED}; see the file @file{COPYING.DOC}.
If not, write to the Free Software Foundation, 59 Temple Place - Suite
330, Boston, MA 02110-1301, USA.

There are no Cover Texts and no Invariant Sections; this text, along
with its equivalent in the printed manual, constitutes the Title Page.
@end copying

@setchapternewpage off

@titlepage
@title @command{sed}, a stream editor
@subtitle version @value{VERSION}, @value{UPDATED}
@author by Ken Pizzini, Paolo Bonzini

@page
@vskip 0pt plus 1filll
Copyright @copyright{} 1998, 1999 Free Software Foundation, Inc.

@insertcopying

Published by the Free Software Foundation, @*
51 Franklin Street, Fifth Floor @*
Boston, MA 02110-1301, USA
@end titlepage


@node Top
@top

@ifnottex
@insertcopying
@end ifnottex

@menu
* Introduction::               Introduction
* Invoking sed::               Invocation
* sed Programs::               @command{sed} programs
* Examples::                   Some sample scripts
* Limitations::                Limitations and (non-)limitations of @value{SSED}
* Other Resources::            Other resources for learning about @command{sed}
* Reporting Bugs::             Reporting bugs

* Extended regexps::           @command{egrep}-style regular expressions
@ifset PERL
* Perl regexps::               Perl-style regular expressions
@end ifset

* Concept Index::              A menu with all the topics in this manual.
* Command and Option Index::   A menu with all @command{sed} commands and
                               command-line options.

@detailmenu
--- The detailed node listing ---

sed Programs:
* Execution Cycle::                 How @command{sed} works
* Addresses::                       Selecting lines with @command{sed}
* Regular Expressions::             Overview of regular expression syntax
* Common Commands::                 Often used commands
* The "s" Command::                 @command{sed}'s Swiss Army Knife
* Other Commands::                  Less frequently used commands
* Programming Commands::            Commands for @command{sed} gurus
* Extended Commands::               Commands specific of @value{SSED}
* Escapes::                         Specifying special characters

Examples:
* Centering lines::
* Increment a number::
* Rename files to lower case::
* Print bash environment::
* Reverse chars of lines::
* tac::                             Reverse lines of files
* cat -n::                          Numbering lines
* cat -b::                          Numbering non-blank lines
* wc -c::                           Counting chars
* wc -w::                           Counting words
* wc -l::                           Counting lines
* head::                            Printing the first lines
* tail::                            Printing the last lines
* uniq::                            Make duplicate lines unique
* uniq -d::                         Print duplicated lines of input
* uniq -u::                         Remove all duplicated lines
* cat -s::                          Squeezing blank lines

@ifset PERL
Perl regexps::                      Perl-style regular expressions
* Backslash::                       Introduces special sequences
* Circumflex/dollar sign/period::   Behave specially with regard to new lines
* Square brackets::                 Are a bit different in strange cases
* Options setting::                 Toggle modifiers in the middle of a regexp
* Non-capturing subpatterns::       Are not counted when backreferencing
* Repetition::                      Allows for non-greedy matching
* Backreferences::                  Allows for more than 10 back references
* Assertions::                      Allows for complex look ahead matches
* Non-backtracking subpatterns::    Often gives more performance
* Conditional subpatterns::         Allows if/then/else branches
* Recursive patterns::              For example to match parentheses
* Comments::                        Because things can get complex...
@end ifset

@end detailmenu
@end menu


@node Introduction
@chapter Introduction

@cindex Stream editor
@command{sed} is a stream editor.
A stream editor is used to perform basic text
transformations on an input stream
(a file or input from a pipeline).
While in some ways similar to an editor which
permits scripted edits (such as @command{ed}),
@command{sed} works by making only one pass over the
input(s), and is consequently more efficient.
But it is @command{sed}'s ability to filter text in a pipeline
which particularly distinguishes it from other types of
editors.


@node Invoking sed
@chapter Invocation

Normally @command{sed} is invoked like this:

@example
sed SCRIPT INPUTFILE...
@end example

The full format for invoking @command{sed} is:

@example
sed OPTIONS... [SCRIPT] [INPUTFILE...]
@end example

If you do not specify @var{INPUTFILE}, or if @var{INPUTFILE} is @file{-},
@command{sed} filters the contents of the standard input.  The @var{script}
is actually the first non-option parameter, which @command{sed} specially
considers a script and not an input file if (and only if) none of the
other @var{options} specifies a script to be executed, that is if neither
of the @option{-e} and @option{-f} options is specified.

@command{sed} may be invoked with the following command-line options:

@table @code
@item --version
@opindex --version
@cindex Version, printing
Print out the version of @command{sed} that is being run and a copyright notice,
then exit.

@item --help
@opindex --help
@cindex Usage summary, printing
Print a usage message briefly summarizing these command-line options
and the bug-reporting address,
then exit.

@item -n
@itemx --quiet
@itemx --silent
@opindex -n
@opindex --quiet
@opindex --silent
@cindex Disabling autoprint, from command line
By default, @command{sed} prints out the pattern space
at the end of each cycle through the script (@pxref{Execution Cycle, ,
How @code{sed} works}).
These options disable this automatic printing,
and @command{sed} only produces output when explicitly told to
via the @code{p} command.

@item -e @var{script}
@itemx --expression=@var{script}
@opindex -e
@opindex --expression
@cindex Script, from command line
Add the commands in @var{script} to the set of commands to be
run while processing the input.

@item -f @var{script-file}
@itemx --file=@var{script-file}
@opindex -f
@opindex --file
@cindex Script, from a file
Add the commands contained in the file @var{script-file}
to the set of commands to be run while processing the input.

@item -i[@var{SUFFIX}]
@itemx --in-place[=@var{SUFFIX}]
@opindex -i
@opindex --in-place
@cindex In-place editing, activating
@cindex @value{SSEDEXT}, in-place editing
This option specifies that files are to be edited in-place.
@value{SSED} does this by creating a temporary file and
sending output to this file rather than to the standard
output.@footnote{This applies to commands such as @code{=},
@code{a}, @code{c}, @code{i}, @code{l}, @code{p}.  You can
still write to the standard output by using the @code{w}
@cindex @value{SSEDEXT}, @file{/dev/stdout} file
or @code{W} commands together with the @file{/dev/stdout}
special file}.

This option implies @option{-s}.

When the end of the file is reached, the temporary file is
renamed to the output file's original name.  The extension,
if supplied, is used to modify the name of the old file
before renaming the temporary file, thereby making a backup
copy@footnote{Note that @value{SSED} creates the backup
file whether or not any output is actually changed.}).

@cindex In-place editing, Perl-style backup file names
This rule is followed: if the extension doesn't contain a @code{*},
then it is appended to the end of the current filename as a
suffix; if the extension does contain one or more @code{*}
characters, then @emph{each} asterisk is replaced with the
current filename.  This allows you to add a prefix to the
backup file, instead of (or in addition to) a suffix, or
even to place backup copies of the original files into another
directory (provided the directory already exists).

If no extension is supplied, the original file is
overwritten without making a backup.

@item -l @var{N}
@itemx --line-length=@var{N}
@opindex -l
@opindex --line-length
@cindex Line length, setting
Specify the default line-wrap length for the @code{l} command.
A length of 0 (zero) means to never wrap long lines.  If
not specified, it is taken to be 70.

@item --posix
@opindex --posix
@cindex @value{SSEDEXT}, disabling
@value{SSED} includes several extensions to @acronym{POSIX}
sed.  In order to simplify writing portable scripts, this
option disables all the extensions that this manual documents,
including additional commands.
@cindex @code{POSIXLY_CORRECT} behavior, enabling
Most of the extensions accept @command{sed} programs that
are outside the syntax mandated by @acronym{POSIX}, but some
of them (such as the behavior of the @command{N} command
described in @pxref{Reporting Bugs}) actually violate the
standard.  If you want to disable only the latter kind of
extension, you can set the @code{POSIXLY_CORRECT} variable
to a non-empty value.

@item -b
@itemx --binary
@opindex -b
@opindex --binary
This option is available on every platform, but is only effective where the
operating system makes a distinction between text files and binary files.
When such a distinction is made---as is the case for MS-DOS, Windows,
Cygwin---text files are composed of lines separated by a carriage return
@emph{and} a line feed character, and @command{sed} does not see the
ending CR.  When this option is specified, @command{sed} will open
input files in binary mode, thus not requesting this special processing
and considering lines to end at a line feed.

@item --follow-symlinks
@opindex --follow-symlinks
This option is available only on platforms that support
symbolic links and has an effect only if option @option{-i}
is specified.  In this case, if the file that is specified
on the command line is a symbolic link, @command{sed} will
follow the link and edit the ultimate destination of the
link.  The default behavior is to break the symbolic link,
so that the link destination will not be modified.

@item -r
@itemx --regexp-extended
@opindex -r
@opindex --regexp-extended
@cindex Extended regular expressions, choosing
@cindex @acronym{GNU} extensions, extended regular expressions
Use extended regular expressions rather than basic
regular expressions.  Extended regexps are those that
@command{egrep} accepts; they can be clearer because they
usually have less backslashes, but are a @acronym{GNU} extension
and hence scripts that use them are not portable.
@xref{Extended regexps, , Extended regular expressions}.

@ifset PERL
@item -R
@itemx --regexp-perl
@opindex -R
@opindex --regexp-perl
@cindex Perl-style regular expressions, choosing
@cindex @value{SSEDEXT}, Perl-style regular expressions
Use Perl-style regular expressions rather than basic
regular expressions.  Perl-style regexps are extremely
powerful but are a @value{SSED} extension and hence scripts that
use it are not portable.  @xref{Perl regexps, ,
Perl-style regular expressions}.
@end ifset

@item -s
@itemx --separate
@opindex -s
@opindex --separate
@cindex Working on separate files
By default, @command{sed} will consider the files specified on the
command line as a single continuous long stream.  This @value{SSED}
extension allows the user to consider them as separate files:
range addresses (such as @samp{/abc/,/def/}) are not allowed
to span several files, line numbers are relative to the start
of each file, @code{$} refers to the last line of each file,
and files invoked from the @code{R} commands are rewound at the
start of each file.

@item -u
@itemx --unbuffered
@opindex -u
@opindex --unbuffered
@cindex Unbuffered I/O, choosing
Buffer both input and output as minimally as practical.
(This is particularly useful if the input is coming from
the likes of @samp{tail -f}, and you wish to see the transformed
output as soon as possible.)

@item -z
@itemx --null-data
@itemx --zero-terminated
@opindex -z
@opindex --null-data
@opindex --zero-terminated
Treat the input as a set of lines, each terminated by a zero byte
(the ASCII @samp{NUL} character) instead of a newline.  This option can
be used with commands like @samp{sort -z} and @samp{find -print0}
to process arbitrary file names.
@end table

If no @option{-e}, @option{-f}, @option{--expression}, or @option{--file}
options are given on the command-line,
then the first non-option argument on the command line is
taken to be the @var{script} to be executed.

@cindex Files to be processed as input
If any command-line parameters remain after processing the above,
these parameters are interpreted as the names of input files to
be processed.
@cindex Standard input, processing as input
A file name of @samp{-} refers to the standard input stream.
The standard input will be processed if no file names are specified.


@node sed Programs
@chapter @command{sed} Programs

@cindex @command{sed} program structure
@cindex Script structure
A @command{sed} program consists of one or more @command{sed} commands,
passed in by one or more of the
@option{-e}, @option{-f}, @option{--expression}, and @option{--file}
options, or the first non-option argument if zero of these
options are used.
This document will refer to ``the'' @command{sed} script;
this is understood to mean the in-order catenation
of all of the @var{script}s and @var{script-file}s passed in.

Commands within a @var{script} or @var{script-file} can be
separated by semicolons (@code{;}) or newlines (ASCII 10).
Some commands, due to their syntax, cannot be followed by semicolons
working as command separators and thus should be terminated
with newlines or be placed at the end of a @var{script} or @var{script-file}.
Commands can also be preceded with optional non-significant
whitespace characters.

Each @code{sed} command consists of an optional address or
address range, followed by a one-character command name
and any additional command-specific code.

@menu
* Execution Cycle::          How @command{sed} works
* Addresses::                Selecting lines with @command{sed}
* Regular Expressions::      Overview of regular expression syntax
* Common Commands::          Often used commands
* The "s" Command::          @command{sed}'s Swiss Army Knife
* Other Commands::           Less frequently used commands
* Programming Commands::     Commands for @command{sed} gurus
* Extended Commands::        Commands specific of @value{SSED}
* Escapes::                  Specifying special characters
@end menu


@node Execution Cycle
@section How @command{sed} Works

@cindex Buffer spaces, pattern and hold
@cindex Spaces, pattern and hold
@cindex Pattern space, definition
@cindex Hold space, definition
@command{sed} maintains two data buffers: the active @emph{pattern} space,
and the auxiliary @emph{hold} space. Both are initially empty.

@command{sed} operates by performing the following cycle on each
line of input: first, @command{sed} reads one line from the input
stream, removes any trailing newline, and places it in the pattern space.
Then commands are executed; each command can have an address associated
to it: addresses are a kind of condition code, and a command is only
executed if the condition is verified before the command is to be
executed.

When the end of the script is reached, unless the @option{-n} option
is in use, the contents of pattern space are printed out to the output
stream, adding back the trailing newline if it was removed.@footnote{Actually,
if @command{sed} prints a line without the terminating newline, it will
nevertheless print the missing newline as soon as more text is sent to
the same output stream, which gives the ``least expected surprise''
even though it does not make commands like @samp{sed -n p} exactly
identical to @command{cat}.} Then the next cycle starts for the next
input line.

Unless special commands (like @samp{D}) are used, the pattern space is
deleted between two cycles. The hold space, on the other hand, keeps
its data between cycles (see commands @samp{h}, @samp{H}, @samp{x},
@samp{g}, @samp{G} to move data between both buffers).


@node Addresses
@section Selecting lines with @command{sed}
@cindex Addresses, in @command{sed} scripts
@cindex Line selection
@cindex Selecting lines to process

Addresses in a @command{sed} script can be in any of the following forms:
@table @code
@item @var{number}
@cindex Address, numeric
@cindex Line, selecting by number
Specifying a line number will match only that line in the input.
(Note that @command{sed} counts lines continuously across all input files
unless @option{-i} or @option{-s} options are specified.)

@item @var{first}~@var{step}
@cindex @acronym{GNU} extensions, @samp{@var{n}~@var{m}} addresses
This @acronym{GNU} extension matches every @var{step}th line
starting with line @var{first}.
In particular, lines will be selected when there exists
a non-negative @var{n} such that the current line-number equals
@var{first} + (@var{n} * @var{step}).
Thus, to select the odd-numbered lines,
one would use @code{1~2};
to pick every third line starting with the second, @samp{2~3} would be used;
to pick every fifth line starting with the tenth, use @samp{10~5};
and @samp{50~0} is just an obscure way of saying @code{50}.

@item $
@cindex Address, last line
@cindex Last line, selecting
@cindex Line, selecting last
This address matches the last line of the last file of input, or
the last line of each file when the @option{-i} or @option{-s} options
are specified.

@item /@var{regexp}/
@cindex Address, as a regular expression
@cindex Line, selecting by regular expression match
This will select any line which matches the regular expression @var{regexp}.
If @var{regexp} itself includes any @code{/} characters,
each must be escaped by a backslash (@code{\}).

@cindex empty regular expression
@cindex @value{SSEDEXT}, modifiers and the empty regular expression
The empty regular expression @samp{//} repeats the last regular
expression match (the same holds if the empty regular expression is
passed to the @code{s} command).  Note that modifiers to regular expressions
are evaluated when the regular expression is compiled, thus it is invalid to
specify them together with the empty regular expression.

@item \%@var{regexp}%
(The @code{%} may be replaced by any other single character.)

@cindex Slash character, in regular expressions
This also matches the regular expression @var{regexp},
but allows one to use a different delimiter than @code{/}.
This is particularly useful if the @var{regexp} itself contains
a lot of slashes, since it avoids the tedious escaping of every @code{/}.
If @var{regexp} itself includes any delimiter characters,
each must be escaped by a backslash (@code{\}).

@item /@var{regexp}/I
@itemx \%@var{regexp}%I
@cindex @acronym{GNU} extensions, @code{I} modifier
@ifset PERL
@cindex Perl-style regular expressions, case-insensitive
@end ifset
The @code{I} modifier to regular-expression matching is a @acronym{GNU}
extension which causes the @var{regexp} to be matched in
a case-insensitive manner.

@item /@var{regexp}/M
@itemx \%@var{regexp}%M
@cindex @value{SSEDEXT}, @code{M} modifier
@ifset PERL
@cindex Perl-style regular expressions, multiline
@end ifset
The @code{M} modifier to regular-expression matching is a @value{SSED}
extension which directs @value{SSED} to match the regular expression
in @cite{multi-line} mode.  The modifier causes @code{^} and @code{$} to
match respectively (in addition to the normal behavior) the empty string
after a newline, and the empty string before a newline.  There are
special character sequences
@ifset PERL
(@code{\A} and @code{\Z} in Perl mode, @code{\`} and @code{\'}
in basic or extended regular expression modes)
@end ifset
@ifclear PERL
(@code{\`} and @code{\'})
@end ifclear
which always match the beginning or the end of the buffer.
In addition,
@ifset PERL
just like in Perl mode without the @code{S} modifier,
@end ifset
the period character does not match a new-line character in
multi-line mode.

@ifset PERL
@item /@var{regexp}/S
@itemx \%@var{regexp}%S
@cindex @value{SSEDEXT}, @code{S} modifier
@cindex Perl-style regular expressions, single line
The @code{S} modifier to regular-expression matching is only valid
in Perl mode and specifies that the dot character (@code{.}) will
match the newline character too.  @code{S} stands for @cite{single-line}.
@end ifset

@ifset PERL
@item /@var{regexp}/X
@itemx \%@var{regexp}%X
@cindex @value{SSEDEXT}, @code{X} modifier
@cindex Perl-style regular expressions, extended
The @code{X} modifier to regular-expression matching is also
valid in Perl mode only.  If it is used, whitespace in the
pattern (other than in a character class) and
characters between a @kbd{#} outside a character class and the
next newline character are ignored. An escaping backslash
can be used to include a whitespace or @kbd{#} character as part
of the pattern.
@end ifset
@end table

If no addresses are given, then all lines are matched;
if one address is given, then only lines matching that
address are matched.

@cindex Range of lines
@cindex Several lines, selecting
An address range can be specified by specifying two addresses
separated by a comma (@code{,}).  An address range matches lines
starting from where the first address matches, and continues
until the second address matches (inclusively).

If the second address is a @var{regexp}, then checking for the
ending match will start with the line @emph{following} the
line which matched the first address: a range will always
span at least two lines (except of course if the input stream
ends).

If the second address is a @var{number} less than (or equal to)
the line matching the first address, then only the one line is
matched.

@cindex Special addressing forms
@cindex Range with start address of zero
@cindex Zero, as range start address
@cindex @var{addr1},+N
@cindex @var{addr1},~N
@cindex @acronym{GNU} extensions, special two-address forms
@cindex @acronym{GNU} extensions, @code{0} address
@cindex @acronym{GNU} extensions, 0,@var{addr2} addressing
@cindex @acronym{GNU} extensions, @var{addr1},+@var{N} addressing
@cindex @acronym{GNU} extensions, @var{addr1},~@var{N} addressing
@value{SSED} also supports some special two-address forms; all these
are @acronym{GNU} extensions:
@table @code
@item 0,/@var{regexp}/
A line number of @code{0} can be used in an address specification like
@code{0,/@var{regexp}/} so that @command{sed} will try to match
@var{regexp} in the first input line too.  In other words,
@code{0,/@var{regexp}/} is similar to @code{1,/@var{regexp}/},
except that if @var{addr2} matches the very first line of input the
@code{0,/@var{regexp}/} form will consider it to end the range, whereas
the @code{1,/@var{regexp}/} form will match the beginning of its range and
hence make the range span up to the @emph{second} occurrence of the
regular expression.

Note that this is the only place where the @code{0} address makes
sense; there is no 0-th line and commands which are given the @code{0}
address in any other way will give an error.

@item @var{addr1},+@var{N}
Matches @var{addr1} and the @var{N} lines following @var{addr1}.

@item @var{addr1},~@var{N}
Matches @var{addr1} and the lines following @var{addr1}
until the next line whose input line number is a multiple of @var{N}.
@end table

@cindex Excluding lines
@cindex Selecting non-matching lines
Appending the @code{!} character to the end of an address
specification negates the sense of the match.
That is, if the @code{!} character follows an address range,
then only lines which do @emph{not} match the address range
will be selected.
This also works for singleton addresses,
and, perhaps perversely, for the null address.


@node Regular Expressions
@section Overview of Regular Expression Syntax

To know how to use @command{sed}, people should understand regular
expressions (@dfn{regexp} for short).  A regular expression
is a pattern that is matched against a
subject string from left to right.  Most characters are
@dfn{ordinary}: they stand for
themselves in a pattern, and match the corresponding characters
in the subject.  As a trivial example, the pattern

@example
The quick brown fox
@end example

@noindent
matches a portion of a subject string that is identical to
itself.  The power of regular expressions comes from the
ability to include alternatives and repetitions in the pattern.
These are encoded in the pattern by the use of @dfn{special characters},
which do not stand for themselves but instead
are interpreted in some special way.  Here is a brief description
of regular expression syntax as used in @command{sed}.

@table @code
@item @var{char}
A single ordinary character matches itself.

@item *
@cindex @acronym{GNU} extensions, to basic regular expressions
Matches a sequence of zero or more instances of matches for the
preceding regular expression, which must be an ordinary character, a
special character preceded by @code{\}, a @code{.}, a grouped regexp
(see below), or a bracket expression.  As a @acronym{GNU} extension, a
postfixed regular expression can also be followed by @code{*}; for
example, @code{a**} is equivalent to @code{a*}.  @acronym{POSIX}
1003.1-2001 says that @code{*} stands for itself when it appears at
the start of a regular expression or subexpression, but many
non@acronym{GNU} implementations do not support this and portable
scripts should instead use @code{\*} in these contexts.

@item \+
@cindex @acronym{GNU} extensions, to basic regular expressions
As @code{*}, but matches one or more.  It is a @acronym{GNU} extension.

@item \?
@cindex @acronym{GNU} extensions, to basic regular expressions
As @code{*}, but only matches zero or one.  It is a @acronym{GNU} extension.

@item \@{@var{i}\@}
As @code{*}, but matches exactly @var{i} sequences (@var{i} is a
decimal integer; for portability, keep it between 0 and 255
inclusive).

@item \@{@var{i},@var{j}\@}
Matches between @var{i} and @var{j}, inclusive, sequences.

@item \@{@var{i},\@}
Matches more than or equal to @var{i} sequences.

@item \(@var{regexp}\)
Groups the inner @var{regexp} as a whole, this is used to: 

@itemize @bullet
@item
@cindex @acronym{GNU} extensions, to basic regular expressions
Apply postfix operators, like @code{\(abcd\)*}:
this will search for zero or more whole sequences 
of @samp{abcd}, while @code{abcd*} would search
for @samp{abc} followed by zero or more occurrences
of @samp{d}.  Note that support for @code{\(abcd\)*} is
required by @acronym{POSIX} 1003.1-2001, but many non-@acronym{GNU}
implementations do not support it and hence it is not universally
portable.         

@item
Use back references (see below).
@end itemize

@item .
Matches any character, including newline.

@item ^
Matches the null string at beginning of the pattern space, i.e. what
appears after the circumflex must appear at the beginning of the
pattern space.

In most scripts, pattern space is initialized to the content of each
line (@pxref{Execution Cycle, , How @code{sed} works}).  So, it is a
useful simplification to think of @code{^#include} as matching only
lines where @samp{#include} is the first thing on line---if there are
spaces before, for example, the match fails.  This simplification is
valid as long as the original content of pattern space is not modified,
for example with an @code{s} command.

@code{^} acts as a special character only at the beginning of the
regular expression or subexpression (that is, after @code{\(} or
@code{\|}).  Portable scripts should avoid @code{^} at the beginning of
a subexpression, though, as @acronym{POSIX} allows implementations that
treat @code{^} as an ordinary character in that context.

@item $
It is the same as @code{^}, but refers to end of pattern space.
@code{$} also acts as a special character only at the end
of the regular expression or subexpression (that is, before @code{\)}
or @code{\|}), and its use at the end of a subexpression is not
portable.


@item [@var{list}]
@itemx [^@var{list}]
Matches any single character in @var{list}: for example,
@code{[aeiou]} matches all vowels.  A list may include
sequences like @code{@var{char1}-@var{char2}}, which
matches any character between (inclusive) @var{char1}
and @var{char2}.

A leading @code{^} reverses the meaning of @var{list}, so that
it matches any single character @emph{not} in @var{list}.  To include
@code{]} in the list, make it the first character (after
the @code{^} if needed), to include @code{-} in the list,
make it the first or last; to include @code{^} put
it after the first character.

@cindex @code{POSIXLY_CORRECT} behavior, bracket expressions
The characters @code{$}, @code{*}, @code{.}, @code{[}, and @code{\}
are normally not special within @var{list}.  For example, @code{[\*]}
matches either @samp{\} or @samp{*}, because the @code{\} is not
special here.  However, strings like @code{[.ch.]}, @code{[=a=]}, and
@code{[:space:]} are special within @var{list} and represent collating
symbols, equivalence classes, and character classes, respectively, and
@code{[} is therefore special within @var{list} when it is followed by
@code{.}, @code{=}, or @code{:}.  Also, when not in
@env{POSIXLY_CORRECT} mode, special escapes like @code{\n} and
@code{\t} are recognized within @var{list}.  @xref{Escapes}.

@item @var{regexp1}\|@var{regexp2}
@cindex @acronym{GNU} extensions, to basic regular expressions
Matches either @var{regexp1} or @var{regexp2}.  Use
parentheses to use complex alternative regular expressions.
The matching process tries each alternative in turn, from
left to right, and the first one that succeeds is used.
It is a @acronym{GNU} extension.

@item @var{regexp1}@var{regexp2}
Matches the concatenation of @var{regexp1} and @var{regexp2}.
Concatenation binds more tightly than @code{\|}, @code{^}, and
@code{$}, but less tightly than the other regular expression
operators.

@item \@var{digit}
Matches the @var{digit}-th @code{\(@dots{}\)} parenthesized
subexpression in the regular expression.  This is called a @dfn{back
reference}.  Subexpressions are implicity numbered by counting
occurrences of @code{\(} left-to-right.

@item \n
Matches the newline character.

@item \@var{char}
Matches @var{char}, where @var{char} is one of @code{$},
@code{*}, @code{.}, @code{[}, @code{\}, or @code{^}.
Note that the only C-like
backslash sequences that you can portably assume to be
interpreted are @code{\n} and @code{\\}; in particular
@code{\t} is not portable, and matches a @samp{t} under most
implementations of @command{sed}, rather than a tab character.

@end table

@cindex Greedy regular expression matching
Note that the regular expression matcher is greedy, i.e., matches
are attempted from left to right and, if two or more matches are
possible starting at the same character, it selects the longest.

@noindent
Examples:
@table @samp
@item abcdef
Matches @samp{abcdef}.

@item a*b
Matches zero or more @samp{a}s followed by a single
@samp{b}.  For example, @samp{b} or @samp{aaaaab}. 

@item a\?b
Matches @samp{b} or @samp{ab}.

@item a\+b\+
Matches one or more @samp{a}s followed by one or more
@samp{b}s: @samp{ab} is the shortest possible match, but
other examples are @samp{aaaab} or @samp{abbbbb} or
@samp{aaaaaabbbbbbb}.

@item .*
@itemx .\+
These two both match all the characters in a string;
however, the first matches every string (including the empty
string), while the second matches only strings containing
at least one character.

@item ^main.*(.*)
This matches a string starting with @samp{main},
followed by an opening and closing
parenthesis.  The @samp{n}, @samp{(} and @samp{)} need not
be adjacent.

@item ^#
This matches a string beginning with @samp{#}.

@item \\$
This matches a string ending with a single backslash.  The
regexp contains two backslashes for escaping.

@item \$
Instead, this matches a string consisting of a single dollar sign,
because it is escaped.

@item [a-zA-Z0-9]
In the C locale, this matches any @acronym{ASCII} letters or digits.

@item [^ @kbd{tab}]\+
(Here @kbd{tab} stands for a single tab character.)
This matches a string of one or more
characters, none of which is a space or a tab.
Usually this means a word.

@item ^\(.*\)\n\1$
This matches a string consisting of two equal substrings separated by
a newline.

@item .\@{9\@}A$
This matches nine characters followed by an @samp{A}.

@item ^.\@{15\@}A
This matches the start of a string that contains 16 characters,
the last of which is an @samp{A}.

@end table



@node Common Commands
@section Often-Used Commands

If you use @command{sed} at all, you will quite likely want to know
these commands.

@table @code
@item #
[No addresses allowed.]

@findex # (comments)
@cindex Comments, in scripts
The @code{#} character begins a comment;
the comment continues until the next newline.

@cindex Portability, comments
If you are concerned about portability, be aware that
some implementations of @command{sed} (which are not @sc{posix}
conformant) may only support a single one-line comment,
and then only when the very first character of the script is a @code{#}.

@findex -n, forcing from within a script
@cindex Caveat --- #n on first line
Warning: if the first two characters of the @command{sed} script
are @code{#n}, then the @option{-n} (no-autoprint) option is forced.
If you want to put a comment in the first line of your script
and that comment begins with the letter @samp{n}
and you do not want this behavior,
then be sure to either use a capital @samp{N},
or place at least one space before the @samp{n}.

@item q [@var{exit-code}]
This command only accepts a single address.

@findex q (quit) command
@cindex @value{SSEDEXT}, returning an exit code
@cindex Quitting
Exit @command{sed} without processing any more commands or input.
Note that the current pattern space is printed if auto-print is
not disabled with the @option{-n} options.  The ability to return
an exit code from the @command{sed} script is a @value{SSED} extension.

@item d
@findex d (delete) command
@cindex Text, deleting
Delete the pattern space;
immediately start next cycle.

@item p
@findex p (print) command
@cindex Text, printing
Print out the pattern space (to the standard output).
This command is usually only used in conjunction with the @option{-n}
command-line option.

@item n
@findex n (next-line) command
@cindex Next input line, replace pattern space with
@cindex Read next input line
If auto-print is not disabled, print the pattern space,
then, regardless, replace the pattern space with the next line of input.
If there is no more input then @command{sed} exits without processing
any more commands.

@item @{ @var{commands} @}
@findex @{@} command grouping
@cindex Grouping commands
@cindex Command groups
A group of commands may be enclosed between
@code{@{} and @code{@}} characters.
This is particularly useful when you want a group of commands
to be triggered by a single address (or address-range) match.

@end table

@node The "s" Command
@section The @code{s} Command

The syntax of the @code{s} (as in substitute) command is
@samp{s/@var{regexp}/@var{replacement}/@var{flags}}.  The @code{/}
characters may be uniformly replaced by any other single
character within any given @code{s} command.  The @code{/}
character (or whatever other character is used in its stead)
can appear in the @var{regexp} or @var{replacement}
only if it is preceded by a @code{\} character.

The @code{s} command is probably the most important in @command{sed}
and has a lot of different options.  Its basic concept is simple:
the @code{s} command attempts to match the pattern
space against the supplied @var{regexp}; if the match is
successful, then that portion of the pattern
space which was matched is replaced with @var{replacement}.

@cindex Backreferences, in regular expressions
@cindex Parenthesized substrings
The @var{replacement} can contain @code{\@var{n}} (@var{n} being
a number from 1 to 9, inclusive) references, which refer to
the portion of the match which is contained between the @var{n}th
@code{\(} and its matching @code{\)}.
Also, the @var{replacement} can contain unescaped @code{&}
characters which reference the whole matched portion
of the pattern space.
@cindex @value{SSEDEXT}, case modifiers in @code{s} commands
Finally, as a @value{SSED} extension, you can include a
special sequence made of a backslash and one of the letters
@code{L}, @code{l}, @code{U}, @code{u}, or @code{E}.
The meaning is as follows:

@table @code
@item \L
Turn the replacement
to lowercase until a @code{\U} or @code{\E} is found,

@item \l
Turn the
next character to lowercase,

@item \U
Turn the replacement to uppercase
until a @code{\L} or @code{\E} is found,

@item \u
Turn the next character
to uppercase,

@item \E
Stop case conversion started by @code{\L} or @code{\U}.
@end table

When the @code{g} flag is being used, case conversion does not
propagate from one occurrence of the regular expression to
another.  For example, when the following command is executed
with @samp{a-b-} in pattern space:
@example
s/\(b\?\)-/x\u\1/g
@end example

@noindent
the output is @samp{axxB}.  When replacing the first @samp{-},
the @samp{\u} sequence only affects the empty replacement of
@samp{\1}.  It does not affect the @code{x} character that is
added to pattern space when replacing @code{b-} with @code{xB}.

On the other hand, @code{\l} and @code{\u} do affect the remainder
of the replacement text if they are followed by an empty substitution.
With @samp{a-b-} in pattern space, the following command:
@example
s/\(b\?\)-/\u\1x/g
@end example

@noindent
will replace @samp{-} with @samp{X} (uppercase) and @samp{b-} with
@samp{Bx}.  If this behavior is undesirable, you can prevent it by
adding a @samp{\E} sequence---after @samp{\1} in this case.

To include a literal @code{\}, @code{&}, or newline in the final
replacement, be sure to precede the desired @code{\}, @code{&},
or newline in the @var{replacement} with a @code{\}.

@findex s command, option flags
@cindex Substitution of text, options
The @code{s} command can be followed by zero or more of the
following @var{flags}:

@table @code
@item g
@cindex Global substitution
@cindex Replacing all text matching regexp in a line
Apply the replacement to @emph{all} matches to the @var{regexp},
not just the first.

@item @var{number}
@cindex Replacing only @var{n}th match of regexp in a line
Only replace the @var{number}th match of the @var{regexp}.

@cindex @acronym{GNU} extensions, @code{g} and @var{number} modifier interaction in @code{s} command
@cindex Mixing @code{g} and @var{number} modifiers in the @code{s} command
Note: the @sc{posix} standard does not specify what should happen
when you mix the @code{g} and @var{number} modifiers,
and currently there is no widely agreed upon meaning
across @command{sed} implementations.
For @value{SSED}, the interaction is defined to be:
ignore matches before the @var{number}th,
and then match and replace all matches from
the @var{number}th on.

@item p
@cindex Text, printing after substitution
If the substitution was made, then print the new pattern space.

Note: when both the @code{p} and @code{e} options are specified,
the relative ordering of the two produces very different results.
In general, @code{ep} (evaluate then print) is what you want,
but operating the other way round can be useful for debugging.
For this reason, the current version of @value{SSED} interprets
specially the presence of @code{p} options both before and after
@code{e}, printing the pattern space before and after evaluation,
while in general flags for the @code{s} command show their
effect just once.  This behavior, although documented, might
change in future versions.

@item w @var{file-name}
@cindex Text, writing to a file after substitution
@cindex @value{SSEDEXT}, @file{/dev/stdout} file
@cindex @value{SSEDEXT}, @file{/dev/stderr} file
If the substitution was made, then write out the result to the named file.
As a @value{SSED} extension, two special values of @var{file-name} are
supported: @file{/dev/stderr}, which writes the result to the standard
error, and @file{/dev/stdout}, which writes to the standard
output.@footnote{This is equivalent to @code{p} unless the @option{-i}
option is being used.}

@item e
@cindex Evaluate Bourne-shell commands, after substitution
@cindex Subprocesses
@cindex @value{SSEDEXT}, evaluating Bourne-shell commands
@cindex @value{SSEDEXT}, subprocesses
This command allows one to pipe input from a shell command
into pattern space.  If a substitution was made, the command
that is found in pattern space is executed and pattern space
is replaced with its output.  A trailing newline is suppressed;
results are undefined if the command to be executed contains
a @sc{nul} character.  This is a @value{SSED} extension.

@item I
@itemx i
@cindex @acronym{GNU} extensions, @code{I} modifier
@cindex Case-insensitive matching
@ifset PERL
@cindex Perl-style regular expressions, case-insensitive
@end ifset
The @code{I} modifier to regular-expression matching is a @acronym{GNU}
extension which makes @command{sed} match @var{regexp} in a
case-insensitive manner.

@item M
@itemx m
@cindex @value{SSEDEXT}, @code{M} modifier
@ifset PERL
@cindex Perl-style regular expressions, multiline
@end ifset
The @code{M} modifier to regular-expression matching is a @value{SSED}
extension which directs @value{SSED} to match the regular expression
in @cite{multi-line} mode.  The modifier causes @code{^} and @code{$} to
match respectively (in addition to the normal behavior) the empty string
after a newline, and the empty string before a newline.  There are
special character sequences
@ifset PERL
(@code{\A} and @code{\Z} in Perl mode, @code{\`} and @code{\'}
in basic or extended regular expression modes)
@end ifset
@ifclear PERL
(@code{\`} and @code{\'})
@end ifclear
which always match the beginning or the end of the buffer.
In addition,
@ifset PERL
just like in Perl mode without the @code{S} modifier,
@end ifset
the period character does not match a new-line character in
multi-line mode.

@ifset PERL
@item S
@itemx s
@cindex @value{SSEDEXT}, @code{S} modifier
@cindex Perl-style regular expressions, single line
The @code{S} modifier to regular-expression matching is only valid
in Perl mode and specifies that the dot character (@code{.}) will
match the newline character too.  @code{S} stands for @cite{single-line}.
@end ifset

@ifset PERL
@item X
@itemx x
@cindex @value{SSEDEXT}, @code{X} modifier
@cindex Perl-style regular expressions, extended
The @code{X} modifier to regular-expression matching is also
valid in Perl mode only.  If it is used, whitespace in the
pattern (other than in a character class) and
characters between a @kbd{#} outside a character class and the
next newline character are ignored. An escaping backslash
can be used to include a whitespace or @kbd{#} character as part
of the pattern.
@end ifset
@end table


@node Other Commands
@section Less Frequently-Used Commands

Though perhaps less frequently used than those in the previous
section, some very small yet useful @command{sed} scripts can be built with
these commands.

@table @code
@item y/@var{source-chars}/@var{dest-chars}/
(The @code{/} characters may be uniformly replaced by
any other single character within any given @code{y} command.)

@findex y (transliterate) command
@cindex Transliteration
Transliterate any characters in the pattern space which match
any of the @var{source-chars} with the corresponding character
in @var{dest-chars}.

Instances of the @code{/} (or whatever other character is used in its stead),
@code{\}, or newlines can appear in the @var{source-chars} or @var{dest-chars}
lists, provide that each instance is escaped by a @code{\}.
The @var{source-chars} and @var{dest-chars} lists @emph{must}
contain the same number of characters (after de-escaping).

@item a\
@itemx @var{text}
@cindex @value{SSEDEXT}, two addresses supported by most commands
As a @acronym{GNU} extension, this command accepts two addresses.

@findex a (append text lines) command
@cindex Appending text after a line
@cindex Text, appending
Queue the lines of text which follow this command
(each but the last ending with a @code{\},
which are removed from the output)
to be output at the end of the current cycle,
or when the next input line is read.

Escape sequences in @var{text} are processed, so you should
use @code{\\} in @var{text} to print a single backslash.

As a @acronym{GNU} extension, if between the @code{a} and the newline there is
other than a whitespace-@code{\} sequence, then the text of this line,
starting at the first non-whitespace character after the @code{a},
is taken as the first line of the @var{text} block.
(This enables a simplification in scripting a one-line add.)
This extension also works with the @code{i} and @code{c} commands.

@item i\
@itemx @var{text}
@cindex @value{SSEDEXT}, two addresses supported by most commands
As a @acronym{GNU} extension, this command accepts two addresses.

@findex i (insert text lines) command
@cindex Inserting text before a line
@cindex Text, insertion
Immediately output the lines of text which follow this command
(each but the last ending with a @code{\},
which are removed from the output).

@item c\
@itemx @var{text}
@findex c (change to text lines) command
@cindex Replacing selected lines with other text
Delete the lines matching the address or address-range,
and output the lines of text which follow this command
(each but the last ending with a @code{\},
which are removed from the output)
in place of the last line
(or in place of each line, if no addresses were specified).
A new cycle is started after this command is done,
since the pattern space will have been deleted.

@item =
@cindex @value{SSEDEXT}, two addresses supported by most commands
As a @acronym{GNU} extension, this command accepts two addresses.

@findex = (print line number) command
@cindex Printing line number
@cindex Line number, printing
Print out the current input line number (with a trailing newline).

@item l @var{n}
@findex l (list unambiguously) command
@cindex List pattern space
@cindex Printing text unambiguously
@cindex Line length, setting
@cindex @value{SSEDEXT}, setting line length
Print the pattern space in an unambiguous form:
non-printable characters (and the @code{\} character)
are printed in C-style escaped form; long lines are split,
with a trailing @code{\} character to indicate the split;
the end of each line is marked with a @code{$}.

@var{n} specifies the desired line-wrap length;
a length of 0 (zero) means to never wrap long lines.  If omitted,
the default as specified on the command line is used.  The @var{n}
parameter is a @value{SSED} extension.

@item r @var{filename}
@cindex @value{SSEDEXT}, two addresses supported by most commands
As a @acronym{GNU} extension, this command accepts two addresses.

@findex r (read file) command
@cindex Read text from a file
@cindex @value{SSEDEXT}, @file{/dev/stdin} file
Queue the contents of @var{filename} to be read and
inserted into the output stream at the end of the current cycle,
or when the next input line is read.
Note that if @var{filename} cannot be read, it is treated as
if it were an empty file, without any error indication.

As a @value{SSED} extension, the special value @file{/dev/stdin}
is supported for the file name, which reads the contents of the
standard input.

@item w @var{filename}
@findex w (write file) command
@cindex Write to a file
@cindex @value{SSEDEXT}, @file{/dev/stdout} file
@cindex @value{SSEDEXT}, @file{/dev/stderr} file
Write the pattern space to @var{filename}.
As a @value{SSED} extension, two special values of @var{file-name} are
supported: @file{/dev/stderr}, which writes the result to the standard
error, and @file{/dev/stdout}, which writes to the standard
output.@footnote{This is equivalent to @code{p} unless the @option{-i}
option is being used.}

The file will be created (or truncated) before the first input line is
read; all @code{w} commands (including instances of the @code{w} flag
on successful @code{s} commands) which refer to the same @var{filename}
are output without closing and reopening the file.

@item D
@findex D (delete first line) command
@cindex Delete first line from pattern space
If pattern space contains no newline, start a normal new cycle as if
the @code{d} command was issued.  Otherwise, delete text in the pattern
space up to the first newline, and restart cycle with the resultant
pattern space, without reading a new line of input.

@item N
@findex N (append Next line) command
@cindex Next input line, append to pattern space
@cindex Append next input line to pattern space
Add a newline to the pattern space,
then append the next line of input to the pattern space.
If there is no more input then @command{sed} exits without processing
any more commands.

@item P
@findex P (print first line) command
@cindex Print first line from pattern space
Print out the portion of the pattern space up to the first newline.

@item h
@findex h (hold) command
@cindex Copy pattern space into hold space
@cindex Replace hold space with copy of pattern space
@cindex Hold space, copying pattern space into
Replace the contents of the hold space with the contents of the pattern space.

@item H
@findex H (append Hold) command
@cindex Append pattern space to hold space
@cindex Hold space, appending from pattern space
Append a newline to the contents of the hold space,
and then append the contents of the pattern space to that of the hold space.

@item g
@findex g (get) command
@cindex Copy hold space into pattern space
@cindex Replace pattern space with copy of hold space
@cindex Hold space, copy into pattern space
Replace the contents of the pattern space with the contents of the hold space.

@item G
@findex G (appending Get) command
@cindex Append hold space to pattern space
@cindex Hold space, appending to pattern space
Append a newline to the contents of the pattern space,
and then append the contents of the hold space to that of the pattern space.

@item x
@findex x (eXchange) command
@cindex Exchange hold space with pattern space
@cindex Hold space, exchange with pattern space
Exchange the contents of the hold and pattern spaces.

@end table


@node Programming Commands
@section Commands for @command{sed} gurus

In most cases, use of these commands indicates that you are
probably better off programming in something like @command{awk}
or Perl.  But occasionally one is committed to sticking
with @command{sed}, and these commands can enable one to write
quite convoluted scripts.

@cindex Flow of control in scripts
@table @code
@item : @var{label}
[No addresses allowed.]

@findex : (label) command
@cindex Labels, in scripts
Specify the location of @var{label} for branch commands.
In all other respects, a no-op.

@item b @var{label}
@findex b (branch) command
@cindex Branch to a label, unconditionally
@cindex Goto, in scripts
Unconditionally branch to @var{label}.
The @var{label} may be omitted, in which case the next cycle is started.

@item t @var{label}
@findex t (test and branch if successful) command
@cindex Branch to a label, if @code{s///} succeeded
@cindex Conditional branch
Branch to @var{label} only if there has been a successful @code{s}ubstitution
since the last input line was read or conditional branch was taken.
The @var{label} may be omitted, in which case the next cycle is started.

@end table

@node Extended Commands
@section Commands Specific to @value{SSED}

These commands are specific to @value{SSED}, so you
must use them with care and only when you are sure that
hindering portability is not evil.  They allow you to check
for @value{SSED} extensions or to do tasks that are required
quite often, yet are unsupported by standard @command{sed}s.

@table @code
@item e [@var{command}]
@findex e (evaluate) command
@cindex Evaluate Bourne-shell commands
@cindex Subprocesses
@cindex @value{SSEDEXT}, evaluating Bourne-shell commands
@cindex @value{SSEDEXT}, subprocesses
This command allows one to pipe input from a shell command
into pattern space.  Without parameters, the @code{e} command
executes the command that is found in pattern space and
replaces the pattern space with the output; a trailing newline
is suppressed.

If a parameter is specified, instead, the @code{e} command
interprets it as a command and sends its output to the output stream.
The command can run across multiple lines, all but the last ending with
a back-slash.

In both cases, the results are undefined if the command to be
executed contains a @sc{nul} character.

Note that, unlike the @code{r} command, the output of the command will
be printed immediately; the @code{r} command instead delays the output
to the end of the current cycle.

@item F
@findex F (File name) command
@cindex Printing file name
@cindex File name, printing
Print out the file name of the current input file (with a trailing
newline).

@item L @var{n}
@findex L (fLow paragraphs) command
@cindex Reformat pattern space
@cindex Reformatting paragraphs
@cindex @value{SSEDEXT}, reformatting paragraphs
@cindex @value{SSEDEXT}, @code{L} command
This @value{SSED} extension fills and joins lines in pattern space
to produce output lines of (at most) @var{n} characters, like
@code{fmt} does; if @var{n} is omitted, the default as specified
on the command line is used.  This command is considered a failed
experiment and unless there is enough request (which seems unlikely)
will be removed in future versions.

@ignore
Blank lines, spaces between words, and indentation are
preserved in the output; successive input lines with different
indentation are not joined; tabs are expanded to 8 columns.

If the pattern space contains multiple lines, they are joined, but
since the pattern space usually contains a single line, the behavior
of a simple @code{L;d} script is the same as @samp{fmt -s} (i.e.,
it does not join short lines to form longer ones).

@var{n} specifies the desired line-wrap length; if omitted,
the default as specified on the command line is used.
@end ignore

@item Q [@var{exit-code}]
This command only accepts a single address.

@findex Q (silent Quit) command
@cindex @value{SSEDEXT}, quitting silently
@cindex @value{SSEDEXT}, returning an exit code
@cindex Quitting
This command is the same as @code{q}, but will not print the
contents of pattern space.  Like @code{q}, it provides the
ability to return an exit code to the caller.

This command can be useful because the only alternative ways
to accomplish this apparently trivial function are to use
the @option{-n} option (which can unnecessarily complicate
your script) or resorting to the following snippet, which
wastes time by reading the whole file without any visible effect:

@example
:eat
$d       @i{@r{Quit silently on the last line}}
N        @i{@r{Read another line, silently}}
g        @i{@r{Overwrite pattern space each time to save memory}}
b eat
@end example

@item R @var{filename}
@findex R (read line) command
@cindex Read text from a file
@cindex @value{SSEDEXT}, reading a file a line at a time
@cindex @value{SSEDEXT}, @code{R} command
@cindex @value{SSEDEXT}, @file{/dev/stdin} file
Queue a line of @var{filename} to be read and
inserted into the output stream at the end of the current cycle,
or when the next input line is read.
Note that if @var{filename} cannot be read, or if its end is
reached, no line is appended, without any error indication.

As with the @code{r} command, the special value @file{/dev/stdin}
is supported for the file name, which reads a line from the
standard input.

@item T @var{label}
@findex T (test and branch if failed) command
@cindex @value{SSEDEXT}, branch if @code{s///} failed
@cindex Branch to a label, if @code{s///} failed
@cindex Conditional branch
Branch to @var{label} only if there have been no successful
@code{s}ubstitutions since the last input line was read or
conditional branch was taken. The @var{label} may be omitted,
in which case the next cycle is started.

@item v @var{version}
@findex v (version) command
@cindex @value{SSEDEXT}, checking for their presence
@cindex Requiring @value{SSED}
This command does nothing, but makes @command{sed} fail if
@value{SSED} extensions are not supported, simply because other
versions of @command{sed} do not implement it.  In addition, you
can specify the version of @command{sed} that your script
requires, such as @code{4.0.5}.  The default is @code{4.0}
because that is the first version that implemented this command.

This command enables all @value{SSEDEXT} even if
@env{POSIXLY_CORRECT} is set in the environment.

@item W @var{filename}
@findex W (write first line) command
@cindex Write first line to a file
@cindex @value{SSEDEXT}, writing first line to a file
Write to the given filename the portion of the pattern space up to
the first newline.  Everything said under the @code{w} command about
file handling holds here too.

@item z
@findex z (Zap) command
@cindex @value{SSEDEXT}, emptying pattern space
@cindex Emptying pattern space
This command empties the content of pattern space.  It is
usually the same as @samp{s/.*//}, but is more efficient
and works in the presence of invalid multibyte sequences
in the input stream.  @sc{posix} mandates that such sequences
are @emph{not} matched by @samp{.}, so that there is no portable
way to clear @command{sed}'s buffers in the middle of the
script in most multibyte locales (including UTF-8 locales).
@end table

@node Escapes
@section @acronym{GNU} Extensions for Escapes in Regular Expressions

@cindex @acronym{GNU} extensions, special escapes
Until this chapter, we have only encountered escapes of the form
@samp{\^}, which tell @command{sed} not to interpret the circumflex
as a special character, but rather to take it literally.  For
example, @samp{\*} matches a single asterisk rather than zero
or more backslashes.

@cindex @code{POSIXLY_CORRECT} behavior, escapes
This chapter introduces another kind of escape@footnote{All
the escapes introduced here are @acronym{GNU}
extensions, with the exception of @code{\n}.  In basic regular
expression mode, setting @code{POSIXLY_CORRECT} disables them inside
bracket expressions.}---that
is, escapes that are applied to a character or sequence of characters
that ordinarily are taken literally, and that @command{sed} replaces
with a special character.  This provides a way
of encoding non-printable characters in patterns in a visible manner.
There is no restriction on the appearance of non-printing characters
in a @command{sed} script but when a script is being prepared in the
shell or by text editing, it is usually easier to use one of
the following escape sequences than the binary character it
represents:

The list of these escapes is:

@table @code
@item \a
Produces or matches a @sc{bel} character, that is an ``alert'' (@sc{ascii} 7).

@item \f
Produces or matches a form feed (@sc{ascii} 12).

@item \n
Produces or matches a newline (@sc{ascii} 10).

@item \r
Produces or matches a carriage return (@sc{ascii} 13).

@item \t
Produces or matches a horizontal tab (@sc{ascii} 9).

@item \v
Produces or matches a so called ``vertical tab'' (@sc{ascii} 11).

@item \c@var{x}
Produces or matches @kbd{@sc{Control}-@var{x}}, where @var{x} is
any character.  The precise effect of @samp{\c@var{x}} is as follows:
if @var{x} is a lower case letter, it is converted to upper case.
Then bit 6 of the character (hex 40) is inverted.  Thus @samp{\cz} becomes
hex 1A, but @samp{\c@{} becomes hex 3B, while @samp{\c;} becomes hex 7B.

@item \d@var{xxx}
Produces or matches a character whose decimal @sc{ascii} value is @var{xxx}.

@item \o@var{xxx}
@ifset PERL
@item \@var{xxx}
@end ifset
Produces or matches a character whose octal @sc{ascii} value is @var{xxx}.
@ifset PERL
The syntax without the @code{o} is active in Perl mode, while the one
with the @code{o} is active in the normal or extended @sc{posix} regular
expression modes.
@end ifset

@item \x@var{xx}
Produces or matches a character whose hexadecimal @sc{ascii} value is @var{xx}.
@end table

@samp{\b} (backspace) was omitted because of the conflict with
the existing ``word boundary'' meaning.

Other escapes match a particular character class and are valid only in
regular expressions:

@table @code
@item \w
Matches any ``word'' character.  A ``word'' character is any
letter or digit or the underscore character.

@item \W
Matches any ``non-word'' character.

@item \b
Matches a word boundary; that is it matches if the character
to the left is a ``word'' character and the character to the
right is a ``non-word'' character, or vice-versa.

@item \B
Matches everywhere but on a word boundary; that is it matches
if the character to the left and the character to the right
are either both ``word'' characters or both ``non-word''
characters.

@item \`
Matches only at the start of pattern space.  This is different
from @code{^} in multi-line mode.

@item \'
Matches only at the end of pattern space.  This is different
from @code{$} in multi-line mode.

@ifset PERL
@item \G
Match only at the start of pattern space or, when doing a global
substitution using the @code{s///g} command and option, at
the end-of-match position of the prior match.  For example,
@samp{s/\Ga/Z/g} will change an initial run of @code{a}s to
a run of @code{Z}s
@end ifset
@end table

@node Examples
@chapter Some Sample Scripts

Here are some @command{sed} scripts to guide you in the art of mastering
@command{sed}.

@menu
Some exotic examples:
* Centering lines::
* Increment a number::
* Rename files to lower case::
* Print bash environment::
* Reverse chars of lines::

Emulating standard utilities:
* tac::                             Reverse lines of files
* cat -n::                          Numbering lines
* cat -b::                          Numbering non-blank lines
* wc -c::                           Counting chars
* wc -w::                           Counting words
* wc -l::                           Counting lines
* head::                            Printing the first lines
* tail::                            Printing the last lines
* uniq::                            Make duplicate lines unique
* uniq -d::                         Print duplicated lines of input
* uniq -u::                         Remove all duplicated lines
* cat -s::                          Squeezing blank lines
@end menu

@node Centering lines
@section Centering Lines

This script centers all lines of a file on a 80 columns width.
To change that width, the number in @code{\@{@dots{}\@}} must be
replaced, and the number of added spaces also must be changed.

Note how the buffer commands are used to separate parts in
the regular expressions to be matched---this is a common
technique.

@c start-------------------------------------------
@example
#!/usr/bin/sed -f

# Put 80 spaces in the buffer
1 @{
  x
  s/^$/          /
  s/^.*$/&&&&&&&&/
  x
@}

# del leading and trailing spaces
y/@kbd{tab}/ /
s/^ *//
s/ *$//

# add a newline and 80 spaces to end of line
G

# keep first 81 chars (80 + a newline)
s/^\(.\@{81\@}\).*$/\1/

# \2 matches half of the spaces, which are moved to the beginning
s/^\(.*\)\n\(.*\)\2/\2\1/
@end example
@c end---------------------------------------------

@node Increment a number
@section Increment a Number

This script is one of a few that demonstrate how to do arithmetic
in @command{sed}.  This is indeed possible,@footnote{@command{sed} guru Greg
Ubben wrote an implementation of the @command{dc} @sc{rpn} calculator!
It is distributed together with sed.} but must be done manually.

To increment one number you just add 1 to last digit, replacing
it by the following digit.  There is one exception: when the digit
is a nine the previous digits must be also incremented until you
don't have a nine.

This solution by Bruno Haible is very clever and smart because
it uses a single buffer; if you don't have this limitation, the
algorithm used in @ref{cat -n, Numbering lines}, is faster.
It works by replacing trailing nines with an underscore, then
using multiple @code{s} commands to increment the last digit,
and then again substituting underscores with zeros.

@c start-------------------------------------------
@example
#!/usr/bin/sed -f

/[^0-9]/ d

# replace all trailing 9s by _ (any other character except digits, could
# be used)
:d
s/9\(_*\)$/_\1/
td

# incr last digit only.  The first line adds a most-significant
# digit of 1 if we have to add a digit.

s/^\(_*\)$/1\1/; tn
s/8\(_*\)$/9\1/; tn
s/7\(_*\)$/8\1/; tn
s/6\(_*\)$/7\1/; tn
s/5\(_*\)$/6\1/; tn
s/4\(_*\)$/5\1/; tn
s/3\(_*\)$/4\1/; tn
s/2\(_*\)$/3\1/; tn
s/1\(_*\)$/2\1/; tn
s/0\(_*\)$/1\1/; tn

:n
y/_/0/
@end example
@c end---------------------------------------------

@node Rename files to lower case
@section Rename Files to Lower Case

This is a pretty strange use of @command{sed}.  We transform text, and
transform it to be shell commands, then just feed them to shell.
Don't worry, even worse hacks are done when using @command{sed}; I have
seen a script converting the output of @command{date} into a @command{bc}
program!

The main body of this is the @command{sed} script, which remaps the name
from lower to upper (or vice-versa) and even checks out 
if the remapped name is the same as the original name.
Note how the script is parameterized using shell
variables and proper quoting.

@c start-------------------------------------------
@example
#! /bin/sh
# rename files to lower/upper case... 
#
# usage: 
#    move-to-lower * 
#    move-to-upper * 
# or
#    move-to-lower -R .
#    move-to-upper -R .
#

help()
@{
        cat << eof
Usage: $0 [-n] [-r] [-h] files...

-n      do nothing, only see what would be done
-R      recursive (use find)
-h      this message
files   files to remap to lower case

Examples:
       $0 -n *        (see if everything is ok, then...)
       $0 *

       $0 -R .

eof
@}

apply_cmd='sh'
finder='echo "$@@" | tr " " "\n"'
files_only=

while :
do
    case "$1" in 
        -n) apply_cmd='cat' ;;
        -R) finder='find "$@@" -type f';;
        -h) help ; exit 1 ;;
        *) break ;;
    esac
    shift
done

if [ -z "$1" ]; then
        echo Usage: $0 [-h] [-n] [-r] files...
        exit 1
fi

LOWER='abcdefghijklmnopqrstuvwxyz'
UPPER='ABCDEFGHIJKLMNOPQRSTUVWXYZ'

case `basename $0` in
        *upper*) TO=$UPPER; FROM=$LOWER ;;
        *)       FROM=$UPPER; TO=$LOWER ;;
esac

eval $finder | sed -n '

# remove all trailing slashes
s/\/*$//

# add ./ if there is no path, only a filename
/\//! s/^/.\//

# save path+filename
h

# remove path
s/.*\///

# do conversion only on filename
y/'$FROM'/'$TO'/

# now line contains original path+file, while
# hold space contains the new filename
x

# add converted file name to line, which now contains
# path/file-name\nconverted-file-name
G

# check if converted file name is equal to original file name,
# if it is, do not print anything
/^.*\/\(.*\)\n\1/b

# escape special characters for the shell
s/["$`\\]/\\&/g

# now, transform path/fromfile\n, into
# mv path/fromfile path/tofile and print it
s/^\(.*\/\)\(.*\)\n\(.*\)$/mv "\1\2" "\1\3"/p

' | $apply_cmd
@end example
@c end---------------------------------------------

@node Print bash environment
@section Print @command{bash} Environment

This script strips the definition of the shell functions
from the output of the @command{set} Bourne-shell command.

@c start-------------------------------------------
@example
#!/bin/sh

set | sed -n '
:x

@ifinfo
# if no occurrence of "=()" print and load next line
@end ifinfo
@ifnotinfo
# if no occurrence of @samp{=()} print and load next line
@end ifnotinfo
/=()/! @{ p; b; @}
/ () $/! @{ p; b; @}

# possible start of functions section
# save the line in case this is a var like FOO="() "
h

# if the next line has a brace, we quit because
# nothing comes after functions
n
/^@{/ q

# print the old line
x; p

# work on the new line now
x; bx
'
@end example
@c end---------------------------------------------

@node Reverse chars of lines
@section Reverse Characters of Lines

This script can be used to reverse the position of characters
in lines.  The technique moves two characters at a time, hence
it is faster than more intuitive implementations.

Note the @code{tx} command before the definition of the label.
This is often needed to reset the flag that is tested by
the @code{t} command.

Imaginative readers will find uses for this script.  An example
is reversing the output of @command{banner}.@footnote{This requires
another script to pad the output of banner; for example

@example
#! /bin/sh

banner -w $1 $2 $3 $4 |
  sed -e :a -e '/^.\@{0,'$1'\@}$/ @{ s/$/ /; ba; @}' |
  ~/sedscripts/reverseline.sed
@end example
}

@c start-------------------------------------------
@example
#!/usr/bin/sed -f

/../! b

# Reverse a line.  Begin embedding the line between two newlines
s/^.*$/\
&\
/

# Move first character at the end.  The regexp matches until
# there are zero or one characters between the markers
tx
:x
s/\(\n.\)\(.*\)\(.\n\)/\3\2\1/
tx

# Remove the newline markers
s/\n//g
@end example
@c end---------------------------------------------

@node tac
@section Reverse Lines of Files

This one begins a series of totally useless (yet interesting)
scripts emulating various Unix commands.  This, in particular,
is a @command{tac} workalike.

Note that on implementations other than @acronym{GNU} @command{sed}
@ifset PERL
and @value{SSED}
@end ifset
this script might easily overflow internal buffers.

@c start-------------------------------------------
@example
#!/usr/bin/sed -nf

# reverse all lines of input, i.e. first line became last, ...

# from the second line, the buffer (which contains all previous lines)
# is *appended* to current line, so, the order will be reversed
1! G

# on the last line we're done -- print everything
$ p

# store everything on the buffer again
h
@end example
@c end---------------------------------------------

@node cat -n
@section Numbering Lines

This script replaces @samp{cat -n}; in fact it formats its output
exactly like @acronym{GNU} @command{cat} does.

Of course this is completely useless and for two reasons:  first,
because somebody else did it in C, second, because the following
Bourne-shell script could be used for the same purpose and would
be much faster:

@c start-------------------------------------------
@example
#! /bin/sh
sed -e "=" $@@ | sed -e '
  s/^/      /
  N
  s/^ *\(......\)\n/\1  /
'
@end example
@c end---------------------------------------------

It uses @command{sed} to print the line number, then groups lines two
by two using @code{N}.  Of course, this script does not teach as much as
the one presented below.

The algorithm used for incrementing uses both buffers, so the line
is printed as soon as possible and then discarded.  The number
is split so that changing digits go in a buffer and unchanged ones go
in the other; the changed digits are modified in a single step
(using a @code{y} command).  The line number for the next line
is then composed and stored in the hold space, to be used in the
next iteration.

@c start-------------------------------------------
@example
#!/usr/bin/sed -nf

# Prime the pump on the first line
x
/^$/ s/^.*$/1/

# Add the correct line number before the pattern
G
h

# Format it and print it
s/^/      /
s/^ *\(......\)\n/\1  /p

# Get the line number from hold space; add a zero
# if we're going to add a digit on the next line
g
s/\n.*$//
/^9*$/ s/^/0/

# separate changing/unchanged digits with an x
s/.9*$/x&/

# keep changing digits in hold space
h
s/^.*x//
y/0123456789/1234567890/
x

# keep unchanged digits in pattern space
s/x.*$//

# compose the new number, remove the newline implicitly added by G
G
s/\n//
h
@end example
@c end---------------------------------------------

@node cat -b
@section Numbering Non-blank Lines

Emulating @samp{cat -b} is almost the same as @samp{cat -n}---we only
have to select which lines are to be numbered and which are not.

The part that is common to this script and the previous one is
not commented to show how important it is to comment @command{sed}
scripts properly...

@c start-------------------------------------------
@example
#!/usr/bin/sed -nf

/^$/ @{
  p
  b
@}

# Same as cat -n from now
x
/^$/ s/^.*$/1/
G
h
s/^/      /
s/^ *\(......\)\n/\1  /p
x
s/\n.*$//
/^9*$/ s/^/0/
s/.9*$/x&/
h
s/^.*x//
y/0123456789/1234567890/
x
s/x.*$//
G
s/\n//
h
@end example
@c end---------------------------------------------

@node wc -c
@section Counting Characters

This script shows another way to do arithmetic with @command{sed}.
In this case we have to add possibly large numbers, so implementing
this by successive increments would not be feasible (and possibly
even more complicated to contrive than this script).

The approach is to map numbers to letters, kind of an abacus
implemented with @command{sed}.  @samp{a}s are units, @samp{b}s are
tens and so on: we simply add the number of characters
on the current line as units, and then propagate the carry
to tens, hundreds, and so on.

As usual, running totals are kept in hold space.

On the last line, we convert the abacus form back to decimal.
For the sake of variety, this is done with a loop rather than
with some 80 @code{s} commands@footnote{Some implementations
have a limit of 199 commands per script}: first we
convert units, removing @samp{a}s from the number; then we
rotate letters so that tens become @samp{a}s, and so on
until no more letters remain.

@c start-------------------------------------------
@example
#!/usr/bin/sed -nf

# Add n+1 a's to hold space (+1 is for the newline)
s/./a/g
H
x
s/\n/a/

# Do the carry.  The t's and b's are not necessary,
# but they do speed up the thing
t a
: a;  s/aaaaaaaaaa/b/g; t b; b done
: b;  s/bbbbbbbbbb/c/g; t c; b done
: c;  s/cccccccccc/d/g; t d; b done
: d;  s/dddddddddd/e/g; t e; b done
: e;  s/eeeeeeeeee/f/g; t f; b done
: f;  s/ffffffffff/g/g; t g; b done
: g;  s/gggggggggg/h/g; t h; b done
: h;  s/hhhhhhhhhh//g

: done
$! @{
  h
  b
@}

# On the last line, convert back to decimal

: loop
/a/! s/[b-h]*/&0/
s/aaaaaaaaa/9/
s/aaaaaaaa/8/
s/aaaaaaa/7/
s/aaaaaa/6/
s/aaaaa/5/
s/aaaa/4/
s/aaa/3/
s/aa/2/
s/a/1/

: next
y/bcdefgh/abcdefg/
/[a-h]/ b loop
p
@end example
@c end---------------------------------------------

@node wc -w
@section Counting Words

This script is almost the same as the previous one, once each
of the words on the line is converted to a single @samp{a}
(in the previous script each letter was changed to an @samp{a}).

It is interesting that real @command{wc} programs have optimized
loops for @samp{wc -c}, so they are much slower at counting
words rather than characters.  This script's bottleneck,
instead, is arithmetic, and hence the word-counting one
is faster (it has to manage smaller numbers).

Again, the common parts are not commented to show the importance
of commenting @command{sed} scripts.

@c start-------------------------------------------
@example
#!/usr/bin/sed -nf

# Convert words to a's
s/[ @kbd{tab}][ @kbd{tab}]*/ /g
s/^/ /
s/ [^ ][^ ]*/a /g
s/ //g

# Append them to hold space
H
x
s/\n//

# From here on it is the same as in wc -c.
/aaaaaaaaaa/! bx;   s/aaaaaaaaaa/b/g
/bbbbbbbbbb/! bx;   s/bbbbbbbbbb/c/g
/cccccccccc/! bx;   s/cccccccccc/d/g
/dddddddddd/! bx;   s/dddddddddd/e/g
/eeeeeeeeee/! bx;   s/eeeeeeeeee/f/g
/ffffffffff/! bx;   s/ffffffffff/g/g
/gggggggggg/! bx;   s/gggggggggg/h/g
s/hhhhhhhhhh//g
:x
$! @{ h; b; @}
:y
/a/! s/[b-h]*/&0/
s/aaaaaaaaa/9/
s/aaaaaaaa/8/
s/aaaaaaa/7/
s/aaaaaa/6/
s/aaaaa/5/
s/aaaa/4/
s/aaa/3/
s/aa/2/
s/a/1/
y/bcdefgh/abcdefg/
/[a-h]/ by
p
@end example
@c end---------------------------------------------

@node wc -l
@section Counting Lines

No strange things are done now, because @command{sed} gives us
@samp{wc -l} functionality for free!!! Look:

@c start-------------------------------------------
@example
#!/usr/bin/sed -nf
$=
@end example
@c end---------------------------------------------

@node head
@section Printing the First Lines

This script is probably the simplest useful @command{sed} script.
It displays the first 10 lines of input; the number of displayed
lines is right before the @code{q} command.

@c start-------------------------------------------
@example
#!/usr/bin/sed -f
10q
@end example
@c end---------------------------------------------

@node tail
@section Printing the Last Lines

Printing the last @var{n} lines rather than the first is more complex
but indeed possible.  @var{n} is encoded in the second line, before
the bang character.

This script is similar to the @command{tac} script in that it keeps the
final output in the hold space and prints it at the end:

@c start-------------------------------------------
@example
#!/usr/bin/sed -nf

1! @{; H; g; @}
1,10 !s/[^\n]*\n//
$p
h
@end example
@c end---------------------------------------------

Mainly, the scripts keeps a window of 10 lines and slides it
by adding a line and deleting the oldest (the substitution command
on the second line works like a @code{D} command but does not
restart the loop).

The ``sliding window'' technique is a very powerful way to write
efficient and complex @command{sed} scripts, because commands like
@code{P} would require a lot of work if implemented manually.

To introduce the technique, which is fully demonstrated in the
rest of this chapter and is based on the @code{N}, @code{P}
and @code{D} commands, here is an implementation of @command{tail}
using a simple ``sliding window.''

This looks complicated but in fact the working is the same as
the last script: after we have kicked in the appropriate number
of lines, however, we stop using the hold space to keep inter-line
state, and instead use @code{N} and @code{D} to slide pattern
space by one line:

@c start-------------------------------------------
@example
#!/usr/bin/sed -f

1h
2,10 @{; H; g; @}
$q
1,9d
N
D
@end example
@c end---------------------------------------------

Note how the first, second and fourth line are inactive after
the first ten lines of input.  After that, all the script does
is: exiting on the last line of input, appending the next input
line to pattern space, and removing the first line.

@node uniq
@section Make Duplicate Lines Unique

This is an example of the art of using the @code{N}, @code{P}
and @code{D} commands, probably the most difficult to master.

@c start-------------------------------------------
@example
#!/usr/bin/sed -f
h

:b
# On the last line, print and exit
$b
N
/^\(.*\)\n\1$/ @{
    # The two lines are identical.  Undo the effect of
    # the n command.
    g
    bb
@}

# If the @code{N} command had added the last line, print and exit
$b

# The lines are different; print the first and go
# back working on the second.
P
D
@end example
@c end---------------------------------------------

As you can see, we mantain a 2-line window using @code{P} and @code{D}.
This technique is often used in advanced @command{sed} scripts.

@node uniq -d
@section Print Duplicated Lines of Input

This script prints only duplicated lines, like @samp{uniq -d}.

@c start-------------------------------------------
@example
#!/usr/bin/sed -nf

$b
N
/^\(.*\)\n\1$/ @{
    # Print the first of the duplicated lines
    s/.*\n//
    p

    # Loop until we get a different line
    :b
    $b
    N
    /^\(.*\)\n\1$/ @{
        s/.*\n//
        bb
    @}
@}

# The last line cannot be followed by duplicates
$b

# Found a different one.  Leave it alone in the pattern space
# and go back to the top, hunting its duplicates
D
@end example
@c end---------------------------------------------

@node uniq -u
@section Remove All Duplicated Lines

This script prints only unique lines, like @samp{uniq -u}.

@c start-------------------------------------------
@example
#!/usr/bin/sed -f

# Search for a duplicate line --- until that, print what you find.
$b
N
/^\(.*\)\n\1$/ ! @{
    P
    D
@}

:c
# Got two equal lines in pattern space.  At the
# end of the file we simply exit
$d

# Else, we keep reading lines with @code{N} until we
# find a different one
s/.*\n//
N
/^\(.*\)\n\1$/ @{
    bc
@}

# Remove the last instance of the duplicate line
# and go back to the top
D
@end example
@c end---------------------------------------------

@node cat -s
@section Squeezing Blank Lines

As a final example, here are three scripts, of increasing complexity
and speed, that implement the same function as @samp{cat -s}, that is
squeezing blank lines.

The first leaves a blank line at the beginning and end if there are
some already.

@c start-------------------------------------------
@example
#!/usr/bin/sed -f

# on empty lines, join with next
# Note there is a star in the regexp
:x
/^\n*$/ @{
N
bx
@}

# now, squeeze all '\n', this can be also done by:
# s/^\(\n\)*/\1/
s/\n*/\
/
@end example
@c end---------------------------------------------

This one is a bit more complex and removes all empty lines
at the beginning.  It does leave a single blank line at end
if one was there.

@c start-------------------------------------------
@example
#!/usr/bin/sed -f

# delete all leading empty lines
1,/^./@{
/./!d
@}

# on an empty line we remove it and all the following
# empty lines, but one
:x
/./!@{
N
s/^\n$//
tx
@}
@end example
@c end---------------------------------------------

This removes leading and trailing blank lines.  It is also the
fastest.  Note that loops are completely done with @code{n} and
@code{b}, without relying on @command{sed} to restart the
the script automatically at the end of a line.

@c start-------------------------------------------
@example
#!/usr/bin/sed -nf

# delete all (leading) blanks
/./!d

# get here: so there is a non empty
:x
# print it
p
# get next
n
# got chars? print it again, etc... 
/./bx

# no, don't have chars: got an empty line
:z
# get next, if last line we finish here so no trailing
# empty lines are written
n
# also empty? then ignore it, and get next... this will
# remove ALL empty lines
/./!bz

# all empty lines were deleted/ignored, but we have a non empty.  As
# what we want to do is to squeeze, insert a blank line artificially
i\

bx
@end example
@c end---------------------------------------------

@node Limitations
@chapter @value{SSED}'s Limitations and Non-limitations

@cindex @acronym{GNU} extensions, unlimited line length
@cindex Portability, line length limitations
For those who want to write portable @command{sed} scripts,
be aware that some implementations have been known to
limit line lengths (for the pattern and hold spaces)
to be no more than 4000 bytes.
The @sc{posix} standard specifies that conforming @command{sed}
implementations shall support at least 8192 byte line lengths.
@value{SSED} has no built-in limit on line length;
as long as it can @code{malloc()} more (virtual) memory,
you can feed or construct lines as long as you like.

However, recursion is used to handle subpatterns and indefinite
repetition.  This means that the available stack space may limit
the size of the buffer that can be processed by certain patterns.

@ifset PERL
There are some size limitations in the regular expression
matcher but it is hoped that they will never in practice
be relevant.  The maximum length of a compiled pattern
is 65539 (sic) bytes.  All values in repeating quantifiers
must be less than 65536.  The maximum nesting depth of
all parenthesized subpatterns, including capturing and
non-capturing subpatterns@footnote{The
distinction is meaningful when referring to Perl-style
regular expressions.}, assertions, and other types of
subpattern, is 200.

Also, @value{SSED} recognizes the @sc{posix} syntax
@code{[.@var{ch}.]} and @code{[=@var{ch}=]}
where @var{ch} is a ``collating element'', but these
are not supported, and an error is given if they are
encountered.

Here are a few distinctions between the real Perl-style
regular expressions and those that @option{-R} recognizes.

@enumerate
@item
Lookahead assertions do not allow repeat quantifiers after them
Perl permits them, but they do not mean what you
might think. For example, @samp{(?!a)@{3@}} does not assert that the
next three characters are not @samp{a}. It just asserts three times that the
next character is not @samp{a} --- a waste of time and nothing else.

@item
Capturing subpatterns that occur inside  negative  lookahead
head  assertions  are  counted,  but  their  entries are counted
as empty in the second half of an @code{s} command.
Perl sets its numerical variables from any such patterns
that are matched before the assertion fails to match
something (thereby succeeding), but only if the negative
lookahead assertion contains just one branch.

@item
The following Perl escape sequences are not supported:
@samp{\l}, @samp{\u}, @samp{\L}, @samp{\U}, @samp{\E},
@samp{\Q}. In fact these are implemented by Perl's general
string-handling and are not part of its pattern matching engine.

@item
The Perl @samp{\G} assertion is not supported as it is not
relevant to single pattern matches.

@item
Fairly obviously, @value{SSED} does not support the @samp{(?@{code@})}
and @samp{(?p@{code@})} constructions. However, there is some experimental
support for recursive patterns using the non-Perl item @samp{(?R)}.

@item
There are at the time of writing some oddities in Perl
5.005_02 concerned with the settings of captured strings
when part of a pattern is repeated. For example, matching
@samp{aba} against the pattern @samp{/^(a(b)?)+$/} sets
@samp{$2}@footnote{@samp{$2} would be @samp{\2} in @value{SSED}.}
to the value @samp{b}, but matching @samp{aabbaa}
against @samp{/^(aa(bb)?)+$/} leaves @samp{$2}
unset.  However, if the pattern is changed to
@samp{/^(aa(b(b))?)+$/} then @samp{$2} (and @samp{$3}) are set.
In Perl 5.004 @samp{$2} is set in both cases, and that is also
true of @value{SSED}.

@item
Another as yet unresolved discrepancy is that in Perl
5.005_02 the pattern @samp{/^(a)?(?(1)a|b)+$/} matches
the string @samp{a}, whereas in @value{SSED} it does not.
However, in both Perl and @value{SSED} @samp{/^(a)?a/} matched
against @samp{a} leaves $1 unset.
@end enumerate
@end ifset

@node Other Resources
@chapter Other Resources for Learning About @command{sed}

@cindex Additional reading about @command{sed}
In addition to several books that have been written about @command{sed}
(either specifically or as chapters in books which discuss
shell programming), one can find out more about @command{sed}
(including suggestions of a few books) from the FAQ
for the @code{sed-users} mailing list, available from:
@display
@uref{http://sed.sourceforge.net/sedfaq.html}
@end display

Also of interest are
@uref{http://www.student.northpark.edu/pemente/sed/index.htm}
and @uref{http://sed.sf.net/grabbag},
which include @command{sed} tutorials and other @command{sed}-related goodies.

The @code{sed-users} mailing list itself maintained by Sven Guckes.
To subscribe, visit @uref{http://groups.yahoo.com} and search
for the @code{sed-users} mailing list.

@node Reporting Bugs
@chapter Reporting Bugs

@cindex Bugs, reporting
Email bug reports to @email{bonzini@@gnu.org}.
Be sure to include the word ``sed'' somewhere in the @code{Subject:} field.
Also, please include the output of @samp{sed --version} in the body
of your report if at all possible.

Please do not send a bug report like this:

@example
@i{@i{@r{while building frobme-1.3.4}}}
$ configure 
@error{} sed: file sedscr line 1: Unknown option to 's'
@end example

If @value{SSED} doesn't configure your favorite package, take a
few extra minutes to identify the specific problem and make a stand-alone
test case.  Unlike other programs such as C compilers, making such test
cases for @command{sed} is quite simple.

A stand-alone test case includes all the data necessary to perform the
test, and the specific invocation of @command{sed} that causes the problem.
The smaller a stand-alone test case is, the better.  A test case should
not involve something as far removed from @command{sed} as ``try to configure
frobme-1.3.4''.  Yes, that is in principle enough information to look
for the bug, but that is not a very practical prospect.

Here are a few commonly reported bugs that are not bugs.

@table @asis
@item @code{N} command on the last line
@cindex Portability, @code{N} command on the last line
@cindex Non-bugs, @code{N} command on the last line

Most versions of @command{sed} exit without printing anything when
the @command{N} command is issued on the last line of a file.
@value{SSED} prints pattern space before exiting unless of course
the @command{-n} command switch has been specified.  This choice is
by design.

For example, the behavior of
@example
sed N foo bar
@end example
@noindent
would depend on whether foo has an even or an odd number of
lines@footnote{which is the actual ``bug'' that prompted the
change in behavior}.  Or, when writing a script to read the
next few lines following a pattern match, traditional
implementations of @code{sed} would force you to write
something like
@example
/foo/@{ $!N; $!N; $!N; $!N; $!N; $!N; $!N; $!N; $!N @}
@end example
@noindent
instead of just
@example
/foo/@{ N;N;N;N;N;N;N;N;N; @}
@end example

@cindex @code{POSIXLY_CORRECT} behavior, @code{N} command
In any case, the simplest workaround is to use @code{$d;N} in
scripts that rely on the traditional behavior, or to set
the @code{POSIXLY_CORRECT} variable to a non-empty value.

@item Regex syntax clashes (problems with backslashes)
@cindex @acronym{GNU} extensions, to basic regular expressions
@cindex Non-bugs, regex syntax clashes
@command{sed} uses the @sc{posix} basic regular expression syntax.  According to
the standard, the meaning of some escape sequences is undefined in
this syntax;  notable in the case of @command{sed} are @code{\|},
@code{\+}, @code{\?}, @code{\`}, @code{\'}, @code{\<},
@code{\>}, @code{\b}, @code{\B}, @code{\w}, and @code{\W}.

As in all @acronym{GNU} programs that use @sc{posix} basic regular
expressions, @command{sed} interprets these escape sequences as special
characters.  So, @code{x\+} matches one or more occurrences of @samp{x}.
@code{abc\|def} matches either @samp{abc} or @samp{def}.

This syntax may cause problems when running scripts written for other
@command{sed}s.  Some @command{sed} programs have been written with the
assumption that @code{\|} and @code{\+} match the literal characters
@code{|} and @code{+}.  Such scripts must be modified by removing the
spurious backslashes if they are to be used with modern implementations
of @command{sed}, like
@ifset PERL
@value{SSED} or
@end ifset
@acronym{GNU} @command{sed}.

On the other hand, some scripts use s|abc\|def||g to remove occurrences
of @emph{either} @code{abc} or @code{def}.  While this worked until
@command{sed} 4.0.x, newer versions interpret this as removing the
string @code{abc|def}.  This is again undefined behavior according to
@acronym{POSIX}, and this interpretation is arguably more robust: older
@command{sed}s, for example, required that the regex matcher parsed
@code{\/} as @code{/} in the common case of escaping a slash, which is
again undefined behavior; the new behavior avoids this, and this is good
because the regex matcher is only partially under our control.

@cindex @acronym{GNU} extensions, special escapes
In addition, this version of @command{sed} supports several escape characters
(some of which are multi-character) to insert non-printable characters
in scripts (@code{\a}, @code{\c}, @code{\d}, @code{\o}, @code{\r},
@code{\t}, @code{\v}, @code{\x}).  These can cause similar problems
with scripts written for other @command{sed}s.

@item @option{-i} clobbers read-only files
@cindex In-place editing
@cindex @value{SSEDEXT}, in-place editing
@cindex Non-bugs, in-place editing

In short, @samp{sed -i} will let you delete the contents of
a read-only file, and in general the @option{-i} option
(@pxref{Invoking sed, , Invocation}) lets you clobber
protected files.  This is not a bug, but rather a consequence
of how the Unix filesystem works.

The permissions on a file say what can happen to the data
in that file, while the permissions on a directory say what can
happen to the list of files in that directory.  @samp{sed -i}
will not ever open for writing  a file that is already on disk.
Rather, it will work on a temporary file that is finally renamed
to the original name: if you rename or delete files, you're actually
modifying the contents of the directory, so the operation depends on
the permissions of the directory, not of the file.  For this same
reason, @command{sed} does not let you use @option{-i} on a writeable file
in a read-only directory, and will break hard or symbolic links when
@option{-i} is used on such a file.

@item @code{0a} does not work (gives an error)
@cindex @code{0} address
@cindex @acronym{GNU} extensions, @code{0} address
@cindex Non-bugs, @code{0} address

There is no line 0.  0 is a special address that is only used to treat
addresses like @code{0,/@var{RE}/} as active when the script starts: if
you write @code{1,/abc/d} and the first line includes the word @samp{abc},
then that match would be ignored because address ranges must span at least
two lines (barring the end of the file); but what you probably wanted is
to delete every line up to the first one including @samp{abc}, and this
is obtained with @code{0,/abc/d}.

@ifclear PERL
@item @code{[a-z]} is case insensitive
@cindex Non-bugs, localization-related

You are encountering problems with locales.  POSIX mandates that @code{[a-z]}
uses the current locale's collation order -- in C parlance, that means using
@code{strcoll(3)} instead of @code{strcmp(3)}.  Some locales have a
case-insensitive collation order, others don't: one of those that have
problems is Estonian.

Another problem is that @code{[a-z]} tries to use collation symbols.
This only happens if you are on the @acronym{GNU} system, using
@acronym{GNU} libc's regular expression matcher instead of compiling the
one supplied with @acronym{GNU} sed.  In a Danish locale, for example,
the regular expression @code{^[a-z]$} matches the string @samp{aa},
because this is a single collating symbol that comes after @samp{a}
and before @samp{b}; @samp{ll} behaves similarly in Spanish
locales, or @samp{ij} in Dutch locales.

To work around these problems, which may cause bugs in shell scripts, set
the @env{LC_COLLATE} and @env{LC_CTYPE} environment variables to @samp{C}.

@item @code{s/.*//} does not clear pattern space
@cindex Non-bugs, localization-related
@cindex @value{SSEDEXT}, emptying pattern space
@cindex Emptying pattern space

This happens if your input stream includes invalid multibyte
sequences.  @sc{posix} mandates that such sequences
are @emph{not} matched by @samp{.}, so that @samp{s/.*//} will not clear
pattern space as you would expect.  In fact, there is no way to clear
sed's buffers in the middle of the script in most multibyte locales
(including UTF-8 locales).  For this reason, @value{SSED} provides a `z'
command (for `zap') as an extension.

To work around these problems, which may cause bugs in shell scripts, set
the @env{LC_COLLATE} and @env{LC_CTYPE} environment variables to @samp{C}.
@end ifclear
@end table


@node Extended regexps
@appendix Extended regular expressions
@cindex Extended regular expressions, syntax

The only difference between basic and extended regular expressions is in
the behavior of a few characters: @samp{?}, @samp{+}, parentheses,
braces (@samp{@{@}}), and @samp{|}.  While basic regular expressions
require these to be escaped if you want them to behave as special
characters, when using extended regular expressions you must escape
them if you want them @emph{to match a literal character}.  @samp{|}
is special here because @samp{\|} is a GNU extension -- standard
basic regular expressions do not provide its functionality.

@noindent
Examples:
@table @code
@item abc?
becomes @samp{abc\?} when using extended regular expressions.  It matches
the literal string @samp{abc?}.

@item c\+
becomes @samp{c+} when using extended regular expressions.  It matches
one or more @samp{c}s.

@item a\@{3,\@}
becomes @samp{a@{3,@}} when using extended regular expressions.  It matches
three or more @samp{a}s.

@item \(abc\)\@{2,3\@}
becomes @samp{(abc)@{2,3@}} when using extended regular expressions.  It
matches either @samp{abcabc} or @samp{abcabcabc}.

@item \(abc*\)\1
becomes @samp{(abc*)\1} when using extended regular expressions.
Backreferences must still be escaped when using extended regular
expressions.
@end table

@ifset PERL
@node Perl regexps
@appendix Perl-style regular expressions
@cindex Perl-style regular expressions, syntax

@emph{This part is taken from the @file{pcre.txt} file distributed together
with the free @sc{pcre} regular expression matcher; it was written by Philip Hazel.}

Perl introduced several extensions to regular expressions, some
of them incompatible with the syntax of regular expressions
accepted by Emacs and other @acronym{GNU} tools (whose matcher was
based on the Emacs matcher).  @value{SSED} implements
both kinds of extensions.

@iftex
Summarizing, we have:

@itemize @bullet
@item
A backslash can introduce several special sequences

@item
The circumflex, dollar sign, and period characters behave specially 
with regard to new lines

@item
Strange uses of square brackets are parsed differently

@item
You can toggle modifiers in the middle of a regular expression

@item
You can specify that a subpattern does not count when numbering backreferences

@item
@cindex Greedy regular expression matching
You can specify greedy or non-greedy matching

@item
You can have more than ten back references

@item
You can do complex look aheads and look behinds (in the spirit of
@code{\b}, but with subpatterns).

@item
You can often improve performance by avoiding that @command{sed} wastes
time with backtracking

@item
You can have if/then/else branches

@item
You can do recursive matches, for example to look for unbalanced parentheses

@item
You can have comments and non-significant whitespace, because things can
get complex...
@end itemize

Most of these extensions are introduced by the special @code{(?}
sequence, which gives special meanings to parenthesized groups.
@end iftex
@menu
Other extensions can be roughly subdivided in two categories
On one hand Perl introduces several more escaped sequences
(that is, sequences introduced by a backslash).  On the other
hand, it specifies that if a question mark follows an open
parentheses it should give a special meaning to the parenthesized
group.

* Backslash::                       Introduces special sequences
* Circumflex/dollar sign/period::   Behave specially with regard to new lines
* Square brackets::                 Are a bit different in strange cases
* Options setting::                 Toggle modifiers in the middle of a regexp
* Non-capturing subpatterns::       Are not counted when backreferencing
* Repetition::                      Allows for non-greedy matching
* Backreferences::                  Allows for more than 10 back references
* Assertions::                      Allows for complex look ahead matches
* Non-backtracking subpatterns::    Often gives more performance
* Conditional subpatterns::         Allows if/then/else branches
* Recursive patterns::              For example to match parentheses
* Comments::                        Because things can get complex...
@end menu

@node Backslash
@appendixsec Backslash
@cindex Perl-style regular expressions, escaped sequences

There are a few difference in the handling of backslashed 
sequences in Perl mode.

First of all, there are no @code{\o} and @code{\d} sequences.
@sc{ascii} values for characters can be specified in octal
with a @code{\@var{xxx}} sequence, where @var{xxx} is a
sequence of up to three octal digits.  If the first digit
is a zero, the treatment of the sequence is straightforward;
just note that if the character that follows the escaped digit
is itself an octal digit, you have to supply three octal digits
for @var{xxx}.  For example @code{\07} is a @sc{bel} character
rather than a @sc{nul} and a literal @code{7} (this sequence is
instead represented by @code{\0007}).

@cindex Perl-style regular expressions, backreferences
The handling of a backslash followed by a digit other than 0
is complicated.  Outside a character class, @command{sed} reads it
and any following digits as a decimal number. If the number
is less than 10, or if there have been at least that many
previous capturing left parentheses in the expression, the
entire sequence is taken as a back reference. A description
of how this works is given later, following the discussion
of parenthesized subpatterns.

Inside a character class, or if the decimal number is
greater than 9 and there have not been that many capturing
subpatterns, @command{sed} re-reads up to three octal digits following
the backslash, and generates a single byte from the
least significant 8 bits of the value. Any subsequent digits
stand for themselves.  For example:

@example
\040  @i{@r{is another way of writing a space}}
\40   @i{@r{is the same, provided there are fewer than 40}}
      @i{@r{previous capturing subpatterns}}
\7    @i{@r{is always a back reference}}
\011  @i{@r{is always a tab}}
\11   @i{@r{might be a back reference, or another way of writing a tab}}
\0113 @i{@r{is a tab followed by the character @samp{3}}}
\113  @i{@r{is the character with octal code 113 (since there}}
      @i{@r{can be no more than 99 back references)}}
\377  @i{@r{is a byte consisting entirely of 1 bits (@sc{ascii} 255)}}
\81   @i{@r{is either a back reference, or a binary zero}}
      @i{@r{followed by the two characters @samp{81}}}
@end example

Note that octal values of 100 or greater must not be introduced
by a leading zero, because no more than three octal
digits are ever read. Note that this applies only to the LHS 
pattern; it is not possible yet to specify more than 9 backreferences 
on the RHS of the `s' command.

All the sequences that define a single byte value can be
used both inside and outside character classes. In addition,
inside a character class, the sequence @code{\b} is interpreted
as the backspace character (hex 08). Outside a character
class it has a different meaning (see below).

In addition, there are four additional escapes specifying
generic character classes (like @code{\w} and @code{\W} do):

@cindex Perl-style regular expressions, character classes
@table @samp
@item \d
Matches any decimal digit

@item \D
Matches any character that is not a decimal digit
@end table

In Perl mode, these character type sequences can appear both inside and
outside character classes. Instead, in @sc{posix} mode these sequences
(as well as @code{\w} and @code{\W}) are treated as two literal characters
(a backslash and a letter) inside square brackets.

Escaped sequences specifying assertions are also different in
Perl mode.  An assertion specifies a condition that has to be met
at a particular point in a match, without consuming any
characters from the subject string. The use of subpatterns
for more complicated assertions is described below.  The
backslashed assertions are

@cindex Perl-style regular expressions, assertions
@table @samp
@item \b
Asserts that the point is at a word boundary.
A word boundary is a position in the subject string where
the current character and the previous character do not both
match @code{\w} or @code{\W} (i.e. one matches @code{\w} and
the other matches @code{\W}), or the start or end of the string
if the first or last character matches @code{\w}, respectively.

@item \B
Asserts that the point is not at a word boundary.

@item \A
Asserts the matcher is at the start of pattern space (independent
of multiline mode).

@item \Z
Asserts the matcher is at the end of pattern space,
or at a newline before the end of pattern space (independent of
multiline mode)

@item \z
Asserts the matcher is at the end of pattern space (independent
of multiline mode)
@end table

These assertions may not appear in character classes (but
note that @code{\b} has a different meaning, namely the
backspace character, inside a character class).
Note that Perl mode does not support directly assertions
for the beginning and the end of word; the @acronym{GNU} extensions
@code{\<} and @code{\>} achieve this purpose in @sc{posix} mode
instead.

The @code{\A}, @code{\Z}, and @code{\z} assertions differ
from the traditional circumflex and dollar sign (described below)
in that they only ever match at the very start and end of the
subject string, whatever options are set; in particular @code{\A}
and @code{\z} are the same as the @acronym{GNU} extensions
@code{\`} and @code{\'} that are active in @sc{posix} mode.

@node Circumflex/dollar sign/period
@appendixsec Circumflex, dollar sign, period
@cindex Perl-style regular expressions, newlines

Outside a character class, in the default matching mode, the
circumflex character is an assertion which is true only if
the current matching point is at the start of the subject
string.  Inside a character class, the circumflex has an entirely
different meaning (see below).

The circumflex need not be the first character of the pattern if
a number of alternatives are involved, but it should be the
first thing in each alternative in which it appears if the
pattern is ever to match that branch. If all possible alternatives,
start with a circumflex, that is, if the pattern is
constrained to match only at the start of the subject, it is
said to be an @dfn{anchored} pattern. (There are also other constructs
structs that can cause a pattern to be anchored.)

A dollar sign is an assertion which is true only if the
current matching point is at the end of the subject string,
or immediately before a newline character that is the last
character in the string (by default).  A dollar sign need not be the
last character of the pattern if a number of alternatives
are involved, but it should be the last item in any branch
in which it appears.  A dollar sign has no special meaning in a
character class.

@cindex Perl-style regular expressions, multiline
The meanings of the circumflex and dollar sign characters are
changed if the @code{M} modifier option is used. When this is
the case, they match immediately after and immediately
before an internal @code{\n} character, respectively, in addition
to matching at the start and end of the subject string.  For
example, the pattern @code{/^abc$/} matches the subject string
@samp{def\nabc} in multiline mode, but not otherwise.  Consequently,
patterns that are anchored in single line mode
because all branches start with @code{^} are not anchored in
multiline mode.

@cindex Perl-style regular expressions, multiline
Note that the sequences @code{\A}, @code{\Z}, and @code{\z}
can be used to match the start and end of the subject in both
modes, and if all branches of a pattern start with @code{\A}
is it always anchored, whether the @code{M} modifier is set or not.

@cindex Perl-style regular expressions, single line
Outside a character class, a dot in the pattern matches any
one character in the subject, including a non-printing character,
but not (by default) newline.  If the @code{S} modifier is used,
dots match newlines as well.  Actually, the handling of
dot is entirely independent of the handling of circumflex
and dollar sign, the only relationship being that they both
involve newline characters. Dot has no special meaning in a
character class.

@node Square brackets
@appendixsec Square brackets
@cindex Perl-style regular expressions, character classes

An opening square bracket introduces a character class, terminated
by a closing square bracket.  A closing square bracket on its own
is not special.  If a closing square bracket is required as a
member of the class, it should be the first data character in
the class (after an initial circumflex, if present) or escaped with a backslash.

A character class matches a single character in the subject;
the character must be in the set of characters defined by
the class, unless the first character in the class is a circumflex,
in which case the subject character must not be in
the set defined by the class. If a circumflex is actually
required as a member of the class, ensure it is not the
first character, or escape it with a backslash.

For example, the character class [aeiou] matches any lower
case vowel, while [^aeiou] matches any character that is not
a lower case vowel. Note that a circumflex is just a convenient
venient notation for specifying the characters which are in
the class by enumerating those that are not. It is not an
assertion: it still consumes a character from the subject
string, and fails if the current pointer is at the end of
the string.

@cindex Perl-style regular expressions, case-insensitive
When caseless matching is set, any letters in a class
represent both their upper case and lower case versions, so
for example, a caseless @code{[aeiou]} matches uppercase
and lowercase @samp{A}s, and a caseless @code{[^aeiou]}
does not match @samp{A}, whereas a case-sensitive version would.

@cindex Perl-style regular expressions, single line
@cindex Perl-style regular expressions, multiline
The newline character is never treated in any special way in
character classes, whatever the setting of the @code{S} and
@code{M} options (modifiers) is.  A class such as @code{[^a]} will
always match a newline.

The minus (hyphen) character can be used to specify a range
of characters in a character class.  For example, @code{[d-m]}
matches any letter between d and m, inclusive.  If a minus
character is required in a class, it must be escaped with a
backslash or appear in a position where it cannot be interpreted
as indicating a range, typically as the first or last
character in the class.

It is not possible to have the literal character @code{]} as the
end character of a range.  A pattern such as @code{[W-]46]} is
interpreted as a class of two characters (@code{W} and @code{-})
followed by a literal string @code{46]}, so it would match
@samp{W46]} or @samp{-46]}. However, if the @code{]} is escaped
with a backslash it is interpreted as the end of range, so
@code{[W-\]46]} is interpreted as a single class containing a
range followed by two separate characters. The octal or
hexadecimal representation of @code{]} can also be used to end a range.

Ranges operate in @sc{ascii} collating sequence. They can also be
used for characters specified numerically, for example
@code{[\000-\037]}. If a range that includes letters is used when
caseless matching is set, it matches the letters in either
case. For example, a caseless @code{[W-c]} is equivalent to
@code{[][\^_`wxyzabc]}, matched caselessly, and if character
tables for the French locale are in use, @code{[\xc8-\xcb]}
matches accented E characters in both cases.

Unlike in @sc{posix} mode, the character types @code{\d},
@code{\D}, @code{\s}, @code{\S}, @code{\w}, and @code{\W}
may also appear in a character class, and add the characters
that they match to the class. For example, @code{[\dABCDEF]} matches any
hexadecimal digit.  A circumflex can conveniently be used
with the upper case character types to specify a more restricted
set of characters than the matching lower case type.
For example, the class @code{[^\W_]} matches any letter or digit,
but not underscore.

All non-alphameric characters other than @code{\}, @code{-},
@code{^} (at the start) and the terminating @code{]}
are non-special in character classes, but it does no harm
if they are escaped.

Perl 5.6 supports the @sc{posix} notation for character classes, which
uses names enclosed by @code{[:} and @code{:]} within the enclosing
square brackets, and @value{SSED} supports this notation as well.
For example,

@example
[01[:alpha:]%]
@end example

@noindent
matches @samp{0}, @samp{1}, any alphabetic character, or @samp{%}.
The supported class names are

@table @code
@item alnum
Matches letters and digits

@item alpha
Matches letters

@item ascii
Matches character codes 0 - 127

@item cntrl
Matches control characters

@item digit
Matches decimal digits (same as \d)

@item graph
Matches printing characters, excluding space

@item lower
Matches lower case letters

@item print
Matches printing characters, including space

@item punct
Matches printing characters, excluding letters and digits

@item space
Matches white space (same as \s)

@item upper
Matches upper case letters

@item word
Matches ``word'' characters (same as \w)

@item xdigit
Matches hexadecimal digits
@end table

The names @code{ascii} and @code{word} are extensions valid only in
Perl mode.  Another Perl extension is negation, which is
indicated by a circumflex character after the colon. For example,

@example
[12[:^digit:]]
@end example

@noindent
matches @samp{1}, @samp{2}, or any non-digit.

@node Options setting
@appendixsec Options setting
@cindex Perl-style regular expressions, toggling options
@cindex Perl-style regular expressions, case-insensitive
@cindex Perl-style regular expressions, multiline
@cindex Perl-style regular expressions, single line
@cindex Perl-style regular expressions, extended

The settings of the @code{I}, @code{M}, @code{S}, @code{X}
modifiers can be changed from within the pattern by
a sequence of Perl option letters enclosed between @code{(?}
and @code{)}. The option letters must be lowercase.

For example, @code{(?im)} sets caseless, multiline matching. It is
also possible to unset these options by preceding the letter
with a hyphen; you can also have combined settings and unsettings:
@code{(?im-sx)} sets caseless and multiline matching,
while unsets single line matching (for dots) and extended
whitespace interpretation.  If a letter appears both before
and after the hyphen, the option is unset.

The scope of these option changes depends on where in the
pattern the setting occurs. For settings that are outside
any subpattern (defined below), the effect is the same as if
the options were set or unset at the start of matching. The
following patterns all behave in exactly the same way:

@example
(?i)abc
a(?i)bc
ab(?i)c
abc(?i)
@end example

which in turn is the same as specifying the pattern abc with
the @code{I} modifier.  In other words, ``top level'' settings
apply to the whole pattern (unless there are other
changes inside subpatterns). If there is more than one setting
of the same option at top level, the rightmost setting
is used.

If an option change occurs inside a subpattern, the effect
is different.  This is a change of behaviour in Perl 5.005.
An option change inside a subpattern affects only that part
of the subpattern @emph{that follows} it, so

@example
(a(?i)b)c
@end example

@noindent
matches abc and aBc and no other  strings  (assuming
case-sensitive matching is used).  By this means, options can
be made to have different settings in different parts of the
pattern.  Any changes made in one alternative do carry on
into subsequent branches within the same subpattern.  For
example,

@example
(a(?i)b|c)
@end example

@noindent
matches @samp{ab}, @samp{aB}, @samp{c}, and @samp{C},
even though when matching @samp{C} the first branch is
abandoned before the option setting.
This is because the effects of option settings happen at
compile time. There would be some very weird behaviour otherwise.

@ignore
There are two PCRE-specific options PCRE_UNGREEDY and PCRE_EXTRA
that can be changed in the same way as the Perl-compatible options by
using the characters U and X respectively.  The (?X) flag
setting is special in that it must always occur earlier in
the pattern than any of the additional features it turns on,
even when it is at top level. It is best put at the start.
@end ignore


@node Non-capturing subpatterns
@appendixsec Non-capturing subpatterns
@cindex Perl-style regular expressions, non-capturing subpatterns

Marking part of a pattern as a subpattern does two things.
On one hand, it localizes a set of alternatives; on the other
hand, it sets up the subpattern as a capturing subpattern (as
defined above).  The subpattern can be backreferenced and
referenced in the right side of @code{s} commands.

For example, if the string @samp{the red king} is matched against
the pattern

@example
the ((red|white) (king|queen))
@end example

@noindent
the captured substrings are @samp{red king}, @samp{red},
and @samp{king}, and are numbered 1, 2, and 3.

The fact that plain parentheses fulfil two functions is not
always helpful.  There are often times when a grouping
subpattern is required without a capturing requirement.  If an
opening parenthesis is followed by @code{?:}, the subpattern does
not do any capturing, and is not counted when computing the
number of any subsequent capturing subpatterns. For example,
if the string @samp{the white queen} is matched against the pattern

@example
the ((?:red|white) (king|queen))
@end example

@noindent
the captured substrings are @samp{white queen} and @samp{queen},
and are numbered 1 and 2. The maximum number of captured
substrings is 99, while the maximum number of all subpatterns,
both capturing and non-capturing, is 200.

As a convenient shorthand, if any option settings are
equired at the start of a non-capturing subpattern, the
option letters may appear between the @code{?} and the
@code{:}.  Thus the two patterns

@example
(?i:saturday|sunday)
(?:(?i)saturday|sunday)
@end example

@noindent
match exactly the same set of strings.  Because alternative
branches are tried from left to right, and options are not
reset until the end of the subpattern is reached, an option
setting in one branch does affect subsequent branches, so
the above patterns match @samp{SUNDAY} as well as @samp{Saturday}.


@node Repetition
@appendixsec Repetition
@cindex Perl-style regular expressions, repetitions

Repetition is specified by quantifiers, which can follow any
of the following items:

@itemize @bullet
@item
a single character, possibly escaped

@item
the @code{.} special character

@item
a character class

@item
a back reference (see next section)

@item
a parenthesized subpattern (unless it is an assertion; @pxref{Assertions})
@end itemize

The general repetition quantifier specifies a minimum and
maximum number of permitted matches, by giving the two
numbers in curly brackets (braces), separated by a comma.
The numbers must be less than 65536, and the first must be
less than or equal to the second. For example:

@example
z@{2,4@}
@end example

@noindent
matches @samp{zz}, @samp{zzz}, or @samp{zzzz}. A closing brace on its own
is not a special character. If the second number is omitted,
but the comma is present, there is no upper limit; if the
second number and the comma are both omitted, the quantifier
specifies an exact number of required matches. Thus

@example
[aeiou]@{3,@}
@end example

@noindent
matches at least 3 successive vowels, but may match many
more, while

@example
\d@{8@}
@end example

@noindent
matches exactly 8 digits.  An opening curly bracket that
appears in a position where a quantifier is not allowed, or
one that does not match the syntax of a quantifier, is taken
as a literal character. For example, @{,6@} is not a quantifier,
but a literal string of four characters.@footnote{It
raises an error if @option{-R} is not used.}

The quantifier @samp{@{0@}} is permitted, causing the expression to
behave as if the previous item and the quantifier were not
present.

For convenience (and historical compatibility) the three
most common quantifiers have single-character abbreviations:

@table @code
@item *
is equivalent to @{0,@}

@item +
is equivalent to @{1,@}

@item ?
is equivalent to @{0,1@}
@end table

It is possible to construct infinite loops by following a
subpattern that can match no characters with a quantifier
that has no upper limit, for example:

@example
(a?)*
@end example

Earlier versions of Perl used to give an error at
compile time for such patterns. However, because there are
cases where this can be useful, such patterns are now
accepted, but if any repetition of the subpattern does in
fact match no characters, the loop is forcibly broken.

@cindex Greedy regular expression matching
@cindex Perl-style regular expressions, stingy repetitions
By default, the quantifiers are @dfn{greedy} like in @sc{posix}
mode, that is, they match as much as possible (up to the maximum
number of permitted times), without causing the rest of the
pattern to fail. The classic example of where this gives problems
is in trying to match comments in C programs. These appear between
the sequences @code{/*} and @code{*/} and within the sequence, individual
@code{*} and @code{/} characters may appear. An attempt to match C
comments by applying the pattern

@example
/\*.*\*/
@end example

@noindent
to the string

@example
/* first command */ not comment /* second comment */
@end example

@noindent

fails, because it matches the entire string owing to the
greediness of the @code{.*} item.

However, if a quantifier is followed by a question mark, it
ceases to be greedy, and instead matches the minimum number
of times possible, so the pattern @code{/\*.*?\*/}
does the right thing with the C comments. The meaning of the
various quantifiers is not otherwise changed, just the preferred
number of matches.  Do not confuse this use of question
mark with its use as a quantifier in its own right.
Because it has two uses, it can sometimes appear doubled, as in

@example
\d??\d
@end example

which matches one digit by preference, but can match two if
that is the only way the rest of the pattern matches.

Note that greediness does not matter when specifying addresses,
but can be nevertheless used to improve performance.

@ignore
If the PCRE_UNGREEDY option is set (an option which is not
available in Perl), the quantifiers are not greedy by
default, but individual ones can be made greedy by following
them with a question mark. In other words, it inverts the
default behaviour.
@end ignore

When a parenthesized subpattern is quantified with a minimum
repeat count that is greater than 1 or with a limited maximum,
more store is required for the compiled pattern, in
proportion to the size of the minimum or maximum.

@cindex Perl-style regular expressions, single line
If a pattern starts with @code{.*} or @code{.@{0,@}} and the
@code{S} modifier is used, the pattern is implicitly anchored,
because whatever follows will be tried against every character
position in the subject string, so there is no point in
retrying the overall match at any position after the first.
PCRE treats such a pattern as though it were preceded by \A.

When a capturing subpattern is repeated, the value captured
is the substring that matched the final iteration. For example,
after

@example
(tweedle[dume]@{3@}\s*)+
@end example

@noindent
has matched @samp{tweedledum tweedledee} the value of the
captured substring is @samp{tweedledee}.  However, if there are
nested capturing subpatterns, the corresponding captured
values may have been set in previous iterations. For example,
after

@example
/(a|(b))+/
@end example

matches @samp{aba}, the value of the second captured substring is
@samp{b}.

@node Backreferences
@appendixsec Backreferences
@cindex Perl-style regular expressions, backreferences

Outside a character class, a backslash followed by a digit
greater than 0 (and possibly further digits) is a back
reference to a capturing subpattern earlier (i.e.  to its
left) in the pattern, provided there have been that many
previous capturing left parentheses.

However, if the decimal number following the backslash is
less than 10, it is always taken as a back reference, and
causes an error only if there are not that many capturing
left parentheses in the entire pattern. In other words, the
parentheses that are referenced need not be to the left of
the reference for numbers less than 10. @ref{Backslash}
for further details of the handling of digits following a backslash.

A back reference matches whatever actually matched the capturing
subpattern in the current subject string, rather than
anything matching the subpattern itself. So the pattern

@example
(sens|respons)e and \1ibility
@end example

@noindent
matches @samp{sense and sensibility} and @samp{response and responsibility},
but not @samp{sense and responsibility}. If caseful
matching is in force at the time of the back reference, the
case of letters is relevant. For example,

@example
((?i)blah)\s+\1
@end example

@noindent
matches @samp{blah blah} and @samp{Blah Blah}, but not
@samp{BLAH blah}, even though the original capturing
subpattern is matched caselessly.

There may be more than one back reference to the same subpattern.
Also, if a subpattern has not actually been used in a
particular match, any back references to it always fail. For
example, the pattern

@example
(a|(bc))\2
@end example

@noindent
always fails if it starts to match @samp{a} rather than
@samp{bc}.  Because there may be up to 99 back references, all
digits following the backslash are taken as part of a potential
back reference number; this is different from what happens
in @sc{posix} mode. If the pattern continues with a digit
character, some delimiter must be used to terminate the back
reference.  If the @code{X} modifier option is set, this can be
whitespace.  Otherwise an empty comment can be used, or the
following character can be expressed in hexadecimal or octal.
Note that this applies only to the LHS pattern; it is
not possible yet to specify more than 9 backreferences on the 
RHS of the `s' command. 

A back reference that occurs inside the parentheses to which
it refers fails when the subpattern is first used, so, for
example, @code{(a\1)} never matches.  However, such references
can be useful inside repeated subpatterns. For example, the
pattern

@example
(a|b\1)+
@end example

@noindent
matches any number of @samp{a}s and also @samp{aba}, @samp{ababbaa},
etc. At each iteration of the subpattern, the back reference matches
the character string corresponding to the previous iteration.  In
order for this to work, the pattern must be such that the first
iteration does not need to match the back reference.  This can be
done using alternation, as in the example above, or by a
quantifier with a minimum of zero.

@node Assertions
@appendixsec Assertions
@cindex Perl-style regular expressions, assertions
@cindex Perl-style regular expressions, asserting subpatterns

An assertion is a test on the characters following or
preceding the current matching point that does not actually
consume any characters. The simple assertions coded as @code{\b},
@code{\B}, @code{\A}, @code{\Z}, @code{\z}, @code{^} and @code{$}
are described above. More complicated assertions are coded as
subpatterns.  There are two kinds: those that look ahead of the
current position in the subject string, and those that look behind it.

@cindex Perl-style regular expressions, lookahead subpatterns
An assertion subpattern is matched in the normal way, except
that it does not cause the current matching position to be
changed. Lookahead assertions start with @code{(?=} for positive
assertions and @code{(?!} for negative assertions. For example,

@example
\w+(?=;)
@end example

@noindent
matches a word followed by a semicolon, but does not include
the semicolon in the match, and

@example
foo(?!bar)
@end example

@noindent
matches any occurrence of @samp{foo} that is not followed by
@samp{bar}.

Note that the apparently similar pattern

@example
(?!foo)bar
@end example

@noindent
@cindex Perl-style regular expressions, lookbehind subpatterns
finds any occurrence of @samp{bar} even if it is preceded by
@samp{foo}, because the assertion @code{(?!foo)} is always true
when the next three characters are @samp{bar}. A lookbehind
assertion is needed to achieve this effect.
Lookbehind assertions start with @code{(?<=} for positive
assertions and @code{(?<!} for negative assertions. So,

@example
(?<!foo)bar
@end example

achieves the required effect of finding an occurrence of
@samp{bar} that is not preceded by @samp{foo}. The contents of a
lookbehind assertion are restricted
such that all the strings it matches must have a fixed
length.  However, if there are several alternatives, they do
not all have to have the same fixed length.  This is an extension
compared with Perl 5.005, which requires all branches to match
the same length of string. Thus

@example
(?<=dogs|cats|)
@end example

@noindent
is permitted, but the apparently equivalent regular expression

@example
(?<!dogs?|cats?)
@end example

@noindent
causes an error at compile time. Branches that match different
length strings are permitted only at the top level of
a lookbehind assertion: an assertion such as

@example
(?<=ab(c|de))
@end example

@noindent
is not permitted, because its single top-level branch can
match two different lengths, but it is acceptable if rewritten
to use two top-level branches:

@example
(?<=abc|abde)
@end example

All this is required because lookbehind assertions simply
move the current position back by the alternative's fixed
width and then try to match.  If there are
insufficient characters before the current position, the
match is deemed to fail.  Lookbehinds, in conjunction with
non-backtracking subpatterns can be particularly useful for
matching at the ends of strings; an example is given at the end
of the section on non-backtracking subpatterns.

Several assertions (of any sort) may occur in succession.
For example,

@example
(?<=\d@{3@})(?<!999)foo
@end example

@noindent
matches @samp{foo} preceded by three digits that are not @samp{999}.
Notice that each of the assertions is applied independently
at the same point in the subject string. First there is a
check that the previous three characters are all digits, and
then there is a check that the same three characters are not
@samp{999}.  This pattern does not match @samp{foo} preceded by six
characters, the first of which are digits and the last three
of which are not @samp{999}.  For example, it doesn't match
@samp{123abcfoo}. A pattern to do that is

@example
(?<=\d@{3@}...)(?<!999)foo
@end example

@noindent
This time the first assertion looks at the preceding six
characters, checking that the first three are digits, and
then the second assertion checks that the preceding three
characters are not @samp{999}.  Actually, assertions can be
nested in any combination, so one can write this as 

@example
(?<=\d@{3@}(?!999)...)foo
@end example

or

@example
(?<=\d@{3@}...(?<!999))foo
@end example

@noindent
both of which might be considered more readable.

Assertion subpatterns are not capturing subpatterns, and may
not be repeated, because it makes no sense to assert the
same thing several times. If any kind of assertion contains
capturing subpatterns within it, these are counted for the
purposes of numbering the capturing subpatterns in the whole
pattern.  However, substring capturing is carried out only
for positive assertions, because it does not make sense for
negative assertions.

Assertions count towards the maximum of 200 parenthesized
subpatterns.

@node Non-backtracking subpatterns
@appendixsec Non-backtracking subpatterns
@cindex Perl-style regular expressions, non-backtracking subpatterns

With both maximizing and minimizing repetition, failure of
what follows normally causes the repeated item to be evaluated
again to see if a different number of repeats allows the
rest of the pattern to match. Sometimes it is useful to
prevent this, either to change the nature of the match, or
to cause it fail earlier than it otherwise might, when the
author of the pattern knows there is no point in carrying
on.

Consider, for example, the pattern @code{\d+foo} when applied to
the subject line

@example
123456bar
@end example

After matching all 6 digits and then failing to match @samp{foo},
the normal action of the matcher is to try again with only 5
digits matching the @code{\d+} item, and then with 4, and so on,
before ultimately failing. Non-backtracking subpatterns
provide the means for specifying that once a portion of the
pattern has matched, it is not to be re-evaluated in this way,
so the matcher would give up immediately on failing to match
@samp{foo} the first time.  The notation is another kind of special
parenthesis, starting with @code{(?>} as in this example:

@example
(?>\d+)bar
@end example

This kind of parenthesis ``locks up'' the part of the pattern
it contains once it has matched, and a failure further into
the pattern is prevented from backtracking into it.
Backtracking past it to previous items, however, works as
normal.

Non-backtracking subpatterns are not capturing subpatterns.  Simple
cases such as the above example can be thought of as a maximizing
repeat that must swallow everything it can.  So,
while both @code{\d+} and @code{\d+?} are prepared to adjust the number of
digits they match in order to make the rest of the pattern
match, @code{(?>\d+)} can only match an entire sequence of digits.

This construction can of course contain arbitrarily complicated
subpatterns, and it can be nested.

@cindex Perl-style regular expressions, lookbehind subpatterns
Non-backtracking subpatterns can be used in conjunction with look-behind
assertions to specify efficient matching at the end
of the subject string. Consider a simple pattern such as

@example
abcd$
@end example

@noindent
when applied to a long string which does not match.  Because
matching proceeds from left to right, @command{sed} will look for
each @samp{a} in the subject and then see if what follows matches
the rest of the pattern. If the pattern is specified as

@example
^.*abcd$
@end example

@noindent
the initial @code{.*} matches the entire string at first, but when
this fails (because there is no following @samp{a}), it backtracks
to match all but the last character, then all but the
last two characters, and so on. Once again the search for
@samp{a} covers the entire string, from right to left, so we are
no better off. However, if the pattern is written as

@example
^(?>.*)(?<=abcd)
@end example

there can be no backtracking for the .* item; it can match
only the entire string. The subsequent lookbehind assertion
does a single test on the last four characters. If it fails,
the match fails immediately. For long strings, this approach
makes a significant difference to the processing time.

When a pattern contains an unlimited repeat inside a subpattern
that can itself be repeated an unlimited number of
times, the use of a once-only subpattern is the only way to
avoid some failing matches taking a very long time
indeed.@footnote{Actually, the matcher embedded in @value{SSED}
tries to do something for this in the simplest cases,
like @code{([^b]*b)*}.  These cases are actually quite
common: they happen for example in a regular expression
like @code{\/\*([^*]*\*)*\/} which matches C comments.}

The pattern

@example
(\D+|<\d+>)*[!?]
@end example

([^0-9<]+<(\d+>)?)*[!?]

@noindent
matches an unlimited number of substrings that either consist
of non-digits, or digits enclosed in angular brackets, followed by
an exclamation or question mark. When it matches, it runs quickly.
However, if it is applied to

@example
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
@end example

@noindent
it takes a long time before reporting failure.  This is
because the string can be divided between the two repeats in
a large number of ways, and all have to be tried.@footnote{The
example used @code{[!?]} rather than a single character at the end,
because both @value{SSED} and Perl have an optimization that allows
for fast failure when a single character is used. They
remember the last single character that is required for a
match, and fail early if it is not present in the string.}

If the pattern is changed to

@example
((?>\D+)|<\d+>)*[!?]
@end example

sequences of non-digits cannot be broken, and failure happens
quickly.

@node Conditional subpatterns
@appendixsec Conditional subpatterns
@cindex Perl-style regular expressions, conditional subpatterns

It is possible to cause the matching process to obey a subpattern
conditionally or to choose between two alternative
subpatterns, depending on the result of an assertion, or
whether a previous capturing subpattern matched or not. The
two possible forms of conditional subpattern are

@example
(?(@var{condition})@var{yes-pattern})
(?(@var{condition})@var{yes-pattern}|@var{no-pattern})
@end example

If the condition is satisfied, the yes-pattern is used; otherwise
the no-pattern (if present) is used. If there are more than two
alternatives in the subpattern, a compile-time error occurs.

There are two kinds of condition. If the text between the
parentheses consists of a sequence of digits, the condition
is satisfied if the capturing subpattern of that number has
previously matched.  The number must be greater than zero.
Consider the following pattern, which contains non-significant
white space to make it more readable (assume the @code{X} modifier)
and to divide it into three parts for ease of discussion:

@example
( \( )?   [^()]+   (?(1) \) )
@end example

The first part matches an optional opening parenthesis, and
if that character is present, sets it as the first captured
substring. The second part matches one or more characters
that are not parentheses. The third part is a conditional
subpattern that tests whether the first set of parentheses
matched or not.  If they did, that is, if subject started
with an opening parenthesis, the condition is true, and so
the yes-pattern is executed and a closing parenthesis is
required. Otherwise, since no-pattern is not present, the
subpattern matches nothing.  In other words, this pattern
matches a sequence of non-parentheses, optionally enclosed
in parentheses.

@cindex Perl-style regular expressions, lookahead subpatterns
If the condition is not a sequence of digits, it must be an
assertion.  This may be a positive or negative lookahead or
lookbehind assertion. Consider this pattern, again containing
non-significant white space, and with the two alternatives
on the second line:

@example
(?(?=...[a-z])
   \d\d-[a-z]@{3@}-\d\d |
   \d\d-\d\d-\d\d )
@end example

The condition is a positive lookahead assertion that matches
a letter that is three characters away from the current point.
If a letter is found, the subject is matched against the first
alternative @samp{@var{dd}-@var{aaa}-@var{dd}} (where @var{aaa} are
letters and @var{dd} are digits); otherwise it is matched against 
the second alternative, @samp{@var{dd}-@var{dd}-@var{dd}}.


@node Recursive patterns
@appendixsec Recursive patterns
@cindex Perl-style regular expressions, recursive patterns
@cindex Perl-style regular expressions, recursion

Consider the problem of matching a string in parentheses,
allowing for unlimited nested parentheses. Without the use
of recursion, the best that can be done is to use a pattern
that matches up to some fixed depth of nesting. It is not
possible to handle an arbitrary nesting depth. Perl 5.6 has
provided an experimental facility that allows regular
expressions to recurse (amongst other things). It does this
by interpolating Perl code in the expression at run time,
and the code can refer to the expression itself. A Perl pattern
tern to solve the parentheses problem can be created like
this:

@example
$re = qr@{\( (?: (?>[^()]+) | (?p@{$re@}) )* \)@}x;
@end example

The @code{(?p@{...@})} item interpolates Perl code at run time,
and in this case refers recursively to the pattern in which it
appears. Obviously, @command{sed} cannot support the interpolation of
Perl code.  Instead, the special item @code{(?R)} is provided for
the specific case of recursion. This pattern solves the
parentheses problem (assume the @code{X} modifier option is used
so that white space is ignored):

@example
\( ( (?>[^()]+) | (?R) )* \)
@end example

First it matches an opening parenthesis. Then it matches any
number of substrings which can either be a sequence of
non-parentheses, or a recursive match of the pattern itself
(i.e. a correctly parenthesized substring). Finally there is
a closing parenthesis.

This particular example pattern contains nested unlimited
repeats, and so the use of a non-backtracking subpattern for
matching strings of non-parentheses is important when applying
the pattern to strings that do not match. For example, when
it is applied to

@example
(aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
@end example

it yields a ``no match'' response quickly. However, if a
standard backtracking subpattern is not used, the match runs
for a very long time indeed because there are so many different
ways the @code{+} and @code{*} repeats can carve up the subject,
and all have to be tested before failure can be reported.

The values set for any capturing subpatterns are those from
the outermost level of the recursion at which the subpattern
value is set. If the pattern above is matched against

@example
(ab(cd)ef)
@end example

@noindent
the value for the capturing parentheses is @samp{ef}, which is
the last value taken on at the top level.

@node Comments
@appendixsec Comments
@cindex Perl-style regular expressions, comments

The sequence (?# marks the start of a comment which continues
ues up to the next closing parenthesis. Nested parentheses
are not permitted. The characters that make up a comment
play no part in the pattern matching at all.

@cindex Perl-style regular expressions, extended
If the @code{X} modifier option is used, an unescaped @code{#} character
outside a character class introduces a comment that continues
up to the next newline character in the pattern.
@end ifset


@page
@node Concept Index
@unnumbered Concept Index

This is a general index of all issues discussed in this manual, with the
exception of the @command{sed} commands and command-line options.

@printindex cp

@page
@node Command and Option Index
@unnumbered Command and Option Index

This is an alphabetical list of all @command{sed} commands and command-line
options.

@printindex fn

@contents
@bye

@c XXX FIXME: the term "cycle" is never defined...