indent.1

GNU 系统开发优化 C 语言程序的应用程序

.\" $Id$
.\"
.TH INDENT 1L 
.\" NAME should be all caps, SECTION should be 1-8, maybe w/ subsection
.\" other parms are allowed: see man(7), man(1)
.SH NAME
indent \- changes the appearance of a C program by inserting or deleting
whitespace.
.SH SYNOPSIS
.nf
.na
indent [-bad] [-bap] [-bbb] [-bc] [-bl] [-bliN] [-br] [-cN] [-cdN] [-cdb]
	[-ce] [-ciN] [-cliN] [-cpN] [-cs] [-bs] [-dN] [-diN] [-fc1] [-fca]
	[-gnu] [-iN] [-ipN] [-kr] [-lN] [-lp] [-nbad] [-nbap] [-nbbb]
	[-nbc] [-ncdb] [-nce] [-ncs] [-nfc1] [-nfca] [-nip] [-nlp] [-npcs]
	[-npsl] [-nsc] [-nsob] [-nss] [-nv] [-orig] [-npro] [-pcs] [-psl]
	[-sc] [-sob] [-ss] [-st] [-T] [-tsN] [-v] [-version] [file]
.ad
.fi
.SH "DESCRIPTION"
This man page is based on the Info file indent.info, produced by Makeinfo-1.47
from the input file indent.texinfo.  This is Edition 0.02, 5 May 1992, of
The `indent' Manual, for Indent Version 1.3.

Copyright (C) 1989, 1992 Free Software Foundation, Inc.
.SH "COPYING"
Permission is granted to make and distribute verbatim copies of this
manual provided the copyright notice and this permission notice are
preserved on all copies.
.SH "INTRODUCTION"
The `indent' program can be used to make code easier to read.  It
can also convert from one style of writing C to another.

`indent' understands a substantial amount about the syntax of C, but
it also attempts to cope with incomplete and misformed syntax.

In version 1.2 and more recent versions, the GNU style of indenting
is the default.
.SH "INVOKING INDENT"

As of version 1.3, the format of the `indent' command is:

.in +5
indent [OPTIONS] [INPUT-FILES]

indent [OPTIONS] [SINGLE-INPUT-FILE] [-o OUTPUT-FILE]

.in -5
This format is different from earlier versions and other versions of
`indent'.

In the first form, one or more input files are specified.  `indent'
makes a backup copy of each file, and the original file is replaced with
its indented version.  *Note Backup files::, for an explanation of how
backups are made.

In the second form, only one input file is specified.  In this case,
or when the standard input is used, you may specify an output file after
the `-o' option.

To cause `indent' to write to standard output, use the `-st' option.
This is only allowed when there is only one input file, or when the
standard input is used.

If no input files are named, the standard input is read for input.
Also, if a filename named `-' is specified, then the standard input is
read.

As an example, each of the following commands will input the program
`slithy_toves.c' and write its indented text to `slithy_toves.out':

.in +5
indent slithy_toves.c -o slithy_toves.out

indent -st slithy_toves.c > slithy_toves.out

cat slithy_toves.c | indent -o slithy_toves.out

.in -5
Most other options to `indent' control how programs are formatted.
As of version 1.2, `indent' also recognizes a long name for each option
name.  Long options are prefixed by either `--' or `+'.  `+' is being
superseded by `--' to maintain consistency with the POSIX standard.
In most of this document, the traditional, short names are used for the
sake of brevity.

Here is another example:

.in +5
indent -br test/metabolism.c -l85

.in -5
This will indent the program `test/metabolism.c' using the `-br' and
`-l85' options, write the output back to `test/metabolism.c', and write
the original contents of `test/metabolism.c' to a backup file in the
directory `test'.

Equivalent invocations using long option names for this example would
be:

.in +5
indent --braces-on-if-line --line-length185 test/metabolism.c

indent +braces-on-if-line +line-length185 test/metabolism.c

.in -5
If you find that you often use `indent' with the same options, you
may put those options into a file called `.indent.pro'. `indent' will
first look for `.indent.pro' in the current directory and use that if
found.  Otherwise, `indent' will search your home directory for
`.indent.pro' and use that file if it is found.  This behaviour is
different from that of other versions of `indent', which load both
files if they both exist.

Command line switches are handled *after* processing `.indent.pro'. 
Options specified later override arguments specified earlier, with one
exception: Explicitly specified options always override background
options (*note Common styles::.).  You can prevent `indent' from
reading an `.indent.pro' file by specifying the `-npro' option.
.SH "BACKUP FILES"
As of version 1.3, GNU `indent' makes GNU--style backup files, the
same way GNU Emacs does.  This means that either "simple" or "numbered"
backup filenames may be made.

Simple backup file names are generated by appending a suffix to the
original file name.  The default for the this suffix is the
one-character string `~' (tilde).  Thus, the backup file for `python.c'
would be `python.c~'.

Instead of the default, you may specify any string as a suffix by
setting the environment variable `SIMPLE_BACKUP_SUFFIX' to your
preferred suffix.

Numbered backup versions of a file `momewraths' look like
`momewraths.c.~23~', where 23 is the version of this particular backup.
When making a numbered backup of the file `src/momewrath.c', the
backup file will be named `src/momewrath.c.~V~', where V is one greater
than the highest version currently existing in the directory `src'.

The type of backup file made is controlled by the value of the
environment variable `VERSION_CONTROL'.  If it is the string `simple',
then only simple backups will be made.  If its value is the string
`numbered', then numbered backups will be made.  If its value is
`numbered-existing', then numbered backups will be made if there
*already exist* numbered backups for the file being indented;
otherwise, a simple backup is made.  If `VERSION_CONTROL' is not set,
then `indent' assumes the behaviour of `numbered-existing'.

Other versions of `indent' use the suffix `.BAK' in naming backup
files.  This behaviour can be emulated by setting
`SIMPLE_BACKUP_SUFFIX' to `.BAK'.

Note also that other versions of `indent' make backups in the
current directory, rather than in the directory of the source file as
GNU `indent' now does.
.SH "COMMON STYLES"
There are several common styles of C code, including the GNU style,
the Kernighan & Ritchie style, and the original Berkeley style.  A
style may be selected with a single "background" option, which
specifies a set of values for all other options.  However, explicitly
specified options always override options implied by a background
option.

As of version 1.2, the default style of GNU `indent' is the GNU
style.  Thus, it is no longer neccessary to specify the option `-gnu'
to obtain this format, although doing so will not cause an error. 
Option settings which correspond to the GNU style are:

.in +5
-nbad -bap -nbbb -nbc -bl -bli2 -c33 -cd33 -ncdb -nce
-cli0 -cp1 -di0 -nfc1 -nfca -i2 -ip5 -lp -pcs -psl
-nsc -nsob -nss -ts8

.in -5
The GNU coding style is that preferred by the GNU project.  It is the
style that the GNU Emacs C mode encourages and which is used in the C
portions of GNU Emacs.  (People interested in writing programs for
Project GNU should get a copy of `The GNU Coding Standards', which also
covers semantic and portability issues such as memory usage, the size
of integers, etc.)

The Kernighan & Ritchie style is used throughout their well-known
book `The C Programming Language'.  It is enabled with the `-kr'
option.  The Kernighan & Ritchie style corresponds to the following set
of options:

.in +5
-nbad -bap -nbbb -nbc -br -c33 -cd33 -ncdb -ce -ci4
-cli0 -cp33 -d0 -di1 -nfc1 -nfca -i4 -ip0 -l75 -lp
-npcs -npsl -nsc -nsob -nss -ts8

.in -5
Kernighan & Ritchie style does not put comments to the right of code in
the same column at all times (nor does it use only one space to the
right of the code), so for this style `indent' has arbitrarily chosen
column 33.

The style of the original Berkeley `indent' may be obtained by
specifying `-orig' (or by specifyfying `--original', using the long
option name).  This style is equivalent to the following settings:

.in +5
-nbap -nbad -nbbb -bc -br -c33 -cd33 -cdb -ce -ci4
-cli0 -cp33 -d4 -di16 -fc1 -fca -i4 -ip4 -l75 -lp
-npcs -psl -sc -nsob -nss -ts8

.in -5
.SH "BLANK LINES"
Various programming styles use blank lines in different places.
`indent' has a number of options to insert or delete blank lines in
specific places.

The `-bad' option causes `indent' to force a blank line after every
block of declarations.  The `-nbad' option causes `indent' not to force
such blank lines.

The `-bap' option forces a blank line after every procedure body.
The `-nbap' option forces no such blank line.

The `-bbb' option forces a blank line before every block comment. A
block comment is one which starts in column one when formatting of such
comments is disabled, or one with `-' or `*' immediately following the
`/*'.  The `-nbbb' option does not force such blank lines.

The `-sob' option causes `indent' to swallow optional blank lines
(that is, any optional blank lines present in the input will be removed
from the output).  If the `-nsob' is specified, any blank lines present
in the input file will be copied to the output file.

The `-bad' option forces a blank line after every block of
declarations.  The `-nbad' option does not add any such blank lines.

For example, given the input

.in +5
.nf
.na
     char *foo;
     char *bar;
     /* This separates blocks of declarations.  */
     int baz;

.in -5
.ad
.fi
`indent -bad' produces


.in +5
.nf
.na
     char *foo;
     char *bar;
     
     /* This separates blocks of declarations.  */
     int baz;

.in -5
.ad
.fi
and `indent -nbad' produces

.in +5
.nf
.na
     char *foo;
     char *bar;
     /* This separates blocks of declarations.  */
     int baz;

.in -5
.ad
.fi
The `-bap' option forces a blank line after every procedure body.
For example, given the input

.in +5
.nf
.na
     int
     foo ()
     {
       puts("Hi");
     }
     /* The procedure bar is even less interesting.  */
     char *
     bar ()
     {
       puts("Hello");
     }

.in -5
.ad
.fi
`indent -bap' produces

.in +5
.nf
.na
     int
     foo ()
     {
       puts ("Hi");
     }
     
     /* The procedure bar is even less interesting.  */
     char *
     bar ()
     {
       puts ("Hello");
     }

.in -5
.ad
.fi
and `indent -nbap' produces

.in +5
.nf
.na
     int
     foo ()
     {
       puts ("Hi");
     }
     /* The procedure bar is even less interesting.  */
     char *
     bar ()
     {
       puts ("Hello");
     }

.in -5
.ad
.fi
No blank line will be added after the procedure `foo'.
.SH "COMMENTS"
Comments are no longer formatted by default as of version 1.2.  This
can be enabled with the `-fca' option.  Doing so will cause newlines in
the comment text to be ignored and the line will be filled up to the
length of a line (which can be modified with `-l').  When formatting is
enabled, blank lines indicate paragraph breaks.

The `-fc1' option enables the formatting of comments which begin in
the first column.  The `-nfc1' option disables the formatting of first
column comments.  When comment formatting is disabled, overall comment
indentation may still be adjusted.

The indentation of comments which do not appear to the right of code
is set by the `-d' option, which specifies the number of spaces to the
left of the surrounding code that the comment appears.  For example,
`-d2' places comments two spaces to the left of code; `-d0' lines up
comments with the code.  The `-cdb' option controls whether the `/*'
and `*/' are placed on blank lines.  With `-cdb', comments look like
this:

.in +5
.nf
.na
     /*
      * this is a comment
      */

.in -5
.ad
.fi
With `-ncdb', comments look like this:

.in +5
.nf
.na
     /* this is a comment */

.in -5
.ad
.fi
`-cdb' only affects block comments, not comments to the right of code. 
The default is `-ncdb'.

Comments which appear on the same line as code are placed to the
right. The column in which comments on code start is controlled by the
`-c' option.  The column in which comments to the right of declarations
start is controlled by the `-cd' option.  By default, they start in the
same column as comments to the right of code, which is column 33.  The
column number for comments to the right of `#else' and `#endif'
statements is controlled by the `-cp' option.  If the code on a line
extends past the comment column, the comment starts further to the
right, and the right margin may be automatically extended in extreme
cases.

If the `-sc' option is specified, `*' is placed at the left edge of
all comments.  For example:

.in +5
.nf
.na
     /* This is a comment which extends from one line
      * onto the next line, thus causing us to consider
      * how it should continue.  */

.in -5
.ad
.fi
instead of

.in +5
.nf
.na
     /* This is a comment which extends from one line
        onto the next line, thus causing us to consider
        how it should continue.  */

.in -5
.ad
.fi
.SH "STATEMENTS"
The `-br' or `-bl' option specifies how to format braces.
The `-br' option formats braces like this:

.in +5
.nf
.na
     if (x > 0) {
       x--;
     }

.in -5
.ad
.fi
The `-bl' option formats them like this:

.in +5
.nf
.na
     if (x > 0)
       {
         x--;
       }

.in -5
.ad
.fi
If you use the `-bl' option, you may also want to specify the `-bli'
option.  This option specifies the number of spaces by which braces are
indented.  `-bli2', the default, gives the result shown above.  `-bli0'
results in the following:

.in +5
.nf
.na
     if (x > 0)
     {
       x--;
     }

.in -5
.ad
.fi
If you are using the `-br' option, you probably want to also use the
`-ce' option.  This causes the `else' in an if-then-else construct to
cuddle up to the immediately preceding `}'.  For example, with `-br
-ce' you get the following:

.in +5
.nf
.na
     if (x > 0) {
       x--;
     } else {
       fprintf (stderr, "...something wrong?\n");
     }

.in -5
.ad
.fi
With `-br -nce' that code would appear as

.in +5
.nf
.na
     if (x > 0) {
       x--;
     }
     else {
       fprintf (stderr, "...something wrong?\n");
     }

.in -5
.ad
.fi
The `-cli' option specifies the number of spaces that case labels
should be indented to the right of the containing `switch' statement.

If a semicolon is on the same line as a `for' or `while' statement,
the `-ss' option will cause a space to be placed before the semicolon. 
This emphasizes the semicolon, making it clear that the body of the
`for' or `while' statement is an empty statement. `-nss' disables this
feature.