ansifold/ansicolrm/ansicut - fold/colrm/cut command handling ANSI terminal sequences
ansifold [ options ]
-w# --width=# Folding width (default 72)
--boundary=word|space Fold on word boundary
--padding[=#] Padding to margin space
--padchar=_ Default padding character
--prefix=string Set prefix string (default empty)
--autoindent=pattern Set auto-indent pattern
--keepindent Preserve indent string
--ambiguous=narrow|wide Unicode ambiguous character handling
-p --paragraph Print extra newline
-r --refill Join paragraph into single line first
--separate=string Set the output separator string (default newline)
-n --nonewline Same as --separate ''
--lb=# --linebreak=mode Line-break mode (all, runin, runout, none)
--runin=# Run-in width (default 4)
--runout=# Run-out width (default 4)
--runlen=# Set run-in and run-out both
--splitwide[=#] Split in the middle of wide character
-s --smart Same as --boundary=word --linebreak=all
-x[#] --expand[=#] Expand tabs
--tabstop=n Tab-stop position (default 8)
--tabhead=char Tab-head character (default space)
--tabspace=char Tab-space character (default space)
--tabstyle=style Tab expansion style (shade, dot, symbol)
--colrm start [ end ] colrm(1) command compatible
-c# --cut list cut(1) command compatible
-h --help Show help message
-v --version Show version
ansicolrm [ options ]
ansicut -c list
Version 1.32
ansifold is a fold(1) compatible command utilizing Text::ANSI::Fold module, which enables to handle ANSI terminal sequences.
ansicolrm works like colrm(1) command. This is an alias for ansifold command and works exactly same except option --colrm is enabled by default.
ansicut works like cut(1) command. This is an alias for ansifold command and works exactly same except default output separator string is set as empty by default. Support only -c (or --cut) option of the original cut(1) command.
ansifold folds lines in 72 column by default. Use option -w to change the folding width.
$ ansifold -w132
Single field is used repeatedly for the same line.
With option --padding, remained columns are filled by padding
character, space by default, or specified by optional value like
--padding=_. Default padding character can be set by --padchar
option.
ansifold handles Unicode multi-byte characters properly. Option --ambiguous takes wide or narrow and it specifies the visual width of Unicode ambiguous characters.
If the last character is full-width and must be wrapped in the middle
of it, it is wrapped just before the character. If padding is
specified, then one padding character is inserted. If you really want
to keep the wrapping position, use the --stripwide option.
If the width argument begins with =, it is interpreted as an RPN
(Reverse Polish Notation) expression with the terminal width as the
initial value. Therefore,
ansifold -w=
will wrap at the width of the terminal, and
ansifold -w=2/
will wrap at half the width of the terminal.
Unlike the original fold(1) command, multiple numbers can be specified.
$ LANG=C date | ansifold -w 3,1,3,1,2 | cat -n
1 Wed
2
3 Dec
4
5 19
With multiple fields, unmatched part is discarded as in the above example. So you can truncate lines by putting comma at the end of single field.
ansifold -w80,
Option -w80, is equivalent to -w80,0. Zero width is ignored
when seen as a final number, but not ignored otherwise.
If the data is shorter and there is no corresponding string for the field, an empty string is returned. If the padding option is specified, the field is padded to the given width.
Negative number fields are discarded.
$ LANG=C date | ansifold -w 3,-1,3,-1,2
Wed
Dec
19
If the final width is negative, it is not discarded but takes all the rest instead. So next commands do the same thing.
$ colrm 7 10
$ ansifold -nw 6,-4,-1
Option --width -1 does nothing effectively. Using it with
--expand option implements ANSI/Unicode aware expand(1) command.
$ ansifold --expand --width -1
This can be written as this.
$ ansifold -xw-1
Number description is handled by Getopt::EX::Numbers module, and
consists of start, end, step and length elements. For
example,
$ echo AABBBBCCCCCCDDDDDDDDEEEEEEEEEE | ansifold -w 2:10:2
is equivalent to:
$ echo AABBBBCCCCCCDDDDDDDDEEEEEEEEEE | ansifold -w 2,4,6,8,10
and produces output like this:
AA
BBBB
CCCCCC
DDDDDDDD
EEEEEEEEEE
Option -n eliminates newlines between columns.
$ LANG=C date | ansifold -w 3,-1,3,-1,2 -n
WedDec19
Option --separate set the output separator string.
$ echo ABCDEF | ansifold --separate=: -w 1,0,1,0,1,-1
A::B::C:DEF
Option -n is a short-cut for --separate ''.
Option --paragraph (or -p) print extra newline after each line. This is convenient when a paragraph is made up of single line, like microsoft word document. The -p option can be repeated multiple times and will output that many newline characters.
If a string is given by --prefix option, that string is inserted at the beginning of each folded text. This is convenient to produce indented text block. Because the first line is not affected, insert appropiate prefix if necessary. Originally made for App::Greple::frame module.
An experimental --autoindent option takes a regex pattern for the
indent label, and set the prefix string as a space string of that
label length. For example, command ps auxgw produce very long line
output and you may want to fold COMMAND portion with appropiate
indentation. In this case use --autoindent option like this:
$ ps axgw | ansifold --autoindent '.*TIME (?=COMMAND)' -w= --boundary=word
PID TT STAT TIME COMMAND
1 ?? Ss 817:25.87 /sbin/launchd
354 ?? S 4:30.01 /System/Applications/TextEdit.app/Contents/
MacOS/TextEdit
522 ?? Ss 2:50.67 /System/Library/PrivateFrameworks/Uninstall.
framework/Resources/uninstalld
If the --keepindent option is specified, the string matched by
--autoindent is inserted at the beginning of the line rather than
indenting with spaces.
Option --refill (or -r) makes the command to run in paragraph mode, which read consecutive non-blank lines at once, and join them into single line before processing. So all paragraphs are reformatted by new text width. You can use this with --autoindent option.
Option -rw-1 will just fill paragraphs without reformatting.
Line break adjustment is supported for ASCII word boundaries. As for Japanese, more complicated prohibition processing is performed. Use option -s to enable everything.
This option prohibit breaking line in the middle of ASCII/Latin word. Context of word is defined by option value; word means alpha-numeric sequence, while space means simply non-space printables.
Option --linebreak takes a value of all, runin, runout or none. Default value is none.
When --linebreak option is enabled, if the cut-off text start with space or prohibited characters (e.g. closing parenthesis), they are ran-in at the end of current line as much as possible.
If the trimmed text end with prohibited characters (e.g. opening parenthesis), they are ran-out to the head of next line, provided it fits to maximum width.
Maximum width of run-in/run-out characters are defined by --runin and --runout option. Default values are 4.
Option --runlen set both run-in/run-out width at once.
If it becomes necessary to break in the middle of a wide character,
split the character into left and right half. Replacement characters
are \N{LEFT HALF BLACK CIRCLE} (◖) and \N{RIGHT HALF BLACK CIRCLE} (◗) by default.
If a parameter is given, the first character is used as the left half. The next character, if any, is used as the right half, otherwise the first character is used.
Option --smart (or simply -s) set both --boundary=word and --linebreak=all, and enables all smart text formatting capability.
Use option --boundary=space if you want the command to behave more like -s option of fold(1) command.
Option --expand (or -x) enables tab character expansion.
$ ansifold --expand
Takes optional number for tabstop and it precedes to --tabstop option.
$ ansifold -x4w-1
If the command is executed with the name ansiexpand, it works as if
the --expand option were given, and set default folding width to
-1. App::ansiexpand is a bit more sophisticated and we recommend
using that one rather.
Each tab character is converted to tabhead and following tabspace characters (both are space by default). They can be specified by --tabhead and --tabspace option. If the option value is longer than single characger, it is evaluated as unicode name. Next example makes tab character visible keeping text layout.
$ ansifold --expand --tabhead="MEDIUM SHADE" --tabspace="LIGHT SHADE"
Option --tabstyle allow to set --tabhead and --tabspace
characters at once according to the given style name. Select from
dot, symbol or shade. Styles are defined in
Text::ANSI::Fold library.
$ ansifold --expand --tabstyle=shade
Option --colrm takes colrm(1) command compatible arguments.
Since the output separator string is not set, use the -n option to get the same result as the colrm(1) command; when invoked as ansicolrm command, the separator string is set to the empty by default.
Next command behave exactly like colrm start end and takes care of
ANSI terminal sequences.
$ ansifold -n --colrm start end
$ ansicolrm start end
Unlike standard colrm(1), start and end can be repeated as
many times as desired. Next command removes column 1-3 and 7-9, and
produces 4560 as a result.
$ echo 1234567890 | ansifold -n --colrm 1 3 7 9
^^^ ^^^
Option --cut (or -c) takes cut(1) command compatible arguments.
Since the output separator string is set, use the -n option to get the same result as the cut(1) command; when invoked as ansicut command, the separator string is set to the empty by default.
Next command behave exactly like cut -c list and takes care of ANSI
terminal sequences.
$ ansifold -n -c list ...
$ ansicut -c list ...
Next command retrieve column 4-6,9- and produces 45690 as a result.
$ echo 1234567890 | ansifold -nc 4-6,9-
^^^ ^^
Unlike cut(1)'s -c option, parameter number is taken as screen columns of the terminal, rather than number of logical characters.
Option --refill will join Hangul string without space. Probably this is not a correct behavior.
-
~/.ansifoldrcStart-up file. See Getopt::EX::Module for format.
$ cpanm App::ansifold
https://www.w3.org/TR/jlreq/: Requirements for Japanese Text Layout, W3C Working Group Note 11 August 2020
Kazumasa Utashiro
Copyright ©︎ 2018-2025 Kazumasa Utashiro
This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.