--- /dev/null
+= 2001/09/29
+
+Philip Newton <Philip.Newton@gmx.net> sent in a clean patch that
+added support for defining words differently; that prevents
+Text::Wrap from untainting strings; and that fixes a documentation
+bug.
+
+So that fill.t can be used in the version included in the perl
+distribution, fill.t no longer uses File::Slurp.
+
+Both Sweth Chandramouli <svc@sweth.net> and Drew Degentesh
+<ddegentesh@daed.com> both objected to the automatic unexpand
+that Text::Wrap does on its results. Drew sent a patch which
+has been integrated.
+
+Way back in '97, Joel Earl <jrearl@VNET.IBM.COM> asked that
+it be possible to use a line separator other than \n when
+adding new lines. There is now support for that.
+
+= 2001/01/30
+
+Bugfix by Michael G Schwern <schwern@pobox.com>: don't add extra
+whitespace when working one an array of input (as opposed to a
+single string).
+
+Performance rewrite: use m/\G/ rather than s///.
+
+You can now specify that words that are too long to wrap can simply
+overflow the line. Feature requested by James Hoagland
+<hoagland@SiliconDefense.com> and by John Porter <jdporter@min.net>.
+
+Documentation changes from Rich Bowen <Rich@cre8tivegroup.com>.
+
+= 1998/11/29
+
+Combined Fill.pm into Wrap.pm. It appears there are versions of
+Wrap.pm with fill in them.
+
+= 1998/11/28
+
+Over the last couple of years, many people sent in various
+rewrites of Text::Wrap. I should have done something about
+updating it long ago. If someone wants to take it over from
+me, discuss it in perl-porters. I'll be happy to hand it
+over.
+
+Anyway, I have a bunch of people to thank. I didn't
+use what any of them sent in, but I did take ideas from
+all of them. Many sent in complete new implamentations.
+
+ Ivan Brawley <ibrawley@awadi.com.au>
+
+ Jacqui Caren <Jacqui.Caren@ig.co.uk>
+
+ Jeff Kowalski <jeff.kowalski@autodesk.com>
+
+ Allen Smith <easmith@beatrice.rutgers.edu>
+
+ Sullivan N. Beck <sbeck@cise.ufl.edu>
+
+The end result is a very slight change in the API. There
+is now an additional package variable: $Text::Wrap::huge.
+When $huge is set to 'die' then long words will cause
+wrap() to die. When it is set to 'wrap', long words will
+be wrapped. The default is 'wrap'.
+
+<shout>LONG WORDS WILL NOW BE WRAPPED BY DEFAULT</shout>.
+This is a change in behavior.
+
+At the bottom of Text::Wrap, there was a function (fill())
+sitting there unpublished. There was a note that Tim Pierce
+had a faster version, but a search on CPAN failed to turn it
+up. Text::Fill is now available.
+
@EXPORT = qw(wrap fill);
@EXPORT_OK = qw($columns $break $huge);
-$VERSION = 2001.0131;
+$VERSION = 2001.0929;
-use vars qw($VERSION $columns $debug $break $huge);
+use vars qw($VERSION $columns $debug $break $huge $unexpand $tabstop
+ $separator);
use strict;
BEGIN {
$debug = 0;
$break = '\s';
$huge = 'wrap'; # alternatively: 'die' or 'overflow'
+ $unexpand = 1;
+ $tabstop = 8;
+ $separator = "\n";
}
use Text::Tabs qw(expand unexpand);
{
my ($ip, $xp, @t) = @_;
+ local($Text::Tabs::tabstop) = $tabstop;
my $r = "";
my $tail = pop(@t);
- my $t = expand(join("", (map { /\s+\Z/ ? ( $_ ) : ($_, ' ') } @t), $tail));
+ my $t = expand(join("", (map { /\s+\z/ ? ( $_ ) : ($_, ' ') } @t), $tail));
my $lead = $ip;
my $ll = $columns - length(expand($ip)) - 1;
my $nll = $columns - length(expand($xp)) - 1;
my $nl = "";
my $remainder = "";
+ use re 'taint';
+
pos($t) = 0;
while ($t !~ /\G\s*\Z/gc) {
- if ($t =~ /\G([^\n]{0,$ll})($break|\Z(?!\n))/xmgc) {
- $r .= unexpand($nl . $lead . $1);
+ if ($t =~ /\G([^\n]{0,$ll})($break|\z)/xmgc) {
+ $r .= $unexpand
+ ? unexpand($nl . $lead . $1)
+ : $nl . $lead . $1;
$remainder = $2;
} elsif ($huge eq 'wrap' && $t =~ /\G([^\n]{$ll})/gc) {
- $r .= unexpand($nl . $lead . $1);
- $remainder = "\n";
- } elsif ($huge eq 'overflow' && $t =~ /\G([^\n]*?)($break|\Z(?!\n))/xmgc) {
- $r .= unexpand($nl . $lead . $1);
+ $r .= $unexpand
+ ? unexpand($nl . $lead . $1)
+ : $nl . $lead . $1;
+ $remainder = $separator;
+ } elsif ($huge eq 'overflow' && $t =~ /\G([^\n]*?)($break|\z)/xmgc) {
+ $r .= $unexpand
+ ? unexpand($nl . $lead . $1)
+ : $nl . $lead . $1;
$remainder = $2;
} elsif ($huge eq 'die') {
die "couldn't wrap '$t'";
$lead = $xp;
$ll = $nll;
- $nl = "\n";
+ $nl = $separator;
}
$r .= $remainder;
=head1 DESCRIPTION
-Text::Wrap::wrap() is a very simple paragraph formatter. It formats a
+C<Text::Wrap::wrap()> is a very simple paragraph formatter. It formats a
single paragraph at a time by breaking lines at word boundries.
Indentation is controlled for the first line (C<$initial_tab>) and
all subsquent lines (C<$subsequent_tab>) independently. Please note:
C<$initial_tab> and C<$subsequent_tab> are the literal strings that will
be used: it is unlikley you would want to pass in a number.
+Text::Wrap::fill() is a simple multi-paragraph formatter. It formats
+each paragraph separately and then joins them together when it's done. It
+will destory any whitespace in the original text. It breaks text into
+paragraphs by looking for whitespace after a newline. In other respects
+it acts like wrap().
+
+=head1 OVERRIDES
+
+C<Text::Wrap::wrap()> has a number of variables that control its behavior.
+Because other modules might be using C<Text::Wrap::wrap()> it is suggested
+that you leave these variables alone! If you can't do that, then
+use C<local($Text::Wrap::VARIABLE) = YOURVALUE> when you change the
+values so that the original value is restored. This C<local()> trick
+will not work if you import the variable into your own namespace.
+
Lines are wrapped at C<$Text::Wrap::columns> columns. C<$Text::Wrap::columns>
should be set to the full width of your output device. In fact,
every resulting line will have length of no more than C<$columns - 1>.
+It is possible to control which characters terminate words by
+modifying C<$Text::Wrap::break>. Set this to a string such as
+C<'[\s:]'> (to break before spaces or colons) or a pre-compiled regexp
+such as C<qr/[\s']/> (to break before spaces or apostrophes). The
+default is simply C<'\s'>; that is, words are terminated by spaces.
+(This means, among other things, that trailing punctuation such as
+full stops or commas stay with the word they are "attached" to.)
+
Beginner note: In example 2, above C<$columns> is imported into
the local namespace, and set locally. In example 3,
C<$Text::Wrap::columns> is set in its own namespace without importing it.
+C<Text::Wrap::wrap()> starts its work by expanding all the tabs in its
+input into spaces. The last thing it does it to turn spaces back
+into tabs. If you do not want tabs in your results, set
+C<$Text::Wrap::unexapand> to a false value. Likewise if you do not
+want to use 8-character tabstops, set C<$Text::Wrap::tabstop> to
+the number of characters you do want for your tabstops.
+
+If you want to separate your lines with something other than C<\n>
+then set C<$Text::Wrap::seporator> to your preference.
+
When words that are longer than C<$columns> are encountered, they
are broken up. C<wrap()> adds a C<"\n"> at column C<$columns>.
This behavior can be overridden by setting C<$huge> to
C<die()> to be called. When set to 'overflow', large words will be
left intact.
-Text::Wrap::fill() is a simple multi-paragraph formatter. It formats
-each paragraph separately and then joins them together when it's done. It
-will destory any whitespace in the original text. It breaks text into
-paragraphs by looking for whitespace after a newline. In other respects
-it acts like wrap().
-
-When called in list context, C<wrap()> will return a list of lines and
-C<fill()> will return a list of paragraphs.
-
-Historical notes: Older versions of C<wrap()> and C<fill()> always
-returned strings. Also, 'die' used to be the default value of
+Historical notes: 'die' used to be the default value of
C<$huge>. Now, 'wrap' is the default value.
=head1 EXAMPLE
projects / p5sagit/p5-mst-13.2.git / commitdiff