=head1 NAME

String::Util -- String processing utilities

=head1 SYNOPSIS

  use String::Util ':all';

  # "crunch" whitespace and remove leading/trailing whitespace
  $val = crunch($val);

  # does this value have "content", i.e. it's defined
  # and has something besides whitespace?
  if (hascontent $val) {...}

  # format for display in a web page
  $val = htmlesc($val);

  # format for display in a web page table cell
  $val = cellfill($val);

  # remove leading/trailing whitespace
  $val = trim($val);

  # ensure defined value
  $val = define($val);

  # repeat string x number of times
  $val = repeat($val, $iterations);

  # remove leading/trailing quotes
  $val = unquote($val);

  # remove all whitespace
  $val = no_space($val);

  # remove trailing \r and \n, regardless of what
  # the OS considers an end-of-line
  $val = fullchomp($val);

  # or call in void context:
  fullchomp $val;

  # encrypt string using random seed
  $val = randcrypt($val);

  # are these two values equal, where two undefs count as "equal"?
  if (eqq $a, $b) {...}

  # are these two values different, where two undefs count as "equal"?
  if (neqq $a, $b) {...}

  # get a random string of some specified length
  $val = randword(10);

=head1 DESCRIPTION

String::Util provides a collection of small, handy utilities for processing
strings.

=head1 INSTALLATION

String::Util can be installed with the usual routine:

 perl Makefile.PL
 make
 make test
 make install

=head1 FUNCTIONS

=head2 collapse(string), crunch(string)

C<collapse()> collapses all whitespace in the string down to single spaces.
Also removes all leading and trailing whitespace.  Undefined input results in
undefined output.

C<crunch()> is the old name for C<collapse()>. I decided that "crunch" never
sounded right. Spaces don't go "crunch", they go "poof" like a collapsing
ballon. However, C<crunch()> will continue to work as an alias for
C<collapse()>.

=head2 hascontent(scalar), nocontent(scalar)

hascontent() returns true if the given argument is defined and contains
something besides whitespace.

An undefined value returns false.  An empty string returns false.  A value
containing nothing but whitespace (spaces, tabs, carriage returns, newlines,
backspace) returns false.  A string containing any other characters (including
zero) returns true.

C<nocontent()> returns the negation of C<hascontent()>.

=head2 trim(string)

Returns the string with all leading and trailing whitespace removed.
Trim on undef returns undef.

So, for example, the following code changes " my string  " to "my string":

 $var = " my string  ";
 $var = trim($var);

trim accepts two optional arguments, 'left' and 'right', both of which
are true by default.  So, to avoid trimming the left side of the string,
set the 'left' argument to false:

 $var = trim($var, left=>0);

To avoid trimming the right side, set 'right' to false:

 $var = trim($var, right=>0);

=head2 ltrim, rtrim

ltrim trims leading whitespace.  rtrim trims trailing whitespace.  They are
exactly equivalent to

 trim($var, left=>0);

and

 trim($var, right=>0);

=head2 no_space(string)

Removes all whitespace characters from the given string.

=head2 htmlesc(string)

Formats a string for literal output in HTML.  An undefined value is returned as
an empty string.

htmlesc() is very similar to CGI.pm's escapeHTML.  However, there are a few
differences. htmlesc() changes an undefined value to an empty string, whereas
escapeHTML() returns undefs as undefs.

=head2 cellfill(string)

Formats a string for literal output in an HTML table cell.  Works just like
htmlesc() except that strings with no content (i.e. are undef or are just
whitespace) are returned as C<&nbsp;>.

=head2 jsquote($string)

Escapes and quotes a string for use in JavaScript.  Escapes single quotes and
surrounds the string in single quotes.  Returns the modified string.

=head2 unquote(string)

If the given string starts and ends with quotes, removes them. Recognizes
single quotes and double quotes.  The value must begin and end with same type
of quotes or nothing is done to the value. Undef input results in undef output.
Some examples and what they return:

 unquote(q|'Hendrix'|);   # Hendrix
 unquote(q|"Hendrix"|);   # Hendrix
 unquote(q|Hendrix|);     # Hendrix
 unquote(q|"Hendrix'|);   # "Hendrix'
 unquote(q|O'Sullivan|);  # O'Sullivan

B<option:> braces

If the braces option is true, surrounding braces such as [] and {} are also
removed. Some examples:

 unquote(q|[Janis]|, braces=>1);  # Janis
 unquote(q|{Janis}|, braces=>1);  # Janis
 unquote(q|(Janis)|, braces=>1);  # Janis

=head2 define(scalar)

Takes a single value as input. If the value is defined, it is returned
unchanged.  If it is not defined, an empty string is returned.

This subroutine is useful for printing when an undef should simply be
represented as an empty string.  Perl already treats undefs as empty strings in
string context, but this subroutine makes the
L<warnings module|http://perldoc.perl.org/warnings.html>
go away.  And you B<ARE> using warnings, right?

=head2 repeat($string, $count)

Returns the given string repeated the given number of times. The following
command outputs "Fred" three times:

 print repeat('Fred', 3), "\n";

Note that repeat() was created a long time based on a misunderstanding of how
the perl operator 'x' works.  The following command using 'x' would perform
exactly the same as the above command.

 print 'Fred' x 3, "\n";

Use whichever you prefer.

=head2 randword(length, %options)

Returns a random string of characters. String will not contain any vowels (to
avoid distracting dirty words). First argument is the length of the return
string. So this code:

 foreach my $idx (1..3) {
   print randword(4), "\n";
 }

would output something like this:

 kBGV
 NCWB
 3tHJ

If the string 'dictionary' is sent instead of an integer, then a word is
randomly selected from a dictionary file.  By default, the dictionary file
is assumed to be at /usr/share/dict/words and the shuf command is used to
pull out a word.  The hash %String::Util::PATHS sets the paths to the
dictionary file and the shuf executable.  Modify that hash to change the paths.
So this code:

 foreach my $idx (1..3) {
   print randword('dictionary'), "\n";
 }

would output something like this:

 mustache
 fronds
 browning

B<option:> alpha

If the alpha option is true, only alphabetic characters are returned, no
numerals. For example, this code:

 foreach my $idx (1..3) {
   print randword(4, alpha=>1), "\n";
 }

would output something like this:

 qrML
 wmWf
 QGvF

B<option:> numerals

If the numerals option is true, only numerals are returned, no alphabetic
characters. So this code:

 foreach my $idx (1..3) {
   print randword(4, numerals=>1), "\n";
 }

would output something like this:

 3981
 4734
 2657

B<option:> strip_vowels

This option is true by default.  If true, vowels are not included in the
returned random string. So this code:

 foreach my $idx (1..3) {
   print randword(4, strip_vowels=>1), "\n";
  }

would output something like this:

 Sk3v
 pV5z
 XhSX

=head2 eqq($val1, $val2)

Returns true if the two given values are equal.  Also returns true if both
are undef.  If only one is undef, or if they are both defined but different,
returns false. Here are some examples and what they return.

 eqq('x', 'x'), "\n";      # 1
 eqq('x', undef), "\n";    # 0
 eqq(undef, undef), "\n";  # 1

=head2 neqq($str1, $str2)

The opposite of neqq, returns true if the two values are *not* the same.
Here are some examples and what they return.

 print neqq('x', 'x'), "\n";      # 0
 print neqq('x', undef), "\n";    # 1
 print neqq(undef, undef), "\n";  # 0

=head2 equndef(), neundef()

equndef() has been renamed to eqq(). neundef() has been renamed to neqq().
Those old names have been kept as aliases.

=head2 fullchomp(string)

Works like chomp, but is a little more thorough about removing \n's and \r's
even if they aren't part of the OS's standard end-of-line.

Undefs are returned as undefs.

=head2 randcrypt(string)

Crypts the given string, seeding the encryption with a random two character
seed.

=head2 randpost(%opts)

Returns a string that sorta looks like one or more paragraphs.

B<option:> word_count

Sets how many words should be in the post.  By default a random number from
1 to 250 is used.

B<option:> par_odds

Sets the odds of starting a new paragraph after any given word.  By default the
value is .05, which means paragraphs will have an average about twenty words.

B<option:> par

Sets the string to put at the end or the start and end of a paragraph. 
Defaults to two newlines for the end of a pargraph.

If this option is a single scalar, that string is added to the end of each
paragraph.

To set both the start and end string, use an array reference.  The first
element should be the string to put at the start of a paragraph, the second
should be the string to put at the end of a paragraph.

B<option:> max_length

Sets the maximum length of the returned string, including paragraph delimiters.

=head2 ords($string)

Returns the given string represented as the ascii value of each character.

For example, this code:

 ords('Hendrix')

returns this string:

 {72}{101}{110}{100}{114}{105}{120}

B<options>

=over 4

=item * convert_spaces=>[true|false]

If convert_spaces is true (which is the default) then spaces are converted to
their matching ord values. So, for example, this code:

 ords('a b', convert_spaces=>1)

returns this:

{97}{32}{98}

This code returns the same thing:

 ords('a b')

If convert_spaces is false, then spaces are just returned as spaces. So this
code:

 ords('a b', convert_spaces=>0);

returns

 {97} {98}

=item * alpha_nums

If the alpha_nums option is false, then characters 0-9, a-z, and A-Z are not
converted. For example, this code:

 ords('a=b', alpha_nums=>0)

returns this:

 a{61}b

=back

=head2 deords($string)

Takes the output from ords() and returns the string that original created that
output.

For example, this command:

 deords('{72}{101}{110}{100}{114}{105}{120}')

returns this string:
 Hendrix

=head2 crunchlines($str)

Compacts contiguous newlines into single newlines.  Whitespace between newlines
is ignored, so that two newlines separated by whitespace is compacted down to a
single newline.

For example, this code:

 crunchlines("x\n\n\nx")

outputs two x's with a single empty line between them:

 x

 x

=head2 spacepad

=head1 TERMS AND CONDITIONS

Copyright (c) 2012-2016 by Miko O'Sullivan.  All rights reserved.  This program
is free software; you can redistribute it and/or modify it under the same terms
as Perl itself. This software comes with B<NO WARRANTY> of any kind.

=head1 AUTHORS

Miko O'Sullivan
F<miko@idocs.com>

=head1 HISTORY

=over 4

=item Version 0.10, December 1, 2005

Initial release

=item Version 0.11, December 22, 2005

This is a non-backwards compatible version.

urldecode, urlencode were removed entirely.  All of the subs that used to
modify values in place were changed so that they do not do so anymore, except
for fullchomp.

See http://www.xray.mpe.mpg.de/mailing-lists/modules/2005-12/msg00112.html
for why these changes were made.

=item Version 1.01, November 7, 2010

Decided it was time to upload five years worth of changes.

=item Version 1.20, July, 2012

Properly listing prerequisites.

=item Version 1.21, July 18, 2012

Fixed error in POD.

=item Version 1.22, July 20, 2012

Fix in documentation for randpost().

Clarified documentation for hascontent() and nocontent().

=item Version 1.23, Sep 1, 2012

Fixed error in META.yml.

=item Version 1.24, December 31, 2014

Cleaned up POD formatting.

Changed file to using Unixish style newlines. I hadn't realized until now that
it was using Windowish newline. How embarrasing.

Added some features to ords().

=item Version 1.25, January 4, 2015

Added parentheses to braces option for unquote. Cleaned up and added to POD.
Minor fixes to comments.

Renamed equndef to eqq, and neundef to neqq. However, the old names have been
kept as aliases.

Minor cleanup of formatting.

=item Version 1.26, Aug 29, 2016

Fixed tests. No significant changes to module.

=back