Multi-byte safe wordwrap() function for UTF-8

user_137067 0
Viewed 14731

PHP's wordwrap() function doesn't work correctly for multi-byte strings like UTF-8.

There are a few examples of mb safe functions in the comments, but with some different test data they all seem to have some problems.

The function should take the exact same parameters as wordwrap().

Specifically be sure it works to:

  • cut mid-word if $cut = true, don't cut mid-word otherwise
  • not insert extra spaces in words if $break = ' '
  • also work for $break = "\n"
  • work for ASCII, and all valid UTF-8
multibyte utf-8 word-wrap php string
9 Answers

I haven't found any working code for me. Here is what I've written. For me it is working, thought it is probably not the fastest.

function mb_wordwrap($str, $width = 75, $break = "\n", $cut = false) {
    $lines = explode($break, $str);
    foreach ($lines as &$line) {
        $line = rtrim($line);
        if (mb_strlen($line) <= $width)
            continue;
        $words = explode(' ', $line);
        $line = '';
        $actual = '';
        foreach ($words as $word) {
            if (mb_strlen($actual.$word) <= $width)
                $actual .= $word.' ';
            else {
                if ($actual != '')
                    $line .= rtrim($actual).$break;
                $actual = $word;
                if ($cut) {
                    while (mb_strlen($actual) > $width) {
                        $line .= mb_substr($actual, 0, $width).$break;
                        $actual = mb_substr($actual, $width);
                    }
                }
                $actual .= ' ';
            }
        }
        $line .= trim($actual);
    }
    return implode($break, $lines);
}
user_6156270 
/**
 * wordwrap for utf8 encoded strings
 *
 * @param string $str
 * @param integer $len
 * @param string $what
 * @return string
 * @author Milian Wolff <mail@milianw.de>
 */

function utf8_wordwrap($str, $width, $break, $cut = false) {
    if (!$cut) {
        $regexp = '#^(?:[\x00-\x7F]|[\xC0-\xFF][\x80-\xBF]+){'.$width.',}\b#U';
    } else {
        $regexp = '#^(?:[\x00-\x7F]|[\xC0-\xFF][\x80-\xBF]+){'.$width.'}#';
    }
    if (function_exists('mb_strlen')) {
        $str_len = mb_strlen($str,'UTF-8');
    } else {
        $str_len = preg_match_all('/[\x00-\x7F\xC0-\xFD]/', $str, $var_empty);
    }
    $while_what = ceil($str_len / $width);
    $i = 1;
    $return = '';
    while ($i < $while_what) {
        preg_match($regexp, $str,$matches);
        $string = $matches[0];
        $return .= $string.$break;
        $str = substr($str, strlen($string));
        $i++;
    }
    return $return.$str;
}

Total time: 0.0020880699 is good time :)

user_6119690

Custom word boundaries

Unicode text has many more potential word boundaries than 8-bit encodings, including 17 space separators, and the full width comma. This solution allows you to customize a list of word boundaries for your application.

Better performance

Have you ever benchmarked the mb_* family of PHP built-ins? They don't scale well at all. By using a custom nextCharUtf8(), we can do the same job, but orders of magnitude faster, especially on large strings.

<?php

function wordWrapUtf8(
  string $phrase,
  int $width = 75,
  string $break = "\n",
  bool $cut = false,
  array $seps = [' ', "\n", "\t", '，']
): string
{
  $chunks = [];
  $chunk = '';
  $len = 0;
  $pointer = 0;
  while (!is_null($char = nextCharUtf8($phrase, $pointer))) {
    $chunk .= $char;
    $len++;
    if (in_array($char, $seps, true) || ($cut && $len === $width)) {
      $chunks[] = [$len, $chunk];
      $len = 0;
      $chunk = '';
    }
  }
  if ($chunk) {
    $chunks[] = [$len, $chunk];
  }
  $line = '';
  $lines = [];
  $lineLen = 0;
  foreach ($chunks as [$len, $chunk]) {
    if ($lineLen + $len > $width) {
      if ($line) {
        $lines[] = $line;
        $lineLen = 0;
        $line = '';
      }
    }
    $line .= $chunk;
    $lineLen += $len;
  }
  if ($line) {
    $lines[] = $line;
  }
  return implode($break, $lines);
}

function nextCharUtf8(&$string, &$pointer)
{
  // EOF
  if (!isset($string[$pointer])) {
    return null;
  }

  // Get the byte value at the pointer
  $char = ord($string[$pointer]);

  // ASCII
  if ($char < 128) {
    return $string[$pointer++];
  }

  // UTF-8
  if ($char < 224) {
    $bytes = 2;
  } elseif ($char < 240) {
    $bytes = 3;
  } elseif ($char < 248) {
    $bytes = 4;
  } elseif ($char == 252) {
    $bytes = 5;
  } else {
    $bytes = 6;
  }

  // Get full multibyte char
  $str = substr($string, $pointer, $bytes);

  // Increment pointer according to length of char
  $pointer += $bytes;

  // Return mb char
  return $str;
}
user_5544060

Just want to share some alternative I found on the net.

<?php
if ( !function_exists('mb_str_split') ) {
    function mb_str_split($string, $split_length = 1)
    {
        mb_internal_encoding('UTF-8'); 
        mb_regex_encoding('UTF-8');  

        $split_length = ($split_length <= 0) ? 1 : $split_length;

        $mb_strlen = mb_strlen($string, 'utf-8');

        $array = array();

        for($i = 0; $i < $mb_strlen; $i += $split_length) {
            $array[] = mb_substr($string, $i, $split_length);
        }

        return $array;
    }
}

Using mb_str_split, you can use join to combine the words with <br>.

<?php
    $text = '<utf-8 content>';

    echo join('<br>', mb_str_split($text, 20));

And finally create your own helper, perhaps mb_textwrap

<?php

if( !function_exists('mb_textwrap') ) {
    function mb_textwrap($text, $length = 20, $concat = '<br>') 
    {
        return join($concat, mb_str_split($text, $length));
    }
}

$text = '<utf-8 content>';
// so simply call
echo mb_textwrap($text);

See screenshot demo: mb_textwrap demo

user_44812350 
function mb_wordwrap($str, $width = 74, $break = "\r\n", $cut = false)
        {
            return preg_replace(
                '~(?P<str>.{' . $width . ',}?' . ($cut ? '(?(?!.+\s+)\s*|\s+)' : '\s+') . ')(?=\S+)~mus',
                '$1' . $break,
                $str
            );
        }
user_27881450
Related