On Fri, Aug 1, 2008 at 9:50 AM, Yeti <[EMAIL PROTECTED]> wrote:
> <?php
> *# Hello Community
> # Internationalisation, a topic discussed more than enough and YES, I am
> looking forward to PHP6.
> # But in reality I still have to develop for PHP4 and that's where the dog
> is burried ^^
> # We have a customer here who is running a small site, but still in five
> different languages.
> # Lately he started complaining about some strange site behaviours:
>
> # He has a discussion board where people can post their ideas, comments etc.
> Nothing special
> # Every post has a maximum length of 2048 characters, which is checked by
> JavaScript at the Browser
> # and after submitting the form by PHP.
>
> # Our mistake was to use strlen();*
> global $cc_strlen; global $cc_mb;
> $cc_strlen = $cc_mb = 0;
> if (array_key_exists('text', $_POST)) {
> $cc_strlen = strlen($_POST['text']);
> $cc_mb = mb_strlen($_POST['text'], 'UTF-8'); *// new code*
> if ($cc_strlen > 2048) { /* snip */ } // do something
> }
>
> /* snip */ // do something
>
> *#this works fine as long as the user only submits single byte charachters,
> but with UTF-8 the whole thing changes ..*
> ?>
> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "
> http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd";>
> <html xmlns="http://www.w3.org/1999/xhtml";>
> <head>
> <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
> <title>test</title>
> </head>
> <body>
> <p>You submitted <?php echo $cc_strlen; ?> characters (STRLEN).</p>
> <p>You submitted <?php echo $cc_mb; ?> characters (MB_STRLEN).</p>
> <p>Characters Left:<span id="remainder">2048</span></p>
> <form action="" method="post" onsubmit="return false;" id="post_form">
> <textarea id="post_text" name="text" onkeydown="check_length();"
> onchange="check_length();" rows="10" cols="50">œŸŒ‡Ņ</textarea><br />
> <input type="submit" value="Submit" id="post_button"
> onclick="submit_form();" />
> </form>
> <script type="text/javascript">
> <!--
> var the_form = document.getElementById('post_form');
> var textarea = document.getElementById('post_text');
> var counter = document.getElementById('remainder');
> function check_length() {
> var remainder = 2048 - textarea.value.length;
> var length_alert = false;
> if (remainder < 0) {
> remainder = 0;
> for (var count = textarea.value.length; (count >= 2048); (count -= 1)) {
> textarea.value = textarea.value.substr(0, 2047);
> counter.style.color = 'red'
> length_alert = true;
> }
> }
> if (length_alert) alert('You are already using 2048 characters.');
> if (document.all) {
> counter.innerText = remainder;
> } else {
> counter.textContent = remainder;
> }
> }
> function submit_form() {
> check_length();
> the_form.submit();
> alert ('You submitted ' + textarea.value.length + ' characters');
> return true;
> }
> -->
> </script>
> <?php
> *# Now as soon as one is starting to submit UTF-8 characters strlen is not
> working proberly any more
> # So we had to work through thousands of lines of code, replacing strlen()
> with mb_strlen();
> # We also found mb_strlen to take about 8 times longer than strlen().*
>
> $s_t = microtime();
> mb_strlen('œŸŒ‡Ņ', 'UTF-8');
> $e_t = microtime();
> echo '<p>MB_STRLEN took : '.(($e_t - $s_t)*1000).' milliseconds</p>';
> $s_t = microtime();
> strlen('œŸŒ‡Ņ');
> $e_t = microtime();
> echo '<p>STRLEN took : '.(($e_t - $s_t)*1000).' milliseconds</p>';
>
> *# So much for internationalisation.
> # Just writing this as a reminder for everyone who is facing similar
> situations.*
> ?>
> </body>
> </html>
>
You can't determine timing by simply calling each function one time. I
changed your script to the following:
<?php
$iterations = 10000;
$s_t = microtime(true);
for ($i = 0; $i < $iterations; ++$i) {
mb_strlen('œŸŒ‡Ņ', 'UTF-8');
}
$e_t = microtime(true);
echo '<p>MB_STRLEN took : '.(($e_t - $s_t)*1000/$iterations).'
milliseconds</p>';
$s_t = microtime(true);
for ($i = 0; $i < $iterations; ++$i) {
strlen('œŸŒ‡Ņ');
}
$e_t = microtime(true);
echo '<p>STRLEN took : '.(($e_t - $s_t)*1000/$iterations).' milliseconds</p>';
?>
I ran this script several times, and the results below are fairly typical:
MB_STRLEN took : 0.054733037948608 milliseconds
STRLEN took : 0.037568092346191 milliseconds
The multi-byte function is slower, but not even by a factor of 2 on my
development machine.
Andrew
