Smarty

[mohrt]

I thought about a preg_match_all solution already, things get tricky when it comes time to replace the upper-cased words back into the original string. You can easily get borked results if a word is a subset of another word. We could keep track of word positions I suppose, but it seemed a bit much. preg_match_all can track matched positions as of 4.3.0, but that isn't early enough

[boots]

Well, then perhaps leave it as preg_replace_callback, but have two versions: one that calls a function that assumes True for the param and one that assumes False.

eg:

[php:1:3778730b93]<?php
function smarty_modifier_capitalize($string, $uc_digits = false)
{
if ($uc_digits)
return preg_replace_callback('!\b\w+\b!', 'smarty_modifier_capitalize_ucfirst', $string);
else
return preg_replace_callback('!\b\w+\b!', 'smarty_modifier_capitalize_ucfirst_nodigs', $string);
}

?>[/php:1:3778730b93]
etc.

What a can of worms

[mohrt]

ok I got that fixed with a static var, not real pretty but better than a constant or global

[boots]

That's the one. I had thought of statics but completely forgot that they could be used outside of a class context!!

Thanks, Monte!

[boots]

I changed the docs to reflect the new optional parameter and updated the example to show the behaviour. I also re-tested with various word forms, whitespacing, linebreaks and character sets and it seems to work great.

Thanks again to both of you!

[messju]

instead of a static variable we could have two replace-functions and use them according to the value of $uc_digits .

[php:1:2f82f89b55]<?php

function smarty_modifier_capitalize($string, $uc_digits = false)
{
$callback = 'smarty_modifier_capitalize_uc' . ( $uc_digits ? 'digits' : 'first' );
return preg_replace_callback('!\b\w+\b!', $callback, $string);
}


function smarty_modifier_capitalize_ucfirst($string)
{
/* uppercase only words that do not contain any digits */
if (!preg_match('!\d!', $string[0]))
return ucfirst($string[0]);
else
return $string[0];
}


function smarty_modifier_capitalize_ucdigits($string)
{
/* uppercase all words */
return ucfirst($string[0]);
}


?>[/php:1:2f82f89b55]

[boots]

@messju: that's essentially what I was trying to suggest earlier. IMHO it is splitting hairs at this point

While we are taking the stick to this horse, if we want to add another nifty feature I propose allowing a second optional parameter which would be a comma separated list of words to ignore.

EG:

[boots]

I made the change to support ignore-words. Should I commit it?