Alternatives to iconv()

[GoNC]

I'm moving everything to a new server. The old server used PHP v. 5.2.17, and the new one is 5.4.25.

One function I used fairly often when importing content from another site was iconv(), like so:

$var = iconv("UTF-8", "ISO-8859-1//IGNORE", $var);

On the new server, though, this is returning false (or maybe just empty), leaving me with nothing but a blank variable. If I remove the line, everything is fine, so iconv() is definitely the problem.

I did a little digging, and found that iconv() does have a bug specifically when using //IGNORE, but unfortunately for me, that's pretty important:

[bugs.php.net...]

I tried the recommended fix, which didn't give any errors, but didn't convert anything, either:

ini_set('mbstring.substitute_character', "none");
$text= mb_convert_encoding($text, 'UTF-8', 'ISO-8859-1');

I don't exactly want to use ini_set() on so many scripts, anyway, so I still don't know if this is the best choice.

Can you guys recommend an alternative to iconv() that would accomplish the same thing?

[penders]

Have you tried

mb_substitute_character("none")

instead of

ini_set()

?

Do you have an example of the content you are trying to convert/import?

[GoNC]

I'm importing RSS from Yahoo News. When it comes in, I occasionally get something like this:

PHILADELPHIA — All Ryan Braun needed to shake his slump was a trip to Philadelphia.

I'm trying to convert the — to something recognizable, or at least a whitespace.

I'm currently getting away with this by using:

$text = str_replace(array('—', '’', 'â?'), ' ', $text);

But of course, this isn't ideal at all, since more odd characters can show up without notice.

[coopster]

Question: rather than convert the characters why not just use utf-8 to store and display them?

If this is not possible you may need to check your locale before the conversion.