Mb_Ereg_Replace Acting Differently Now It's Moved To A Function

[timski72]

Hi Guys,

 

I was using mb_ereg_replace in a script. It was using it to convert Greek characters to their latin ahplabet equivalent. That was working fine. Then I moved it to a function in a new script and now it doesn't seem to work.

 

The following lines of code:

$input = "η φιλι";
$input = mb_ereg_replace("τζ", "dz", $input);

used to leave $input untouched as "τζ" doesn't appear in $input.

 

However when I moved this to a function $input was changed by this line of code. I've pasted some screenshots that I made whilst debugging this in phpDesigner to show what I mean.

 

OK this is a little difficult to show as the Greek characters show as goggledegook in the debugger but the first screenshot shows that I set $input to "η φιλι" which shows as I've highlighted in yellow in the screenshot below in the debugger. If you echo $input it shows as "η φιλι". I'm not sure why it shows as highlighted in yellow below in the debugger but that's not the problem. I presume phpDesigner doesn't display unicode properly.

 

 

Anyway, once the content of $input has passed through mb_ereg_replace within the same script you can see the content has remained the same (see below). This is as expected as "τζ" isn't present in $input.

 

 

 

However, once I moved this code to a new script and put it in a function it behaved differently (see below). As you can see I've set $input to "η φιλι" again and that is the same as above.

 

Now though when it passes into the function it starts out ok, i.e. the content of $input is still the same.

However, once I step over that line of code and $input goes through mb_ereg_replace the content is changed as below.

 

As "τζ" doesn't appear in $input I would expect it to remain unchanged as it did when this line of code was in a single script. I don't understand why this has changed by moving it to a new script and function?

 

Any help would be very welcome.

Thank you,

Tim.

[PFMaBiSmAd]

For the code that doesn't work, have you set the mb_internal_encoding anywhere in the script?

[timski72]

Yes. I hadn't at the time that I made the screenshot but now the code includes this line

 

mb_internal_encoding("UTF-8");

 

but still the content is being 'corrupted'.

 

Thanks,

Tim.

[Christian F.]

Verify that the file is indeed saved as UTF-8, and that your editor/debugger is treating it as such. Not to mention HTTP headers, and everything else involved, of course.

 

Also, why are you using two mb_strtolower () calls after each other, on the same variable? Doesn't make a whole lot of sense, to be honest. A further improvement you could do is to replace mb_ereg_replace () with a straight str_replace () (it works with multi-byte strings), seeing as you're not using an actual Regular Expression in your replacements.