Re: Strange 'Â' character output when using simplexml load string

On 25 Feb, 10:56, Toby A Inkster <usenet200...@xxxxxxxxxxxxxxxxx>
Andy Hassall wrote:
bizt <bissa...@xxxxxxxxxxx> wrote:

I converting an XML string using simplexml_load_string function. It is
giving me a  character for some reason dotted around the text.

 simplexml always outputs in UTF-8. Is your page's encoding UTF-8?

At a guess, ISO-8859-1 or perhaps ISO-8859-15.

In UTF-8, a "prefix" of an 0xC2 byte is used to access the top half of the
"Latin-1 Supplement" block which includes a lot of juicy characters such
as currency symbols, fractions, superscript 2 and 3, the copyright and
registered trademark symbols, and the non-breaking space.

However in ISO-8859-1 and -15, the byte 0xC2 represents an Â, so if UTF-8
is misinterpreted as one of those, then you get  followed by some other
nonsense character.

Probably the easiest solution would be to take the output from SimpleXML
and pass it through iconv():

        $xmlout = iconv('UTF-8', 'ISO-8859-15//TRANSLIT', $xmlout);

Note that UTF-8 is capable of representing a far greater range of
characters than ISO-8859-1/-15 are, so certain characters may not properly
survive conversion. (Using the '//TRANSLIT' option tells iconv to do its
best, and if, say, a particular accented character is not available in
ISO-8859-1, then to substitute an unaccented one in its place.)

Toby A Inkster BSc (Hons) ARCS
[Geek of HTML/SQL/Perl/PHP/Python/Apache/Linux]
[OS: Linux, up 26 days, 15:55.]

                               Bottled Water

Hi, ive tried what you said which worked for one of my pages but when
i tried it on another i got the following:

Notice: iconv() [function.iconv]: Detected an illegal character in
input string in /home/public_html/search_apartments.php on line 67

Im using the following to convert my XML string which is fetched via

$result = iconv('UTF-8', 'ISO-8859-15//TRANSLIT', $result);

Would it be the case that my $result string, im not providing the
iconv() with the correct input encoding? If so, is there a way for me
to detect the input encoding?