XML encoding

Hi,

I've written a PHP class which can output a UK position as XML.

The code snippet is as follows:

$crlf = "\r\n";

$xml = "";
$xml = $xml."<?xml version=\"1.0\" encoding=\"UTF-8\"?>".$crlf;
$xml = $xml."<UK_Location xmlns:xsi=\"www.w3.org\">".$crlf;
$xml = $xml."<OS_X>".$this->osX."</OS_X>".$crlf;
....snip...
// Convert the XML into the UTF-8 text encoding scheme
$xml = utf8_encode($xml);

The output from Firefox is as expected and as I want:

<?xml version="1.0" encoding="UTF-8"?>
<UK_Location xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Postcode>W1D 5BT</Postcode>
<OS_X>529827</OS_X>
<OS_Y>180975</OS_Y>
<Latitude>N51:30:46</Latitude>
<Longitude>W0:07:50</Longitude>
<Latitude_Decimal>51.512793</Latitude_Decimal>
<Longitude_Decimal>-0.130487</Longitude_Decimal>
</UK_Location>

But from IE, I get a '-' at the start of the 2nd line.

<?xml version="1.0" encoding="UTF-8" ?>
- <UK_Location xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Postcode>W1D 5BT</Postcode>

This makes the XML invalid. Why do I get this and what can I do about it?
Please note the '-' is there even if I leave the utf8_encode() line out,
my initial supposition was that the utf8_encode() was causing the problem.

Thanks.

however I just discovered if I save the file from IE, the '-' disappears!
This PC only has IE 6.0 on it, might it just be an older version of IE XML
rendering problem?

> however I just discovered if I save the file from IE, the '-'
> disappears! This PC only has IE 6.0 on it, might it just be an older
> version of IE XML rendering problem?

No, it is IE's way of telling you that you can unfold that region. Just
click the '-'.

Have a nice weekend,
--
Willem Bogaerts

Application smith
Kratz B.V.
http://www.kratz.nl/

Willem Bogaerts emailed this:
>> however I just discovered if I save the file from IE, the '-'
>> disappears! This PC only has IE 6.0 on it, might it just be an older
>> version of IE XML rendering problem?
>
> No, it is IE's way of telling you that you can unfold that region. Just
> click the '-'.
>
> Have a nice weekend,

Oops. Spot the person who almost never uses IE !!

Thanks for that.

..oO(Matthew)

>Willem Bogaerts emailed this:
>>> however I just discovered if I save the file from IE, the '-'
>>> disappears! This PC only has IE 6.0 on it, might it just be an older
>>> version of IE XML rendering problem?
>>
>> No, it is IE's way of telling you that you can unfold that region. Just
>> click the '-'.
>>
>> Have a nice weekend,
>
>Oops. Spot the person who almost never uses IE !!

Just some additional "style hints":

1) Is there any specific need for a CRLF line break? Usually a LF ("\n")
should be more than enough. You could also define your line break as a
constant, which is what I usually prefer in such cases:

define('CRLF', "\r\n");

2) There are many ways to write an XML string with embedded variables.
Personally I find the one you use the least readable because of all the
concatenations and the escaping. Even with syntax highlighting it still
looks ugly to me. SGML/XML also allow single quotes around attribute
values, so some possible alternatives would be for example:

$xml .= '<?xml version="1.0" encoding="UTF-8"?>'.CRLF;

or simply

$xml .= "<?xml version='1.0' encoding='UTF-8'?>\n";

sprintf() can also be very helpful for more complex strings with many
embedded variables or expressions.

3) You could also have a look at the DOM extension to create your XML.

Just my 3 cents.

Micha

> 1) Is there any specific need for a CRLF line break? Usually a LF ("\n")
> should be more than enough. You could also define your line break as a
> constant, which is what I usually prefer in such cases:

If I recall correctly, End-Of-Line-characters are always LineFeeds in
xml. It is in the XML standard.


Best regards

..oO(Dikkie Dik)

>> 1) Is there any specific need for a CRLF line break? Usually a LF ("\n")
>> should be more than enough. You could also define your line break as a
>> constant, which is what I usually prefer in such cases:
>
>If I recall correctly, End-Of-Line-characters are always LineFeeds in
>xml. It is in the XML standard.

Yes, it's how an XML processor is supposed to handle the line breaks,
even if they're written as CRLF.

2.11 End-of-Line Handling
http://www.w3.org/TR/xml/#sec-line-ends

But of course you could also write all your XML without any line breaks
on a single line.

Micha

Thanks for the info and hints. I'm new to PHP so your comments are helpful.

> Just some additional "style hints":
>
> 1) Is there any specific need for a CRLF line break? Usually a LF ("\n")
> should be more than enough. You could also define your line break as a
> constant, which is what I usually prefer in such cases:
>
> define('CRLF', "\r\n");

I've been using something similar to define '<br />' but didn't bother
making a define for one function's benefit.


> 2) There are many ways to write an XML string with embedded variables.
> Personally I find the one you use the least readable because of all the
> concatenations and the escaping. Even with syntax highlighting it still
> looks ugly to me. SGML/XML also allow single quotes around attribute
> values, so some possible alternatives would be for example:
>
> $xml .= '<?xml version="1.0" encoding="UTF-8"?>'.CRLF;
>
> or simply
>
> $xml .= "<?xml version='1.0' encoding='UTF-8'?>\n";

Both are much clearer than mine, I'll use the second and modify my code.
Didn't know there was a .= operator, useful.


> sprintf() can also be very helpful for more complex strings with many
> embedded variables or expressions.

I remember that from my C days, many moons ago.


> 3) You could also have a look at the DOM extension to create your XML.

Will do.

Thanks a lot.

Dikkie Dik emailed this:
>> 1) Is there any specific need for a CRLF line break? Usually a LF ("\n")
>> should be more than enough. You could also define your line break as a
>> constant, which is what I usually prefer in such cases:
>
> If I recall correctly, End-Of-Line-characters are always LineFeeds in
> xml. It is in the XML standard.

Thanks Dikkie.

Michael Fesser emailed this:
> .oO(Dikkie Dik)
>
>>> 1) Is there any specific need for a CRLF line break? Usually a LF ("\n")
>>> should be more than enough. You could also define your line break as a
>>> constant, which is what I usually prefer in such cases:
>> If I recall correctly, End-Of-Line-characters are always LineFeeds in
>> xml. It is in the XML standard.
>
> Yes, it's how an XML processor is supposed to handle the line breaks,
> even if they're written as CRLF.
>
> 2.11 End-of-Line Handling
> http://www.w3.org/TR/xml/#sec-line-ends
>
> But of course you could also write all your XML without any line breaks
> on a single line.

But not very human readable. Hence my line breaks.

Cheers again.