base64 decode French characters

We are getting base64 encoded (XML) data from a third party. If the XML data is in English, everything works fine, I am able do base64 decode, and parse the XML. If the XML is all lower case French characters, everything works fine. But if the开发者_运维知识库 xml data contains upper case French characters (like &Agrave), if I do base64 decode and try to parse it, the parser fails. Any suggestions on how to fix this problem?

Thanks.

Base64 is a method to encode 8-bit binary data using 7-bits/US-ASCII charachters. After the Base64 decode you should have a standard XML file. Probably this XML file contains illegal characters, or does not correctly specify the character encoding it uses.

You mention À, an HTML-specific (not-XML) representation of À. If the XML contains the HTML encoded string À, there should also be a reference in the XML to an entity table specifying how to decode that string.

Alternatively, if your XML contains the À character directly, encoded using (for example) the ISO-8859-1 character set, either your XML should specify this encoding (<?xml version="1.0" encoding="ISO-8859-1"?>), or you should specify it yourself when decoding it. Failing that, the parser may assume (e.g) UTF-8 encoding is used, and will fail when trying to decode the À.

The exact error message should tell you what the problem is.

[update: À directly]: Sounds like the XML is invalid then; that they say UTF-8 but are actually using a different encoding. Check the XML bytes (after the base 64 decode) for this; if the À is encoded as one byte, it is definitely not UTF-8.

[update: how to fix?] If they incorrectly specify it in the XML header, they should really replace the false header (<?xml version="1.0" encoding="UTF-8"?>) with the correct one (<?xml version="1.0" encoding="windows-1252"?>). If they don't specify anything, it looks like the iconv function may be your best bet. I haven't really needed it, so I'm not 100 % sure about this, but looks like you could use: $data = iconv("ISO-8859-1", "UTF-8", $data) after the base64_decode and before the simplexml_load_string. I don't know of a way to specify the encoding directly while decoding the XML.

I'm not really experienced with the PHP specifics of character encoding, so I'm not giving any guarantees...

What's the XML character encoding? Maybe it's not UTF-8 and your parser is trying to parse the XML string as UTF-8.