Printable View
Hi,
I am trying to parse an XML file using Libxml2 parser. When I am parsing the file, I get the following exception:
parser error : Input is not proper UTF-8, indicate encoding !
Bytes: 0xC7 0x41 0x49 0x53 <Name>AMITIÉS FRAN╟AISES ANTWERPEN</Name>
I would want to handle the exception such that the code handles this exception and procedes to the next node for processing without exiting the application.
Is this possible in C programming?
In C sure, if you write your own parser.
But you specified that you are using Libxml2. So you should look at the docs for that to figure out how to do that.
maybe it is easier just to fix the file - to make it proper xml before parsing. I'm not sure the libxml parser is supposed to partially parse broken xml
Any suggestions on how to fix the file ? I cannot make out what encoding is used in the XML file. Is there a way to convert the XML file to a standard encoding, say utf-8 without knowing what the actual encoded format is?
This is imperative as I cannot ascertain the encoding from the XML source.
Have you tried to open file in the textpad and then Save AS... and chose the UTF-8 encoding?Quote:
Originally Posted by dunxton
or Add at the begining of the xml-file string
with the correct version and encoding?Code:<?xml version="1.0" encoding="UTF-8"?>
That encoding looks like 0-255 ASCII. The vertical bar before the "AIS" is the bad character - it is not UTF-8. It appears to be extended ASCII, like from the old DOS days. See here: http://www.asciitable.com/
0XC7 is decimal 199, and that matches the table at that link. (Scroll down to see the high codes).
edit: Try this code page: 437. http://www.microsoft.com/globaldev/r...e/oem/437.mspx
dunxton, why are you having a hard time accepting the fact that it's INVALID XML? Since it's invalid, you can treat it however you want. The best you can do is ask the "3rd party" how they are incorrectly encoding the text so that you can make a best guess as to what the original text was supposed to be. Perhaps it really is just ISO-8859-1 and they are only using the encoding declaration - who knows.
Or better yet, tell them to fix it!
http://www.codeguru.com/forum/showthread.php?t=469816
gg
Here's an example of how easy this would be to correct with a Java program:
http://bytes.com/groups/net-c/229845...p437-text-file
Changing between code-pages and Unicode is one thing (which, yes, is easier in Java). But the only way to "correct" the XML is know how the author incorrectly encoded it. If dunxton (kmajay) is lucky, the author simply used a UTF8 encoding declaration, when they should of used ISO-8859-1.
ISO-8859-1 code page in Windows (28591): http://www.microsoft.com/globaldev/r...iso/28591.mspx
Latin Capital Letter C With Cedilla in UTF8 and other code pages: http://www.tachyonsoft.com/uc0000.htm#U00C7
gg
Why does the XML have to be corrected? Why wouldn't this work?
(I guess in answer to my own question "why might it not work", might be because the parser does not recognize that encoding.)Code:<?xml version="1.0" encoding="cp437"?>
Oooooooh, I get it now. That is supposed to be a C (of some sort). cp437 is not what you want, but probably what Codeplug suggested. So, change the header to read
which is a pretty standard codpage, and the parser should have no issue with it.Code:<?xml version="1.0" encoding="ISO-8859-1"?>
Without encoding="..." specified, libxml2 is probably defaulting to UTF-8.
Thanks guys for the help.
Codeplug, I very well acknowledge the fact that it is an invalid XML. But I am in no position to go back to the Vendor of the XML file for correction. So I was looking for alternative options. But then , as you said, there is probably no other option other than guess the encoding and try.
But what I dont understand is:
The input string is "AMITIÉS FRANÇAISES ANTWERPEN" . This string has two special characters: É and Ç and both are stored as 2 bytes in UTF-8 and as one byte in iso-8859-1. However, when I parse it, the error is happening only with Ç and not with É . If the parser fails to parse Ç as UTF-8 then it should fail to parse É as well. Why is this not happening?
Sounds like a question for the author of the libxml2 library.
>> But I am in no position to go back to the Vendor of the XML file for correction.
What would stop you? Do they not exist anymore?
>> no other option other than guess
Yep. Or you could ask the Vender what they're doing. If that truly isn't an option then with a few more data points you could at least make an educated guess.
>> when I parse it, the error is happening only with Ç and not with É
Open up the file in a hex editor. What byte(s) do you see for the É character? Here are the possibilities: http://www.tachyonsoft.com/uc0000.htm#U00C9
gg