We Keep Coding

Issue with soap call and content length

I have written a java swing app that is sending SOAP requests based on this code here
Overall it is working great, however I have just started testing it when parsing Chinese characters in the soap:BODY and this causes an error where I get a 400 response from the Web Server:
s.AddParameter("xml", "班");
Using wireshark i eventually tracked it down to the Content-Length value that was constructed being incorrect when parsing these Chinese characters (and i am assuming any other multibyte(?) character).
I have proven this by overriding the content length generation by simply changing the code to this:
out.println("Content-Length: " + String.valueOf(postData.length()+2));
Obviously this is not a solution as it only proved my very isolated test case of sending a single character, but i believe the issue is that the postData.length() is calculated first and then on posting the data my 班 character is then converted to \347\217\255, throwing the content-length out and causing the request to fail.
So I am asking for advice on how to resolve this issue?
Is it possible for me to encode the value first, obtain the content length and suppress the encoding on the post? I am unsure what is actually encoding it; the PrintWriter i am assuming?
Regards.

Does javascript by default have support for UTF-8

I want to know whether javascript is by default UTF-8 compliant or not. If no, then how to make it UTF-8 compliant.
I am sharing you the snapshot, i have used german charterers in input type email as : αβγδεζηθ in the image shown below, which when entered shows ok, but after the request is made to java to save this value in database, the value in this field shows unknown charterers.
Is this javascript issue or java issue ..

javascript supports unicode.
Proof:
ﾟωﾟﾉ=/｀ｍ´）ﾉ~┻━┻//*´∇｀*/['_'];o=(ﾟｰﾟ)=_=3;c=(ﾟΘﾟ)=(ﾟｰﾟ)-(ﾟｰﾟ);(ﾟДﾟ)=(ﾟΘﾟ)=(o^_^o)/(o^_^o);(ﾟДﾟ)={ﾟΘﾟ:'_',ﾟωﾟﾉ:((ﾟωﾟﾉ==3)+'_')[ﾟΘﾟ],ﾟｰﾟﾉ:(ﾟωﾟﾉ+'_')[o^_^o-(ﾟΘﾟ)],ﾟДﾟﾉ:((ﾟｰﾟ==3)+'_')[ﾟｰﾟ]};(ﾟДﾟ)[ﾟΘﾟ]=((ﾟωﾟﾉ==3)+'_')[c^_^o];(ﾟДﾟ)['c']=((ﾟДﾟ)+'_')[(ﾟｰﾟ)+(ﾟｰﾟ)-(ﾟΘﾟ)];(ﾟДﾟ)['o']=((ﾟДﾟ)+'_')[ﾟΘﾟ];(ﾟoﾟ)=(ﾟДﾟ)['c']+(ﾟДﾟ)['o']+(ﾟωﾟﾉ+'_')[ﾟΘﾟ]+((ﾟωﾟﾉ==3)+'_')[ﾟｰﾟ]+((ﾟДﾟ)+'_')[(ﾟｰﾟ)+(ﾟｰﾟ)]+((ﾟｰﾟ==3)+'_')[ﾟΘﾟ]+((ﾟｰﾟ==3)+'_')[(ﾟｰﾟ)-(ﾟΘﾟ)]+(ﾟДﾟ)['c']+((ﾟДﾟ)+'_')[(ﾟｰﾟ)+(ﾟｰﾟ)]+(ﾟДﾟ)['o']+((ﾟｰﾟ==3)+'_')[ﾟΘﾟ];(ﾟДﾟ)['_']=(o^_^o)[ﾟoﾟ][ﾟoﾟ];(ﾟεﾟ)=((ﾟｰﾟ==3)+'_')[ﾟΘﾟ]+(ﾟДﾟ).ﾟДﾟﾉ+((ﾟДﾟ)+'_')[(ﾟｰﾟ)+(ﾟｰﾟ)]+((ﾟｰﾟ==3)+'_')[o^_^o-ﾟΘﾟ]+((ﾟｰﾟ==3)+'_')[ﾟΘﾟ]+(ﾟωﾟﾉ+'_')[ﾟΘﾟ];(ﾟｰﾟ)+=(ﾟΘﾟ);(ﾟДﾟ)[ﾟεﾟ]='\\';(ﾟДﾟ).ﾟΘﾟﾉ=(ﾟДﾟ+ﾟｰﾟ)[o^_^o-(ﾟΘﾟ)];(oﾟｰﾟo)=(ﾟωﾟﾉ+'_')[c^_^o];(ﾟДﾟ)[ﾟoﾟ]='\"';(ﾟДﾟ)['_']((ﾟДﾟ)['_'](ﾟεﾟ+(ﾟДﾟ)[ﾟoﾟ]+(ﾟДﾟ)[ﾟεﾟ]+(ﾟΘﾟ)+(ﾟｰﾟ)+(ﾟΘﾟ)+(ﾟДﾟ)[ﾟεﾟ]+(ﾟΘﾟ)+((ﾟｰﾟ)+(ﾟΘﾟ))+(ﾟｰﾟ)+(ﾟДﾟ)[ﾟεﾟ]+(ﾟΘﾟ)+(ﾟｰﾟ)+((ﾟｰﾟ)+(ﾟΘﾟ))+(ﾟДﾟ)[ﾟεﾟ]+(ﾟΘﾟ)+((o^_^o)+(o^_^o))+((o^_^o)-(ﾟΘﾟ))+(ﾟДﾟ)[ﾟεﾟ]+(ﾟΘﾟ)+((o^_^o)+(o^_^o))+(ﾟｰﾟ)+(ﾟДﾟ)[ﾟεﾟ]+((ﾟｰﾟ)+(ﾟΘﾟ))+(c^_^o)+(ﾟДﾟ)[ﾟεﾟ]+(ﾟｰﾟ)+((o^_^o)-(ﾟΘﾟ))+(ﾟДﾟ)[ﾟεﾟ]+(ﾟΘﾟ)+(ﾟΘﾟ)+(c^_^o)+(ﾟДﾟ)[ﾟεﾟ]+(ﾟΘﾟ)+(ﾟｰﾟ)+((ﾟｰﾟ)+(ﾟΘﾟ))+(ﾟДﾟ)[ﾟεﾟ]+(ﾟΘﾟ)+((ﾟｰﾟ)+(ﾟΘﾟ))+(ﾟｰﾟ)+(ﾟДﾟ)[ﾟεﾟ]+(ﾟΘﾟ)+((ﾟｰﾟ)+(ﾟΘﾟ))+(ﾟｰﾟ)+(ﾟДﾟ)[ﾟεﾟ]+(ﾟΘﾟ)+((ﾟｰﾟ)+(ﾟΘﾟ))+((ﾟｰﾟ)+(o^_^o))+(ﾟДﾟ)[ﾟεﾟ]+(ﾟｰﾟ)+(c^_^o)+(ﾟДﾟ)[ﾟεﾟ]+(ﾟΘﾟ)+((o^_^o)-(ﾟΘﾟ))+((ﾟｰﾟ)+(o^_^o))+(ﾟДﾟ)[ﾟεﾟ]+(ﾟΘﾟ)+((ﾟｰﾟ)+(ﾟΘﾟ))+((ﾟｰﾟ)+(o^_^o))+(ﾟДﾟ)[ﾟεﾟ]+(ﾟΘﾟ)+((o^_^o)+(o^_^o))+((o^_^o)-(ﾟΘﾟ))+(ﾟДﾟ)[ﾟεﾟ]+(ﾟΘﾟ)+((ﾟｰﾟ)+(ﾟΘﾟ))+(ﾟｰﾟ)+(ﾟДﾟ)[ﾟεﾟ]+(ﾟΘﾟ)+(ﾟｰﾟ)+(ﾟｰﾟ)+(ﾟДﾟ)[ﾟεﾟ]+(ﾟｰﾟ)+((o^_^o)-(ﾟΘﾟ))+(ﾟДﾟ)[ﾟεﾟ]+((ﾟｰﾟ)+(ﾟΘﾟ))+(ﾟΘﾟ)+(ﾟДﾟ)[ﾟoﾟ])(ﾟΘﾟ))('_');
JsFiddle
The above code prooves that js indeed supports unicode. The above code alerts "Hello World" and is taken from here
So, it is not js' fault. Fault is somewhere else, or the system is unable to show those characters.

JavaScript is definitely UTF-8 compliant.
However, there are many points of failure for UTF-8. You will need to make sure that your server-side code can handle UTF-8. Then check that your database columns use UTF-8 as their character set. Then ensure that the connection between your server-side code and your database is using UTF-8.
The "conversion" of characters to ?s suggests to me that it's the database that's missing encoding information.

JavaScript (the language) comes with Unicode support (all strings are inherently Unicode), but not with default UTF8 support. That is part of the APIs and their implementation. For example, in a browser the DOM does support it.
When you are sending data from a browser to a server, a) you need to set the correct headers b) send the correct data and the browser will do as expected. Then, c) the server needs to understand this request and use the correct encoding for deserialising the data. If something does not work, ask a question that shows your code.

Handle multiple language encoding

In my application, I read tweets from twitter, but the tweets are not language restricted. So when I am trying to send response for a Chinese/Japanese tweet the content is not displayed correctly. I have currently set the
response.setContentType("text/html;charset=UTF-8");
before sending the response.
How can we handle multiple languages?
i can see the message sent
{"lastPost":{"lastUpdate":"毋成金口","pubDate":"Fri Aug 12 00:39:09 UTC 2011","message_id":101814948329562112}
this is a json string and added to the response..
on my client i.e iphone the lastpost is "????"

Telling the browser that the page is UTF-8 is a good thing, but useless unless you make sure that you are actually writing only UTF-8 in the page.
To make sure this happens :
Whenever you read, from twitter or whatever, always require UTF-8 data, make sure you are receiving UTF-8 bytes.
When you create a string from raw bytes, Java by default uses the "platform default encoding" which could be anything. Bytes to string conversion happens when creating a new String from a byte array or when using a Reader. Both these methods allow you to explicitly define the encpding you expect the bytes to be. Once point 1 is checked and you are receiving UTF-8 byes, make sure everywhere in your application you are specifying to use UTF-8 when converting bytes to strings.
when using a Writer, to convert strings to bytes sent for example to the browser (the servlet writer), the same rules apply : try to be explicit and always specify UTF-8
If you store stuff in databases, then you have two encoding problems. The first one is which encpding you database is using when talking to your application (connection encoding), the second is which encoding the database is actually storing strings in (storage encoding). Usually, you can specify only the connection encoding from Java, while the storage encoding is specified in the database when it is created (search for "collation" if you are using mysql).
Detecting where a string that is supposed to be UTF-8 gets reencoded badly is a hard task. 99% of the times, it is being converted to ISO-latin or similar encoding somewhere, which causes special characters like à or ì appear as two chars of garbage. Often debugging is the only way to find out where this happens.

the problem was with the client encoding.. it was set to ISO-

a question related to URL

Dear all,Now i have this question in my java program,I think it should be classified as URL problem,but not 100% sure.If you think I am wrong,feel free to recategorize this problem,thanks.
I would state my problem as simply as possible.
I did a search on the famouse Chinese search engine baidu.com for a Chinese key word "奥巴马" (Obama in English),and the way I do that is to pass a URL (in a Java Program)to the browser like:
http://news.baidu.com/ns?word=奥巴马
and it works perfectly just like I input the "奥巴马”keyword in the text field on baidu.com.
However,now my advisor wants another thing.Since he can not read the Chinese webpages,but he wants to make sure the webpages I got from Baidu.com is related to "Obama",he asked me to google translate it back,i.e,using google translate and translate the Chinese webpage to English one.
This sounds straightforward.However,I met my problem here.
If I simply pass the URL "http://news.baidu.com/ns?word=奥巴马" into Google Translate and tick "Chinese to English" translating option,the result looks awful.(I don't know the clue here,maybe related to Chinese character encoding).
Alternatively,if now my browser opens ""http://news.baidu.com/ns?word=奥巴马" webpage,but I click on the "百度一下" button (that simply means "search"),you will notice the URL will get changed,now if I pass this URL into the Google translate and do the same thing,the result works much better.
I hope I am not making this problem sound too complicated,and I appologize for some Chinese words invovled,but I really need your guys' help here.Becasue I did all this in a Java program,I couldn't figure out how to realize that "百度一下"(pressing search button) step then get the new URL.If I could get that new URL,things are easy,I could just call Google translate in my Java code,and pops out the new window to show my advisor.
Please share any of your idea or thougts here.Thanks a lot.
Robert

You could use
URLEncoder.encode("http://news.baidu.com/ns?word=奥巴马", "utf-8")
then pass the resulting URL to Google Translate like:
http://translate.google.com/translate?js=y&prev=_t&hl=en&ie=UTF-8&layout=1&eotf=1&sl=zh-CN&tl=en&u=YOUR_URL
Cheers

When you press the search button, the browser encodes the search term into %E5%A5%A5%E5%B7%B4%E9%A9%AC, which is the UTF-8 encoding for 奥巴马. It does this because UTF-8 is the default encoding for HTML forms.
Java uses a UTF-16 encoding internally, so it’s possible that the URL library builds a request in that encoding if you do not specify anything.
However, I could not reproduce your problem with Google translate — pasting that URL appeared to work correctly no matter how I did it.

Try calling
URLEncoder.encode("http://news.baidu.com/ns?word=奥巴马", "utf-8")
(or utf-16; I'm not quite familiar with the Chinese characters representation)

URLs can contain only ASCII characters. All other characters must be converted to bytes then %-encoded in ASCII. However there is no mandate on what charset is used to convert chars to bytes. UTF-8 is recommended, but not required. As long as a server expresses its preference on charset, the client should respect that and use the same charset for encoding.
You can see from page info that baidu uses gb2312 encoding. The characters 奥巴马 in a form on its page will be converted to bytes in gb2312: B0C2 B0CD C2ED, then %-encoded to %B0%C2%B0%CD%C2%ED. That is what actually sent to baidu server, http://www.baidu.com/s?wd=%B0%C2%B0%CD%C2%ED
Your OS happens to be configured to use gb2312 by default, therefore when you paste http://news.baidu.com/ns?word= 奥巴马 to the browser, browser does the same thing, and baidu gets the correct chars. When I paste that URL in my browser, it screws up, because my OS uses UTF-8, and the browser encodes these chinese characters in UTF-8, not something baidu expectes. (when entering a URL directly in a browser, the browser may not have communicated to the server and does not know the charset the server prefers, therefore the browser uses platform default charset)
Now, Google uses UTF-8. That's why if you paste the URL to google form, it will screw up just like on my OS. The chars are encoded in UTF-8, and baidu will try to parse it as gb2312, and gets totally wrong words.
Solution is easy. Just encode the parameter in the way that the server expects:
"http://news.baidu.com/ns?word=" + URLEncoder.encode("奥巴马", "gb2312")

Soap body is utf-8 encoded twice

We use a web service which expects UTF-8. The framework we use on the client is Apache Axis2. We call the web service and the soap body contains strings in UTF-8. The problem is that it seems like the body is "double encoded". I.e we have the character 'å'. The utf-8 representation of 'å' in utf-8 is C3 A5 however we see in our logs that the (double) encoded value sent is C3 83 C2 A5.
Has anyone experienced similiar problems?

It's not entirely clear how you're calling the web service. Does the method in the web service just take a string? If so, what does your string look like in Java? All strings in Java are UTF-16 encoded - if you're converting the UTF-8 binary representation into a string by taking each byte and turning it into a character, then that's the problem.
If you could show what the method you're calling looks like, and how you're calling it, that would help a lot.
For what it's worth, I've used Axis with non-ASCII strings with no problem in the past. I strongly suspect this is a problem with how you're using it rather than with Axis itself, although I'm willing to be proved wrong :)
EDIT: Based on your comment, it sounds like you've got problems receiving the HTML form data, before you hit the web service. If the user has typed "å" into the form, then that's what you should see when you debug in Eclipse. If you're putting bad data into your web service, it's no wonder you're getting bad data out at the other end. I suggest you run WireShark to see exactly what the browser is sending you, both in terms of the raw bytes and also what content encoding it's specifying. My guess is that your web server is treating it as ISO-8859-1 but it's actually UTF-8.
Once you've got the string correctly from the form, I suspect you'll find there are no problems at all in passing it on to the web service.