tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From André Warnier>
Subject Re: Problem with Transfer-Encoding
Date Sat, 05 Jul 2014 23:36:56 GMT
Sushil Prusty wrote:
> Dear User
> Thanks for you input.

You're welcome.
First, a foreword : I will try my best to help you, but doing this is very difficult, and

doing it via email is even more difficult.
I was not kidding when I wrote earlier that even looking at the data may make it change.
Of course, that is not really true, but the fact of cutting and pasting this data, from 
your saved HTTPFox trace into an email that you send to the Tomcat list, and then the 
Tomcat list server forwarding this to other people in a new email, may again decode and 
re-encode this data several times, and confuse the situation totally.
So we need to be very, very systematic, and make sure that what we see is really what we 
get, ok ?

What you should really do, is to save the original HttpFox data to a file, then save that

file, then zip that file, then post it somewhere where we can get this zip-file.
So that we can download it, unzip it, and then be sure that we are really seeing the same

data as you do.

In the meantime, a question :

> I just debugged using HttpFox here is below you find header
> (Request-Line)    POST /test/

The above request line is triggered by something.
By what ?
Is that a link or button on a HTML page which is currently loaded in your browser ?

If yes, then before you actually click this link, can you in your browser use the 
"View..Character set" function, and tells us what the browser thinks about the current 
page loaded in the browser, before you even send this request to the server ?

The reason why I am asking, is that this is the character set which the browser will most

probably use to encode the text data that it sends to the server (when you click the link).

Then see the note below, in the text.

> HTTP/1.1
> Host    **********
> User-Agent    Mozilla/5.0 (Macintosh; Intel Mac OS X 10.8; rv:30.0)
> Gecko/20100101 Firefox/30.0
> Accept    text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
> Accept-Language    en-US,en;q=0.5
> Accept-Encoding    gzip, deflate
> Referer    https://s
> ************
> Cookie    JSESSIONID=******************; doNotShowStartupOnLoad=true
> Connection    keep-alive
> Content-Type    multipart/form-data;
> boundary=---------------------------*******************
> Content-Length    4039
> In Post body
> -----------------------------1550434539176507601876254213
> Content-Disposition: form-data; name="disclaimerText"
> Zażółć gęślÄ jaźń! ta funkcjonalność nie jest wspierana

The line above may or may not have been further corrupted (compared to the original that 
you see), by the simple fact of copying this text into your email.
But assuming for a moment that it was not, and that it really is what it looks like above,

there is some kind of a problem :

(You'll have to follow carefully here)
If I take the original text line which you posted in your first message :

Zażółć gęślą jaźń! ta funkcjonalność nie jest wspierana*

and I imagine that internally, this is encoded as UTF-8;
Then if I look at that same series of UTF-8 characters, but now examine the *bytes* that 
compose these characters and view them in ASCII, I should see this :

Zażółć gęślą jaźń! ta funkcjonalność nie jest wspierana*

But if you compare this carefully, with the string as it appears in your HttpFox trace, 
you will see that it does not match exactly. For example, look at the last 2 letters of 
the word "funkcjonalność", in both versions.

So there appears to be some discrepancy between the character set which your browser is 
really using (to send data to the server), and the UTF-8 that your server seems to expect.

Furthermore (and put this on account of my suspicious nature if you want) :
The second part of that message, in Polish, means : "This functionality is not supported".
Which triggers the question : what kind of HTML page would be sending this phrase, as part

of the data, in a POST to a server ?  Can you give us some context as to what you are 
trying to do here ?

> -----------------------------1550434539176507601876254213
> I believe psot data got changed before   reaching   to server .
>   Do you have any ideas what's wrong here, where the error might be ?
> On Sat, Jul 5, 2014 at 9:08 PM, André Warnier <> wrote:
>> Konstantin Kolinko wrote:
>>> 2014-07-05 9:24 GMT+04:00 Sushil Prusty <>:
>>>> Hello,
>>>> summary of my Problem:
>>>> When a client POSTs Tranfer-Encoding data   using browser ,
>>>> my server is not processing the request character encoding properly .
>>>> I send the following request:
>>>> Content-Type text/html;charset=UTF-8
>>>> Date Sat, 05 Jul 2014 05:10:09 GMT
>>>> Server Apache-Coyote/1.1
>>>> Transfer-Encoding chunked
>>>> *disclaimerTextZażółć gęślą jaźń! ta funkcjonalność nie jest wspierana*
>>>>   Full details:
>>>> My application running on  apache-tomcat-7.0.40
>>>>  and Java
>>>> 1.6.x)  in linux box.
>>>> Below response is changed once it's reach to my controller
>>>> *ZażóÅÄ gÄÅlÄ jaźÅ! ta funkcjonalnoÅÄ nie jest wspierana*
>>>> I have below configuration  in server.xml
>>>>  <Connector port="80" protocol="HTTP/1.1" connectionTimeout="20000"
>>>> maxPostSize="5242880" maxParameterCount="25000"/>
>>>>     <Connector
>>>>             port="443"
>>>>             protocol="HTTP/1.1"
>>>>             scheme="https"
>>>>             noCompressionUserAgents="gozilla, traviata"
>>>> compressableMimeType="text/html,text/xml,text/javascript,
>>>> text/css,application/javascript,application/json"
>>>>             URIEncoding="UTF-8"
>>>>     />
>>>> and in my
>>>> set JAVA_OPTS=-Djavax.servlet.request.encoding=UTF-8
>>>> -Dfile.encoding=UTF-8
>>>> (...)
>>> As a sanity check:
>>> 1) That "I send the following request" listing looks more like a
>>> response, not a request. (E.g. the "Server Apache-Coyote/1.1" header
>>> makes no sense in a request).
>>> So you are lying somewhere.
>>> There is no point for me to try guessing what you are doing. You may
>>> have confused "reading" with "writing" somewhere, and without source
>>> code one cannot verify your words.
>>> You have to provide a step-by-step instruction and enough source code
>>> so that a person who is not familiar with your system were able to
>>> reproduce your problem.
>>> 2) Content-Type says "text/html", but that line of text is not a valid
>>> HTML document.
>> +1
>> Character encoding/decoding issues are hell to debug as it is, because
>> they are like quantum physics : even looking at them can change them.(*)
>> So you need to provide *accurate* and "raw" information, otherwise it is
>> just a loss of time for everyone.
>> Use a browser plugin like HttpFox, LiveHttpHeaders, HttpFox or similar to
>> monitor the requests being sent and responses being received, at the
>> browser level.  All these plugins allow you to selectively dump
>> requests/responses to a file.  Do that.
>> Also, check in your browser that when you receive a response page back
>> from the server, your browser is really seeing this response in the proper
>> character set (use "View.. Character encoding..").
>> "Transfer Encoding" has nothing to do with the *character encoding* of
>> either the request or the response.  The little imprecise data that the OP
>> provided above /suggests/ that there is some double encoding taking place
>> /somewhere/, but so far it could as well be in the email client that he
>> used to post to the list, as anywhere else.
>> (*) with the wrong editor, or the wrong locale e.g.
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail:
>> For additional commands, e-mail:

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message