hc-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Idan Sheinberg (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HTTPCLIENT-1708) Issues With Extended ASCII ( Latin Chars ) Escaping with HttpClient 4.5.1
Date Mon, 28 Dec 2015 17:43:49 GMT

    [ https://issues.apache.org/jira/browse/HTTPCLIENT-1708?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15072946#comment-15072946

Idan Sheinberg commented on HTTPCLIENT-1708:


Before closing this as not a problem , can you please hear me out a bit more ??

I understand there's no particular charset encoding but that's not the issue here.
Using the toASCIIString() method causes character value manipulation - that the issue :

%CC%81  ( which the browser , URLEncode , curl and other tools generate  ) stands for UTF
hex code cc 81 , Unicode U+0301 :
first row in the table here  - http://www.utf8-chartable.de/unicode-utf8-table.pl?start=769

While %C3%A9 encoding ( which the toASCIIString() method generates ) stands for UTF hex code
c3 a9 , Unicode U+00E9 :
first row in the table here  - http://www.utf8-chartable.de/unicode-utf8-table.pl?start=233

These are different chars, different symbols/values with different meanings !

Don't you think it's a problem that line of code you library is using can trigger character
manipulation in certain cases such as this ( where U+301 becomes U+00E9 , due to the Normalizer
usage ?) 

> Issues With Extended ASCII ( Latin Chars ) Escaping with HttpClient 4.5.1
> -------------------------------------------------------------------------
>                 Key: HTTPCLIENT-1708
>                 URL: https://issues.apache.org/jira/browse/HTTPCLIENT-1708
>             Project: HttpComponents HttpClient
>          Issue Type: Bug
>          Components: HttpClient
>    Affects Versions: 4.5.1
>         Environment: Ubuntu Variant 15.04
> Netbeans 8.1 IDE
> Java 8 ( Oracle JDK 1.8.6x)
>            Reporter: Idan Sheinberg
> Hey Guys
> Trying to send an httpclient HttpHead request for the following url  - 'http://some.domain.com/amnetcanadaplatform/HomeDépot_Électros_WEB22s_video.mp4'
> I Notice it gets escaped as 'http://some.domain.com/amnetcanadaplatform/HomeD%C3%A9pot_%C3%89lectros_WEB22s_video.mp4'
> While other programs/utilities/frameworks expect it to be 'http://some.domain.com/amnetcanadaplatform/HomeDe%CC%81pot_E%CC%81lectros_WEB22s_video.mp4'
> I've done some digging up the source code and tracked the issue down  to  "toASCIIString()"
of the Java URI ( WHICH OF COURSE IS NOT YOUR RESPONSIBILITY )object being called in order
to retrieve the request line
> Class : org.apache.http.client.methods.HttpRequestWrapper
> Method : getRequestLine() 
> Line : 113
> Internally the line 'String ns = Normalizer.normalize(s, Normalizer.Form.NFC);' manipulates
the chars so their unicode value changes , which causes the 'inappropriate' values to be displayed
for the escaped URI 
> Class : java.net.URI
> Method : encode(String s)
> Line : 2723
> Now I know this would extra hard to even get to a point where I find out if this is a
java language issue , but I don't believe unicode manipulation  of bytes is a desired behavior
 for url encoding. Is there any known specific why you used the "toASCIIString()" method instead
of the plain "toString()" method
> Do you think there's a chance this issue can be resolved on your end ?

This message was sent by Atlassian JIRA

To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org

View raw message