Re: Handling malformed uri with WebContent

Hi Jérôme,

Thank you for reporting this problem.

I've created an issue to fix this problem:
https://issues.apache.org/jira/browse/APA-47

I think RewritableHttpReverseProxyServiceImpl may need to replace some disallowed characters such as space before creating HTTP Client Methods.
I think it's a good time to release apa-webcontent-1.2 after fixing this problem.
I will try to release 1.2 in a couple of weeks.

By the way, you may consider to provide a custom proxyPathMapper by setting it to a request attribute.
RewritableHttpReverseProxyServiceImpl checks if there exists one custom one by the following line:


    HttpReverseProxyPathMapper proxyPathMapper = (HttpReverseProxyPathMapper) request.getAttribute(HttpReverseProxyConstants.PATH_MAPPER);

    // ...
    String proxyTargetURL = proxyPathMapper.getRemoteURL(localPathInfo);

And, it invokes proxyPathMapper.getRemoteURL(). So, you may implement a custom proxyPathMapper for the specific website in order to replace disallowed characters (in the overriden method). You can set it to request attribute in the pre-processing filter.
But, anyway, this doesn't look a nice and easy solution.

I'll try to fix it and keep you updated.

Regards,

Woonsan



----- Original Message -----
> From: "[email protected]" <[email protected]>
> To: [email protected]
> Cc: 
> Sent: Friday, January 13, 2012 10:31 AM
> Subject: Handling malformed uri with WebContent
> 
> 
> Hello
> 
> First, just a word to say that our application has been released and is
> running in "real life", and it works!
> 
> Second, I currenlty work on an integration with Lotus Notes, the goal is to
> show a view  stored in Notes to see them through jetspeed so that users
> cannot see the Notes machines.
> 
> The first step is to provide public document (no autentification stuff now,
> just access through a proxy)
> 
> For that, I configued a proxy with the reverseProxyIFrame portlet: document
> have there url.
> Basically it works fine, I can even navigate through links inside Notes,
> thanks to the reverse proxy feature.
> I have a problem left, it's that some uri of documents are malformed and
> contain space, and webcontent proxyx (I use the 1.1 version) doesn't encode
> them with the default configuration.
> 
> Is there a mean to encode the url before parsing it?
> 
> here is the return exception
> java.lang.IllegalArgumentException
>     java.net.URI.create(URI.java:842)
>     org.apache.http.client.methods.HttpGet.<init>(HttpGet.java:69)
> 
> org.apache.portals.applications.webcontent.proxy.impl.RewritableHttpReverseProxyServiceImpl.invoke
> (RewritableHttpReverseProxyServiceImpl.java:289)
> 
> org.apache.portals.applications.webcontent.proxy.impl.DefaultHttpReverseProxyServlet.service
> (DefaultHttpReverseProxyServlet.java:436)
>     javax.servlet.http.HttpServlet.service(HttpServlet.java:717)
> 
> cause m�re
> 
> java.net.URISyntaxException: Illegal character in path at index 108:
> http://SECRETSERVER.bnf.fr/DSI/PiloteProdV6.nsf/0/F8C57293124EDBF9C125783200376835/$FILE/EXO-ADM-2011-000020-01
> (p2).doc
>     java.net.URI$Parser.fail(URI.java:2809)
>     java.net.URI$Parser.checkChars(URI.java:2982)
>     java.net.URI$Parser.parseHierarchical(URI.java:3066)
>     java.net.URI$Parser.parse(URI.java:3014)
>     java.net.URI.<init>(URI.java:578)
>     java.net.URI.create(URI.java:840)
>     org.apache.http.client.methods.HttpGet.<init>(HttpGet.java:69)
> 
> Cordialement,
> -----------------------------------------------
> Jérôme Dupont
> Bibliothèque Nationale de France
> Département des Systèmes d'Information
> Tour T3 - Quai François Mauriac
> 75706 Paris Cedex 13
> téléphone: 33 (0)1 53 79 45 40
> e-mail: [email protected]
> -----------------------------------------------
> 
> Exposition  Casanova, la passion de la Liberté  - du 15 novembre 2011 au 19 
> février 2012 - BnF - François-Mitterrand / Grande Galerie Avant d'imprimer, 
> pensez à l'environnement. 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>