It would be very fine if BASIC Authentication would be enforced by default when
one opens an URLConnection in OpenVMS and other platforms. But that is not the
case. Instead it goes into some kind of NTLM conversation and fails in OpenVMS
on JDK 5.0 (that can be seen when I run Ethereal)
I have found the sources below on the web: And I like to see now how it is
going for BASIC Authentication on my OpenVMS machine. So far when I try to
debug I can see the code called in rt.jar from sun. I can see code because
src.zip is in the source class path. But as the rt.jar was not complied with
the -g option I can't see any argument values but bunches of asynchronity when
doing single steping; Hence I cannot see the argument values passed so deep in
code from the java system properties into critical rt.jar routines used to open
an connection with the proper authorization and proxy data, called.
How would you go from there to see argument values passed if arguments belong
to sun rt.jar code not complied with the -g option; I can only see argument
values in my own short code. Do I get somewhere an rt.jar which was compiled
with the -g option or do I have to make that myself a project?
/*
* TestJDKNTLM.java
*
* Created on 30. August 2010, 15:19
*
*/
package axawl.spezpla.jdkntlmauth;
import java.io.BufferedReader;
import java.io.IOException;
import java.io.InputStreamReader;
import java.net.URL;
import java.net.URLConnection;
/**
*
* @author C770817
*/
public class TestJDKNTLM {
private void execute() throws IOException {
#1 System.setProperty("proxySet", "true");
#2 System.setProperty("http.proxyHost", "bcproxy.ch.winterthur.com");
#3 System.setProperty("http.proxyPort", "8080");
#4 System.setProperty("http.proxyUser", "C770817");
#5 System.setProperty("http.proxyPassword", "mydomainpassword");
#6 System.setProperty("http.auth.ntlm.domain", "ourdomain");
URL url = new URL("http://www.ks-net.ch";);
URLConnection conn = url.openConnection();
conn.setAllowUserInteraction(true);
BufferedReader reader = new BufferedReader(new
InputStreamReader(conn.getInputStream()));
int read = 0;
String body = null;
do {
body = reader.readLine();
System.out.println(body);
} while (body != null);
}
/**
* @param args the command line arguments
*/
public static void main(String[] args) throws IOException {
new TestJDKNTLM().execute();
}
}
Commenting #1 above out shows no effect on Vista and on OpenVMS
Commenting #4, #5, #6 out has no effect on Vista and on OpenVMS
Commenting #2 and/or #3 out shows an effect on OpenVMS and VISTA
I have removed all properties from files like build.properties;
So ANT and JAVA does not know this properties unless code makes em known.
Effe4ct: on VISTA NTLM takes place and all works due to NTLM works, on OpenVMS
I get 407 errors and al fails!
That is to say: NTLM authentication still works because the magic stuff is at
the sun corner for VISTA and fails short for the OpenVMS part, but is done in
the JDK itself with little to no effect doing such things here in code.
But what is going on there, that remains the big question?
To make it worse - stepping through remote code at OpenVMS shows asynchronity
at the rt.jar level; some lines are hit as they should and some are not. That
means the sources and or JDK versions used on VISTA and Open VMS are the same
but not really! JDK's and/or sources of JDK 1.5 may vary in minor ways, making
it hard to debug using a modern IDE like NetBeans and remote debugging. Guess I
have to use stone time jdb on OpenVMS to not be enforced to remote debugging
and asynchronity when doing single stepping.
So far the nice properties;
On OpenVMS I can conclude: That BASIC authentication does not work if the proxy
server asks for NTLM and BASIC because the client side code at OpenVMS does not
know how to handle it. And this magic belongs to rt.jar (so far quality and
portability among platforms and testing).
Josef
-----Ursprüngliche Nachricht-----
Von: [email protected] [mailto:[email protected]] Im Auftrag
von Anders Hammar
Gesendet: Montag, 30. August 2010 15:30
An: Maven Users List
Betreff: Re: maven-1.1 build-bootstrap fails with OpenVMS
AFAIK, basic auth is what is used out-of-the-box. If you want something
else, then you need some configuration. So you should be fine.
/Anders
On Mon, Aug 30, 2010 at 15:15, Stadelmann Josef <
[email protected]> wrote:
> Anders
>
> Thanks for the info, I went through that already, A DN it is formaven2 and
> maven3
>
> However as you can read as well .. just below the box on tghe mentioned
> page/like
> http://maven.apache.org/guides/mini/guide-proxies.html
>
> Please note that currently NTLM proxies are not supported as they have not
> been tested.
> You may be able to use the relevant system properties on JDK 1.4+ to make
> this work.
>
> ++> would be nice to know how?
>
> So what java system properties do I have to set to make JDK work and force
> to use only BASIC authentication?
>
> Josef
>
>
> -----Ursprüngliche Nachricht-----
> Von: [email protected] [mailto:[email protected]] Im
> Auftrag von Anders Hammar
> Gesendet: Montag, 30. August 2010 14:46
> An: Maven Users List
> Betreff: Re: maven-1.1 build-bootstrap fails with OpenVMS
>
> http://maven.apache.org/guides/mini/guide-proxies.html
>
> /Anders
>
> On Mon, Aug 30, 2010 at 14:34, Stadelmann Josef <
> [email protected]> wrote:
>
> > Stephen
> >
> > I agree with some of your points,
> >
> > But please tell me how authentication with a proxy server is handled in
> > maven-2 or maven-3?
> >
> > If authentication toward a proxy server is handled the same way as in
> > maven-1
> > then there is at least one pile left in maven-2 or maven-3 :-),
> >
> > But given there are no such piles left; how then is authentication
> handled
> > in maven-2 or maven-3?
> >
> > Most important for me:
> > Can I force maven-2 or maven-3 to use only BASIC authentication (user
> name
> > AND password) but not
> > NTLM or DIGEST toward a proxy server to reach out for artifact
> downloading,
> > this because I run
> > from OpenVMS, which might be not as good as Linux or Windows is in
> dealing
> > with NTLM Authentication.
> >
> > Josef
> >
> >
> >
> >
> > -----Ursprüngliche Nachricht-----
> > Von: Stephen Connolly [mailto:[email protected]]
> > Gesendet: Montag, 30. August 2010 12:45
> > An: Maven Users List
> > Betreff: Re: maven-1.1 build-bootstrap fails with OpenVMS
> >
> > There was a ____MAJOR____ refactoring between Maven 1.x and Maven 2.x
> >
> > If a project is built using Maven 1.x you have no hope to build it with
> > Maven 2.x unless you convert the build process yourself (or unless
> somebody
> > else has converted it for you)
> >
> > Maven 1.x development pretty much ceased at least 4-5 years ago
> >
> > I think you are out of luck unless you can use a Maven Repository Manager
> > (e.g. Nexus) to proxy the requests for you
> >
> > -Stephen
> >
> > P.S. Stephen has never done any development work with Maven 1.x. Stephen
> > thinks Maven 1.x is a stinking pile of excrement... which makes an
> > excellent
> > fertiliser from which the roses that are Maven 2.x and Maven 3.x have
> grown
> > ;-)
> >
> > On 30 August 2010 09:56, Stadelmann Josef <
> > [email protected]> wrote:
> >
> > > We are using a feature with Axis2-1.2 which is called
> scope="soapsession"
> > > That is to say, we have a state-full web service realized and
> interfaced
> > > with Microsoft WCF 3.5 Clients.
> > > State-full web services with long lasting user sessions is a thing
> which
> > > makes every bode I've meet so far cry.
> > >
> > > Hence I never got real help on that issues when I need it, hence I have
> > > built axis2-1.2 in the past from scratch myself just to be able to
> single
> > > step through the engine code to master this features; sorry but we use
> > long
> > > lasting session because our legacy server code some 120'000 line of
> > OpenVMS
> > > Pascal code is as it is. And yes: all is running fine. But since some
> few
> > > weeks my destroy() method is no longer called when the user session
> > breaks.
> > >
> > > Axis2-1.2 needs at a max maven-1.1 for building.
> > >
> > > I do it routinely at home not sitting behind a fire wall proxy
> demanding
> > > username and password and domain name to reach out for artifacts; but
> > here
> > > in the AXA company I can't reach out for artifacts unless this NTLM
> > > authentication is properly supported.
> > >
> > >
> > > Also: WHO OF YOU MAVEN GUYS CAN TELL ME how to best approach?
> > >
> > > 1. Does Axis2-1.5 and greater support scope="soapsession" correctly;
> I've
> > > read somewhere that this is not any longer the case; Unfortunately I
> lost
> > my
> > > pointer to this information, hence I can't ask the author. And again;
> > this
> > > feature seems to be so exotic but is there, that nobody helps me.
> > >
> > >
> > > 2. WHO OF YOU MAVEN DEVELOPERS can tell me that maven-2.2.1 and above
> > used
> > > to build Axis2-1.5 and above can reach through proxies using username
> > > password and domain name and is able to deal with NTLM authentication
> > > correctly?
> > >
> > > If this does not run on maven-1.1 why should it run on maven-2 and
> > maven-3?
> > >
> > >
> > > I will certainly go any build maven-2.2.1 and port it to OpenVMS and
> hope
> > > that it can reach through proxy protected networks to get artifacts
> > > downloaded.
> > >
> > > Josef
> > >
> > >
> > >
> > >
> > >
> > > -----Ursprüngliche Nachricht-----
> > > Von: Kathryn Huxtable [mailto:[email protected]]
> > > Gesendet: Dienstag, 24. August 2010 19:14
> > > An: Maven Users List
> > > Betreff: Re: maven-1.1 build-bootstrap fails with OpenVMS
> > >
> > > Egads, is anyone still using Maven 1? If so, why? -K
> > >
> > > On Aug 24, 2010, at 11:40 AM, Wayne Fay wrote:
> > >
> > > >> Wasting time?
> > > >>
> > > >> How do I force BASIC authentication using maven-1.1 either with
> > > >> maven-1.1 or maybe I can do it outside maven-1.1-code through a
> > > >
> > > > Let me be more plain... you are WASTING YOUR TIME trying to fix
> > > > Maven1. Perhaps you understand me better now?
> > > >
> > > > Find the simplest thing that could possibly work -- just download the
> > > > artifacts into the proper spot in your local filesystem -- and get on
> > > > with the more important things, which is apparently debugging your
> > > > problems Axis.
> > > >
> > > > Wayne
> > > >
> > > > ---------------------------------------------------------------------
> > > > To unsubscribe, e-mail: [email protected]
> > > > For additional commands, e-mail: [email protected]
> > >
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: [email protected]
> > > For additional commands, e-mail: [email protected]
> > >
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: [email protected]
> > > For additional commands, e-mail: [email protected]
> > >
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [email protected]
> > For additional commands, e-mail: [email protected]
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
