NDT-DEV email list created

[Dominic Hamon <>]

Committed as r787.

Thanks for all the great feedback :)

- dominic

On Sun, Oct 28, 2012 at 8:14 AM, Kavitha Kumar <> wrote:

On Oct 25, 2012, at 6:32 PM, Dominic Hamon wrote:

Ah, nice catch. I've changed it a little from your suggestion:

-  if (args.length != 1) {
+  if (args.length < 1 || args.length > 2) {
+       System.out.println("Usage: java -jar Tcpbw100.jar <hostname> [client-id]");
       System.exit(0);
    }

WDYT?

Perfect! 

I'd like the data to be available in the 'generic' section so it sits alongside similar information - is there any reason to remove it from 'additional data'?

No, pls commit ! Thanks, Dominic.

Once you confirm you agree with this, I'll commit.

- dominic

On Thu, Oct 25, 2012 at 11:15 AM, Kavitha Kumar <> wrote:

Hi Dominic,

  In your latest addition, the lines in the Applet that check for the number of command-line arguments will fail, since it necessitates just one argument.
These are the lines:
-               if (args.length != 1) {
-                       System.out.println("Usage: java -jar Tcpbw100.jar " + "HOST");
+                       System.out.println("Usage: java -jar Tcpbw100.jar <hostname> [client-id]");
                }

I modified this, and ran tests for clt/applet with and without the "client" param, and saw the client-name logged into the .meta file.

So you should be good to commit once these changes are incorporated.

I had another minor (mostly aesthetic) suggestion: with your changes, the client name is logged into both the "generic" and "additional data" sections of the .meta file
with no difference in the actual values/data being logged. If you'd be happy with this data being present in just the "additional data" section of the .meta file, then your changes in the logging.c and test_meta_srv.c files dont need to go in at all. I'm sorry I did'nt suggest this earlier. Its not a necessary change, so, going either way is fine with us here at I2.

Thanks,
Kavitha

------- my changes were these ---------------------------------------------

-  if (args.length != 1) {
+  if (args.length < 1) {
+       System.out.println("Usage: java -jar Tcpbw100.jar <hostname> [client-id]");
       System.exit(0);
    }

---------------------------------------------------------------------------

----- Original Message -----
> From: "Dominic Hamon" <>

> To: "Kavitha Kumar" <>
> Cc:

> Sent: Wednesday, October 24, 2012 4:23:49 PM
> Subject: Re: [ndt-dev] Re: M-Lab metadata patch
>
> Error on my part. New patch takes optional second command-line
> argument to set the client identifier.
>
>
>
>
>
>
> On Wed, Oct 24, 2012 at 1:11 PM, Kavitha Kumar <
> > wrote:
>
>
> Errata:
> "Assigned applet._sClient to args[0]" should have been
> "Assigned args[0] to applet._sClient"
>
>
>
>
> ----- Original Message -----
> > From: "Kavitha Kumar" < >
> > To: "Dominic Hamon" < >
> > Cc:
> > Sent: Wednesday, October 24, 2012 4:01:52 PM
> > Subject: Re: [ndt-dev] Re: M-Lab metadata patch
> >
> > Hi Dominic,
> >
> > This one addresses my questions, thanks! Here're my remaining
> > questions:
> > 1. NDTConstants.java:
> > We need to maintain EIGHT as a "double". This helps stop overflow
> > leading to negative test results. So your change to this line
> > will have to be backed off.
> > public static final double EIGHT = 8.0; // Used in octal number,
> > conversions from
> > - // Bytes-> bits etc
> > + public static final int EIGHT = 8; // Used in octal number,
> > conversions from Bytes-> bits etc
> >
> > 2. Tcpbw100.java:
> >
> > You assigned applet._sClient to args[0], which is the server
> > hostname. If this was intentional, please add a comment as to why
> > you assign it to the host name by default.
> >
> > thanks,
> > Kavitha
> >
> > ----- Original Message -----
> > > From: "Dominic Hamon" < >
> > > To:
> > > Sent: Wednesday, October 24, 2012 3:10:25 PM
> > > Subject: [ndt-dev] Re: M-Lab metadata patch
> > >
> > > I've updated the patch as there were some merge issues with the
> > > last
> > > one.
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > On Wed, Oct 24, 2012 at 9:34 AM, Dominic Hamon <
> > >
> > > >
> > > wrote:
> > >
> > >
> > > Hello
> > >
> > >
> > > As discussed previously, we have a patch that we use to identify
> > > clients running NDT against M-Lab servers. I'd like to get this
> > > patch into trunk so that we can stop maintaining our own branch
> > > and
> > > encourage developers to use the trunk version of NDT when
> > > building
> > > tools to run on M-Lab.
> > >
> > >
> > > I've attached the patch here and would appreciate your opinion.
> > >
> > >
> > > This patch has been rebuilt from a branch that was behind trunk
> > > by
> > > some way, so it's possible that there are some errors. I have
> > > built
> > > and tested locally, however, and did not spot anything obvious.
> > >
> > >
> > >
> > >
> > >
> > >
> > > - dominic
> > >
> > >
>
>