IPv6 Guide - C/Java samples not comparable

[Bernd Eckenfels]

Hello,

Am 13.06.2013, 11:39 Uhr, schrieb Alan Bateman <Alan.Bateman at oracle.com>:
> There are details on how to configure it at the end of this page:
> http://docs.oracle.com/javase/7/docs/technotes/guides/net/ipv6_guide/index.html

Speaking of this guide, there are two problems with this as well:

First of all the Title is Java 5, but it is published in JDK 7  
Documentation set as well. For the Java 8 I would recommend to overhaul at  
least the title.

When touching that document anyway, there is one thing I find very  
disturbing, the comparision between a Java Program and a "corresponding"  
C-Program is not fair, as the new/old version of the C-Program both loop  
over all possible addresses and the Java program does not. So in fact the  
Java program does not work very well in Dual-Stack environements whereas  
the C-program follow best practice.

Eighter add the Loop for the Java program as well (but then the notion  
"sample from tutorial works" is a bit wrong) or maybe add something like  
"however you might want to add additional support for trying different  
potential addresses like...")

For the ServerSocket C-Code I would add ioctl(V6ONLY,0) to make that  
obvious that Java is doing that always.

In the guide the is some mentioning of a V6ONLY option. Is there a plan to  
offer it, or should maybe the guide be reworked a bit?

I also think we could remove the part about compatibility addresses, they  
have been deprecated in RFC 5156 (unfortunately :)

What is the original source format of that document, is it somewhere  
available I would take a shot on changing it.

Greetings
Bernd

PS: the lack of Happy Eyeballs support in Java will be a seperate Thread.
-- 
http://bernd.eckenfels.net