<?xml version="1.0" encoding="windows-1252"?>
<node id="898031" title="Re^7: Socket hang. (Windows or Perl? Solutions?) (Updated)" created="2011-04-07 08:19:45" updated="2011-04-07 08:19:45">
<type id="11">
note</type>
<author id="640184">
Illuminatus</author>
<data>
<field name="doctext">
&lt;blockquote&gt;
  &lt;i&gt;Maybe IPv6 has corrected some of these archaic leftovers?&lt;/i&gt;
&lt;/blockquote&gt;
That would be a logical assumption, which would make it highly unlikely :).  IPv6 changes nothing about TCP.  Well OK, it changes how the pseudoheader checksum is calculated (part of TCP header).  From an OSI-model point of view, they really had little choice about changing TCP, as this is at a different layer, and each layer is supposed to be independent of the others.  I worked on our IPv6 implementation, and the pseudoheader was the only TCP code that we had to touch.  They even pretty much left the 1500-byte MTU alone, mostly because of legacy devices.&lt;p&gt;&lt;b&gt;fnord&lt;/b&gt;&lt;/p&gt;
</field>
<field name="root_node">
897591</field>
<field name="parent_node">
897912</field>
</data>
</node>
