2009-05-15 20:01:56 +00:00
|
|
|
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
|
|
|
|
|
<HTML>
|
|
|
|
|
<HEAD>
|
|
|
|
|
<META http-equiv="Content-Type" content="text/html; charset=ISO-8859-1">
|
|
|
|
|
<META http-equiv="Content-Style-Type" content="text/css">
|
|
|
|
|
<META NAME="Language" content='en'>
|
|
|
|
|
<META name="author" lang="en" content="Michiel Broek">
|
|
|
|
|
<META name="copyright" lang="en" content="Copyright Michiel Broek">
|
|
|
|
|
<META name="description" lang="en" content="MBSE BBS Manual">
|
|
|
|
|
<META name="keywords" lang="en" content="MBSE BBS, MBSE, BBS, manual, fido, fidonet, gateway, tosser, mail, tic, mailer">
|
|
|
|
|
<TITLE>Nodelist and Nidediff processing.</TITLE>
|
|
|
|
|
<LINK rel=stylesheet HREF="manual.css">
|
|
|
|
|
</HEAD>
|
|
|
|
|
<BODY>
|
|
|
|
|
<BLOCKQUOTE>
|
2011-05-22 19:21:15 +00:00
|
|
|
<!-- MBSEADVERT -->
|
2009-05-15 20:01:56 +00:00
|
|
|
<div align='right'><h5>Last update 22-Jan-2002</h5></div>
|
|
|
|
|
<div align='center'><H1>Nodelist and Nodediff processing</H1></div>
|
|
|
|
|
|
|
|
|
|
<H3>Introduction</H3>
|
|
|
|
|
<P>
|
|
|
|
|
A received a lot of questions about nodelist and nodediff processing, so
|
|
|
|
|
I will describe here the setup of the development system for the Fidonet
|
|
|
|
|
nodelist. First of all, it is <strong>very important</strong> that you
|
|
|
|
|
use three separate directories to do the nodelist processing. This is to
|
|
|
|
|
make sure that all stages are independent of each other, and if something
|
|
|
|
|
goes wrong, you still have a working system. The three directories are:<br>
|
|
|
|
|
<ol>
|
|
|
|
|
<li><strong>/opt/mbse/ftp/pub/fido/nodelist</strong>, this is the public
|
|
|
|
|
download area, the received diff's are stored here as well as the final
|
|
|
|
|
compressed nodelists for download.
|
|
|
|
|
<li><strong>/opt/mbse/tmp/nlwork</strong>, this is the working directory
|
|
|
|
|
to apply diffs to the previous nodelist. This directory should allways
|
|
|
|
|
contain the latest <strong>uncompressed</strong> nodelist.
|
|
|
|
|
<li><strong>/opt/mbse/var/nodelist</strong>, this is the systems nodelist
|
|
|
|
|
directory defined in mbsetup, menu 1.4.4
|
|
|
|
|
</ol>
|
|
|
|
|
In short the steps to process the nodediff's is as follows:
|
|
|
|
|
<ol>
|
|
|
|
|
<li>Receive the nodediff and store it for download.
|
|
|
|
|
<li>Apply the diff to the latest nodelist.
|
|
|
|
|
<li>Hatch the new compressed nodelist.
|
|
|
|
|
<li>Store the new nodelist for download.
|
|
|
|
|
<li>Unpack the new nodelist in the nodelist compiler directory.
|
|
|
|
|
<li>Set the compile semafore.
|
|
|
|
|
<li>Compile the nodelists.
|
|
|
|
|
</ol>
|
|
|
|
|
Next I will describe these steps in detail.
|
|
|
|
|
<P> </P>
|
|
|
|
|
|
|
|
|
|
<H3>The download area</H3>
|
|
|
|
|
<P>
|
|
|
|
|
First define the download area for the bbs. In my case, this is area 20. From
|
|
|
|
|
here users can download the nodelists and nodediffs, files to the downlinks
|
|
|
|
|
are send from here. Below is the example of my system.
|
|
|
|
|
<P>
|
|
|
|
|
<IMG SRC="images/nodelist.png" Border="0" Alt="BBS download area 20">
|
|
|
|
|
<P> </P>
|
|
|
|
|
|
|
|
|
|
<H3>The NODEDIFF tic area</H3>
|
|
|
|
|
<P>
|
|
|
|
|
From your uplinks you usually receive NODEDIFF or NODEDIZZ files. Create a tic area for
|
|
|
|
|
that purpose. I have keep# set to 5, this means the last 5 diff's are stored
|
|
|
|
|
in the download directory, older ones are removed. Now you can receive
|
|
|
|
|
nodediff files, store them for download, and send them to other nodes.
|
|
|
|
|
<P>
|
|
|
|
|
<IMG SRC="images/nodelist1.png" Border="0" Alt="NODEDIZZ tic area">
|
|
|
|
|
<P> </P>
|
|
|
|
|
|
|
|
|
|
<H3>Apply the diff</H3>
|
|
|
|
|
<P>
|
|
|
|
|
We do this with the tic <strong>magic</strong> processor. In this example
|
|
|
|
|
I have NODELIST.007 in the <strong>/opt/mbse/tmp/nlwork</strong> directory.
|
|
|
|
|
Note that this filename is uppercase, they are usually stored and distributed
|
|
|
|
|
as uppercase names. As I receive the diff files as zip, the filemask on
|
|
|
|
|
my system is <strong>nodediff.z##</strong>.
|
|
|
|
|
This means that the file with the name nodediff.z14 in the area NODEDIZZ
|
|
|
|
|
is a match. The command that is executed expands to
|
|
|
|
|
<strong>mbdiff /opt/mbse/tmp/nlwork/NODELIST /var/spool/mbse/ftp/pub/fido/nodelist/nodediff.z14 -quiet</strong>
|
|
|
|
|
if the received nodediff is
|
|
|
|
|
<strong>nodediff.z14</strong>.<br>The mbdiff program applies
|
|
|
|
|
<strong>nodediff.z14</strong> against <strong>NODELIST.007</strong> in the
|
|
|
|
|
<strong>/opt/mbse/tmp/nlwork</strong> directory. If this is successfull, a
|
|
|
|
|
new <strong>NODELIST.014</strong> is created there, a compressed <strong>
|
|
|
|
|
nodelist.z14</strong> is created there and <strong>NODELIST.007</strong> is
|
|
|
|
|
removed.<br> If this operation fails, only <b>NODELIST.007</b> will stay
|
|
|
|
|
in that directory.
|
|
|
|
|
Because the ARC program for GNU/Linux isn't good for files, I
|
|
|
|
|
left the Arc files command empty in the archiver setup. As a fallback the
|
|
|
|
|
mbdiff program uses <strong>zip</strong> to create the compressed archive.<br>
|
|
|
|
|
If creating the new nodelist fails for some reason, a missed diff or so,
|
|
|
|
|
the whole processing stops here. The previous nodelist is still here and
|
|
|
|
|
you can manually correct the situation. So, if you missed a diff, see that
|
|
|
|
|
you get it and manually give the <strong>mbdiff</strong> commands as user
|
|
|
|
|
<strong>mbse</strong> until you are up to date. Or, place the latest
|
|
|
|
|
uncompressed nodelist in the directory <strong>/opt/mbse/tmp/nlwork</strong>.
|
|
|
|
|
<P>
|
|
|
|
|
<IMG SRC="images/nodelist2.png" Border="0" Alt="Magic to apply the nodelist diff file">
|
|
|
|
|
<P> </P>
|
|
|
|
|
|
|
|
|
|
<H3>Processing the new nodelist</H3>
|
|
|
|
|
<P>
|
|
|
|
|
Now that we have created the new compressed nodelist, it has to go somewhere.
|
|
|
|
|
The file <strong>nodelist.z14</strong> is in the directory
|
|
|
|
|
<strong>/opt/mbse/tmp/nlwork</strong>. The example for the hatch manager
|
|
|
|
|
is shown below. The hatch manager runs automatic with the comand
|
|
|
|
|
<strong>mbfido tic</strong>. This setup will hatch the new nodelist in the
|
|
|
|
|
tic area <strong>NODELIST</strong> The two screens below show the hatch and
|
|
|
|
|
tic setup for this area.
|
|
|
|
|
<P>
|
|
|
|
|
<IMG SRC="images/nodelist3.png" Border="0" Alt="Hatch new nodelist">
|
|
|
|
|
<IMG SRC="images/nodelist4.png" Border="0" Alt="TIC area for new nodelist">
|
|
|
|
|
<P>
|
|
|
|
|
Now that we have hatched the new nodelist and processed it the new nodelist
|
|
|
|
|
is stored in in the download area, and maybe send some copies to downlinks.
|
|
|
|
|
We now have to feed it to the nodelist compiler for our own system.
|
|
|
|
|
We use a tic <strong>magic</strong> command to do
|
|
|
|
|
that. In this case we unpack the nodelist in <strong>/opt/mbse/var/nodelist</strong>
|
|
|
|
|
and set the <strong>compile</strong> semafore so that the <strong>mbindex
|
|
|
|
|
</strong> will compile the new nodelist. Don't be afraid that the unpacked
|
|
|
|
|
nodelists will acumulate in the nodelist directory, <strong>mbindex</strong>
|
|
|
|
|
will handle that, only the latest two nodelists are kept there. The
|
|
|
|
|
<strong>mbindex</strong> program is started by the taskmanager <strong>
|
|
|
|
|
mbtask</strong>.
|
|
|
|
|
<P>
|
|
|
|
|
<IMG SRC="images/nodelist5.png" Border="0" Alt="Magic copy nodelist and trigger compiler">
|
|
|
|
|
<P> </P>
|
|
|
|
|
|
|
|
|
|
<A HREF="index.htm"><IMG SRC="images/b_arrow.png" ALT="Back" Border="0">Go Back</A>
|
|
|
|
|
</BLOCKQUOTE>
|
|
|
|
|
</BODY>
|
|
|
|
|
</HTML>
|
|
|
|
|
|
