MaraDNS snapshot update

I've updated the MaraDNS FAQ to add an entry describing my support policies regarding MaraDNS name resolution bugs. I have already posted a blog describing the extent of unpaid support I supply for these issues here:

http://maradns.blogspot.com/2009/05/alexa-top-500-list.html

RadioGatún C++ class

I have been working on an off with a C++ class for the 32-bit version of the RadioGatún hash for a couple of weeks so I can become more familiar with C++ classes. I finally, after some trouble, got this class to correctly implement RadioGatún-32. It can be downloaded here:

http://www.samiam.org/rg32

Deadwood 2.3.05 released

I have released Deadwood 2.3.05 today. This is a stable release of Deadwood; the only changes from Deadwood 2.3.04 are bug fixes. To wit:

• Spurious fflush() removed
• Deadwood now forwards on to the client (but doesn't cache) unusual DNS packets, as long as the ID and question match. This can be disabled by setting a new dwood2rc parameter, deliver_all, to 0
• Slight revision to code so it will compile in GCC 4.4 without any warnings
• Documentation and SQA tests updated. Most notably, the deliver_all parameter is documented and the BUGS section of the manpage points out having an incomplete last line is a syntax error.

It can be downloaded here:

http://www.maradns.org/deadwood

Deadwood short-term plans

Here are my short-term plans for Deadwood:

• There is a bug where dwood2rc files that don't end with a proper linefeed cause there to be a syntax error reported on the last line of the file. Fix.
• I have updated my development system from CentOS 5.3 to CentOS 5.4; the biggest change is that the current version of GCC is now GCC 4.4 instead of GCC 4.3; update the "no warnings when compiled with the latest GCC" test to use GCC 4.4, and fix any compile-time warnings found. This needs to be done for both the 2.3 and 2.4 branches of Deadwood.
• Do full SQA recressions, then release Deadwood 2.3.05 and Deadwood 2.4.08

People's Tactics: A fun and free wargame

Back in the 1970s and 1980s, the type of game to play if you were a hardcore player of board games were these hex-based games that simulated wars, usually historical battles. I remember having a lot of fun playing Avalon Hill games with my father, brother, and grandmother as well as a couple of those MetaGaming microgames.

Well, the hobby died out; the rules for these games were too complicated to have a broad market; the market for these wargames dried up in the early 1980s when home computer video games started entering the scene and Dungeons and Dragons took over. The only game with rules simple enough to have general appeal was Axis and Allies, which was published by the Milton Bradley Company in the mid-1980s; the game is still around and is published by what little is left of Avalon Hill.

However, the hobby still lives because there are a number of high-quality, free computer games out there that handle the complicated rules for you, have fog of war—the ability to not always see what your opponent is doing—something that never was possible with those old board games, the ability to easily save games to load at a later time, random map and scenario generators, and play-by-email support.

In my experience, the Windows-only games that aren’t open-source have the most compelling game play (but see postscript below). My favorite game right now is People’s Tactics which is the free “teaser” version of the far more advanced and flexible Advanced Tactics. The game comes with a number of scenarios, has a random map/scenario generator, and even allows you to make your own scenarios with control over pretty much any game parameter.

The rules are simple enough that I was able to go through the tutorial in about 30 minutes. Typical of turn-based strategy games, gameplay is slow: it takes three or four hours to finish a small randomly generated 23x23 map, and maps can be up to 100x100 in size. The game is nice because it runs in a window instead of using full-screen mode, allowing me to easily check my email and perform other tasks on my computer while playing the game.

The game offers a lot of play and, when I get a job with a living wage again, I will pay the $40-$50 to support the author and get the more extensive Advanced Tactics.

Another series of hex-based wargames that are free are games in the Steel Panthers line. “Steel Panthers: World at War” is a tactical-level hex-based turn based simulation of World War II, complete with random terrain and random scenario generators. There is a thriving community with hundreds of user-created mods at spwaw.com. Another related free game is “Steel Panthers: World War II”. I have not had a chance to learn these games yet to give a full review, but SP:WaW looks very popular based on the huge number of scenarios and mods available for it.

To download People’s tactics, go to this webpage and look for “Download People’s Tactics” on the left (scroll the page down) or the box with the label “People’s Tactics” on the right (again, you will need to scroll the page down):

http://www.advancedtactics.org

Edit: Thinking about it some more, I neglected to remember two compelling open-source turn-based strategy games: Battle for Wesnoth which I enjoy greatly, and, while I haven't played it, Eight Kingdoms looks compelling.

New Deadwood and MaraDNS snapshots

I’ve made updates to both the “head” branch of both Deadwood and MaraDNS. In the Deadwood update, I have updated the documentation for the 2.4 branch to describe a new dwood2rc parameter added because of yesterday’s bug fix. In the MaraDNS update, I have updated the text of the MaraDNS.org webpage to spell out how exactly to get MaraDNS support by subscribing then posting to the list.

http://maradns.org/deadwood/snap/

http://maradns.org/download/1.3/snap/200910

Deadwood snapshot update

Deadwood will now forward on DNS packets upstream which it thinks are invalid packets, since some DNS servers send "name error" packets without a SOA record in the NS section of the reply.

I have made sure that this fix does not allow DNS packets which would cause security problems to be forwarded to the DNS stub resolver (we still make sure the ID and question are the same).

I would like to thank Jakob Blomer at CERN laboratories for reporting this problem, and for supplying a patch.

It can be looked at here:

http://maradns.org/deadwood/snap/