I sent this to the list a yesterday, but it got bounced because I attached a jpeg. The jpeg and a couple other things are here: http://futeki.homelinux.com/fireconf/ Take a look at the code plan, which I wrote last night. I think that will help us get organized and not clobber each others work. I should have the project up on TriForge (see below) over the weekend. Until then, if you're doing some coding, make sure it hasn't been done before. Check Tyler Larson's work here: http://www.tlarson.com/fireconf.tar.gz and Jeffrey Ollie's work here: http://www.ollie.clive.ia.us/iptables/ For those of you who've already started coding, start moving your code into alignment with the code plan. Make all classes start with fc_ and put docstrings on all your methods. See here about docstrings: http://www.python.org/doc/current/tut/node6.html#SECTION006600000000000000000 At the bottom of the code plan is a section on recursive parsing. It looks like Jefferey has started to do this with his code. I still would want the fc_iptables.py file to interact with file and pipe I/O. But if we start putting these parse methods into each of the core classes, it will make it tremendously easier to add all the proto/target/match extensions later. On Wed, 2004-02-18 at 12:03, Tyler Larson wrote: > Truth is, the only way this is going to happen > as fast as I would like to have it is by getting a lot of help from the rest of > you. I agree and would like to get a project site going with a code oriented design spec so we can easily establish what needs to be done and delegate those tasks. I put together a simplistic diagram of how this might be done. I've attached a jpeg. Each of the main sections should be a independent set of classes to make delegation easier. I think what you've done in iptables.py would end up mostly in the parser area. And my pseudo code will become the main core (with added methods specific to our program). Does this make sense? > I like what you've described about how you'd like to do the GUI, and I'd like > to see the Glade project you've put together. I'll email you off list with my modified fireconf tarball later today. > > In the mean time, I think my best bet is to start scheduling my time better-- > I'm going to try to put in at least a few hours a week working on this project. > First thing I'll do is convert the pseudo-code datastructures to Python, then > I'll work on some of the GUI stuff. Expect some visible results within a week. Sounds good. I'll try to work on just getting the project planned and tasks set up. > > Next thing: we need a repository. I've heard bad stuff about sourceforge's > servers, and I don't like setting up a public CVS repo on my own servers > because of the security implications. I could set up a different type of repo > (the only other one I've played with is subversion, but I'm open to options). > If RedHat's going to have their public CVS server up any time soon, we could > use that too. Suggestions are, of course, welcome. My LUG (TriLUG) has a forge that we could use. The guys said it's still a bit beta, but we can try it out. http://forge.trilug.org It would allow us to have our own mailing list, CVS, etc. The trilug admins are _very_ knowledgeable and anything we had trouble with, we could probably get help pretty quickly. Also, I can post a message to the TriLUG list to recruit some people if necessary. I registered a project for us, so we can start playing with it. I'll post here again when the project is accepted. -Tim -- Timothy A. Chagnon <tchagnon@xxxxxxxxx>