On Mon, Feb 16, 2004 at 12:28:19PM -0500, Kevin L. Mitchell wrote: > On Mon, 2004-02-16 at 02:22, Diane Bruce wrote: > > 1) I wrote the original ircd-parser.y, I find it offensive that someone > > else is claiming ownership of this. The least they should have done > > is add a "derived from" notice to the comments. It never made > > sense to me to add my copyright name on a GPL piece of code. > > Oops...I wonder who's to blame for this blunder! Sorry :(
np. Even if it had been stated as "Original by Diane Bruce, changes copyright by..." or something that would have been fine. > > > 2) you need a #include <sys/types.h> for adns to compile on freebsd > > frankly, adns is a piece of shit and should be ripped out. > > Agreed, but thankfully we're only using it in .12... Please don't. Let me rip adns out. Please? We have far better resolver. I just need to find some docs on your event code. > > > 3) the anti join/part spambot code. I've tried to make it fit into > > ircu standards, it would need another pass by someone but it works > > as is. I tested it. It should patch against .11 as well easily. > > It has been extremely effective in stopping join/spam/part bots on > > EFnet for many years. > > Interesting. Do you know what its effectiveness is compared to our > target limit code? (Channel joins count as a target, and I often find > myself target limited when joining the huge number of testnet channels I > join...) Its very very very very effective. I see spambots on undernet that I never seen happen on efnet. The target limiting code attacks a different problem, and btw, sounds eerily similar to something I had tried a few years ago. I may steal it. ;-) (Its less of a problem having off channel spamming, since I can go umode +g, so haven't been to worried) They attack two different types of problem. Jeekay has been playing with a variant in services. I think the numbers I use for timers etc. works very well, but still he has found at least one spambot with it. I think his timers/counters will produce a lot more false positives than the ones I use. > -- > Kevin L. Mitchell <[EMAIL PROTECTED]> P.S. I'll have a go at the hybrid resolver if you want. Lemme rip out adns. ;-) and I'll wait until I know which tree I should cvs checkout... -db
