00:14:45 Unstable branch on cbro.berotato.org updated to: 0.30-a0-358-g0d69a1922f (34) 01:31:27 Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-4850-gd9e8576752 03:05:40 -!- daethoven83 is now known as daethoven8 04:22:32 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-4850-gd9e8576752 05:32:34 ??wiki 05:32:34 wiki[1/3]: Wiki for players: http://crawl.chaosforge.org/. Ask dvwalter151@gmail.com or spudwalt on reddit for an account to make it less bad! 05:32:56 no idea if that's current though, sorry 05:33:05 (wiki is completely unofficial) 05:33:15 ^ sockthot 05:43:28 <06advil> hm does anyone know what 60.234.103.73 is? It seems to be pulling all cao logfiles/milestones about once a second each 05:43:59 <06advil> "Wget/1.20.3 (linux-gnu)" 05:48:55 <06advil> 45.76.231.194 (default apache page, user agent "python-requests/2.26.0") is also quite aggressive, once per 20s 05:49:01 <06advil> even sequell is only once per 2min 06:05:15 <06advil> I guess I will probably block one or both of those unless I can figure out who they are to ask nicely 06:05:27 <06advil> the only possibility I'm really aware of is the dcss vercel stats app 06:22:52 <06advil> and it sounds like it's not vercel, he's actually getting his data from shallott, and polling once every 5min 06:27:22 <06advil> hm, I wonder if it would be better if cao scoring also used centralized logfiles from shallott (https://loom.shalott.org/server-xlogs/) 06:35:57 <06advil> github keeps showing me this, which tbh always looks like some sort of spammy malwarebytes popup (I just cannot get that thing off of my laptop) or whatever, but maybe actually corresponds to some useful features. In particular, I think it could hard prevent force pushes to master 06:35:57 <06advil> https://cdn.discordapp.com/attachments/747522859361894521/1039533737450405978/Screen_Shot_2022-11-08_at_8.34.38_AM.png 06:36:39 Webtiles server stopped. 06:36:44 Webtiles server started. 06:38:49 <06advil> I think all this improved logging on cao is just going to essentially confirm that it is disk IO that has been the problem all along 06:39:26 <06advil> I guess I've gotten some specifics at least 06:40:04 <06advil> like, sometimes just the open call for the dgl status file blocks for 200ms up to like 1500ms, I have no idea why 06:40:21 <06advil> what is that file even used for 07:06:19 <06advil> ok, I've redirected both of those IPs to http://crawl.akrasiac.org/403bot.html for now 07:06:34 <06advil> I guess also @gammafunk in case someone contacts you 07:07:00 <06advil> I picked 5min there kind of arbitrarily but that's what the vercel stats dev is using 07:22:45 -!- Stella[OotC] is now known as aoei[TNNT] 07:26:25 -!- daethoven85 is now known as daethoven8 07:27:16 <10PleasingFungus> yeah, i think it’s fairly common practice to turn on ‘no force pushes to master’ 07:28:55 <06advil> also, I guess we never did rename master -> trunk 07:30:57 <10PleasingFungus> github uses ‘main’ for new projects these days. or maybe git does. something 07:31:12 <06advil> seems like a reasonable name 08:02:08 <06advil> hm, of course doing a 403 redirect doesn't really stop the flood of requests from these bots 08:17:33 <09gammafunk> Interesting that this app is just pulling from shallot 08:17:56 <06advil> that turned out to be wrong 08:18:21 <06advil> I think he may have pulled old data from shallott but for cao at least he's pulling directly from cao on a 5min cycle 10:26:26 <09gammafunk> Yeah that makes sense, I told the dev in discord about the shalott url when we were talking about how certain servers like cjr/cszo etc were missing. 10:28:45 <09gammafunk> No idea if it's a problem pulling from shalott for Cao scoring but greensnark could give the green light if contacted 10:38:22 <06advil> I actually think it would be a lot better if we had something centralized like that for game data and I've contemplated it before, but one wants to know what it actually costs (bandwidth/money/time) and it def seems worrying to put it by default on an inactive devteam member 16:01:52 -!- Discord|2 is now known as Discord| 16:24:01 Unstable branch on underhound.eu updated to: 0.30-a0-358-g0d69a1922f (34) 16:47:04 <06advil> it sort of looks like cutting off those bot accounts made some difference to cao (I also did a pass through bots we were seeing that claim to obey robots.txt too) 16:47:20 <06advil> of course, now on to the next mystery source of disk latency 16:47:34 <09gammafunk> Responding a bit late, but a couple years ago I posed the question about Sequell maintenance to greensnark when there was an outage and you had to run NotSequell for a while (apparently another in your dynasty of Not bots): > gammafunk: Incidentally, if you feel hosting the Sequell server is a thing you won't have time for long-term, we can always look into hosting ourselves. We would only do so with your blessing, of course, seeing 16:47:34 as you are the reason we have a Sequell at all. Thanks again for all the work you've done on DCSS stuff in general! > > greensnark: Long term I'm not sure I'll be able to maintain Sequell, but I'll > first look to dockerize it so it's easier for someone else to host. Not really saying much, especially since it's been two years since that convo and he's still maintaining sequell, but yeah maybe we shouldn't count on that happening forever and might have 16:47:35 to host it ourselves. 16:48:05 <06advil> heh 16:48:16 <06advil> sometimes it is impressive how set-and-forget (mostly) some of these things are 16:48:20 <09gammafunk> He hasn't dockerized it, so we still have time! 16:48:54 <06advil> like, cao scoring was(/is) a mess, but it basically ran autonomously with no one editing the code besides config for like 10 years or something 16:49:03 <06advil> and sequell is way better in that respect I think 16:49:25 <09gammafunk> yeah, an far more complicated a patchwork of technologies too 16:49:34 <06advil> that is definitely true 16:51:05 <09gammafunk> I think if it weren't for you and greensnark we wouldn't really have much infrastructure beyond the game servers themselves now that neil hasn't had much time to help; he was the one primarily keeping things together for a good while 16:51:37 <09gammafunk> Maybe someone else would have appeared to save us though (not me as I lack the expertise) 16:52:36 <06advil> one can gain the expertise! 16:53:13 <06advil> for cao scoring possibly one of the other efforts would just have become the default, sort of happening a bit with the vercel one tbh 16:53:57 <09gammafunk> right, that vercel app is pretty sophisticated 16:54:13 <06advil> I update cao scoring some every once in a while but it is .... still not modern 16:55:01 <06advil> I wonder if any players ever did notice http://crawl.akrasiac.org/scoring/fastest-wins-time.html#current and http://crawl.akrasiac.org/scoring/fastest-wins-turns.html#current 16:55:14 <09gammafunk> oh nice 16:58:31 <06advil> I guess I never got around to adding tabs to the top score page though 16:59:45 <09gammafunk> not terribly important, but not sure about this page: http://crawl.akrasiac.org/scoring/gkills.html 16:59:50 <09gammafunk> oh that's milestones, right 16:59:58 <09gammafunk> a casualty of the change to not have cao load those I'm sure 17:00:03 <06advil> lol 17:00:15 <06advil> there's a bug there that I thought I fixed, which is that the sort order is flipped 17:00:17 <09gammafunk> er no 17:00:23 <09gammafunk> that is in logfile 17:00:27 <09gammafunk> ah, ok 17:00:36 <06advil> maybe I never restarted scoring after fixing it? (though that was months ago) 17:01:00 <09gammafunk> !lg * cikiller=player_ghost ikiller~~congressman 17:01:06 <04Cerebot> 516. slowdac the Conjurer (L6 DECj of Ru), mangled by Congressman's ghost (kmap: gammafunk_ghost_split) on D:3 on 2022-10-08 07:24:13, with 286 points after 3089 turns and 0:08:16. 17:01:10 <09gammafunk> oh neat 17:01:12 <09gammafunk> they're still at it 17:01:20 <06advil> yeah logfiles have enough info to get that right I think 17:01:52 <06advil> it's just kills per ghost name, so you can pull the ghost names out of the cikiller and invert 17:02:39 <06advil> possibly I should cutoff that page at 10 kills or something 17:03:00 <09gammafunk> nicolae and I suggest without any bias that if you want to compete with vercel app, the next thing you should add is comprehensive vault death and vault author rankings 17:03:11 <06advil> haha 17:03:30 <09gammafunk> yeah that page takes a bit to load for me, like several of the cao summary pages do 17:03:30 <06advil> really what it needs is some form of bookmarking 17:03:36 <06advil> and search 17:04:07 <06advil> unfortunately the trick that tourney uses for search probably won't work 17:04:31 <06advil> because there are however many unique names 17:05:11 <09gammafunk> yeah I think I get what you mean, and even the tournament pages are pretty slow at filling in the pagination 17:05:30 <09gammafunk> it takes a number of seconds for those to properly render the pagination after loading 17:05:43 <06advil> right, even just the plain search page may not work though 17:05:51 <06advil> it has a static list of every username 17:06:13 <06advil> well, it'll work in the sense that right now you have to load the all player list and use ctrl-f, this would be similar 17:06:28 <06advil> hm, maybe given that I should just try implementing 23:34:56 Unstable branch on crawl.develz.org updated to: 0.30-a0-358-g0d69a1922f (34) 23:56:44 Windows builds of master branch on crawl.develz.org updated to: 0.30-a0-358-g0d69a1922f