00:30:22 <11Odds> Due to an old old bug, monsters don't currently follow the player down abyssal stairs (I stumbled across this because I noticed an ancestor wasn't following, but it's true generally). All monsters who can use stairs should follow you in the abyss too, right? 00:31:28 <11Odds> (We do place such monsters in transit to Abyss:N if that's where you are going, but the abyss only reads from transits to Abyss:1) 00:42:46 <04dracoomega> I'm honestly inclined to call that 'not a bug'. Or rather, monsters don't follow you into portals and this has kind of the same feel to me. (Deliberately diving in the Abyss is usually a bad call - especially these days - so it seems fine to me if you can get a momentary reprieve from things chasing you that way) 00:43:10 <04dracoomega> (I mean, it's possible that this behavior is because of a bug, but I think I prefer it this way?) 00:43:20 <04dracoomega> Though we could at least keep your allies with you 00:48:39 <11Odds> Yeah companions should certainly follow you. Hostile monsters... yeah I probably do like the gameplay of "dive out of this nasty situation into a long-term scarier place". In terms of consistency, I guess Pan is the closest analogue? And we don't do following in Pan. 00:48:53 <11Odds> So in other words I agree. I'll make the exception explicit rather than an unintended-seeming consequence of the code, and fix for companions. 01:35:56 03CrawlOdds02 07* 0.35-a0-752-g9a013e2f2f: Make companions follow through the Abyss 10(9 hours ago, 2 files, 13+ 6-) 13https://github.com/crawl/crawl/commit/9a013e2f2fc9 03:33:30 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5261-gd9800d219b 07:07:01 New branch created: pull/5346 (3 commits) 13https://github.com/crawl/crawl/pull/5346 07:07:03 03CrawlOdds02 07https://github.com/crawl/crawl/pull/5346 * 0.35-a0-753-g3cf34e93d8: Make mapstat reproducible in-game and log seeds 10(2 days ago, 3 files, 88+ 18-) 13https://github.com/crawl/crawl/commit/3cf34e93d887 07:07:03 03CrawlOdds02 07https://github.com/crawl/crawl/pull/5346 * 0.35-a0-754-g1768e9c8eb: Move stone stair fixups before connectivity checks 10(3 days ago, 1 file, 7+ 7-) 13https://github.com/crawl/crawl/commit/1768e9c8eb74 07:07:03 03CrawlOdds02 07https://github.com/crawl/crawl/pull/5346 * 0.35-a0-755-ga06fae9508: Add a connectivity check for totally isolated dungeon areas. 10(3 days ago, 2 files, 180+ 122-) 13https://github.com/crawl/crawl/commit/a06fae950846 07:08:16 <11Odds> (Planning to merge that tomorrow or so unless anyone has cunning ideas to improve it; mostly made the PR to have somewhere to put some images of what it does) 09:40:05 <09gammafunk> hrm, is that to be used in the specific case of running mapstat with a given seed? 09:44:10 03CrawlOdds02 07https://github.com/crawl/crawl/pull/5346 * 0.35-a0-756-gd7cd69e227: Remove an obsolete test 10(81 seconds ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/d7cd69e227e6 09:44:35 <11Odds> No, the mapstat bit is to be used when you want to find a seed that produces a certain veto 09:45:07 <11Odds> Although the reproducibility bit could let you mapstat a seed you'd found from a game in a matching way, if that's useful 09:46:09 <11Odds> (My workflow for this new veto was mapstat to find vetoes -> look at them by booting up a game) 09:47:19 <09gammafunk> I guess I'm wondering how this works as a mode because normally when running mapstat you dictate a specific set of levels 09:48:34 <11Odds> Ah.... I should go make sure it plays well with that. It should be fine, we just need to skip the levels we haven't asked for 09:49:01 <11Odds> (And in that case it won't match a seeded game if you don't ask for a prefix of generation order) 09:49:28 <09gammafunk> yeah 09:49:42 <09gammafunk> are you running mapstat with just a default level set then? 09:50:03 <11Odds> Yeah, for the use case of "change dungeon generation" that's the way that's generally useful I think 09:50:16 <11Odds> For the use case "build vaults" I guess you build the levels they appear on 09:50:33 <09gammafunk> hrm, and I guess this default set does match the default generation order of a normal game? 09:51:09 <09gammafunk> I'm not actually sure what the default generation set of mapstat/objstat is 09:51:30 <11Odds> It's "everything", previously in a random order but now in a game-matching order 09:52:21 <09gammafunk> ok, that does seem helpful then. I'm not sure if there's anything we can do for the common case of running a specific subset set of levels in terms of seed 09:52:46 <09gammafunk> but maybe that's fine 09:53:08 <11Odds> Yeah if we wanted seeded there, we'd just have to generate everything and drop the levels we don't care about 09:53:26 <11Odds> Guessing people would not to do that so as to make things faster 09:53:39 <09gammafunk> right, that's basically the whole point of specifying a specific set 09:55:11 <11Odds> (Ah while what I did mostly played nicely with a specific set I had forgotten not to generate portals off the side of it) 09:56:03 <09gammafunk> and looking at your diff, it seems you did implement this specifically/only for mapstat, and not objstat 09:56:34 <09gammafunk> or maybe not 09:58:52 <11Odds> No I believe they share this machinery 10:18:30 <09gammafunk> @Odds one thing to take a look at is objstat_generate_stats() and specifically the code block that populates the static stat_levels; this set of level_ids is important for objstat to initialize and populate its object data and must be consistent with the static generated_levels in that it must ultimately be the exact same set of levels. I think what you're doing is fine in that regard, since your PR is about the order of that vector. 10:18:31 We do use iterations of stat_levels to write out object stat files but again I don't think this will be relevant to any ordering of generated_levels 10:19:05 <11Odds> Cheers, I'll take a much closer look at objstat (I've checked it runs but that's about it) 10:20:35 <09gammafunk> also I apparently didn't read my own code: cpp // Populate a vector of the levels ids for levels we're tabulating. ...actually I see a vector of levels under that comment that appears to be unused.... 10:20:37 <11Odds> Actually I'm fairly likely to have messed that up due to portals generation alongside 10:21:29 <11Odds> But yeah I was definitely thinking about this through a mapstat lens and will get to grips with objstat too 10:21:44 <09gammafunk> ok, thanks 10:22:31 <09gammafunk> I tried to improve the seed behavior of mapstat/objstat with 10:22:34 <09gammafunk> %git 14726095691d0599cc0f1f99f3733f71cec6a3b5 10:22:34 <04Cerebot> gammafunk * 0.34-a0-1913-g1472609569: Properly use seeds for objstat and mapstat (7 months ago, 3 files, 27+ 4-) https://github.com/crawl/crawl/commit/14726095691d 10:23:25 <11Odds> Ah nice, that was definitely super needed for any of what I've done to work 10:23:30 <09gammafunk> I guess this is actually related to your work in that the seed...yeah 10:23:33 spl-cast.cc:2341:13: warning: enumeration value 'seen_hups' not handled in switch [-Wswitch] 10:24:06 <11Odds> For the dungeon generation stuff it's really nice to be able to detect and investigate specific seeds in-game 10:24:34 <09gammafunk> yeah, agree that the workflow you outlined above sounds good for tracking down layout problems 12:41:35 03CrawlOdds02 07https://github.com/crawl/crawl/pull/5346 * 0.35-a0-757-gb65ab701fb: Fixup: fix mapstat/objstat portal generation 10(2 hours ago, 1 file, 23+ 9-) 13https://github.com/crawl/crawl/commit/b65ab701fb65 13:21:23 <05icoson> the seed explorer might also do some of what you want: https://github.com/crawl/crawl/blob/master/crawl-ref/source/scripts/seed_explorer.lua 13:22:01 <05icoson> I always thought of it as, mapstat/objstat are bulk testing that's supposed to run fast at scale, seed explorer is for fine-grained testing that matches levelgen behavior as close as possible 13:22:29 <05icoson> seed explorer has only ever been maintained by me afaik so it's probably out of date a bit by now 13:23:05 <05icoson> fast vs levelgen-accurate has always been a bit of a tradeoff 13:23:09 <11Odds> Oooh, I wasn't aware of that, thanks! 13:23:42 <05icoson> https://github.com/crawl/crawl/blob/master/crawl-ref/source/scripts/placement.lua could also be interesting for you 13:23:58 <11Odds> I'm hoping that mapstat is now both! 15:43:17 Unstable branch on underhound.eu updated to: 0.35-a0-752-g9a013e2f2f (34) 16:19:13 Odds (L27 NaHu) ASSERT(!env.markers.need_activate()) in 'main.cc' at line 2576 failed. (Spider:3) 16:19:38 <11Odds> ^ Wizmode testing, please ignore 16:21:53 New branch created: pull/5347 (1 commit) 13https://github.com/crawl/crawl/pull/5347 16:21:53 03Aliscans02 07https://github.com/crawl/crawl/pull/5347 * 0.35-a0-753-g9bd5c652e1: Only say the ground has solidified when it has. 10(80 minutes ago, 1 file, 8+ 3-) 13https://github.com/crawl/crawl/commit/9bd5c652e104 22:35:43 Unstable branch on crawl.develz.org updated to: 0.35-a0-752-g9a013e2f2f (34) 23:00:28 Windows builds of master branch on crawl.develz.org updated to: 0.35-a0-752-g9a013e2f2f 23:12:50 Unstable branch on cbro.berotato.org updated to: 0.35-a0-752-g9a013e2f2f (34) 23:56:00 Monster database of master branch on crawl.develz.org updated to: 0.35-a0-752-g9a013e2f2f