03:34:01 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5261-gd9800d219b 09:09:46 <11O​dds> I was having a look at https://github.com/crawl/crawl/issues/5515, and to my surprise when I use that seed on that version I don't get that Vaults:3. I do always get the same one, but it's not that level. 09:30:35 my {inugami,dog,canine familiar} attacks my dormant clockwork bee. Is this intentional? 09:31:50 seems odd to me, but maybe I'm missing some of the story around each summon? 09:32:08 <11O​dds> Sounds very much like a bug to me 09:33:55 <11O​dds> I can't repro that though? 09:34:35 nope sorry nevermind, it attacked itself because it is confused. noob issue here nothing to worry about haha 09:34:49 * nerdcore hides 09:38:41 <11O​dds> Ah good 🙂 09:51:52 <11O​dds> This is a strange one - the issue has a save on it, which my repro attempt matches in dungeon generation until V:1, but V:2 goes wrong. So looks like there's some rather rare bug in seed reproducibility. 10:18:33 <08o​____0> my v:3 is a crystal wall level when I try it 10:21:17 <11O​dds> Huh.... mine is not but also doesn't match the issue 10:22:47 <08o​____0> My V:1 does have glaive of the guard so I assume it's the same as the save haha 10:24:52 <11O​dds> Could you send me a screenshot of your V:2 and V:3 minimaps? 10:33:27 <08o​____0> https://cdn.discordapp.com/attachments/747522859361894521/1561432032708657162/v2.png?ex=6ad65466&is=6ad502e6&hm=6632c6dc5ae2bf6ac511c5466cde429ec786010ef611675761da6d3962c44bbb& 10:33:28 <08o​____0> https://cdn.discordapp.com/attachments/747522859361894521/1561432033069105182/v3.png?ex=6ad65466&is=6ad502e6&hm=1979bf3e79bab5721582eb98091cc031a8ba00d5555b02ceb789743b85f1a8d9& 10:37:33 <08o​____0> my spider 4 is also different from thiers 10:39:05 <08o​____0> yeah and shoals:4 so it's something earlier than vaults 10:39:21 <08o​____0> (though vaults:1 is the same) 10:41:11 <08o​____0> Mine is also consistent so I guess I guess it's different on different machines? 🙃 10:55:35 <11O​dds> Hmmm so yours diverges before mine… just double checking you used the version in the bug? 10:57:58 <08o​____0> Seeded DCSS version 0.35-a0-1289-g0f3b0f9dcb (tiles) character file. Game seed: 12585028045717431110 (custom seed) 10:58:41 <11O​dds> Odd. I also tried downloading the trunk build, and got the same result as my local build 11:00:16 <08o​____0> ??trunk[2 11:00:17 <04C​erebot> trunk[2/7]: learndb jargon for "that version that's trunk when I wrote this two years ago, naturally you automatically know which one that is. Duh." 11:07:30 <12g​e0ff> I can't replicate this bug: the first tutorial finishes w/o error in Local Tiles/77a9b5a. Exiting with ctrl-q works too. (the coordinates from the lua error, (43, 22), belong to the exit at the very end of the first lesson) 11:13:09 <08o​____0> I tried running the windows console build in wine but I think it's just giving errors because it says all the vault walls are stone :P 11:14:03 <08o​____0> Oh they are stone. I thought vault walls were metal all this time 11:43:38 <11O​dds> All very strange. I can repro this as a difference between a trunk seeded game on CAO and one on my machine, with the same seed (though it produces different dungeons)... but a couple of other seeds match throughout 11:45:59 <11O​dds> @o____0 could you send me a save with that seed and version, that you ran in a debug build? 11:46:48 <08o​____0> You want me to start a new game in a debug build? 11:47:25 <11O​dds> Unless your last one was in one, yes please 11:47:31 <11O​dds> (The save will then have level builder logs) 11:49:26 <08o​____0> https://cdn.discordapp.com/attachments/747522859361894521/1561451155962466386/6z97fkh.cs?ex=6ad66635&is=6ad514b5&hm=2bc4dbdd733bfb7f3e18bf4b1b57c6dbf32e57d1c1a160eb22603c7f171d1dc8& 13:02:45 <11O​dds> OK, so I think what is happening with that seed mismatch is that nicolae_vaults_variations_on_themes rolls some tags when building the *.idx file. So games that generate this vault only match if they rolled the same theme at that point - hence stability on one build but not between them. I think this can be fixed by something like adding a constant vaults_room tag and not running this line if is_validating: 13:02:46 https://github.com/crawl/crawl/blob/77a9b5a25eb2a2dea80ed447a925931ba8783c73/crawl-ref/source/dat/des/branches/vaults_rooms_standard.des#L2949 13:04:02 <11O​dds> (Ah searching back for nicolae_vaults_variations_on_themes this may explain another seed mismatch event late last year which AFAIC was never resolved) 13:05:32 <11O​dds> Not fixing this one myself as I'm not sufficiently confident I understand the consequences of not adding these tags at build time 13:39:08 <08o​____0> amazing good catch! 13:40:08 <11O​dds> Kinda lucky, I cleared the map files because I was suspicious they were somehow stale and got yet another variant 🙂 15:32:44 03hellmonk02 07[conducttitles2] * 0.35-a0-1321-gf374c8ce07: Medusa form titles 10(69 minutes ago, 1 file, 8+ 0-) 13https://github.com/crawl/crawl/commit/f374c8ce0795 15:32:44 03hellmonk02 07[conducttitles2] * 0.35-a0-1322-gaf28474766: Werewolf form titles 10(62 minutes ago, 1 file, 9+ 0-) 13https://github.com/crawl/crawl/commit/af2847476632 15:32:44 03hellmonk02 07[conducttitles2] * 0.35-a0-1323-gf2a33d9c58: Dr exclusive dragon form title 10(55 minutes ago, 1 file, 7+ 0-) 13https://github.com/crawl/crawl/commit/f2a33d9c5894 15:32:44 03hellmonk02 07[conducttitles2] * 0.35-a0-1324-geb26e43f8c: Serpent form 15 rune title 10(42 minutes ago, 1 file, 8+ 0-) 13https://github.com/crawl/crawl/commit/eb26e43f8c96 15:32:44 03hellmonk02 07[conducttitles2] * 0.35-a0-1325-ge22810455d: Rime Yak title 10(23 minutes ago, 1 file, 7+ 0-) 13https://github.com/crawl/crawl/commit/e22810455dc6 15:32:44 03hellmonk02 07[conducttitles2] * 0.35-a0-1326-g615c11f6c4: Blade form title 10(19 minutes ago, 1 file, 7+ 0-) 13https://github.com/crawl/crawl/commit/615c11f6c4d6 15:32:44 03hellmonk02 07[conducttitles2] * 0.35-a0-1327-g6fe04dd699: Aqua form title 10(13 minutes ago, 1 file, 7+ 0-) 13https://github.com/crawl/crawl/commit/6fe04dd69998 15:32:44 03hellmonk02 07[conducttitles2] * 0.35-a0-1328-g045a5ed4e9: Adjust requirements and cleanup 10(7 minutes ago, 1 file, 13+ 10-) 13https://github.com/crawl/crawl/commit/045a5ed4e963 15:32:44 03hellmonk02 07[conducttitles2] * 0.35-a0-1329-g94a4412292: Two more titles 10(2 minutes ago, 1 file, 4+ 0-) 13https://github.com/crawl/crawl/commit/94a441229258 15:52:01 New branch created: pull/5520 (32 commits) 13https://github.com/crawl/crawl/pull/5520 15:52:01 Branch pull/5520 updated to be equal with conducttitles2: 13https://github.com/crawl/crawl/pull/5520 16:12:25 03hellmonk02 07[conducttitles2] * 0.35-a0-1330-gcc394a5e85: Change some title requirements 10(2 minutes ago, 1 file, 6+ 16-) 13https://github.com/crawl/crawl/commit/cc394a5e85a8 16:12:25 Branch pull/5520 updated to be equal with conducttitles2: 13https://github.com/crawl/crawl/pull/5520 16:36:02 03hellmonk02 07[conducttitles2] * 0.35-a0-1331-g3b35640abc: One last title 10(2 minutes ago, 1 file, 2+ 0-) 13https://github.com/crawl/crawl/commit/3b35640abc4e 16:36:03 Branch pull/5520 updated to be equal with conducttitles2: 13https://github.com/crawl/crawl/pull/5520 16:37:47 <09h​ellmonk> ok, I think this is finally ready for review 16:38:17 <09h​ellmonk> kind of got carried away, hopefully someone can cut this into something reasonable 17:52:07 <04d​racoomega> Wow, fascinating (and a little scary). I wouldn't be confident this is the only case where something like that could happen, either... (I am reminded of some depth-based logic not working for some vaults because it was run on validation, where the player was always at a particular default depth, rather than caring about the depth when the vault is placed) 22:55:45 <11O​dds> Ah interesting. I looked for more random instances (found one, in a bazaar so won’t affect seed stability at least) but yeah I should probably also check how player position affects what we build 22:56:39 <11O​dds> (And yeah definitely a little scary, I wonder if there’s any way to catch such things automatically by checking for consistency across a few builds) 23:45:15 Monster database of master branch on crawl.develz.org updated to: 0.35-a0-1308-g77a9b5a25e