00:00:42 -!- fooobarrr has quit [Read error: Operation timed out] 00:03:15 Unstable branch on crawl.develz.org updated to: 0.12-a0-1842-gb9bc147 (34) 00:04:16 -!- st_ has quit [Ping timeout: 256 seconds] 00:05:11 Unstable branch on crawl.s-z.org updated to: 0.12-a0-1842-gb9bc147 (34) 00:07:24 -!- shirish has quit [Ping timeout: 276 seconds] 00:12:23 -!- GuraKKa has quit [Read error: Connection reset by peer] 00:18:20 Windows builds of master branch on crawl.develz.org updated to: 0.12-a0-1842-gb9bc147 00:18:30 -!- remyroy1 has quit [Read error: Connection reset by peer] 00:20:20 -!- shirish has joined ##crawl-dev 00:20:28 -!- kekekela has quit [Ping timeout: 245 seconds] 00:39:07 -!- valrus has quit [Quit: Computer has gone to sleep.] 00:39:47 -!- cocofalco has quit [Quit: Page closed] 00:43:04 -!- caleba has quit [Quit: caleba] 01:00:03 -!- Amilir has quit [Ping timeout: 245 seconds] 01:02:50 -!- ophanim has quit [Read error: Connection reset by peer] 01:05:02 -!- pelotr0n has quit [Ping timeout: 256 seconds] 01:13:09 -!- andrewhl has quit [Remote host closed the connection] 01:37:15 -!- _dd has joined ##crawl-dev 01:44:28 -!- motorbit has quit [Quit: Page closed] 01:45:45 -!- caleba has quit [Quit: caleba] 01:53:00 -!- minqmay has quit [Quit: Leaving] 01:54:01 -!- crate has quit [Read error: Connection reset by peer] 01:54:19 -!- Heteroy has quit [Ping timeout: 246 seconds] 01:57:23 -!- crate_ is now known as crate 01:59:45 -!- ghallber1 has joined ##crawl-dev 02:00:35 -!- mikee_ has quit [Quit: HydraIRC -> http://www.hydrairc.com <- Organize your IRC] 02:08:06 -!- ghallber1 has left ##crawl-dev 02:17:06 -!- shirish has quit [Ping timeout: 252 seconds] 02:20:57 -!- ussdefiant_ has joined ##crawl-dev 02:21:51 -!- ussdefiant has quit [Ping timeout: 252 seconds] 02:21:55 -!- ussdefiant_ is now known as ussdefiant 02:22:21 -!- Zifmia has quit [Read error: Connection reset by peer] 02:25:30 -!- Morphy has quit [Client Quit] 02:28:33 -!- bza has quit [Quit: Leaving] 02:30:20 -!- shirish has joined ##crawl-dev 02:30:41 -!- Zifmia has quit [Read error: Connection reset by peer] 02:32:22 -!- ontoclasm has quit [Quit: Leaving.] 02:33:16 -!- rwbarton has quit [Ping timeout: 240 seconds] 02:33:27 -!- Zifmia_ has quit [Read error: Connection reset by peer] 02:33:36 -!- crate has quit [Ping timeout: 252 seconds] 02:36:11 -!- frontfelloff has quit [Quit: Page closed] 02:38:07 -!- crate_ is now known as crate 02:43:58 -!- WalkerBoh has quit [Remote host closed the connection] 03:25:46 -!- Sky__ has quit [Quit: Page closed] 03:27:03 -!- inspector071 has quit [Read error: Connection reset by peer] 03:28:43 -!- Zifmia has quit [Quit: AndroidIrc Disconnecting] 03:31:48 -!- dtsund has quit [Quit: dtsund] 03:33:06 -!- bhaak has quit [Ping timeout: 276 seconds] 03:34:32 -!- bhaak has joined ##crawl-dev 03:36:23 -!- descol has quit [Quit: Page closed] 03:38:16 -!- shirish has quit [Ping timeout: 246 seconds] 03:51:21 -!- shirish has joined ##crawl-dev 03:54:24 -!- bonghitz has quit [Quit: bonghitz] 03:55:24 -!- rast has quit [Read error: Connection reset by peer] 04:04:27 -!- syllogism has joined ##crawl-dev 04:06:05 -!- edlothiol has joined ##crawl-dev 04:10:18 -!- crate has quit [Quit: Leaving] 04:11:52 -!- ussdefiant has quit [Ping timeout: 252 seconds] 04:23:14 -!- Sorbius has quit [Ping timeout: 252 seconds] 04:39:13 -!- syraine has quit [Ping timeout: 245 seconds] 04:42:59 -!- alefury has joined ##crawl-dev 04:43:16 -!- zero_one has quit [Disconnected by services] 04:43:26 -!- derp has quit [Client Quit] 04:44:12 -!- Sprort has quit [Disconnected by services] 04:44:16 -!- zero_one is now known as Sprort 04:44:39 -!- zero_one has quit [Client Quit] 04:54:23 Liquorice (L13 MiFi) ASSERT(!at_branch_bottom()) in 'stairs.cc' at line 533 failed. (Abyss (Sprint)) 04:54:49 -!- nooodl has joined ##crawl-dev 05:14:51 -!- scummos has quit [Ping timeout: 276 seconds] 05:19:00 -!- Ragdoll has joined ##crawl-dev 05:35:53 -!- Blade- has quit [Ping timeout: 245 seconds] 05:50:11 -!- Souljazz\unfoog has quit [Quit: ChatZilla 0.9.89 [Firefox 18.0.1/20130116073211]] 06:12:44 -!- hurdos has quit [Ping timeout: 256 seconds] 06:20:38 -!- st_ has joined ##crawl-dev 06:34:09 Orb of destruction does not hit orb of fire and passes through it (https://crawl.develz.org/mantis/view.php?id=6574) by inspector071 06:42:02 -!- monqy has quit [Quit: hello] 06:58:01 -!- DracoOmega has quit [Quit: Leaving] 07:00:45 -!- Psyknux has quit [Read error: Connection reset by peer] 07:31:22 -!- Heteroy has quit [Ping timeout: 246 seconds] 07:37:26 -!- dagonfive has quit [Read error: Connection reset by peer] 07:37:51 -!- Blazinghand has quit [Ping timeout: 276 seconds] 07:53:59 -!- wheals has joined ##crawl-dev 07:56:07 ??sure blade 07:56:07 sure blade[1/3]: Level 2 Hexes/Charms spell that gives much higher to-hit but only works with short blades. Very useful, provided you're using short blades. 07:56:14 -!- kek has joined ##crawl-dev 07:56:33 whoops, thought this was my Henzell tab 07:57:56 -!- Fhqwhgads_ has quit [Remote host closed the connection] 08:00:58 -!- kek has quit [Ping timeout: 256 seconds] 08:01:59 -!- kek has joined ##crawl-dev 08:07:46 -!- kek has quit [Ping timeout: 256 seconds] 08:08:25 -!- kek has joined ##crawl-dev 08:16:16 -!- kek has quit [Ping timeout: 256 seconds] 08:19:06 -!- Zaba has quit [Ping timeout: 256 seconds] 08:22:33 -!- Engwar has quit [Ping timeout: 245 seconds] 08:28:23 is it really useful to have lots of different portals to hell in D 08:28:35 -!- kek has joined ##crawl-dev 08:31:58 -!- kek has quit [Client Quit] 08:32:08 marginally if using vestibule as a stash (in other words, no) 08:32:20 -!- kek has joined ##crawl-dev 08:33:53 i think its cool 08:38:01 -!- Zaba has joined ##crawl-dev 08:44:56 -!- Yermak has quit [Quit: Page closed] 08:52:22 -!- Elkan has quit [Ping timeout: 252 seconds] 08:59:24 -!- shirish has quit [Ping timeout: 276 seconds] 09:11:16 -!- rwbarton has quit [Ping timeout: 245 seconds] 09:15:24 -!- magistern has quit [Quit: This computer has gone to sleep] 09:31:25 -!- wheals has quit [Read error: Connection reset by peer] 09:32:53 -!- Engwar has quit [Quit: Page closed] 09:39:23 Where is Wensley these days? 09:40:10 !seen wensley 09:40:11 I last saw Wensley at Thu Jan 17 14:40:14 2013 UTC (1w 3d 59m 57s ago) quitting with message Quit: leaving. 09:40:52 ah 09:40:53 -!- shirish has joined ##crawl-dev 09:41:43 i remember when i originally wrote a test version of the newvaults layout (ages ago) he hosted a version so people could test it easily 09:46:36 how about testing it by pushing it to trunk :P 09:47:04 ^ 09:48:10 i'm sure elliptic would be happy to assist 09:48:57 I try not to push code I don't understand :( 09:49:14 who pushed newvaults? 09:49:19 elliptic: you should "understand" it by pressing &^R a few times and noting how much better than newvaults it looks 09:49:22 but it does seem like newnewvaults is pretty close to ready to trunkable? 09:49:30 I think it's been trunkable for a day or two? 09:49:46 AFAIK the only issue is that it's sometimes a bit slow to generate levels, but it doesn't seem to use lots of CPU or anything, and my computer is really slow. 09:49:59 mumra: evilmike I think 09:50:28 ok we need evilmike ;) 09:50:47 but seriously, the only thing i'm worried about with merging is the occasionally excessive level generation times 09:50:52 yeah, evilmike merged the branch 09:51:07 i might have solved that by fixing some vaults that potentially caused vetos 09:51:20 I think the level generation times will be better on a server than a 1.33 GHz ULV Core 2 Duo from a few years ago. 09:51:38 yeah but sometimes it's been unacceptable locally as well 09:51:45 on a brand new quad-core gaming laptop 09:52:47 have you profiled it 09:52:54 right now it seems good, but still a bit slower than newvaults 09:53:02 freefall: no, can Lua be profiled? 09:53:33 how large a level generation time are we talking about? on the servers D:27 already often takes a second or two 09:53:40 yeah, there are a number of tools 09:53:56 the worst for me was like 30 seconds, but that's on a slow machine and an outlier 09:54:05 when it's not nearly instant it's most often about 5 seconds or so -- again on my slow machine 09:54:18 okay that is sort of bad 09:54:29 need to figure out what's causing that and fix it before merging, certainly 09:54:30 not the kind of thing that should happen in a stable release, but 5 times per game doesn't seem so bad for testing in trunk 09:54:50 elliott: i'm not seeing any outliers like that at the moment. usually it's max 5 seconds, average time 1 or 2 secs 09:55:47 mumra: https://code.google.com/p/shinyprofiler/ may be useful 09:55:51 Zaba: I think it's just repeatedly vetoing the level? 09:55:55 and it'd be very easy to improve by reducing the number of tries at placing rooms. could result in less interesting layouts sometimes but it'd outright prevent silly times 09:55:59 elliott, yes, and that's not normal 09:56:27 elliott: at the moment we don't know if the problem is still happening, i haven't reproduced it for a while 09:56:54 honestly the D:27 thing already bothers me a fair amount and I am used to it 09:57:25 and even then, i don't know for certain if it's repeated veto, or if just specific layouts (probably the two last maze variants) are more problematic 09:57:53 clearly crawl should generate the next few levels you could reach in parallel while you play 10:00:16 I've been meaning to come through newnewvaults and see what/if needs tweaking and/or what's causing the reported long level generation times. 10:00:21 s/come/somb/ 10:00:22 ... 10:00:24 s/somb/comb/ 10:00:30 I apparently am not awake yet. 10:01:22 freefall: tbh i don't think profiling would help here, it's very easy to understand where chokepoints might exist in the code. there are up to 2500 attempts to place a room, if that number is getting reached then it will be slow. this could depend on how the level fills out and what rooms are still available (i.e. depending on allow_dup) 10:02:05 i think having more small rooms will help by allowing areas to connect up where large rooms can't fit. 10:02:12 mumra: yeah, in that case it's probably more than enough to log some simple stats by hand and just do lots of test runs 10:02:25 but a short-term solution would be just reducing the total number of tries 10:02:44 freefall: yeah that's what i was planning to do today and get a better picture what's going on inside it 10:03:03 I do think it'd be a big shame to miss out on some of the more intricate level designs for that 10:03:24 Grunt: any time you spend on it would be great, the code is much MUCH simpler to understand now, and it would be good if someone other than me had looked at it as well ! 10:03:26 is there a test harness that lets you run certain parts of the code easily and repeatedly without having to actually fire up the whole game 10:04:10 freefall: i think the stress test does that already 10:07:08 -!- Sobieck has quit [Ping timeout: 245 seconds] 10:07:48 i agree that it's a bad idea to merge code that nobody on the dev team understands, and that if newvaults were subjected to this same stringency there might not have been a mess to fix, although possibly then we would never have had newnewvaults ;) 10:10:20 Well, at the time that newvaults was merged, we had two or three devs on hand who each thought they half-understood it. <_< 10:13:59 the good thing about newvaults is that it's entirely self-contained 10:14:37 heh :) they maybe didn't look at it long enough ... i spent a while trying to understand it and it's one of those "the more you know, the less you understand" things 10:15:10 Zaba: why is that a good thing? you were recommending moving parts of newnewvaults into c++ ... 10:15:39 Zaba: (currently, newnewvaults is also completely self-contained) 10:15:48 mumra, I was recommending improving the lua API, because everyone would benefit from that 10:16:21 -!- _dd has quit [Remote host closed the connection] 10:16:34 Zaba: I'm not exactly sure what you mean by "self-contained" ... surely all Lua layouts are by definition self-contained? 10:16:52 (except for the ones that e.g. call old C++ layout functions) 10:16:52 (except for the ones that e.g. call old C++ layout functions) 10:17:05 the C++ layouts are pretty self-contained in dgn-layouts.cc :P 10:17:15 -!- Heteroy has quit [Ping timeout: 240 seconds] 10:17:22 can you give an example of a layout that's *not* self-contained? 10:17:28 but I meant that it didn't involve adding 2000 lines to dungeon.cc/maps.cc/mapdef.cc :P 10:18:04 Zaba: neither does newnewvaults. it does however make some useful extensions to the Lua api, and this involved adding a couple of dozen lines to dungeon.cc. 10:20:10 the main one is an inspect_map function which allows you to check the features that a vault will produce after it has been resolved, without actually modifying the dungeon grid in any way 10:21:23 newnewvaults is otherwise around 1000 lines of Lua, which means a total saving of about 7000 lines of Lua ... 10:22:23 if someone does want to take a look at the code i can help explain it, it's mostly extremely straightforward 10:23:46 -!- ark__ has joined ##crawl-dev 10:24:47 did you think newnewvaults was all in dungeon.cc or something? 10:25:24 actually, I never implied that newnewvaults wasn't self-contained 10:25:51 I only subtly pointed out that that was probably the only good thing ever about newvaults :P 10:26:26 -!- Zifmia has quit [] 10:26:30 ah heh, sorry completely misunderstood what you were getting at :P 10:26:33 -!- Vizer_ has quit [Quit: bye] 10:29:30 -!- _dd has joined ##crawl-dev 10:33:48 -!- tophat has quit [Ping timeout: 245 seconds] 10:35:58 elliott: i just tried significantly reducing the number of tries at placing rooms. layouts are consistently generating in under a second and still look great. 10:36:22 i see nothing unmergable about this ;) 10:39:42 I'll take your word for it 10:39:53 hopefully somebody figures out how to megaoptimise it at some point and it can try placing them a billion times just in case :P 10:40:22 elliptic: can I bribe you to merge newnewvaults with vehpatch/trappatch 10:40:30 -!- shirish has quit [Ping timeout: 276 seconds] 10:43:15 -!- codile has quit [Ping timeout: 240 seconds] 10:44:19 -!- ussdefiant has joined ##crawl-dev 10:44:42 The exploding dart explodes! 10:44:43 The explosion of exploding dart fragments engulfs you! 10:44:50 st_: i just found out why you sometimes saw too many hard rooms etc. - should be working with next push 10:45:46 todo: have the average V level generate with V:$ level of difficulty <_< 10:46:06 elliott: there are a couple more ways i can optimise things but they're highly non-trivial, i want to find out whether it needs it first 10:46:29 ...todo: make up a vaults stair ambush vault with number of vault guards vs., say, ordinary humans varying with depth >_> 10:47:16 although, can anyone comment on how slow static_cast<..> is? if it's really bad then there is a small optimisation that will help a bit 10:47:54 static_cast should not really incur a performance hit, to the best of my knowledge. 10:48:04 Grunt: there is a already a vault guard stair vault, didn't you make that one? ;) 10:48:19 mumra: _checkpoint? :b 10:48:37 I mean something in the style of the V:$ welcoming party, sort of like that one entry vault. 10:48:40 s/entry/Vaults entry/ 10:48:48 the only c++ cast with a real runtime hit is dynamic_cast 10:49:53 static_cast is... static :) 10:50:12 elliott, killed by a static_cast discharge. 10:50:39 -!- frogbotherer has joined ##crawl-dev 10:51:03 mumra, I just got a freeze on level generation. Sadly this isn't a debug build I'm trying it on, so I don't know what froze... 10:51:13 cool. i thought i remembered someone commenting on performance with it. but clearly they were talking about dynamic 10:52:57 -!- dtsund has joined ##crawl-dev 10:53:21 * Grunt makes a debug build to attempt to regenerate the crash and trace it... 10:53:30 s/crash/freeze/ 10:57:18 -!- frogbotherer has quit [Remote host closed the connection] 10:59:01 Grunt: I fixed a specific occasional Lua error since that last commit so hopefully that's the problem. just tweaking some more performance related stuff then will push and let you know 11:00:00 -!- anidude has quit [Remote host closed the connection] 11:03:38 -!- ussdefiant has quit [Ping timeout: 255 seconds] 11:06:39 -!- ontoclasm has joined ##crawl-dev 11:06:51 -!- Elkan has quit [Ping timeout: 276 seconds] 11:07:25 -!- HangedMan has joined ##crawl-dev 11:09:21 Grunt: pushed some changes, see if you can make it happen now 11:10:06 i modified the placement loop so it will exit much sooner if it's not placing anything. level generation times seem *highly* conistent now. 11:13:21 -!- Stelpa has quit [Ping timeout: 245 seconds] 11:15:50 10:02:05 i think having more small rooms will help by allowing areas to connect up where large rooms can't fit. 11:16:05 how small would they need to be? 11:18:20 -!- casmith_789 has quit [Read error: Connection reset by peer] 11:19:06 between 6x6 and 10x10. a lot of st_'s new ones are 11x11 but frequently (particularly towards the end of layout generation) there won't be room for stuff that big 11:19:57 also narrowish rooms could work e.g. 4x8 - 8x16 11:20:41 -!- Sorbius has quit [Client Quit] 11:20:52 -!- Gungr has quit [Quit: ChatZilla 0.9.89 [Firefox 18.0.1/20130116073211]] 11:20:57 some small rooms with allow_dup are showing up multiple times per level which is a strong indicator of this 11:21:27 * elliott remembers when mumra threw away all the small rooms :p 11:21:36 hehe 11:22:00 they were 4x4 rooms, and yes there could be an argument for bringing them back. but i don't miss 'em ;) 11:22:24 hmm 11:22:38 I'll see if I can make some smaller ones 11:23:08 today's project was just screwing around with a silly encompass vault 11:23:55 cool 11:25:55 i think 6x6 is the absolute minimum size required to actually make a room interesting, and there's still very little that can be done at that size. if i do want smaller rooms i'll just allow floor_vault to make small empty rooms, but i think gaps that tiny can be plugged in a more directed way (i am thinking of carving short corridors between some rooms when they are that close) 11:26:34 well if they don't have walls you could at least make a geometric thing in 5x5 11:26:52 I was struggling after a certain point to make 7x7 subvaults for my portal vault 11:26:53 walls are excluded from the sizes i am talking about, yeah 11:27:17 thin long vaults sounds good to play around with though 11:27:49 5x5 maybe but there aren't a lot of geometries you could fit in that size 11:29:04 corridor vaults could be a good way to break into unused areas of the map that are otherwise blocked off by room positioning (e.g. by rooms with only one entrance wall) ... it's a bit like scrabble 11:29:33 obviously will just make 2x68 vaults 11:30:44 i will also be making floor vault generate some empty corridors like corners and t-junctions, this should help as well but also make some generally interesting layouts 11:31:02 mmm 11:31:14 there's pretty much no end to the things that floor vault can do ;) 11:37:36 -!- pantaril has quit [Quit: leaving] 11:44:44 -!- _D_ has quit [Quit: leaving] 11:44:47 loads of example map output from newnewvaults: http://pbrd.co/UuXrEf 11:45:06 for anyone that hasn't seen it first-hand as yet 11:47:45 mumra: I like the "staggered" bit in the bottom-left of the third map in the second row 11:48:15 that's one of hangedman's vaults 11:48:54 it does nicely show how non-standard shapes are fine now 11:51:17 pff, I can't find half of my maps in any of these, clearly worst example maps possible 11:52:58 alternatively my vaults need to not be the minority but that'd take foreverrrrr 11:52:59 haha 11:53:13 can set WEIGHT higher 11:53:30 nah, not necessary 11:53:51 just thought things like _hat or _corner would place more 11:53:53 but, it proves that we have enough maps already for things to not be repetetive, if i can generate 24 layouts and not even hit half your map 11:54:11 yes, a good sign overall 11:54:57 HangedMan: _hat is in the map at (3,4) 11:55:17 and at (5,1) 11:55:22 you need to look closer ;) 11:56:11 sometimes maps blend in quite well so they don't really look like distinct entities 11:56:16 -!- evilmike has joined ##crawl-dev 11:56:36 -!- ihalpi has quit [Quit: Page closed] 11:56:38 yes, have made _too_ good maps 11:57:02 hello evilmike, look at http://pasteboard.co/2DSP5n5Y.jpg 11:57:10 still causing trouble, are you? 11:57:32 oh, nice. mumra sure seems to have done a good job with that 11:58:47 that one diamond shape seems to like popping up a lot :) 11:59:19 -!- HangedMan has quit [Quit: leaving] 12:00:25 -!- blackcustard has joined ##crawl-dev 12:00:27 clustering illusion :P 12:03:03 -!- mreisz_ has quit [Quit: Page closed] 12:03:22 also, reasons i stated why we need more small vaults. we also need more big vaults. but more more small vaults than more big vaults. 12:06:29 need more more small new newnewvaults vaults. i cannot state this enough times. 12:07:37 btw I think newnewvaults needs a flashy name like inception 12:08:17 elliptic: it has one! 12:08:26 elliott: is it "newnewvaults"? 12:08:30 no ask mumra 12:08:50 elliptic: it's "Vaults: Legacy" 12:09:09 ??legacy 12:09:09 I don't have a page labeled legacy in my learndb. 12:09:51 I feel like I am missing some reference, but okay :P 12:10:07 tron sequel 12:10:07 elliptic: where are my 500 crawler's thanks by the way?? 12:10:27 yeah, it's not as obvious a reference as Inception 12:10:37 st_: elliptic is withholding them until he merges newnevaults 12:10:39 which he will never do!! 12:10:40 st_: well they won't thank you until they have seen it 12:10:50 which only a few people have 12:11:04 but certainly you and mumra have my thanks for working on this :) 12:20:05 -!- y2s82 has quit [Remote host closed the connection] 12:30:37 -!- Vizerr has quit [Ping timeout: 246 seconds] 12:31:16 -!- Fhqwhgads has quit [Ping timeout: 245 seconds] 12:31:28 -!- Fhqwhgads_ is now known as Fhqwhgads 12:32:52 -!- DainDwarf has quit [Changing host] 12:34:04 -!- MarvinPA has joined ##crawl-dev 12:38:39 -!- shirish has joined ##crawl-dev 12:40:13 -!- puppyknuckled has quit [Quit: Page closed] 12:48:42 -!- SurpriseTRex has quit [Ping timeout: 256 seconds] 12:49:51 -!- codile has quit [Ping timeout: 276 seconds] 12:50:17 -!- Elkan has quit [Read error: Operation timed out] 12:59:16 -!- Kalir has quit [Changing host] 13:10:56 -!- BlastHardcheese has quit [Quit: Talk is cheap because the supply exceeds the demand.] 13:25:03 -!- PsyMar has quit [Quit: Be excellent to one another, and party on dudes!] 13:29:33 -!- shirish has quit [Ping timeout: 252 seconds] 13:33:23 -!- Kellhus has quit [Ping timeout: 245 seconds] 13:35:52 http://pastebin.com/Juq5qZ1n 13:35:54 <_< 13:38:34 -!- Rebthor has quit [Ping timeout: 256 seconds] 13:43:48 Grunt: you're an evil man 13:44:48 ontoclasm: why, thank you! 13:45:07 not necessarily a compliment there 13:46:24 -!- CampinSam has quit [Read error: Connection reset by peer] 13:48:17 It is if I say it is. :b 13:50:10 -!- CampinSam has quit [Remote host closed the connection] 13:50:52 -!- Blade- has quit [Quit: Page closed] 13:51:31 -!- bonghitz has quit [Quit: bonghitz] 13:55:16 -!- shirish has joined ##crawl-dev 13:56:14 I don't think we should have subvaults that place specific stuff like that 13:56:39 we went down that road with newvaults 13:58:55 i think there can be some with very low weight; it's fun to see specific stuff now and then, just not every game 13:59:13 -!- Sobieck has quit [Ping timeout: 245 seconds] 13:59:15 it should probably be a vaults_hard anyway 13:59:32 -!- scummos has quit [Ping timeout: 256 seconds] 13:59:48 although not sure about how uniques work ... for instance what happens if that unique has already appeared somewhere? 14:00:15 It moves on to another selection. 14:00:36 I agree with st_ 14:00:37 There are other vaults that do something similar (that one vaults entry vault immediately comes to mind) 14:00:40 on the first account, I mean 14:01:17 it's pretty much exactly the same as that entry vault except with necromancers and wizards 14:02:12 yeah but compare to vaults in the main dungeon ... occasionally there is very specific stuff and a lot of players like that because they tend to be quite memory ... but it only works if you'll only see that stuff once in dozens or hundreds of games 14:02:32 -!- WalkerBoh has quit [Remote host closed the connection] 14:02:59 (although i'm making no comment about this particular vault, it is obviously stealing V:$'s gimmick) 14:03:21 not every branch needs to provide the same things 14:04:36 Grunt: a comment on the stairs - they might not function how you expect. they will all get removed when the vault is placed. then some *might* be restored at the end of layout building. however which number staircase they are or whether they are up or down is irrelevant. they might as well all just be ( . 14:06:32 Grunt: having a higher number of staircases does make a room more likely to get stairs at the end right now however (but this might change) 14:09:19 I see. 14:13:53 -!- codile_ has quit [Ping timeout: 255 seconds] 14:16:03 -!- syllogism has quit [] 14:25:48 -!- DracoOmega has joined ##crawl-dev 14:25:53 Grunt: we already have an entry vault and vaults:$ ambush, I don't see much value in repeating the gimmick yet again inside V 14:26:29 vault guards are basically the most boring monsters anyway 14:27:00 -!- bonghitz has quit [Ping timeout: 276 seconds] 14:27:00 -!- bonghitz_ is now known as bonghitz 14:27:17 cyprian (L11 SpEn) (D (Sprint)) 14:27:33 elliptic: Thanks for the new feedback on the conjurer patch. I'm glad things seem to be working out much better 14:27:34 DracoOmega: You have 2 messages. Use !messages to read them. 14:27:54 Those bugs you mention sound easy to fix, I think. I'll get on them shortly 14:28:58 DracoOmega: newfamiliar actually reminds me a bit of tornado 14:29:03 Oh? 14:29:07 Howso? 14:29:27 does it synergise with tornado 14:29:38 phew ... i finally find out *why* layout generation is sometimes taking forever ... but i don't know how to fix it (yet) 14:29:42 -!- Heteroy has quit [Ping timeout: 256 seconds] 14:30:04 the good news is it's nothing to do with my lua :) ... and i actually found some further optimisations and fixed something while i was looking for it 14:30:05 mumra: do tell? 14:30:16 DracoOmega: good damage/MP ratio compared with other spells of similar level, but you have to wait for the damage 14:30:27 sometimes resolve_map is taking an extraordinary amount of time, on certain type of layout 14:30:30 Ah, in that sense. Yeah. 14:30:44 because resolve_map has to find somewhere it can *potentially* place the map in the level in order to work 14:31:00 and on the "maze" type levels there is basically no open space where it can easily do that 14:31:34 DracoOmega: (I view this as a good feature both of tornado and of familiar, if that isn't clear... helps to differentiate them from other spells) 14:31:41 Yeah, I agree 14:31:58 so basically i need to find a way for resolve_map to not require level space and we're all good 14:32:14 As for that bug, familiar definitely works fine when you target an empty space with most spells (as I have tested this a bazillion times). But there was some special-casing for explosions back when it used to mirror them, and it's possible this has somehow become obsolete in a bad way 14:32:31 But I think it's probably as simple as removing that 14:33:07 DracoOmega: yeah, I spent a while trying to break familiar with simple projectile spells and failed 14:33:47 It took ages to finally get it working right, but I do think it is pretty consistent now (and it sounds like it also was in practice) 14:35:17 Yeah, I think this is as simple as removing a '&& !testbeam.is_explosion' from one if statement 14:36:23 By the way, the bit about dazzling spray not choosing optimal target numbers... isn't that true currently for a lot of more conventional AoE spells, too? 14:36:32 yes it is! 14:36:52 Like, I didn't think that was anything unique to this spell (I have not looked into how default targets are chosen in the first place) 14:36:52 won;t stop you fixing all of them!!! 14:36:58 heh 14:37:21 DracoOmega: yes, it isn't something that would be easy to fix either 14:37:27 -!- Pthing has quit [Remote host closed the connection] 14:37:28 just something a bit unfortunate 14:37:36 So, that bit wasn't a 'try to fix this before I merge' deal? :) 14:37:59 no, though if you ever want to try to improve default targetting for spells in general then that would be cool :P 14:39:07 I may take a look at some point. I am not sure how hard that will be to fix in practice, though. Less for code and more for algorithmic reasons 14:39:46 Though I suppose 'better than present' is still fine even if it's not 'optimal' 14:40:02 optimality optimal 14:40:38 -!- Zaba has quit [Read error: Operation timed out] 14:40:40 fwiw, the most dumb things that strike me all the time in practice are (a) target defaulting to plant or mushroom, and (b) cloud spells saying "no monsters in range" when there are monsters, because they're already in a cloud 14:40:42 Yup, that looks like it fixed it 14:40:46 those might be easy cases to fix 14:40:57 The conditional was left over from when it did something totally different, I think 14:41:14 And no longer made any sense 14:42:12 -!- Zaba has joined ##crawl-dev 14:43:03 -!- caleba has quit [Quit: caleba] 14:44:47 -!- Drahnier has quit [Client Quit] 14:49:06 why do I have to say "Y" when i'm walking through clouds of scalding steam, when it doesn't hurt at all? 14:49:38 Didn't my patch for that already get put in ages ago? 14:49:53 perhaps it's not in 0.11 14:50:28 %git :/steam 14:50:43 03Eronarn * 0.12-a0-1475-g23c7d36: Lava orcs cool down in water, making steam. However, lava orcs don't cool down when in a rage. 10(1 year, 10 months ago, 1 file, 24+ 0-) 13http://git.develz.org/?p=crawl.git;a=commitdiff;h=23c7d36932e7 14:50:43 %git a18b63e 14:50:43 03MarvinPA * 0.12-a0-393-ga18b63e: Don't prompt for walking into steam unless on low HP (DracoOmega) 10(5 months ago, 1 file, 11+ 0-) 13http://git.develz.org/?p=crawl.git;a=commitdiff;h=a18b63e3bf48 14:50:48 Ah 14:51:00 Somehow I thought it was, but I guess not 14:51:13 5 months ago but after 0.11 feature freeze I guess 14:51:18 stable versions, even more ancient than you'd assume 14:51:35 what's the estimate for 0.12 freeze 14:51:36 ok, thanks 14:51:41 if there is one 14:52:18 -!- Elkan has quit [Ping timeout: 244 seconds] 14:52:51 ontoclasm: "soon after newnewvaults and newconjurer are in" would be my guess 14:53:36 Are newnewvaults not in yet 14:53:56 no 14:54:02 Maybe we should shorten it to (new^2)vaults => N2vaults => nitrovaults 14:54:38 is there a branch i can checkout that has newnewvaults? I can't seem to find it 14:54:55 it's on mumra's account 14:54:58 evilmike: https://gitorious.org/~mumra/crawl/newnewvaults 14:54:59 (I'm just curious and want to take a look, I'm still inactive) 14:55:00 ah 14:55:01 https://gitorious.org/~mumra/crawl/newnewvaults/commits/newnewvaults 14:55:08 thank you 14:55:42 evilmike: every now and then it will freeze up for a long time during layout generation. it'll get there eventually. i've narrowed down the problem to resolve_map 14:56:20 is it the same problem that sometimes freezes late in D, especially on D:27? that issue has been around for a while 14:56:38 that's entirely possible, or at least related 14:56:39 in debug mode i noticed it happens when it's trying and failing to place a minivault. that's all i know though 14:56:54 that's basically what's happening here but probably for a different reason 14:59:04 evilmike lives! o_O 14:59:47 yeah i'm using irc again because someone sent me an email the other day amounting to "Are you still alive?". don't expect me to be here on weekdays though 15:00:29 elliptic: It looks like the issue of dazzling spray falsly warning you about hitting your familiar is actually a wider bug that just shows up rarely 15:00:58 elliptic: The hit warning doesn't seem to know that it can't actually hit it. (It also will warn you about using mottled draconian breath next to an orb of destruction, for example) 15:01:03 -!- codile has quit [Quit: WeeChat 0.4.0] 15:01:04 evilmike could possibly be an IRC-using zombie 15:01:25 * Grunt throws his arms wide. evilmike the human zombie comes into view. The dead are walking! 15:01:29 really anyone here could be a zombie 15:01:53 DracoOmega: hm, it doesn't warn you when you try to fire a magic dart through a familiar though 15:01:59 -!- DainDwarf has quit [Quit: Dain's default quit message] 15:02:03 Yes, that uses a different type of target warning 15:02:06 speaking of tohit warnings i was just about to look into disabling those for fire vortices because ugh it is the worst 15:02:08 Seperate codepaths, I believe 15:02:14 -!- monqy has joined ##crawl-dev 15:02:25 But the unique hitfunc of the spray needed the other routine, I believe 15:02:31 (As with mottled breath) 15:02:36 -!- magistern has quit [Quit: This computer has gone to sleep] 15:02:50 -!- Sobieck has quit [Quit: Page closed] 15:03:07 MarvinPA: As in let you shoot through them, or just not bother to warn you if you ARE going to hit them? 15:03:14 just the latter 15:03:17 -!- rkd has quit [Read error: Connection reset by peer] 15:03:41 i guess the problem there is shooting through them as opposed to them being behind your target 15:03:58 Debian builds of 0.11 branch on crawl.develz.org updated to: 0.11.2 15:04:07 since there should probably still be a warning if they're blocking your line of fire to the target 15:04:38 Yeah 15:07:48 make them hostile 15:11:28 -!- bmfx has quit [Read error: Connection reset by peer] 15:14:47 -!- Sombrero_Mott has joined ##crawl-dev 15:15:24 -!- SamB has quit [Read error: Connection reset by peer] 15:16:50 Well, I can make dazzling spray not prompt on familiars easily enough, but this feels like an imcomplete fix. Unfortunately, a more complete one looks fairly tricky. 15:17:02 (Also, it looks like mottled breath has a few other cases of oddness, too) 15:17:33 Basically, the code for checking whether an arbitrary hit_func will affect something is almost wholely seperate from beam tracer code 15:17:49 Meaning things which the beam code itself says "I can't hit" are ignored 15:18:26 -!- Mottie has quit [Ping timeout: 256 seconds] 15:18:52 Dammit... this probably means that dazzling spray warns about shooting through plants with Fedhas, too =/ 15:19:34 Oh, no it doesn't 15:19:38 Though this confuses me more 15:19:50 how did you assume similar stuff goes through similar code paths 15:19:53 its crawl after all 15:22:56 Oops, nevermind 15:23:02 It DOES prompt unnecessarily for Fedhas plants 15:23:09 I just had the relevant bit commented out at the time, while testing something 15:25:37 Need to figure out the most sensible way to fix this that doesn't just duplicate a whole bunch of code in another place 15:25:53 -!- rkd2 has quit [] 15:25:53 -!- rkd has quit [] 15:27:17 -!- |amethys1 has joined ##crawl-dev 15:27:17 -!- |amethys1 has quit [Client Quit] 15:29:02 -!- Alheris has quit [Ping timeout: 256 seconds] 15:29:35 -!- Crazylemon64 has quit [Ping timeout: 255 seconds] 15:35:56 sweet ... fixed the newnewvaults freezing problem. level generation is always fast now. could make it a bit faster at the expense of not always trying to hard to place more awkward rooms, if people find it's too slow. 15:36:06 turns out there was another undocumented lua parameter 15:38:03 mmm 15:39:04 -!- |amethys1 has joined ##crawl-dev 15:39:22 -!- |amethyst has quit [Disconnected by services] 15:39:35 -!- |amethys1 is now known as \amethyst 15:40:29 What did this undocumented parameter do? 15:44:27 -!- bob123 has quit [Quit: Page closed] 15:44:49 pfui 15:45:03 i have to walk through miasma to get the rune of swamp? 15:45:12 how sucky is that... 15:47:09 Well, sometimes you can blink or scblink across 15:48:54 DracoOmega: the parameter was "check collisions" 15:50:21 -!- Kalir has quit [Ping timeout: 252 seconds] 15:53:45 -!- kek has quit [Quit: Kicked by KickServ] 15:53:59 -!- Kalir has quit [Changing host] 15:54:06 -!- HangedMan has joined ##crawl-dev 15:55:14 -!- Mandevil has quit [Quit: Leaving] 16:01:01 -!- keszocze has quit [Ping timeout: 244 seconds] 16:01:57 -!- SamB has joined ##crawl-dev 16:07:13 -!- Jayrays has quit [Ping timeout: 244 seconds] 16:07:41 -!- \amethyst has quit [Quit: leaving] 16:10:43 -!- gffa has quit [Quit: sleep] 16:16:14 -!- Nightbeer has quit [Ping timeout: 256 seconds] 16:17:25 -!- |amethyst has joined ##crawl-dev 16:20:43 -!- WalkerBoh has quit [Remote host closed the connection] 16:25:08 -!- dtsund has quit [Ping timeout: 245 seconds] 16:27:37 I may have to go with a slightly hacky solution here, since a more proper one seems to require deeper refactoring of beam code 16:27:55 Which is another issue altogether 16:28:26 ??goodcode 16:28:27 goodcode[1/1]: beem.is_beam = false 16:28:48 beem.is_beem 16:28:51 Since there isn't currently a way to poll whether a beam can affect a given monster, independly of firing that beam 16:29:01 elliott: would it still be false 16:29:08 copy the state of the world, fire the beam, observe the results, put the old world back 16:29:25 Yes, somehow I think just specifically excluding a couple monsters here is less hacky than that :P 16:29:51 But there definitely should be a more generalizable solution 16:30:04 (But I'm inclined to think that shouldn't hold up the conjurer patch currently) 16:30:20 -!- hayenne has quit [Ping timeout: 255 seconds] 16:31:37 Since who knows how much work and problems trying to do that could uncover? 16:32:17 -!- Crazylemon65 has quit [Ping timeout: 252 seconds] 16:33:28 -!- oxeimon has quit [Client Quit] 16:35:57 DracoOmega: the beam code is an ancient monstrosity that hasn't ever been rewritten 16:36:14 which because of Crawl's history means something 16:36:41 it's nowhere as bad as, say, teleport/blink or throwing code, but still 16:37:05 Yes, there are a few conspicuous limitations on how it works here 16:39:04 But I think I can fix this specific bug in this specific case without too much hackishness 16:39:19 Even if it partially duplicates a check in the beam code 16:39:38 (Because there is currently no way to access that check outside of firing the beam, which raises other problems) 16:40:53 -!- andrewhl has quit [Remote host closed the connection] 16:42:48 I'm afraid there's no real way other than refactoring. Or at least adding some hacks. 16:43:31 Got abyssed, decided it was time to a break. Later on I was unable to login and others have tried aswell. (https://crawl.develz.org/mantis/view.php?id=6575) by timbw 16:43:37 Yeah 16:43:55 Which probably should be done eventually. But is a little big in scope to do just for this, I think 16:44:19 apparently it was crashing for timbw when he tried to load the game 16:45:27 http://dobrazupa.org/morgue/timbw/ 16:45:37 what is all this crash-recursive stuff here 16:45:58 iirc blade got a message like that when moving out of vaults was timing out 16:52:33 I've been getting cloud-related crashes in debug mode stress tests (ie, with an assert sweep every turn), since from some time before newnewabyss merge. 16:52:50 -!- Elkan has quit [Ping timeout: 260 seconds] 16:53:32 no idea about the cause, but at least, the assert should come before saving, not after 16:53:50 (so there'd be a crash but before the save is corrupted) 17:00:51 -!- Kalir has quit [Quit: CADENZA-CLASS CLOCKWORK KNIGHT--OPERATION TERMINATED] 17:02:04 I made it so; this will probably greatly increase the number of crashes though 17:02:06 -!- Ragdoll has quit [Quit: Ik ben weg] 17:03:33 Abyss keeps getting saved over and over but hardly ever read (only when you actually exit the game or use X[]), so we happily wrote bogus data that would result in corrupted games 17:04:01 03kilobyte 07* 0.12-a0-1843-g6b9242b: Crash games with bogus clouds when saving rather than loading. 10(4 minutes ago, 1 file, 10+ 0-) 13http://git.develz.org/?p=crawl.git;a=commitdiff;h=6b9242bb7284 17:04:11 -!- Koolguydude has quit [Quit: ChatZilla 0.9.89 [Firefox 18.0.1/20130116073211]] 17:05:20 Unstable branch on crawl.s-z.org updated to: 0.12-a0-1843-g6b9242b (34) 17:05:37 -!- Nightbeer has quit [Quit: Verlassend] 17:08:51 -!- sgiratch has quit [Ping timeout: 276 seconds] 17:10:22 -!- NeremWorld has quit [Read error: Connection reset by peer] 17:13:25 -!- Nightbeer has quit [Client Quit] 17:15:25 elliptic: Okay, I have fixed the bugs you mentioned (and one other I discovered in the process) 17:15:44 Should I rebase and squash a bunch of this stuff together for a somewhat tidier patch that might be mergable then? 17:16:32 if doing that isn't too hard then that sounds good, yeah 17:16:51 Yeah, it shouldn't be hard to squash some of the bugfixes into the main spell commits, I think 17:16:56 If you feel it makes sense to do this 17:21:36 arena and toadstools: should they be removed altogether, or excluded from faction counts? The latter sounds like a more proper fix, but I'm afraid that monsters may look at friendly toadstools not exactly the same as at hostile ones. 17:22:26 new attitude for toadstools 17:22:37 ATT_TOADSTOOL 17:22:59 or at least ATT_NEUTRAL, which is already bogus enough that no one cares 17:23:38 (ie, AI is mostly disabled, which removes any concerns about fairness) 17:32:10 -!- animegrampa has quit [Ping timeout: 256 seconds] 17:33:20 i like the toadstools 17:35:53 is removing them from crawl an option 17:36:14 because i'd vote for that (or at the very least bring back bow bashing) 17:36:21 ChrisOelmueller: +1 17:36:22 it's horrible on the edge to losing sanity currently 17:36:31 toadstool (04f) | Spd: 0 | HD: 1 | HP: 2-4 | AC/EV: 1/0 | 03plant | Res: 13magic(immune), 03poison, 12drown, 13neg+++, 13torm | XP: 0 | Sz: tiny | Int: plant. 17:36:31 %??toadstool 17:36:51 nerf their stats 17:37:00 speed -5 17:37:03 bowbashing still wouldn't work would it 17:37:31 it's not like an orc should grow man-sized mushrooms on its corpse before you finish off the orc next to him 17:37:52 well, size tiny, which means spriggan-sized 17:37:55 (Fedhas invocation excluded, which is magic) 17:37:57 bat-sized 17:38:14 big-enough-to-block-passage-sized, at least 17:38:15 it is as large as a tiny 17:38:29 kilobyte: rat-sized? 17:38:39 butterfly-sized 17:38:54 everything in crawl is mutated to be slightly larger than normal 17:39:00 you see, a flying dragon can't pass because of a rat on the ground, right 17:39:15 it'd just be rude 17:39:38 -!- minqmay has quit [Read error: Operation timed out] 17:39:52 HangedMan: except for dragons and giants, which still fit the same amount of space, and can't squish tiny gnats like humans by just stomping without looking 17:40:15 ~~~***abstraction***~~~ 17:40:39 it works like that in my game! too bad i am too busy to work on it ever :( 17:40:50 -!- mikee_ has joined ##crawl-dev 17:41:06 yeah, and this abstraction means fungus growing on a corpse blocks your way 17:41:32 obviously just make all corpses already rotting have implied fungi mould things on them 17:42:22 fr: mushroom type token so ghouls can get shiitake or button when they eat rotten corpses 17:42:31 mushroom option 17:42:42 Eronarn: :p 17:43:27 anyway, what would you guys say about making toadstools an unthing except for Fedhas' magic? 17:44:46 please do 17:44:59 fedhas' magic and a handful of entry vaults 17:45:00 -!- caleba has quit [Quit: caleba] 17:45:02 silly entry vaults 17:45:46 -!- nooodl has quit [Ping timeout: 256 seconds] 17:46:17 toadstools not "fungus" 17:46:26 oh whoops right 17:46:29 ie, that little shit that lives only several turns 17:47:46 that would have the pleasant side effect of stopping auto cloud targeting from focusing them 17:48:18 kilobyte: add them back in via forest wyrm breath please :) 17:49:30 Eronarn: axing generation on corpses doesn't remove them completely 17:50:05 (sadly, it's not like we can't let Fedhas grow wandering mushrooms from corpses directly) 17:50:12 i would also love them to not spawn from corpses without fedhas, yeah 17:50:26 -!- Koolguydude has quit [Quit: ChatZilla 0.9.89 [Firefox 18.0.1/20130116073211]] 17:51:17 +1 to unthinging non-fedhas toadstools 17:51:32 -!- ussdefiant has joined ##crawl-dev 17:55:24 -!- edlothiol has quit [Ping timeout: 256 seconds] 17:58:01 -!- inspector071 has quit [Quit: inspector071] 17:59:20 -!- Jayrays has quit [Ping timeout: 244 seconds] 18:01:08 -!- _dd has quit [Ping timeout: 255 seconds] 18:07:58 -!- Kellhus has quit [Ping timeout: 245 seconds] 18:09:26