00:05:25 -!- paxed has quit [Ping timeout: 276 seconds] 00:06:22 -!- paxed has joined ##crawl-dev 00:29:28 -!- paxed has quit [Ping timeout: 276 seconds] 00:30:58 -!- paxed has joined ##crawl-dev 00:52:12 -!- paxed has quit [Ping timeout: 276 seconds] 00:53:33 -!- paxed has joined ##crawl-dev 01:09:32 -!- eith has quit [Ping timeout: 265 seconds] 01:15:37 -!- paxed has quit [Ping timeout: 276 seconds] 01:17:08 -!- paxed has joined ##crawl-dev 01:32:31 -!- Zaba has quit [Ping timeout: 276 seconds] 01:38:22 -!- paxed has quit [Ping timeout: 276 seconds] 01:39:42 -!- paxed has joined ##crawl-dev 01:43:02 -!- Zaba has joined ##crawl-dev 01:54:16 moin 02:01:06 -!- paxed has quit [Ping timeout: 276 seconds] 02:02:16 -!- paxed has joined ##crawl-dev 02:02:59 greensnark, ping 02:03:10 Miaow 02:04:08 greensnark, regarding building primary vaults after layouts... It can be done, and I believe that with some tweaking, it can be beaten into looking good (_build_rooms produces a lot of nasty dead ends, Dis:7 has messed up rooms, but all that can be worked around with some luck) 02:04:50 greensnark, but another issue is cleaning up the code path.. We only need to build a layout when the primary vault is non-encompassing, but if we build the layout before we pick the vault, how do we know it's going to be encompassing? 02:04:54 Zaba: Well, the main advantage would be that we wouldn't have to use _build_rooms at all if _build_rooms sucks 02:05:10 yes, _plan_3 can wholly replace it with minor tweaking 02:05:32 Zaba: If the vault is an encompass vault, we can just destroy the built level 02:05:39 but that's wasteful 02:06:08 Or you can pick the vault, see if it's encompass; if not build skeleton + vault else build vault 02:06:20 Minor details are all yours :P 02:07:19 greensnark, also.. it'd be nice if one could choose the layout for the level with random_map_for_tag("layout") or so 02:08:08 existing C++ layouts could be wrapped as they are into API functions, which would be called from otherwise empty layout vault definitions.. Which would exist solely for the purposes of their depth specification. 02:08:22 that'd be ugly as hell, but I think it'd work 02:08:30 Zaba: Well, you can do that right now 02:08:34 greensnark, not really 02:08:38 Why so 02:08:53 Or why not rather 02:08:54 I can't specify a layout that has, say, weight 5 on D:5-22, and chance 50% in Vault:* 02:09:11 I can do two different layout vaults, but I don't want to. 02:09:17 because, bloody hell. 02:17:11 greensnark, and I'm not sure how to implement -that- 02:19:36 Why can't re re-implement all current builders, if not into Lua, at least into Lua-wrapped-C++? 02:20:18 due, because we are lazy 02:20:36 due, all _new_ builders should be implemented in lua, or in lots-of-additions-to-the-lua-api+a-bit-of-lua 02:21:31 Well, the infrastructure is not quite there at the moment to make that less than painful 02:21:57 greensnark, well, _that_ should be fixed; implementing more layouts in C++ is definitely not the way to go, given the situation with _builder_* functions we have now 02:23:04 Ideally yes, but I don't think we should hold up new layout generation until we get our Lua ducks in a row 02:23:14 *new layout creation 02:23:41 oh, I wasn't saying anything about creation of new layouts 02:23:49 I mean if someone has a cool new layout we shouldn't make them worry about Lua and des files and stuff 02:23:52 but it might be that it will be necessary to hold down their inclusion into the game 02:23:52 -!- paxed has quit [Ping timeout: 276 seconds] 02:24:51 -!- paxed has joined ##crawl-dev 02:26:59 greensnark, wrapping them in single API functions, and creating layout vaults that call these functions, solely for the purposes of their DEPTH/etc. specification is ugly, but acceptable, I guess. 02:28:11 greensnark, would require the aforementioned ability to give a single vault different chance/weight on different depths, though.. 02:29:28 greensnark, that would have the major advantage of having no ugly layout selection code in dungeon.cc. At the moment, ti's bloody craziness that I do _not_ want to add anything to. 02:32:18 greensnark, regarding a better infrastructure for lua layouts: Any obvious things that could be improved? 02:32:37 I don't see anything when glacing at the existing layouts, but then, it might be that I'm secretly a masochist. 02:42:40 seriously though, where exactly's the painfulness? 02:50:31 -!- monky has quit [Quit: hello] 03:02:32 Zaba: Before you can write a nontrivial layout you have to spend time wrapping a bunch of C++ functions 03:02:45 i.e. if you're doing something computationally intensive 03:02:54 for instance the Shoals builder would be a pain to write in Lua 03:03:02 yes, and as long as said wrapping is done in a reusable manner, you only have to do it once 03:03:05 With the current Lua api 03:04:15 greensnark, so effort should be directed to improving the api, rather than reducing the motivation to do so. 03:06:05 Who's reducing the motivation to do so :P 03:06:11 By all means improve the api 03:06:38 anybody who would now write a new, complex layout in C++ would reduce said motivation 03:06:42 and also make me go spare 03:07:21 I don't see any reason to force a layout writer to learn the Lua api 03:07:28 We can always Lua-ize layouts after the fact 03:07:59 but well. 03:08:06 Solution: write a new language for layouts? 03:08:25 holy shit _no_ 03:08:29 ;D 03:10:23 I've actually not been happy with the disconnect between the .des file system of abstract symbols with no place information vs dungeon.cc's explicit features and everything-has-its-placeness 03:10:40 -!- paxed has quit [Ping timeout: 276 seconds] 03:10:52 This makes .des files somewhat fundamentally unsuitable for layouts because you have to bolt on a notion of actual physical coordinates on the real map grid 03:11:15 Why can't we unbolt physical coordinates from map building? 03:11:26 In dungeon.cc, you mean 03:11:29 Don';t the Lua builders use maplines instead? 03:11:36 dungeon.cc has to operate in the constraints of the map grid 03:11:43 Well, yeah. 03:11:50 -!- paxed has joined ##crawl-dev 03:12:01 greensnark, well.. extend_map { width = gxm, height = gym } and using mapgrd seems close enough to using grd directly 03:12:02 But we should just be giving dungeon.cc the final product: exact size of the map, glyphs at which coordinates, and say, "here, place this somewhere". 03:12:17 greensnark, it's not like the dungeon builder has any choice about where to put it 03:12:46 due: That's exactly what happens now. The vault builder runs the Lua, then places that chunk of mapline somewhere it fits 03:12:59 what's the problem with that? 03:13:30 Nothing wrong with that 03:13:41 But it is one level of abstraction away from the real dungeon level 03:14:06 well, I don't see the problem with that. 03:14:26 So if I want to check map masks from Lua, I have obvious issues 03:14:50 that is one point, but seeing how a layout is the first thing built on the level, why'd you want to do that anyway :P 03:15:28 Well, for the Shoals, I want to place plants after vaults are placed 03:15:35 I can't very well do that in layouts now, can I 03:15:56 turn plants into vaults? :P 03:16:16 and let the map mask be checked by the usual map placement code :P 03:16:36 Unsuitable if I want clumping around Shoals rocks 03:16:42 then implement the custom vault place finder functions, and employ such for desirable effect :P 03:17:28 There are many things that involve looking at the actual grid that are far from easy in the vault model 03:17:53 but.. well, yeah. 03:17:55 I could place Shoals plants using maps, but the gymnastics involved would be about 100x the effort of the current implementation 03:18:13 So the idea of moving all-custom-shoals-levelgen into .des files is not really practical with the Lua api as it stands 03:19:19 Personally, I don't think .des files are the answer to all level-gen; I think it's fine to expose low-level map grid ops to Lua and write layouts in plain Lua unrelated to the .des file vault model 03:19:31 we already expose those ops 03:19:39 the zig level builder actively uses them, too 03:19:42 dungeon.cc should eventually become a collection of low-level primitive routines 03:20:02 And a top-level Lua function that decides what to build and how to build it 03:20:24 yes, that's something I'd like to see, too. 03:20:45 That way, when I want to write a layout, I don't have to screw around with a .des file syntax that was really intended for canned scenery, not procedural generation 03:21:18 however, the vault model would be useful for picking layouts by DEPTH/etc. 03:21:39 as to using the grid directly instead of maplines, you can already do that 03:21:46 I think the vault model is unnecessarily heavy just for picking stuff by DEPTH 03:21:55 The trouble with vaults is that they only describe a small piece of the level 03:22:20 the level builder would be a lot simpler if every vault was encompass 03:24:28 greensnark, it should be very well possible to write a layout generator that uses dgn.grid() instead of mapgrd 03:24:38 Zaba: Of course 03:24:45 I did that in the zig gen as you noted :P 03:25:20 But at that point I don't think that's a great use of vaults 03:25:30 yes.. being a vault would be irrelevant for such a builder 03:26:26 (apart from the part where it'd be able to be chosen by tag/depth/etc., as I've mentioned above) 03:27:37 We need first-class access to map_def creation from Lua, that'd be a good step up 03:27:57 It'd probably need swig or something equivalent to not be incredibly annoying to write though 03:28:54 I've been considering trying Luabind 03:29:04 regarding existing lua layouts, it doesn't seem to me that it matters much if they use dgn.grid or mapgrd 03:29:19 Yeah, I don't think it matters much 03:29:38 Using mapgrd has the advantage that you can use shuffles and substs, but it's not clear that anybody bothers 03:29:46 yes, and subvaults 03:30:03 but anyway.. 03:31:08 shuffles and substs could be used instead of some of the hardcoded randomization that is currently present in dungeon.cc 03:32:49 greensnark, what exactly would first-class access to map_def creation let one do? 03:33:14 Zaba: define_vault("my_awesome_layout", layout_generation_function) 03:33:25 -!- paxed has quit [Ping timeout: 276 seconds] 03:33:55 ah. 03:34:04 hm. 03:34:11 Stuff like define_vault_based_on(find_map_by_name("cow").clone(), function() end) 03:34:24 -!- paxed has joined ##crawl-dev 03:34:57 and 'layout_generation_function' would be a lua chunk akin to those that follow NAME: in .des files? 03:35:19 Yes, you can set up CHANCE and DEPTH if you want and then call the actual builder routine 03:35:38 Actually both the uses I've just described would be pretty trivial even without first class vault access 03:35:54 hm. 03:36:25 I'm kind of confused as to how that would help reducing the amount of crappiness in dungeon.cc, though :P 03:36:43 -!- syllogism has joined ##crawl-dev 03:37:06 Zaba: You could define layouts as simple functions in a Lua file instead of shoehorning them into .des files 03:37:21 hm, okay, that sounds good 03:37:40 hrm. 03:38:01 anyway.. what should I do now. 03:38:16 Recover from your illness :P 03:38:57 I'm already feeling better, thanks. But what should I do with regard to the dungeon builder. 03:39:17 * Napkin whispers hackaholic * 03:39:24 moin guys :) 03:39:32 Napkin: You got snow there too? :) 03:39:32 moin Napkin 03:40:05 omg, don't ask - lot of new snow from during the night :-O 03:40:18 hi napking 03:40:29 Napkin: Yeah, bhaak was saying it's snowing where he's at, so I wondered :) 03:40:32 so glad i did everything left yesterday and don't need to leave the house today :D 03:40:48 moin due :) 03:40:48 bhaak obviously planned his long drive for a suitably snowy day 03:40:57 oh, pity 03:41:44 we have about 60cm snow with strong winds 03:41:58 Zaba: Pick something that looks like fun and do it :) 03:42:13 That's what I usually do with Crawl 03:42:18 if you never ever hear from me, I got snowed in 03:42:38 bhaak: Just in case, take a St. Bernard with flask of whisky along 03:42:42 well.. I'd like to kill _builder_{by_branch,normal} dead, but that's beyond me in certain regards 03:43:01 how about the rain in india? did it stop for good recently? 03:43:23 Napkin: No rain in Chennai for the last couple of weeks 03:43:33 that's great 03:43:36 And a very pleasant 20-30C temperature range :) 03:43:53 ahhhhhh! i want, i want! :D 03:44:02 I imagine due has some warm weather too :) 03:44:12 how's the storm, due? 03:44:13 greensnark: does this one count as a St. Bernard? http://bhaak.dyndns.org/misc/pics/bilbo_shiva2.jpg 03:44:39 lol - if they team up, yes ;) 03:44:41 bhaak: As long as it can carry a big flask of whisky on its collar 03:45:11 he can't even carry a newspaper, so I don't think so 03:45:55 Wikipedia tells me the St Bernard stereotype does brandy, not whisky, my mistake 03:46:13 'A Punch magazine cartoon from 1949 depicts a man with a St. Bernard and several puppies, all of which are wearing neck casks. The man explains, "Of course, I only breed them for the brandy."' 03:46:19 Wikipedia is sometimes awesome 03:46:48 hey greensnark - Ford is out-sourcing from Chennai to the Phillippines - i hope that's not a trend? 03:47:11 hum.. the fact _build_rooms builds ten more rooms when the primary vault is ORIENT: float is so.. weird 03:47:24 or rather, makes ten more tries to build a room 03:47:26 It's actually getting reasonably common, living costs in India are rising 03:47:29 it never builds more than 30 rooms anyway 03:48:06 and I think it makes sense to try to build some more for primary vaults with any ORIENT... 03:48:56 geez... after about year the guys in chennai finally knew what they are doing.. and now we have to deal with noobs again... very annoying. 03:50:10 Napkin: Hehe 03:53:19 hmm. 03:55:30 -!- paxed has quit [Ping timeout: 276 seconds] 03:56:58 -!- paxed has joined ##crawl-dev 04:18:16 -!- paxed has quit [Ping timeout: 276 seconds] 04:19:32 -!- paxed has joined ##crawl-dev 04:37:39 -!- Mu_ has joined ##crawl-dev 04:40:35 -!- elliptic has quit [Ping timeout: 240 seconds] 04:40:49 -!- paxed has quit [Ping timeout: 260 seconds] 04:42:04 -!- paxed has joined ##crawl-dev 05:03:07 -!- paxed has quit [Ping timeout: 276 seconds] 05:04:39 -!- paxed has joined ##crawl-dev 05:17:59 03kilobyte * r1da7669f35d2 10/crawl-ref/source/itemname.cc: Fix "the spriggan corpse monk". 05:18:33 spriggan corpse monk <3 05:19:54 <3 05:20:17 corpse monk is a cool role 05:20:27 "the spriggan corpse annihilator" is better :) 05:20:52 <333333 05:21:07 I want to be a corpse annihilator too 05:21:27 all uses of name_suffix are in Wucad Mu's wizlab where they are summoned, so this bug didn't manifest before 05:21:48 * greensnark proposes corpse annihilator as a new class^Wjob^Wbackground^Wmétier 05:22:23 m'etier? 05:22:33 kilobyte: ahhh 05:25:21 -!- paxed has quit [Ping timeout: 255 seconds] 05:27:14 -!- paxed has joined ##crawl-dev 05:32:26 03kilobyte * r9e1ecaeda202 10/crawl-ref/source/monster.cc: Sheep fires (for greensnark) 05:32:27 Ahahahaha 05:33:21 !!! 05:49:16 -!- paxed has quit [Ping timeout: 276 seconds] 05:50:41 That is hideously awesome. 05:50:48 -!- paxed has joined ##crawl-dev 05:57:22 gunpowder sheep 05:57:34 Can we add sheep pyromancers 05:57:47 sheep firebugs 06:00:06 bonine aviation 06:00:09 aka flying sheeps 06:42:23 -!- dpeg has joined ##crawl-dev 06:42:29 cheers 06:42:29 dpeg: You have 1 message. Use !messages to read it. 06:42:33 !messages 06:42:33 (1/1) monky said (10h 32m 45s ago): 14, didn't work for picking up part of a stack locally either, with both numrow and numlock with either numlock configuration. 06:42:43 thanks 06:42:58 I somehow expected that, and using 5 would've been impossible in any case 06:48:13 sheep fires crashing the game, i think 06:48:19 (tiles) 06:51:56 !tell greensnark What a beautiful christmas present :) 06:51:56 dpeg: OK, I'll let greensnark know. 06:59:16 All I want for Christmas is explosive sheep 06:59:16 greensnark: You have 1 message. Use !messages to read it. 07:03:49 greensnark: <3 07:04:04 ...my two front sheep 07:09:33 killed 19 sheep with one sticky flame 07:10:06 RangerC on the forum about Ogres: "The devs did a good job making one interesting challenge race out of two bad races." <3 07:10:27 I assume that at least some god should dislike that? 07:14:27 -!- pointless_ has joined ##crawl-dev 07:16:13 -!- blackpenguin has quit [Ping timeout: 276 seconds] 07:18:51 -!- blackpenguin has joined ##crawl-dev 07:20:49 -!- MarvinPA has joined ##crawl-dev 07:24:46 Napkin: We neeeeed you :) 07:25:02 I guess the new sprint map should go into CDO asap. 07:26:15 sweet there's a new sprint map? 07:26:39 oh? what's up? 07:27:04 will a make install do or do i need to manually copy the map? 07:27:47 Mu_: yes 07:28:00 I am somehow forced into this christmas business over here. 07:28:24 Mu_: could you have a quick look at the map (it is by zipcode, so it will be good), and add it to trunk. 07:28:33 -!- MakMorn has joined ##crawl-dev 07:28:36 ok 07:29:05 Napkin: after this, would you update CDO, so that we can play it? 07:29:22 sure - beep me when you are done, Mu 07:29:27 You rock, folks. 07:29:38 haha he has customized oklob plant spellcasters 07:29:49 Could then someone spread the word on ##crawl, and urge people to win as fast as possible? 07:30:14 Weihnachten calling, later folks 07:30:20 enjoy! :) 07:41:06 ooo 07:41:15 03pointless_ 07wandering * r122391d31ea0 10/crawl-ref/source/mon-movetarget.cc: Restructure herd checks 07:41:30 there are a lot of shops but not much gold 07:42:02 back for a moment 07:42:19 if I add a bunch of LOS checks for each sheep on a level each turn, is that going to have too much performance impact? 07:42:20 Mu_: this may not be so bad 07:45:11 pointless_: <3 07:45:32 Not for the LOS checks, but for actually tackling monster movement code :) 07:45:53 well I'm not actually improving the general situation, just complicating it 07:46:12 Making herds stick together? 07:46:19 that's the idea 07:46:33 and I'm only looking at beh_wander really 07:47:56 Mu_: how much gold is there on the map? 07:49:01 a few hundred 07:49:16 Hot damn, a new Sprint map? 07:50:29 yes, by zipcode! 07:50:39 03Mu * r5bf008974226 10/crawl-ref/source/dat/des/sprint/sprint_fedhas.des: New sprint map by 78291 07:50:39 Mu_: ah, that sounds scarce indeed 07:50:45 <3 07:50:56 easy to add if it's needed 07:51:29 yes 07:51:35 it might also need some food too, can tweak such things later 07:51:55 lots of oklobs :) 07:52:06 players should be dying, not eating or buying 07:52:15 haha 07:52:24 Napkin - beep! 07:52:44 compile running :) 07:52:49 ;-) 07:54:29 Dungeon looks way more interesting if you disable dummy vaults :P 07:56:58 the new spring is very themed 07:56:58 Unstable branch on CDO updated to: 0.8.0-a0-4044-g5bf0089 (31.18) 07:57:53 Spring? :P 07:58:06 I think winter weighs heavily on our Finnish friends :P 07:58:17 They try to talk about Sprint and keep referring to Spring instead 07:58:35 :) 07:58:59 did someone tell the ##crawl crowd? 07:59:22 did gretell announce the build to ##crawl too, btw? 08:00:22 death to bumblebee 08:01:03 Napkin: Add it to ##crawl topic? 08:01:36 While we're about it, can we remove the really ancient reference to 0.7 release 08:01:52 pardon? 08:02:04 you want me to add it there? 08:02:04 Napkin: ##crawl topic update! 08:02:10 Also remove tourney notice :) 08:02:29 i was just curious if gretell autmatically announced the build in ##crawl too 08:02:33 there's a spriggan berserker that has BiA 08:02:36 Napkin: Yes 08:02:46 Napkin: But please update the topic! 08:02:51 Aha 08:03:00 Aargh 08:03:03 I don't want +o :P 08:03:07 too late 08:03:14 sorry, busy here ;) 08:03:23 Merry Christmas :P 08:03:35 exactly :D 08:03:41 and to you too :) 08:03:54 !send Napkin schnapps 08:03:55 Sending schnapps to Napkin. 08:04:09 Have fun :) 08:04:17 hehe, thanks ;-P 08:06:28 dpeg going for the first win already :P 08:06:33 or not 08:06:37 level 8 08:07:20 03pointless_ 07wandering * r7d44ab2b8cbb 10/crawl-ref/source/mon-movetarget.cc: Make isolated herd monsters wander less spastically 08:07:38 oklob plant meteorologist 08:08:56 What, really? 08:09:01 yes 08:09:54 awesome 08:10:02 syllogism: playing or reading? 08:10:10 reading, probably not going to play 08:10:48 dpeg: I'm going to be out of town for a few weeks, could you get someone to review the "wandering" branch for performance? 08:12:09 pointless_: yes, will do. Expect to be !told. 08:12:13 syllogism: not playing anymore? 08:12:44 the new sprint doesnt look like my thing 08:13:23 it's a short and basic, nothing wrong with that 08:14:30 ah, ok 08:20:13 Revelling in nature, these protectors of forests are the doom of anyone who as 08:20:16 much as snaps a twig. 08:34:43 13921 dpeg the Brawler (level 11, -59/154 HPs) 08:34:46 sniff 08:35:03 those BiA spriggans are going to be basically impossible if you dont get lucky 08:35:10 yes, of course 08:35:28 I should have called my own brothers in advance. 08:35:47 It feels slightly unthematic to have Trog spriggans in a Fedhas map. 08:39:55 felids are pretty silly in sprint 08:42:54 I can imagine. 08:43:32 it might make more sense if felids always respawned back on the starting staircase when they die 08:43:34 should specialcase sprint so when you die it puts you back to the stairs 08:43:39 heh 09:04:00 -!- casmith789 has joined ##crawl-dev 09:11:07 -!- Hehfiel has quit [Remote host closed the connection] 09:13:49 -!- Hehfiel has joined ##crawl-dev 09:22:58 -!- Lollipop has joined ##crawl-dev 09:27:50 -!- pointless__ has joined ##crawl-dev 09:29:46 -!- pointless_ has quit [Ping timeout: 260 seconds] 09:29:47 -!- pointless__ is now known as pointless_ 09:49:44 -!- cw_ has joined ##crawl-dev 09:51:25 -!- eith has joined ##crawl-dev 09:55:48 -!- pointless_ has quit [Quit: ChatZilla 0.9.86 [Firefox 3.6.7/20100713130626]] 10:17:37 -!- eith has quit [Ping timeout: 260 seconds] 10:30:07 -!- eith has joined ##crawl-dev 10:43:38 03greensnark * r6a36820c5d0b 10/crawl-ref/source/stuff.cc: Don't try to use curses in print_error_screen if curses hasn't been initialised yet. 10:43:40 03greensnark * r9e6e6919463b 10/crawl-ref/ (16 files in 3 dirs): Allow maps to specify different CHANCE and WEIGHT for different depth ranges (Zaba). 10:43:41 03greensnark * r53fa0ff3f1cd 10/crawl-ref/source/prebuilt/ (levcomp.lex.cc levcomp.tab.cc levcomp.tab.h): Update canned level compiler for depth-constrained CHANCE and WEIGHT changes. 10:53:24 greensnark, yay 10:53:39 greensnark, hrm, now I get .../dat/des/portals/lab.des:63: Bad weight: -112 (must be >= 0) 10:54:42 Really? :) 10:54:49 Did you rebuild 10:54:52 yes 10:56:26 ..and I've no idea what the heck it's referring to. 10:56:58 I'm rebuilding, will check in a moment 10:59:30 Zaba: I totally don't get that error message 10:59:46 Does throwing out your saves/des dir fix it 11:00:37 hrm 11:01:19 no, it persists 11:01:32 Interesting 11:01:48 What's your bison version 11:01:54 2.4.2 11:02:06 That's pretty recent 11:03:04 levcomp.tab.cc is identical to the prebuilt one 11:03:14 well, apart from the bison version #define 11:03:19 and such 11:03:30 hrm. 11:03:38 and I've no local changes that could affect that. 11:04:55 And the error message references lab.des:63? 11:04:58 That's pretty weird 11:05:39 -!- elliptic has joined ##crawl-dev 11:05:47 and where does it systematically get -112 from 11:05:54 Can you start crawl with ./crawl -dump-maps 2>err and pastebin the Lua dumped 11:06:49 it's plain ridiculous: http://sprunge.us/KUiF 11:07:40 the map has 'WEIGHT: 20' here 11:08:11 Mmm 11:08:30 Is your .lex.cc the same as in prebuilt 11:08:41 yes 11:09:20 * greensnark is perplexed. 11:09:50 Oh, your bug is real 11:10:03 Fortunately you triggered it with a negative junk value 11:10:05 Here's mine :P 11:10:31 http://sprunge.us/RIjU 11:10:39 oh nice. 11:10:41 I must have screwed up something, let me check 11:10:50 hm, valgrind didn't tell me about anything of interest 11:11:06 Oh, I get it 11:11:16 I changed all number parsing to use doubles 11:11:27 Which is unfortunate when I then stick them into formats intended for ints 11:11:33 ah, I think I see how that can cause problems 11:11:44 And even more unfortunate with warnings disabled for those files 11:11:51 Let's re-enable warnings 11:13:39 I love how every little change forces a full recompile 11:13:50 yeah, that's cute 11:15:13 * greensnark needs a faster laptop. 11:23:18 03greensnark * r2cc57573dbdc 10/crawl-ref/source/makefile: Re-enable warnings for level-compiler lexer and parser. 11:23:19 03greensnark * re506600e331a 10/crawl-ref/source/ (files.cc main.cc stuff.cc stuff.h): Call print_error_screen from end to void duplication. 11:23:22 03greensnark * r79ded102a182 10/crawl-ref/source/ (maps.h util/levcomp.ypp): Fix doubles being passed to %d format strings in level compiler (Zaba). 11:23:23 03greensnark * r11fc1d9aebae 10/crawl-ref/source/prebuilt/levcomp.tab.cc: Update canned level compiler with WEIGHT fix. 11:24:28 aaaand now it's recmopile time again. 11:24:43 luckily, it only takes one and a half minute here 11:27:15 03greensnark * ree51c36eb9ab 10/crawl-ref/source/ (prebuilt/levcomp.tab.cc util/levcomp.ypp): Fix std::string being passed directly to %s format in level compiler. 11:34:08 hrm. 11:34:31 Something else broken? :p 11:35:01 nope 11:35:09 just had to remove saves/des a few times, and now it works :P 11:37:19 03greensnark * reef6327f18a7 10/crawl-ref/source/ (makefile util/Makefile): Tell flex not to generate yyunput, use -Wno-sign-compare to kill level-compiler warnings with recent flex and bison. 11:37:50 03zaba * r4ad043f1511b 10/crawl-ref/source/dat/des/builder/layout.des: layout_roguey: Simplify make_inner_room. 11:37:52 03zaba * r4791b70c6c74 10/crawl-ref/source/dungeon.cc: Try to build more rooms in _build_postvault_level regardless of the vault's ORIENT. 11:37:55 03zaba * r797f9c5bcd3f 10/crawl-ref/source/dat/des/branches/abyss.des: Merge abyss_entry and abyss_entry_guaranteed. 11:37:58 03zaba * r216b87e806be 10/crawl-ref/source/dat/des/branches/pan.des: Merge pan_entry and pan_entry_guaranteed. 11:38:17 there, now we are also exercising the new feature :P 11:39:26 Coool 11:39:40 regarding its use for picking layouts, there are still open questions, viz. is it really worth the ugliness wrt the C++ layouts? 11:40:28 and.. primary vault building would have to be converted into seconday vault building beforehand, I guess. 11:40:45 one can fix _plan_4 so it never generates a room that would be cut in half by the Dis:7 vault :P 11:41:46 but _build_rooms has more problems: Not only it gets ugly dead ends, it can also get inadvertly disconnected zones as a result of vault placement. 11:42:13 that becomes much more than a nuisance on branch bottoms 11:42:48 or indeed single-level branches, such as Blade, where, due to the nature of its vault, the issue also manifests itself quite often 11:44:07 of course, _technically_ there is nothing that prevents such problems from arising with seconday vaults, apart from.. well.. the nature of seconday vaults that we have. 11:44:44 at least the guaranteed primary vaults, and maybe all that have ORIENT but not encompass or float, make assumptions regarding the level layout 11:47:38 03greensnark * rd84ce67adf91 10/crawl-ref/source/stuff.cc: Call print_error_screen only for actual errors (oops). 11:47:40 03chriscampbell89 * r8868fa58c897 10/crawl-ref/source/ (spl-data.h zap-data.h): Improve some Earth spells 11:49:34 -!- Lollipop_ has joined ##crawl-dev 11:52:29 -!- Lollipop has quit [Ping timeout: 260 seconds] 11:59:04 03chriscampbell89 * r5497f9eaba1e 10/crawl-ref/source/spl-data.h: Necromancy level tweaks 11:59:07 03chriscampbell89 * rdbd6454ca150 10/crawl-ref/source/book-data.h: Necromancy book tweaks 11:59:08 03chriscampbell89 * rf4e646046e9b 10/crawl-ref/source/spl-summoning.cc: Reduce Haunt duration 12:03:17 03chriscampbell89 * r587abec32692 10/crawl-ref/source/ (book-data.h spl-data.h): Rebalance Tloc spell levels 12:26:34 MarvinPA: your changes? 12:38:46 -!- paxed has quit [Ping timeout: 276 seconds] 12:39:36 -!- eith has quit [Ping timeout: 255 seconds] 12:39:43 -!- paxed has joined ##crawl-dev 12:39:48 casmith789: yep :) 12:40:03 well, most of them were other people's suggestions on the wiki 12:40:23 Monsters which shouldn't travel at normal speed in shallow water (https://crawl.develz.org/mantis/view.php?id=3035) by casmith789 12:48:20 further proposals: a radical one: let's give Lehudib's Crystal Spear a range of 9 12:48:35 sounds liek fun 12:48:40 it has been reduced from a staple to a gimmick spell no one ever uses 12:48:51 what's its range now? 12:48:52 and for its level it would be still relatively weak 12:49:06 4 12:49:12 make it at least 6... 12:49:44 I'm thinking of going the whole hog rather bumping it just a notch. 12:50:16 well, it is rather powerful, isn't it? 12:50:54 and Iron Shot could take the shotgun targetting that was proposed for Magma but with a twist of mostly ignoring EV but being very weak vs AC (hundreds of tiny pellets) 12:51:44 i was tempted to include a range increase for LCS in the earth magic patch, yeah 12:51:44 LCS is single target, doing damage weaker than what spells just one level higher do in a large radius 12:52:03