00:02:16 -!- upsy has joined ##crawl-dev 00:22:44 -!- TGWi has quit [Read error: Connection reset by peer] 00:37:15 Autotravel to stairs fails when the nearest staircase is under an exclusion (https://crawl.develz.org/mantis/view.php?id=2911) by smock 00:47:45 -!- valrus has joined ##crawl-dev 00:50:40 -!- Textmode has quit [Ping timeout: 245 seconds] 00:52:16 -!- valrus has quit [Ping timeout: 255 seconds] 01:13:20 -!- ahpla has quit [Changing host] 01:13:20 -!- ahpla has joined ##crawl-dev 01:35:35 -!- Paroid has quit [Ping timeout: 265 seconds] 02:08:47 goooood morning. 02:20:20 moin 02:37:29 -!- casmith789 has quit [Remote host closed the connection] 02:56:11 -!- syllogism has joined ##crawl-dev 03:00:20 -!- Icelos has joined ##crawl-dev 03:01:01 -!- Icelos has left ##crawl-dev 03:01:21 03galehar * rb6fa7e9dabc5 10/crawl-ref/source/ (player.cc transform.cc transform.h): Grey draconians can walk in deep water, but they don't actually swim. 04:01:17 -!- monky has quit [Quit: hello] 04:03:59 -!- eith has joined ##crawl-dev 04:37:28 good morning 04:37:45 Napkin: meow! 04:37:49 kilobyte :) 04:38:04 rebase failed tonight, just fixed that 04:38:10 but, got a compile error: 04:38:16 dgl-message.cc:18: error: aggregate ‘stat mfilestat’ has incomplete type and cannot be defined 04:38:19 dgl-message.cc: In function ‘void check_messages()’: 04:38:22 dgl-message.cc:138: error: aggregate ‘stat st’ has incomplete type and cannot be defined 04:38:25 dgl-message.cc:139: error: invalid use of incomplete type ‘struct stat’ 04:38:27 dgl-message.cc:18: error: forward declaration of ‘struct stat’ 04:38:30 --- 04:38:45 yay 04:38:49 (rebase failed because there is no longer need to adjust permissions for public and private files 04:38:52 ) 04:39:29 happens :) 04:39:33 !coffee 04:39:33 * Henzell hands Napkin a pot of caf au lait, brewed by Sigmund. 04:39:36 !coffee kilobyte 04:39:37 * Henzell hands kilobyte a cup of irish coffee, brewed by Nikola. 04:40:17 the rebase is a good thing -- since Crawl no longer tried to chmod anything by itself, you don't need to disable that anymore 04:41:24 about dgl-message.cc: er, what? 04:42:16 -!- Mu_ has joined ##crawl-dev 04:42:32 the rebase failed in the patch for AppHdr.h, because it's no longer required 04:42:39 I changed three calls of lock_file_handle() -- all of them used a constant from and now take a bool 04:43:12 according to the rebase fcntl.h is no longer included, I think 04:43:18 If I saw that correctly 04:43:18 I see 04:43:23 or did you move it elsewhere? 04:43:51 sorry, just got up and didn't pay too close attention :-$ 04:46:35 #include fixes it... let's wait for the rest of the compilation 04:46:51 cool :) 04:47:38 when was cdo last updated? I forgot how to use "git show" so can't check :P 04:47:59 03kilobyte * r79ee111e399b 10/crawl-ref/source/dgl-message.cc: Fix DGL compilation. 04:48:00 on thursday morning (UTC), syllogism 04:48:20 check the changelog when in dgl 04:48:40 oo 04:49:01 or the version string 04:49:07 each version has its own view changelog option 04:49:14 ah so the ponderousness change isn't in yet :( 04:49:15 should I try again, kilobyte? 04:49:30 of course 04:49:36 it's probably being added to cdo while we speak, syllogism ;) 04:49:47 on it 04:49:52 yay 04:50:01 it will be in several build failures later 04:52:10 nah, i bet on no more build failures :) 04:53:10 Unstable branch on CDO updated to: 0.8.0-a0-3560-g79ee111 (31.14) 04:53:31 good job, kilobyte :) 04:53:54 hey, who screwed it in the first place?! 04:54:16 depends on how you define screwed :> 04:54:33 you rid me of a patch - I think that's great :-P 05:06:32 -!- Icelos has joined ##crawl-dev 05:26:19 napking <3 05:26:23 What patch did we kill from CDO? 05:26:38 the one to adjust permissions for shared files :) 05:27:33 Nice :) 05:28:27 yes :) 05:29:37 sup guys 05:30:30 eronarn around? 05:32:57 not likely 05:34:16 on his side of the pond, he's most likely still sleeping. 05:34:21 !seen dpeg 05:34:22 I last saw dpeg at Tue Nov 30 21:48:25 2010 UTC (3d 13h 45m 57s ago) quitting with message Quit: leaving. 05:35:18 Hrm, I could use our Dear Leader for an arbitrary but visible choice. 05:35:41 max player's name length: in some places, it's 10, 14, 29, 30 or infinity 05:35:43 he mentioned wanting to try fixing armour :O 05:35:53 soon 05:36:05 Vandal: Eronarn has massesof grandiose plans. 05:36:08 this is why he is not a developer ;) 05:36:15 crawl would implode, basically. 05:36:18 nice 05:36:42 it can barely stand my 4529279823498 half-done projects already 05:37:00 what I don't understand currently is why armour makes it harder to cast spells AND attack things in melee 05:37:01 kilobyte: nor my featuer mimics + holyhaul. 05:37:21 as if the first penalty wasn't huge enough 05:37:22 Vandal: because the defense boni are such that it needs these negatives? 05:37:47 without it, there would be never any reason to wear scale 05:37:57 apparently not any more? 05:38:13 What wouldn't be a reason to wear scale? 05:38:31 the attack penalties? 05:39:12 but then, I've heard of people who found Dragon King's on a high-level MDFi and died in a lousy scale because they "didn't have good enough Armour skill yet" 05:40:05 I don't see why you don't just make Str more mandatory for armour 05:40:19 Take away the massive penalties and just don't let the player wear the armor without a minimum str 05:40:26 penalties to attack, I mean 05:41:02 the penalties are ok, I think... the problem IMO lies in armour giving lousy protection 05:41:15 It is both 05:41:26 Why on earth would you give up spellcasting and attacking? 05:41:27 the 0.6 pendulum swinged HARD and moved hardly a notch in 0.7 05:41:35 It doesn't make any sense at all 05:42:00 Make casting spells even harder then if you want there to be some massive penalty 05:42:18 I don't see how it is justified though when EV is so incredibly good and doesn't restrict any actions at all 05:42:39 because you can get a massive chunk of HP takeen out regardless of how good your EV is? 05:42:46 It is just silly that as soon as you put on a shield and some armor as a new char you are like a turtle on your back 05:43:02 -!- eith has quit [Ping timeout: 250 seconds] 05:43:21 Have you ever put on a shield and some armour in real life? 05:43:26 is this real life? 05:43:31 -!- Radoslav has joined ##crawl-dev 05:43:44 I have never spammed Magic Dart for 5 dungeon levels in real life 05:43:48 but it sure works well in Crawl 05:44:41 Eh, this is a pointless conversation. You seem to reject the basic notions behind the principles. 05:44:45 * due goes away. 05:45:01 The basic notion is that because armour and shields are unwieldy in real life, they must also be in the game? 05:45:47 Or are you suggesting these aren't problems because it would be better if AC gave more protection than EV? 05:46:08 I don't see how that will fix anything 05:46:30 It will just be like it was before 05:51:20 vandal, i believe your style of arguing ("doesn't make any sense", "it's just silly", "is this real life") won't get you anywhere. 05:51:29 You're probably right 05:51:49 I'm just confused which basic notions I am rejecting 05:52:44 then sit down, and write an email to crd with arguments (leaving out those kind of evaluations i just mentioned) 05:53:05 Well I thought it was already acknowledged that there is something wrong with Armour 05:53:19 There is a very good writeup in the wiki 05:54:54 Is it still a goal to make stats matter more? 05:55:12 besides int 06:10:42 -!- st_ has quit [] 06:15:31 -!- st_ has joined ##crawl-dev 06:19:32 03kilobyte * r59ff4a95deac 10/crawl-ref/source/ (externs.h ghost.cc ng-input.cc): Fix two off-by-one bugs for player's name length. 06:19:33 03kilobyte * r281db821fc2a 10/crawl-ref/source/ (startup.cc tags.cc): Assert-crash rather than corrupt saves. 06:19:37 03kilobyte * rb74532d2ae3a 10/crawl-ref/source/ (libutil.cc libutil.h): Add stubs for strwidth(). 06:19:48 03kilobyte * r2cb71e782391 10/crawl-ref/source/ (ghost.cc ng-input.cc startup.cc): Restrict the player's name length properly in all(?) cases. 06:27:54 !tell dpeg What should be the max length of player name? We had limits of 10, 14, 29, 30 and infinity, leading to corrupted saves. I fixed the bugs (I think), it's time to decide between 14 and 30. The former allows reasonable screen layouts, the latter is the current limit for local play. 06:27:54 kilobyte: OK, I'll let dpeg know. 06:43:59 -!- paxed has quit [Ping timeout: 265 seconds] 06:45:42 -!- paxed has joined ##crawl-dev 06:51:54 hi depax. 07:02:45 kilobyte, you there? 07:03:22 Fangorn_: meow? 07:03:57 hi...that checkin to packag.cc wrt #define DO_FSYNC is breaking compilation under Mac OS X 07:04:01 package.cc 07:07:26 all it does is call fdatasync() 07:07:35 are you going to tell me Mac lacks that? 07:07:45 BSD has it 07:10:46 google is telling me mac's don't have it....i'm looking for a solution 07:11:23 it appears that indeed that's the case... 07:11:29 we can use fsync() instead 07:11:55 just an unnecessary seek+write more, you can survive that 07:12:36 that seems to be the solution i see too 07:13:16 i'll create a patch for mac in syscalls.* to that effect 07:19:43 03tigen 07stone_soup-0.7 * r10931a1aea41 10/crawl-ref/source/libunix.cc: fix compile error with setting escdelay with new ncurses 07:19:51 03kilobyte 07stone_soup-0.7 * r1f05fee7164d 10/crawl-ref/source/ (AppHdr.h libunix.cc makefile): Use the system locale rather than hardcoded en_US.UTF-8 (twb). 07:20:02 03kilobyte 07stone_soup-0.7 * rec71d2a370f4 10/crawl-ref/source/tags.cc: Fix save corruption for too long player names. 07:21:26 I see reports that Mac 10.3 doesn't have it but 10.6 does 07:26:33 oh? 07:30:47 -!- Radoslav_ has joined ##crawl-dev 07:32:26 i see that SYS_fdatasync is defined in syscall.h, but don't sync fdatasync defined in unistd.h 07:32:51 -!- Radoslav has quit [Ping timeout: 240 seconds] 07:33:31 -!- Radoslav__ has joined ##crawl-dev 07:34:24 fsync() is declared in unistd.h, but not fdatasync() 07:36:05 -!- Radoslav_ has quit [Ping timeout: 245 seconds] 07:39:35 When falling through a shaft or using an escape hatch, you are always positioned in the top-left corner (https://crawl.develz.org/mantis/view.php?id=2912) by doublep 08:00:32 -!- valrus has joined ##crawl-dev 08:04:44 fdatasync() missing from unistd.h (https://crawl.develz.org/mantis/view.php?id=2913) by Fangorn 08:04:50 -!- valrus has quit [Ping timeout: 245 seconds] 08:05:20 patch attached to issue in mantis 08:21:35 03matt.tadd * refea59875442 10/crawl-ref/source/ (syscalls.cc syscalls.h): fdatasync() implementation for Mac OS X 08:21:38 03kilobyte * rcd43e9c8c86d 10/crawl-ref/source/syscalls.cc: A comment about fdatasync() on new fruits. 08:40:35 heh, neat. at least under MacOS X 10.6 SDK, I can just add an extern "C" declare for fdatasync() in syscalls.h, and it'll use the symbol you mentioned in libc. No idea if that's a no-op or not though. 08:41:16 that fdatasync() change mentioned earlier broke compilation on freebsd for me 08:41:19 http://pastebin.com/KrRXe05A 08:42:51 i think freebsd suffers from the same problem as Mac OS X. 08:47:19 i applied your patch and changed TARGET_OS_MACOSX to TARGET_OS_FREEBSD and it compiles now 09:02:28 -!- MarvinPA has joined ##crawl-dev 09:05:35 ahpla, i added a new patch to the issue that'll fix both mac os x and free bsd now 09:06:15 ahpla: oh, I didn't know. The BSD box I sshed into had it. 09:06:24 not even sure which of the BSDs it is :p 09:07:11 perhaps a link test might be better 09:08:37 -!- Icelos has quit [Quit: Ex-Chat] 09:20:13 03zaba * rc441d869ebb8 10/crawl-ref/source/dungeon.cc: Move the call to dgn_place_stone_stairs from _roguey_level to _builder_normal. 09:20:15 03zaba * r20f1a82bcd53 10/crawl-ref/source/dungeon.cc: Remove some code that is never reached. 09:20:21 03zaba * rf851d9fd8791 10/crawl-ref/source/dungeon.cc: Remove erroneous parameters from some _build_seconday_vault calls. 09:20:22 03zaba * r72abb07dea66 10/crawl-ref/source/dungeon.cc: Remove a no longer used 'stair' argument from _place_portal_vault. 09:20:26 03zaba * r5506f36894b0 10/crawl-ref/source/monster.cc: Make the "loading spellbook" message more informative, and use dprf for it. 09:20:27 03zaba * r5b6fda32fd9f 10/crawl-ref/source/abyss.cc: Remove arguments of _abyss_place_map that are the same in all cases. 09:20:30 03zaba * r86290769e6f7 10/crawl-ref/source/dungeon.cc: Don't needlessly use dgn_region::absolute where we already have the coord_defs. 09:20:37 03zaba * r456705e9d539 10/crawl-ref/source/dungeon.cc: Change _plan_4 to take a dgn_region_list* for exclusion zones. 09:20:38 03zaba * rd6b28ddc4416 10/crawl-ref/source/dungeon.cc: Remove some now-unused variables from _build_vault_impl. 09:21:16 Fangorn_: thanks for the new patch 09:21:21 kilobyte: this is 8.1-STABLE fyi 09:22:26 no sweat 09:26:35 -!- TGWi has joined ##crawl-dev 09:59:11 Vandal: i'm around now 10:22:37 hm. 10:23:08 is there any substantial difference between the odds of 138 in 210, and 2 in 3? 10:24:12 .657 versus .667 10:24:16 so, no 10:24:41 where is this? 10:25:21 _plan_4 10:25:25 it's one whole freaking percent of difference. If you change it, the world will end. 10:25:52 it uses that to decide whether it should build just a box, or a room with doors 10:26:22 in a large enough sample, it could be substantial 10:26:41 _plan_4 is only common in Vaults 10:27:01 2/3 of vaults levels are built using it, I reckon 10:27:28 it is occasionally used elsewhere, too 10:28:28 does that count as a large enough sample? 10:28:34 I doubt 10:28:54 in a large enough room, it would change one doored room to a boxed one (or vice versa) 10:31:04 it's not relevant 10:31:46 just change it :| 10:32:12 changed 10:32:39 I shall accumulate some more commits before pushing, though :P 10:33:09 no, this is important 10:33:11 do it ASAP 10:33:32 I don't want the world to end just yet. 10:36:05 it's ridiculous how _plan_main is biased towards _plan_3 so heavily 10:36:26 what's _plan_3? 10:36:31 layout_rooms 10:37:37 Zaba: there's a new layout algorithm, layout_delve. It's nicely parametrized too, which can produce pretty distinct maps. 10:38:10 where can it be found? 10:38:19 too bad by's implementation in lua can take several seconds (!) to run, even though it's an O(n) (n being the area) algorithm 10:38:50 huh, lua's generally faster than that. 10:39:08 roguebasin has a page about it, but the sample code is hosted on some service that requires you to install some Windows spyware 10:39:33 I did do it on a virtual machine and it got immediately pwned :p 10:39:41 fortunately I do have the source 10:39:49 hm, there's a branch in the crawl repo for it, too 10:40:01 by the way, where did by go? 10:40:18 it's in pretty weird style of C, but I can rewrite it -- I just never got around to it 10:40:40 the branch holds that lua implementation -- it's slow and buggier than hell 10:40:57 but I don't really understand why it could be that slow 10:41:19 does appending to a list in Lua have some terrible complexity? 10:41:26 it shouldn't.. 10:41:34 well.. depending on how it's done, I guess. I'm no lua expert. 10:41:51 I have several idea about fancy variants of the algorithm 10:41:56 the only reason why I support layouts in lua is because that reduces the amount of clutter in dungeon.cc 10:42:05 but if I do it in C++, would you scream? 10:42:08 no 10:42:11 in a separate file of course 10:42:19 as long as it's nicely separated.. it's fine 10:42:56 the biggest problem with current C++ layouts is not that they're in C++, but that they are so badly intertwined with things that you can't easily convert them 10:43:22 for example: Lua layouts don't support exclusions. That's needed for _plan_4 and _build_rooms. 10:43:31 actually, I'd love to get support for that 10:43:34 speaking of new layouts... generating a layout in a triangle and transforming it from polar coordinates looks interesting 10:43:42 even though a bit gimmicky 10:43:46 got a link to that article on roguebasin you mentioned? 10:44:07 kilobyte, making lua layouts support taking a list of regions of the level that they shouldn't touch would be -awesome-. 10:44:27 http://roguebasin.roguelikedevelopment.org/index.php?title=Delving_a_connected_cavern 10:44:40 of course, it wouldn't immediately make them all support making use of that information, but it'd be a start 10:44:49 -!- Radoslav_ has joined ##crawl-dev 10:45:42 actually, it'd be a nightmare to make such a feature useful 10:46:10 because the only case where that is used is _build_vault_impl... which is also used for placing lua layouts. 10:46:11 -!- Radoslav__ has quit [Ping timeout: 255 seconds] 10:47:41 Zaba: about layout_misc: it has two versions with hardly any similarity 10:47:58 I know.. 10:48:12 one is a bunch of corridors, and another is something named spotty 10:48:39 'spotty' comes from spotty_level being applied to it 10:49:18 level generation is so cool 10:49:33 I think the chance of it being picked by _plan_main is 1/108 or so 10:50:01 for the bunch of corridors variant, all are horizontal or vertical. I guess giving it a chance to have corridors at arbitrary angle -- using Bresenham's without diagonals. 10:50:44 strictly diagonal corridors suck, diagonal at not exactly 45⁰ can be good 10:52:06 kilobyte, I agree that it needs to be brought into a better shape. If you've got ideas, surely go ahead 10:52:19 kilobyte, if it turns out interesting, the chances of its use could probably be raised, too 10:52:59 no, actually, Bresenham's with diagonals looks good too 10:53:02 http://sprunge.us/XQgB 10:54:56 the lines are respectively: 45⁰, Bresenham's w/o diagonals at a non-cardinal angle, Bresenham's with diagonals on a similar angle 10:55:29 a layout builder using just lines (without subsequent spotty_level) would need more than that, though 10:55:58 yeah, the non-spotty layout_misc is not very good 10:56:10 ok at its infinitessimal chance, though 10:59:15 kilobyte, bringing it to shape would essentially be writing a new layout generator 10:59:31 since right now it's practically as good as missing anyway 11:02:43 myself, I don't really have enough clue to do that, but I'm all for more variety in the dungeon 12:04:50 03zaba * r48d5f58bb2fe 10/crawl-ref/source/dungeon.cc: Simplify the odds in _plan_4. 12:04:51 03zaba * ra44c70fe6334 10/crawl-ref/source/dungeon.cc: Use dprf instead of #ifdef'd mprf(MSGCH_DIAGNOSTICS, ...) in some more places. 12:17:11 Zaba: Along with cleanup, pushing dungeon layout to Lua also makes it possible to have vaults that use those layouts. 12:17:34 <3 any effort to clean that code up and keep moving that effort forward. 12:17:51 enne, how can vaults use those layouts? 12:18:56 I mean, what exactly are you referring to 12:21:37 Ok, so not only would you convert it to lua, but you'd also have to modify it to be subvaults instead of just writing to grd directly. 12:21:57 ah 12:22:00 You're right that layout.des couldn't be applied to a vault as currently written. 12:22:06 yeah. 12:22:14 ...but it's a step in the right direction. :) 12:22:25 enne, one thing keeps bothering me about the dungeon builder 12:22:30 (My apologies. I had been thinking about doing that at some point, but forgot that I never did. :P) 12:22:34 enne, it's the way levels that have non-encompass primary vaults are built 12:23:38 How so? 12:23:40 namely, how all of them apart from Dis:7 are built the same way 12:24:18 it partially defeats the purpose of having many different layout generators 12:24:49 (Oh, actually, I lied. layout.des doesn't write to grd anymore. You just need to take out the extend to gxm, gym logic and they'd work just fine.) 12:25:11 (Clearly I need more coffee this morning to remember what I had or hadn't done years ago.) 12:25:23 because things like branch ends with vaults, and D:1 with entry vault, and occasional other levels with primary vaults are all constructed by _build_rooms 12:25:30 which is quite boring and even ugly 12:25:54 You want almost branch-specific weights and code? 12:26:06 Dis:7 uses _plan_4, and no other level generator is suitable for such cases 12:26:21 enne, no.. It bothers me how only _two_ level generators can work on levels with primary vaults. 12:26:48 What do you mean by "primary vault"? 12:27:05 And why don't other level generators work? I'm not familiar with why that is. 12:27:10 something placed by _build_primary_vault 12:27:59 enne, no other level generators can be told to avoid generating things on specific regions of the level 12:28:14 That seems very fixable! 12:28:39 they all assume they're working on an empty level; _plan_4 and _build_rooms are the only ones that don't (in case of _plan_4, optionally) 12:29:14 enne, how would one pass a list of dgn_regions to lua layout generators? 12:29:32 -!- monky has joined ##crawl-dev 12:29:59 enne, and I'm not sure how well recursive _build_vault_impl would work, since lua layouts use it, too 12:30:30 of course, it's all fixable... 12:31:35 If it were me, I'd probably use the subvault hammer. 12:31:46 well. 12:31:53 I didn't really look at the subvault code much. 12:32:00 and I try not to do massive rewrites, you see 12:32:17 I can screw things up even on small rewrites, I daren't attempt large ones :P 12:32:28 Subvaults get a rectangle to generate a vault in, but also a mask. 12:32:38 The mask specifies whether or not a given cell will get written back to the main vault. 12:32:44 hm. 12:32:51 This lets you have non-rectangular regions, or alternatively regions you want to not touch. 12:33:38 but just letting them generate things normally and then disregarding a region where the primary vault is won't work well 12:33:46 But you can query it, if you want. 12:33:46 that will look ugly and won't guarantee connectivity 12:34:56 Part of the difficulty here is that you place a vault and it sets some vault flag on those grd cells. 12:35:32 ...but when you're generating a vault (in Lua), you're generating a vault which may be flipped, mirrored, or rotated, so you don't know for sure what vault cell corresponds to what grd cell. 12:35:49 it'd be nice if less things depended on grd cells 12:35:58 Absolutely agreed. 12:36:55 enne, still, all that sounds like massive changes I've got problems wrapping my mind around 12:39:12 So, we already have some Lua code in layout.des that does dungeon generation. 12:39:19 ...it just doesn't work with primary vaults, as you point out, yeah? 12:39:45 enne, neither do other C++ layout generators, aside from _build_rooms and _plan_4. 12:40:24 _roguey_level, spotty_level, _builder_basic, to name a few 12:40:43 enne, another ugliness that keeps bothering me is special room handling 12:40:59 So, a first step might be to modify existing C++ generators to consider the VAULT flag and be smart about it. 12:41:17 If those get pushed to Lua, the same logic will apply. 12:41:47 Just start with one and see how it goes. 12:41:57 What's up with special rooms? 12:42:38 they can be created and placed either in _roguey_level (which prevents it from being converted to lua, among other things) or _builder_normal (which does a lot of other things) 12:42:56 and _roguey_level does a part of _special_room's job on its own, whereas _builder_normal just calls it 12:43:21 if it has been placed by _builder_normal, it's 'hooked up' (by _special_room_hook_up) only in _build_dungeon_level 12:43:35 they use special dungeon features to mark their floor and walls, which is ugly 12:43:57 Special rooms should really be pushed to Lua as subvaults. 12:44:08 enne, the catch with special rooms is that they're of variable size 12:44:11 It'd let vault makers specify a part of a vault as being filled by what's normally in a special room. 12:44:17 Subvaults can handle that. 12:44:26 enne, in case of _roguey_level, it just uses one of its already generated rooms as special 12:44:37 in case of _builder_normal, it randomly constructs a room somewhere 12:44:54 and.. _builder_normal is not even a layout generator. It just invokes layout generators. 12:45:18 dungeon.cc is a mess. 12:45:23 it used to be worse 12:45:24 :> 12:45:32 Indeed. :) 12:46:41 enne, by the way, the aforementioned handling of special rooms in _roguey_level has been fixed by me a while ago. I've no idea how long it's been broken for. 12:47:15 -!- bmh has joined ##crawl-dev 12:47:36 enne, I'm sure it's better this way, though. 12:47:36 -!- bmh has quit [Client Quit] 12:47:46 and whatever refactoring happens to special rooms, such handling should stay 12:58:56 enne, I wonder whether converting special rooms to subvaults is realistic without enormous amount of reworking of other code 13:01:17 I'm not sure. They're already vaults. They just need to not write to grd directly. 13:01:46 enne, what about dungeon.cc:_special_room/_special_room_hook_up? 13:01:50 and things that call them 13:04:19 The wall construction could get pushed to Lua as well. 13:04:23 I'm not sure about the hookup. 13:04:45 There's a lot of global lua_special_room_foo variables that look messy. 13:05:01 enne, that is used to pass special room data to the lua code that fills them 13:05:13 Oh, ok. 13:05:13 -!- Noom has joined ##crawl-dev 13:06:00 enne, alternate way of building them is in _roguey_level, which only calls _special_room for the actual filling 13:06:08 I'm not sure how that'd be handled, either 13:07:21 In general, it still feels like special rooms should just be subvaults. Use the subvault mask to pass which cells should be filled. 13:07:29 Then just apply that vault to the level, like you would any other vault. 13:07:44 If it needs walls, make variant Lua functions that do that. 13:08:37 oh well 13:08:52 I still don't know nearly enough about subvaults to say anything senseful 13:09:16 they seem to be more of a mapdef.cc thing that dungeon.cc thing 13:09:27 I didn't read mapdef.cc all that carefully :P 13:09:55 I should sleep now 13:12:40 They are more of a mapdef thing, but I consider that a positive. :) 13:13:16 Take a look at how Vault:8 works. It's the best subvault example, I think. 13:25:10 -!- valrus has joined ##crawl-dev 13:28:01 -!- casmith789 has joined ##crawl-dev 13:29:28 -!- valrus has quit [Ping timeout: 245 seconds] 13:34:36 -!- Textmode has joined ##crawl-dev 13:43:53 -!- eith has joined ##crawl-dev 13:44:48 -!- eith has left ##crawl-dev 13:52:35 03kilobyte * r1efa36264b85 10/crawl-ref/source/fight.cc: Fix claws when transformed. 13:52:46 03kilobyte * r15e0500f4fe2 10/crawl-ref/source/ (36 files): Get rid of you.attribute[ATTR_TRANSFORMATION]. 14:06:48 click on the splash screen can select a menu option (https://crawl.develz.org/mantis/view.php?id=2914) by KiloByte 14:07:38 -!- Pseudonut has joined ##crawl-dev 15:20:27 03kilobyte * r3f72f361209f 10/crawl-ref/source/tilepick.cc: Make formatting slightly more consistent. 15:20:30 03kilobyte * r634543b568a8 10/crawl-ref/source/tilemcache.cc: Let spriggan berserkers show their weapons. 15:20:31 03kilobyte * r199d9155ccbb 10/crawl-ref/source/tilepick-p.cc: Fix a warning. 15:23:56 I think I found a bug with the shopping list 15:24:17 there are two books of dreams in a single shop, and when I try to shoppinglist/unshoppinglist one of them it does the same to the other too 15:24:32 known? 15:32:32 -!- dpeg has joined ##crawl-dev 15:32:34 Evenings 15:32:34 dpeg: You have 2 messages. Use !messages to read them. 15:34:43 !messages 15:34:43 (1/2) Cryp71c said (3d 4h 43m 17s ago): poke poke, I poke you. 15:34:47 !messages 15:34:47 (1/1) kilobyte said (9h 6m 53s ago): What should be the max length of player name? We had limits of 10, 14, 29, 30 and infinity, leading to corrupted saves. I fixed the bugs (I think), it's time to decide between 14 and 30. The former allows reasonable screen layouts, the latter is the current limit for local play. 15:34:59 kilobyte: around? 15:35:12 I vote for 14. Only nobles need longer names, and we don't like them. 15:35:40 !tell Cryp71c Hi, I am here. Been busy etc. You can always write a mail btw :) 15:35:41 dpeg: OK, I'll let Cryp71c know. 15:42:48 "sir Dpeg" is still only 8 letters long :p 15:44:06 :) 15:44:10 14 ok? 15:44:56 no one said anything when I asked before 15:45:14 14 is better for a number of reasons, imo 15:45:26 also ghosts and (later) player gods 15:46:03 If I recall correclty, the limit on cao is 10, but the limit on cdo is 20; how much of cdo has >14-char names? 15:47:03 (and how would this affect those people--I'd personally be a slight bit irked with a truncated name, but maybe I'm petty) 15:47:22 hmm... then where it is 14 then? 15:47:36 You have this religiousness indicated by your nick :) 15:48:20 I think "Linley Henzell" and "MrMisterMonkey" are 14 15:48:20 kilobyte: it could very well be that 14 is a number I pulled out off my hat... but hey, it is a good number for the purpose. 15:48:56 sure 15:49:52 it of course should be 14 character cells, not 14 bytes 15:50:02 yes 15:50:08 for unicode 15:50:27 currently we sometimes allow non-ASCII sometimes we don't 15:50:40 but in any case it often gets corrupted later 15:51:40 I guess we want to block them eg. on CDO... but blocking them in character notes or such would be weird 15:52:01 I noticed that most Germans use English in notes but most Finns do not 15:52:24 that is an interesting observation 15:53:21 not very scientific, just what I remember from glancing at a few morgues 15:54:46 -!- Pseudonut has quit [Quit: Computer went to sleep] 16:06:45 another issue: monster berserkers somehow are dangerous summoners. That feels really weird and unthematic: I'd imagine them as solitary tough warriors rather than casters. 16:07:18 because they spawn berserkers instead of going berserk? 16:07:29 kilobyte: player berserkers are dangerous summoners 16:08:18 you call the bros like ten or so times in a game 16:08:35 kilobyte: yes, monsters could do it with a very very low probability 16:08:49 or do it only once, and then berserk afterwards 16:08:52 (at least I do, some people get away with being less conservative) 16:09:12 kilobyte: the player is the strongest threat any given monster has ever lived to see :P 16:11:12 Eronarn: we don't expect monsters to play perfectly 16:11:59 make them do it once and then berserk 16:12:36 make them only do it after berserk has worn off 16:18:08 kilobyte: so there should be a thematic solution without kicking the effect completely 16:19:03 yeah... I wonder which one is the best 16:19:56 st_'s is pretty cool too 16:20:09 encouraging the player to fight them while they're berserked is neat 16:22:43 both are fine 16:22:54 no brothers if the monster didn't berserk already 16:22:59 and lower chances anyway 16:24:48 could also make them only do it if they have no allies in los 16:27:21 they make tough enemies in Zigs, but mostly because of a bug that makes the brothers spawn next to you instead of next to the caster 16:27:55 and with that bug fixed, most of the difficulty would be gone 16:28:25 is haunt a monster spell? 16:28:30 but then, Zigs are not something that is really relevant to balancing monsters 16:28:32 yes 16:28:33 it is 16:28:38 oh 16:30:01 spider levels should be rarer than average in zigs by the way 16:30:34 they're mostly just an annoyance 16:30:53 st_: do ghost moths work? 16:31:21 they're supposed to drain all your mana with smite targetting 16:31:34 yes, but tornado means int loss isn't really an issue (but loltornado) 16:31:34 that doesn't really make spiders difficult, though, does it? 16:31:41 mp loss* 16:31:53 also apparently you can use invis to avoid mp drain 16:32:35 a zig level that's really easy, but you have to do it without MP is just annoying 16:33:00 rPois is necessary though, red wasps all over the place 16:41:35 I think that they do provide an interesting situation, but it's tiresome when you get a bunch of them close together 16:42:51 imo that's true of any zig type 16:43:40 I don't think you can avoid that unless you come up with 27 different types of levels (and using each exactly once). 16:50:28 dpeg: could just make it so if a zig level appears, you won't get it again for (1 / a few) levels 16:50:57 doesn't make it tremendously less random, but does mean you won't run into 3 snake levels in a row or whatever 16:51:51 yes 17:00:48 03kilobyte * rcf132f9d1c9c 10/crawl-ref/source/ (5 files): Fix several bugs wrt monster Brothers in Arms. 17:01:28 -!- valrus has joined ##crawl-dev 17:05:43 -!- valrus has quit [Ping timeout: 245 seconds] 17:07:18 -!- valrus has joined ##crawl-dev 17:08:27 -!- valrus has quit [Remote host closed the connection] 17:09:22 -!- valrus has joined ##crawl-dev 17:09:47 -!- valrus has quit [Remote host closed the connection] 17:11:12 could the cao site get a learndb list with bad_ideas filtered out? There's 2600 of the things and it gets in the way of searching for anything 17:11:30 plain "ideas" has 300, that could go too 17:13:11 -!- Zaba has quit [Ping timeout: 240 seconds] 17:17:14 -!- Zaba has joined ##crawl-dev 17:17:29 delete them from the learndb entirely tbh 17:24:36 -!- Zaba has quit [Ping timeout: 250 seconds] 17:25:39 -!- Paroid has joined ##crawl-dev 17:33:14 -!- Zaba has joined ##crawl-dev 17:46:00 03j-p-e-g * rc30785472394 10/crawl-ref/ (6 files in 3 dirs): Re-add the old ghoul tiles, thus fixing #2815. 17:57:02 -!- valrus has joined ##crawl-dev 18:05:06 -!- dpeg has quit [Quit: Lost terminal] 18:15:22 -!- Pseudonut has joined ##crawl-dev 18:18:13 Small information leak with invis + clouds (https://crawl.develz.org/mantis/view.php?id=2915) by sorear 18:51:50 -!- syllogism has quit [] 19:45:49 -!- keithburgun has joined ##crawl-dev 19:46:14 hi all 19:46:43 hey, I'm the lead designer of 100 Rogues (an iPhone RL) and I'd really like to make an item that references Crawl 19:46:43 hello, keithburgun. 19:46:53 anyone have any ideas/suggestions for what it should be? 19:47:12 an eveningstar or a snozzcumber 19:47:22 snozzcumber you say 19:47:27 ??snozzcumber 19:47:27 snozzcumber[1/1]: The Snozzcumber resembles a cucumber but differs in various respects. The snozzcumber varies from nine to twelve feet long. It is striped lengthwise black and white, with wart-like growths all over its surface. It has a repulsive flavor. 19:47:37 I dont think Ive encountered one of those. Is there a picture someone online? 19:47:46 I think it's made up for crawl 19:47:50 it's a food item 19:47:53 but I mean, a pic of it in crawl? 19:47:54 how it looks? 19:47:59 % 19:48:06 heh, I mean in Tiles? 19:48:06 isn't it from discworld or something? 19:48:11 ... oh, right, we have tiles now 19:48:32 -!- Mu_ has quit [Quit: Defecator, may everything turn out okay so that you can leave this place.] 19:48:44 oh my god - wait 19:48:52 Snozzcumber appears to be from Roald Dahl's book the BFG 19:48:55 according to a google search 19:49:09 so then I'd be referencing two awesome things =] 19:49:13 keithburgun: somehing referring Sigmund would also be cool :0 19:49:15 use a scythe 19:49:15 *:) 19:49:16 bah 19:49:21 oh yeah Sigmund 19:49:23 thats a better idea actually 19:49:44 could have a rune of Zot somewhere 19:50:04 a scythe-shaped rune 19:50:05 True, yeah. Maybe it's just like, a useless item though in 100 Rogues? There's like tons of them all over the place =] 19:50:13 a rune-shaped scythe 19:50:15 A bit like the joke in Fallout 2 where there are water chips all over the place 19:50:23 sigmunds scytche of distortion hits you for over 9000 damage 19:50:37 Felirx, I think that's the item we shall have 19:51:36 maybe someday I can convince Ploog to add in a "Crusader's Helmet" that looks like this: http://redkingsdream.com/wp-content/uploads/2010/11/100-Rogues-crusader.jpg 19:51:57 I have heard of 100 Rogues but have bee nunable to play. 19:52:14 * due makes sign of the cross against iPhones. 19:52:35 yeah, i don't blame you sir. We're actually porting it to Android now, though 19:52:40 oooh. 19:52:44 Good, I have two android phones. :D 20:57:06 :O 21:01:49 Yes, I want to try this too 21:09:20 -!- Keith_burgun has joined ##crawl-dev 21:09:20 -!- keithburgun has quit [Read error: Connection reset by peer] 21:27:59 -!- Cryp71c has joined ##crawl-dev 21:31:29 -!- Noom has quit [Ping timeout: 255 seconds] 22:05:56 Keith_burgun: 22:05:58 ??knife of accuracy 22:05:59 knife of accuracy[1/1]: +27,-1 thin dagger. Yes, it's a dagger not a knife - stab away. Note that weapon to-hit enchantments don't affect stabbing. 22:06:02 it even has a +27 in it 22:06:27 knice of accuracy isn't a very crawly thing, is it? 22:06:34 Well yeah it just doesn't sound memorable 22:06:38 unless the reference is supposed to be really really subtle 22:06:51 I'd like people to think, "hey, that's from crawl!" 22:07:00 xom 22:07:02 The scythe is a good idea, but... art wise it's gonna be tough 22:07:03 then he'd be in three roguelikes 22:07:15 Ha, XOM is a funny idea, dunno how we'd put him in though 22:07:21 The Rune of Zot would really be the easiest 22:08:11 xom is also in two of linley's shmups 22:08:22 doesn't xom only have two shmups 22:08:26 linley 22:08:31 (linley is the same as xom) 22:08:57 there's at least excellent bifurcation, and I recall yet another 22:08:58 if there's a third I'd like to see it, white butterfly was fun 22:09:01 oh, ok 22:27:23 -!- Cryp71c has quit [Quit: Leaving] 22:31:25 I didnt know linley did smups now. Anyone have a link? 22:31:38 google "white butterfly" 22:34:53 http://www.allegro.cc/members/linley.henzell/projects I should try transdimensional hellspider sometime 22:35:11 crawl needs a transdimensional hellspider monster 22:35:36 I think someone suggested paying homage to transdimesnional hellspider on the Spiders' Nest wiki page 22:36:20 you know what else we need? 22:36:21 helllions 22:41:55 hell lions 22:42:07 it was pretty obvious 22:43:45 i'm serious 22:43:50 add lions and then add helllions 22:44:02 I actually read that as "hellions" 22:44:10 I thought I was making the dumb joke first 22:44:18 hellfelidae 22:44:21 when have you ever known me to lag behind on something dumb 22:45:02 also: distortion squid 22:45:11 you're ahead of your time 22:45:16 (in terms of dumb stuff) 22:45:22 it is like a kraken, but it sticks its tentacles through gateways it opens up all around you 22:45:31 shows up in the abyss only 22:45:40 making abyss nifty would be nice 22:46:15 (it's not like malign gateway because it moves and makes new portals, it's more like a monster that you can't outrun directly) 22:46:31 abyss needs more 50's slang 22:47:33 abyss needs less dungeon monsters 22:47:40 and more ones limited to it 22:49:58 abyss needs more excitement and less turns expected to get an exit or rune 22:50:03 so boring 22:50:32 yeah it's a bit overused right now 22:51:13 i think it would be fine the way it is if it only showed up a handful of times in a 15 rune win 22:51:13 distinguishing between early-game and extended-game abyssal would be nice, perhaps by ramping up the difficulty after passing up exits or something 22:51:38 I put a few dumb ideas on the wiki page, but I'd prefer them over the status quo 22:51:44 I might go back and tidy them up sometime 23:15:31 -!- valrus has quit [Remote host closed the connection] 23:55:07 Sonja blinked after degenerating into a pulsating lump of flesh (https://crawl.develz.org/mantis/view.php?id=2916) by smock