00:45:27 -!- Moriasc has quit [Ping timeout: 255 seconds] 00:50:35 -!- Twinge has quit [Ping timeout: 240 seconds] 01:11:55 -!- Twinge has joined ##crawl-dev 01:25:36 !tell greensnark moin moin! the links on http://crawl.akrasiac.org/scoring-stable/overview.html are broken: they all point to /scoring/ instead of /scoring-stable/ 01:25:37 Napkin: OK, I'll let greensnark know. 01:29:52 !tell greensnark also, links to morgues of 0.7.1 point to /stable/ instead of /0.7/ - morges from other versions might too, didn't thoroughly check 01:29:52 Napkin: OK, I'll let greensnark know. 01:51:11 !lg * 0.7 win cdo 01:51:12 greensnark: You have 2 messages. Use !messages to read them. 01:51:12 94. Surr the Swashbuckler (L25 KoHe), worshipper of Jiyva, escaped with the Orb and 3 runes on 2010-12-24, with 1606913 points after 60130 turns and 5:12:40. 02:26:47 -!- galehar has joined ##crawl-dev 02:27:04 hi 02:36:28 hiii 02:36:37 hi Napkin! 02:44:13 Crashing on startup in tiles during the first load-up (https://crawl.develz.org/mantis/view.php?id=3094) by nicolae 02:45:16 :( 02:46:59 Good morning! 02:47:40 Hi-No! 02:47:51 (Yes, I just made a portmanteaux of your name and "Hi".) 02:48:00 That's awesome :D 02:48:26 "Ei" pronounces exactly like "Hi" without the 'h'. 02:49:06 I'll now sing it like the 7 dwarves: "Hi-No, Hi-No, it's off to crawl we go!"? :D 02:49:57 I am speechless. 02:50:18 Excelent! 02:50:20 * due hugs. 02:50:30 hugs hugs! ^_^ 02:50:36 How do you think we should represent Liquefaction in tiles? 02:50:44 I tried doing a transparent ripple overlay but it is ugly. 02:52:11 Moin! :) 02:54:35 Hi napkin! 02:55:18 due: Perhaps felirx will code up some cool distortion graphical effect for that :) 02:55:37 (he wants to do an effects engine/stuff for spells and so on) 02:57:50 C:\Crawl\dat/des/portals/sewer.des:281: Bad item name: 'potion of levitationanyscroll' 02:58:04 who gave commit rights to this dpeg guy? ;) 02:58:58 ha! 02:59:03 yes, he really needs a lecture 02:59:23 also comitted those temple entries without an actual entry to the temple ;) 03:00:10 due: check out my idea at the bottom: https://crawl.develz.org/mantis/view.php?id=3064 03:01:57 lol @ coolio's lab rats - tey're fantastic!! 03:03:52 moin 03:04:34 Hi Zaba! 03:04:43 what's up 03:05:57 Keskitalo, regarding no_shop_gen: There was some reason why I didn't remove it immediately, but I forgot... 03:06:32 Keskitalo: Yes; we're going to have "armourd laboratory rat", etc :D 03:09:57 hehe 03:10:32 Zaba: by its description in syntax.txt it feels like it would veto shops from the **whole** level, not just the vault, but i might misinterpret that 03:11:38 Keskitalo, you are misinterpreting that, yes 03:11:45 Okay, good. :) 03:11:54 it marks the vault with MMT_NO_SHOP, which used to be checked by the old shop placement code 03:11:58 !tell MarvinPA Thanks for all the patches and bug reports! 03:11:59 Keskitalo: OK, I'll let MarvinPA know. 03:13:06 Keskitalo: I'm not sure coolio realises, but he just came up with the flavour implementation list :) 03:13:15 due: Aesome :) 03:13:35 I love when e.g. new tiles inspire new features (as long as it's good of course!) 03:13:38 I need to investigate ghost structure first. 03:13:48 should rename that 03:13:53 Yes, agree. 03:13:59 "complex_monster" structure, eperhaps ): 03:14:05 yeah, something like that 03:14:15 I love the word "dynamic" =) 03:14:28 "dynamic_monster" 03:14:31 maybe silly :) 03:15:00 due: Sounds like the lab rats would be a lot like ugly things 03:15:17 which also means they should be carefully made to be not too much like them 03:15:45 where do laboratory rats come from in a bloody _dungeon_, anyway? 03:15:46 due: Also good call replacing grey rats with porcupines 03:15:56 Zaba: Why, from izard labs of course 03:15:57 or are we getting an underground research facility branch next 03:17:37 place_spec_shop checks MMT_NO_SHOP. Hmm.. So, in theory, wizmode shop placement would not work in no_shop_gen vaults. 03:17:42 that's about the only effect it has now, I think... 03:17:47 due: please ready my query :D 03:17:59 I guess it can go 03:19:01 Zaba: hehe :) 03:19:25 I love that we can now create some lovely shop vaults for random shop placement 03:19:41 I really like it how there are just now two places that call place_spec_shop, and how one of them is from wizmode code 03:19:44 (right? just have to be careful about chances)? 03:19:57 Keskitalo, umm.. not with the current implementation, but it'd be easy to make that possible 03:20:10 Keskitalo, you'd need to pick the shop vault by tag rather than name 03:20:40 just need to change how dgn.place_maps is called slightly, and the vault definition itself 03:20:56 Napkin: <3. 03:21:56 Keskitalo: Thanks. If there's one thing I hate about Lair, it's the "OMG RATS!!!". 03:22:12 yeah, they're pretty damn boring 03:22:20 due: I had an idea from the rat king discussion 03:22:25 Go on! 03:22:35 due: but not sure if it's good since it duplicates slime creatures a bit 03:22:55 the idea was that one square could contain several rats 03:23:00 well, that's about it 03:23:15 Hm. 03:23:22 Yeah, trying to avoid duplication. 03:23:24 they could all hit you, but you culd only hit one at a time 03:23:35 bah, don't 03:23:36 I was reading the user-submitted holy monster ideas... 03:23:45 that's a bit different in the sense that with slime creatures, there's still only one monster 03:24:02 A lot of them are basically "zombies, but holy" "x, but holy" or slightly different versions of what we already have. 03:24:46 Keskitalo, shop vaults was one thing I didn't really think about at all, somehow. 03:24:56 also, an actual rat king idea: rat kings contain several rats, and have hydra-like attacking. severing attacks tear rats separate from the rat king, but they're not killed, just "spawn" next to the rat king 03:25:16 Now that's interesting! 03:25:18 Zaba: Heh, well is that easy(ish) or natural to implement? 03:25:25 -!- Textmode has quit [Ping timeout: 276 seconds] 03:25:29 Keskitalo, should be, yeah. 03:25:41 as I said, just need to change it to pick the vault by tag rather than the name 03:25:47 yeah, great! 03:26:07 I always imagined rat kings working like kraken but with rats instead of tentacles, rat tail segments instead of tentacle segments, and a big knotted mess instead of a kraken 03:26:27 maybe a bit less coordinated too 03:26:30 monky: That is cool. 03:26:31 hehe, that'd be cool but otoh i think rats are too smallfor it to work like that 03:26:38 yeah :( 03:26:39 but maybe it's better to suspend disbelief 03:26:58 because the idea of rats in the rat king being a little confused about direction is a *great* one 03:27:05 Keskitalo: Anyway: once lab rats are in, sewer lab rats! 03:27:15 (btw required reading: The Amazing Maurice and his Educated Rodents) 03:27:20 due: Ooh, nice! 03:27:45 btw, we really need monster ideas that are hard for casters and archers 03:27:56 its much much easier to come up with cool ideas that are hard for melee characters 03:28:27 One idea for the thunder snail was that it would retaliate ranged attacks with lightning 03:28:32 I think that was from OG17? 03:28:47 Interesting... 03:28:52 due: btw 03:28:53 A bit like the whelk from Final Fantasy vI? 03:28:59 due: Yes, exactly 03:29:09 due: Also, could you give agate snails reflection? 03:29:17 Hm, should be able to! 03:29:46 due: Finally, what do you think of my "spit goo" idea for slugs and snails? I feel it's a great one, but I think I've gotten like zero dev feedback about it. 03:29:49 :/ 03:30:02 Some non-dev comments though <3 03:30:20 Where was the idea? 03:30:27 I have't read the wiki much recently, I confess. 03:30:35 In the "Tweaks to existing monsters" page 03:30:37 Ahhh 03:30:40 I'll recap here 03:30:49 Might be a good idea to air links etc with some regularity in here? 03:31:08 Because the slugs and snails are slow, they get a ranged attack that ties their prey into place, so they can catch up 03:31:23 basically, it's like a throwing net 03:31:42 I have this "held and restricted" page that lists ideas about different types of such attacks 03:31:53 so that each are a little different and more dangerous to some char builds than to others 03:32:05 oooh, nice! 03:32:15 for instance, spider webs require STR to pull away from, so that they're harder for casters (as Spider is designed to be) 03:32:42 I forget what the special thing for the goo spit was 03:32:57 probably that if you teleport/blink, you are still covered in it (unlike with nets) 03:33:39 adding different types would need a bit of code generalization, i did some research for it a while a go but it was too big a project for me ultimately, i didn't get *really* started 03:33:52 i did send jpeg some email about it so maybe my thoughts are transferable 03:34:44 !tell jpeg a **big** thank you for updating the report page!! 03:34:45 Keskitalo: OK, I'll let jpeg know. 03:34:46 Leda's Liquefaction was a ... simple spell to start with :/ 03:35:01 and now there's a lot of special cases, ? :) 03:35:04 Then I ended up writing a function to find the centre of an area etc. I did wonder how I'd got into that. 03:35:15 No, just had to write a hunk of new code to steal kilobyte's swirly pattern. 03:35:17 aha 03:35:22 hehe 03:36:19 argh, managed to close browser with 1000 tabs by accident 03:36:19 due, is there going to be a wizlab for Leda 03:37:10 Zaba: Don't know. 03:37:27 what. Where's the 'absolutely yes'? 03:37:38 Zaba: I don't currently have a good idea for one yet. 03:37:51 Zaba: Of course, whirlpools of sand everywhere done in Lua == good idea. 03:38:00 "Liquefaction marker", clearly. 03:39:01 03zaba * rd41c96c44fca 10/crawl-ref/source/dat/des/portals/sewer.des: Correct whitespace errors in sewer.des 03:39:05 03zaba * r871dc77d5f77 10/crawl-ref/ (7 files in 5 dirs): Remove the no_shop_gen vault tag. 03:39:06 03zaba * r593e65b6bb6e 10/crawl-ref/source/dat/des/portals/sewer.des: Correct a typo in sewer.des. 03:39:34 -!- syllogism has joined ##crawl-dev 03:45:39 The cocoon new monster idea could be cool for a summoning spell 03:45:51 instead of ready monsters, it summon cocoons that hatch after a delay 03:45:59 (and I'm thinking a monster spell, primarily) 03:49:00 So long as it's not hatching into spiders I'm happy :) 03:49:10 djnrempel's idea about the weapon display, similar to spell display is GREAT 03:49:22 could hatch into wasps 03:49:23 due: No, that's be eggs :> 03:49:26 mmm wasps 03:49:42 perhaps Foxfire card could be nerfed with cocoon delay 03:49:44 Keskitalo: What do you think of the new lurking horror idea? 03:49:44 :) 03:49:55 due: Can you give the elevator pitch? 03:50:00 Elevator pitch? 03:50:54 describe it shortly 03:51:18 Sure... describe what? :)) 03:51:22 idea: some physical conjuration could be severing and cut hydra heads 03:51:30 due: lurking horror idea? 03:51:33 Sorry, sorry. 03:51:36 I meant the tile! 03:51:40 hmm. 03:51:50 oh! haven't seen it, can you link? 03:51:58 https://crawl.develz.org/mantis/file_download.php?file_id=1352&type=bug 03:52:14 (The lurking horror is a cloud of darkness with spotty eyes... it hides around corners and then sneaks upp on your and tries to devour you.) 03:52:15 also i don't know the monster! 03:52:16 thanks 03:52:25 It's one of the abyssal horrors Mu and I have been working on 03:52:34 ah ywah, I saw that and I thought that probably wasn't what you were going for 03:52:35 The tiles are being developed cocurrently with the idea, so each can influence the other. 03:52:40 it's hella cute though 03:52:49 There's an addition o that image! 03:52:55 Loook below that one by three squares. 03:52:56 it looks playful 03:53:23 It reminds me of the demon from Sandman... can't remember which one though. 03:55:59 due: og that 03:56:00 yes azrael 03:56:17 due: the old version 03:56:20 even with the old version 03:56:32 i immediately thought that's a monster tile that'd be cool if it was animated :D 03:56:34 Azrael! 03:56:45 but not sure if its good to aniate some monsters and not others 03:57:10 That new version... 03:57:12 due: The new one is better, though less well drawn. the black should be blacker too 03:57:15 no wait.. 03:57:20 It's too wel defined 03:57:23 needs more contrast I think.. 03:57:27 Yeah! 03:57:37 http://pousse.rapiere.free.fr/tome/tiles/DO/tome-domonsterstiles.htm 03:57:46 I confess to wanting something like the "void elemental" on that page under Elementals. 03:57:52 (We can't use those, they're CC-licensed.) 04:00:42 hmm. 04:00:51 something about gnoll camp maps bothers me but I'm not sure what it is 04:03:15 !tell pointless_ If I didn't already say so, <3 <3 <3 for looking at the herd AI 04:03:15 Keskitalo: OK, I'll let pointless_ know. 04:03:40 Keskitalo: Oh, I had a brilliant idea for that the other day... 04:03:51 Zaba: generation or the vaults themselves? 04:03:53 But if pontless is doing stuff with it I will leave him to it :D 04:04:10 the vaults themselves 04:04:19 But my idea was basically: when a yak spots you, it finds all other members of its band and sets their foe as you and their behaviour as BEH_SEEK. 04:04:26 oh good - if you can give feedback I can tweak them :) 04:04:49 03j-p-e-g * r8428bacddce4 10/crawl-ref/source/dat/des/portals/sewer.des: Add weight to the sewer vault's potion of levitation. 04:05:08 i tried to give them a distinct look by the stone wall badding around the rock walls 04:05:17 do you think it looks good? 04:05:33 due: something like that would be good 04:05:47 due: also would work excellent for wolves - the howling, you know 04:05:58 of course wolves would need to be pack animals first :) 04:06:06 Keskitalo, they're too.. static, not enough randomness. 04:06:24 I think 04:06:35 hmm.. do you think they need more variants, or randomness in the vaults themselves? 04:06:42 randomness in the vaults themselves 04:06:42 one thing is that they will be rare in normal games 04:06:54 so a player might not notice 04:07:17 but yeah, adding a little more fuzziness can't be bad 04:07:27 still, they're rather bland as they are 04:07:29 also 04:07:40 Zaba: yeah, I think that's the unfortunate case too 04:07:42 especially the smaller ones 04:08:15 Keskitalo, perhaps rock walls could be mixed with stone ones, with odds biased towards rock? 04:08:27 Zaba: Sounds good to me 04:08:37 if you'd like to tweak, and add more vaults 04:08:40 instead of the current static badding 04:08:43 and remove bad ones 04:08:50 that would be great! i have little time :) 04:08:59 Keskitalo, also, the largest one seems too large. Too much empty space and all. 04:09:05 I have to go now, too, brb 04:09:12 k, see you! 04:09:40 due: Have you looked at serial vaults? 04:10:22 Yes, they're interesting! 04:10:38 due, Zaba, greensnark, what do you think of using serial vaults for city layouts, including the idea for Vaults? 04:10:56 I don't think it would work quite as well... 04:10:57 due you worked on the Vaults thing so you can probably say something here 04:11:07 of course, serial vautls being randomly placed right now 04:11:11 I haven't looked at it clearly enough, but I thin kthe placement would be too random. 04:11:16 would create not so citylike maps 04:11:20 Yeah. 04:11:30 but!! it'd be great if you could control that by figuring out a "street layout" first 04:11:41 then using that plan to place the vaults 04:11:43 well, placing in a city-like layout wouldn't be... too difficult 04:11:47 shouldn't be too hard, yeah 04:11:48 using dgn.place_city_map instead 04:11:56 it couldn't be stateless 04:12:11 but it wouldn't be too hard to store the values (eek) statically. 04:12:16 and the rooms don't need to be handade vaults, can use algorithms to create them 04:12:20 the same ones that are in place now 04:12:29 just put them into des scripts you know 04:12:37 or would that be too cpu intensive? 04:13:09 It's only being done once so I don't think so? 04:13:39 well, it can be done a lot on the server.. but yeah, probably not too bad? 04:14:14 I don't think the load increase would be that dramatic. 04:14:18 It's done a lot, but it's not done often enough. 04:14:29 alright, good, good 04:14:40 Change level -> (x time for the builder) -> explore level -> change level -> ... 04:15:04 So, even if it was slightly more expensive, it would be relatively deferred expense unless someone decided to try doing builder statistics. :) 04:15:18 the fun thing for me to thinkg about is that.. with the city layout using serial vaults, you basically start with an empty layout before adding the vaults 04:15:24 that's just funny to me 04:15:34 Why? 04:15:42 of course, you actually need the street plan, but you don't actually place any walls or so 04:16:01 due: I can't explain ... 04:16:13 it just feels so simple that it's funny 04:16:30 also i haven't slept properly for a couple of weeks =) 04:16:42 I see :) 04:16:48 New parent syndrome, etc? 04:17:09 yeah 04:17:23 I finished shedu :) 04:17:30 And ophan, actually... 04:17:47 \o/ 04:18:04 due: What did you think of my "lone shedu" unique idea? =) 04:18:11 if you looked at that patch item link 04:18:18 i mean we got two shedu tiles 04:18:25 could use the other for a new unique 04:18:39 "lone shedu" in itself is already a great backstory 04:19:12 anyawy, i need to go now.. new parent thing! cheers! 04:19:28 @??shedu 04:19:28 shedu (16H) | Speed: 19 | HD: 13 | Health: 91-130 | AC/EV: 2/10 | Damage: 19, 23 | Flags: 08holy, fly, !sil | Res: 06magic(52), 03poison | XP: 3106 | Sp: resurrection, heal other. 04:19:43 @??ophan 04:19:45 ophan (05G) | 04UNFINISHED | Speed: 10 | HD: 15 | Health: 74-117 | AC/EV: 10/10 | Flags: 08holy, see invisible, lev, !sil | Res: 13magic(immune), asphyx | XP: 1852 | Sp: b.fire (3d24), holy flames. 04:20:05 are the stats ok as well? 04:20:21 I think so. 04:20:29 (well, hard to tell that without knowing their target place) 04:20:35 Yeah. 04:20:41 However, they're fully implemented now. 04:20:48 Need to nuke the unfinished flag for opahn. 04:21:07 Will work on phoenixes next... not quite sure how to do that outside of creating a new marker which is painful but I suppose not too painful. 04:21:08 I added it since they 1. spawned in zotdef, 2. had buggy corpses. 04:21:22 Ophan? 04:21:32 Huh. 04:21:35 yeah, but it seems to be an easy fix 04:21:37 I didn't test corpses, I confess. 04:24:56 Bah, I want to fix up trees in swamp with tiles but I'm keeping in mind what Enne said... okay, time for some serious hacking! 04:27:10 * due touches enum.h :( 04:43:22 -!- Mu_ has joined ##crawl-dev 04:45:26 due: do you really lack a bigger machine somewhere? 04:45:56 kilobyte: Actually, yeah. but also lack of committment. 04:46:32 kilobyte: I don't have any other *nix machines at home. 04:46:35 distcc can be bad if your dev machine has a slow link -- preprocessed sources are many megabytes big, so pushing them to and fro would take longer than the actual compilation 04:46:48 Yeah. 04:46:56 If I had some other *nix machines to work with, I'd switch. 04:47:05 The other option is to push my changes to my work machine and then compile and test there. 04:47:24 that's what I do 04:47:31 I don't think I can distcc with my two work machines either though, one of them is 64bit and the other is 32bit. 04:47:34 But otherwise that woul dbe nice. 04:47:54 everything except tiles, which are unusable over a worse-than-lan connection 04:48:14 Yeah. 04:48:19 I'm hacking tiles at the mintue anyway. :0 04:48:41 my dev machine is 64 bit and the ancient home desktop which just let out the smoke was 32 bit 04:49:18 a proper solution would be cross compiling, but that would require a tiny bit of reading, so I just use a 32 bit chroot 04:51:54 Can someone sanity-check: http://sprunge.us/ZIJV 04:51:59 I'm not sure the C++ is right. 04:52:10 that's unhelpful. 04:52:27 -!- monky has quit [Quit: hello] 04:52:38 http://pastebin.ca/2034098 04:52:57 diff was a bit confusing I think. 04:53:12 Specifically directed at kilobyte: is that the proper way to initialise yourself from another instance of yourself? 04:53:29 "yourself" being...? 04:53:37 say 04:53:45 packed_cell newcell = packed_cell(&old_cell)? 04:55:00 for structs, plain assigment should copy all (most?) fields, and if you do that in such a function, you may do some additional processing as well 04:55:19 Well, I'm just copying over the relevant stuff. 04:55:35 explicitely copying all fields by hand should work, too 04:55:42 Yeah. 04:55:53 The other stuff is all explicitly handled by the initialisation code elsewhere. 04:56:59 why do you use an enum for a single value? Isn't a #define or a const simpler? 05:09:50 galehar: Existing code. 05:11:32 galehar: It works, so I'm not touching that in case I break it :) 05:15:03 Title graphic & logo (https://crawl.develz.org/mantis/view.php?id=3095) by Denzi 05:15:19 ! :D 05:16:25 oh aweeesome. 05:16:30 Denzi is astonishing. 05:17:48 -!- Adeon is now known as MaryPoppins 05:18:06 MaryPoppins: I loved you in that movie! 05:18:50 I'm disturbed 05:20:59 ooh, it's not broken yet! yay! 05:23:41 Hm, trying to watch movie and compile at the same time is unrecommended behaviour. 05:32:52 re 05:32:57 hi Zaba! 05:37:29 using subvaults for Vault layouts is possible 05:37:54 or rather, serial vaults 05:38:12 dgn.place_map can take the desired vault coordinates, so you could calculate the positions of the rooms and place them in lua 05:38:54 of course, the actual calculation might be pretty involved, but nothing that is impossible 05:39:43 "found two bug-filled silvery potions" 05:40:10 oooh, how exciting! It is done! 05:40:24 kilobyte: That's odd. 05:40:39 dear tiles: please don't crash, love, jude. 05:41:45 does hard-locking the host machine from inside a VM count as crash? 05:42:32 Well, it crasshed but becauseof something else. 05:43:13 Weird visual distortions. 05:44:21 -!- cw_ has quit [Remote host closed the connection] 05:44:45 O...kay. That's a bit odd. Why does it do that. 05:45:24 -!- cw_ has joined ##crawl-dev 05:46:05 That solves Issue #1. 05:46:08 Issue #2 is... odder still. 05:46:19 -!- dpeg has joined ##crawl-dev 05:46:36 hi ploog 05:47:30 moin dpeg 05:48:24 Hi 05:48:34 Zaba: awesome stuff on the shop generation 05:48:46 one question: can shops still be generated in Vaults:8? 05:48:52 no 05:49:01 Do we want shops there? 05:49:06 no idea :P 05:49:22 but whoever designs (sub)vaults for vault:8 probably doesn't -expect- them there 05:49:31 probably okay to have no shops in V:8 05:49:39 Zaba: would it be possible to have a shoppy subvault? 05:49:44 is that just a coincidence, or do Vaults:8 produce way more shops than usually? 05:49:45 of course 05:49:58 kilobyte: coincidence 05:50:03 I've only changed generation of random shops 05:50:03 I can't remember a game without any, at least 05:50:05 I didn't notice this in my games 05:50:07 vaults can place shops all the hell they want 05:50:20 I mean, not more but there seems to be always some 05:50:35 kilobyte: I had plenty of shopless Vault ends. 05:50:42 there certainly aren't always shops on v:8 05:50:44 actually so many that the occasional shop sticks out 05:50:45 right, a stroke of luck then 05:51:06 If I get the time, I can make a shop subvault. Would be fun if most of the loot is in shops. 05:51:20 it does feel to me like there are more random shops there than normal, but that probably isn't actually true 05:52:32 dpeg, an interesting possibility that Keskitalo has noticed but I haven't, is creating shop vaults 05:52:46 yeah, would be cool 05:52:48 could be employed for some additional flavour 05:52:54 Zaba: for ordinary game? Sure! 05:53:08 instead of shops randomly strewn all around the levels, they could form a mall 05:53:20 yes 05:53:26 kilobyte, well.. each vault would need to contain a single shop, though 05:53:35 why? 05:53:42 you mean for serial shops vaults? 05:53:45 yes 05:53:49 or just flavour stuff around the shop glyph? 05:54:07 btw, TOME has taken off in the poll 05:54:15 but I doubt they have 400 dedicated players :) 05:54:17 any vault that would be a substitude for the 'shop' vault in dat/des/builder/shops.des would need to contain only one shop 05:54:27 Zaba: ah, I see 05:55:00 Zaba: I want to make some. Do I just name them "shop"??? 05:55:04 no 05:55:43 we need a tag? 05:55:52 yes 05:56:18 wait. 05:56:35 Zaba: I have time pressure right now... Would you file a dummy shop vault? I'd add more right away. 05:57:11 -!- casmith789 has joined ##crawl-dev 05:59:49 03zaba * r5824792c4bc7 10/crawl-ref/source/dat/des/builder/shops.des: Select the shop vault by tag. 05:59:53 dpeg, there.. 06:00:29 cool 06:00:43 I believe it's not selectable by anything aside from 'shop' tag... would be pretty broken if it turns out that I'm wrong. 06:01:06 oh, how do I set the weights? 06:01:17 I want the new shop vaults to be rarer than the standard one 06:01:27 just WEIGHT should work 06:01:29 ah, I see 06:01:32 If it's selecting by tag you should just use WEIGHT. 06:02:06 just make sure you don't add DEPTH/PLACE :P 06:02:35 -!- elliptic has quit [Ping timeout: 240 seconds] 06:03:38 03j-p-e-g * rb0a2d7e72b0b 10/crawl-ref/ (5 files in 3 dirs): New temporary option tile_allow_detached_montab. 06:03:47 Nooooooooooooo. 06:05:26 :) 06:10:10 This sorta works but it doesn't clear. 06:12:13 kilobyte: I'm implementing multi-variant laboratory rats using ghost structure. It's on my TODO :) 06:12:31 I'm not sure coolio realises but they just wrote the list of rat flavours :) 06:13:11 -!- jpeg has joined ##crawl-dev 06:13:28 awesome, I still got my plus 06:13:28 jpeg: You have 1 message. Use !messages to read it. 06:13:34 <3 elly 06:13:37 !messages 06:13:38 (1/1) Keskitalo said (2h 38m 53s ago): a **big** thank you for updating the report page!! 06:13:41 :) 06:14:42 due: how's it going? 06:15:45 moin jpeg 06:15:53 jpeg: I've managed to reimplement it so we can just extend the packed_cell structure instead of requiring new flags. 06:16:01 Zaba: hi! 06:16:24 due: so I've seen, but does it work? 06:16:25 jpeg: However current issue is that the packed_cell isn't refreshed often enough to deal with halos, etc... I don't understand *why* it isn't. I think there's a "check if changed, if not do nothing". 06:18:20 03dpeg * r8de0841eee90 10/crawl-ref/source/dat/des/builder/shops.des: Some shop vaults (Zaba, Eino). 06:18:48 due: hmm... you don't even set the halo in packed_cell 06:18:59 or was that just an example? 06:19:32 jpeg: Ah, that was a mistake. 06:20:51 dpeg: how about extending resistances as boegh's messiah to followers? 06:20:55 ah, got it. 06:21:00 dpeg: see https://crawl.develz.org/tavern/viewtopic.php?f=8&t=134 06:21:16 dpeg, looks good 06:21:43 -!- ortoslon has joined ##crawl-dev 06:21:45 the packed_cell is reused; if you set the halo flag in a cell once, it remains until you set it as false; 06:21:57 the previous code worked because the tile was *always* wiped clean and then the flags were applied. 06:23:44 when i try to run http://crawl.develz.org/trunk/ , i get this error: D:\temp\crawl_tiles-0.8.0-a0-4194\dat/des/portals/sewer.des:281: Bad item name: 'potion of levitationany scroll' 06:23:57 ortoslon: naughty dpeg made a typo :) 06:24:05 another one? 06:24:25 ah, no, cdo is just not _that_ quick in updating :) 06:24:53 ortoslon: don't you get any mails if a bug report of yours gets resolved? 06:25:09 i do 06:25:28 i haven't reported this typo 06:25:55 ah, sorry 06:26:05 could someone update windows builds then? kilobyte? Napkin? 06:26:13 ortoslon: mm'kay 06:26:20 i was checking out two bug reports simultaneously, and got you mixed up 06:26:27 erm, seriously? 06:26:42 updating like every 8 hours now, ortoslon? 06:26:42 jpeg: if you cross your eyes on two bugs, you see a spider! 06:26:46 Napkin: you can't currently even start the game 06:27:04 all dpeg's fault of course 06:27:07 Napkin: yeah, not my fault/OCD 06:27:12 (not like I haven't done that before) 06:27:45 not so many changes since 06:37 this morning :-O 06:27:47 ortoslon: :) what happens if I'm juggling half a dozen? 06:28:03 centipede. duh. 06:28:03 breaking the game horribly is approximately 50% of all the fun and joy that's involved in being a developer 06:28:34 Zaba: yeah, the best part is always committing a big change, forgetting to add a new file, and then going to sleep :D 06:28:35 reporting bugs is 10% of the fun of a trunk player 06:28:45 ortoslon: <3 06:29:09 and losing your game to my bugs, how big a percentage is that? 06:29:20 yelling is fun too 06:29:32 kilobyte: hey, most games are recoverable, aren't they? 06:29:43 1.8gb of builds.. i'll keep the last 5 - is that ok? 06:30:01 you just have to treat asserts as OOD creatures 06:30:06 walk around carefully 06:30:14 ha, players love to bitch 06:30:23 with trunk, we give them a valid reason to do so 06:30:54 Napkin: actually, I think two would be fine... 06:31:04 Napkin: you mean the archived builds? 06:31:30 Are any really necessary? 06:31:35 i guess it mostly matters if some change breaks save compatibility 06:31:38 otherwise not 06:31:45 jpeg, I prefer committing a big change, realizing how broken it really is only once it hits cdo and then quietly reverting it a few days later :P 06:31:48 they are all in one folder - it's just the website that categorizes them as archive 06:31:56 there are legitimate delays, like loading the page just before a rebuild and clicking on the download just after, so the previous one should be kept at least for several minutes if not one hour 06:32:46 Napkin: just keep the previous one 06:33:27 as if anyone except me downloads them... 06:35:18 done 06:35:58 ortoslon: I don't think most trunk players build from source 06:36:09 i was joking 06:37:26 Napkin: thanks! 06:37:33 936 downloads in 4 days, ortoslon 06:37:50 stop your download roboter ;) 06:37:56 woah, so why are we behind ToME again? 06:38:07 probably because people don'tknow about the poll? :) 06:38:09 since yesterday, jpeg 06:38:24 perhaps we should naughtily link to the poll? 06:38:34 on Tuesday tome was just 2 ahead 06:38:58 didn't stone soup win once already? 06:39:07 yes, last year, I think. 06:39:12 a bunch of cheaters, those ToME guys 06:39:14 :) 06:39:17 ;D 06:39:28 2008, just before 0.5 came out 06:39:30 the poll is just "how many people are pointed to the poll", not "how many think X is better" 06:39:39 ... which is, like, *ages* ago 06:39:59 should I vote from all my IPs? :p 06:40:09 actually, it seems it's enough to drop the cookie 06:40:10 CDO has 16 too ;) 06:40:39 nah, we'll probably get an honorary mention either way 06:40:51 Windows development builds on CDO updated to: 0.8.0-a0-4201-g8de0841 06:41:18 thanks, kilobyte 06:41:32 sadly, I no longer have root in a couple of local ISPs I consluted for in the past :p 06:42:00 they found your cronjob? :> 06:43:37 naah, legitimate root not hacked 06:44:05 I didn't abuse it, too -- just joking about massive vote stuffing 06:44:22 hehe 06:44:45 -!- Unrodney has joined ##crawl-dev 06:44:50 but a single person doing a big of skullduggery can sway the poll completely in some way 06:44:55 !asciidreams_vote 06:44:55 bhaak: ToME 4.0.0, 292, Dungeon Crawl Stone Soup, 268, Dwarf Fortress, 168, DoomRL, 93, Rogue Survivor, 68, Angband, 44, Goblin Camp, 34, UnNetHack, 32, POWDER, 31, Brogue, 30 (Votes: 860, Hours left to vote: 20) 06:45:07 you don't have much time left to skew the poll :-) 06:45:10 and I kind of fail to believe that among 200+ voters no one cheats 06:45:14 -!- Unrodney has left ##crawl-dev 06:45:27 s/big/bit/ 06:45:31 Just spam it in ##crawl regularly and stick it up on the CDO download page? 06:45:46 so unrodney is bhaak's golem or what? 06:45:48 kilobyte: there are no rules, so there are no cheaters? 06:46:11 put the link on the download page, yeah 06:46:32 jpeg: yes, he only follows my orders. But I didn't put a piece of paper into his mouth :) 06:46:33 where's the poll again? 06:46:42 What we need is a XSS to vote on the ppoll whenever peopleclick download. :D 06:46:50 http://roguelikedeveloper.blogspot.com/ 06:46:53 due: :) 06:47:06 we need XSS to vote on the poll whenever somebody visits cdo 06:47:14 see, people want to vote, they just don't know where to go 06:47:26 tome winning wouldnt be too bad, darkgod has worked at a crazy pace and alone 06:47:26 wow, Goblin Camp is in the top list? nice 06:47:26 we need XSS to vote whenever someone thinks of visiting 06:47:43 lol, jpeg 06:48:11 because 06:48:14 it's the thought that counts 06:48:35 syllogism: I've never played TOME before 06:48:42 syllogism: I tried to get it running but it was very graphic heavy, IIRC. 06:48:49 tome 4 is yes 06:49:01 jpeg: Oh, I'll send you an email when I do the first push to git hub... I've decided to call it MurderRL :) 06:49:10 i tried tome4, got lost in the overworld and returned to the deadly comfort of crawl 06:49:18 (If you're still interested) 06:49:35 due: great, and I am! 06:49:43 due: MurderRL <3 06:49:47 :D 06:49:52 I tried to play Angband this week, but it looped on creating the first level :) 06:49:52 i enjoyed tomenet for a while - but geez such a messy code 06:50:36 -!- Unrodney has joined ##crawl-dev 06:50:45 wb Unrodney 06:50:49 Unrodney: whoami 06:50:49 bhaak: you are my boss 06:50:54 Unrodney: whoami 06:50:54 due: you are *due_612101448 06:50:56 :( 06:50:56 -!- Unrodney has left ##crawl-dev 06:51:03 612101448? 06:51:11 There are no other dues! 06:51:22 for him, you are just a number 06:51:25 :( 06:51:50 but not to us! 06:51:52 for us, the one and only 06:52:24 :D 06:52:56 I like Denzi's new art. 06:53:19 yeah, the dragon doesn't suck 06:53:31 still kinda partial to the old logo 06:53:38 unlike that @#$%^&*() pixellated title 06:53:46 the "pre-DCSS" one? 06:53:51 kilobyte: wow, that's some praise 06:53:55 indeed, that one is good too 06:54:08 * due bops kilobyte 06:54:25 I do like the logo a lot... though the letters are awfully far apart in "Stone Soup". 06:54:28 might just be me 06:54:34 kilobyte: the one in the first titel we had, with "Stone Soup" on some kind of stony background 06:54:42 *title 06:56:56 jpeg: good idea! 06:57:37 dpeg: I actually *thank* thanked RangerC for that post! <3 06:59:47 -!- casmith789 is now known as casvalk 06:59:57 -!- casvalk is now known as casmith789 07:05:10 bah! laptopdroppeddead. 07:05:50 jpeg: If you are interested in developing I'll probably be licensing it with MIT or something... and I confess, my programming skills are entirely self-taught so expect no great things from me :) 07:06:24 now... are the packed cells ever cleaerd? 07:06:37 due: so are mine, self-taught I mean :p 07:08:08 jpeg: :) sometimes I feel overwhelmed at the comprehension other devs have of programming... they seem to grasp things quickly that I take ages to pick up. 07:08:17 alternately, they've had more experience than me :D 07:08:38 due: uh, looks like they're only used locally inside a for-loop in tilereg-dgn/mon 07:08:46 due: I know the feeling :| 07:09:18 due, you've been writing a roguelike? 07:09:31 Zaba: I am starting to write a murder-mystery roguelike 07:09:34 due: I feel stupid if I don't understand those bit flags right away 07:09:41 -!- Twinge has quit [Ping timeout: 264 seconds] 07:09:52 due, in what language? 07:09:57 Zaba: python :) 07:10:02 due: ha, so you've taken the step from design -> development? 07:10:19 yes, well, I like to dable :) 07:11:19 jpeg: self-taught? Do you mean the "BASIC at the age of 5, assembler at 6" kind? 07:12:07 kilobyte: not quite, I didn't even have a computer until I was 10 :p 07:12:36 kilobyte: more the "learned C by tweaking NH code" variety 07:13:17 didn't know Neverhood was written in C :P 07:13:24 personally, I can't recollect the act of learning C at all. At some point, I've just started knowing it... 07:13:45 Zaba: mots of the c++ I know is by example from DCSS. 07:14:03 I actually have a C++ book. 07:14:31 Oh, I have books galore. 07:14:57 i think it's the only programming language book that I have.. 07:15:02 due: I've got two, but 90% of the stuff I know, I picked up via Stonesoup 07:15:20 I use mine for reference when there are people like kilobyte, sorear, etc, to pester. 07:15:26 I also have the pragmatic programmer, but it's mostly language-agnostic 07:15:52 I haven't used C++ before Crawl either... and I don't use it outside it, too 07:15:56 I have the TeXbook, and that's the extent of remarkable computer-related literature that I possess 07:16:14 so I am not really a C++ pro 07:16:34 I don't really like C++, but with crawl, you gotta deal with it (or lua, but I don't really like lua, either) 07:16:48 I don't mind Lua but it always seems so ... unfinished. 07:16:50 Zaba: so what do you like? 07:16:57 Or at least, it lacks the things I need to use, like... continue :) 07:16:59 jpeg, C and Perl 07:17:18 yeah, lua is meant for simple scripts... if you try to do anything more complex, it ends up badly unreadable 07:17:30 far worse than C++ at least 07:17:39 Zaba: +1 07:17:43 ahhh, worked it out! 07:17:50 kilobyte, in most cases with lua, you can wrap things into separate functions until it looks more or less OK 07:17:52 <3 Perl 07:18:28 Zaba: separate Lua function or external C code? 07:18:34 separate lua functions 07:18:44 ah, okay then 07:18:45 it ends up long and verbose, but at least it doesn't hurt eyes 07:19:00 * galehar loves perl too 07:19:07 let's port Crawl to perl! :) 07:20:39 Perl itself is not really that adorable, but hardly anything can beat cpan 07:21:29 where are the levels stored in the save? 07:21:39 grepping for level or map is not particularly fruitful 07:27:03 -!- Zaba has quit [Ping timeout: 240 seconds] 07:29:34 jpeg: not sure, sorry, I'venever touched that 07:29:40 jpeg: kilobyte would know, I think 07:30:04 hmm... env should do 07:30:30 however, where the heck are map tile definitions parsed? 07:30:59 mapdef.cc I think 07:31:02 I've only found the monster_tile spec 07:31:31 the save file keeps only actual ready levels, not vault definitions 07:31:44 -!- Zaba has joined ##crawl-dev 07:32:02 kilobyte: that's what I'm trying to change 07:32:18 kilobyte: storing tile definitions for features 07:33:09 oh 07:33:23 dgn_tile in l_dgn.cc 07:33:30 er 07:33:33 l_dgntil.cc rather 07:34:21 thanks <3 07:34:34 a special tile-only file, eh? 07:34:43 well, changing that's for later 07:35:12 for now, I'd be happy if vault tiles didn't get trashed whenever new dngn tiles get added 07:37:09 jpeg: thanks for your work on keeping tiles sane when upgrading! 07:37:46 well, I really wanted to add the hellbinder tiles 07:37:54 but doing so meant wrecking my current game 07:38:09 so the change was born of entirely selfish reasons :D 07:38:35 jpeg: do you plan on going all the way and implement kilobyte's idea of using a tile palette? 07:39:04 galehar: not sure yet, I'm still working out where the code needs to go 07:39:05 galehar: it depends if she stores the tiles the current way or using props 07:39:12 galehar: but yeah I think so 07:39:20 the former+pallette would be more effective but more complex as well 07:39:48 wait, what do you mean by palette? 07:40:11 yesterday due said something about storing the index to a vector of strings 07:40:46 yes, something like that 07:41:07 it would be much cleaner too I think. 07:41:27 what's the alternative? storing strings per cell? 07:41:38 that's crazy! 07:41:46 galehar: I agree! :D 07:41:56 yeah 07:42:00 strings per cell is crazy 07:42:10 ... which is why I'm asking what other ways you guys came up with 07:42:26 the tricky part is filling the vector in the first place 07:42:41 probably best to use a temporary hash for that 07:43:01 this is really annoying; I can't see any reason why it shouldn't owrk, but it doesn';t 07:43:21 ... but if we hardly ever have > 20 tiles, that might be overkill 07:46:26 fresh eyes in the morning may help is tihs doesn't fix it :) 07:46:59 yay segfalut :D 07:48:03 that's a nice typo, could be a panlord name 07:48:15 All hail Sigfalut! 07:48:21 :) 07:48:50 idea: Segfalut's wizlab that randomly crashes 07:48:56 HAHAHA. 07:48:59 they do that already! ;) 07:49:05 which reminds me that lone shedu are buggy 07:49:28 aw, they're just lonely 07:49:53 http://sprunge.us/DjZV 07:49:58 I give up for now. I can't see any reason why it doesn't work. 07:50:06 Feel free to investigate! Sleep for the wicked! bye all. 07:50:22 night! 07:50:58 I might give it a try 07:56:32 Ancient zyme tile (https://crawl.develz.org/mantis/view.php?id=3096) by coolio 08:00:44 nice... whatever it is :) 08:05:07 -!- MarvinPA has quit [Ping timeout: 265 seconds] 08:21:41 New title logo (https://crawl.develz.org/mantis/view.php?id=3097) by galehar 08:26:12 -!- ortoslon has quit [Quit: bye] 08:30:27 -!- MarvinPA has joined ##crawl-dev 08:32:56 kilobyte: did you add mid to monster struct? 08:33:41 because sometimes: 08:33:41 ASSERT(m->mid > 0) in 'dbg-scan.cc' at line 334 failed. 08:35:41 galehar: do you have a clue when that happens? 08:36:12 no, but I've got a savefile that does it upon loading 08:37:11 wizmode, climbing up stairs to an unexplored level, plenty of monsters in sight (and even more out of sight as the level is unexplored) 08:37:43 -!- Cryp71c has joined ##crawl-dev 08:40:19 by an unexplored level you mean one that actually exists in the save, or not yet? 08:40:51 (ie, one visited but not explored vs one that hasn't been generated yet) 08:41:36 I didn't went to it at all. Just teleported to orcish mines, and when climbing back to the main dungeon, crashes. 08:41:52 but it saved the level before the crash 08:41:57 so it is reproducible 08:42:19 debugging: p env.mons[39] 08:42:21 type = MONS_ORC, mid = 0 08:42:26 this orc has no id 08:44:37 should I open a mantis and upload the save? 08:49:16 is it a save before or after the corruption? 08:49:27 of course, I'd like to take a look 08:54:09 I think the corruption occured when the level was generated, so I'd say after 08:54:14 I'm opening a bug 09:01:56 crash in debug mode (https://crawl.develz.org/mantis/view.php?id=3098) by galehar 09:22:27 -!- pointless_ has joined ##crawl-dev 09:24:10 I don't suppose anyone has a degree in mathematics and/or experience with statistics and statistical analysis? 09:28:13 Cryp71c: I have a degree but statistics is not my forte. 09:29:03 dpeg, ah, :( trying to find info on breaking down sets of data to identify outliers and confidence values, given an expected distribution 09:29:10 but ##statistics and #math are no help. 09:33:04 Cryp71c: sorry, that's very far from my areas of expertise 09:33:12 :( 09:33:13 sadface 09:49:02 -!- coolio_mc has joined ##crawl-dev 09:56:05 -!- Vandal has quit [Remote host closed the connection] 10:00:39 -!- Iainuki has quit [Quit: Iainuki] 10:26:44 dpeg, ping 10:27:51 Cryp71c: i know a bit 10:28:05 hmmm.. do the staves of conjuration and enchantment still work? checking I/I i don't see any difference with my conjurations or enchantments 10:28:19 the statistical test you will want to use will depend on the nature of the data + what kind of distribution it is expected to have 10:29:05 Napkin: they affect only power, not success 10:29:59 oh, well, nevermind - i thought i checked. at least the staff of enchanment changes power so much, that II can display it 10:30:10 sorry & thanks 10:34:56 Eronarn, some of the distributions will be roughly bell, but other distributions will be set-specific and could be a variety of linearly-shaped (positive or negative) or unpatterned. 10:40:59 Eronarn, any reading topics you can give me on this kind of subject would be appreciated 10:41:05 Cryp71c: you can probably use a simple t-test on the ones approximating a normal distribution 10:41:23 http://en.wikipedia.org/wiki/T-test 10:43:07 the other ones you may want to perform a regression http://en.wikipedia.org/wiki/Statistical_regression#Underlying_assumptions 10:43:56 http://en.wikipedia.org/wiki/File:Linear_regression.png this is what a regression looks like in most statistics programs 10:45:38 you would be looking for these things: 1) are there any dots very far away from the line? 2) does the slope of the line generally make sense, or appear to be an artifact? 3) what is the actual predictive value of the correlation 4) would the distribution appear to fit a non-linear function better 10:46:29 for an example of 4, see http://en.wikipedia.org/wiki/Curve_fitting 10:47:21 -!- elliptic has joined ##crawl-dev 10:51:56 03j-p-e-g * r03c708493e38 10/crawl-ref/source/dat/ (clua/lm_1way.lua des/portals/sewer.des): Use rusty sewer tile when a sewer times out. 10:54:18 -!- galehar has quit [Quit: Page closed] 10:59:28 -!- Twinge has joined ##crawl-dev 11:13:41 Eronarn, so are there standard T-tests that can be used, or do you have to formulate specific T-tests based on the expected distribution of the dataset. 11:13:41 ? 11:15:52 Cryp71c: there are a few standard ones depending on what you are comparing 11:16:22 you want a one-sample because you are comparing some observations to a known value for the population, and then trying to figure out if the observations differ from that population value - yes? 11:17:55 Eronarn, you mean like comparing numerical values of the scores to the mean of all scores in the dataset? 11:21:23 sure, that's one thing you could do with it (the scores of some subset of the population) 11:21:33 Eronarn, also (in mathematics), what would you call a data set which has only 1 subset of values (eg array(12, 45, 87, 34, 98, 100, 34))? One dimensional? 11:21:34 (vs (array(array(12, 34, 23), array(234, 23, 45), etc.) 11:21:37 at least i'm pretty sure, been a while since stats class - can't recall if you need to do anything to process it first 11:22:03 Eronarn, well normal distributions are centered on the mean, aren't they? 11:22:04 that is a one dimensional array in programming, i'm not sure if it is called something in mathematics 11:22:30 Cryp71c: kind of... you can have a normal distribution with skew 11:27:47 Sword of Cerebov autoinscribes to {flame,} (https://crawl.develz.org/mantis/view.php?id=3099) by kurper 11:30:54 03galehar * r47c22999268e 10/crawl-ref/source/tilereg-mon.cc: Update monster tab before drawing it. 11:36:15 Eronarn, is there a set formula as to how much skew is applied to achieve a normal distribution? Do all normal distributions use mean + skew, or only particular situations of normal distributions? 11:38:10 Cryp71c: a true normal distribution has no skew, but see http://en.wikipedia.org/wiki/Skew_normal_distribution 11:40:35 Eronarn, ah, I'm finding that normal distributions are set so that 50% values are below and 50% values are above, so normal distributions use median, not mean? 11:43:32 Cryp71c: an exactly normal distribution has mean = median = mode 11:53:56 -!- Twilight-1 has quit [Read error: Connection reset by peer] 11:54:16 -!- Twilight-1 has joined ##crawl-dev 11:54:18 quite a pity that slime pits have been made so inaccessible :( 12:13:42 -!- Cryp71c has quit [Quit: Leaving] 12:31:43 Narretz (L16 HOMo) (Vault:1) 13:11:59 Napkin: pointless_ will fix that 13:13:10 fix what? 13:13:15 slime? 13:13:15 Slime 13:13:16 :) 13:13:18 \o/ 13:13:26 go, pointless, go :)) 13:13:47 (eventually) (when I have some free time) 13:13:54 I'm one of the few people who liked it as it is, but he outlined a proposal that keeps the good, throws out stuff that most people find bad. Also volunteered to do the changes 13:14:24 awesome 13:14:28 My way to fix Slime would've been to remove the Dig spell, which had fairly little support :) 13:14:47 erm, i think i'm rather on the other side of the fence 13:14:58 hehe :) 13:15:13 it's been ages since i had a char that actually was able to spend more than 10 turns in that slime pit hell 13:15:22 I think I've done the new slime three times, and died once horribly 13:15:30 I really enjoyed it 13:15:50 I thought it restricted moving around, going back to stash, etc in an interesting way 13:16:19 Also, squeezin through thight placec not knowing what the opposition was to be on the otherside and arriving hurt was great 13:17:01 but pointless_ had a convincing argument that doing that was actually a voluntary challenge, not actually necessiated by the game.. i forget the details 13:17:22 so the proper way to play it is to learn Dig and always dig three-wide corridors 13:17:51 spell-caster only dungeon then? :-/ 13:17:57 now, removing the Dig spell as an inexhaustable source of digging would fix that (or "fix" if you don't like my way of thinking) 13:18:15 so you'd have to plan your wand charges 13:18:28 also i proposed that a single charge would dig three-wide corridors 13:18:52 hmm 13:19:06 Napkin: I'm not sure; I did it two times without Dig (got the rune), but I think they were casters 13:19:18 well, i'll keep my opinion to myself and let you designers change something ;) 13:19:57 I think the stuff that I liked about it would be great for a portal vault some day 13:20:11 but as it stands, it's a hell hole... i actually _never_ go inside unless i have only artefact equipment, which happens only when playing okawaru, which doesn't happen anymore ;) 13:20:19 :) 13:20:31 you know, the dungeon is full of spare equipment! 13:21:13 -!- MarvinPA has quit [Ping timeout: 265 seconds] 13:21:14 nah, not going to argue ;-P 13:21:42 but i totally hear that a lot of people just didn't and don't enjoy the changes 13:21:49 so it'll be good 13:22:07 it'll be a great branch after the changes 13:22:10 Keskitalo: stomach portal vault with undiggable acid walls ;) 13:22:30 Eronarn: Yeah! 13:23:06 it'd also be good for a Hell I think 13:24:17 Keskitalo: mnoleg level could have acid walls 13:24:39 cool :) looking forward to seeing acid a bit more "playable" :) 13:24:54 !vault mnoleg 13:24:55 one thing that would help is just plain toning down the damage 13:24:57 Lines pasted to http://paste.scsys.co.uk/77117 13:24:58 it does FAR too much 13:33:23 napkin: the slime walls don't corrode equipment any more, at least 13:33:29 originally they did and that was awful 13:33:42 <- playing stable aka 0.7.1 13:33:49 so i'm still screwed ;) 13:34:18 and you dare come here and complain about that? ;) 13:34:22 oh, right, I guess that was only fixed in trunk 13:35:48 we generally do too few bug fix releases, but that one wasn't really a bug and thus probably wouldn't have been included anyway 13:37:21 if you get abyssed in (0.7.1) slime you get nice slimy walls in abyss 13:37:45 to be fair, that only got fixed in trunk yesterday 13:42:12 hey elliptic, what's your take on an affliction (tentatively titled Malaise) that's no-MP-regen 13:42:17 it came up in brainstorming about mummy curses 13:42:47 so like sickness but for MP (and no stat drain presumably)? 13:46:14 I'm already not fond of how sickness works, and this doesn't sound any more fun if that's all you mean 13:46:59 -!- pointless_ has quit [Ping timeout: 240 seconds] 13:47:44 elliptic: just because of being able to mash 5, or something else? 13:48:34 the fact that it usually just means more time mashing 5, yes 13:49:26 mp regen is too slow for that to be very interesting, I think 13:49:49 i suppose you might get ambushed while resting it away but still 13:49:53 yeah, temporary no-MP-regen seems even less interesting to me 13:51:05 i liked the idea that some monster would make you forget some of your spells temporarily 13:51:18 yeah, that one definitely has legs 13:53:37 Kogan (L18 DEWz) (Vault:1) 13:58:15 I can't believe all these tiles contributions 13:58:19 vaults, too 13:59:32 we've got industrious players :) 14:01:49 just saw the power creep post on the forum *chuckle* 14:02:23 Keskitalo: in regards to digging, unless the player is unlucky they get all the charges they need on wands of digging and don't need the spell at all. 14:05:20 -!- Moriasc has joined ##crawl-dev 14:05:33 removing the spell would be seriously annoying to felids, by the way 14:05:51 oh right 14:05:52 yes 14:05:56 digging is not really required, is it? 14:06:05 well it's a good tactical choice 14:06:11 nice escape item 14:06:58 this is most evident on the orb run but it is also very useful for disconnected + portal vault 14:07:26 sorry, these are different points 14:07:37 if you magic map an area (or ash) and find a portal vault, digging is often the only? way of getting there short of a lucky random tele, as ctele is rare 14:07:58 jpeg: digging of some form can be necessary to get the bone rune, for example 14:08:05 that doesn't sound right 14:08:08 LRD at decent power or shatter would still work, but that's annoying 14:08:15 or iood :) 14:08:30 what about the stone of earth 14:08:33 you could use that 14:08:38 sure, that works also 14:08:42 not exactly common 14:09:00 would be like relying on a RoR 14:09:48 it's much more common than that 14:09:49 well, if the spell got removed, there might be other changes to make up for it 14:09:49 also, if it makes felids into a challenge race, that might be a good thing 14:10:11 I've never liked how stone of earth elementals is an infinite digging tool, by the way 14:10:49 how did you arrive at the spell getting removed anyway? 14:10:57 I don't know why dig is considered being removed, it's not exactly used for scumming, except slime 14:10:58 elliptic: yeah, the misc. items are in need of an overhaul since stonesoup 0.1 :/ 14:11:04 dpeg wants it removed, or at least did at one point 14:11:07 now, removing the Dig spell as an inexhaustable source of digging would fix that (or "fix" if you don't like my way of thinking) 14:11:29 was the comment earlier that brought it up 14:11:30 re: slime walls 14:11:30 ah, I remember 14:11:30 and using it to do slime is like using shatter 14:11:35 I don't exactly disagree either 14:11:38 i don't like it as a spell because you can make permanent changes with a nonexhausting resource, however small 14:11:54 Keskitalo: so shatter, lrd, iood? 14:11:54 you can still achieve the same using Lee's, it's just more tedious 14:12:04 casmith789: that point is so unfair ;) 14:12:05 I hardly ever bother with the spell except on felids because wands of digging are so plentiful 14:12:19 you can't really fix the issue unless you make the walls completely undiggable, which obviously isn't currently possible 14:12:21 Keskitalo: last time I did slime I used shatter, and I've seen someone use iood 14:12:40 casmith789: ouf, cool 14:12:40 so just LRD to go :p 14:12:40 anyway we are changing slime 14:12:50 if slime is being changed I don't really see the need 14:12:53 there was some talk about making the walls be only temporarily removed 14:13:06 casmith789: yeah 14:13:27 digging "permanent" changes are mostly used in pan, and I (and as far as I know) other people who scum pan for a hobby just used wands 14:13:32 so they'd grow back? that's cool! 14:13:41 yes 14:13:41 yeah i like that 14:14:20 one dig doing three-wide openings and them growing back would be very nice 14:14:34 the problem with this is that you could trap yourself 14:15:33 i guess they could grow back to single-cell corridors for the wall damage, at least 14:15:35 i think i proposed that the actual beam path wouldn't grow back 14:15:43 :) 14:15:44 that would work 14:15:54 * Keskitalo highfives jpeg 14:16:03 * jpeg highfives Keskitalo 14:16:27 flavourwise, this wouldn't actually dig three-wide openings 14:16:54 more like, the walls shrinking back from the beam, then returning to their original state 14:17:02 well, it would kind of work, as in practice you could still trap yourself :P 14:17:05 (except for the ones that were actually blasted, of course) 14:17:11 due to the damage being fairly ridiculous if the path is long 14:17:21 heh, yeah 14:17:22 that applies to Crawl in general 14:17:59 you might as well say you trap yourself if you dive into the dungeon because you're more likely to die down there 14:18:21 yes and I guess after doing it once you would be prepared 14:19:16 Keskitalo: why would that be any different, you could just dig another one wide corridor next to it afterwards 14:19:57 oh yeah, that idea did need no Dig spell 14:20:08 casmith789: hmm, good point 14:20:08 and i you did want to dig a permanent corridor, youd have to wait around 14:20:17 which wouold be tedious 14:20:25 yes in fact it would be more tedious than before! 14:20:32 aye! 14:20:32 it would always leave only one tile wide path 14:20:45 how is that even possible to implement 14:21:01 the main points I dig in slime are just finding the > without going next to acid walls 14:21:57 the main problems with this are random # (which this would not stop) or if the other area is completely disconnected where this digging would not stop, it would just lead to more scumming 14:23:40 i think i agree, yeah 14:24:09 well, with the spell at least 14:24:25 spells like you pointed out 14:25:26 fwiw i don't think wall-destroying is a major part of lrd or iood, that could just as well go 14:25:30 haven't seen shatter 14:25:32 hmm 14:25:50 or in lrds case, destroying the walls sometimes.. you sometimes lose a resource 14:25:58 that's quite cool 14:26:22 haha the entire point of lrd is destroying the walls 14:26:25 isn't shatter about shattering walls and features? 14:26:38 and the monsters more of an add-on? 14:26:43 yes 14:26:47 or am I mixing this up with disintegration? 14:27:00 shatter is fun precisely because it destroys all walls 14:27:00 or lrd, yeah? 14:27:14 LRD's flavour is to hit the monsters with wall fragments 14:27:19 casmith789: i'm saying its a visual effect 14:27:19 not that this does any damage at all 14:27:24 disclaimer: I haven't actually tried any of these spells 14:27:39 it's not just a visual effect 14:27:50 it's more, LRD is basically a destroy stone spell 14:28:08 else why would it be level 5, it's about L1 power 14:28:14 well, you know i don't like that in a spell :) 14:28:17 hehe 14:28:44 shatter destroying walls is actually a drawback as it's extremely noisy so you can easily get surrounded 14:28:49 Mennas tile (https://crawl.develz.org/mantis/view.php?id=3100) by coolio 14:28:56 what about L1 spells with level 5 power? 14:28:56 yeah, it's a cool drawback 14:29:00 keskitalo: basically nobody uses LRD for damage 14:29:02 that was the good thing about it 14:29:09 I used shatter to destroy cerebov's walls and then tornado him in the open 14:29:21 boo hiss 14:29:22 I admit this was showing off but its not always a drawback :p 14:29:28 *it's 14:30:16 since the entire school is all about earth, it makes sense that there are occasional side effects on the dungeon 14:30:29 could do with a quicksand spell, oh wait... 14:30:31 :) 14:30:39 we need a gravity flux spell 14:30:46 and a stalactite-generating spell 14:30:47 jpeg: i don't disagree about that, but i think the effects on gameplay are slightly bad 14:30:52 I dont 14:31:06 you know what that means? 14:31:12 buuut i go on and on this because it breaks this new slime thing which i like 14:31:15 we need to get rid of Earth spells 14:31:25 haha, it would make EE better :p 14:31:28 it wouldn't be that badly missed a school 14:31:33 it's so awful right now 14:31:45 well it's just been improved, give it a chance :p 14:31:55 let's wait for pointless changes to slime before raising a fuss about this 14:31:59 (earth would really benefit from having more synergy with other schools... there's no earth/sum, earth/nec, earth/tloc, earth/ench, etc.) 14:32:18 jpeg: yeah, i'm just chattering idly 14:32:27 there's earth/tmig 14:32:36 Keskitalo: if you look at what these "break the game" earth spells are used for so far, there is (afaict) basically no abuse in main game 14:32:36 most schools dont have much synergy 14:32:44 what no pulling bones from the grounds? no summoning earth worms? no heating the earth to create lava? 14:32:56 Keskitalo: it's basically only used for fun stunts involving stone walls and sprint :) 14:33:18 these glass-encased weapon vaults are annoying 14:33:22 syllogism: fire has more overlap with different schools than earth does i think 14:33:22 replace dig with Summon boring beetle! 14:33:26 probably the main use of lrd 14:33:31 casmith789: yeah, i know, it just blocks a cool feature from being added (but very few people actually like the cool feature, so i lose :) ) 14:33:45 which is the cool feature? :) 14:33:50 syllogism: sumon boring beetles would be such an awesome monster spell. 14:34:00 casmith789: acid walls, I think 14:34:01 casmith789: the new Slie :P 14:34:02 jpeg: the lack of ways for non-EEs to get through stone is pretty awful, yes 14:34:10 the new slime? 14:34:11 there's... does IOOD go through stone? 14:34:14 pretty sure that was added 14:34:25 telecontrol, and ??? not sure if there is even anything else 14:34:26 iood does not go through stone 14:46:19 pity really, i wonder if it could go through stone/metal at high damage (i.e. long range and high spell power) or if that would be too powerful 14:49:10 a lot of vaults rely on stone being difficult to penetrate 14:50:28 i certainly think lrd should do competitive daage :P 15:00:56 i remember using it in crypt a bit, it's nice when it instantly kills a skeletal warrior or bone dragon 15:03:57 Mu_: nice, but bad design 15:04:09 it should do higher damage to them, but not instakill them 15:05:03 LRD on fleshy things definitely should do more damage, on the other hand 15:05:29 getting your bones shattered is kind of a big deal, even if you have a lot of flesh covering those bones 15:06:05 (though this makes me want an Earth/Nec spell where you make enemies into marionettes!) 15:07:46 plants have synergy with Earth, so there could be a plant-summoning Earth/Summ spell 15:07:48 oh, random import from ##crawl: apparently orbs of fire are not insubstantial; bug, or are they orbs of *condensed* fire? i could see it going either way 15:09:32 jpeg: that would seem a bit of a reach for 'any old plant', but i could totally see some magical plant made out of earthy materials - like one that is an endless tangle of razored edges, which fires leaves at enemies 15:10:21 one thing i have wanted for a while is a plant that starts at some point, and grows on walls (only) very rapidly - so it can follow you but with limitations 15:10:24 not any old plant, no, that would be boring 15:10:37 a moving creeper - cool! 15:11:05 "quicksilver kudzu" 15:16:55 < Daekdroom> FR: Sky Beast Form 15:17:35 we definitely still need this, but one question I'd have: should it flicker randomly, or would it be more interesting if it *glowed* you randomly? so you were invisible the whole time, but either grey glow or not 15:17:53 why do we need that 15:19:04 there's no air/tmut spells and it's an interesting idea for one 15:25:05 -!- cw_ has quit [Remote host closed the connection] 15:27:37 -!- galehar has joined ##crawl-dev 15:28:01 hey 15:30:19 how do I set a define via make again? -D doesn't work 15:30:26 galehar: hi! 15:34:00 -!- monky has joined ##crawl-dev 15:46:56 back 15:47:24 Zaba? 15:47:36 Keskitalo: what are the plans for Slime? I'm not informed. 15:51:09 Eronarn: is this from ##crawl? has ##crawl stopped being an intolerably unpleasant place? 15:52:16 Hey, Denzi is making jokes now :) 15:53:59 sorear: no, i am just a very tolerant person 16:10:08 [the correct answer was yes, I am just a very tolerant person] :) 16:18:41 fsim crashes wizmode, is this news? 16:19:31 &f that is 16:19:33 https://crawl.develz.org/mantis/view.php?id=2985 16:20:01 oh cool 16:20:52 you on Tiles or ASCII? 16:21:13 console, just made a note 16:31:14 -!- dpeg has quit [Quit: zzz] 16:32:44 03kilobyte * r60ef9db0a56a 10/crawl-ref/source/ (makefile.obj spl-damage.cc spl-tornado.cc): Split Tornado's code away from spl-damage.cc 16:32:46 03kilobyte * r18929323f28c 10/crawl-ref/source/spl-tornado.cc: Don't make Tornado levitation linger after its effect, no matter permalev. 16:32:46 03kilobyte * r1a270545b74a 10/crawl-ref/source/initfile.cc: Simplify access to ~/.crawl/macros.txt 16:32:49 03kilobyte * r5bf6e398b825 10/crawl-ref/source/abl-show.cc: Allow recasting of Injury Mirror. 16:32:50 03kilobyte * rb7ab6e75b036 10/crawl-ref/source/religion.cc: Cancel Injury Mirror on abandoning Yred. 16:32:53 03kilobyte * r3c3f58b29316 10/crawl-ref/source/ (ng-setup.cc player.cc): Fix a valgrind crash. 16:32:54 03kilobyte * rf0091b1b9e92 10/crawl-ref/source/player.cc: Fix a valgrind warning (only in --test, not regular games). 16:32:57 03kilobyte * r5755333cd5cc 10/crawl-ref/source/stuff.cc: With DEBUG_GLOBALS, explicitely free and zero the five big globals to catch leaks. 16:33:05 03kilobyte * ref678e2bfe9d 10/crawl-ref/source/ (player.cc stuff.cc): Plug valgrind errors reachable from no-op --builddb. 16:33:11 03kilobyte * r1b2363230e2b 10/crawl-ref/source/ (l_option.cc main.cc options.h stuff.cc): Include Options among globals handled by DEBUG_GLOBALS. 16:33:22 03kilobyte * r340fef0858cf 10/crawl-ref/source/tags.cc: Check for saves with broken monster ids. 16:33:25 03kilobyte * r1ec7cf2689ff 10/crawl-ref/source/ (mon-act.cc mon-clone.cc spl-cast.cc): When cloning a monster, give it an unique id. 16:33:26 03kilobyte * r9ea26faa6595 10/crawl-ref/source/ (tag-version.h tags.cc): Fix mon.number being 32 bit in memory but only 16 in saves. 16:35:01 galehar: I still have no clue how to reproduce your crash 16:35:43 monster clones produced duplicate ids but never blank ones 16:37:26 hint: it's possible that a friendly orc came up the stairs with me 16:38:02 it still would get an id previously 17:01:40 crap, the Ubuntu import freeze just passed, I was sure it's mid-January rather than yesterday 17:02:08 no 0.7 fixes for them ): 17:13:35 -!- jpeg has quit [Quit: Page closed] 17:14:27 morning 17:14:33 enne: ping! 17:16:18 due: Howdy. :) 17:17:44 ennne: http://sprunge.us/DjZV <-- This is the work I did on packed cell; basically, I decided to store a packed_cell in screen_cell_t instead of having to extend it every time we made an addition to packed_cell. The code works perfectly, but unfortunately it seems that if a square is ever haloed, silenced, etc, it remains that way and is never cleared. 17:18:20 enne: It's either an issue with my nnot clearing it properly, or bad programming... either way, I can't work it out :( 17:19:35 That seems like a reasonable approach. 17:21:09 I can't tell from a brief inspection why it's not working. 17:22:32 From what I can gather, when tile_apply_properties sets the boolean values in the member of screen_cell_t, they're set "permanently": the next iteration over goes "oh, new tile, yay!" and clears the tile aspect, but it doesn't clear the booleans. 17:22:37 -!- Twinge has quit [Ping timeout: 276 seconds] 17:22:51 There's a snippet of testing code in there that involves setting it to false... but that doesn't seem t owork either. :/ 17:39:24 -!- galehar has quit [Quit: disconnect] 17:47:31 -!- pointless_ has joined ##crawl-dev 17:55:06 -!- Twinge has joined ##crawl-dev 18:05:35 -!- jspiro has joined ##crawl-dev 18:10:46 -!- Textmode has joined ##crawl-dev 18:10:53 hi all. I'm playing Crawl 0.7.1. My laptop lost power, so Crawl was terminated unexpectedly. IIRC I just exited an ossuary. When I try to load the leftover save, Crawl terminates with an assertion failure in the travel cache loading code. A) Is this a bug? B) Should I report it and attach a zipfile of my save? 18:10:58 C) Tiles doesn't show assertion failures like some other Windows apps do; Tiles just terminates. Why? D) Isn't this a usability problem? 18:11:12 delete travel cache 18:11:29 ("just exited" means "maybe a few minutes before power loss). 18:11:37 check if you actually have a save 18:11:40 syllogism: shouldn't Crawl delete it for me? :) 18:11:43 well you can still report 18:11:46 sorear: I do. There are like a dozen files. 18:12:02 jspiro: but do they contain any data, or are they just lots of zeroes? 18:12:09 sorear: lemme see 18:12:34 are you sure it doesnt show the assertion failure message? 18:12:59 on windows console it gets printed near the top of the screen 18:13:09 syllogism: in Tiles? No, the game just disappears offscreen and immediately quits. This is Windows tiles. 18:13:47 By the way why does --edit-save report that I typed an invalid option? 18:14:08 sorear: I looked at Y.01a and it's not all zeroes. It contains lots of data. 18:14:27 what about the character file 18:14:38 .cs or whatever it is nowadays 18:14:53 .cs is a 0.8 thing 18:15:03 oh he is playing 0.7 18:15:15 jspiro: 0.7.1 doesn't use --edit-save 18:15:48 try renaming/deleting .tc and see if it works 18:15:54 sorear: is there an alternative to --edit-save for non-trunk users? 18:16:00 0.8 keeps all save data in a single file to increase crash resistance; --edit-save lets you manipulate components 18:16:04 syllogism: I plan to do that once I get back to a machine with Crawl on it. 18:16:10 jspiro: yes, just manipulate the component files directly 18:16:16 sorear: ah. 18:18:09 by the way wouldn't it be cool if Crawl saves were zipped text (or zipped XML like OpenOffice.org does)? 18:18:23 that'd make them easier to edit. 18:19:13 Erm 18:19:23 idea has been evaluated and rejected in the past 18:19:28 Most of what crawl saves is either binary information or numbers. 18:20:04 -!- Textmode has quit [Quit: Over the ages, man has been certain of a great many things. some of these things were even true.] 18:20:42 numbers are far easier to edit with a text editor than a hex editor. :) Also, spell lists, mutation lists, etc. could be text. 18:20:55 -!- Textmode has joined ##crawl-dev 18:22:01 -!- vimpulse has joined ##crawl-dev 18:22:53 there's wizmode for that 18:23:27 jspiro: ... why should they be in text? 18:23:36 -!- vimpulse has quit [Client Quit] 18:23:40 jspiro: They are represented using enum constants in the game. 18:23:50 1 sec 18:23:53 -!- jspiro has left ##crawl-dev 18:24:00 -!- jspiro has joined ##crawl-dev 18:24:15 that's true. It would be much more work for it to be text. And wizmode does let you do most everything. 18:24:19 It would be nice if we could have a more easily edited save format, but it's not currently compatible with the way crawl works. 18:26:44 Fair; thanks for explaining. What about my initial questions C) and D) btw? 18:27:46 Though saving the actual text values would go some way to removing save-compatability issues, it would require converting all enums into strings before outputting, and then translating them back again at run time. It would also vastly increase the size of saves. 18:27:51 * due scrolls up. 18:28:23 jspiro: If you start Tiles from the command line I believe that it will report assertions there. 18:28:26 it'd severely slow down things, would increase the size of saves several times (even with compression), and I don't think it would be possible to have it crash-safe at all 18:28:51 jspiro: And you're right, this is a usability issue; the lack of crash-dumps on Windows is a constant source of consternation to us. 18:29:05 kilobyte: how would storing enums as text affect crash safety at all? 18:29:24 jspiro: windows and tiles are both second-class support. Most of the devs use neither 18:29:44 We should probably have a tiles-specific assert function that opens a dialog in windows with the assertion message... 18:29:47 due: the lack of crash dumps (i.e. core files)? Or the lack of displaying assertion failure messages where users can see them? 18:30:19 jspiro: The latter; on *nix sources were do a complete dump of the game state into a text file, complete with a backtrace. This isn't possible on windows at the minute for a ... variety of reasons. 18:30:26 Does the Windows C runtime have a function to pop up an assertion failure dialog and let you continue or abort? 18:30:28 sorear: how can you do transactions in a text file? The only way I can think is rewriting everything to a temp file and doing a rename 18:31:07 jspiro: Well, assertion failures are an abort. It's a "we've reached this point and x is y, and it shouldn't be, and I can't recover". 18:31:09 kilobyte: oh, you're still on XML 18:31:24 kilobyte: I was thinking marshalEnum stuff 18:31:36 kilobyte: transactions? ooh. I know. You could use a SQL database. SQLite is ACID-compliant, according to Wikipedia. Firefox uses it. 18:31:43 And it's embeddable. 18:31:54 We already use SQLite for the vaults and so on. :) 18:31:58 *Does the Windows C runtime have a function to pop up an assertion failure dialog and force you to abort? 18:32:12 jspiro: Oh. I forgot to add "don't know" :) 18:32:34 jspiro: it's considered bad form to randomly drop in and tell people what technology they should be using 18:32:40 enne? Do you know if we can do a dialog for Windows tiles on assertions? 18:32:52 sorear: enums are a tiny part of save compatibility, and one that's easy to do 18:32:58 sorear: Suggestions of what people could use != telling people what technology they should use. 18:33:43 -!- MarvinPA has joined ##crawl-dev 18:36:28 I've never used the windows API before. 18:36:32 enne: Assertions should get written out to crash dumps. 18:37:02 (Hmm, talking to myself, apparently.) 18:37:08 (First sign of madness!) 18:37:29 http://msdn.microsoft.com/en-us/library/ww5t02fa(v=VS.100).aspx says that Microsoft Visual C++ includes assertion support in each of the MFC framework, the ATL framework, and the C run-time library, among others. http://msdn.microsoft.com/en-us/library/05d5fakw(v=vs.71).aspx says that assertion failure dialogs let you choose between three options. One is to ignore the assertion failure and continue execution. 18:37:41 speaking of sqlite, I see that abandoning it could give us a massive reduction of startup times. It's really not a fault of sqlite itself, merely the way we use it. 18:37:41 enne: what crash dumps? 18:37:59 jspiro: We don't really use Visual C++, sorry. 18:38:12 jspiro: The windows builds are cross-compiled using MingW. 18:38:41 due: maybe MinGW lets you use the Microsoft C runtime. Every modern Windows PC has a copy already installed. 18:38:41 ... which means that we're obviously going to have issues with the windows API. 18:39:10 jspiro: They're cross-compiled using MinGW on a Linux box, though :) 18:39:14 ignoring assertions is rarely ever a sane idea 18:39:33 According to Wikipedia MinGW ships with the libraries and header files for the Windows API. 18:39:38 just crashing with a proper dump should be enough 18:40:01 the "proper dump" is what we lack :( 18:40:49 Don't we at least dump the assertion, even if we don't have a call stack? 18:40:54 jspiro: mingw builds against msvc runtime, yeah 18:41:07 "Software and hardware manufacturers may access their error reports using Microsoft's Winqual program." --"Windows Error Reporting", Wikipedia 18:41:23 anyway, got to go. All the best. Thanks again. 18:41:34 enne: Pretty sure we don't do dumps at all ever on Windows. Could be wrong. 18:41:45 Yeah. 18:41:52 Confer crash-w.cc. 18:41:57 The article continues, though: "In order to ensure that error reporting data only goes to the engineers responsible for the product, Microsoft requires that interested vendors obtain a VeriSign Class 3 Digital ID." 18:41:59 * Summary: Windows specific crash handling functions. 18:42:09 20 lines in total, three stubbed out functions. :) 18:42:18 -!- jspiro has quit [Remote host closed the connection] 18:42:19 Poor, sad Windows. :( 18:42:26 Con: It doesn't work 18:42:27 That was rather odd. 18:42:29 Pro: It's efficient! 18:42:58 Anyway, they do raise a good point. A "we crashed with an assertion" dialog would be excellent for Windows, because otherwise "the game just exits when I do !". 18:43:20 Being able to give a rough idea as to what the assertion fail was or rather, where it was, would improve the quality of bug reporting with Windows tiles builds. 18:43:29 do the Digital IDs still cost 5000$ a year? 18:44:04 or do I mistake them for another type of MS signing key? 18:46:37 Who knows. 18:49:28 the Wikipedia page for winqual links to a page which in turn links to something which lists a way to get such certificates for "just" 116$ a year -- but I'm not sure if that one is good enough 18:50:23 oh, and you need to fax them a proof of incorporation 18:50:35 -!- pointless_ has quit [Ping timeout: 240 seconds] 18:51:05 $116 is $116 more than I'm willing to pay to Microsoft for the privelege. 18:52:08 you can make a fake document to fax in minutes if you have a sample, too 18:53:13 so... in a #if defined(TARGET_OS_WINDOWS) && defined(USE_TILE) preprocessor block we could define our own assert macro for displaying an alert dialog. 18:56:24 sounds like a good idea 18:56:33 All we need is the MessageBox routine. 18:56:44 and windows.h provides a block of NO for streamlining compillation as well. 18:56:46 and ExitProcess 18:57:09 we could also extend ASSERT to accept additional parameters 18:57:30 so it would give a better message if it fails 18:57:57 yeah 18:58:07 As opposed to the current ASSERT(!"error message")? 18:58:46 variadic macros are a C99/C++0x thingy, but we already use them 18:59:33 -!- MarvinPA has quit [Ping timeout: 265 seconds] 18:59:33 macros are scary 18:59:50 I mean ASSERT(x=0, "bad X: %d", x) 19:00:44 oh that would be awesome 19:04:17 there's a mantis discussion about not abusing assertions for failing 19:04:26 Yeah. 19:04:51 But are't asserts enabled for CDO, windows tiles builds, CAO, etc? 19:04:58 We may as well just disable disabling asserts. 19:05:40 ASSERT(!"error message) always fails, but it might be done more cleanly 19:06:57 I added fail() already, but it uses exceptions rather than exiting outright, thus not being a good idea for corruption 19:08:14 Huh, I can't actually find a copy of windows.h. 19:08:27 it is meant for, say, invalid saves -- a failure that doesn't suggest Crawl itself has gone bonkers 19:08:37 If there's one included in mingw32 dpkg -L mingw32 can't find it. 19:08:51 Ah! 19:08:58 It's in mingw32-runtime. 19:08:58 due: it's in one of dependencies 19:09:07 Yeah, just realised that. 19:09:37 it's a massive include fest rather than a single file, too 19:10:35 it's like if you took every single glibc include and linked to them from a single .h 19:10:43 * due shudders. 19:12:37 if you don't give a damn about compilation speed it's actually handy 19:14:08 And if you do you have to manually specify what you don't want. :/ 19:15:53 So, ... all we need is "MessageBox(NULL, (LPCTSR)"Assertion text", (LPCTSR)"Assertion fail", MB_OK|MB_ICONERROR)" 19:16:24 So, do we find every single NO define and iinclude that and then start removing them until it works? :D 19:16:53 That's exceedingly simplistic. Could it be so simple? 19:17:04 -!- MarvinPA has joined ##crawl-dev 19:19:44 you don't need to cast strings, too 19:20:26 passing NULL as the first argument is not a good idea though if there's another window 19:21:15 clearly we should just be using a memory-safe language where exceptions can be thrown at any time 19:21:54 note that there's no rule saying we can't just copy the MB_OK, MB_ICONERROR, LPCTSTR, and MessageBox definitions into our own header 19:21:57 kilobyte: It's wanting an "owner" window. 19:22:08 sorear: This also is true. 19:22:13 Microsoft has already committed to eternal binary compatibility... 19:22:14 the window manager will show the message box as another program and will allow switching back to the original window which is blocked by the MessageBox call 19:22:26 Ah. 19:22:54 I'm uncertain as to how to achieve the pointer to the main window via Tiles. 19:23:01 * due hasn't gone near that code before. 19:24:10 HAH! 19:24:24 The joys of using mingw32 to cross compile something into a windows compatible binary and then testing it via Wine. 19:25:45 not sure if we can get the window reliably in console too 19:26:00 Do we need to worry about in console? 19:29:52 -!- coolio_mc has quit [Read error: Connection reset by peer] 19:30:44 sorear: Is it that simple t ocopy the definitions over? 19:31:14 -!- ortoslon has joined ##crawl-dev 19:32:00 -!- Twinge has quit [Ping timeout: 265 seconds] 19:38:48 -!- Cryp71c has joined ##crawl-dev 19:53:31 Hm, the NO_blah defines seem to do nothing on mingw32. That's to be expected, I guess. 19:54:01 -!- Twinge has joined ##crawl-dev 19:59:32 due: yes 20:00:00 Wine is a lousy testing environment imo 20:01:01 sorear: Well, yeah. 20:04:31 -!- ais523 has joined ##crawl-dev 20:05:45 Monsters slept by recite can wake up immediately (https://crawl.develz.org/mantis/view.php?id=3101) by MarvintheParanoidAndroid 20:06:36 VirtualBox is the best way of dealing with Windows I know 20:06:55 kilobyte: and the one I prefer. 20:07:05 (besides Ripley's way, of course) 20:08:25 Ripley's? 20:09:05 nuke it from the orbit, that's the only way to be sure 20:09:27 Ah. :) 20:11:50 kilobyte: I can't find where to work out what the handler is for the main tiles window. :/ 20:13:45 next question please :p 20:13:55 enne? 20:13:57 * due makes puppy eyes at enne. 20:14:12 I can't resist the puppy eyes! 20:14:16 :D! 20:14:23 * due offers a jar of puppy eyes! 20:14:31 Ew! 20:14:39 What do you mean by handler? 20:15:38 SDL handles all of the window nonsense itself. 20:15:44 I'm sure you can ask SDL for whatever you're looking for? 20:17:51 Okay, so to the google mobile. 20:19:07 Easy! 20:19:12 http://www.sdltutorials.com/sdl-tip-window-handle/ 20:21:56 -!- Mu_ has quit [Quit: Defecator, may everything turn out okay so that you can leave this place.] 20:26:11 -!- Cryp71c has quit [Ping timeout: 240 seconds] 20:29:40 Okayy, so TARGET_COMPILER_MINGW = windows cross compile. 20:33:21 -!- syllogism has quit [] 20:37:08 kilobyte: http://sprunge.us/ZGWK 20:37:23 I don't really have a sane way of testing that. 20:38:01 Or... not. Wrong bit. 20:39:00 Oh no, it's right. 20:42:56 If someone could test that it would be great. 20:52:13 -!- Cryp71c has joined ##crawl-dev 21:18:35 -!- Cryp71c has quit [Ping timeout: 240 seconds] 21:28:35 -!- Cryp71c has joined ##crawl-dev 21:47:16 !tell dpeg Slime plan == Charles' proposal for tweaking the levgen a bit (which now seems like a much smaller change than I thought previously) 21:47:16 Keskitalo: OK, I'll let dpeg know. 21:48:07 Keskitalo: You build on windows, yes? 21:48:22 nope, not for a while now 21:48:25 Oh, hm. 21:48:34 I should really just cross-compile but I don't have the inclination. 21:48:41 I'll stick the patch up on the tracker for people to poke at a lot then. 21:50:07 I' on a windows achine right ow if you have a build to tes, i just don't have the build env 21:51:17 Don't have a build to test unfortunately. Not even sure if it *will* build. 21:51:19 Napkin, are you around? 21:55:11 -!- psyshvl has joined ##crawl-dev 21:57:11 ok, i'll try to sleep instead :) 21:59:55 -!- Cryp71c has quit [Quit: Leaving] 22:00:47 :) 22:06:24 Monster bands come upstairs together (https://crawl.develz.org/mantis/view.php?id=3102) by MarvintheParanoidAndroid 22:14:31 -!- Twinge_ has joined ##crawl-dev 22:16:17 -!- Twinge has quit [Ping timeout: 264 seconds] 22:28:27 MarvinPA (L11 SEEE) ASSERT(in_bounds(src_pos) || src_pos.origin()) in 'mon-behv.cc' at line 770 failed. (Lair:1) 22:28:35 Fun fun fun! 22:29:18 ow 22:29:28 i tried to lrd next to an oklob 22:29:33 :( 22:29:54 !lm . -log 22:29:54 10333. MarvinPA, XL11 SEEE, T:17448 (milestone): http://crawl.develz.org/morgues/trunk/MarvinPA/crash-MarvinPA-20101231-042828.txt 22:34:25 MarvinPA (L11 SEEE) ASSERT(in_bounds(src_pos) || src_pos.origin()) in 'mon-behv.cc' at line 770 failed. (Lair:1) 22:36:29 odd 22:36:34 [Patch] Provide a dialog message upon assertion for Windows Tiles builds. (https://crawl.develz.org/mantis/view.php?id=3103) by bookofjude 22:36:35 Can you do a savebackup? 22:36:36 looks like it crashes if you lrd the level border 22:36:42 but only if there's a monster there to hit 22:37:10 also you can shatter the level border and turn it into floor, but it stays unpassable 22:37:10 Hm. 22:37:20 Ah. 22:37:25 It checks the source position. 22:37:36 It checks that the noise is cmoing from within bounds. 22:37:42 http://crawl.develz.org/saves/dumps/MarvinPA-9319fad-101231-0437.tar.bz2 is the save, if you still need it 22:38:41 Well, beam code should be doing bounds checking before affecting out of bounds positions. 22:41:36 Crash when LRDing the level boundary (https://crawl.develz.org/mantis/view.php?id=3104) by MarvintheParanoidAndroid 22:42:42 <3 thanks 22:48:45 no problem :) 22:51:40 LRD can turn the level boundary into floor (https://crawl.develz.org/mantis/view.php?id=3105) by MarvintheParanoidAndroid 23:00:05 so reaching still can't reach diagonal huh 23:00:21 does ! work 23:00:31 yes 23:00:41 but autofight doesn't know that 23:01:58 you could probably put a modified autofight in your config 23:02:03 (one that knows that) 23:11:32 Or modify autofight and submit a patch? <3 23:11:46 [Patch] Re-implement tile flag handling (Tiles) with packed_cell. (https://crawl.develz.org/mantis/view.php?id=3106) by bookofjude 23:13:34 well it'd just be a horrible hack to get around circular ranges for reaching 23:14:04 but I guess it wouldn't be too horrible and it'd still work if/when that's fixed 23:14:14 I think squarelos' ship has sailed. 23:14:48 right, but reaching should be able to reach 2 squares diagonally still 23:14:49 -!- Twinge_ has quit [Read error: Connection reset by peer] 23:14:56 or it should be unable to 23:15:00 but not halfway between like it is now :P 23:15:10 (in that it can, but only by cheating) 23:15:11 solution: remove reaching ): 23:15:26 reaching should be innate in all polearms 23:15:39 -!- Twinge has joined ##crawl-dev 23:31:53 Troves don't pick which item to take from you very intelligently if you have multiple items meeting the requirements. (https://crawl.develz.org/mantis/view.php?id=3107) by elliptic 23:38:32 03due * r51afd0a504b4 10/crawl-ref/source/dat/clua/lm_trove.lua: Resolve #3107: Trove portals shouldn't inspect the ground (elliptic). 23:38:46 elliptic: <3