00:09:15 -!- valrus has quit [Remote host closed the connection] 00:17:31 LuckyNed the Slayer (L23 MiBe) (Elf:1) 01:05:26 -!- MarvinPA has quit [Ping timeout: 255 seconds] 01:21:03 -!- gaffe has joined ##crawl-dev 01:41:31 output.cc:824: warning: comparison between signed and unsigned integer expressions 01:41:31 output.cc:827: warning: comparison between signed and unsigned integer expressions 01:41:53 kilobyte, that seems to be related to your string width fixes 01:58:46 moin 02:09:49 -!- syllogism has joined ##crawl-dev 02:09:55 -!- valrus has joined ##crawl-dev 02:36:02 delicious string width! 02:40:17 -!- Kurper has quit [Remote host closed the connection] 03:06:03 -!- valrus has quit [Remote host closed the connection] 03:08:27 -!- eith has joined ##crawl-dev 03:30:41 -!- bmh has joined ##crawl-dev 03:30:47 -!- bmh has quit [Changing host] 03:30:47 -!- bmh has joined ##crawl-dev 03:31:38 -!- bmh has quit [Client Quit] 03:32:49 Zaba: this appears only in gcc-4.4, modern versions don't shout there 03:35:24 omg, 2010 isn't modern enough? ;) 03:37:07 2010, as in the year Crawl 0.5 was in? 03:39:04 kilobyte, okay 03:39:28 kilobyte, I wonder whether clang complains about that 03:39:57 gcc-4.4: April 21, 2009, gcc-4.5: April 14, 2010, gcc-4.6: March 25, 2011 03:40:49 actually, latest released clang (at least that from Debian unstable) can't even compile Crawl, you need to pull one from SVN 03:41:25 gcc-4.4.5 --version -> Copyright (C) 2010 Free Software Foundation, Inc. 03:41:52 it seems the FSF doesn't believe in copyright ranges 03:44:23 the point release .5 was indeed released in 2010 03:46:03 gcc-4.4 fails to compile crawl at least on powerpc, too 03:46:25 but even for that the --version isn't exact enough. if I want to know the real date of the point release, I'll have to go to the hp, too :/ 03:46:48 here on x86, gcc-4.3.5 works fine 03:47:14 i386 is quite mainstream 03:48:05 but even armel, being, by numbers, far more mainstream, tends to be badly supported 03:48:32 03kilobyte * r6b1a9735450c 10/crawl-ref/source/contrib/Makefile: Advertise native dependencies over contribs in the error message. 03:48:32 crawl 0.7 got no bug reports of being useless on armel for half a year after release :( 03:48:42 03kilobyte * ra004d6f0768c 10/crawl-ref/source/output.cc: Hush gcc-4.4 warnings. 03:49:45 does the compiling fail (C++ features not supported on some architectures?) or is the compiled binary faulty? 03:52:37 powerpc: ICE (gcc's fault, fixed in 4.5), armel: segfaults (our fault, fixed in 0.7.2) 03:54:04 although the problem on armel should be probably blamed on whoever had the brilliant idea of making "char" have different signedness on different architecture 03:55:14 fun thing is, we had -fsinged-char but it was set in a makefile variable which was then overwritten :p 03:55:23 life wouldn't be fun without such bugs 04:11:09 -!- Zaba has quit [Ping timeout: 240 seconds] 04:19:08 -!- monqy has quit [Quit: hello] 04:20:02 -!- Zaba has joined ##crawl-dev 05:13:02 -!- Pseudonut has quit [Remote host closed the connection] 05:30:37 -!- upsy has joined ##crawl-dev 05:38:58 -!- Mu_ has joined ##crawl-dev 06:09:10 -!- bmh has joined ##crawl-dev 06:09:23 -!- bmh has quit [Changing host] 06:09:23 -!- bmh has joined ##crawl-dev 06:09:41 It's been a while. I seem to recall someone saying that we spend a lot of CPU time generating random numbers. Is this true? 06:22:14 not a noticeable percentage, but way more than we should 06:22:56 on servers, that is -- local games use mt19937 and that's it 06:25:38 I think, instead of running mt19937 and then taking a costly hash of the result, we might just take consecutive numbers and AES them 06:26:25 with a long enough starting state it should be secure and fast 06:33:25 Mmmmersenne twister! 06:33:28 It's like some sort of dance. 06:36:35 fwiw, I recently implemented a bunch of generators that are statistically robust and outperform a solid MT implementation by an order of magnitude 06:38:49 I have a special little hatred for MT. It's a big silly generator that *still* fails some tests 06:40:42 -!- edlothiol has joined ##crawl-dev 06:45:02 -!- evilmike has quit [] 06:45:11 -!- edlothiol has quit [Ping timeout: 252 seconds] 06:46:02 -!- edlothiol has joined ##crawl-dev 06:55:07 bmh: are your generators of cryptographical quality? 06:56:18 kilobyte: heck no :-) 06:56:26 I mean, there are many good generators that provide randomness good for any simulations/etc, but they let people decode their state 06:56:49 sure, like MT 06:57:03 right 06:58:35 MT's speed is not a problem, it's about MT+SHA256's speed 06:59:10 we should be using a real cipher then 06:59:45 how did the state decoding attack work against CAO/CDO? 07:01:10 there are two ways of obtaining random numbers from the generator: 1. observing visible game state, 2. lua 07:01:17 -!- edlothiol has quit [Ping timeout: 250 seconds] 07:02:03 lua might be plugged by having two generators: one for the game itself and one for user interface 07:02:34 but game state, while somewhat tedious, is something I don't really know how to hide 07:03:23 MT is really easy to decode, because once you've observed enough outputs, you've actually captured the entire internal state 07:06:56 I'd feel comfortable running the output from a generator through any old block cipher in CBC mode, or any other reasonable mode 07:08:56 from what I understand, even running consecutive integers through a secure cipher would be good... but then, I'm not a crypto buff 07:09:28 yeah, running a block cipher in counter mode would work 07:11:56 I don't suppose we care much about statistical robustness 07:18:25 03j-p-e-g * r046d0de71e2f 10/crawl-ref/source/show.cc: Fix #3608: leaking of items/features when clearing monster trails. 07:28:38 reminds me, that cao has no /dev/random or /dev/urandom 07:29:00 yuck 07:29:14 how does it initialize the rng? 07:29:25 I don't know 07:29:37 I was told to create them for CDO a while back 07:31:09 kilobyte? 07:31:52 should we ask rax or greensnark to create those device files? 07:37:15 -!- bmh has quit [Quit: bmh] 07:44:08 -!- edlothiol has joined ##crawl-dev 07:54:46 Using numlock crashes game. (https://crawl.develz.org/mantis/view.php?id=3705) by Gigalith 08:03:03 Napkin: if there's no /dev/{u,,s,a}random, we use just time()+times()+getpid() 08:04:19 which is the prefered method? 08:05:13 in that order 08:05:38 is there any difference in cpu usage? 08:06:02 not for that little data as we take 08:06:13 /dev/random can block though 08:07:03 regular i386/amd64 produce a good bit of entropy even in a headless server, but this is probably not true in a Xen guest 08:26:36 -!- dpeg has joined ##crawl-dev 08:31:45 Hey, Hi dpeg :) 08:35:38 Hi 08:35:38 dpeg: You have 1 message. Use !messages to read it. 08:35:42 only a short visit 08:35:44 !messages 08:35:45 (1/1) Eronarn said (3d 11h 50m 9s ago): http://eronarn.info/casmith789.txt 08:43:16 -!- bmh has joined ##crawl-dev 08:43:16 -!- bmh has quit [Changing host] 08:43:16 -!- bmh has joined ##crawl-dev 08:43:21 hey dpeg! 08:44:14 Hi 08:46:32 judging by the logs, I think you missed the discussion about randomness in crawl 08:48:00 dpeg: meow! 08:48:39 kilobyte: I'm looking at sha256_genrand() -- that's a gross function. 08:49:28 it was a quick hack to plug MT's vulnerability 08:50:07 I'd be happy to provide an alternative implementation, if this is actually causing load on the server 08:50:40 well, undue load 08:51:51 about using a regular cipher: if my understanding is right, counter mode is in the worst case (fixed starting point) a known plaintext... ie, something modern ciphers are secure against 08:52:19 would using a non-crypto RNG as the base would provide any additional security or just slow down things? 08:52:59 I'd be more comfortable taking a non-crypto RNG with good statistical properties and XORing the output with a cryptographically secure stream cipher 08:53:43 A 'true random' stream XORed with any independent stream, should be just as random as the original stream 08:54:42 03kilobyte * rd39bda75610f 10/crawl-ref/source/food.cc: Revert no ZotDef hunger for non-Sp. 08:54:43 03kilobyte * r3c9e0a35eb2f 10/crawl-ref/source/windowmanager-sdl.cc: Don't assert on function keys that produce a symbol. 09:06:51 -!- valrus has joined ##crawl-dev 09:11:36 -!- valrus has quit [Ping timeout: 276 seconds] 09:55:28 Peaceful sea monsters around Elyvilon altars shouldn't block auto-explore (https://crawl.develz.org/mantis/view.php?id=3706) by Nexos 10:02:14 -!- MarvinPA has joined ##crawl-dev 10:45:13 -!- monqy has joined ##crawl-dev 11:07:32 -!- valrus has joined ##crawl-dev 11:20:15 dpeg: shall we take that awesome crypt end by evilmike? (https://crawl.develz.org/mantis/view.php?id=3704) 11:22:05 -!- valrus has quit [Remote host closed the connection] 11:25:49 kilobyte: yes, but I won't have time to check it in. 11:26:41 kilobyte: I see you tested it... just apply whatever tweaks you consider necessary, like the grave colours, and commit. 11:27:02 I tested the looks not balance, to be honest 11:27:09 I see :) 11:30:35 03kilobyte * re9d51cb4f618 10/crawl-ref/settings/init.txt: Mention ~/.crawl/init.txt in the sample config. 11:30:46 03kilobyte * r0273f5765bf3 10/crawl-ref/source/dat/des/branches/crypt.des: Make gravestones light grey rather than bright white. 11:30:47 03kilobyte * rd56517754dfc 10/crawl-ref/source/dat/descript/species.txt: Correct kobolds' short desc: they're obligate carnivores (Soadreqm) 11:30:48 03kilobyte * rcaef20d8c4e4 10/crawl-ref/source/ (4 files in 3 dirs): Some (crappy) tornado effect tiles. 11:30:48 03kilobyte * rfad21f77efd3 10/crawl-ref/source/dat/des/branches/crypt.des: An awesome Crypt end by evilmike. 11:35:01 cool, many thanks 11:35:51 thanks should go to evilmike, it's a masterpiece 11:36:05 he made good vaults before, I know :) 11:37:23 there has been a flood of forest vaults recently, though. Two Lair ends, my forest_0.8 and now this. 11:39:20 forest_0.8 had WEIGHT:999999 for a while (and it proved fruitful), I turned it down to WEIGHT:10 and toned down the Trog's temple area, but it can still hit people hard 11:39:21 * Zaba ponders making a forest vault 11:39:50 I haven't played since before we got a forest! 11:40:33 dpeg: you might want to take a look if it's totally out of whack for a regular vault balance-wise 11:40:45 Zaba: just a vault? You're thinking low! 11:41:02 orc tiles acting strange in arena (https://crawl.develz.org/mantis/view.php?id=3707) by ghend 11:41:05 kilobyte, what else? 11:41:13 uhm, branch? :p 11:41:17 kilobyte, should I write a level generator for the branch? 11:41:38 (in lua) 11:42:48 Zaba: I've been mulling with one, but I have issues coming up with a generator that can 1. regenerate the area out of sight while not leaving visible seams, and 2. plays well with many minivaults 11:43:44 the idea is to have different levels work with different rules: a static level at the bottom, at least one level that changes all the time, perhaps more... 11:43:51 regenerating the area out of sight, kind of like in lab? 11:44:03 (except lab doesn't actually regenerate, but still does stuff out of sight) 11:44:06 yeah, just more aggressive 11:44:13 kilobyte: I'll have a quick look later today. Cannot expect evilmike to completely screw it up :) 11:44:14 that sounds wicked 11:44:21 it's trivial for an acyclic maze 11:44:52 dpeg: I mean, at my stuff... evilmike can be trusted to do things more reasonably :p 11:45:18 Zaba: actually, I think 1. and 2. might be the same... especially if we place vaults _first_ 11:46:11 the generator can look for places to attach paths to and calculate their local direction 11:46:38 a vault with an open edge is no different from a clearing 11:48:03 the first algorithm I came up with can't do this, though :( 11:49:23 placing minivault before generating the level kind of defeats the purpose of them being minivaults 11:51:05 Crash on using number keys to move (https://crawl.develz.org/mantis/view.php?id=3708) by christhawk 11:52:07 Just so it can bring some insight: 1. make a grid graph with edges ~10 long; I think hexagonal would be best, 2. "shake" it for some time: take a vertex, move it randomly to a place that doesn't cause any edges to cross -- for quite a number of iterations; 3. remove some edges, leaving the minimal spawning tree and some percentage of others; 4. draw the edges with some fuzz (paths); 5. bloat some nodes (clearings) 11:54:05 overcomplicated, probably 11:54:07 why have a separate shaking stage instead of randomly picking the initial vertex positions? 11:54:39 as in, doing triangulation? 11:56:43 well, I was thinking of just determining the possible range of cooridinates for each vertex based on the number of them in the grid and the map size and randomly picking them within that range 11:57:45 I don't think you can refer to that as triangulation... 11:58:14 03j-p-e-g * rad2e507f97a0 10/crawl-ref/source/describe.cc: Fix #3441: inventory actions occasionally affecting the wrong item. 11:58:42 ok, please fasten your seatbelts.. tryng to activate eaccelerator on the webserver 11:59:55 I have another idea, though: 1. take a random starting point (a vault's exit, etc) and guess its local direction; 2. keep two angles: a "big-scale" and a "small-scale" one; 3. walk forward, randomly altering the angles. The small one can fluctuate wildly but is biased to return to the big one; 4. walk until hitting a map edge or another path, sometimes make a clearing at that spot; 5. goto 1 12:00:31 Zaba: so, in your variant, edges would often intersect, right? 12:00:48 kilobyte, why? The ranges of the coordinates wouldn't overlap 12:01:31 the graph would look quite artificial then... that "shaking" version can make them wander quite far while preserving the graph's topology 12:02:04 I didn't even try actually implementing this yet, though 12:02:29 0.8 and all that stuff... 12:02:31 or it could let edges intersect, I don't think the extra chaos would actually hurt anything 12:02:49 perhaps, yeah 12:02:49 the vertex density could also be random 12:03:23 heck, even if we come up with multiple competing generators it wouldn't be a bad thing 12:03:40 yeah, we would just pick one at random 12:04:12 from what I've seen of forests, they don't normally have paths laid out in regular grids, so making something that looks like roguey_level would indeed be bad (and forest might as well use roguey) 12:04:22 now, if it was a Park branch... 12:04:28 that'd be different :> 12:06:05 -!- jpeg has joined ##crawl-dev 12:09:20 ok, 200mb cache for pre-compiled php 12:09:26 let's see how that goes 12:09:44 -!- edlothiol has quit [Ping timeout: 240 seconds] 12:09:50 Hi jpeg :) 12:10:06 hi :) 12:10:22 we need a trunk update yet again now the numpad issues are fixed 12:10:25 pretty please? 12:10:36 sure! 12:10:43 it's running 12:11:07 Napkin: ah, good you told me _before_ I pressed the button 12:11:16 try, kilobyte ;) 12:11:42 good! 12:11:47 by chance i clicked that report to check if adding php still works with this caching thingy and saw the report :) 12:11:57 *and saw the need 12:12:09 wtf, forget that sentence :D 12:12:57 !send Napkin ?amnesia 12:12:58 Sending ?amnesia to Napkin. 12:13:11 hehe 12:13:22 thanks ;) 12:13:26 anything else we need to do before merging the branch? 12:14:46 nothing you can't cherry-pick, I think 12:15:55 Windows development builds on CDO updated to: 0.8.0-a0-6177-gad2e507 12:16:25 issues on my table: 1. formatted_string is one place left that is wonky with non-ASCII, 2. Trog seems to be giving far too much unrandarts, 3. glyphs that can't be currently drawn in tiles could use a fallback 12:17:21 292 php scripts cached 12:17:41 kilobyte: what do you mean by 3.? 12:17:44 using 47mb.. easy! 12:21:07 jpeg: having proper font fallback would take quite a bit of effort and/or adding fat dependencies 12:21:37 you mean for the messages? 12:21:41 thus my current code can draw only characters 0..0xff (ISO-8859-1 range) 12:21:55 any text written in tiles 12:22:26 it's better than 0.7 which mangled anything non-ASCII in most cases, but still not good 12:22:42 fortunately, this hardly ever happens 12:22:49 I'm trying to come up with cases where this will matter - inscriptions? 12:23:19 and if a Tiles player were to remap the monster/feature glyphs, I guess 12:23:27 "just don't do that then" -- ie, you can't write inscriptions in Chinese yet, unlike on console. Not a big problem. 12:23:40 the real issues are remapping and system messages about paths 12:23:56 do you remember those Russians whose paths were inaccessible? 12:24:31 currently tiles will show those reversed question marks :( 12:24:32 I gather the paths themselves work now, but the messages ("init file read from ...") will look off 12:24:41 exactly 12:24:47 I rather like the reversed question marks :) 12:24:59 much nicer than the oddly apostrophed i it was before :P 12:25:13 * kilobyte would prefer actual Cyrillic there :p 12:25:27 and 0.9 will have that 12:25:33 oh well, we'll get bug reports about that, but it is a huge improvement already 12:25:36 thank you! 12:27:36 kilobyte: you know that the current default settings don't work for most people, right? dozens of people on ##crawl have needed to be told about char_set=ascii to make their games playable 12:27:45 there are two ways for doing font fallback: 1. a load of platform specific code, 2. pango 12:28:17 -!- stabwound has joined ##crawl-dev 12:28:48 imho, the default settings should be the really basic ones 12:29:01 with an option to load a single settings file to replace them with fancy stuff 12:29:09 (including changing the char_set option) 12:29:09 elliptic: was there something that hasn't been fixed already? The only problem left I know of is when you change to a specific font (Terminal) on putty. 12:29:56 kilobyte: I don't know what has been fixed in the last day or so, but the usual problem is that you end up with single square using up multiple characters... or at least that's my problem :) 12:30:14 03kilobyte * ra9ab6c87ea49 10/crawl-ref/source/mon-data.h: Allow raising trapdoor spider zombies. 12:30:26 03kilobyte * re1eefe161870 10/crawl-ref/source/viewchar.cc: Make clouds paragraph marks by default (Eronarn) 12:30:57 elliptic: lemme see... CDO, right? And what is your terminal? 12:31:40 -!- edlothiol has joined ##crawl-dev 12:32:01 elliptic: (and which of your chars, if they use different configs) 12:32:42 whatever the default xfce terminal is called 12:33:29 @rc hyperelliptic 12:33:32 this was as hyperelliptic, though I've never put in any special stuff in my config for charsets (until now) 12:33:55 http://crawl.develz.org/configs/trunk/hyperelliptic.rc 12:34:12 the char_set = ascii was added in to fix things 12:34:26 I can remove that and go find some water or something 12:34:41 to see whether there is still a problem for me 12:36:02 well, d:10 seems okay for me now 12:36:26 it was where I first ran into trouble with the character yesterday 12:36:33 ok, so it was one of the bugs fixed, great :) 12:36:51 I'll continue testing without the char_set = ascii, then :) 12:38:04 the idea is to have unicode _capability_ on, but using it by default only when there's a reason to. Like, clouds or trees, but not anything that has a decent glyph in ASCII already. 12:38:17 water is questionable 12:40:21 ah right, clouds are a problem 12:40:38 See https://crawl.develz.org/mantis/view.php?id=2454 12:40:44 actually, char_set = ascii does Unicode too, just defaulting to all ASCII glyphs, including the abysmal \ for portals. So you can use mon_glyph = orc priest {ö} 12:41:32 (except that neither of DGL's editors will allow you to put literal characters into your config yet... but numeric values work as they used to) 12:43:56 jpeg: Eronarn suggested §, and I put it as the default for now. I personally use ະ which looks a lot better, but being a Thai character it won't work for terminals without font fallback. 12:44:26 I only see a rectangle glyph here on freenode 12:44:38 § is both in CP437 and WGL4 12:44:52 I meant the Thai glyph :) 12:45:03 (still better than re-using #, I admit) 12:45:04 ah, that's ok. 12:46:00 I mean, not ok but expected. Neither xterm, rxvt nor putty can fallback to other fonts. 12:49:23 elliptic: please tell me which glyphs can be contentious. Getting clouds out of '#' is good, trees out of '7' is good too (new players won't confuse it with demons), portals at '\', fountains at '}' looked just wrong, but there are glyphs I'm not sure about. 12:50:11 if you're getting trees out of 7, what about statues/golems both on 8 12:50:13 of course, some players who got used to '7' can be disoriented... but when assaulted by green deaths and midges, the change is good 12:51:06 having monster and feature statues on different glyphs would be nice 12:51:08 a lot of people seem to dislike the new tree glyph, but I don't have a better suggestion 12:51:16 what is it? 12:51:20 monqy: good point 12:51:38 what is wrong with portals at \, by the way? 12:51:40 elliptic: trees? They didn't change. 12:51:48 they aren't 7 12:51:56 -!- edlothiol has quit [Ping timeout: 255 seconds] 12:52:06 elliptic: doesn't look like a portal, at all 12:52:20 well, most things in ascii don't look like what they are 12:52:24 confusable with wands, mostly 12:54:18 I didn't get far enough earlier before needing to set char_set=ascii to know exactly what changed 12:56:04 I started from old cset_unicode so I ponder changing books from ∞ to : 12:56:26 lemme convert the numbers to glyphs 12:57:20 oh, books are ∞? that sounds contentious 12:58:41 it is - I've been complaining about it for ages 12:58:55 but since I play tiles I only see the borked glyphs in the command help 12:59:41 in general I think that items that you can pick up should be the standard ascii in the default settings... makes more sense to do fancy glyphs for features 13:00:34 I agree 13:00:52 portal: \ → ∩, fountain: } → ⌠, water: ~ → ≈, item_detect: { → ⌂ (not ∆ because of CP437), book: : → ∞, cloud: # → §, tree: 7 → ♣, explosion: # → Θ 13:01:05 items should use the same glyph as the selection command in the pickup/drop screen 13:01:09 ie, book... sure, will change 13:01:32 what is CP437? 13:01:50 kilobyte: nice, except for the book ;) 13:01:53 (I liked ∆ better) 13:01:55 not sure about explosion, too... this hardly ever appears: for a split second for fireball, while you --more-- messages during firestorm 13:02:39 elliptic: it is better, yeah, but might not work on Windows with bitmap fonts 13:02:57 also, anything special for monster detection? 13:03:02 but then, the bitmap fonts there are broken in most locales... and it's not a new breakage 13:03:23 doh, { is not only invis_exposed but also Ash 13:03:43 (didn't notice since I use circled digits) 13:04:45 lemme see how ∆ looks on English win2k/XP... since this is the only locale where bitmap fonts actually work 13:05:23 Θ looks very much like 0 in my terminal font 13:07:30 upsy: right... not sure if it needs to be distinguished from # badly enough, since it never stays even for a whole turn 13:08:05 why not re-use the cloud glyph for explosions? 13:09:32 not sure how you change the glyph for things like explosions either, since you can't examine that ingame for the name 13:09:36 do "feature = explosion {}" and "feature = beam {}" work? 13:10:18 upsy: no... and they are not changeable with feature = either 13:10:28 only cset = 13:10:56 oh right, that would make much more sense 13:11:13 should be overridable this way, but no one did it yet 13:14:00 and looking at options_guide.txt at least, I don't see a cset for those 13:15:05 back to bugfixing... 13:15:10 bye 13:15:13 -!- jpeg has quit [] 13:15:22 by the way, can someone tell me why the hell ANYONE ever uses Windows 7? http://angband.pl/tmp/win7_lan.png -- this is copying to a computer in the basement, 1 wired ethernet switch away. Win2k did it in a small fraction of a second. 13:15:35 s/to/from/ 13:17:12 both in a virtual machine, but at at least two physical machines Win7 does that too -- when talking to any system other than another win7. Including XP and what not. 13:17:45 by which i mean http://pastebin.com/GMknJjxX 13:20:23 uh, indeed 13:24:09 kilobyte: is it intended that beam spells use §? 13:28:43 a solid-type symmetric glyph would be better 13:28:57 it's showing that the square is filled by a cloud, not that there's some object in the square 13:30:07 elliptic: uhm, DCHAR_FIRED_ZAP is '*' 13:30:28 is anything on # now, even 13:30:34 branded bolts/darts use it 13:30:37 elliptic: in all predefined charsets 13:30:43 upsy: uhm, walls :p 13:30:52 in unicode 13:31:16 (unless you use char_set = unicode (traditional), ibm or dec) 13:31:49 kilobyte: well, I haven't seen it myself but people have been commenting in ##crawl on stuff other than clouds being § 13:33:46 kilobyte: I'm watching it right now, branded bolts from yaktaurs are making lines of § 13:34:00 I guess I misunderstood when people were talking about beams earlier 13:34:32 still fighting with Win7, I'll check in a second 13:34:43 it managed to fail the copy 13:34:48 wow 13:36:10 since, you know, expecting a wired ethernet with a gigabit card from such as unknown network chip maker as Realtek on a network with 0% packet loss to have a bandwidth better than an acoustic modem is unheard at in Redmond 13:40:20 uh.. my FN keys macros don't work anymore :-O 13:41:12 Took a while to parse that as not referring to the Fn key of laptops. 13:41:15 I saw some changes about the f-keys.. should I update CDO to make it work again or reconfigure my macros? 13:41:46 yeah, sorry 13:41:56 I mean the ordinary F-keys 13:42:00 kilobyte? 13:42:29 Napkin: no, it was arrow keys+NumLock in tiles 13:42:29 yeah, what is up with macros? a lot of people on ##crawl have been complaining about them getting messed up 13:43:40 my F1 is suddenly reporting: \{-1011} 13:45:32 previously it was reporting \{1011} 13:46:51 ok.. they all just give negative values now 13:47:14 F12: \{276} -> \{-276} 13:48:16 strange 13:48:40 Napkin: yeah, I had to do that since +1011 is Greek yot 13:49:12 ah, ok 13:49:22 numbers between 0 and 0x10ffff (1114111) are taken... 13:50:04 I could migrate old macro files but it would be problematic if you have an actual greek keyboard and want to macro one of those letters... 13:50:40 nah, no prob, i just negated the few that i had 13:51:58 elliptic: the bolts were a mistake, already fixed 13:52:36 03kilobyte * rd769e99db472 10/crawl-ref/source/viewchar.cc: Make default books ':' again, use different glyphs for detected items on Windows. 13:52:38 I guess I'll update CDO so we don't get outdated complaints 13:54:42 Unstable branch on CDO updated to: 0.8.0-a0-6180-gd769e99 (32) 13:56:34 I think a dev blog post would be nice to reduce confusion 13:56:37 -!- ais523 has joined ##crawl-dev 13:56:48 complaints are good when they have a chance of being constructive 13:59:45 -!- Kurper has joined ##crawl-dev 14:10:14 03kilobyte * r63f08da3a980 10/crawl-ref/ (docs/options_guide.txt source/viewchar.cc): Add missing cset overrides. 14:20:49 I'll never get over the fact.. that a level like snake4 can have uniques like D20 or something 14:39:55 -!- OG17 has quit [] 14:45:01 -!- elliptic has quit [Quit: Leaving] 14:47:33 -!- OG17 has joined ##crawl-dev 14:54:35 -!- minced has joined ##crawl-dev 15:06:01 -!- ZChris13 has joined ##crawl-dev 15:12:06 KiloByte the Ruinous (L7 OgCj) (D:5) 15:18:49 -!- bmh has quit [Quit: bmh] 15:22:59 KiloByte the Ruinous (L7 OgCj) (D:6) 15:29:35 -!- bmh has joined ##crawl-dev 15:32:02 -!- bmh has quit [Client Quit] 15:46:03 Unstable branch on crawl.akrasiac.org updated to: 0.8.0-a0-6181-g63f08da (32) 16:09:01 unlimited xp pool is killing the layout: 16:09:03 XL: 27 Exp: 105174Place: Pandemonium 16:09:17 s/killing/ruining/ 16:10:24 eventually it runs into place, e.g. Exp: 1051749lace 16:17:08 -!- dpeg has quit [Quit: Lost terminal] 16:17:25 Autoexplore walked into conjured flame with no warning (https://crawl.develz.org/mantis/view.php?id=3709) by effo 16:19:53 -!- minced has quit [Ping timeout: 240 seconds] 16:23:38 -!- minced has joined ##crawl-dev 16:29:23 -!- valrus has joined ##crawl-dev 16:34:49 -!- syllogism has quit [] 16:35:33 -!- MarvinPA has quit [Ping timeout: 276 seconds] 16:41:52 -!- ZChris13 has quit [Quit: ChatZilla 0.9.86.1 [Firefox 4.0/20110318052756]] 16:42:33 Felid respawning does not respect no_tele_into (https://crawl.develz.org/mantis/view.php?id=3710) by Mu 16:50:03 -!- galehar has joined ##crawl-dev 16:54:37 03kilobyte * r1b530ed1ed63 10/crawl-ref/source/ (60 files): Doxygenization and junk purging that was missed by my script. 16:54:48 03kilobyte * r1097d773f78f 10/crawl-ref/source/ (libgui.cc libgui.h libunix.cc libunix.h libw32c.cc libw32c.h): Don't rely on undocumented behaviour of ncursesw. 16:54:48 03kilobyte * rec3a84509275 10/crawl-ref/source/misc.cc: Obey no_tele_into when bringing you a new fresh kitty. 17:02:38 LRD on translucent rock walls makes brown explosion (https://crawl.develz.org/mantis/view.php?id=3711) by effo 17:10:38 thanks for the quick fix kilobyte! 17:11:49 not a really taxing one, though 17:12:00 well all the same :) 17:17:11 -!- bmh has joined ##crawl-dev 17:19:45 -!- MarvinPA has joined ##crawl-dev 17:24:50 -!- Pseudonut has joined ##crawl-dev 17:37:49 After death, ESC is very slow to exit. (https://crawl.develz.org/mantis/view.php?id=3712) by XuaXua 17:41:35 # Q: How is there a dead forest in the Crypt? 17:41:36 # A: ~use ur Imagination~ 17:41:37 lol 17:43:36 use pilcrow for trees 17:43:44 it looks better than ♣ 17:45:10 here's a good tree glyph 7 17:45:28 yeah, 7 is fine really 18:06:12 -!- galehar has quit [Remote host closed the connection] 18:20:26 -!- minced has quit [Ping timeout: 255 seconds] 18:24:35 03kilobyte * raa8822dde835 10/crawl-ref/source/ (acquire.cc dungeon.h makeitem.cc): Reduce the amount of unrands among god gifts. 18:24:47 03kilobyte * re6cfca1ae901 10/crawl-ref/source/spl-damage.cc: Correct the colour of LRDed glass. 18:30:27 -!- minced has joined ##crawl-dev 18:31:45 -!- minced_ has joined ##crawl-dev 18:32:52 !tell Zaba there's an ugly hack to force stone in Zot to masquerade as rock: a number of places are hard-coded to look for player_in_branch(BRANCH_HALL_OF_ZOT) to do so. I'd clear it but I wonder what we _want_ stone in Zot to look like. 18:32:52 kilobyte: OK, I'll let Zaba know. 18:35:17 -!- minced has quit [Ping timeout: 255 seconds] 18:35:19 "This is a wall of Zot." 18:35:25 -!- minced_ is now known as minced 18:51:41 kilobyte: should probably be distinct from all normal wall types, and Zot rock 18:58:37 might be just permarock... it's used only in a few vaults 19:00:26 chukamok (L14 DSWz) (Hive:1) 19:14:26 -!- minced has quit [Ping timeout: 255 seconds] 19:17:53 -!- edlothiol has joined ##crawl-dev 19:21:24 -!- minced has joined ##crawl-dev 19:25:51 -!- evilmike has joined ##crawl-dev 19:31:28 -!- minced_ has joined ##crawl-dev 19:32:34 -!- minced__ has joined ##crawl-dev 19:32:53 -!- minced has quit [Ping timeout: 255 seconds] 19:33:00 -!- minced__ is now known as minced 19:36:29 -!- minced_ has quit [Ping timeout: 255 seconds] 19:44:08 -!- Mu_ has quit [Quit: Defecator, may everything turn out okay so that you can leave this place.] 20:01:21 hey, can I use Ely's Purification to remove Mara's mislead effect? 20:03:44 last I heard a potion of healing removes mislead 20:04:41 -!- eith has quit [Ping timeout: 240 seconds] 20:08:18 -!- Pseudonut has quit [Ping timeout: 264 seconds] 20:31:53 -!- edlothiol has quit [Ping timeout: 240 seconds] 20:51:13 -!- valrus has quit [Remote host closed the connection] 21:00:49 -!- valrus has joined ##crawl-dev 21:04:46 -!- minced has quit [Read error: Connection reset by peer] 21:04:50 -!- Pseudonut has joined ##crawl-dev 21:04:53 -!- minced has joined ##crawl-dev 21:17:24 -!- MarvinPA has quit [] 21:43:09 -!- Cryp71c has joined ##crawl-dev 21:50:08 -!- minced has quit [Ping timeout: 255 seconds] 22:36:40 -!- minced has joined ##crawl-dev 22:43:58 -!- ortoslon has joined ##crawl-dev 22:54:34 -!- Jordan7hm has quit [Ping timeout: 246 seconds] 23:04:49 -!- OG17 has quit [Read error: Connection reset by peer] 23:12:44 -!- OG17 has joined ##crawl-dev 23:18:27 -!- minced has quit [Quit: ChatZilla 0.9.86.1 [Firefox 4.0/20110318052756]] 23:20:21 -!- due has quit [Read error: Operation timed out] 23:21:39 -!- due has joined ##crawl-dev 23:25:34 -!- ZChris13 has joined ##crawl-dev 23:25:39 -!- valrus has quit [Remote host closed the connection] 23:33:51 Rawr 23:40:28 -!- due has quit [Read error: Operation timed out] 23:45:08 -!- due has joined ##crawl-dev 23:45:08 -!- due has quit [Changing host] 23:45:08 -!- due has joined ##crawl-dev 23:57:55 -!- Cryp71c has quit [Quit: Leaving]