00:03:21 Unstable branch on crawl.develz.org updated to: 0.11-a0-1252-gccceac4 (32) 00:10:43 -!- shirish has joined ##crawl-dev 00:16:10 -!- G-Flex has quit [Read error: Connection reset by peer] 00:16:42 -!- nrook has quit [Quit: ChatZilla 0.9.88.2 [Firefox 11.0/20120312181643]] 00:18:14 Windows builds of master branch on crawl.develz.org updated to: 0.11-a0-1252-gccceac4 00:18:33 -!- G-Flex has joined ##crawl-dev 00:38:38 -!- evilmike has quit [Quit: something happened] 01:05:41 -!- G-Flex has quit [Quit: Time for us to bug out.] 01:11:57 03kilobyte * rd090918256bb 10/crawl-ref/source/dat/descript/monsters.txt: Add a desc for "crazed wizard" (hellish Lair end). 01:11:57 03kilobyte * raaea51835ffd 10/crawl-ref/source/ (32 files): Mindexicide: behaviour_event(). 01:11:57 03kilobyte * r888cf270cf25 10/crawl-ref/source/mon-stuff.cc: Don't let n_adj/n_suf survive polymorphs. 01:11:58 03kilobyte * r22539c8b98c4 10/crawl-ref/source/ (item_use.cc quiver.cc quiver.h): Split =f/+f into =f/+f (autoquiver) and =F/+F (ammo cycling). 01:11:58 03kilobyte * re7a5462849e7 10/crawl-ref/docs/macros_guide.txt: Don't call Tab unused in ?~ 01:13:59 -!- shirish has quit [Ping timeout: 246 seconds] 01:33:07 -!- Wensley has quit [Ping timeout: 276 seconds] 01:36:28 -!- ZChris13 has quit [Quit: ChatZilla 0.9.88.2 [Firefox 11.0/20120312181643]] 01:42:54 -!- G-Flex has joined ##crawl-dev 01:44:12 -!- blabber has joined ##crawl-dev 01:45:26 -!- Felyza__ has quit [Read error: Connection reset by peer] 01:45:48 -!- Felyza__ has joined ##crawl-dev 02:04:37 -!- MarvinPA_ has joined ##crawl-dev 02:06:54 -!- MarvinPA has quit [Ping timeout: 250 seconds] 02:06:54 -!- MarvinPA_ is now known as MarvinPA 02:20:22 -!- ais523 has joined ##crawl-dev 02:24:16 -!- monqy has quit [Quit: hello] 02:29:27 U(un7ul1nuX 02:30:04 oops.. 02:30:50 it might have been a cat, but I'll write that down on a list of people's passwords, just in case :p 02:31:15 In case you ever get your hands on my laptop I guess you're golden.. 02:31:35 you mean, it actually was one? 02:31:55 Yep. 02:32:23 well, I've done something similar before too 02:32:35 Some 1337-5p33k and some capitalization, along with an easy to remember phrase -> instant pw 02:33:08 xkcd recently pointed out some flaws there 02:33:14 Yeah, it's not perfect. 02:33:48 -!- magistern has quit [Quit: This computer has gone to sleep] 02:33:58 I'm no good at remembering/coming up with long phrases 02:35:43 Anywaym it's not supposed to stop brute forcing, it's suposed to stop guessing by classmates. 02:35:49 So, it works :) 02:36:22 http://xkcd.com/538/ 02:36:26 Also related :) 02:36:52 my personal way includes some mappings I know by heart: like, Doom2 map names (16 -> The Suburbs), ASCII codes (e -> 101), and generating every password by an algorithm that depends on something related to the place it is for 02:37:13 -!- headzone has joined ##crawl-dev 02:37:21 Mine is something like that yeah. 02:37:28 as long as the algorithm is unobvious enough, you have a different password for every use, and can't forget them 02:38:15 I have a standard password for random game forums and such as well. 02:38:30 Stuff I don't care about. 02:39:17 hell yeah, obviously 02:46:19 -!- Pingas has joined ##crawl-dev 03:20:18 I just use lastpass :p 03:22:24 Chousuke: so any time you manage to lose your data, you lose every single password as well? 03:23:07 or when there's a boo-boo in your browser, or that lastpass thingy. And browser vulnerabilities come in monthly. 03:28:25 the passwords are stored encrypted on a server; the plugin fetches them to the local machine and decrypts them there. 03:29:24 I think it's more secure than that, I'll have to look up how it works again 03:29:25 you mean, "the cloud" is safer than a single not backed up disk? 03:29:45 well, I'm pretty sure it is :P 03:30:25 the service can't read my passwords, they just store them. 03:30:26 http://blogs.computerworld.com/18265/four_things_you_should_know_about_lastpass 03:30:36 'LastPass's Grid feature generated a random grid of numbers which I printed out and carry in my wallet. It randomly asks for the numbers from different positions on that grid every time I log in. No numbers, no passwords.' 03:30:42 I've heard about that kind of thing before 03:30:44 it looks pretty good 03:32:10 ie, your wallet get stolen, *poof*. Some other bad guys get a copy, *poof*. Some Keystone Kop wants to sniff you, *poof*. 03:32:20 LastPass INC goes down, *poof* 03:32:36 LastPass decides this service is no longer profitable, *poof* 03:32:48 -!- ToBeFree has joined ##crawl-dev 03:32:59 Well, it depends how paranoid you are 03:33:11 you can just export your passwords if lastpass shuts down 03:33:17 presumably they wouldn't do so without notice 03:33:25 LastPass goes away forever - any service that authenticated itself using lastpass can fall back to email authentication until they shuffle over to something else 03:33:48 -!- dtsund has quit [Quit: dtsund] 03:34:10 LastPass's databases are hacked - won't give hackers enough information to determine your password (and if it uses something like scrypt salting, will be prohibitively expensive to guess and check too) (also, if there's additional authentication like with a yubikey, they won't have enough information - and they also don't have master passwords, etc) 03:34:16 assuming they do have a fallback plan 03:34:41 Your wallet is stolen - the wallet stealer will have your yubikey/equivalent but lack your master password. You'd need a seriously determined criminal to get both at once, if you're paranoid about it 03:34:50 Well, they might not HAVE one, but lastpass isn't the only service of its kind 03:36:16 did you remember to have another couple of copies of that number grid hidden somewhere else? 03:38:01 if you lose it you can just log in from a trusted computer and invalidate it I guess 03:38:27 I don't use the grid so I'm not sure how it works exactly 03:38:35 basically, you get a big grid of numbers 03:38:57 "trusted computer", aha. Ie, depending on their setup, either something that's likely to fail, or something anyone can impersonate. 03:38:59 when you log in, lastpass decides 'I want line X starting at character A or column Y starting at character B' 03:39:03 and you have to get your card out and type in only those numbers 03:39:16 that way, even if your data is being intercepted, they only know a small portion of the card 03:39:16 kilobyte: it all depends on how YOU set it up 03:39:28 and if services lock you out after X failed attempts, they don't have enough information to reliably masquerade as you 03:40:55 you can be super paranoid and use two-factor authentication and one-time-passwords and what have you or you can worry less and just use the master password to access your password vault from wherever you are. 03:41:51 they provide a screen keyboard to get around keyloggers too :P 03:42:01 looking at the page you linked, it says access to the primary e-mail account gives full override to everything, including that master password 03:42:21 including recovery of "forgotten" master password 03:42:25 Chousuke: that doesn't really work nowadays, modern keyloggers take screenshots on mouse clicks 03:42:41 ais523: I suppose. 03:42:46 make your cursor invisible 03:43:15 Patashu: or make the symbols on the buttons flash on and off, and you can click them when they're off 03:43:59 or, type using an alternative typing tool, like that one where you point into an infinite tree of all possible sentences 03:44:03 forget what it's called now 03:44:45 dasher? 03:44:49 yes, dasher 03:44:54 anyway, :paranoia 03:49:36 hm. Access to the email address only enables you to get a local OTP so that you can access the vault using the OTP from a computer that has been previously used to access it 03:50:43 if you have access to a computer, you also have access to email (or will have it soon) 03:51:12 how so? 03:51:25 -!- alefury has joined ##crawl-dev 03:53:02 most folks have some kind of autologin, and for those who don't, the keylogger will provide the ssh key passphrase/etc 03:53:31 or really, even paranoid folks tend to have ssh-agent or such, so a keylogger is not needed 03:53:49 (of course, if you can access ssh-agent you can set up a keylogger, it's merely a matter of convenience) 03:53:59 i'm basically using two-way auth with gmail, and the second they have me, my phone and my computer they may as well do it 03:54:06 but computer alone is not sufficient 03:55:50 it is, they'll just wait for when you're logged in 03:56:30 you can't protect against everything 03:57:06 we're getting into territory that's not really going to happen to most people :P If you have secrets that would make attackers go to such lengths, don't store them in lastpass 03:57:14 of course, typically you do everything from a single computer, but it's good to have some alternate means that are TOTALLY unconnected 03:57:25 hmm, true :p 03:59:12 still, even though I haven't, ehm, accessed a remote machine without authorization, in many years (5?), I'm not going to ever get v&ed if I'd ever do that again but forget about some stupid shit like this 04:01:00 -!- tholmes has quit [Read error: Connection reset by peer] 04:01:24 -!- tholmes has joined ##crawl-dev 04:06:25 03kilobyte * r6755d9ef5326 10/crawl-ref/source/util/optimize-pngs: Make util/optimize-pngs work with /bin/sh != bash. 04:06:25 03kilobyte * r25140de0ac14 10/crawl-ref/source/ (18 files in 3 dirs): roctavian's tiles for Spider walls. 04:16:14 03kilobyte 07stone_soup-0.10 * r4feebba05a8e 10/crawl-ref/source/libutil.cc: Mention the region size when asserting out on invalid screen writes. 04:16:17 03elliptic 07stone_soup-0.10 * re6334b3e1df8 10/crawl-ref/source/mon-stuff.cc: Fix Powered by Death mutation sometimes not triggering (Patashu). 04:27:35 -!- Ragdoll has joined ##crawl-dev 04:32:31 -!- elliptic has quit [Ping timeout: 276 seconds] 04:33:06 how do i kill local commits again? i tried git reset --hard HEAD, but that didnt work 04:33:16 -!- st_ has joined ##crawl-dev 04:34:38 HEAD^ 04:34:58 reset --hard HEAD means you drop uncommitted changes 04:35:49 -!- Pacra has quit [Quit: This computer has gone to sleep] 04:36:19 if you have several commits to trash, try reset --hard origin/HEAD to reset everything to the remote head 04:36:28 alefury: reset --hard then the hash of the newest commit you want to keep 04:37:46 -!- elliptic has joined ##crawl-dev 04:40:42 kilobyte, ChrisOelmueller, Patashu, Chousuke and anyone else have you heard of "rubberhose de-cryption" or "rubberhose cryptanalysis"? 04:42:55 that's when you beat the encryption keys (or plaintext) out of a person with a rubber hose 04:42:59 that's what we call "social engineering" 04:43:25 social engineering has a rnage of possibilities, yes 04:46:08 -!- mumra has joined ##crawl-dev 04:46:13 phyphor: that's why it's important to keep your identities far enough from each other that they won't be connected, including not being connected with your person 04:46:28 Or not being interesting enough 04:46:45 thanks, either HEAD^ or origin/HEAD worked :) 04:47:03 not sure which 04:47:16 origin/HEAD is not the same -- unless you have precisely one commit compared to origin 04:47:42 (if shit happens, reflog to the rescue!) 04:47:56 ahhhh 04:48:05 it was precisely one commit, so im probably fine 04:48:30 this cant be so hard to do, right? i mean, its a pretty common scenario... 04:49:02 not really, depends on your workflow 04:49:18 im new to git, i dont have a workflow 04:49:21 if you work in feature branches and decide to abandon them, you'd just not care or delete the branch 04:49:39 but not everybody is most productive with that, especially when just starting 04:49:40 right, i know that now 04:49:43 you don't kill commits, you just reset the branch :) 04:51:05 Chousuke: thats what i tried to do, to the single branch i used... 04:51:11 Chousuke: unless you don't know about reflog 04:51:12 the general rule is you don't delete, edit or rebase anything that you've shown to someone else 04:51:54 alternatively you can getr away with it as long as you stick your fingers in your hears when yo do it and say "lalalal, this never happened" 04:52:04 there's no need to be too careful with git though 04:52:05 of course that requires 3 hands so it's a barely used exception 04:52:13 if you've committed it once, it's hard to get rid of 04:52:29 octopode_reasons 04:52:37 hahah 04:52:38 Chousuke: not really; reset it away from all branches, then garbage-collect 04:52:50 ais523: that's hard enough! 04:53:00 it's two commands in the most common case 04:53:07 and the second one will get run automatically eventually 04:53:28 yeah, but the reflog prevents stuff from being gcäd for 90 days IIRC. 04:53:34 s/ä/' 04:54:36 two commands with a 90-day duration between them 04:54:40 still two commands ;) 04:57:10 kilobyte: regarding 5582, are unknown items in a pile even worth having? I mean, whats the point? 04:57:42 that you don't benefit from xv on every pile you see 04:57:55 yes, but the tradeoff is that you benefit from walking to every pile you see 04:58:05 which is not better imo 04:58:05 it makes the game less fun! 04:59:03 walking to every pile is automated with autoexplore, but thats really the only thing it has going for it imo 04:59:06 alefury: the difference between walking to piles and doing xv is that one of them is an action that takes time in-game, the other one is tedious interface 04:59:31 why should view items on a pile have an ingame cost, while viewing spread out items doesnt? 04:59:47 why can't you automatically see everything in a pile, anyway 04:59:49 without xv 04:59:58 oh my god 04:59:58 phyphor 05:00:04 I was right about what "rubber-hose cryptanalysis" is 05:00:08 OSX builds of master branch on crawl.develz.org updated to: 0.11-a0-1257-g22539c8 05:00:08 I seriously just made that up 05:00:09 yes, yes you were 05:00:12 phyphor: probably the interface awkwardness 05:00:13 off the top of my head 05:00:17 I was joking :( 05:00:31 ais523: you see a dwarven dagger on top of 32 darts 05:00:59 phyphor: well... 24 items come into view 05:01:13 it'd get ridiculous with some of the loot piles you get in NetHack, after killing a castle full of soliders 05:01:15 *soldiers 05:01:21 phyphor: but what about the more common situation where a monster with multiple items dies 05:01:23 -more- 05:01:28 you see a dwarven dagger on top of 32 darts on top of a runed orcish scale mail on top of a runed dart on top of a poisoned needle on top of an uncursed glowing blowgun 05:01:49 ais523: in DoomRL it can get even worse 05:01:50 randomly exploring and seeing a stack of items is quite rare 05:01:56 but you also avoid the situationw here you can't see a large shield under a dart 05:02:03 ais523: and there items don't even floodfill but spill through walls 05:02:39 in Pokémon Mystery Dungeon, items can only be one per square and can only spill up to two squares, that works quite well 05:02:52 (especially as it's impossible to make stashes in that game) 05:03:03 (so you don't have the problem of stashing being awkward) 05:03:07 ais523: ugh, does that mean items are destroyed if there isn't room for them? 05:03:09 brogue also has only one item per square (and stashing is impractical because it actually has a food clock) 05:03:18 elliptic: yes, but you'd have to be really trying to get that many items in an area 05:03:56 it'd take about half your inventory to fill an area enough that items started getting destroyed 05:04:27 well, games like crawl have many more items... a single monster could have too many items to fit in a corridor with that rule 05:04:55 yeah, it surely wouldnt work in crawl 05:05:17 indeed 05:05:39 in PMD monsters can only carry one item at a time, and (on earlier levels) don't even know how to use them, they just pick them up in the hope they do something 05:05:47 fr: auto-squelching in crawl 05:05:49 ancient versions of angband worked like this in combination with monsters who could drop up to 15 items... so you always wanted to lead them into rooms before killing them 05:05:50 ais523: interface awkwardness vs play awkwardness 05:06:06 phyphor: right 05:06:36 people can modify an interface using options 05:07:05 I guess the standard Crawl method would simply be to remove items that aren't relevant by the time you reach them 05:07:26 -!- syllogism has joined ##crawl-dev 05:08:21 how do i update the sql contrib? so the crawl i compile can actually run? 05:09:05 git submodule update, I think 05:09:16 i tried that, it said it did something, but it still doesnt work 05:09:31 after make contrib and make TILES=y 05:09:56 oh, maybe i have to pull again? 05:10:07 nope... 05:12:16 alefury: are you on Windows? If not, contribs are always a bad idea. 05:12:20 i am 05:13:51 anyway, the first time i ran update it said it checked out some sqlite version, but make contrib said nothing to be done for contrib, and make TILES=y also doesnt really do anything 05:14:42 ill try redoing it after make clean :/ 05:16:41 make clean-contrib, just in case 05:19:38 some crap about permission denied, and Error 1, Error 2 05:19:55 -!- Mottie has quit [Ping timeout: 264 seconds] 05:21:32 after restarting the console it worked 05:21:38 wow, i did not miss this crap 05:22:12 only common way to get a permission denied when you own all the files involved is if you try to execute something that isn't an executable, or isn't flagged as one 05:22:37 i couldnt open that folder with windows explorer either 05:23:36 ais523: unless it's Windows 05:23:48 ais523: deleting files seem to randomly fail often 05:23:58 kilobyte: oh, you can't delete an open file in Windows 05:24:09 just like you can't delete an open directory in Linux 05:24:26 ais523: ? Try it. 05:24:29 (this is also the reason that daemons do a cd / once they're done setting up; it's so you can delete the directory they're in 05:24:57 mkdir a && cd a && rmdir ../a && pwd 05:25:05 huh, you can delete open directories sometimes 05:25:16 perhaps it's something else you can't do with them 05:25:24 always can, or it'd violate POSIX 05:25:29 $ ls ../b 05:25:31 ls: cannot access ../b: No such file or directory 05:25:39 can't unmount a busy filesystem, though 05:25:42 $ pwd 05:25:43 /tmp/a/b 05:25:57 actually you can, there's a switch to umount that tells it to unmount it the moment it becomes not-busy 05:26:13 but it's not unmounted until then 05:27:16 indeed 05:27:26 likewise, the directory isn't actually deleted until you stop using it 05:27:30 it's just unlinked 05:27:59 -!- Textmode has quit [Ping timeout: 255 seconds] 05:28:47 the difference is, the directory is gone from the tree visible to anyone who doesn't have an open descriptor to it 05:29:10 it's still accessible via /proc 05:29:12 could be better to mount a placeholder over the pending-umount filesystem 05:29:16 -!- Textmode_ has joined ##crawl-dev 05:29:33 hmm yeah, you can use that to borrow someone else's descriptor 05:33:09 $ cd .. 05:33:10 cd: error retrieving current directory: getcwd: cannot access parent directories: No such file or directory 05:33:22 $ pwd 05:33:25 /tmp/a/b/.. 05:33:27 OK, this is increasingly making no sense 05:34:22 creating a subdirectory in a deleted one makes it even more convoluted, yeah 05:34:56 this is BTW the reason git's upstream rejected my patch to allow using "git bisect" from a subdirectory 05:37:28 they care about cases like checking out Crawl, sitting in crawl/crawl-ref/source/, bisecting to a revision that didn't use this scheme (ie, commits from 2006) then try running an automated bisection script 05:38:26 I think I'd be far better to just scream and throw an error in such a case 05:39:41 -!- Textmode_ has quit [Ping timeout: 255 seconds] 05:42:07 -!- shmup has quit [Ping timeout: 264 seconds] 05:48:53 -!- shmup has joined ##crawl-dev 05:49:02 -!- Textmode_ has joined ##crawl-dev 06:21:59 -!- MarvinPA has quit [*.net *.split] 06:21:59 -!- Patashu has quit [*.net *.split] 06:21:59 -!- Snowclone has quit [*.net *.split] 06:24:42 -!- Textmode_ has quit [Ping timeout: 255 seconds] 06:28:32 -!- MarvinPA has joined ##crawl-dev 06:28:32 -!- Patashu has joined ##crawl-dev 06:28:32 -!- Snowclone has joined ##crawl-dev 06:29:26 -!- Pacra has joined ##crawl-dev 06:39:24 -!- Textmode_ has joined ##crawl-dev 07:09:03 -!- MarvinPA has quit [Read error: Connection reset by peer] 07:09:28 -!- MarvinPA has joined ##crawl-dev 07:28:09 -!- MarvinPA_ has joined ##crawl-dev 07:30:39 -!- MarvinPA has quit [Ping timeout: 256 seconds] 07:30:39 -!- MarvinPA_ is now known as MarvinPA 07:39:32 -!- ToBeFree has quit [Read error: Connection reset by peer] 07:42:32 03dolorous * rb7668d5b5df4 10/crawl-ref/source/dat/descript/monsters.txt: Tweak wording of crazed wizard description. 07:44:12 -!- jeanjacques has joined ##crawl-dev 07:48:33 -!- syllogism- has joined ##crawl-dev 07:50:09 -!- syllogism has quit [Ping timeout: 245 seconds] 07:54:40 -!- jarpiain has quit [Ping timeout: 260 seconds] 07:54:44 -!- jarpiain has joined ##crawl-dev 07:55:11 -!- jarpiain is now known as Guest35775 08:31:00 -!- Xiberia has joined ##crawl-dev 08:43:26 -!- blabber has quit [Read error: Connection reset by peer] 08:44:28 -!- blabber has joined ##crawl-dev 08:48:47 -!- magistern has joined ##crawl-dev 08:52:19 -!- mumra has quit [Read error: Connection reset by peer] 09:57:37 Morban (L12 OpWr) ASSERT(y >= 1 && y <= sz.y) in 'libutil.cc' at line 942 failed. (D:12) 10:07:56 <|amethyst> !lm morban crash x=v -log 10:07:57 1. Morban, XL12 OpWr, T:18531 (milestone): http://crawl.develz.org/morgues/0.10/Morban/crash-Morban-20120421-145729.txt 10:08:00 <|amethyst> !lm morban crash x=v 10:08:00 1. [2012-04-21] [v=0.10.0] Morban the Covert (L12 OpWr) ASSERT(y >= 1 && y <= sz.y) in 'libutil.cc' at line 942 failed on turn 18531. (D:12) 10:08:24 <|amethyst> !lm morban crash x=tiles 10:08:24 1. [2012-04-21] [tiles=y] Morban the Covert (L12 OpWr) ASSERT(y >= 1 && y <= sz.y) in 'libutil.cc' at line 942 failed on turn 18531. (D:12) 10:08:46 <|amethyst> this one involved Passage of Golubria 10:09:21 Unstable branch on crawl.akrasiac.org updated to: 0.11-a0-1260-gb7668d5 (32) 10:10:14 -!- Textmode_ has quit [Ping timeout: 255 seconds] 10:10:54 <|amethyst> I think there is a crash still in tunk when a monster was at the very edge of LOS, you get trampled so that it's LOS, and it moves back in LOS in the same turn 10:11:53 <|amethyst> err, *when* the edge of LOS is the edge of the screen, anyway 10:12:16 <|amethyst> I was thinking about maybe making view_update_at() return immediately if pos is outside the screen, maybe even when pos is outside LOS 10:12:58 <|amethyst> view_update_at does have a very explicit comment that it should only be called for squares in LOS, though 10:13:17 <|amethyst> so maybe it's monster::check_redraw()'s fault for calling it on a square out of LOS 10:13:34 <|amethyst> OTOH, it does that for a reason---so the monster doesn't leave a trail when it enters LOS 10:14:22 <|amethyst> so maybe monster::check_redraw() should verify that old is onscreen before doing view_update_at(old) 10:16:05 <|amethyst> hm 10:16:13 <|amethyst> no, it's a little more complicated than that 10:16:30 <|amethyst> the situations when it triggers, that is 10:25:08 -!- Textmode_ has joined ##crawl-dev 10:34:22 <|amethyst> eh, I think I'll just have view_update_at avoid drawing offscreen cells (still doing the show_update_at() for map knowledge etc) 10:38:06 -!- Blade- has joined ##crawl-dev 10:39:30 -!- Textmode_ has quit [Ping timeout: 255 seconds] 10:47:15 -!- mikee_ has quit [Quit: HydraIRC -> http://www.hydrairc.com <- Like it? Visit #hydrairc on EFNet] 10:57:56 -!- Textmode_ has joined ##crawl-dev 11:02:46 -!- Wensley has joined ##crawl-dev 11:03:41 -!- nrook has joined ##crawl-dev 11:26:50 -!- elliptic has quit [Ping timeout: 246 seconds] 11:38:50 -!- Mottie has joined ##crawl-dev 11:39:45 -!- elliptic has joined ##crawl-dev 11:41:37 03|amethyst * r0321078a4479 10/crawl-ref/source/view.cc: Don't draw off-screen cells. 11:42:18 03|amethyst 07stone_soup-0.10 * rd961752debb5 10/crawl-ref/source/view.cc: Don't draw off-screen cells. 11:45:45 -!- elliptic has quit [Ping timeout: 246 seconds] 12:04:10 Sosloow (L8 CeBe) (D:7) 12:08:43 -!- HangedMan has joined ##crawl-dev 12:14:13 -!- shirish has joined ##crawl-dev 12:14:22 03policarpocp * rcdc7f72a3fc6 10/crawl-ref/source/l_item.cc: New clua binding that returns an array of floor items. 12:14:32 03|amethyst * r198c943b116a 10/crawl-ref/source/l_item.cc: Avoid info leak, get_items_at -> get_item_at. 12:17:09 -!- G-Flex has quit [Read error: Connection reset by peer] 12:18:40 -!- G-Flex has joined ##crawl-dev 12:21:59 -!- elliptic has joined ##crawl-dev 12:35:48 -!- Dixbert has joined ##crawl-dev 12:36:51 -!- blabber has quit [Quit: leaving] 13:03:57 -!- MarvinPA has quit [Read error: Connection reset by peer] 13:16:50 03|amethyst * rb306b9965d0e 10/crawl-ref/source/ (main.cc misc.cc): Prompt for moving next to allies etc while confused. 13:18:10 -!- ToBeFree has joined ##crawl-dev 13:22:03 -!- medgno has joined ##crawl-dev 13:30:47 -!- medgno has quit [Quit: medgno] 13:44:05 -!- bb|kek has joined ##crawl-dev 13:52:35 03|amethyst * re285e6384794 10/crawl-ref/source/rltiles/dc-dngn/floor/ (12 files): Darker spider floor tiles (roctavian). 13:57:00 -!- elliptic has quit [Ping timeout: 246 seconds] 14:01:55 -!- dtsund has joined ##crawl-dev 14:05:09 Stable branch on crawl.develz.org updated to: 0.10.2-17-gd961752 (32) 14:17:16 -!- elliptic has joined ##crawl-dev 14:30:56 -!- elliptic has quit [Ping timeout: 246 seconds] 14:40:24 -!- evilmike has joined ##crawl-dev 14:44:28 Debian builds of 0.10 branch on crawl.develz.org updated to: 0.10.2-17-gd961752 14:49:33 -!- elliptic has joined ##crawl-dev 15:17:01 03kilobyte * r68d7b4c9d7cf 10/crawl-ref/source/l_you.cc: Remove a stray prototype that doesn't belong in that file. 15:17:11 03kilobyte * r2efd080bc107 10/crawl-ref/source/l_item.cc: Revert get_items_at -> get_item_at, fix it not working out of LOS, fix crashes. 15:18:29 -!- heteroy has joined ##crawl-dev 15:20:51 <|amethyst> kilobyte: thanks... I had missed the typedef, didn't realize that item_info really was an item_def 15:23:54 |amethyst: still not sure how to handle stash items 15:23:54 <|amethyst> kilobyte: glad your commit made it in before mine, since yours does more :) 15:24:06 oops, sorry for a conflict 15:24:10 <|amethyst> no problem 15:24:24 <|amethyst> literally every change in mine was also in yours 15:24:51 you said something about stashes... how far did you get investigating them? 15:25:39 I see it holds item_defs too 15:25:40 <|amethyst> kilobyte: I implemented something that worked like the code in full_describe_view() 15:25:57 <|amethyst> the problem is, that whole API copies item_defs around a lot 15:26:16 <|amethyst> so you end up with a copy of a copy of a copy, which is of course allocated on the stack 15:26:32 <|amethyst> so you can't safely put a pointer to it in the userdata you're pushing 15:27:39 also, ANY lua item function can break if there's wizmode involved 15:27:58 <|amethyst> break how? 15:28:05 <|amethyst> oh 15:28:13 <|amethyst> because they have a pointer to the mitm struct 15:28:19 <|amethyst> s/struct/array/ 15:28:20 there's an assumption all lua references last until the turn counter is incremented 15:28:25 yeah 15:28:50 <|amethyst> even out of wizmode 15:29:09 does that Chei ability bump the turn count? 15:29:16 <|amethyst> well 15:29:22 (should IMO, it's an action even if it takes 0 aut) 15:29:27 <|amethyst> I was thinking you could still have the item more than a turn later 15:29:33 <|amethyst> oh 15:29:38 <|amethyst> it stores the turn 15:29:42 <|amethyst> so it can be invalidated 15:29:42 <|amethyst> okay 15:30:28 <|amethyst> when I looked at it earlier I had worried "what if you keep this lua object around for a long time and the item gets replaced in mitm?" 15:31:11 <|amethyst> if I knew the LUA-C API a little better I might be able to come up with something 15:31:40 -!- MarvinPA has joined ##crawl-dev 15:31:48 AFAIK nothing the player can do in non-wizmode can possibly alter an item without bumping the turn (got to check Chei) 15:31:53 <|amethyst> It seems you should be able to new() or malloc your item_def and tell lua to free it when the wrapper is discarded 15:31:54 wait, Ashenzari's scrying too 15:32:21 what chei ability are you even talking about? 15:33:08 |amethyst: actually, that sounds like a good idea. We could store item_defs referenced. Not sure when to free them though. 15:35:09 -!- bb|kek has quit [Quit: Leaving] 15:36:25 got it: the wrapper could store a copy of the item rather than the live version 15:36:52 what'd you say about that? Somewhat wasteful, but safe. 15:36:57 <|amethyst> hm 15:37:09 <|amethyst> does dlua use the same interface to manipulate objects? 15:37:18 <|amethyst> because then you do want a pointer to the real object 15:37:29 pan exits are far too common imo. Pan has lost any semblance it had of being difficult to leave, which was the entire point of the branch 15:39:51 st_: yeah, fully agree 15:40:09 st_: I somehow reduced the frequency on portal_branches, but it's still way too high, I guess 15:40:15 it would be nice if the random exits were much less common but the chances for exits in pan lord vaults was kept 15:40:48 |amethyst: I'm not sure dlua can even modify objects at all currently 15:41:44 <|amethyst> kilobyte: it looks like all the do_* should work with copies 15:42:32 item_property_set() 15:42:33 <|amethyst> kilobyte: until something changes, anyway (because then their link won't match the real item) 15:43:19 <|amethyst> ah 15:43:45 <|amethyst> what about instead getting the list from the stash, and using index() to get the "real" pointer 15:44:37 index()? Not sure what you mean. 15:44:59 <|amethyst> item_def::index() // Returns index in mitm array. Results are undefined if this item is not in the array! 15:45:17 works only while in mitm 15:46:05 <|amethyst> oh 15:46:06 primary use for not in mitm is player's inventory (not sure if lua can access any other uses) 15:46:14 <|amethyst> it just subtracts 15:47:44 <|amethyst> hm... in the cases where it is a copy, you could use lua_newuserdata to allocate a copy of the item_def 15:48:05 <|amethyst> then lua does the collection 15:48:07 if I want pretty printer goodness, I have to compile gdb, right? There doesn't seem to be deb packages for 7.3 15:48:58 by the way on pan, in older versions we used to have those big open hex levels (they had abyss gates in the middle) but for some reason they were removed, we don't need the abyss gates any more but I don't see why the level type can't be there 15:49:08 galehar: there's some major configuration magic which I don't understand but copied as-is, otherwise it works fine for me 15:49:27 galehar: the python stuff is in something-dbg 15:49:30 kilobyte: which version of gdb are you using? 15:49:50 galehar: I guess libc??-dbg or libstdc++??-dbg 15:49:59 kilobyte: I found this stuff, but I think it needs gdb >= 7.3 15:50:17 <|amethyst> what version of what distribution? 15:50:26 debian 15:50:39 err unstable 15:50:48 <|amethyst> 7.4 is in wheezy and sid, and 7.3 in squeeze-backports 15:50:55 (can't remember the codenames...) 15:50:56 hmm, lemme check, I don't run stable on any machine I mess with Crawl on 15:51:51 squeeze 15:52:31 <|amethyst> You can add squeeze-backports to your sources (see http://backports-master.debian.org/) 15:52:42 <|amethyst> then you should be able to get 7.3 15:52:48 thanks 15:52:52 st_: yeah, i dunno what happened to those 15:53:04 they existed for a while with the abyss exits removed i think? 15:53:07 <|amethyst> http://backports-master.debian.org/Instructions/ is the page I should have linked, sorry 15:53:09 but then disappeared somehow 15:53:22 (I thought I installed unstable...) 15:54:19 <|amethyst> anyway, I'm out for the evening... later 15:54:47 |amethyst: hrm, bad to see you go. I guess, let's talk about this later then. 16:01:15 MarvinPA: I think it's quite cool to have pan and abyss connected, thematically I liked it that going through the abyss was the best way to exit pan (but it's bad gameplay wise) and apparently there used to be pan entrances in abyss which would be fun for weird speedrun/challenge stuff 16:02:06 st_: I went into a pan entrance in abyss whenever I found one while abyssal rune scumming 16:02:10 regarding pan exits, they do tend to be rather common, but its also annoying if you just want to escape pan, but keep getting unlucky 16:02:23 especially if you've reached the point where pan is no longer a threat (eg you just finished a zig) 16:02:23 the typical SpAK can survive around four or five levels in Pan before getting oneshotted by something that they don't have a chance to run away from 16:02:47 so, i'd say if pan exits were made less common, it should also be easier to identify which levels they are on. using octagon levels for that is a good idea 16:03:28 i don't think tying it to a specific layout is particularly good 16:03:28 -!- HangedMan has quit [] 16:04:39 seems pretty spoilery, and kinda defies the point of having them in pan lord vaults if you can just dive through floors until you get that layout 16:05:08 using pan lord vaults is good, but even with those it isnt perfect 16:06:09 on my last pan visit I did a zig, and then spent maybe 10 (just a guess, but it took a while) levels looking for an exit. I was unlucky, but it felt really tedious and there was no challenge at all 16:06:20 just peek into the vault, see if there's a portal, go to next level 16:07:08 yeah, having exits in pan lord vaults is only really interesting for weak characters who can't killdudes everything 16:09:31 maybe if the chance of a random exit (1/30 per upstair right now, I think) was tied to the number of pan runes you have, or something 16:11:31 I was thinking about that too 16:12:06 personally though I find it much more tedious to get a demonic rune than to leave Pan... 16:13:27 the demonic rune can be a bit annoying, but I usually don't have trouble with it. The way I play, I explore every level until I find a demonic rune... I think I've only had one or two games where I found it after the four unique pan runes 16:13:53 helps a lot that its guaranteed in a couple of vaults (somewhat arbitrarily) 16:14:43 IMO if you never need to abandon a pan level because there are too many nasty 1s, then pan is pretty boring anyway :P 16:14:52 but yeah, the vaults where it is guaranteed help 16:15:30 hells have increased in difficulty a lot over the past few versions, but pan has maybe gotten easier 16:16:09 st_: evilmike's demonic rune vaults definitely made pan a bit harder; I'm curious how you think pan is easier? 16:16:34 gloorx and mnoleg have also gotten harder in the last few versions 16:16:39 easier to exit mainly 16:17:05 I did forget about the demonic rune vaults 16:17:07 yeah, exits are more common at least, though that comes with not being able to get out with distortion 16:19:55 pan could maybe use some vaults containing the portals to the next level 16:20:26 currently the portals to the next level are always just scattered around unguarded; maybe sometimes some of them could be guarded 16:20:34 you mean like, put all three stairs in it? (could still get hatches outside) 16:20:53 sometimes all three, maybe sometimes just one or two... not sure 16:22:10 it feels too easy to run to a new level... the only drawback of leaving a level mostly unexplored is that you might miss out on the guaranteed demonic rune vaults 16:22:13 one thing I'd like to see is a better sense of progression for the branch. Harder demons as you go "deeper" 16:22:23 might be better to just make it get harder the more levels you've seen, yeah 16:23:15 there's also the possibility of making it finite... it gets harder and harder, and after some number of levels (81?) you are kicked out (just exits) 16:23:42 with persistent levels, perhaps? 16:23:42 with a guaranteed demonic rune on the last level, of course, and guarantees of getting the rune levels before then 16:23:47 I don't mind it being infinite, myself 16:24:06 elliptic: yeah, one of my thoughts about pan was to have a set number of levels, and once you enter you can't exit until the end (maybe you can go through multiple times) 16:24:21 kilobyte: what do you mean? letting people go backwards, or letting people go in again and start from the beginning? 16:26:44 although, 81 levels is a huge number, way above the average number of levels you'll see 16:26:47 elliptic: one of ideas that has been discussed a couple years ago is to have 81 levels (yeah, you're not the first one with that number! :p), and have you visit a random level every time you move 16:27:11 ie, the more you sit there, the more often you end up in places you have already been to 16:27:21 kilobyte: hm, that could work too... you'd want to give the special levels extra weight I guess 16:27:38 yeah, that was the primary concern 16:27:39 27 is roughly how many levels it takes to find all 5 runes, for reference 16:27:53 and fudging the chance up sounds good 16:28:10 the other concern are Ziggurats 16:28:18 bringing you back to empty levels repeatedly sounds kinda weird 16:28:29 my main concern about that is that I foresee people being really annoyed when they keep on going to already cleared levels, yeah 16:28:35 yeah, i dont like that (even if autotravel is made to work in pan) 16:28:36 most folks currently don't explore them fully 16:29:18 elliptic: 81 levels means you're scumming 16:29:36 (or didn't explore the levels, in which case they're still mostly fresh) 16:29:50 anyway it could go well with the idea for having ziggurats be infinite 16:29:53 kilobyte: yeah, I guess... but even in 27 you'll still get several duplicates 16:30:11 X> 16:30:25 (BTW, isn't X\ supposed to work? I doesn't seem to) 16:30:34 -!- medgno has joined ##crawl-dev 16:30:50 tab is cycle through portals on X 16:30:52 it might play decently in practice 16:31:24 I like finite zigs because it gives a clear goal, and lets people say they've cleared a zig 16:31:37 not to deny that infinite ones sound interesting, though 16:31:59 we could just guarantee some number of zigs in pan, and not worry about people who want to clear 1348059 zigs in one game 16:32:23 I feel like the infinite zig concept has more appeal to people who have cleared several already (for the record, I've only ever done 2 zigs down to 27) 16:32:27 seriously, if those 81 levels contain 6 zigs or so, I think it would be fine 16:32:47 elliptic: oh, also, there's the idea of marking Pan transits: you are told the name of the owner of that realm 16:32:55 well then it could get unbelievably tedious to find the last zig, or whatever 16:33:14 print a message when a zig is on the level 16:33:24 MarvinPA: well, don't guarantee an exact number... just usually have around that many 16:33:29 there's precident with bazaars and labyrinths, although zigs are untimed 16:33:41 MarvinPA: could be fudged the same way named Pan lords could be 16:33:42 oh, it would also be good to remove the gold cost of zigs 16:33:51 I had some idea of coming up with other possible costs 16:34:00 elliptic: it stops people entering D:6 zigs 16:34:08 kilobyte: I mean for pan zigs 16:34:20 of course, no one cares for Zigs spawning on D:6 16:34:23 <|amethyst> FR: descend to pan:81 with the orb to win (after fighting a panlord based on the last player who won this way) 16:35:23 |amethyst: hah, sounds more interesting than my idea of copying DoomRL's Ao100 directly :p 16:36:15 <|amethyst> and now I'm off for real, until tomorrow. kilobyte: I think using lua_new_userdata() to copy the stash item_defs should work... probably want to lift the items-at-pos code from full_describe_view into a new function and use that 16:36:29 a player pan lord boss would probably be kind of easy 16:36:42 <|amethyst> multiply the HP etc by 3 :) 16:37:08 <|amethyst> I can try to implement the stash thing tomorrow, unless you want to do it first 16:37:11 would also have to forbid followers of good gods from being made into pan lords... they could become some kind of A though 16:37:11 <|amethyst> anyway, later 16:37:23 |amethyst: bye! 16:37:54 steal an idea from ancient crawl and have the player fight avatars of various gods 16:38:00 (cerebov used to be okawaru, etc...) 16:38:14 -!- Dixbert has quit [Ping timeout: 250 seconds] 16:39:06 -!- MarvinPA_ has joined ##crawl-dev 16:39:33 Kikugloorxaqudgha 16:39:44 evilmike: that still blows my mind that crawl used to have that 16:39:45 -!- elliptic has quit [Ping timeout: 246 seconds] 16:39:46 (she had his level) 16:39:51 evilmike: were they avatars of some sort, or the gods themselves? 16:40:05 Here: Cerebov wielding the -1 club of Okawaru {god gift} 16:40:11 G-Flex: regular Pan lords with the names currently used for gods 16:40:13 G-Flex: given that this is from 10+ year old versions, I don't think anyone can say 16:40:19 st_: <3 16:40:30 I interpret it as avatars of gods, because killing "okawaru" for good is a bit... excessive 16:40:33 -!- elliptic has joined ##crawl-dev 16:40:41 st_: and wearing a -2 animal skin 16:40:45 wait no 16:40:50 -!- MarvinPA has quit [Ping timeout: 250 seconds] 16:40:50 -!- MarvinPA_ is now known as MarvinPA 16:41:00 the -2 animal skin of existentialism {-1 Int, +MUT, god gift} 16:42:25 hah, can't ever browse those sources without laughing at "Lugafu the Hairy" 16:42:28 poor Lucy 16:43:52 https://gitorious.org/crawl/crawl-ancient/blobs/ancient-2.72/religion.cc 16:45:13 yeah, Okawaru is both a good and a Pan lord you can kill 16:45:53 "the Hairy"? 16:46:03 -!- Blade has joined ##crawl-dev 16:46:15 there's not much I was ever able to dig up about Lugafu, kind of looks like a prototypical trog though 16:46:43 I want to worship Illegal God 16:48:08 beh, you can't ever change gods there 16:50:48 -!- Blade has quit [Ping timeout: 272 seconds] 16:52:01 2.80 is the first version that has altars, Lugafu is removed in all but one place 16:52:26 we need a beard god 16:52:36 does it still have that big commented out portion of abyss.cc? 16:53:18 -!- Blade- has quit [Quit: Page closed] 16:55:16 oh, there isn't even in abyss.cc in the version there 16:56:21 kilobyte: how did gods work before altars then? 16:58:38 Kal (L14 OpSu) (Snake:1) 17:00:30 -!- Ragdoll has quit [Quit: Ragdoll] 17:01:29 G-Flex: not at all, they were dead code 17:05:49 oh, okay 17:30:35 -!- Pingas has quit [Read error: Connection reset by peer] 17:32:13 "You kill the azure jelly! Trog accepts your kill. Trog appreciates your killing of a magic user." 17:32:57 -!- elliptic has quit [Ping timeout: 246 seconds] 17:34:11 could it have been a shapeshifter? 17:35:12 possibly 17:35:48 shapeshifters are magical? i've never noticed him award me for killing others. 17:35:57