00:08:54 shouldn't we run the 0.8 branch on cao/cdo instead of trunk, for the time being? 00:12:22 probably would be a good idea 00:12:32 0.9?????? 00:13:41 -!- ZChris13 has quit [Read error: Connection reset by peer] 00:14:05 -!- ZChris13 has joined ##crawl-dev 00:16:19 -!- valrus has quit [Remote host closed the connection] 00:25:46 also, I tagged 0.9.0-a0 just yesterday, so all builds in the meantime had confusing version numbers 00:26:18 you had to use the hash part to determine which branch it is from 00:38:05 due: do you remember that Windows crash dump tool you investigated a year or so ago? 00:41:35 argh, kinda 00:41:59 well, I was around when it was discussed I don't remember if I actually investigated it or not 00:42:10 let me check my logs 00:42:58 I think it was Matthew Cline 00:43:01 ahah! 00:43:13 https://crawl.develz.org/mantis/view.php?id=98 00:43:47 -!- Twilight-1 has quit [Quit: ...those special autumn days, when the air smells like smoke, and the twilights are orange and ash gray...] 00:44:33 kilobyte: was that it 00:44:36 *was that it? 00:46:05 possible, I don't know the issue 00:46:15 heck, I was sure it is your fault! :p 00:46:39 03kilobyte * r66798554a9ec 10/crawl-ref/source/ (3 files in 3 dirs): Denzi's tile for pressure plate trap triggers. 00:46:51 i know nothing! 00:46:51 03kilobyte * r219b0bdeb457 10/crawl-ref/source/fight.cc: Use the same message for bashing with a ranged weapon as for bashing with non-weapons. 00:46:53 03kilobyte * rb69b35e2d62a 10/crawl-ref/source/rng.cc: Don't read 28 bytes of randomness into a 16 byte array. 00:46:54 03kilobyte * rbbd3c2531c99 10/crawl-ref/source/rltiles/ (25 files in 3 dirs): Move all trap tiles into a dir of their own. 00:46:58 * due proclaims innocence. 00:48:21 -!- Twilight-1 has joined ##crawl-dev 00:48:21 kilobyte: i did do the windows alert dialog with assertion messages though 00:48:27 kilobyte: which is kinda related? 00:49:11 next to stack dump in the code, yeah 00:49:33 stack dumps are a little beyond me -- at least with C. 00:49:51 i understand the mechanics but i've no idea how one wouldinteract with the stack to determine the reelvant information in C/C++. 00:52:23 the debug info includes information that function X's code starts from address 0x1234 to 0x1256 00:53:13 so if EIP on the stack points to anything in that range, you know that this stack frame was inside that function 00:53:57 line numbers work by the same scheme, they just take substantially more disk space 00:56:58 moin 00:57:36 oh hi Zaba! 00:58:10 what's up 00:58:11 what do you think about layout_bigger_room in Crypt? I'd disable it. 00:58:32 does it have pools of water? 00:58:33 the bug report was about it producing live trees, but pools don't look that hot there too 00:58:55 yeah, it probably should be renamed many_pools like in its code 00:59:19 it would be nice to have a crypt special case for many_pools' pool material, even if bigger_room is disabled in it, though 00:59:26 because it can still use used on builder_basic levels 00:59:36 be used* 00:59:40 ah 01:00:10 (a special case akin to those for gehenna/cocytus. Not sure what material would be fitting in crypt, though. Tombstones? :>) 01:01:45 graveyards would make a lot of sense, but I doubt they can be done well with pools 01:01:59 yeah, it'd be ugly and hacky. 01:02:47 kilobyte, _build_river/_build_lake return early when called in crypt and tomb, by the way 01:02:52 maybe _many_pools could do the same 01:03:01 right 01:03:14 I didn't notice it's called outside bigger_room, too 01:03:30 should bigger_room be generated without the pools? 01:03:36 no 01:04:01 but many_pools should generate nothing in crypt 01:04:32 we could do gravestones, but to do that without nasty hacks it would have to be a feature on its own rather than a redefined statue 01:05:39 after many_pools is made to generate nothing in crypt, it makes perfect sense to disable it entirely in crypt.. because just an open level is boring 01:07:52 right, graveyards can be regular vaults 01:08:09 the whole _builder_extras should be somehow refactored at some point 01:08:41 (_build_{lake,river}, _many_pools) 01:18:52 kilobyte, I'll do the special case/disabling of bigger_room, then 01:18:58 (unless you're already doing it..>) 01:20:14 -!- ortoslon has joined ##crawl-dev 01:21:43 you know the matter so much better than me 01:28:26 03zaba * r200e02be736e 10/crawl-ref/source/dungeon.cc: Disable _many_pools in Crypt. 01:28:36 03zaba * raf5ecec0b93c 10/crawl-ref/source/dat/des/builder/layout.des: Don't generate bigger_room levels in Crypt. 01:29:02 kilobyte, this is supposed to fix some bug, right? 01:29:04 -!- ais523 has joined ##crawl-dev 01:39:57 https://crawl.develz.org/mantis/view.php?id=3769 01:42:32 ok 02:12:07 -!- ZChris13 has quit [Quit: ChatZilla 0.9.86.1 [Firefox 4.0/20110318052756]] 02:16:54 -!- valrus has joined ##crawl-dev 02:54:23 -!- dexap is now known as paxed 02:57:09 how are chaos knights changing in 0.8? 03:01:15 they always worship xom 03:01:47 CK of lugonu was split off into a seperate class as abyssal knight, and makhleb isn't available as a starting god anymore 03:03:13 -!- ortoslon has quit [Read error: Connection reset by peer] 03:05:40 ah. I was afraid you'd be dropping Lugonu as a starting god. 03:05:43 That would make me sad 03:07:46 Why the Makhleb change? Is he too powerful/vanilla to have as a starting god? 03:09:04 he doesn't really add a lot to your playstyle early on so it's fairly uninteresting/unnecessary 03:09:32 what about Deep Dwarves? 03:09:35 trog, yred, zin all have abilities that you basically use straight away on d:1 and define your playstyle, makhleb doesn't really have that 03:10:06 deep dwarves start with a wand of healing :P 03:10:44 -!- monqy has quit [Quit: hello] 03:21:27 megabats? I haven't played in a long time... 03:29:25 bmh: renamed giant bats 03:30:21 ghallberg: I gathered 03:41:40 -!- syllogism has joined ##crawl-dev 03:48:46 -!- casmith789 has joined ##crawl-dev 03:52:36 !tell galehar Ok, right now it works like this: If you try to butcher with no sharp weapons in inv it asks "Switch to pen knife for butchering?", if you have sharp weapons you get "What would you like to use? (- for pen knife, esc to cancel)?" and the regular menu for choosing chopping-tool. 03:52:37 ghallberg: OK, I'll let galehar know. 04:34:45 -!- Twilight-1 has quit [*.net *.split] 04:35:41 -!- Twilight-1 has joined ##crawl-dev 04:49:47 -!- Zaba has quit [Ping timeout: 276 seconds] 04:57:42 -!- Zaba has joined ##crawl-dev 05:18:39 -!- eith has joined ##crawl-dev 05:38:06 -!- jpeg has joined ##crawl-dev 05:43:24 -!- Ashenzari has quit [Ping timeout: 264 seconds] 05:44:10 -!- Ashenzari has joined ##crawl-dev 05:44:10 -!- The topic of ##crawl-dev is: Logs: http://tozt.net/crawl | People with +v have commit access. | Please keep general crawl-related chit-chat to ##crawl. | Dev wiki: http://crawl.develz.org/wiki | Long stuff to pastebin.ca please | Immortal Warwalrus and Crazy Yiuf forever. | Tentative 0.8 release date April 24th 06:02:44 03j-p-e-g 07stone_soup-0.8 * reecabb2fbcc2 10/crawl-ref/source/cloud.cc: Fix #3156, #3709: cloud safety checks getting randomized AC reductions. 06:02:56 03kilobyte 07stone_soup-0.8 * rb2a27bfbfc9d 10/crawl-ref/source/rltiles/ (25 files in 3 dirs): Move all trap tiles into a dir of their own. 06:02:57 03kilobyte 07stone_soup-0.8 * r30a3f24817af 10/crawl-ref/source/fight.cc: Fix missing checks for players bashing with bows, monsters with rods. 06:02:58 03zaba 07stone_soup-0.8 * r553ebf841590 10/crawl-ref/source/dungeon.cc: Disable _many_pools in Crypt. 06:02:58 03kilobyte 07stone_soup-0.8 * r66ad99461831 10/crawl-ref/source/ (3 files in 3 dirs): Denzi's tile for pressure plate trap triggers. 06:02:59 03kilobyte 07stone_soup-0.8 * rcb4cb6d19a2f 10/crawl-ref/source/fight.cc: Use the same message for bashing with a ranged weapon as for bashing with non-weapons. 06:03:00 03zaba 07stone_soup-0.8 * ra02c1e396ebb 10/crawl-ref/source/dat/des/builder/layout.des: Don't generate bigger_room levels in Crypt. 06:10:33 -!- upsy has joined ##crawl-dev 06:14:24 -!- valrus_ has joined ##crawl-dev 06:27:26 Ragdoll (L16 MuBe) (Vault:1) 06:32:31 -!- edlothiol has joined ##crawl-dev 06:46:10 -!- valrus_ has quit [Remote host closed the connection] 07:03:05 Anyone around? 07:06:42 hmm, what do penknives do? 07:06:48 butcher slower, or something? 07:08:15 ais523: Nah, it's prettymuch emptyhanded butchering. 07:08:28 ais523: Fallback if you don't have a sharp weapon. 07:08:31 why would you ever butcher with anything else, then? 07:08:48 Well you wouldn't. 07:09:17 That's kinda the point, removing the tedium of looking for a knife in the early game. 07:09:22 so why is there even an option to do so? 07:09:43 Because you might have a cursed dagger or sth in your hand 07:10:04 Then you can use that to butcher but you couldn't if it was a cursed mace. 07:19:02 -!- ortoslon has joined ##crawl-dev 07:47:15 -!- Mu_ has joined ##crawl-dev 07:58:32 03j-p-e-g * r440b1b0876ef 10/crawl-ref/source/output.cc: Colour unavailable quiver slot darkgrey. 07:58:43 03j-p-e-g * r7d9ed24700f8 10/crawl-ref/source/output.cc: Skip Qv line for Felids. 08:25:03 -!- Cryp71c has joined ##crawl-dev 08:30:07 -!- valrus has quit [Read error: Connection timed out] 08:32:44 -!- valrus has joined ##crawl-dev 08:38:00 -!- ortoslon has quit [Read error: Connection reset by peer] 09:06:29 -!- valrus has quit [Read error: Connection timed out] 09:07:57 -!- valrus has joined ##crawl-dev 09:22:16 -!- bmh has quit [Quit: bmh] 09:43:59 -!- ZChris13 has joined ##crawl-dev 10:28:29 03galehar * r2bf5c1088a77 10/crawl-ref/source/ (main.cc player.cc player.h transform.cc): Prompt the player before moving to a deadly cell with an expiring status. 10:30:28 !tell galehar When you get back, drop me a note on how to convey my changes to you for checking. 10:30:29 ghallberg: OK, I'll let galehar know. 10:38:27 thought A: we should really just permit ranged stabbing, even if it's worse; it'd make the skill more relevant on e.g. hunters 10:39:11 thought B: i wonder whether Stealth skill should affect noise from fighting/casting, afaik it only affects your odds of being spotted right now and has no impact on noise 10:41:59 -!- valrus has quit [Read error: Connection timed out] 10:42:24 03galehar * r867ad5103757 10/crawl-ref/source/main.cc: Fix verb with the new confirmation prompt whith expiring status. 10:43:09 -!- valrus has joined ##crawl-dev 10:44:14 -!- casmith789 is now known as camith789 10:44:19 -!- camith789 is now known as casmith789 10:44:34 thought C: what if we exposed Ashenzari tension tiers to the player? this would be useful in A) a general /consider type of function B) prompts (e.g., default of --more-- when something REALLY dangerous comes into view C) scripting (e.g., ignore trivial monsters for autopickup) 11:00:49 Crawl.exe crashes almost instantly (https://crawl.develz.org/mantis/view.php?id=3770) by XuaXua 11:10:53 Miscellaneous Lava Orc Bugs (https://crawl.develz.org/mantis/view.php?id=3771) by casmith789 11:18:10 Do we really want to use Mantis for lava orc bugs 11:18:24 greensnark, I wonder, too 11:18:32 they're -not- in mainline crawl, right?... 11:18:41 I don't know, I'm pulling now 11:19:23 Eronarn told me to 11:19:38 Some of these obsolete branches should be dropped from the repo: I think articulated kraken is already in 11:19:50 if it's not in mainline crawl, it can be easily closed as invalid :P 11:20:02 Looks like someone merged octopi into master 11:20:03 hey lava orcs are in .9 development clearly 11:20:10 -!- ZChris13 has quit [Quit: ChatZilla 0.9.86.1 [Firefox 4.0/20110318052756]] 11:20:43 but I don't care much unless there appears a lava orc branch or something 11:20:45 Whoever's merging new species into the official repo should really get dpeg to sign off first 11:21:18 I can stop reporting lava orc bugs on mantis if you want 11:21:47 casmith789: Thanks. I know Eronarn needs some way to collect bug reports, but using Mantis is just going to tick off dpeg 11:22:13 Unless we can get a new product on Mantis 11:22:19 oh apparently I can't close my own bug report 11:23:36 * octopus is a branch in gitorious, not in master, sorry 11:23:54 they are interesting earlygame 11:24:08 because of 0ac, they take ridiculous damage from everything 11:24:12 Nice 11:25:02 Just for the record: I'm not pro or anti LO/Oct, I put them on CAO so Eronarn can get feedback, etc. and eventually the team can decide if the new species look good 11:26:03 -!- edlothiol has quit [Read error: Operation timed out] 11:28:11 well HO are bland right now, LO seems to differentiate them slightly, although they are clearly overpowered and octopodes are unbalanced 11:28:34 LO can worship beogh, right? 11:28:42 yes but not well 11:28:47 why not? 11:28:53 when they get fired up they kill the orcs 11:28:59 do they burn up allies or something 11:29:04 heh. 11:29:04 apparently Eronarn is going to code beogh protection 11:29:49 !hs * lo god=beogh 11:29:49 16. Pseudonut the Cleaver (L10 LOPr), worshipper of Beogh, slain by an orc warrior (a +0,-1 glaive of reaching) on D:11 on 2011-04-08, with 8879 points after 11195 turns and 0:42:43. 11:31:25 are there going to be lava orc monsters? :P 11:31:39 apparently so 11:35:52 casmith789: they must be at least dummies because of show_player_species 11:36:20 03j-p-e-g * r4e4c5b93a483 10/crawl-ref/source/tilepick.cc: Fix #3764: Replace an assert with some more sensible handling. 11:36:23 which doesn't work properly so far, they have some elemental colour or something which shows as green 11:37:50 well, that's because they are not coded, at least on in the version I looked at last 11:38:21 something should be done about Shuffle, in my opinion. Even if it doesn't necessarily hurt the character all that much, it's still basically permanent and not fun at all 11:38:45 I guess if you are not as greedy as I am, you can just triple draw :P 11:39:32 you mean, trolls with 40 int? 11:39:54 not sure if that card is a good idea at all 11:39:59 that's a fun gimmick but not really feasible anymore as shuffles are rarer than they used to be 11:40:26 if a caster shuffles int, you just kind of lose the will to play 11:40:34 especially if you aren't intending to do a quick 3 runer 11:42:40 my MDFi drew it, losing all str, forcing me to use PDA+Summon Dragon instead of CPM :p 11:44:12 as a non-Nemelex char, I scummed early Zigs for decks, without luck 11:44:19 greensnark: the world is coming to an end 11:44:46 greensnark: we found a japanese guy who did a xml savefile branch of jnethack 11:44:47 but then, a weaker char would indeed have to indeed go Ashenzari ASAP or #quit 11:46:17 bhaak: not that hard, actually. I once considered writing a text version of saves for NetHack too (not XML, I wouldn't insult the devil calling XML his work) 11:48:21 kilobyte: in principle, it's not hard. but this guy gets bonus points for doing also an perl parser of the structs to dynamically add new bits and flags 11:50:17 NetHack saves are pretty small, though 11:50:42 Crawl tends to be tens of megabytes uncompressed 11:51:04 Typo causes keysym crash in windowmanager-sdl (https://crawl.develz.org/mantis/view.php?id=3772) by jpeg 11:51:08 the structures are way more complex, too 11:52:07 Windows development builds on CDO updated to: 0.9.0-a0-36-g4e4c5b9 11:52:07 -!- jpeg has joined ##crawl-dev 11:52:55 hmm, does crawl use objects and inheritance for it's internal structure? 11:55:49 kilobyte: it appears all the unicode bugs are fixed now (we are still running 0.8 on CDO, right?) 11:55:55 greensnark: i'm fine with people using github, or my email, to get in touch instead of mantis. i've been telling them that any of the above are fine but i can drop the mantis part if the bugs there are annoying 11:56:06 kilobyte: so it's probably time to cherry-pick the changes 11:56:08 bhakk: not really, there's a crapload of structs though 11:56:28 jpeg: 0.9, there was some confusion about branches and it's still not switched 11:56:39 kilobyte: are there any among your recent related commits that should *not* go into 0.8? 11:57:12 mostly because I forgot to tag something as 0.9.foo so both branches had "git describe" say 0.8.foo 11:57:57 even greensnark when he did the same error for 0.7 caught it in two days 11:57:59 tagging 0.8 would have been my job, I guess, but I completely forgot about it 11:58:35 also would you be willing to take over as release manage when I quit the team? 11:58:57 jpeg: I don't think there is a reason not to, they are not vital though, since Unicode texts are quite unlikely to happen 12:00:10 possibly 12:00:32 so all those formatted and tagged string changes don't need to go into 0.8? 12:01:00 lemme check if there's string truncation, this one should go if it's not already in 12:02:31 word wrapping causes far more benign errors: since number of bytes is never smaller than the width, visible errors will be: 1. wrapping too soon, and 2. sometimes showing a broken character at the end and the beginning of a new line 12:02:54 Eronarn: github tracker or get Napking to add a new project or game version on Mantis 12:05:32 also, galehar proposed disabling ZotDef for 0.8, since it's too buggy (no one bothered to fix it...) 12:07:14 jpeg: quit the team?? 12:08:28 I start a job the week after next 12:08:53 so? ;) 12:09:01 we all have jobs :) 12:09:02 it's a game development company, and the contract forbids working for the "competition", unpaid or not 12:09:05 (hopefully) 12:09:21 arg.. that's sucky 12:09:40 I can still come here to chat with my friends, but I'd rather not do anything that could get me into trouble 12:10:03 jpeg: Congrats, the job sounds like fun :) 12:10:06 i understand 12:10:38 greensnark: it does, hope it'll be :) 12:10:39 Focussing on the job is a good idea anyway, Crawl can wait 12:10:49 sure 12:10:52 there have been many law suits about what is "competition" and what contracts are allowed to prohibit you 12:11:44 but of course, greensnark is right :) 12:12:01 greensnark: didn't you have a similar reason to stop being listed as a team member? 12:12:10 even though you still occasionally fix bugs? 12:12:29 Yes, my reasons for dropping off the team are similar :) 12:12:41 Although I do not have a no-compete clause for open-source stuff 12:12:48 Because I don't work in game dev 12:13:25 yes, they do have the usual "you should relax in your free time" clauses, but all that's fairly benign 12:13:37 it's really just a problem with game development 12:13:46 which is all the coding I ever do in my free time 12:13:57 so I guess I'll have to go out and get a life ;) 12:14:03 heresy! 12:14:19 (I know, I should get a life too...) 12:14:40 it's the end of an era... *sigh* 12:14:59 (heretic thoughts spreading through the channel at the speed of gossip!) 12:15:26 but then, watching my young sis serve a young terrorist all the time kind of discourages anything of that kind 12:17:41 bad relationship? 12:18:00 no, crying kid 12:18:12 hehehe 12:18:23 :) 12:18:36 11 months old, and she has no shred of time for anything 12:19:45 only one side of the medal, kilobyte :) 12:20:02 jpeg: congrats on the job btw. But i'll only be able to buy your games if you do a console version. you can't expect me to boot into windows :) 12:20:34 sorry, Windows and XBox only :P 12:20:43 I'll have to get used to Visual Studio 12:20:43 pfui! ;) 12:20:53 kilobyte: i've seen something similar with my sister and my mother, but i've seen something completely different with my sister-in-law. it depends on the mother :) 12:20:55 hehe - you'll hate it ;) 12:21:13 Napkin: I'll come here whenever I need to complain :) 12:21:23 most welcome! :) 12:21:34 lol, so you don't get a salary but a compensation of mental damages :) 12:21:42 well put :p 12:21:45 and those contracts sound more harsh than they are lived in reality 12:22:00 -!- monqy has joined ##crawl-dev 12:22:02 well, I'd rather err on the side of caution 12:22:15 start out with caution, of course :) 12:22:25 bhaak: also thanks! :D 12:22:40 do you want me to change the credits entries similar to your brother? 12:23:12 Napkin: I'd like to be listed in the 0.8 credits as a developer (that one week doesn't count) 12:23:31 but afterwards, I'll have to be shunted down to the list of contributors 12:23:34 oh, there i have no write permissions 12:23:41 i was talking about the wordpress 12:23:45 just saying 12:23:48 yeah 12:24:18 sorry, Crawl is like Alcoholics Anonymous, there's no way out 12:24:28 sshh! ;> 12:24:44 Napkin: no, best to remove me completely, just in case they check (Crawl was in my references, after all) 12:24:47 * Napkin grins at greensnark * 12:25:28 also, once I've got a healthy dose of 3d graphics, I don't think I can come back to tiles ;) 12:25:49 the list of contributers is timeless 12:26:11 then you'll fall back to ascii :) that can't be beaten by 3D :) 12:26:18 3d ascii 12:27:02 kilobyte: the truncation fix probably is dba9c785, but there are some other commits where you reference 0.8 12:27:25 only they all seem to be interlocked 12:28:53 especially the word wrapper was followed by a load of bug fixes 12:29:03 Eronarn, greensnark: a new version would probably be better than a separate project? 12:29:51 ah, I checked truncation of inscriptions which was already applied to 0.8 12:30:18 Eronarn, greensnark: but there will be a permissions issue.. 12:31:18 two choices: 1. cherry-picking all the commits, 2. blocking all non-ASCII character names, notes and inscriptions 12:31:59 Napkin: New version sounds fine to me 12:32:10 Can't help with the permissions issue, unfortunately :P 12:32:12 I think the former can be safe enough, though 12:33:09 just thinking.. subproject maybe? but the "report issue" page is highly modified for DCSS project only 12:33:35 -!- MakMorn has quit [Remote host closed the connection] 12:33:40 greensnark: well, let's see what is the definition of a "dev"... 12:34:49 kilobyte: okay 12:36:09 -!- MakMorn has joined ##crawl-dev 12:38:50 the impact of these bugs is pretty small, so even inaction wouldn't cause world to explode 12:39:09 well, the corruption issue does need fixing 12:40:04 unless we explicitly forbid non-ascii character names 12:41:09 ah right 12:45:02 jpeg: on Windows, what does Alt-Delete usually do? 12:45:33 no idea, doesn't trigger anything for me 12:45:44 (the crash there happens only for the "grey" delete, not the keypad one) 12:46:14 ah, I don't have a numpad, so that hadn't occured to me 12:49:59 greensnark: heh, actually, I think having it as a subproject or even separate project will just totally confuse people and bug reports will land in the wrong project definitely ;D 12:51:14 the default bug report points to dcss, I guess 12:51:28 a separate version might be easier for people to change 12:51:32 oh.. i'm at a point of creating a new version for 0.9... which means "longterm planning (0.9+)" needs to be increased... 1.0 or 0.10? :D 12:51:47 (though that wouldn't filter it out from the main view) 12:52:05 Napkin: heh, that's the big question 12:52:17 not sure there's anything special planned for the new version after 0.9 12:52:22 there is no default project unless you made it. new users will be asked where to report to and "mark as default?" - existing users will not realize the new project at all 12:52:24 how about we skip the issue and start versioning with letters? 12:52:39 hahaha 12:52:49 27 > 9 :D 12:53:12 0.27 sounds smaller than 0.3 12:53:17 10 > 9 12:53:52 all version number parsers I know consider the version string to be numbers rather than decimals 12:54:23 1.0 it is? 12:55:13 kilobyte's argument would make it 0.10 ... 12:55:30 you know, I think dpeg left on purpose, precisely to avoid this discussion :P 12:55:38 :D 12:55:38 I checked: Debian's, GNU's, Microsoft's 12:55:57 0.99.. 12:56:02 some programs have gone from 0.9 to 0.10 12:56:28 it shouldn't be 1.0, anyway, that implies a rewrite or a done-with-development 12:56:36 yeah 12:56:36 you could always go up to 0.91... 12:56:48 kernel 2.6.10 was after 2.6.9 ... 12:57:28 in what decade do you expect to hit 1.0? ;-) 12:57:38 well, people generally raise the major version when something is achieved - not because it's the next 12:57:48 Microsoft's version syntax is simplest and most rigid: there are always four numbers, all within 0..65535 12:58:01 you probably could have raised the major version at least twice :) 12:58:09 agreed, bhaak 12:58:28 you could do it the mozilla way, throw out 3 major versions in this year 12:58:37 yeah, once for tiles, and once around 0.6 or so 12:59:46 Napkin: use 0.10, it'll be easy to change if we decide on 1.0, after all 13:00:49 default_fixed_in_version = 0.9 development branch 13:01:17 - longterm planning (0.10+) 13:01:24 - 0.9 development branch 13:01:32 - 0.8 prerelease branch 13:01:37 - 0.7 stable branch 13:01:47 --- 13:02:04 sounds good 13:02:32 the stats are a bit wrong now though - we should have created 0.9 at the same time we started 0.8 for cherry picking (created the branch) 13:03:06 why? 13:03:09 -> everything up until 0.9 was tagged is "fixed in 0.8" according to mantis 13:03:43 sounds about right 13:03:53 where in reality there was already the branch before that 13:04:07 but - again - i didn't think about it 13:04:22 all the new bugs don't appear in 0.8 because the commits introducing them haven't been cherry-picked 13:05:04 well, a bug report closed yesterday is marked with "fixed in 0.8" - but that bug didn't even existed yet in 0.8 13:05:04 to the player, all that matters is whether a given problem happens in 0.8 or not 13:05:16 of course - but it's interesting to you 13:05:37 -!- TGWi has joined ##crawl-dev 13:05:40 we could update the fixed version if it really bothers you :) 13:05:46 the version after 0.9 is 0.A 13:05:47 to you as release manager 13:05:50 pretty sure that's how it works 13:06:10 well, if it doesn't bother you - good :) 13:07:17 ah, you mean the 'change log'? 13:07:25 not only that - also the search 13:07:27 i've never been able to work out how to use the statistics 13:07:59 you could have told mantis.. "show me all fixed in 0.8" and then cherry pick from it 13:08:09 *you could have told mantis.. "show me all fixed in 0.9" and then cherry pick from it 13:08:29 -!- TGWi has left ##crawl-dev 13:10:18 anyways, you can still do that now 13:10:54 damn, no power (on phone currently) 13:11:37 jpeg: can't fix the keysym bug right now but it should be safe to ignore the key 13:12:07 so just quieten the assert? 13:12:17 I was reading SDL docs before power died, and didn't find the cause 13:12:19 it's not urgent, you can fix it when you have the time 13:12:29 03j-p-e-g 07stone_soup-0.8 * r47c4b3907f9c 10/crawl-ref/source/tilepick.cc: Fix #3764: Replace an assert with some more sensible handling. 13:12:29 I'd return from the function 13:12:31 03galehar 07stone_soup-0.8 * r594915b5d592 10/crawl-ref/source/ (travel.cc travel.h): Fix travel failing to start if standing right next to a cloud. 13:12:31 03kilobyte 07stone_soup-0.8 * rcc5cac1701fc 10/crawl-ref/source/ (output.cc spl-book.cc viewmap.cc): Text truncation fixes. 13:15:43 Napkin: I see I've been using the Mantis search wrong all along - mostly looking at open issues etc. ;) 13:16:32 ;D 13:21:19 the severities set are totally random 13:22:54 one bug I'd care about the most is the one where if a db rebuild is interrupted, games will break in mysterious way until you delete saves/{db,des} 13:24:01 for the db, I'd store a timestamp in the database itself instead of statting the file 13:24:10 no idea about maps 13:24:17 it's currently marked as 'block', so not exactly low-priority 13:24:59 though I agree that a lot of the minor/major issues could easily be put into the other category 13:25:07 most other bugs are random, though 13:25:44 on slow machines db rebuild can take ages, too 13:26:07 also, we appear to have a lot 'Crawl won't start, crashes immediately' types of reports 13:26:32 how often does the db get rebuilt? 13:27:31 nearly a minute on my old machine on XP, much faster on that box on Linux... >20 mins(!) on win7 over SMB on a fast machine (XP got that ok) 13:27:51 jpeg: after every upgrade that changes the timestamps 13:28:26 I think the packaging does cp instead of cp -p, too 13:28:32 well, for me it is much faster than the actual compilation, which is why I probably don't really notice 13:28:55 hey, now I've got a job I can finally buy a new computer! 13:29:15 (it's not that easy, we'd have to use last commit since git doesn't set mtimes) 13:30:02 well, on a fast machine you don't notice it really, unless it's a network drive 13:30:55 people sometimes think the game is hung since there's no progress bar 13:36:21 at least the tiles splash screen hints at loading progress 13:36:29 yeah 13:36:41 in console there's just a cursor 13:37:10 we could list the file that's being worked on 13:37:17 reticulating splines 13:48:42 -!- hayenne has quit [Ping timeout: 248 seconds] 13:54:34 -!- Jordan7hm has quit [Ping timeout: 248 seconds] 13:54:41 -!- eith has quit [Ping timeout: 252 seconds] 13:58:20 03j-p-e-g * r0b625a5a8587 10/crawl-ref/source/windowmanager-sdl.cc: Fix #3772: Don't "die" if an invalid key symbol is entered. 14:00:13 03j-p-e-g 07stone_soup-0.8 * r22a73bfc7c8f 10/crawl-ref/source/windowmanager-sdl.cc: Fix #3772: Don't "die" if an invalid key symbol is entered. 14:00:31 -!- casmith789 has left ##crawl-dev 14:01:56 -!- jpeg has quit [] 14:02:13 -!- Jordan7hm has joined ##crawl-dev 14:11:51 -!- bmh has joined ##crawl-dev 15:07:51 would there be any interest in a vault full of clouds (benign and dangerous) instead of walls? 15:20:00 cloud generators are already overused IMO 15:21:33 -!- valrus has quit [Read error: Connection timed out] 15:23:14 -!- valrus has joined ##crawl-dev 15:26:34 bmh: i proposed a portal vault like that a long while ago with amnesia-clouds and a wrapping map 15:43:02 Eronarn: did you do anything that would alter the m screen in erocrawl? not sure if this bug is an erocrawl thing or not 15:45:08 -!- hayenne has joined ##crawl-dev 15:46:43 st_: are you using show_gold_real_turns or whatever 15:46:43 Eronarn: You have 14 messages. Use !messages to read them. 15:46:45 because that's broken 15:47:00 oh wait, m screen 15:47:15 ican't think of anything i did to the m screen, but that doesn't mean i didn't accidentally do something 15:47:19 two of my apts have disappeared on it, not sure why though 15:47:54 i add an extra space on the main screen for LO temperature, so if one line is missing, that would be one place to investigate 15:48:24 probably don't report it unless it showsu p in trunk though 15:51:27 do we really need to still show the conversion from the new aptitude scale to the old 0.5 one on the apts screen? 15:51:33 or 0.6, whenever it was 15:51:50 I thought it was quite recent, maybe even 0.7 15:51:59 oh, so 0.6 before the change 15:52:05 yeah 16:00:36 03MarvinPA * r86f4b0ab2602 10/crawl-ref/ (docs/template/apt-tmpl.txt source/spl-cast.cc): Rename some casting/aptitude adjectives for clarity 16:03:27 -!- DrPraetor has joined ##crawl-dev 16:03:51 Some trunkments. The improved AI makes the late-early/early-mid game, already not a cake walk for many builds, significantly harder. 16:03:57 Eronarn: how would you distinguish such a thing from the Abyss? 16:04:39 The autoexplore needs to notice when you encounter an unusual wall-type... or, preferably, we should lose the special vaults which are enclosed in secret doors behind funny-colored walls. 16:04:58 (there's a trove vault that works this way, and a mausoleum vault... probably others) 16:05:13 there was some discussion long ago about having different door types 16:05:22 DrPraetor: the way i favor this is to have a marker that can be added that stops autoexplore 16:05:46 Ernoarn - find, but at that rate, why have the secret door at all? It's just a waste of people's time. 16:05:48 that way it can be added for anything where the vaultbuilder wants to say 'hey, this is interesting' 16:06:02 e.g., bloodstains 16:06:04 Eronarn - I agree we should have such a thing. 16:06:26 as for the secret door - secret doors are not great but that's not really any worse than any other application of them 16:06:27 BUT that just reverses the problem; if you blunder into such a thing with autoexplore OFF, will you notice? 16:07:06 Myself, I would be happy to do away with secret doors entirely. 16:07:18 Systematically replace them all with normal doors. 16:07:32 BUT, failing that, they cause a particular problem in these funny-walled vaults. 16:28:42 -!- bmh has quit [Quit: bmh] 17:12:24 -!- eith has joined ##crawl-dev 17:38:37 03MarvinPA * r129d0f2a6bff 10/crawl-ref/source/spl-transloc.cc: Don't avoid glow by cancelling semicontrolled blinks with enter instead of escape 17:41:21 -!- Cryp71c has quit [Read error: Connection reset by peer] 18:02:50 Deep qwarf damage reflection message is bad. (https://crawl.develz.org/mantis/view.php?id=3774) by elliptic 18:02:50 Manuals care whether a skill is turned on or off. (https://crawl.develz.org/mantis/view.php?id=3773) by elliptic 18:07:52 Trog's hand doesn't have a message when cast by monsters (https://crawl.develz.org/mantis/view.php?id=3775) by 78291 18:19:08 -!- syllogism has quit [] 18:37:22 -!- Kurper has joined ##crawl-dev 18:44:10 -!- casmith789 has joined ##crawl-dev 18:49:18 Has stab training been fixed in trunk yet? 19:05:12 Also, "Vehumet is extending the range of your destructive magics" is incorrect. Freeze *is* destructive magic (he boosts success chance), but Vehumet does not boost the range :(. 19:05:30 So either give Freeze (and, where apropos, summoning spells) a range boost, or fix the in-game text. 19:06:17 -!- casmith789 has quit [Remote host closed the connection] 19:07:23 i don't think any summoning spells have limited range 19:07:59 freeze is touch-range, not range 1 19:08:05 that could be made more clear i suppose 19:08:06 and touch spells aren't range-extended 19:08:09 but yeah, that 19:08:17 it's pretty clear from the targeting system though 19:12:19 03MarvinPA * r82a4c24d46b8 10/crawl-ref/source/dat/database/monspell.txt: Adjust Roxanne's casting messages, since she can't move 19:13:55 -!- ais523 has quit [Remote host closed the connection] 19:23:28 drpraetor: stab training? I wasn't aware there was a problem with it 19:24:09 elliptic: It trains when wielding weapons that give negligible bonus damage when stabbing (i.e. halberds). 19:24:17 so? 19:24:19 s/i.e./e.g. 19:24:32 dodging trains in GDA 19:24:39 this isn't a problem 19:24:53 Yes it is. It causes inverse-victory-dancing. 19:24:59 huh? 19:25:10 it means you'll get like 2-3 stabbing skill 19:25:15 which is completely negligible 19:25:23 You avoid hitting confused enemies because, if you do, you'll waste XP on stabbing. Which, on top of being a waste of XP, *raises the cost of all your other skills IT IS NOT NEGLIGIBLE*. 19:25:24 just wasting a tiny bit of xp 19:25:42 it seems fine to me 19:25:45 it doesn't raise the cost of all your other skills in any significant way at all 19:25:52 you probably don't understand how that works 19:26:07 i was an unarmed stabber on one of my recent characters, it was useful 19:26:27 cost of raising skills increases based on how much XP you've spent on skills 19:26:52 so if you've spent a negligible amount of XP (you have), then the effects are negligible too 19:26:56 so stop worrying about it 19:27:14 I don't think any of us, enclosed as we are in a mere three dimensions, can really be said to *understand* how skill training works, but: if you have a lot of skills, +1 stabbing can take a lot of XP. Once +1 stabbing takes a lot of XP, it adds a lot of XP to the amount-of-XP-you've spent and has a noticeable impact on the cost of your other skills. 19:28:08 I ran some simulations on the actual code at one point and, while it wasn't *huge*, it ended up costing you thousands of XP, basically for nothing. 19:28:45 if you're stabbing things, you're getting a damage bonus from it, so it's not like it's doing nothing 19:28:46 the first level of a skill costs very little, i really don't think it'd be significant even to the biggest minmaxer 19:28:50 +1 stabbing never takes a lot of XP, yeah 19:28:55 if you're not stabbing things, you won't be training it 19:29:01 -!- Mu_ has quit [Quit: Defecator, may everything turn out okay so that you can leave this place.] 19:30:08 I suppose I'll check out another copy of trunk and run it on the current code, but if you spend even a modest amount of time killing elves in meph clouds, it's hard to avoid a stabbing of 2 or 3. 19:30:19 i don't think the game should be saying "well, i don't think you *should* be stabbing with this weapon, so i'm not going to let you" 19:30:42 'thousands of XP' is nothing 19:31:03 yeah, thousands of XP is one slime creature pack, say 19:31:03 also, what eronarn said 19:31:12 -!- edlothiol has joined ##crawl-dev 19:31:26 doy: in that case, the stabbing bonus should be roughly the same for most weapons. Having *clubs* be an inneficient tool for *clubbing unconscious people to death* has never made a lot of sense to me. 19:31:53 i don't see how that follows at all 19:31:56 uh, clubs are an efficient tool for clubbing unconsious people to death 19:32:01 you need a club to club people 19:32:13 they aren't an efficient tool for stabbing people 19:32:44 clubs have their cute confusion-on-stab now too, not that it's hugely useful 19:33:27 Look, either the "stabbing" skill subsumes clubbing-to-death-while-unconscious, *or* it doesn't. If it does, than clubs should not be noticeably inferior to GREAT SWORDS when dealing with unconscious people. 19:34:02 why would a club be better than a great sword? hitting an unconscious person with a great sword sounds pretty good to me 19:34:19 tabbing[1/4]: A succesful stab multiplies damage done by a factor proportional to stabbing skill. Expect about double damage at 10 skill. Short and long blades get two multipliers, 3-4 times normal damage at 10 skill. Finally, short blades get a bonus of up to 30 added *before* multiplying. 19:34:33 ? 19:34:38 we know how stabbing works 19:34:52 SO. A great sword gets a *proportionally larger bonus* than, say, a great axe or a club. 19:35:02 it does, yes... and what's the problem? 19:35:21 swords are pointy 19:35:28 pointy lets you aim at specific areas 19:35:35 not all swords are 19:35:43 all crawl swords are 19:35:52 Unlike, say, a great axe? Which cannot be aimed at an unconscious person's, let us say, neck? 19:36:25 great axes do get a damage boost though 19:36:34 (they also don't exist, but that's another matter) 19:36:36 what's a great axe 19:36:39 right 19:37:05 Meant an executioner's axe. Look, I say again: either stabbing is the skill of *literally* stabbing people, in which case, fine, only blades get a big benefit, BUT also only blades train with it. 19:37:33 currently everyone gets a benefit, and blade-users get a bigger benefit 19:37:35 OR, alternatively, stabbing is the skill of *figureatively* stabbing people, including vulnerable strikes with other weapons, in which case the set of weapons that get a bigger bonus makes no sense. 19:38:12 maybe short blades are bad at melee but good at stabbing and other weapons are bad at stabbing and good in melee 19:38:16 ??realism of crawl 19:38:16 realism of crawl[1/6]: whacking KILLER FUCKING BEES with a halberd and throwing javelins at them 19:39:08 Ah, okay. In that case, we should make butterflies explode, because it is *desirable* that the rules should be capriciously arbitrary and ungrounded in the underlying mechanics. 19:39:37 Myself, I am unclear why people are attached to the increased stabbing bonus for great swords. 19:40:00 it'd be fine if long blades didn't get a bonus, it's ignorable anyway 19:40:27 drpraetor: people aren't attached to it, in a recent version 1-handed polearms were given the boost too 19:41:05 OG17 - so you're saying that *doubling* the bonus is ignorable, but the half-bonus-you-still-get is desirable enough for axe wielders that they should train it by default? 19:41:22 what 19:41:38 if stabbing skill does something, it should train 19:41:40 Long blades get, last I checked, double the bonus of a club or axe. 19:41:45 stabbing skill is doing something 19:41:48 therefore it trains 19:41:52 what is the problem 19:41:58 Stabbing skill should do something *worthwhile* if it is going to do anything at all. 19:42:04 it is worthwhile 19:42:05 extra damage is worthwhile 19:42:17 have you considered turning off stabbing 19:42:25 if you don't think it's worthwhile you can turn off the skill and then you won't train it anymore 19:42:26 not worth training for that, but it isn't useless either 19:42:38 2 19:42:42 I do turn off stabbing, I'll run some more simuations, but the 2-3 points of stabbing I end up with anyway.... 19:42:52 2-3 skill takes so little XP that you should consider it worthwhile 19:42:54 how much xp is three points of stabbing 19:42:58 that's what we are trying to tell you 19:42:58 crawl doesn't let you assign 100% of your xp exactly where you would like it 19:43:01 is that like five ogres 19:43:01 this is not new 19:43:11 (except ashenzari but whatever) 19:43:24 INVERSE VICTORY DANCING IS BAD. 19:43:32 your arguments are bad 19:43:37 please leave 19:43:46 wtf is inverse victory dancing 19:43:47 is inverse victory dancing "skills training when you use them"? 19:43:51 yes 19:43:54 is this argument still happening 19:43:55 because yeah 19:44:02 Inverse victory dancing is where you avoid hitting a confused enemy in order to avoid wasting XP on stabbing. 19:44:08 drpraetor: if it makes you feel better, the skill training system might change drastically in 0.9 19:44:11 you could just melee things and not care that a hilariously trivial amount of xp is going to be diverted over the course of a game 19:44:15 kilobyte has some proposal 19:44:24 skill training overhauls <3 19:44:29 anyway 19:44:29 and then you could avoid ever gaining stabbing skill if you wanted 19:44:30 !learn add DrPraetor < DrPraetor> Inverse victory dancing is where you avoid hitting a confused enemy in order to avoid wasting XP on stabbing. 19:44:31 drpraetor[4/4]: < DrPraetor> Inverse victory dancing is where you avoid hitting a confused enemy in order to avoid wasting XP on stabbing. 19:44:45 can someone be useful and suggest a message for when monsters cast trog's hand? 19:45:07 what's the player message? 19:45:17 "you skin crawls you feel resistant to magic"? 19:45:21 just "your skin crawls. you feel more resistant to hostile.." yes 19:45:21 OG17 - I'll run some simulations, but last I checked, which was 0.2 versions ago, that 3 stabbing could have been +2 evocations for a late game enhancer-staff wielder. 19:45:27 Which is hardly negligible. 19:45:48 +2 evocations meaning 2 instead of 0 19:46:13 @monster@ grins madly as the Hand of Trog touches upon it! 19:46:18 No, +2 evocations meaning 10 instead of 8, actually. 19:46:44 that's not possible 19:46:50 were you the one that handled the armor nerf? 19:47:04 also go change the player message i guess 19:47:13 the skin crawl / resistant to magic are boring anyways 19:47:20 yeah 19:47:30 I was the one who wrote up dpeg's request for the armour nerf. I disagreed with the implementation and then had a conversation MUCH LIKE THIS ONE when I said it was nerfing armour too hard. 19:48:07 perhaps if your conversation was MUCH LIKE THIS ONE that's why nobody listened to you 19:48:07 didn't you say you had no idea there was a problem a few months ago 19:48:10 yes 19:48:33 this really isn't going anywhere productive 19:48:46 eronarn: I don't really see berserkers as the "grins madly" type 19:48:48 sorry, yeah 19:49:01 I said that I'd been assured there wasn't a problem by the people making the design decisions and what can I do? I *did* write the AEVP component of the new armour rules, that was my idea if anyone wishes to complain about that. 19:49:33 spriggan berserkers grinning madly is a good mental image 19:49:37 but yeah, i'm not sure 19:49:41 elliptic: i picture trogites as being cackling and screaming all the time 19:49:44 less good for the qwarves, I think 19:49:47 heh 19:50:03 but yeah if you see them as GRR ANGRY then maybe they should shout instead 19:50:07 (the ability could be a warcry, etc.) 19:50:21 not a big nerf on it to make the player version make noise or whatever 19:51:47 orcs already use a warcry for something completely different 19:51:51 and vaguely berserkish 19:51:54 might be confusing 19:52:14 MarvinPA: go remove the slaying penalty from bracers of archery, it's dumb 19:52:39 bracers of archery are weird 19:53:04 they should always identify or something 19:53:08 or have another possible brand 19:53:28 also it's only -1, -1, what's wrong with it? 19:53:42 maybe berserkers should just have passive mr and regen, hand isn't very noticable or something that a player's going to react to anyway 19:54:00 yeah, I don't mind the slaying penalty, but it is bad that you always know what they are on sight and yet most chars don't want them ever 19:55:43 does anyone know why the map in the morgue file disappeared? 19:56:04 MarvinPA: it's just a weird little penalty for no reason, melee primary characters already don't have any particular reason to choose it 19:56:07 I guess that might have something to do with the unicode stuff 19:56:31 it was being buggy in some cases so maybe that's why 19:57:42 it is really annoying though since now I actually have to watch recordings of games if I want any clue about what the ending situation was like 19:57:43 -!- eith has quit [Ping timeout: 258 seconds] 19:57:47 Why did Wights get changed to genus z? I see how this can be viewed an improvement since they are corporeal, but I don't think it is. 19:58:02 because they are corporeal 19:58:13 because they are corporeal, unlike every other W 20:00:07 also they use equipment 20:02:43 for posterity, getting to three stabbing takes around 70 xp 20:03:18 and that's while training fighting sbl and stealth 20:10:05 OG17 - now try the same test with a spellcasting of 20 and a conjurations of 16. 20:10:48 Or, better yet, when I get around to downloading the current trunk I'll get some hard numbers. 20:12:27 Q - why can ghouls see brown chunks now but not purple ones? Or does the sky beast not have purple chunks? 20:14:29 A - ghouls can eat purple chunks 20:14:34 it was a trick question 20:14:38 Ghouls can eat brown chunks as well. 20:14:43 I could try the same test to see if that 70 xp raises high-level evo skills by two 20:15:33 OG17 - I'm unaware of the source of your hostility to any of my several underlying complaints, but one of them is of course that the skill training system is opaque. So are you beign deliberately dense? 20:15:50 i have a feeling brown chunks are preferable for ghouls 20:16:01 aren't they more likely to heal rot? 20:16:34 If that is the case, they should probably show up in blue or some other "good" color for ghouls? Like how rotten chunks get to be yellow? 20:17:20 answer: because rotten chunks get to be yellow 20:17:41 Right, so if you're a ghoul... is any other item in the game color-coded brown? 20:17:49 could make all non-rotting chunks lightgrey for ghouls if they don't do something special 20:18:12 and gourmand should make everything yellow because it's not obvious 20:18:19 I mean for ghouls of course 20:18:31 monqy - that seems to be the case, but if the purpose of the color-coding is to communicate that contaminated chunks are *better* for ghouls, then contaminated chunks should get a *good* colour. 20:18:36 And you're right about gourmand as well. 20:18:46 DrPraetor: a good colour like what 20:19:02 Blue? Blue is better than gray, yeah? 20:19:04 ghouls can pretend brown is good 20:19:19 Or rotting chunks could be artifact-white while contaminated chunks get to be yellow? 20:19:27 yellow is supposed to be "emergency" i think 20:19:34 yellow is all sorts of stuff 20:19:35 which isn't quite right for chunks 20:19:37 that too 20:19:50 for food it's not "emergency" 20:19:57 I don't even know what it is 20:20:05 it differs a lot 20:20:12 "yum you have gourmand" 20:20:17 or something 20:20:20 oh wait no, not even that 20:20:23 for carnivores it means meat, for herbivores it means veggies, for ghouls it means rotting 20:20:34 honeycombs and royal jellies for all 20:20:36 except also honeycombs/jellies 20:20:46 mummies don't get yellow food 20:20:53 also vampires 20:21:14 I guess blood is yellow for them 20:22:16 are you sure that contaminated chunks matter for ghouls? 20:22:16 03MarvinPA * r4be022e33964 10/crawl-ref/source/dat/des/variable/float.des: Don't autoexplore into lemuel_bear_cage 20:22:28 03MarvinPA * rb071afa2f719 10/crawl-ref/source/mon-cast.cc: Add a message for monsters casting Trog's Hand 20:22:41 i was getting "this chunk tastes delicious" as opposed to "good" for brown chunks sometimes 20:23:07 fuck. 20:23:19 Well, at least I died more to pore luck than to actual stupidity that time. 20:23:53 wonder if it sometimes acts as rotting or if it's just flavor 20:24:21 be simpler to only have rotting chunks act as rotting if it's the former 20:32:59 -!- upsy has quit [Quit: Leaving] 20:35:31 !learn add elliptic < elliptic> ... hitting an unconscious person with a great sword sounds pretty good to me 20:35:32 elliptic[3/3]: < elliptic> ... hitting an unconscious person with a great sword sounds pretty good to me 20:35:37 I love quoting people out of context 20:35:54 :) 20:36:00 Actually, in that case I think the quote contains all needed context. Does crawl compile under cygwin? 20:38:49 I want everything under X11, Math, Science and Devel. 20:41:24 greensnark: planning to do a lorc build tonight? i have some stuff in the branches that's not merged, so i'd want to start it sooner rather than later if so 20:42:50 Sure, let me know when I should pull 20:42:56 I'll be around another 2 hrs or so 20:43:16 Is the amulet of conservation gone? I just realized I haven't seen one in like ten games. 20:43:24 no, it is still around 20:43:40 ??rng[4] 20:43:41 rng[4/10]: The RNG has it out for you. Really. 20:43:44 ??rng[5] 20:43:45 stupid death[3/3]: nil the Executioner (L21 MDBe), worshipper of Trog, forgot to breathe caused by wearing the shield of Ignorance on Vault:8, with 285561 points after 71561 turns and 11:19:06. 20:43:47 ??rng[6] 20:43:47 rng[6/10]: Technically speaking, it's the Mersenne Twister God. 20:43:51 ??rng[3] 20:43:51 rng[3/10]: More properly a *pseudo*-random number generator. Unlike true random numbers, pseudo-random numbers are generated out of spite, in order to make you cry. 20:46:18 I don't particularly want to generate a Sifite and check, but is new-Sif finalized in trunk or are we still opening the field to suggestions? 20:46:47 new-sif? what 20:46:53 sif is still sif 20:47:40 Well, sif is going to lose the amnesia power and her mana-boost is going to contaminate you rather than hungering you, is that not correct? 20:47:54 Those are pretty substantial changes. 20:47:59 not as far as i knew, i don't think anyone's in a rush to overhaul sif 20:48:30 no 20:48:35 i know dpeg's keen on a library of some sort instead of book gifts, that's the only change i can think of that's been planned but not in any great depth 20:48:36 you do know that stuff on the wiki is just proposals, right 20:48:38 Huh. I'd thought the hunger->contamination change, at least, was a done deal. 20:48:53 Well, yes :) 20:53:39 Increase rate of training for recite (https://crawl.develz.org/mantis/view.php?id=3776) by Eronarn 21:05:32 But that brings my next question - is it a good idea to do some sif revision in the next build, or no? I had some thoughts on making Sifites more "coherent in mechanics", that is, so that they 1) care more about invocations skill, and 2) care more about spending their piety. 21:06:17 By the way, the charm/hex split has given ghouls a substantial buff, since their Ench skill used to be -2 but they're now -1 in Charms, which was 99% of what a Ghoul would ever care about? 21:07:32 the other 1% is confusing touch 21:07:34 -2 to -1 isn't a substantial buff but sure 21:07:53 why did ghouls' charms go up, anyway, if that's what happened 21:09:26 no idea 21:11:04 -2 to -1 makes it a lot easier to cast CTele, for example. I'd say it's a pretty big deal, on the scale of such things. 21:12:00 -!- hayenne has quit [Ping timeout: 248 seconds] 21:12:00 -!- edlothiol has quit [Ping timeout: 264 seconds] 21:13:20 It's certainly not a game breaker, like being forced to sink some XP into stabbing for profoundly marginal benefit. 21:19:19 !apt gh 21:19:20 Gh: Air: -2, Armour: -1, Axes: -1, Bows: -1, Charms: -1, Conj: -2, Xbows: -1, Dodge: -1, Earth: 1, Evo: -1, Exp: 120, Fighting: 1, Fire: -2, Hexes: -2, Ice: 1, Inv: 0, Long: -1, Maces: -1, Nec: 0, Poison: 0, Polearms: -1, Shields: -1, Short: -1, Slings: -1, Splcast: -1, Stab: 0, Staves: -1, Stealth: 1, Summ: -1, Throw: -1, Tloc: -1, Tmut: -1, Traps: -1, Unarmed: 1! 21:19:26 yep 21:19:37 the Hexes And Charms Have To Be Different For Some Reason principle at work 21:19:52 all kinds of races have slight apt differences in them with absolutely zero justification 21:21:21 I'm sure kilobyte had reasons for all the apts 21:23:47 I'm fine with hexes != charms, but surely Ghouls, who are full of spite and malevolence, would be good at Hexes? Because they love all the Voodoo you do? 21:25:25 witches aren't ghouls 21:26:12 Speaking of which, the Prince of Hunger is a unique n, he has the hungry ghost effect on his claws, and he gets a spellbook with some Lugonu effects and hexes. 21:26:32 elliptic: But *ARE GHOULS WITCHES*???? They totally are. 21:27:06 not sure about that :P 21:27:25 !tell greensnark pushed to github; some squarelos and lorc fixes, think this one should be compatible too 21:27:25 Eronarn: OK, I'll let greensnark know. 21:27:40 elliptic: this fixes both pois cloud and autofight, i hope 21:28:19 great :) any hope of square tornado anytime soon? ;) 21:28:34 are square haloes/silence in? 21:29:30 square haloes/silence are, but tornados are a lost cause, i'd have to completely ditch that part of the code 21:29:41 because it's its own custom thing, not normal roundlos stuff 21:47:12 ??tukima's dance 21:47:13 tukimas dance[1/2]: A level 3 enchantment spell that turns your wielded (non-fixedart melee) weapon into a pet temporarily. It will be hostile if cursed (and will remain cursed). The weapon can be recovered after the spell ends. Is essentially a [@?dancing weapon] wielding the chosen weapon. Does not bypass distortion unwield effects. 21:47:20 Is tukima's dance Hex or Charm? 21:47:51 hex 21:48:07 Ah, okay. Well, that's a damn good spell and Ghouls are unhappy to lose it, so there you go. 21:48:46 Now, does this mean that I should just leave all this reaching crap lying around where I found it? 21:49:35 you realize that tukima has been nerfed multiple times since 0.5, right? 21:49:42 -!- Cryp71c has joined ##crawl-dev 21:49:43 also, a -2 apt doesn't mean you can't use a spell 21:51:24 Evening 21:52:53 Yeah, it's still useful. 21:53:07 Well, with reaching weapons *combined* with shadow creatures, it is useful. 21:53:21 But I'm not going to get my hexes high enough to... dancing weapons are still banishable, right? 21:53:59 dancing weapons are magic-immune but distortion can banish them as of recently 21:54:07 (it used to bug out) 21:54:10 Sorry, meant *abjurable*. 21:54:16 they're not abjurable 21:54:17 But glad that was fixed. 21:54:27 but they're kind of weak 21:54:33 I wouldn't bother with it 21:54:40 Even so, with a hexes of only 3 or so, I'm not going to get enough duration. 21:54:54 the dancing weapon strength depends on spell power 21:54:55 so 21:54:57 Well, they're also only 3 MP each. So back when they were also fearsome, that was insane. 21:55:08 Ooooooh. I didn't know that. 21:55:23 they got nerfed multiple times 21:56:41 Possibly they've been overnerfed, then. I remember thinking the first nerf was judicious. 21:56:49 they're not overnerfed 21:56:57 it's a L3 single school spell or whatever 21:58:13 Anyway, the bailey, with the two islands and the orc knights and the xbows? That bailey should have really nice treasure in it, right now it's a spoily ripoff (the proper course of action is just to flee). 21:59:09 ...what 21:59:15 you have no idea what you're talking about 21:59:18 ?? 21:59:27 I was just in it. Did I miss a secret door? 21:59:40 Oh, I suppose with Meph Cloud it becomes a pile of free XP instead. 21:59:45 i don't even think the doors are secret, there's two loot rooms 22:00:01 They had nothing in them just now. A glowing spear and a few $. 22:00:30 the monsters pick up stuff in them 22:00:42 also it has less loot than other baileys because it's shorter and easier 22:00:45 I should I guess have pointed this out while I was inside, but the loot rooms were empty. Well, whatever they picked up, it was none of it scrolls or potions. 22:01:25 Easier than the kobold bailey? No way. 22:01:42 It's easier than the other bailey with orc knights, and it's easier than the deep elf bailey, no Q. 22:01:57 But it was much harder than the (better payed) gnoll bailey, or the hobgoblin one, or the kobold one. 22:02:13 It has xbow wielding orc knights in it. How is that easy? 22:02:52 -!- valrus has quit [Remote host closed the connection] 22:03:56 they can't melee you and you can get out of LOS of a bunch of them easily and you can drown them 22:06:54 Eronarn: CAO updated 22:06:54 greensnark: You have 1 message. Use !messages to read it. 22:07:09 Eronarn: Do you have a preferred homepage that I can point at from the dgl menu 22:07:41 So that players can get more info on lava orcs and stuff even if they're not on irc 22:07:47 Yes, if you have meph cloud, it's a cakewalk. 22:07:52 greensnark: lava orc wiki page, i suppose 22:08:02 octopodes are more of a side show at this point 22:08:13 though i am surprised at how much people like them 22:08:21 the 0 AC thing is getting great reviews 22:08:32 Again, I don't see the point of octopodes until schoolgirls are implemented. 22:08:55 this isn't really the channel for that kind of comment 22:09:00 Eronarn: Yeah, they seemed more popular than lava orcs on ##crawl :P 22:09:27 greensnark: people do like lava orcs, just that the temperature is necessarily screwy because of how hard it is to debug tension 22:09:43 it'd be nice to get tension calculations logged or something 22:09:55 so that it could be monitored looking at actual games 22:11:06 -!- Cryp71c has quit [Read error: Connection reset by peer] 22:11:29 -!- valrus has joined ##crawl-dev 22:14:32 eronarn: personally I don't think lava orcs are really playable at the moment 22:14:37 -!- Cryp71c has joined ##crawl-dev 22:14:40 they feel just like hill orcs with rF+ 22:15:08 elliptic: did temp not come up at all, or? 22:15:25 because some people are definitely making use of temp effects 22:15:33 temperature basically comes up with early uniques and that's all 22:15:43 hmm, maybe it doesn't scale well enough with level 22:16:34 Eronarn, you're using the XP system to determine value, right? 22:16:37 (tension value) 22:16:56 Make lava orcs really good at casting spells in heavy armor. That's an underserved niche. 22:17:40 Cryp71c: it's the existing tension system with some further processing to try and make it feel more organic 22:17:46 and also with a tweak to tension calculation 22:18:06 By the way: out of depth skeletal warrior packs? Not cool, dude. 22:18:11 i changed the formula for low HP scaling to sqr(% of HP lost) 22:18:27 so now it can go up to x10 or so 22:18:33 whereas before it just barely hit x2 22:18:56 i think my processing might be excessive, but i really need to get raw tension numbers from a wide variety of situations, which is hard to do in wizmode 22:19:07 Eronarn, that's good, though I've noticed (through my proximity to the guardian stuff) that there are some scenarios where tension calculation is unnecessarily high and (more frequently) drastically low. 22:19:31 yeah 22:19:42 some stuff is a lot more threatening than its XP value would indicate 22:20:12 i'd say maybe look per-monster to see what capabilities it has that could make a situation go bad fast 22:20:33 Part of this is unavoidable; its the result of not being able to identify that a hasting-might casting unique of HD == Player XL is MUCH more dangerous to the player than a unique with an HD (and subsequently more XP) 5 or 6 values higher than the player's XL just because of how that player's loaded out. 22:20:47 also possibly scale tension by distance, especially for melee monsters 22:21:20 Part of the issue is how XP scaling occurs, which is problematic in and of itself. 22:21:45 I would be that although there is very little tension occuring in the mid to late game that all extended end games and zot:5 action would see dramatic bursts of tension 22:21:59 If for nothing else than because of how much XP is yielded for killing stuff like draconians and orb guardians. 22:22:29 Since I'm about to die at D~=10-15 for the fifth consecutive time I disagree that the mid-game is definitionally low tension :). You can get yourself killed if you try. 22:22:52 Oh, you're saying that the *XP value* spikes tremendously in the late game. Yeah, that's true. 22:23:07 Sorry, inverted your meaning. 22:23:15 -!- Kurper has quit [Remote host closed the connection] 22:23:43 You could base the tension scale on the *actual damage output* of monsters. So hydras are very high tension. And then spellcasters and such contribute more tension at a distance? 22:24:04 "You gently float upwards from the wall." (https://crawl.develz.org/mantis/view.php?id=3777) by reid 22:24:34 Cryp71c: yeah the xp scaling mechanic seems a bit off 22:24:50 how did you set up the formulas in the first place? trial and error? 22:24:53 The underlying math would be complicated but something like: for each monster, tension contributes: total damage*speed/(distance from char). Any monster holding a missile weapon is distance 1, and spellcasting monsters contribute additional tension depending on the nominal level of their spells? 22:25:11 Eronarn, perhaps Praetor has a reasonable idea. for each turn, for each monster in LOS, calculate their tension by faking combat and comparing the amount of damage that would've been done with the player's total hp. 22:25:40 that sounds really, really messy 22:25:48 Eronarn, just an idea. 22:26:02 It has an interesting approach, its the implementation of combat that makes faking it nasty. 22:26:06 I think you'd have to use some relatively simple approximation, rather than actually trying to simulate combat. 22:26:30 you'd have to make sure all the code that needed to be run like that could be isolated without affecting anything else, and you'd have to have a huge sample size on account of stuf like monsters using random spells 22:26:33 So a high EV character would overestimate the tension from monsters with the inaccurate tag, okay. Perfect vs. good. 22:26:43 Eronarn, yeah it was mostly trial and error coupled with some source diving on xp calculations and referencing commonly-encountered uniques / monsters at various XLs. 22:27:10 You can either make some simple calculation from monster stats that works better than monster XL/XP, or you can't. 22:27:14 it's probably better to just use existing tension but look at some specific monster capabilities the same way that it looks at whether the monster has specific enchants 22:27:19 Eronarn, yeah, there could be a high variance in the amount of tension a particular monster added, based on how the fake combat *happened* to play out that particular turn. 22:27:44 Having the tension kick around by some random factor would be desirable in and of itself, actually. 22:27:48 like... can the monster slow you, paralyze you, etc.? then it's basically acting as a potential multiplier to the entire rest of the combat 22:27:57 no it wouldn't 22:27:58 Eronarn, tension has a lot of potential for fairly rare (but unique) effects, like your heat. I'd like it to be as effective a system as possible. 22:28:03 greensnark: a quick question about crawl lua if you have time: if I'm using crawl.sendkeys() to send keypresses, is there any way to get crawl to process those keypresses inside the lua? coroutine.yield() doesn't actually seem to do anything as far as I can tell 22:28:26 Cryp71c: yeah, it's a fairly central effect, effort should be put in to making it as refined as possible 22:28:32 Eronarn: I take it that you haven't heard the good news about Xom? 22:28:36 elliptic: I thought you'd already solved that for the bot 22:29:01 what good news 22:29:05 can't cling to a closed door (https://crawl.develz.org/mantis/view.php?id=3778) by reid 22:29:33 Eronarn, perhaps it involves looking at the XP calculation system...whatever we do, tension can't be built upon a foundation of either a) high-cost (execution wise) code AND/OR b) inaccurate code (regarding approximating monster threat) 22:29:34 iirc coroutine.yield transfers control to crawl, which resumes after it updates state, except sometimes crawl doesn't actually resume because ????? 22:29:34 You have hooves. Also, you're on fire. 22:29:37 greensnark: well, currently it just processes the strings of keypresses in between presses of the key macroed to the lua function, as far as I can tell 22:29:51 what 22:30:13 Cryp71c: well, XP calculation does actually already do this sort of thing 22:30:20 I'd say that it's more important to be unscummable rather than inaccurate. 22:30:23 the question is whether XP and tension should be doing this with separate goals 22:30:28 greensnark: which works, but I was curious about whether it is possible to eliminate holding down the macroed key 22:30:57 elliptic: Well, sendkeys just fills the input buffer. You need something like process_command to actually react to those keys 22:31:01 Like, should casting Ozo's armor cause the tension level to plummet? 22:31:05 for example, merged slime creatures next to you might be very high tension, but not any more XP than the separate slime creatures 22:31:11 Eronarn, the question I mean to arrive at is whether tension is off (for certain periods of the game) because of a failure in the tension code or a failure in the xp-calculation code...or perhaps minor ones in each. 22:31:20 There's crawl.process_keys which does sendkeys + process_command 22:31:33 But it also flushes the input buffer before the sendkeys, which may be non-ideal 22:31:42 Cryp71c: it feels like both to me - the tension code has some pretty clear problems, but it's based on XP, which also does 22:31:46 And process_command just processes one key 22:31:50 @??boggart 22:31:51 Er, one command 22:31:51 boggart (06g) | Speed: 12 | HD: 2 | Health: 6-16 | AC/EV: 0/12 | Damage: 5 | Flags: spellcaster, see invisible | Res: 06magic(18) | XP: 21 | Sp: confuse, slow, invisibility, blink, shadow creatures. 22:32:01 21 XP just because they're so low HD 22:32:06 @??gnoll 22:32:06 gnoll (08g) | Speed: 10 | HD: 2 | Health: 8-18 | AC/EV: 2/9 | Damage: 9 | Res: 06magic(8) | Chunks: 07contaminated | XP: 14. 22:32:25 Well, okay. If we fix the XP values of the monsters to match their respective threat levels, would that fix tension? Because if that's the only issue, we should just do that and leave the tension code alone. 22:32:25 boggarts are way, way, way worse than a gnoll and a half, even if they're not actually really bad 22:32:33 @??ogre 22:32:34 ogre (07O) | Speed: 10 | HD: 5 | Health: 17-39 | AC/EV: 1/6 | Damage: 17 | Res: 06magic(20) | Chunks: 07contaminated | XP: 119. 22:32:37 @??blofly 22:32:37 unknown monster: "blofly" 22:32:41 @??blowfly 22:32:41 unknown monster: "blowfly" 22:32:48 @??giant blofly 22:32:48 unknown monster: "giant blofly" 22:32:53 @??giant blowfly 22:32:53 unknown monster: "giant blowfly" 22:33:04 @??red wasp 22:33:05 red wasp (05y) | Speed: 15 | HD: 8 | Health: 29-59 | AC/EV: 7/14 | Damage: 2304(paralyse) | Flags: fly | Res: 06magic(32) | Vul: 09poison | Chunks: 09poisonous | XP: 543. 22:33:07 @??hill giant 22:33:08 hill giant (04C) | Speed: 10 | HD: 11 | Health: 44-76 | AC/EV: 3/4 | Damage: 30 | Res: 06magic(44), 12drown | Chunks: 07contaminated | XP: 652. 22:33:29 DrPraetor: it wouldn't fix tension no... for example, an ogre 8 away from you is the same as an ogre 1 away from you. being at 1 HP is only twice as bad as being at full HP. having allies on screen makes stuff less tense, which may not always be appropriate 22:34:07 can't extend spider form duration while clinging over deep water (https://crawl.develz.org/mantis/view.php?id=3779) by reid 22:34:10 DrPraetor, changing xp values also involves rebalancing other aspects of the game that are already (mostly, relatively) balanced. We don't want that atm. We have enough on our plate. 22:34:14 greensnark: I've made some attempt to use crawl.process_keys, yes... but it seems that crawl.process_keys tries to translate the first key in your string into a command and then processes that 22:34:27 and this seems to work rather poorly in practice 22:34:37 I agree that, flawed as the XP values are, they're not actually *broken*. 22:34:53 But the ys and the summoners are chronically undervalued, and should get a boost. 22:35:04 so I was wondering if there was any alternative way 22:35:25 Na, the 50% penalty for summoners is just the beginning of making them fair. 22:35:55 (example of it not working... it doesn't know how to translate '#' into a command and gives an error message) 22:36:30 i think rather than boosting summoner XP we should nerf summoners 22:36:30 Cryp71c: I meant that monster summoners are worth too little XP, not that the penalty for indirect kills was the wrong size. 22:36:44 it's not fun to have a big bag of XP that you have to run from if you can't drop it quickly 22:37:20 throwing more XP at them won't change that they can be pushovers if they don't summon or run-or-you're-dead if they do 22:37:28 Meh. I don't think summoners need a nerf. Crawl is not particularly well balanced between different character types. Summoners are boring and relatively easy. 22:37:38 we're not talking about player summoners 22:37:48 -!- Kurper has joined ##crawl-dev 22:37:53 Cryp71c seemed to be. 22:38:06 DrPraetor, I only got on it from your comment. Back to tension. 22:38:26 But yeah, let's not change subject. If you kill a summoner before it summons, you should get more XP than it was worth because you showed good tactical foresight. How can that be a problem? 22:38:43 But the summoner should definitely contribute a lot to tension as long as it is aware of you and might summon. 22:39:13 ...what 22:39:19 i don't even know where to start with that 22:39:47 Suppose Boggarts are worth 500XP. If you're smart and canny and kill them quick, you get a bunch of XP for minimal effort outlay. 22:40:05 If you end up running from their summons, you messed up, sucks to be you. 22:40:24 DrPraetor, that's relating back to xp bonuses and random stuff we're neither interested in touching or considering in tnsion. 22:40:26 tension* 22:40:59 Right, tension. Well, an unaware boggart is zero threat. Any aware monster with a summoning spell in it's spellbook is a big threat. Best fix for this would be to assign a tension value to individual *spells*. 22:41:13 So crystal spear would contribute a lot of tension because, yeah, you can die any second while the archlich can see you. 22:42:23 We can't special case individual spells like that either, though. That's a lot of (inefficient) work. 22:43:05 It wouldn't be special case. Every spell would get a tension value. I'll allow that it's a lot of work, but if it's important for the tension mechanic to work well, there's no way around it. 22:43:36 We could set up some general formula where the tension of a spell was a function of the damage it did and/or the total HD of monsters it was expected to summon, if we wanted an imperfect approximation. 22:43:52 DrPraetor, there are hundreds of spells. There is a more efficient way to do it. Also, applying a specific value to spells doesn't touch on tension levels for non-spellcasters. 22:44:26 I still think that the best, non-CPU insane approximation is some function of their total damage. 22:45:16 You could so something like: sqrt (sum of squared damage for each attack) * speed * to-hit-EV-10. Then you divide by distance (for melee) or don't (for ranged, or for stuff in spellbooks.) 22:45:28 This could all be precalculated for each monster in the beastiary and held in memory until needed. 22:45:31 this is a dumb conversation 22:46:15 Well, either we really, really want tension to be a good reflection of the total threat level: in which case we need to put in the effort to make some good estimate of how threatening things actually are. 22:46:29 OR, we're willing to make do with some approximation. In that case, I still think damage output is a better indicator than XP value. 22:47:29 * DrPraetor shrugs. 22:47:33 ATM, only non-wandering, aware monsters at (or near) LOS affect tension. Tension is calculated almost directly from XP value, tension is boosted or minimized based on monster HP and current status's (berserk, haste, slow, etc. It covers almost all of them) 22:48:21 Allied monsters reduce tension by the same amount (IIRC) that they would've added to tension (was always problematic, since the amount they would've added to tension is still based on their XP value to the player) 22:48:59 So it's tracking distance, awareness and LOS already? Would it be CPU intensive to track clear-line-of-fire (i.e., not through allies)? 22:49:11 "Cling" HUD status light is extremely irritating (https://crawl.develz.org/mantis/view.php?id=3780) by reid 22:49:21 I know the monster AI is already tracking this every round. 22:49:23 DrPraetor, current system doesn't factor in potential damage in any way. 22:49:47 DrPraetor, monster AI runs each time the system indicates that the monster may make an action, we cannot probe in to determine what that action may or may not be. 22:50:16 Cryp71c: No, we can't. But when the monsters actually *do* make decisions, they can change the tension value at the same time, without much CPU overhead. 22:50:33 So when a monster *sees* you, the tension goes up. When a monster says "oh, I can't see him anymore", tension goes down. etc. 22:50:39 So as far as we know (or care to know), we know monsters and what they're capable of, but not necessarily what they're doing their turn. 22:51:05 I'm saying, from a purely procedural level, that the tension could actually be calculated *as monsters make decisions*. 22:51:32 DrPraetor, in retrospect to our conversation, tension should not vary for the same type of monster under normal circumstances. For a player of certain XL, an orc wizard should always result in the same tension value, regardless of whether he casted the "awesomeness" spell, or just moved. 22:52:19 Should the tension from the orc wizard be the same whether or not he can see you? 22:52:54 I'm saying... I suspect it doesn't actually matter what I'm saying since it's Eronarn I gather who is interested in working on tension and he thought the conversation wasn't useful. 22:52:59 DrPraetor, if you can see him, he can see you (presuming awareness), if you're invisible, tension is already factored in for monsters that can't see or sense invis (I believe) 22:53:38 Right, what about line of fire. Is the tension for an orc wizard the same if there's a kobold between you and the orc wizard so the orc wizard can't shoot you? 22:53:47 Or won't, in any case? 22:54:08 DrPraetor, ideas are always welcome, but Eronarn's perspective on this conversation is based on the observation that your initial suggestions are infeasible (which we know because our familiarity with the source lends us this insight, but which you're not) 22:54:17 elliptic: Does process_command work? 22:54:18 DrPraetor, yes, the tenson is the same. 22:54:32 process_command does require that you know how many commands are queued though, so you know when to stop calling it 22:54:49 elliptic: You'll probably have to patch Crawl :) 22:55:09 DrPraetor, tension does not vary under different combat scenarios. If a monster is in sight and is aware of you, it has the same tension whether it can attack you or not. 22:55:11 There's also the whole force-interrupt thing which is used in several places and will kill the bot-macro 22:55:13 my perspective on this: tension just needs some minor tweaks to account for stuff like 'spell that can paralyze' or 'spell that can slow', and the only question is whether those tweaks should only occur for tension or should occur at the XP level 22:55:35 monsters already get a boost to XP for knowing certain spells 22:55:48 Eronarn, I think they're factored in artificially when determining a monster's HD and spells, yeah. 22:55:49 however, it may not be large enough; or XP : tension may not be a 1:1 ratio 22:55:57 greensnark: I think I was actually a little confused earlier and process_keys might work if I can figure out how to get a control-F key in lua 22:55:58 Eronarn, I find the latter. 22:56:28 Ah, so in answer to *my* earlier question: "can we fix this just by fiddling with some XP values", the answer is not "no" but in fact "yes"? 22:56:31 since before I was using *f and that gets into trouble with key_to_command stuff 22:56:46 perhaps something like a sphinx needs to give X amount of XP, but more than that to tension - for example, its high speed means it has more chances to go twice in one turn, which alters the distribution of random unlucky events in its favor 22:56:54 DrPraetor, we're not changing XP values. 22:57:03 ... 22:57:31 tension has some non-XP-related stuff that could stand to be looked at also, like distance, or whether LOS exists 22:57:36 DrPraetor, but we're not obligated to using JUST XP values in calculating tension. We can say that xp value contributes 75% of tension and some other (easily performed) calculation determines the other 25% 22:57:39 but monster XP has been known to be not so great for a while 22:57:56 I'll let you two work this out, then. FWIW, you could give some monsters an MR-dependent tension contributer tag. There are few enough paralyze casters that, if they're the big tension issue, that'd fix it. 22:59:03 Eronarn, high speed and having spells which can drastically turn the tide of a situation (before they actually DO turn the tide of the situation) are probably the biggest issues that I can think of. 22:59:14 Other advantage being that tension wouldn't skyrocket whenever a Lvl 27 deep elf saw an orc wizard who knew confusion. But OTOH, if we moved some of the tension code into the monster decision-making code block (and yes, I do have *some* familiarity with how crawl is written), this might give better handlign for your own summons being-in-the-way. 22:59:14 @??waglaf 22:59:14 unknown monster: "waglaf" 22:59:38 Who's the unique okawaru dwarf w/ haste and might? 22:59:51 he's a good example. Fairly unthreatening, but once hasted and mighted, is big trouble. 22:59:56 Agnes is a good speed-example. 23:00:37 DrPraetor, well, the DE vs orc wizard is already handled since (right now) a or wizard gives negligible XP and - thus - negligible tension. 23:00:45 s/a or/an orc 23:01:28 Cryp71c: Okay, I wasn't clear. *If* we add a tension modifier for every monster that knows confuse, that tension modifier ought to be dependent on the players MR (in roughly the same way that all tension modifiers ought to be dependent, more than they are, on the players hit points.) 23:01:48 So an Orc W, regardless of how many XP it's worth, *should* contribute a lot of tension to a low-MR character, but little tension to a high-MR character. 23:02:06 Again, if we care about tension enough to put in the effort to make it work really well. 23:02:12 no it shouldn't 23:02:25 a dragon doesn't become lower tension just because you have rF+ 23:02:34 there's absolutely no way we could account for stuff like that 23:02:59 elliptic: ^A is 1, ^B is 2, etc., so ^F is 6 23:03:18 Eronarn: Perfect is the enemy of the good. You're already proposing that we need a special case for paralyze. Myself, I'm perfectly happy to have a special case for paralyze but not for crystal spear. 23:03:48 So if you're saying you want a special case for paralyze, don't tell me that it'll be inconsistent to make that special case work well. 23:04:15 Clinging protects from mechanical traps which aren't next to walls (https://crawl.develz.org/mantis/view.php?id=3781) by reid 23:04:21 elliptic: Weren't you already sending ^F for the bot? 23:04:22 So, I stand by my special case shortcut - give some of the monsters a tension-vs-MR flag. 23:04:37 I was using *f 23:04:42 Oh right, missed that 23:04:57 DrPraetor: cool, go code it 23:04:59 see you in 2030 23:05:15 elliptic: But process_keys still processes just one command 23:05:20 meanwhile i'll spend my time on stuff that sees tangible benefits before thousands of manhours are spent 23:05:34 So you'd need to make multiple calls, don't know if that's an issue 23:05:45 hii 23:05:53 Morning due on the grass 23:05:54 your hat :o 23:05:55 You're in a good mood. "perhaps something like a sphinx needs to give X amount of XP, but more than that to tension - for example, its high speed means it has more chances to go twice in one turn, which alters the distribution of random unlucky events in its favor" 23:06:38 So you dont' want any advice on how to implement that well? I'm really pissing people off today for no reason I can discern. 23:08:15 a) you are really obnoxious b) you clearly haven't played crawl regularly in a while because you have all kinds of wrong ideas about the game c) you're putting forward all kinds of false dilemmas d) you're proposing stuff that has nothing to do with what we're trying to work on e) you are really obnoxious 23:08:20 i think that about sums it up 23:09:22 DrPraetor, Speed we can universally identify. If a monster is slower than the player, they really aren't THAT threatening, even if they have ranged attacks. 23:09:26 Eronarn: Cool it, if he's annoying you ignore him for a while 23:09:38 -!- hayenne has joined ##crawl-dev 23:10:16 DrPraetor, other spells (like paralyze) are already taken into account (there's a fairly small list of them that drastically alter a current fighting scenario) but only AFTER they've already taken effect, by which time its too late to factor their impact on tension. 23:10:48 DrPraetor, (at least for some effects, particularly Eronarn's heat concept) 23:11:18 greensnark: yeah, I know process_keys is just for one command at a time... currently I'm being confused by the behavior of crawl.process_keys("5") though, which causes the message _You start resting. but doesn't actually seem to do any resting 23:11:22 -!- varmin has quit [Remote host closed the connection] 23:11:22 -!- greensnark has quit [Remote host closed the connection] 23:11:22 -!- Sequell has quit [Remote host closed the connection] 23:11:35 DrPraetor, So handling the *possibility* of this fairly small list of effects which drastically alter a current fighting scenario (coupled with a speed and xp-considering system) cover the bases more adequately than the current (xp only) system 23:12:06 but do not incur the expensive overhead of long hash's of monster threat values or itterate through every spell calculating its potential threat. 23:12:21 Cryp71c: Myself, I think I already gave a practical, modest-output solution to this (just flag all the monsters that know paralyze or confuse, and have them make more tension depending on the player's current MR), although I'm not going to lecture anyone on how easy something is to code when I'm not doing it myself. 23:12:47 did I miss something? 23:13:13 due, not rly. Discussing tension as it relates to LO heat. Eronarn is getting aggrivated, I'm explaning things to DrPraetor 23:13:36 huh. 23:15:27 -!- greensnark has joined ##crawl-dev 23:15:27 Also, red wasps should be worth a lot more tension than hill giants. 23:15:29 DrPraetor, there's lots of variables that you're not considering though, some of which are that there is not a predefined list of monsters that have paralyze or confuse. (Its much easier to just check to see if the current monster knows either of these spells anyways). Also, paralyze and confusion aren't necessarily even the biggest threats we're referring to (although i guess they're somewhere on the list, confusion fairly low, p 23:15:29 aralyze fairly in the middle). 23:15:57 Fine, you can have an MR-tension flag on some of the *spells*, which was my original suggestion. 23:16:04 Trying to factor in MR is a slippery slope. 23:16:36 Where does it stop? Do we factor in rF for dragons? rC for ice dragons, ice fiends? Torment resistance? Acid resistance for acidy-monsters? 23:16:45 The "slippery slope" argument si something I just don't buy. If a special case is important, you treat it. This doesn't obligate you to add all of those... 23:16:51 except for torment resistance yes we should do that!@ 23:17:03 Huge code-base is huge :/ 23:17:07 DrPraetor, slippery slope is ALWAYS pertinent because we're not coding in the theoretical sense. 23:17:11 Practical world is practical. 23:17:21 We're not just coding for the moment either. 23:17:38 Maintainability is bigger than the "current" picture, and that sort of implementation affects maintainability to a huge degree. 23:17:41 Huge, often-changing code-base is huge, for that matter. It's very difficult to cover every eventuality in a way that is persistent and won't result in bugs, bugs, bugs. 23:17:42 Yeah, so an architecture to *allow* every spell to contribute some unique tension factor would be nice to have. 23:18:05 DrPraetor, yes, but its impractical, so we have to come up with a more elegantly designed solution. 23:18:06 Are we talking about monster spells, here? 23:18:39 due, sort of. are you familiar with tension? We're only (really) caring about spell effects that we're already factoring in (paralyze, sleep, berserk, haste, might, slow, etc. etc.) 23:18:43 Yeah, monster spells. Again, perhaps this was what pissed off Eronarn because he thinks this is a false dillemma? If paralyze-casting mosnters *really do* need to add more tension, we need to add a flag for this purpose *somewhere*. 23:18:55 Except right now we only care AFTER the effect has been applied, which (often times) is too late to start ramping up LO heat. 23:19:13 It can be on the spell itself (which I prefer), or it can be on the monster. To me, approxmations are fine, and there can be monsters with random spellbooks who fail to add tension. 23:19:20 So tension calculations should at least factor in a monster having one of those effects in its arsenal and boost its tension appropriately, even before the spell is cast. 23:19:38 we don't have a data collection framework to pull tension data right now 23:19:50 putting in anything more than an afternoon of fixes without that is insane 23:19:55 Per-monster tension would make more sense, at least for paralysis, in the sense that paralysis isn't *always* caused by spells. 23:20:12 due, per-monster is the current system. 23:20:29 We loop through monsters in (or near) LOS, calculating individual tension levels. 23:20:35 Summing them to give the current tension. 23:20:38 due: also monsters where the combination of spells makes them more threatening than those spells in isolation - like para + summon 23:21:04 Fine, here's my proposal. I think we've covered the imperfections in this, so by proposing this, I'm saying, "those problems are minor in the context of the issue so we're ignoring them." 23:21:14 Eronarn: as in, you cannot apply a value to spells because sometimes it is the combination of spells that is multiplicative rather than additive? 23:21:29 -!- Sequell has joined ##crawl-dev 23:21:40 wb greensnark. 23:21:43 Eronarn, trying to couple those effects seems an equally fruitless endeavor, but seperately they could have their own scaling factors that when scaled upon each other, result in a semi exponential scaling (1.2 for each, so 1.2 * 1.2) 23:22:16 due: right... or consider a monster with one really bad spell, and several ones that don't matter 23:22:28 it'll use that one spell less 23:22:28 -!- varmin has joined ##crawl-dev 23:22:51 I think the simple solution is that there isn't really a *simple*, workable solution. 23:22:53 or monsters that are very threatening early on, but not as much later 23:22:53 Proposal: Monsters *may choose to have* special tension flags, which contribute extra tension which can be scaled down by one of the players resistances. For the two dozen or so monsters for which this is an issue, we add the flag. If dragons don't get the rF flag, that's not perfect, but the XP values aren't perfect either and the game doesn't explode. 23:23:29 Cryp71c: problem is that if you use 1.2x or whatever, you're just building off of XP 23:23:35 like with boggarts 23:23:57 say, with cerebov - him summoning 1s is actually not threatening at all because he'd otherwise be casting firestorm 23:24:33 Eronarn, either we use a multiplicative scaling system (which must eventually be based off some core value upon which the scaling takes effect) OR we do additive and fail to encorporate monsters having a particularly nasty array of spells which complement themsevles well (paralyze + summon, might + haste, etc) 23:24:49 Cerebov is easy. If we just raise the tension contribution from low hitpoints (which we should do anyway if I'm reading this mess properly), cerebov will contribute more tension just before he kills you :) 23:25:37 Don't might and haste already contribute to tension? Just make might and haste multiplicate and you'll solve that. 23:25:38 DrPraetor, that works fine for the first tension-based effects (Demonic Guardian) because tension calculations take effect the turn they're calculated. Eronarn's system ramps up tension and smooths out these spikes. 23:25:48 Cryp71c: non-xor - i think it makes sense for summ spells to add a flat boost to tension, but other stuff to be multipliers 23:25:51 DrPraetor, they factor in only after they're applied to a monster. 23:25:56 though i don't think this needs to be a formal system, either 23:26:05 Eronarn, no, it woud be too bulky. 23:26:11 it would be way faster to just look at the XP value of monsters and say 'hey, that monster should really be way more tension-causing than its XP value would indicate' 23:26:22 and assign them either a flat boost or a multiplier 23:26:29 Cryp71c: I think that's fine, actually. If you meph cloud Agnes before she hastes herself, should the tension really go up because she could in theory cast haste? 23:26:31 can just do a big switch statement 23:26:57 Cryp71c: Alternatively, for the few monsters who know haste and might, they could have a flat bonus to tension if we really care. 23:27:36 DrPraetor, why give flat bonuses when we can cover the whole field of monsters (as well as any that might get implemented down the road) with a less-flaggy, more (simply) predictive model. 23:28:09 Cryp71c: I think we're running in circles here. Such a predictive model would start with tension flags for the individual spells.... yes? 23:28:28 Eronarn, IMO, take the list of effects we already factor into a monster's tension and build values for how they should add (or multiply) tension for that monster if the monster is capable of effecting those effects. 23:28:41 Cryp71c: Or are you proposing something more similar to the expected damage output thing I mentioned originally? I don't really think that's practical. 23:28:44 DrPraetor, nope, merely checking for the possibility of a spell is sufficient, which would not require flags. 23:28:55 Cryp71c: i think that this works ok for some stuff, but that we should try to keep the list of effects pretty small 23:29:02 Cryp71c: It would require flags *on the splels* so that you know which spells to check. 23:29:04 Eronarn, it is fairly small. 23:29:15 DrPraetor, just hard code the list as they're coded already in the tension system. 23:29:20 and i think that we need to address monsters regardless because some monsters just plain stand out 23:29:25 Eronarn, anyways, so build this array of values for monsters which have these dangerous spells. 23:29:38 like giant eyes 23:29:43 they should give a ton of tension 23:29:47 Also, as due pointed out, this wouldn't work for red wasps (for example), which contribute a tiny fraction of the tension they should. 23:29:50 And floating eyes. 23:30:05 Eronarn, correct, and calculate the threat they represent with spells factored in and compare it to the threat value based on xp (and modified however else we're going to modify it) 23:30:08 And those blinking eye things even more so, since they confuse *and* they're tougher to kill. 23:31:04 Eronarn, if one is unusually low (eg, more than two standard deviations) always use the higher tension value. 23:31:39 Eronarn, get how I mean? Do the two simple calculations and compare them, if they're SUPER off, we're probably considering something with an unusually low XP value and should switch to the "backup" tension mode. 23:32:27 Cryp71c: Anyway, I don't want to ignore you if you address things to me, but I wanted to bow out some time ago so hopefully you and Eronarn can work out something useful? 23:33:19 DrPraetor, sure, thanks for your input, don't worry about Eronarn :) as a dev its frustrating to hear ideas that you know won't work but the proposer doesn't. 23:33:59 DrPraetor, and that's not bad, ideas drive the development of crawl. As someone with less knowledge of the source than we, you're held to fewer limitations in your ideas. we tend to think inside the box because we know we don't want to code outside of it :D 23:35:14 it would probably be a good idea to make a wiki page for tension if there isn't one already 23:36:05 Nothing in brainstorming. 23:39:15 !tell kilobyte Nice Shoals splat, very tasteful <3 23:39:16 greensnark: OK, I'll let kilobyte know. 23:39:25 butchering message while in spider form is odd (https://crawl.develz.org/mantis/view.php?id=3782) by reid 23:41:06 greensnark: currently I'm trying to figure out why resting or traveling via process_keys doesn't seem to work 23:41:13 LuckyNed the Cleaver (L19 DSCr) (Snake:2) 23:41:31 How does it fail 23:41:42 !lm * crash -log 23:41:43 827. LuckyNed, XL19 DSCr, T:42894 (milestone): http://crawl.akrasiac.org/rawdata/LuckyNed/crash-LuckyNed-20110410-044112.txt 23:42:23 greensnark: for resting (via crawl.process_keys("5"), say) you just get the message _You start resting. with no turns passing 23:42:25 I recall having troubles with it too 23:42:54 -!- Cryp71c has quit [Quit: Leaving] 23:42:55 I didn't get a message but in retrospect that's probably because I muted it 23:43:06 elliptic: Does coroutine.yield not help? 23:44:04 I remember in a few cases, but I don't know about the exact conditions, the macro never gets resumed after yielding 23:44:39 greensnark: doesn't seem to matter 23:44:52 Sounds like the rest is getting interrupted by something 23:45:32 elliptic: Can you file a bug and attach the code that tries to rest 23:45:49 I'm off now, but I'll see if I can check it out tomorrow 23:46:06 sure, will do... and thanks for all the help :) 23:48:50 thanks for the lorc stuff, btw, greensnark 23:49:28 unmelding a vampiric weapon is likely instadeath (https://crawl.develz.org/mantis/view.php?id=3783) by reid 23:50:11 nice 23:52:50 -!- Pseudonut has joined ##crawl-dev 23:59:31 crawl.process_keys("5") doesn't work (https://crawl.develz.org/mantis/view.php?id=3784) by elliptic 23:59:43 @?giant snail 23:59:44 unknown monster: "giant snail" 23:59:53 @?elephant snail 23:59:54 unknown monster: "elephant snail" 23:59:58 @?agate snail 23:59:59 agate snail (09j) | Speed: 4 | HD: 14 | Health: 77-102 | AC/EV: 7/2 | Damage: 18 | Flags: amphibious | Res: 06magic(56) | Chunks: 09poisonous | XP: 83.