00:00:55 -!- RichardHawk has quit [Ping timeout: 258 seconds] 00:01:05 Unstable branch on crawl.develz.org updated to: 0.10-a0-763-ge0aca7a (32) 00:07:21 seemingly random assert in store.cc (https://crawl.develz.org/mantis/view.php?id=4570) by Galefury 00:09:01 Windows builds of master branch on crawl.develz.org updated to: 0.10-a0-763-ge0aca7a 00:19:13 guess that should be assert failure 00:19:17 oh well 00:19:19 good night 00:19:35 not fixed in the just updated build btw 00:19:51 -!- Galefury has quit [Quit: Stay sane inside insanity!] 00:48:30 -!- cbus has joined ##crawl-dev 01:07:05 -!- ldierk_ has quit [Ping timeout: 260 seconds] 02:23:14 -!- elliptic has quit [Ping timeout: 276 seconds] 02:27:31 -!- ainsophyao has joined ##crawl-dev 02:31:58 -!- galehar has joined ##crawl-dev 02:32:05 hi 02:32:06 galehar: You have 1 message. Use !messages to read it. 02:32:12 !messages 02:32:12 (1/1) kilobyte said (8h 16m 56s ago): uhm, ede91018e189 uses progress to next level, not to next skill rank 02:32:47 !tell kilobyte yeah, I realized that this morning. Sorry for the idiotic commit :S 02:32:48 galehar: OK, I'll let kilobyte know. 02:41:24 Unstable branch on crawl.akrasiac.org updated to: 0.10-a0-763-ge0aca7a (32) 03:06:17 galehar: no problem, combat does crash at certain skill values, so yours is not the only error before this can go in anyway 03:06:17 kilobyte: You have 3 messages. Use !messages to read them. 03:07:39 Ash's skill boost and spell slots will need to be addressed too. 03:09:12 you gain one per 0.5 skill, seems good enough to me 03:09:58 this can't be randomized, so there needs to be _some_ discrete point you gain an additional slot 03:13:22 Teleported into doorless room (https://crawl.develz.org/mantis/view.php?id=4571) by dashcunning 03:16:45 kilobyte: we can also factor in partial XL. Something like: (you.xl(100) + you.skill(SK_SPELLCASTING, 200)) / 100 03:17:27 but in both case we probably need to put a message for when a slot is gained (unless you don't have spellcasting or you worship trog) 03:37:32 -!- ais523 has joined ##crawl-dev 03:38:09 shouldn't we move gnolls a bit deeper with the new reaching? Or maybe spawn them alone on first levels and in pack deeper. 03:38:53 galehar: Isn't that already how bands are handled/ 03:38:55 or at least, ood bands 03:40:03 not sure, will look it up 03:40:36 -!- monqy has quit [Quit: hello] 03:40:36 From memory, you can't have gnolls pack on D:1, but they do show up on D:2, which might be too deadly with reaching. 03:40:46 yeah, the extra turn of dealing damage, and having to fight two+ gnolls at a time makes them a good bit more dangerous 03:46:34 -!- ais523 has quit [Read error: Connection reset by peer] 03:46:54 -!- galehar has quit [Quit: Page closed] 03:47:11 -!- ais523 has joined ##crawl-dev 03:47:56 -!- galehar has joined ##crawl-dev 04:23:16 -!- mumra has joined ##crawl-dev 04:28:05 -!- st_ has quit [Ping timeout: 260 seconds] 04:29:20 I mentioned that days ago but no one listened :( 04:31:54 kilobyte: in _olgreb_pluses you didn't use skill_rdiv. Oversight? 04:32:34 ghallberg: it happens. You just need to keep shouting :P 04:32:43 :) 04:33:36 galehar: no, it's important to not use randomness in places like that 04:34:24 you.skill() is deterministic, you.skill_rdiv() is not. Perhaps they could be named differently to make this more clear? 04:34:52 03kilobyte * r786963fadc3d 10/crawl-ref/source/util/art-data.pl: Add a missing dire warning to dc-unrand.txt 04:34:56 I think function names are ok 04:35:03 03kilobyte * r41475ae75f4d 10/crawl-ref/source/ (10 files in 2 dirs): Remove the code for ARTP_PONDEROUS, fix the hat of Pondering not working. 04:35:03 03kilobyte * rfb00bbc03743 10/crawl-ref/source/items.cc: Tell Cheibriadites their Orb run spawns are not as fast. 04:35:03 03kilobyte * r304cd6d8e2b6 10/crawl-ref/source/store.cc: A more helpful crash message on non-existant prop asserts. 04:36:12 would it change anything to use something like this? you.skill(SK_POISON_MAGIC, 10) / 30 (or some other scaling) 04:36:54 no, not at all -- you'll still get an extra plus only at whole 3, 6, 9, 12, ... skill. 04:37:46 anything that affect targetting range and can be freely cancelled (like spell power) can't be randomized (haven't ran into it yet, just sayin') 04:39:48 yeah, that's why I'm using skill() everywhere unless I feel safe with randomness 04:40:04 in _WUCAD_MU_evoke you use div_rand_round(you.skill). Why not using skill_rdiv right away? 04:41:00 because I edited art-func.h first, before adding skill_rdiv() :p 04:42:12 fixed 04:44:53 I think duration calculation for BEAM_ENSLAVE_SOUL can use skill_rdiv 04:45:34 moin 04:47:19 hi 04:48:21 galehar: in general, randomness in such places is a bit bad, but you're right, here you might get an extra turn, that's better. Fixed. 04:49:23 galehar: did you change player_exp_progress to skill progress already (and didn't push yet), or not yet? 04:49:38 no I haven't, you can fix it 04:51:06 but maybe keep the player_exp_progress function and add another one for skill (or do it right in you.skill) 04:52:08 hmm. Well skill progress can have a function on its own, since it will be called both by you.skill and the skill screen (progress will still be visible in wizmode) 04:52:58 yeah, get_skill_percentage() 04:54:55 another occurence of a possible use of skill_rdiv: evoke.cc:928 04:55:28 maybe call it get_skill_progress() since the scale might be different than 100 04:58:59 ... that's how I did name it :p 05:00:05 OSX builds of master branch on crawl.develz.org updated to: 0.10-a0-763-ge0aca7a 05:00:07 sorry wrong line number. In evoke.cc, I think skill_rdiv can be used at lines 391,402,501 05:00:55 kilobyte: is ponderous armour brand being removed entirely? 05:01:37 mumra: it can still spawn on plate mails -- they get huge initial enchantment in return 05:01:42 aaa 05:02:05 and there's the Hat of Pondering and Black Knight's horse barding 05:02:10 kilobyte: just checking; it was referenced in vault docs 05:16:55 -!- Amonchakad has joined ##crawl-dev 05:24:17 I forget, did we figure out what the credentials needed for accessing backed up saves on CAO are? I want to try to debug #4511.... 05:25:03 Zaba: I seem to have access, link me to the save/ 05:25:04 *? 05:25:19 https://crawl.akrasiac.org/saves/dumps/Tenaya-crawl-git-04fb28ffe7-110908-1639.tar.bz2 05:26:28 got it 05:27:36 http://crawl.develz.org/other/Tenaya-crawl-git-04fb28ffe7-110908-1639.tar.bz2 05:27:51 thanks 05:27:58 !tell Napkin BTW, ~crawl/website/other is 1.2gb :0 05:27:58 due: OK, I'll let Napkin know. 05:34:15 -!- Amonchakad has quit [Quit: Page closed] 05:47:31 yay, it works! A starting DrTm can memorize Blade Hands at Useless with 2.25 Tmut skill. With 2.70 it is at Terrible. 05:48:34 03kilobyte 07smooth_skills * r548fc43300bc 10/crawl-ref/source/ (19 files): Re-allow shortening you.skill(X, 1) to you.skill(x). 05:48:34 03kilobyte 07smooth_skills * r6f9267b15d22 10/crawl-ref/source/skills2.cc: Remove two ancient unused sets of titles. 05:48:34 03kilobyte 07smooth_skills * r7458b757a99f 10/crawl-ref/source/art-func.h: Shorten a div_rand_round(you.skill()) to you.skill_rdiv(). 05:48:35 03kilobyte 07smooth_skills * r2cae4c2dafe1 10/crawl-ref/source/ (player.cc skills2.cc skills2.h): Actually return partial skill values, not progress to next character level. 05:48:35 03kilobyte 07smooth_skills * r0146b630395f 10/crawl-ref/source/ (beam.cc evoke.cc): Use skill_rdiv() in a few more places where randomness is ok. 05:48:35 03kilobyte 07smooth_skills * ra51d98d6c606 10/crawl-ref/source/random-var.cc: Fix random_var division by a number other than 2. 05:48:45 03kilobyte 07smooth_skills * r44e7ef5f7973 10/crawl-ref/source/ (fight.cc random-var.cc random-var.h): Implement div_rand_round() for random_var, use it for combat delay. 05:58:33 is there a git shortcut to point to the branch source (the commit where it diverges from the master)? To easily see the whole diff of a branch. 05:59:38 not sure if there's a faster way, but I use git diff `git merge-base HEAD master` HEAD 06:00:09 doesn't "git --format-patch [branch]" generate a patch of current branch against [branch] ? 06:00:22 (at least, multiple diffs) 06:01:07 (hmm, i realise that's not what you're after) 06:01:19 kilobyte: thanks 06:01:25 mumra: if you want multiple diffs, yeah. This is for squashed. Especially useful if you want, say, differences in fight.cc to merge into the combat rewrite 06:10:43 kilobyte: why base damaged of unarmed attacks can't have random rounding? It's already the case for forms. 06:12:16 they can, it just needs to return a random_var 06:12:19 -!- jle has joined ##crawl-dev 06:12:19 oh, and you're right about forms 06:12:45 the function doesn't use random_var, does it? 06:14:33 btw, shouldn't we tone down UC a bit? It's nice that it's usable in the endgame, but the fact that it's better than most weapons is a bit weird. 06:14:44 oh, indeed damage estimations don't use random_var yet 06:14:57 or rather, we have no damage estimations, just attack speed ones 06:15:02 galehar: +1 06:16:24 UC is an odd bird. It's better than most everything on the face of it, but it can get no brands or bonuses apart from slaying. 06:16:25 and to_hit which is disabled for some reason 06:16:47 it has forms instead of brands 06:17:55 I think UC in the endgame should be viable only for Tr/Gh and forms 06:17:55 Forms tend to have other tradeoffs, like items you can wear. 06:18:08 03kilobyte 07smooth_skills * r656daccb5fed 10/crawl-ref/source/fight.cc: Randomly round skills in unarmed damage calculations. 06:18:37 Blade Hands, Necromutation... 06:18:43 kilobyte: a very minor simplification possible at fight.cc:3552 06:19:01 In the endgame, the appropriate comparison would seem to be a foo of holy wrath. 06:19:23 Er, extended I should say. 06:19:56 galehar: ah, 2*100, I see 06:22:08 03kilobyte 07smooth_skills * rb0a0e1540638 10/crawl-ref/source/fight.cc: Shave off an unneeded multiplication. 06:22:27 -!- ais523 has quit [Read error: Connection reset by peer] 06:23:25 -!- Fail_Bot has quit [Ping timeout: 260 seconds] 06:24:44 what with the display? Showing the skill as 2.1 looks better to me over other proposals by quite a lot. 06:25:37 -!- ais523 has joined ##crawl-dev 06:29:04 -!- ais523_ has joined ##crawl-dev 06:29:04 -!- ais523 has quit [Disconnected by services] 06:29:04 -!- ais523_ is now known as ais523 06:42:59 I was thinking of just hiding the decimal 06:43:12 it never has any real tactical or strategical value 06:44:00 if it did not, we wouldn't be smoothing skills 06:44:51 if you show it, it means player will consider it as a significant information and will try to make something out of it. The point is that you don't have to bother if you're at 7.2 or 7.8. 06:46:25 it does tell people they don't have to bother about 7.5+7.5 vs 7.0+8.0 06:46:32 knowing that you are between 7 and 8 is enough for any decision. 06:46:43 Is skill smoothing currently going to be just that, or are there e.g. balance changes for weapons as well? 06:46:48 might be 06:46:50 -!- syllogism has joined ##crawl-dev 06:47:22 jle: skill smoothing is just that. We might introduce other changes and balance, but they won't be related to it. 06:47:24 jle: as of 0.9, players tend to micromanage skills so there's no big change 06:48:03 the point is that training skill X has the same return on investment wether you're at 7.1 or 7.9. Which isn't the case at all right now 06:48:08 jle: for micromanagers (ie, advanced spoiled players), the difference is on the average half of a point in a single skill 06:48:36 I cna finally stop waiting for the skill-up message. 06:49:05 I KNOW the XP to get to 5/10% doesn't matter but it's still non-optimal... 06:49:08 the display question aside, do you think we can risk merging already? 06:49:15 kilobyte: Yeah, I like this skill fuzzing idea. The difference in e.g. success rates for spells can be pretty major for a mere one skill point. 06:49:49 kilobyte: I'm not opposed to experimenting with showing the decimal, but I think ultimately, it will just be worthless detail and we'll be better without it. 06:52:31 kilobyte: not much risk in merging if you tested it a bit. There's still details to be addressed, but we can do it in trunk as well. 06:52:32 galehar: The one issue with no decimals I can think of offhand is spellcasting giving spell slots. Presumably now you will get one slot every half level, since iirc you get two for each level now. 06:52:32 exactly. If you show the decimal, then players will micromanage their spellcasting skill to focus it when it's at .4 or .9 06:52:32 the bit in "testing a bit" is way too small, but using poor trunk players is more efficient. For us, not them :p 06:52:42 of course :P 06:53:56 I'm all for :F 06:54:09 Let there be blood. 06:54:44 godabil.cc:2999 constify skill variable 06:54:58 -!- elliptic has joined ##crawl-dev 06:55:07 hi elliptic 06:55:15 hi 06:55:15 elliptic: You have 1 message. Use !messages to read it. 06:57:00 elliptic: what do you think of having a single upstairs on the first level of branches? One entry, one exit. 06:57:44 We keep hatches for connectivity. Might make them normal hatches or keep converting them to branch exits. 06:59:11 galehar: it would be best if there were exactly one branch exit, of course 06:59:17 galehar: on a local scalar with no pointer to it, being const vs non-const has precisely zero effect 06:59:39 galehar: maybe have a single branch exit and make all hatches point downwards? 06:59:40 it might at most be documentation 06:59:54 kilobyte: ok 07:00:45 elliptic: if it doesn't break connectivity, sure it's better. 07:01:19 right, I don't know what the connectivity issues are exactly... might be hard to make work e.g. in Orc 07:01:41 it does bother me that there currently are multiple < which take you to the same > 07:02:08 03kilobyte * rd45d96fbf38c 10/crawl-ref/source/ (36 files): Merge branch 'smooth_skills'. 07:02:23 yeah, I'm not sure either. Seems there's still some connectivity issues since the builder sometimes go crazy and put hatches everywhere. 07:02:57 * galehar stashes his AI work and goes back to working on skills 07:03:54 galehar, kilobyte: good work with the skill smoothing :) 07:08:59 connectivity issues would be in orc and slime, yes 07:09:36 some branches that end up with a layout_basic level for their first level, too.. 07:17:37 Zaba: btw, I don't understand why some vaults can disrupt the connectivity checking and make the builder put hatches everywhere since connectivity is done with a floodfill. 07:18:20 unless a vault is tagged 'transparent', it's assumed to be opaque for connectivity purposes 07:18:28 regardless of its actual terrain 07:19:27 it's clearly done that way on purpose, but don't ask -me- what the purpose is :P 07:22:21 doesn't make much sense to me. 07:22:27 maybe because since vaults can be dangerous, we don't want to force players through them. Except that we do. 07:23:07 galehar: i'd like it if entry vaults (some, at least) contained three staircases that connected to each of the 3 staircases in the branch ... so you have options if, say, the normal staircase is surrounded by mobs 07:23:49 maybe it should be the other way around. By default, they are traversable, but if they contain something very nasty like mutagenic clouds, you tag them opaque 07:24:05 03kilobyte * r88c6798f5a5d 10/crawl-ref/source/ (output.cc player.cc player.h): Revert player_exp_progress(), fixing build failure. 07:24:12 mumra: but there's only one downstairs 07:24:15 branch entry 07:24:16 galehar, a comment in dungeon.cc says, "Don't peek inside MMT_OPAQUE vaults (all vaults are opaque by default) because vaults may choose to create isolated regions, or otherwise cause connectivity issues even if the map terrain is travel-passable." 07:24:31 which is rather vague, but isn't talking about gameplay purposes 07:24:47 galehar: yes, it would mean changing entry vaults so they could have 3 entrances to the branch 07:25:34 mumra: I find it thematic that there is a single entrance to a branch. I like that we changed Zot to have a single portal for example. 07:25:53 Zaba: imagine if on D:1, all the downstairs were on the other side of one of those flame cloud crossroad vaults 07:25:57 the game would probably be unwinnable 07:26:18 galehar: certainly i didn't think it would be appropriate for all branches (i didn't know about that zot change tho) 07:26:23 (flame cloud vaults create travel exclusions only when you see them, so the game would think the level was connected until that point) 07:26:40 I like mumra's idea. Aside from places specifically designed to give you a nasty surprise (Elf, Vaults), it's not particularly fun when a branch is blockaded by some random unique or, let's say, hydras. 07:26:51 ais523: well, such vaults could be tagged opaque. I'm not saying the mechanism is useless, just that the default behaviour should be reveresed 07:27:07 Unstable branch on crawl.develz.org updated to: 0.10-a0-780-g88c6798 (32) 07:27:11 jle: especially if rune lock happens :) 07:27:12 galehar, I've thought of that, too 07:27:32 jle: the most fun I've had in Crawl was when Sonja was near the only entrance stair in Lair 07:27:41 galehar, it would take a lot of seeking out the vaults that need to be tagged 'opaque' and testing, though 07:27:49 as it forced me to actually come up with a plan for fighting her, rather than run away, as is the best strategy for use against basically every unique in existence 07:27:52 primarily testing 07:27:52 apart from the really easy ones 07:28:12 having three stairs between levels, pretty much its only gameplay effect is to make running away better 07:28:23 and providing annoyance in trying to explore levels fully 07:28:32 Zaba: sure, but's that the perfect timing for this kind of changes. We have time to test before release. 07:29:26 I suppose you could write a script to tag all vaults opaque, change the default 07:29:39 then start removing opaque tags by hand when you decided that the vault didn't cause a connectivity problem 07:29:44 that'd be a safe way to solve the issue 07:30:40 ais523: otoh, that could've been game over if your character for whatever reason just wasn't good enough to take sonja on or progress deeper in D 07:30:43 ais523: I can see this in places where it's your own fault for going there (blade, elf, vaults) but several uniques at the moment (sonja for example) are a case mainly of finding the right resistance. 07:31:02 mumra: well, yes 07:31:16 but the thing is I actually got to use tactics, which is very rare in Crawl normally (it's normally a game purely about strategy) 07:31:27 ais523: i think many would disagree 07:31:43 perhaps 07:31:47 ais523: whenever i meet a unique, i figure out if there's a tactical way to beat them; and if it's possible, i do 07:31:59 well, I agree that having a single entry to a branch sometimes make things interesting. 07:32:08 mumra: oh, I consider "having a plan to beat single tough monsters" to be part of strategy 07:32:25 the thing is that you come up with the plan in advance, and it doesn't really matter whether the opponent is, say, Sigmund or Duvessa 07:32:31 your tactics are likely to be the same either way 07:35:39 I suppose autorobin is the extreme example, where the tactics are defined before the start of the game and never deviated from 07:39:51 -!- Mu_ has joined ##crawl-dev 07:45:18 -!- galehar has quit [Ping timeout: 252 seconds] 07:55:37 * greensnark mews at Mu_ 07:57:01 I like the double tourney peak for 2011 07:57:18 * due waves greensnark. 07:57:48 Hey due 07:57:57 Good evening 07:59:30 hi 07:59:49 I need to play crawl again, so many exciting changes :) 08:00:12 I feel like Rip van Winkle 08:02:09 -!- ainsophyao has quit [Ping timeout: 265 seconds] 08:04:40 -!- valrus has joined ##crawl-dev 08:05:55 -!- edlothiol has joined ##crawl-dev 08:09:00 -!- galehar has joined ##crawl-dev 08:19:15 -!- greatzebu has quit [Ping timeout: 258 seconds] 08:19:51 what's with the new warnings? 08:19:52 warning: format not a string literal and no format arguments 08:20:15 galehar: it's a potential security bug 08:20:29 if you do mprf(foo), and foo contains "%n", then it ends up trashing memory 08:20:41 you should use mpr(foo) or mprf("%s", foo) instead 08:20:53 likewise with sprintf, and other -f functions 08:21:02 and it's pretty easy to fix, too 08:21:23 I remember fixing all the cases of that in NetHack, there were over a hundred but fixing them was pretty trivial by just adding "%s", in all the appropriate places 08:21:25 I only got them recently. Is it a code change or a compiler upgrade? 08:21:30 compiler upgrade, I imagine 08:21:51 gcc started doing those warnings even when warnings weren't requested, because they're a) potentially really bad, and b) trivial to fix 08:22:52 -!- Twilight-1 has quit [Read error: Connection reset by peer] 08:22:54 ok thanks 08:24:09 -!- Twilight-1 has joined ##crawl-dev 08:25:10 but this issue must have been there since printf came into existence. it only took them ~20 years to put that as default into -Wall. 08:25:18 heh 08:25:26 and I seem to remember you "fixed" it in Un by turning the warning of 08:25:27 *off 08:25:31 it really isn't that hard to fix, you know 08:25:48 (it's probably easier to fix than to work out if %n can possibly get into the format string or not) 08:27:05 yes, i know, but touching that many lines of code just for something this trivial. it's just not motivating. there is not even a vanilla bug number for that 08:28:02 I don't really measure bugs in terms of number of lines of code touched, but in difficulty to fix 08:28:06 and %n security bugs are very easy to fix 08:30:40 the question is, *why* does it trash memory? surely it's a compiler or *f problem in the first place? 08:31:03 -!- RichardHawk has joined ##crawl-dev 08:33:22 because %n means "place the number of characters read in the memory pointed to by the next argument given" 08:33:44 and there isn't a next argument, so it interprets whatever happens to be next on the stack as an address and writes to that 08:33:59 -!- Pingas has joined ##crawl-dev 08:35:12 oic 08:36:51 i'm pretty sure there could still be an issue ... e.g. mprf(foo, bar) ... foo could contain "%s %n", right? 08:36:55 -!- dpeg has joined ##crawl-dev 08:36:58 cheers 08:37:30 mumra: yes, but normally when you're giving more than one argument, it's because you know that the first is formatted like a printf format string 08:37:34 and so will contain the right number of elements 08:37:35 -!- Textmode has joined ##crawl-dev 08:37:56 whereas when you're giving one, often you're just trying to print a string and have forgotten that % is interpreted specially 08:38:53 would still be safer to have mprf etc. check the number of arguments ... it'd be really easy to pass in a string with one too many 08:39:42 it can't, based on how C argument-passing conventions work 08:39:53 if you consider this a deficiency in C, you're probably right 08:40:01 i guess that's why i usually work in more modern languages where that kind of thing gets caught :) 08:40:03 and C++ has the same problem 08:42:49 -!- MarvinPA has quit [Read error: Connection reset by peer] 08:46:50 mumra: the first argument to printf, resp. mprf is usually not coming from user input, so you don't have this issue. 08:47:08 even arguments other than %n can be an information leak 08:49:38 bhaak: yes but it could still be coming from other functions, or from external text files 08:54:11 03galehar * rda0fd2e87351 10/crawl-ref/source/ (main.cc newgame.cc player.cc): Fix some compiler warnings. 08:54:15 03galehar * rdba47fdf7d3b 10/crawl-ref/source/ (main.cc output.cc player.cc player.h): Reintroduce get_exp_progress(), properly this time. 09:00:43 -!- valrus has quit [Remote host closed the connection] 09:09:57 hyperbolic (L13 OpHe) ASSERT(smc) in 'mon-util.cc' at line 1496 failed. (D:13) 09:10:08 hm, not sure what happened there 09:10:20 !lm . crash -log 09:10:22 13. hyperbolic, XL13 OpHe, T:26421 (milestone): http://crawl.develz.org/morgues/trunk/hyperbolic/crash-hyperbolic-20110919-140953.txt 09:14:51 doesn't the dump tells which monster is acting usually? 09:27:33 Ozymandias the Skirmisher (L1 DsCK) ASSERT(!crawl_state.prev_cmd_keys.empty()) in 'main.cc' at line 4481 failed on turn 0. (D:1) 09:37:18 galehar: apparently ash skill boosts got messed up by the smooth skills 09:39:49 looks like it is just a matter of ash_skill_boost not scaling level 09:51:05 -!- ainsophyao has joined ##crawl-dev 10:02:38 -!- MarvinPA has joined ##crawl-dev 10:03:11 I need help with git :( 10:05:47 -!- greatzebu has joined ##crawl-dev 10:10:34 03elliptic * r45d1378db704 10/crawl-ref/source/godpassive.cc: Fix Ash skill boost to handle skill scaling. 10:11:35 mumra: well, usually the format argument is passed in to *something* as a string argument 10:11:52 er, s/string argument/string constant/ 10:12:46 someone with the ability to make CDO update (due, kilobyte, Napkin, ?), could you push the button? since currently ash is unplayable there 10:14:14 -!- greatzebu has quit [Read error: Connection reset by peer] 10:14:25 -!- greatzebu has joined ##crawl-dev 10:14:45 (and, given the way varargs work in C (and C++), that has to be the one function that actually has ... in its parameter list) 10:15:47 so we try to tell GCC about all such functions so it can check the format strings 10:16:08 dpeg: what's the issue? 10:16:14 I'm far from a git expert, but might be able to help 10:20:37 mumra: not necessarily perfect -- that obviously doesn't help with MSC builds, and some format specifiers may even have portability issues that only kick in there, or we could have typedefs that stand for different types there -- but it's at least better than nothing... 10:21:24 mumra: so if you run into any printf-style functions that lack the special incantations in their declarations, just give a hollar 10:22:18 elliptic: updating CDO right now. 10:22:58 elliptic: I want to change to formula so that it gives skill points instead of skill level. 10:23:08 mumra: and any places where the format argument to a printf-style function is coming from outside should probably have a "%s" there instead... 10:23:25 *probably* 10:23:44 galehar: you mean instead of the minus level/4 stuff? 10:24:07 that might make more sense, yes 10:25:07 Unstable branch on crawl.develz.org updated to: 0.10-a0-783-g45d1378 (32) 10:26:32 -!- jooosa has joined ##crawl-dev 10:26:33 w00t 10:26:43 an actual merge commit! 10:27:10 elliptic: yes, it would be smoother 10:27:12 sorry, work called 10:27:21 just have to find a formula. 10:27:34 -!- joosa has quit [Write error: Broken pipe] 10:27:42 -!- jooosa is now known as joosa 10:28:12 I have pulled mumra's demigods branch (and it works), but now I am unable to pull on master. "It looks like git-am is in progress. Cannot rebase." (This is probably unrelated to branch switching.) 10:29:43 elliptic: the change will probably have some effect on balance. In your opinion, is the Ash skill boost too strong or too weak at certain skill levels? 10:30:18 dpeg: it sounds like it's in the middle of a patching application, and hit a conflict 10:30:34 ais523: yes. How can it tell git to abort and pull? 10:30:37 you can cancel the whole thing with git am --abort 10:30:42 thank you 10:30:49 and then the pull should work normally 10:31:30 It does. So the demigods adventure just added to my confusion but not to the problem. 10:32:14 galehar: currently the spellcasting boosts are extremely strong when you first get them at max (or near-max) strength, since they'll raise a lot of skills from 7 to 12 and such 10:33:09 From the reactions I've seen they seem a bit too good, I agree. 10:33:10 galehar: that's partly because the ratio in skill points needed to get L7 vs L12 is fairly large, I think... perhaps it would be good to have the boosts be multiplicative? 10:33:45 that is, you get 20% more skill points or something rather than a fixed increase 10:34:29 the skill boosts continue to be good all game after that, too, but they do definitely get worse 10:34:57 -!- greatzebu has quit [Ping timeout: 258 seconds] 10:35:15 the boost gets worse also because many skills are not worth training to L27 10:35:32 thanks for the feedback, will try to run some numbers later 10:35:34 whats the reasoning behind nerfing evaporate? :) 10:35:50 cbus: really need to ask? 10:36:01 galehar, well if you play chars that have problems casting it I do :) 10:36:18 its one thing if its like a SE or something but Op have issues to get to spider form :) 10:36:53 it's L2 and dual school. It's much easier to cast than mephitic 10:36:55 galehar: well, even if you want to train skills to 27, the difference between 24 and 27 is always going to be much less significant than the difference between 7 and 12 10:37:12 and xp is plentiful later on 10:38:14 fully bounded can even raise your skill from 7 to 14. 10:38:49 so, I'll aim for a flatter bell curve 10:50:10 03kilobyte * r6cbf123c677b 10/crawl-ref/source/misc.cc: Refactor stop_attack_prompt(). 10:50:23 -!- blueDave has quit [] 11:05:32 -!- bmh has joined ##crawl-dev 11:05:49 afternoon 11:06:45 Hi! 11:09:23 -!- edlothiol has quit [Quit: Page closed] 11:12:40 dpeg: did I miss any exciting crawlly stuff while I was away? on my end, I managed to make it out of Nepal right before the earthquake 11:13:00 bmh: did the earthquake cause casualties? 11:13:49 dpeg: a handful in Kathmandu, more in India, but probably no worse than the weekly toll from road fatalities 11:14:35 New abyss is in, with all effects, I think. It did drew complaints, however. (Forum, I think.) Cannot really judge, though. 11:14:57 dpeg: link? 11:15:28 https://crawl.develz.org/tavern/viewtopic.php?f=5&t=2581 is one 11:16:25 "I once lost a high-level Vampire caster in the Abyss, because I was resting in bat form over a lava tile in one of those little corners that was closed off" 11:16:36 That's the fault of the player entirely. 11:17:00 https://crawl.develz.org/wiki/doku.php?id=dcss:brainstorm:dungeon:branch:abyss#abyss1 is possibly more useful feedback 11:17:10 MarvinPA: ah, right, that was the other source. Thanks! 11:19:20 hm... I think I can answer st's complaint by adding a 'dampening zone' 11:19:46 reduce the abyss shift around the player, perhaps in an ellipsoid centered on the player and rotated based on direction of travel 11:20:01 bmh: galehar and kilobyte did a lot for the Abyss lately, perhaps talk with them 11:21:06 dpeg: will do 11:21:47 the latest version is live on CDO? 11:21:52 bmh: that's a pretty cool idea! 11:22:40 -!- Zaba has quit [Ping timeout: 259 seconds] 11:22:40 I have zero sympathy for the guy who died in batform over lava. 11:22:57 Just say NO to fried bats! 11:23:26 There's an abyss rune vault with a kraken? 11:23:28 That's just funny. 11:23:31 should be "roasted", right? I am not good with the cuisine 11:23:52 Players complain about lack of abyssical vaults... instead of sending some. (Well, we got a few.) 11:24:11 there's some pretty cool ones, i think they're still quite rare though 11:24:15 bmh: also, instead of call _abyss_wipe_square, just move the items to the nearest floor cell. Or morphing just eats all the loot 11:24:57 galehar: ah-ha. Are you special casing vaults so that the player can't sit outside and wait for the toys to fall out? 11:25:19 vaults are completely immune to morphing now 11:25:34 might add a tag to allow some morphing to some of them 11:25:50 but since most of them are either rune or exit, I'm not sure it's worth it 11:26:00 MarvinPA: just up their WEIGHTs! 11:26:21 or add more :D 11:26:45 in the past people seemed to be of the opinion that vaults in the abyss *should* be rare 11:26:49 we will get more vaults once players see vaults in action ==> increase WEIGHT to get more :) 11:26:57 but yeah, they probably should be pretty rare 11:27:00 because abyss is supposed to be mostly unstructured 11:27:08 and i've not done abyss recently to see what it's like 11:27:08 dpeg: I woke up in a very jetlaggy state this morning and thought it might be nice to have a Kiku "pile of bone" overflow altar 11:27:10 elliptic: yes, but players want easier access to rune/exit 11:27:38 dpeg: not convinced that we should give this to them :P 11:27:44 dpeg: I'm of the mind that the abyssal rune should be as hard to get as any rune. 11:27:52 bmh: I think we recently got a second Kikubaaqudgha vault with a lot of lua trickery. 11:28:13 elliptic: that's the spirit! But you read the long rants "I spend five hours in the Abyss and then died" :) 11:28:38 bmh: well, that is served best with vaults (no matter how rare). Currently, it's more about tedium, afaiu. 11:28:54 -!- Zaba has joined ##crawl-dev 11:29:12 galehar: can you help me? 11:29:21 sorry, not right now 11:29:26 have to take care of the kids 11:29:26 dpeg: I avoid reading tavern unless someone calls my attention to a specific thread... :P 11:29:31 so no slime entry vault today 11:29:44 galehar: good call! 11:30:06 dpeg: !tell me or email what's it about, I'll have a look later 11:30:17 cool 11:31:18 would anyone be opposed to scaling abyss difficulty based on the source of banishment? 11:31:36 that sounds weird to me 11:31:48 abyss should be abyss :) 11:31:55 MarvinPA: Hah, now that you're here: we got informed that you have abused LRD! That's absolutely okay, as long as you help with nerfing it now!! :) 11:32:06 oh, i did? :o 11:32:27 what's the problem with LRD? 11:32:39 elliptic: miscasting summon ice beast probably shouldn't throw you as deep into the abyss as getting slammed by Roxanne 11:33:00 MarvinPA: doesn't matter if the rumour is true or not, as long as the nerf goes in. I've seen it in Elf:7 (with Ash, but you can also do it with antenna, and perhaps blind) and there is a story about a spectacular Zot:5 abuse. 11:33:15 oh right, hitting things through walls with it 11:33:55 hitting things through walls is somewhat abusable, yeah... but LRD has a habit of breaking the walls, so it isn't as bad as it could be 11:34:17 i did abuse that in sprint 4 i guess 11:34:33 bmh: well, I guess that could be okay if we were making the abyss get harder with time (and only a chance at the rune once it gets sufficiently hard) 11:34:34 MarvinPA: so what we need is to make sure that LRD will only damage monsters on the player's side of the wall. Two proposals (there might be more): (a) can only affect monsters which are in sight of the player, where all walls are treated as opaque. (b) Form line @ to wall, compute perpendiculars, only affect monsters in your half space. Does any one of them sound good enough? 11:34:49 bmh: but I don't see any need for it 11:35:27 elliptic: harder with something seems good to me... currently, we encourage players, or at least don't disencourage them, to stroll the Abyss for items. 11:35:34 elliptic: re: comments on the tavern. Players are complaining about their Pan Lord eating characters being abyssed to death, and others are griping that they miscast a low level summon and die. 11:35:44 elliptic: I find "harder with time spent" to be a natural metric. 11:35:49 bmh: don't listen to them :P 11:36:03 -!- Twilight has joined ##crawl-dev 11:36:11 people complain about anything that kills them 11:36:22 Miscasting Call Canine Familiar really shouldn't be able to cause a trip to the abyss, though 11:36:25 abyss is no different from anything else 11:36:28 -!- Twilight is now known as Guest94446 11:36:30 Very disproportionate compared to other schools 11:36:37 ??summoning miscast 11:36:38 summoning miscast[1/1]: 1: 5-13 dmg / sp. vortex / imp 2: 2-4 sp. vorts / common demon / 2-3 lesser 3: abomination / greater demon / 2-3 common / banish / hostile malign gateway 11:36:52 Popular complaint topics: Swamp, Abyss, labyrinths. 11:36:53 need a level 3 miscast for abyssing, same as with translocation 11:37:10 dpeg: harder with time spent is good, sure 11:37:27 elliptic: Saw it happen in ##crawl the other day. 11:37:36 You are adapting to the Abyss, and the Abyss adapts to you. 11:37:39 Perhaps low level spells should not be able to get L3 miscasts at all. 11:37:56 RichardHawk: sure, it happens... but people learn that casting spells at terrible is a bad idea 11:38:00 wouldn't you have to be trying to cast it at terrible to get that? 11:38:03 yes 11:38:11 for canine at least 11:38:36 in which case you could just as easily be getting level 3 tmut miscasts and berserkitis, or greater demons, or something equally nasty 11:38:46 -!- Twilight-1 has quit [Ping timeout: 252 seconds] 11:38:47 Hmm, never knew miscast severity is actually influenced by your success rate. 11:39:08 it is influenced extremely extremely strongly by your success rate :P 11:39:29 I would have assumed that severity was determined by something like: success rate and spell level -- anything else in the soup? 11:39:37 nope 11:40:30 anyway, making bad miscasts less lethal is a completely separate issue IMO... as marvinpa says, there are plenty of non-abyss ways for miscasts to kill you 11:41:03 Last time I had no luck, but perhaps this time: any opinions on (a) vs (b). It is an actual abuse, I've witnessed it, so I want a remedy for 0.10. 11:42:24 I don't really like either (a) or (b), to be honest 11:42:47 Same 11:42:52 I'd prefer something like this: 11:45:16 (c) LRD doesn't damage monsters unless there is a clear path to them (after the explosion possibly removes a wall) 11:45:30 m1nced (L10 DsIE) ASSERT(prog < scale) in 'skills2.cc' at line 150 failed. (Lair:1) 11:45:38 clear path on map or with LOS radius? 11:45:44 *within 11:45:53 elliptic: I like option (c) 11:45:57 Hmm. 11:46:02 where a direct path is a line from you to the explosion followed by a path within the 3x3 square centered at the explosion 11:46:04 All of these would deal with Elf and Zot, so I am fine with any. 11:46:06 or something like that 11:46:28 elliptic: we need specifics! :) 11:46:42 I am concerned about non-abuse scenarios 11:46:51 Such as this (pardon my ascii, tiles player) 11:46:53 ... 11:46:53 OOO 11:46:53 .=. 11:46:53 ... 11:46:53 ... 11:46:54 .@. 11:47:03 It should hit all three monsters here, IMO 11:47:15 But nothing beyond a solid wall elf/zot style. 11:47:17 dpeg: I know, it is a bit tricky to define... but I'd like to preserve the current behavior in all non-abusive situations 11:47:40 hm, we can explicitly distinguish between wall types 11:47:42 -!- st_ has joined ##crawl-dev 11:48:03 OOO 11:48:06 ### 11:48:07 ... 11:48:08 .. 11:48:13 .@. 11:48:13 But I phrased my version so restrictedly because I am afraid a monster on the other side of a wall will just stand there LRDed, not moving. 11:48:18 Ah, # is wall 11:48:38 dpeg: the point is that if they are really on the other side of a wall, they should only be damaged if the wall is destroyed 11:48:48 elliptic: yes, I like that about (c). 11:48:59 elliptic: Ah, that makes sense. 11:49:03 so if you target the center # in my picture, it shouldn't hurt the Os unless the wall is destroyed 11:49:07 It does mean that the middle ogre might be unharmed in RichardHawk's example, though. 11:49:11 no 11:49:19 true! 11:49:20 in his example they should all be hit regardless 11:49:24 direct path :) 11:50:33 I think the way to define this actually is that you pick the square adjacent to the LRD target which is in the direction of the player 11:51:08 and then you only affect squares that are locally connected to those after the explosion 11:51:21 elliptic: your proposal is conceptually really good because it treats a single rock wall as having large size (when it does not crumble, so does not damage monsters on the other side) or as thin (other case). 11:51:55 Hm, tasty "locally connected". Let's all hear the "semi-locally simply connected" now. 11:53:01 actually, replace "square adjacent to the LRD target which is in the direction of the player" with "square(s) adjacent to the LRD target that the player can see 11:53:04 " 11:53:33 elliptic: can you formalise it on the development plan? This should be a nerf to go into 0.10. Your version is so much better than mine. 11:53:45 sure 11:54:09 no link or nothing, just put it in a paragraph on its own 11:55:02 Foonesh (L7 SpEn) (D:6) 12:02:01 -!- ais523 has quit [Remote host closed the connection] 12:11:51 arrrrrr 12:12:14 elliptic: thank you! Btw, is it still possible to lose a turn by targeting a monster with LRD? 12:12:48 yeah... maybe even by targeting floor 12:12:49 if you target a monster, it does actually do some damage 12:12:57 yes (very little though) 12:12:58 just a really tiny amount 12:13:02 it's probably not worth keeping 12:13:20 except for skeletal things etc that actually shatter 12:13:43 I found this annoying in the tournament... wasn't aware it's a feature. So we best get rid of it? 12:14:22 MarvinPA: ah, sure. So targeting for LRD should do walls (near monsters) and brittle monsters. 12:14:42 there is some question about what it should do about invisible brittle monsters, I guess 12:15:05 ugh 12:15:08 z vs Z? 12:15:10 since it would be nice if it aborted if you targeted an empty square, but then you lose the option of exploding invisible brittle monsters... 12:15:45 haha 12:15:53 that'd just be petrified invisible things? 12:16:11 or skeletons that you zapped with a wand! 12:16:15 elliptic: you could do it with Z 12:16:29 I'd be more inclined to just insist that you have to be able to see the target 12:16:36 or change the rule, yes 12:16:38 it makes some sense, really 12:16:49 fine by me 12:17:09 Yesterday there was speculation about whether Shatter is similarly broken. But I cannot tell. Is it? 12:17:42 so yeah, LRD targeting should just abort without loss if you target an empty square or a non-brittle monster 12:17:55 well, shatter breaks walls rather reliably 12:18:06 so the fact that it does damage through walls isn't abusable for long 12:18:25 (I tried to use this to kill cerebov once... it didn't work, he was out and firestorming me almost immediately 12:18:30 elliptic: also not on Zot:5? There are those scummy types :) 12:18:50 it occurs to me that it might be good on zot:5, yeah :/ 12:18:58 But if you say that it's no big deal, also okay. 12:19:08 because you can dig around the permanent walls and such 12:19:18 you can't do that anymore 12:19:25 oh, that's good at least 12:19:47 shatter could just not go through permarock, if that's possible 12:19:47 so you'd have to actually be in the vault or near the entrance... and it is really really loud so stuff will come find you 12:20:15 Next one: is noise propagation broken in 0.9? 12:20:40 noise doesn't seem to travel through walls at all 12:20:43 this is sort of bad 12:20:59 messes up at least one vault (an ossuary, with flying skull sentries) 12:21:35 aside from that it is probably fine, I'd say 12:21:58 I don't understand the noise system at all but I'd say that noise should propagate through walls, just with greater loss 12:23:53 elliptic's proposal for LRD is basically what i suggested but better worded, btw 12:26:54 for shatter: one possible side effect of the status quo is that it might actually favor reducing power, because you'll still do appreciable damage through walls, but you'll blow up walls less 12:26:59 have not looked at the relevant formulas though 12:28:52 eronarn: well, chance of breaking a wall with shatter is proportional to power, which is proportional to damage dealt 12:29:31 so you might break half as many walls each cast, but you'll also do half as much damage 12:30:34 proportional? i'm not okay with that, where is the crazy formula :( 12:30:42 haha 12:30:43 (did anyone fix airstrike damage yet) 12:31:23 airstrike damage formula really isn't that bad... average damage is still approximately proportional to power :P 12:31:40 i think it would be good to refactor a lot of the spell power code to just take a probability function, but i also don't see that ever happening in crawl 12:31:58 if i were making a roguelike from scratch that's what i'd do though 12:44:51 Monsters no longer show weapon on attack (https://crawl.develz.org/mantis/view.php?id=4572) by minmay 12:47:20 -!- ais523 has joined ##crawl-dev 13:12:14 03elliptic * r04d47c8a1433 10/crawl-ref/source/ (itemprop.cc makeitem.cc): Tweak early weapon generation. 13:14:40 back 13:15:36 elliptic: noise not propagating through walls is a bug. No idea if Darshan had plans with it (I think he fixed noise propagation elsewhere). 13:22:05 oh i kind of liked that noise doesn't go through walls 13:23:04 well, hopefully there's some sort of middle ground between now and when making the tiniest noise in crypt made the entire level come for you all at once :P 13:25:33 That is the hope, yes. 13:30:29 tomb is a real problem after the branch noise changes, it's so reliant on stairdancing and you have *no choice* about doing it (unless you teleport and skip stuff) 13:30:35 keep the soundproof walls in sprint maybe :P 13:35:09 perhaps it should depend on the material of which the wall is made? 13:35:25 (but it should definately go around corners) 13:46:23 dpeg: i had some sif ideas; well, one specifically (but i made it into a general proposal) 14:06:41 mumra: I just was about to applaud :) 14:07:06 mumra: is this in any way related to what Wensley is doing? He is avoiding my questions for weeks now :) 14:07:11 hehe 14:07:23 i suggested the "silence" theme to him - he said he might try and work it in 14:07:48 and the collecting books thing was something i thought of but it turned out he was planning something similar 14:08:28 I like it a lot. Will change the appeal and audience a lot, of course. Ergo many "I will never play my casters with Sif again" comments. 14:08:31 but, i have no idea about the details of his proposal - since it hadn't appeared and i was thinking about it, i thought i'd just go ahead and write it ;) 14:08:40 many thanks for that! 14:09:31 i think sif needs to change audience in some way to get any real character 14:10:49 -!- ncampion has quit [Quit: Linkinus - http://linkinus.com] 14:16:56 -!- ncampion has joined ##crawl-dev 14:18:40 mumra: I agree. Just got the Sif page a little uplift, it was terribly dated in places, for example right at the start. 14:18:56 dpeg: btw, how did you get on with demigods? are you still having git troubles? 14:19:26 -!- ncampion has quit [Client Quit] 14:19:28 mumra: I could play them fine, but not very much. 14:19:43 You need feedback to go on, right? 14:20:44 -!- ncampion has joined ##crawl-dev 14:20:56 no, not really - still need to get a skeleton for all the minions and names in place 14:24:14 back 14:24:16 it's probably hard to gauge the difficulty level and item rewards until they're set up better 14:24:19 dpeg: still need help? 14:24:52 -!- ainsophyao has quit [Remote host closed the connection] 14:25:17 OT: Is the yes 14:25:19 oops 14:25:20 yes 14:25:30 what's up with #4572? Wasn't it always like that? 14:26:12 Don't know. But mentioning the weapon is a good idea. :) 14:28:38 -!- ussdefiant has quit [Ping timeout: 276 seconds] 14:29:54 dpeg: so what do you need help with? git? 14:31:04 yes 14:31:24 galehar: can you read messages in your irc client? 14:45:08 -!- dtsund has joined ##crawl-dev 14:53:31 !seen kilobyte 14:53:31 I last saw kilobyte at Mon Sep 19 13:47:08 2011 UTC (6h 6m 23s ago) saying even arguments other than %n can be an information leak on ##crawl-dev. 14:54:07 !tell kilobyte seems like #4572 is caused by 4f6f50a0. Was it intentional? 14:54:07 galehar: OK, I'll let kilobyte know. 14:56:35 We have to decide how to indicate (reaching) weapons. Listing all weapons in the monster list won't work. "o - 6 ogres (3 reaching)" ? 14:58:49 ah, about reaching weapons 14:58:53 dpeg: i think you can just say it if even one of them has reach 14:59:04 because that implies you will want to check 14:59:14 in tiles you click the weapons to (un)equip them 14:59:28 -!- syllogism has quit [] 14:59:29 but on reaching weapons you evoke them 14:59:34 petete: we're talking about monsters' weapons 14:59:38 its a bit inconsistent 15:00:17 petete: also inconsistent in console 15:00:40 in tiles, you cannot click a monster two squares away? 15:01:07 i mean clickin in the inventory 15:01:21 petete: well, mouse interface is a balance between convenient and consistent. Click triggers default action which can vary. Ctrl+click and shift+click are more consistent. The popup always tells you the actions (or should) 15:02:23 i think the popup does not know about the polearms change then 15:02:27 but you're right, nobody would click on a weapon to attack a monster. You click on the monster for that 15:03:02 galehar: yes. If you really want to close in to the monster, click on the empty square instead. 15:04:17 monster click is quite consistent now. It attacks it if you can (fire, reach or melee) else it moves towards it 15:04:44 ah, so it _was_ in 15:05:33 I'm not sure about reaching and the monster list. your example is not too bad. 15:05:38 Has anyone ever contemplated using the mouse wheel to scroll through ammunition and spells? 15:05:50 you just need to ^x to see the details 15:06:05 I'd like something better, eg. showing reaching monsters with some glyph, but it's not possible. 15:06:28 Not sure a branding is appropriate, ö would be much better for orc with reaching. 15:06:46 -!- bmh has quit [Ping timeout: 252 seconds] 15:06:53 I don't think the wheel does anything right now, but it should be quite easy to do. 15:07:25 dpeg: yeah, but what about a Troll with reaching? 15:07:36 In the same vein, would it make sense to have wand possession indicated in the UI somehow? 15:07:41 does unicode supports it? 15:08:31 galehar: yes, but console doesn't, if I understood kilobyte correctly 15:08:48 jle: I doubt so. What about armour and shield? Sometimes you have to eXamine! 15:09:58 galehar: Having an armour and a shield is not going to murder you as quickly on low levels as possessing a wand of fire/disint/cold. 15:11:37 you don't know about the wand until the monster has zapped it anyway. On the other hand, you can see weapons of monsters as soon as they come into view. 15:12:02 Yeah, that's true. I was just thinking it would be easier to spot in a pack. 15:12:44 Though you're right that in that case you generally take a good look at who's who. 15:14:04 We could have a general "dangerous" branding on monsters which they get if they carry unusual equipment. I am not sure it will work very well, though. 15:14:20 the problem with polearms, it's that the information is available if you look for it, but the game doesn't push it to you. So you have to manually check every monster if you want to have the information as soon as it is available. 15:15:02 another possibility is to tell about polearms when the monste comes into view, just like Ash's warning for branded weapons. 15:15:11 -!- Xiberia has quit [Quit: ChatZilla 0.9.87 [Firefox 6.0.2/20110902133214]] 15:16:02 galehar: that is good in any case 15:16:12 but seeing which orc on screen has a spear is still good 15:19:26 of course it is, but we still have no idea how. 15:20:43 Maybe give it a black ( on the health-icon? 15:22:37 it's only a bit better than dpeg's earlier proposal, since you still don't know which orc has the spear. 15:23:02 you can only tell that it's the unharmed one (as usually with those cowards with reaching weapons) 15:23:28 Maybe split it out into a separate line in the list? 15:23:38 ghallberg: I tried to avoid that 15:23:42 ok 15:23:45 sorry 15:23:50 no sweat 15:24:06 if we only use the monster list, then my proposal saves a line whereas your sometimes gives more information 15:24:15 galehar: there is branding, after all 15:24:24 wouldn't hurt to add a "reaching" brand 15:24:54 dpeg, galehar: I thought the plan for this was to list the wielded weapon on the monster list 15:24:57 which would be good anyway 15:26:03 -!- evilmike has joined ##crawl-dev 15:26:47 elliptic: my console only has five lines for monsters 15:26:58 elliptic: how? We merge monster groups only if they wield same weapon? 15:27:07 I think it'll be a step backwards for me 15:27:16 galehar: you suggested before merging monster groups only if necessary 15:27:31 elliptic: ah, that would be okay 15:27:36 yeah, I remember suggesting that :P 15:27:46 If there are five orcs, we can mention everyorc's weapon 15:27:55 I'm not sure how it would actually play but it sounds better to me than using branding 15:28:02 dpeg: or you can use a bigger terminal 15:28:07 if a gnoll enters, we collect orcs, keeping the reaching ones separate 15:28:13 galehar: not an option 15:28:19 dpeg: and if there are two orc warriors and five orcs, you stack the orcs and separate the warriors 15:28:23 I would die for my 80x24 console. 15:28:28 elliptic: yes 15:28:49 stuff like that... you'd have to make the system fairly smart, but it seems doable to me 15:28:53 elliptic: but is it more important to know the weapons of the two orc warlords or that one orc has reaching? :) 15:29:04 the orc warlords 15:29:08 I know! 15:29:16 Wasn't there a separate suggestion for that (5 orcs, 3 reaching)? 15:29:19 could compare by damage, and give additional weight for reaching 15:29:31 jle: yes, the first one on this topic today 15:29:39 Or is this when there are more than 5 groups of monsters on the screen? 15:29:43 dpeg: this is going too far. Weighting to choose how to group. 15:29:55 dpeg: just compare monsters by danger level, try to keep monsters of higher danger level ungrouped 15:29:56 branding! 15:30:01 don't special-case reaching please 15:30:14 the current grouping system is already decent. Can be improved some more of course. 15:30:17 that will just cause lots of trouble when other weapon types get special abilities and we want to know about them too 15:30:27 ok 15:30:42 I fully trust you guys, we should just have some announcement for 0.10 =] 15:31:16 after playing a few games with polearm reaching, I can say that it hasn't been bothering me even with the current interface as much as I expected 15:31:25 we have to support 80x24 but it doesn't have to drive every interface design either. 15:32:19 if you have many different monsters in sight, it's going to be more convenient with a bigger terminal. 15:32:46 * dpeg feels like an outdated device. 15:32:58 smaller termsizes are better since you can use larger fonts :P 15:33:22 how about both? Do you use a 14" screen? 15:33:52 I mean both bit termsize and big fonts 15:34:07 17'' screen and my terminal takes up about 2/3 of each dimension 15:34:42 ... 15:35:05 well, play in full screen then :P 15:35:23 nice thing about default termsizes is it makes ttyrecs more watchable... footv doesn't play nice with large terminals. More footv's problem than anything, but it doesn't look like it will be changed any time soon 15:35:31 imo a big thing about playing online is being able to watch replays of other games 15:36:16 galehar: well, full screen means I can't do stuff like do things in a terminal while keeping an eye on IRC... 15:36:34 elliptic: of course, I'm just joking 15:36:49 -!- ussdefiant has joined ##crawl-dev 15:36:58 evilmike: sure, but as you said it's more footv problem's than crawl. But it's important enough that we can't ignore it. 15:37:49 it's just that we're in 2011 and feeling constraint to 80x24 when designing interface feels weird. I'll go back coding on tiles... 15:38:24 -!- ais523 has quit [Remote host closed the connection] 15:40:21 galehar: I agree. Just provide a branding, console players can sort the rest by themselves. 15:41:30 -!- ussdefiant has quit [Read error: Connection reset by peer] 15:41:54 someone asked for a hostile summon branding on the tavern not long ago. I looked it up and it wasn't as easy as I thought to implement. Some curses flag were saturated already. 15:41:56 having the ability to configure branding based on the full descriptive string for the monsters (i.e. including weapon wielded) seems reasonable 15:42:03 -!- ussdefiant has joined ##crawl-dev 15:42:31 but please don't force new brandings on people :) 15:42:31 elliptic: that is a great idea 15:42:50 would also allow us to do the summons 15:43:03 elliptic: what is wrong with them? 15:43:34 more clutter on the screen unless you specifically want them 15:44:16 elliptic: which brands do you turn off? 15:44:42 none currently 15:45:08 but I think I'm near my limit ;) 15:45:20 sleeping and allies are crucial 15:45:29 COLFLAG in defines.h seems to be saturated 15:45:39 distracted and pile not as much, but I wouldn't want to play without them 15:46:07 I don't know how to add one static brand, so configurable based on the full descriptive string? 15:46:42 * galehar let the bearded old guys play with their tiny terminals 15:47:32 * dpeg thinks that 80x24 is the first and foremost rule of rogueliking. 15:47:49 As is having a beard. 15:48:32 do you also oppose to 256 colours terminals? 15:49:51 * bhaak hasn't shaved in days, but denies that is has to do with being a roguelike developer 15:49:57 Now, a full colour beard sounds like a good idea. 15:50:40 4096 colors should be enough for everybody. you can quote me on that 15:50:55 galehar: the lack of colours was very painful at times. However, I am convinced that a palette too large would have undesirable consequences. Less is more. Except for facial hair, of course. Down with the moustache! 15:53:04 Especially annoying is that your standard console will only accept 8 out of 16 colours for the background. You wonder what they've been doing back in the day. Probably cursing all the time while stroking their beards entangled manifoldly into punch cards. 15:53:32 I thought about stuff like using shades to show how likely a cell is going to be affected by a cloud spell, but then I wondered if someone would analyse the colour and reverse engineer to find out the exact spell power. 15:54:13 galehar: The answer is probably yes. 15:54:21 well you can find out the exact spell power from looking at your skills and int anyway :P 15:54:33 oh, less exactly with fuzzing i suppose 15:54:36 Folks who draw labyrinth maps on a sheet of paper would do anything. 15:55:00 Right, I don't have much time for stuff but I found a pet peeve I would like to change. 15:55:07 MarvinPA: not anymore with smooth skills (well, soon when I have removed skill progress from the menu) 15:55:08 Trog shouldn't gift ranged weapons/ammo. 15:55:13 It makes NO sense. 15:55:22 also what would that even be useful for? surely the only use for calculating exact power would be, say, to figure out how likely a cell is to be affected by a cloud spell 15:55:23 Opinions? 15:55:29 so if you're displaying that anyway... 15:55:31 ghallberg: because you can't use them while berserk? 15:55:58 elliptic: Pretty much, he's focused on aggresive fighting. 15:56:13 That's my understanding of it anyway. 15:56:14 dpeg: no one draws labyrinth maps 15:56:16 MarvinPA: yeah, knowing the exact spell power is useless. But some folks seem to value it high when I started talking about skill fuzzing. 15:56:25 evilmike: well, unfortunately I know better. 15:56:39 who does it, then? 15:56:45 but if it doesn't help you to calculate it, why does it matter if you can calculate it? 15:57:16 galehar: One thing I thought about concerning fuzzing. Will there still be "gained a skill level" messages? 15:57:25 MarvinPA: my conclusion was: let them do it. 256 colours can still be useful! 15:57:25 It was someone I value highly and it was under special circumstances (tournament, clan, old labyrinth, no food). Only that one time, of course. 15:57:34 ghallberg: yes 15:57:36 ah 15:57:41 galehar: ah okay, yeah :) 15:58:01 galehar: Ok, cool. Was worried that I would have to chek skill screen all the time to see when I wanted to turn somethign off. 15:58:27 dpeg: I changed maprot to make it less agressive. labs might be a bit easier. 15:58:32 So, does anyone agree about Trog, or am I alone? 15:58:42 galehar: I don't think they were ever hard :) 15:58:59 ghallberg: I agree but I not happy with all weapon and armour gifts in the current state. 15:59:02 *am 15:59:09 ghallberg: you certainly have a point, but I don't think it would be healthy for oka to be the only god giving ranged gifts 15:59:20 hm, should mimics spawn in slime? 15:59:23 elliptic: I would prefer oka only gifting armour. 15:59:35 There doesn't have to be ranged gifts. 15:59:37 it doesn't seem like their natural habitat as such 15:59:39 are you talking about mimics in slime 6...? 15:59:40 but what about the poor slingers? where will they get their ammo? 15:59:51 no, just stair mimics in earlier slime 16:00:07 elliptic: Increase ammo drop rate, differentiate Oka and Trog more. 16:00:28 Maybe it's not feasible and mostly a product of my annoyance... 16:00:50 MarvinPA: No! 16:01:01 I don't really care about Tor gifting me bad weapons, he's not supposed to be smart, but ranged is just out of character and pretty bad.. 16:01:01 ghallberg: it is feasible probably, but would require a fair amount of rebalancing I think 16:01:09 MarvinPA: oh, Slime:6? Sure! 16:01:12 the ranged weapons are useful with trog though 16:01:17 maybe it is trog's flavour that should change 16:01:22 it was a staircase mimic on slime:2 16:01:26 if that's the issue here :P 16:01:40 elliptic: Or let it be harder, Trog isn't exactly too hard as it is. 16:01:50 marvinpa: what's wrong with a staircase mimic in slime? 16:01:57 MarvinPA: yes, sounds okay 16:01:58 well, i dunno 16:02:15 i suppose i just think of mimics as a different type of trap, and all the traps in slime are eaten away 16:02:32 mechanical traps at least 16:02:34 MarvinPA: might feel odd to have something masquerade right there, but should be fun for the player to discover 16:02:58 my policy when implementing mimics was: add them everywhere, and remove them later from places they are really not welcome 16:03:06 yeah, sounds fair enough 16:03:17 so excluded places are only temple and vestibule 16:03:37 I added no staircase mimics in tombs because it messes the layout 16:04:05 I'll think this god stuff through a bit and come back :) 16:04:19 Could we have an orcish idol mimic? I want to be chased by Beogh. 16:04:31 not sure about slime. It certainly feels like a special place. 16:04:44 dpeg: I think you can 16:04:52 -!- monqy has joined ##crawl-dev 16:04:54 can mimics be corroded by acid walls? 16:05:05 if so, that's probably reason enough to exclude them from slime 16:05:07 i think they're acidproof? 16:05:08 evilmike: I think so 16:05:10 @??feature mimic 16:05:10 feature mimic (09X) | Speed: 10 | HD: 8 | Health: 29-59 | AC/EV: 5/1 | Damage: 1208(poison), 12, 12 | Flags: 11non-living | Res: 06magic(32), 05fire, 02cold, 10elec, 03poison, 12drown, 04rot+++, 13neg+++, 13torm | XP: 465. 16:05:14 oh okay, maybe not 16:05:17 it just wasn't by a wall 16:05:28 oh, if they aren't acidproof they shouldn't be in slime 16:05:42 yes 16:05:47 fine, I'll remove them from there 16:05:53 only feature mimics though 16:05:54 but the @?? line looks as if they want to be acidproof 16:06:08 you'll still have item mimics in the vault loot :) 16:06:46 i've only run into door and stair mimics this game i think, i'll have to hope it stays that way :P 16:07:00 dpeg: I confirm. orcish idols can be mimiced. 16:07:20 item mimics are too rare? 16:07:24 tell me! :P 16:08:28 possibly, apparently i've killed 9 features and no items so far 16:08:40 03dpeg * r0d2606434749 10/crawl-ref/source/dat/des/branches/slime.des: Some slime entries with slimy walls. 16:09:48 any other feedback on item mimic frequency? 16:10:01 I got a few item mimics this game 16:11:02 The frequency is 1 in 500 item. But there's so many items, I can't even tell if this is rare or not. 16:11:42 seems okay to me 16:12:30 MarvinPA: you're not greedy enough. Go get the loot! :P 16:22:19 -!- Fail_Bot has joined ##crawl-dev 16:25:16 mumra: don't take the comments personal. It is often like this. You propose something and then your proposal gets all kinds of negative reactions. The only thing worse is o reactions. 16:25:21 *no 16:26:00 dpeg: i knew not everyone would like it ;) 16:32:00 if I said anything personal there I apologize, I didn't mean to do anything but state my own opinion, which probably isn't worth much 16:33:09 I think the piety hits of the proposal are just way too high. 16:33:15 yeah 16:33:26 what proposal is this? 16:34:02 they were just the first numbers i came up with; i certainly didn't think them through enough 16:34:02 Got a couple one-line high-score fixes/tweaks for you: https://gitorious.org/crawl/crawl/merge_requests/5 16:34:06 What is piety actually used for later on? Just to lend new spells, right? So once you're happy with your spell selection, you can turn piety into loud spells. Just have to be a little more conservative than your hellfire slinging Vehumite who'd use Storms on rat packs. 16:34:49 elliptic or galehar: could you check SamB's patch? He keeps committing and not getting reactions :O 16:34:59 I think you really need to spam those spells to make the most effective use of them, especially now that they are more resistable 16:35:04 https://crawl.develz.org/wiki/doku.php?id=dcss:brainstorm:god:sif 16:35:10 dpeg: oh, I usually get stuff 16:35:14 dpeg: that's the idea; the piety hit could be toned down (or, the piety cost for silent casting could be very low, and its duration increased etc. to compensate) 16:35:25 silent casting is an interesting perk... but it doesn't really make up for the loss of being able to cast them whenever you want, unless the piety hit were fairly low 16:35:26 that is, most of my stuff has been accepted 16:35:29 evilmike: spamming is okay in a short burst, like a battle 16:35:29 evilmike: you can still spam them several times while wordless casting is enabled 16:35:37 sorry, need to finish some skill tweaking then sleep 16:35:43 except for the luadoc stuff and one small change I've since dropped 16:35:44 galehar: okay 16:35:51 poke me again tomorrow if nobody else has looked at it 16:35:57 SamB: ^ 16:36:00 mumra: ah, I see. I was thinking of wordless casting as being on a per-spell basis, rather than a duration based thing 16:36:09 SamB: No-one's really familiar with doxygen apart from due, I am afraid. 16:36:23 evilmike: no, per spell would also be way too spammy. 16:36:37 well, it would have the advantage of letting it be a passive ability 16:36:45 evilmike: yeah, i didn't really specify; i'd only just thought of it so i hadn't exactly decided; but a status effect would be the best thing 16:36:56 yeah, or a passive ability 16:37:04 but that could be hugely overpowered 16:37:06 mumra: I have some more questions, perhaps better asked here: The god would use Invocations? Does lending go down upon getting a book? 16:37:13 oh yeah something i just noticed... your sif proposal seems to lack channeling? 16:37:16 you could have a silent spectre following you around and be able to cast any spell constantly 16:37:25 evilmike: on purpose, I think :) 16:37:55 * dpeg believes that channeling is a huge problem, right there with cTele. 16:37:56 evilmike: i start the proposal saying "this is a very different sif, i'm dropping all the existing abilities" ... so yeah, i was expected reactions :P 16:37:57 anyway, these are to (a) fix high-score handling when restart_after_game=true and (b) allow printing high-scores from abnormal gametypes on the command line 16:38:13 SamB: thank you 16:38:14 what about channeling that acts kind of like the regeneration spell, but for mp? 16:38:22 one line each 16:38:41 dpeg: Invocations could determine a) strength and duration of silence effects b) # of spells in the books sif lends you (as examples) 16:38:42 dpeg: channeling is mainly just something to make the spellcasting interface less annoying 16:38:50 by which I mean not spending 4180758109 turns resting for MP 16:38:53 the tricky part was figuring out where to put that line ;-) 16:39:13 elliptic: I am always afraid it turns casters into machines who never run out of gas. This seems so wrong to me. 16:39:23 it doesn't really work like that 16:39:30 elliptic: a god power designed to eliminate an interface problem doesn't sound that interesting; besides, there are other more interesting methods of channeling 16:39:32 elliptic: promised? :) 16:39:58 mumra: sure, it isn't the most interesting ability... I just don't think it will bring about the death of crawl like dpeg does ;) 16:40:26 mumra: can lending in a school ever go down? 16:40:28 oh by the way, one of mumra's comments refers to makhleb as a caster god, which is just wrong 16:40:31 meant another god, maybe? 16:41:02 Makhleb is very versatile, but the powers befit only special hybridising casters, I guess. 16:41:09 ok, i just don't play makhleb (i assumed MP-for-kills meant some sort of caster god) 16:41:16 dpeg: even with mummies with free channeling, I hardly use channeling in combat much at all... mainly good for recovering more quickly in between fights 16:41:24 makleb is good for all character builds, but doesn't favour casters in particular 16:41:58 i still want to get rid of makh's mp-for-kills :/ 16:42:14 oh is that what that thing meant 16:42:17 dpeg: I recently played a chei caster, and needing to rest for MP all the time was really annoying compared with sif's channeling or veh's MP-on-kills 16:42:32 that's what I use channeling for too... I do use it sometimes in combat with low mp, but in those cases it winds up putting like a 2-3 turn delay between casting each spell, which is usually pretty bad 16:42:48 and make his invocations cost hp since that would also fit with his mortification of flesh thing 16:42:56 ouch 16:42:57 dpeg: the other thing i mentioned invocations for in the proposal, is how long you get to keep a book for (which could be relevant if you borrowed a book which turned out to contain another spell you were after, but you wanted or had to wait a short while to get extra spell slots) 16:43:04 that would not work with mirror damage... 16:43:29 hm? 16:43:34 not early, anyway 16:43:36 MarvinPA: I like the latter but I feel the former reduces the audience too much. What if powers would cost both HP and MP? 16:43:40 elliptic: but what about a staff of channeling? or a crystal ball? both of which have the benefit of using evocations skill, which has other uses as well... 16:44:27 well, i sort of feel like invocations from hp would encourage casters who take him to use the invocations more 16:44:29 staff of channeling is a lot slower than sif muna channeling... it's probably the worst method of mp recovery 16:44:29 mumra: right, that chei char raised evocations really high to use the crystal ball late game, and things became less annoying... but the evoc options are not really reasonable earlier on 16:44:41 MarvinPA: Ouch, it's a cool idea but DDs will hate you :) 16:44:46 since those would come from hp (which makh restores on kills) and they could also still cast from mp 16:44:56 MarvinPA: good point 16:44:57 and yeah it has the excellent side-effect of nerfing DD of mahkleb a bit 16:45:06 Yes, it's pretty neat :) 16:45:11 and dpeg should really hate crystal ball a lot more than sif channeling, since it acts a lot more like never running out of MP (aside from confusing you and draining your MP to zero occasionally) 16:45:17 He does. 16:45:35 Hmm, there should be a god that can petrify monsters and turn them into ammo :) 16:45:52 ellpitic: i was using a staff on my 15-runer; i was carrying another weapon and i kept switching from one to the other to recharge; i really didn't find a problem with it 16:46:09 elliptic: of course, i was with vehumet, so a lot of the time mp was manageable 16:46:23 What if mumra-Sif would passively increase MP regeneration depending how quiet it is. Quick recovery after battle, no effect during battle. 16:46:31 dpeg: i like that a lot 16:46:39 dpeg: that could be cool 16:46:54 that does sound interesting 16:47:02 Would fit the theme and somehow fit what you guys tell me. 16:47:07 how would it react with ambient noise? 16:47:26 Real men play Kikubaaqudgha exclusively. And maybe Yredelemnul and Lugonu in weak moments. 16:47:28 eg crypt is quieter than other branches 16:47:30 heh, quicker mp regen in tomb, slower in hive? could probably not worry about background noise :P 16:47:38 yes 16:47:46 yeah, make it relative to ambient noise levels 16:47:59 The follower does not affect background noise, after all. 16:48:02 on the other hand, i was going to add a note about portal timers - sif likes it when you stop them making that racket :) 16:48:08 haha 16:48:22 hmm, what *do* followers do to noise levels? 16:48:47 not cast, shout, fight (stabbing would be okay)? 16:48:54 not cast loud spells, sorry 16:49:33 * SamB was wondering if their shuffling about would make noise 16:50:54 I like the flavour of a "quiet" god, however I'm not sure crawl's noise system is really ready to handle it 16:51:26 basically, i was thinking "librarian" and the first thing that came to me was "ssh, be quiet!" ;) 16:52:08 03galehar * recaf6b2bf323 10/crawl-ref/ (5 files in 2 dirs): New option: skill_focus. 16:52:09 03galehar * rf6afd2c6903d 10/crawl-ref/source/dungeon.cc: No feature mimics in slime. 16:52:09 03galehar * r130b8f916c3b 10/crawl-ref/source/ (skill_menu.cc skills2.cc): Only show pogress to next skill level in wizmode. 16:52:10 but yeah when i started having a look thru the spell noise code, there is very little to differentiate what is and isn't noisy 16:52:10 03galehar * r9e8371aef827 10/crawl-ref/source/ (skill_menu.cc skills.cc skills.h): Hide training percentage for unknown skills. 16:52:11 03galehar * rffdae314ec41 10/crawl-ref/source/dat/database/help.txt: Update the skill screen help. 16:52:29 it would be best to limit the noise-type conduct to just a few obvious things rather than trying to think about how noisy every single part of the game is, I think 16:52:35 mumra: they hate it even more if they lose books, in my experience. Not just theft, misplacing is almost as bad. 16:52:43 galehar: awe-some! 16:53:10 dpeg: well of course; that's why there's a huge piety hit if you fail to return a book you've borrowed (even more than just destroying a book you've found) 16:53:11 like, explosions are bad, shouting is bad... 16:53:21 elliptic: I agree. Piety hits for shouts, spells (depending on volume) and battle (this does cause noise) seems good enough. 16:53:28 no need to worry about the base spell noise for higher-level spells 16:53:50 only the really really noisy ones like shatter and such should matter 16:53:59 elliptic: but wouldn't it be interesting if you bias your spell list by volume? 16:54:05 ah, fair enough 16:54:07 There are only about 6 or 7 spells that have that noise level 16:54:17 elliptic: actually, shatter is just as noisy as any level 9 spell 16:54:26 uh, no 16:54:26 the level 9 spells, lightning bolt, conjure ball lightning (oddly enough) are the only ones I can think of 16:54:28 If the list is short enough, those could be on the Sif anti-list 16:54:42 Why is CBL so loud? 16:54:42 shatter is far noisier than summon dragon, say 16:54:52 indeed 16:54:56 elliptic: ok, change the above to level 9 *conjurations* 16:54:58 oh yeah I forgot summon dragon was level 9 :P 16:54:58 dpeg: spell noise levels are really arbitrary at the moment 16:54:59 galehar: now that skill focus is actually useful, we may use the three-way setting normally 16:54:59 kilobyte: You have 1 message. Use !messages to read it. 16:55:20 isn't ball lightning usually earily quiet? 16:55:23 kilobyte: I agree but I think there were some objections. 16:55:24 elliptic: all conjurations are the same noise level, which is twice as loud as summons 16:55:24 would reduce the amount of switches needed 16:55:29 kilobyte: uh, why is skill focuse useful? 16:55:34 And would free a line on the m screen, I think. 16:55:35 kilobyte: I just commited that 16:55:38 mumra: uh, lots of stuff is special-cased 16:55:40 \o/ 16:55:42 elliptic: *unless* noise_mod is non-zero (but very few spells have that, and it's usually a -ve amount) 16:55:49 the new skill_focus options default to true 16:55:50 you are just looking at the base noise 16:55:56 elliptic: ah well; i couldn't see that from from spl-data.h 16:55:58 galehar: it is useful since today morning :) 16:56:03 haha 16:56:35 why is skill focus useful now? I don't understand 16:56:40 galehar: with smooth skills, you no longer have to turn off/on skills the moment one reaches the next level 16:57:03 elliptic: you can set it and not look at the skills screen for several levels 16:57:09 kilobyte: you mean elliptic right? 16:57:47 elliptic: Is the point that you're still better off manually training skills one by one? 16:58:27 dpeg: well, I don't foresee my habit of training skills one at a time changing 16:59:16 the main reason why I ever had two skills on at once was to avoid accidentally going too much over a skill breakpoint... now that skills are smooth, I won't be worrying about that 16:59:37 galehar: oh, I see 17:00:00 galehar: I understood you as if you just implemented skill focus 17:00:17 elliptic: If you want to level two skills but one mroe than the other, (for example) 17:00:25 I think that since progress to the next skill level isn't shown, it's probably better to train them one at a time (or at least turn them off when they reach the next level). Makes it easier to remember "where" the skill is 17:00:44 galehar: found my first item mimic on zot:4 :P 17:00:59 evilmike: It doesn't matter since you will benefit from training inbetween "levels" aswell. 17:01:22 evilmike: I've heard that before, and I don't see the point of knowing "where your skill is" 17:01:26 -!- Pacra has joined ##crawl-dev 17:01:29 I know. But it's nice to at least have a precise idea of where they are, in between levels 17:01:47 galehar: Should just remove the levels... 17:01:56 Or add more of them. 17:02:11 just put that back in as decimal points ;-) 17:02:37 It would make sense to make skills out of 100 or 200 or something, but this seems like one of those "won't do" type things 17:03:01 (er, I mean, digits after the decimal) 17:03:17 I think the main argument for giving people more exact information about where their skills are is that it helps them figure out how much xp they are getting for things 17:03:28 This is only psychological, it has absolutely no gameplay meaning. 17:03:29 for example, if I drink a potion of xp 17:03:33 except the skill scale isn't linear. Levels get more expensive. 17:03:51 A scale of 27 is arbitrary, but not so bad actually. 10 would be too low, I guess and 100 is too fine. 17:03:57 sure, but now I might drink a potion of xp and not be able to tell whether it did anything to my skills 17:04:01 dpeg: you get XP % to next level on the main screen now 17:04:03 and report it as a bug 17:04:06 the psychological factor is huge. I honestly think crawl is more fun because of the xp percentage on the main GUI 17:04:18 I like that, too 17:04:24 elliptic: I agree, yeah 17:04:43 galehar: but that is a bit older, isn't it? I like that. Piety stars and xp until next XL were often checked by players via %. This is not necessary anymore. 17:04:50 putting in tenths seems reasonable to me, say 17:04:56 solution: hide skill level numbers full stop. replace with meaningless messages like "you have gained some foo skill", "you have gained lots of foo skill" etc. at completely random intervals 17:05:24 I think I'd start taking out-of-game notes in that case 17:05:26 as I said earlier, I am not opposed to testing things with a decimal. 17:05:28 * SamB hopes that mumra is attempting reductio ad absurdium 17:05:31 XL is discrete, skill levels are not anymore. 17:05:51 galehar: I like the decimal, or multiplying by 10 or so. 17:05:54 Ultimately, I think it will be useless, but at least while we're testing it might be useful 17:05:55 ugh 17:05:58 Except multiplication breaks 27... 17:06:20 dpeg: this is reminding me that I still want to change the XL display to "XL: 15.35" rather than "XL: 15 Next: 35%" 17:06:26 I also would like the decimal 17:06:35 elliptic: hmm, but that's *not* smooth 17:06:53 These are the same people who answer maths/physics exercises with "the height of the tree is 15.273 metres" :P 17:06:55 sure, but I think it would look nicer :P 17:07:10 dpeg: hey, I'd be happy enough with just one digit after the decimal 17:07:29 also, recall that I suck at crawl, so I'm often rather low in skill level 17:07:44 elliptic: I wouldn't mind. 17:08:05 ok, I'm adding a decimal to skill level 17:08:38 -!- Textmode has quit [Ping timeout: 258 seconds] 17:08:38 * dpeg opens another pack of valium. 17:08:46 elliptic: I like 15.35 for XL 17:09:06 galehar: maybe I'll add it as an option 17:09:13 I'm not sure I'd actually mind using decimals for the XL myself, but I think it gives the wrong idea 17:09:46 -!- Pacra has quit [Read error: Connection reset by peer] 17:10:08 SamB: good point. The partial level for XL does nothing, but it does for skill now. So the fact that skill level has a decimal but xl isn't has a meaning. 17:10:17 galehar: oh, by the way, another reason to keep a little precision in the skill display is so that people are less confused by crosstraining/antitraining 17:10:33 I like decimals for skills but ot XL since skills are continous but levels are still discrete. 17:10:48 ghallberg: exactly 17:10:56 I was slow, but yeah :) 17:11:37 * SamB is still waiting on his high-score one-liners from https://gitorious.org/crawl/crawl/merge_requests/5 17:12:06 -!- Wensley has joined ##crawl-dev 17:12:31 elliptic: you're XP proposal is still good as an option default off. 17:12:48 Shun the option! 17:13:10 Wensley: hi! Is your Sif by any chance similar to https://crawl.develz.org/wiki/doku.php?id=dcss:brainstorm:god:sif#librarian_proposal by mumra? 17:13:13 yeah, I would leave the current display as the default of course 17:13:47 We have too many options. If we think something is better than status quo, just throw out the old version unless we really know why not to do so. 17:14:04 * dpeg suspects that coders are paid by how many options they've added. 17:14:08 * SamB wonders if kilobyte would like to merge it, since he evidently likes merge proposals so much 17:14:26 kilobyte seems busy tonight. 17:14:43 dpeg: I think it's just that they're scared that some users won't be able to handle not keeping the old way as a choice 17:14:50 dpeg: I've just said why not good as default. But it's still nice. Especially if we can make good use of the freed space. 17:15:12 don't we have some space constraints? ;) 17:15:13 galehar: if we make good use of the space, everyone should have his share :) 17:16:08 SamB: yes, this is why we have so insanely many options. Understandably, coders want to code and not argue with angry users who want their silly old state back, being the conservative sobs they are. We need the New Man. 17:16:25 sometimes it's not so silly 17:16:33 regarding options, is there a reason why pickup_mode = multi isn't default? 17:16:42 the default pickup mode is so... clunky in comparison 17:16:53 good question 17:16:58 why is there even a choice? 17:17:05 I suppose the answer must be "because its always been that way" 17:17:12 Yes, not all options are bad. I didn't say that. But many of Crawl's options could be trashed with just a little effort. The options guide is threating to eat the unsuspecting reader whole. 17:17:27 the options guide is... somewhat daunting, yeah 17:17:35 dpeg: I wholly support trashing the options file 17:17:37 dpeg: Good idea for a revision then... 17:17:37 dpeg: I think display options aren't so bad, really... it is the options for interacting with crawl that make things complicated 17:17:50 MarvinPA: you should have seen the original version. I think I got into the devteam solely by cleaning it up. 17:18:32 SamB: damn busy, lemme take a quick glance, but I'm not promising anything 17:18:40 elliptic: yes, display is usually a good place for options. More annoying are new interface features which clearly improve matters but that insuranced with an option. 17:18:47 I think the whole option code could use some refactoring and some code generated from text file describing options could help. 17:19:31 Haran promised an in-game options editor. ("Even Nethack has one.") I bet he still has a slightly bad conscience about that one. :) 17:20:01 that sounds hard 17:20:15 how do you avoid making it eat wanted comments? 17:21:40 It surely can be done, but I don't think it is worth the effort. 17:21:52 I think it is 17:22:00 Some options also depend on playing style (of your current character) those should be available in game 17:22:17 like pickup options -- hey, there is even an implementable for this one :) 17:23:25 SamB: why exactly does the first commit include a tarball with a saves/ dir? 17:23:33 kilobyte: what? 17:23:38 that was an accident 17:23:38 well, that's the point. If we refactor and have part of the code generated from a structured self documented text file it will be much easier to maintain and edit options in-game. 17:23:53 sorry! 17:25:22 galehar: I am not sure I understand. The descriptions should be in the same file as the options? You can do that, but then the options file becomes insanely big. 17:27:08 it's like a csv file, from it you generate options.h, options-guide.txt, some parts of initfile.cc 17:27:49 SamB: thanks for the fixes! 17:28:02 you're quite welcome 17:28:11 03naesten * rc49ef905431d 10/crawl-ref/source/hiscores.cc: Fix odd high-score behaviour with restart_after_game=true set. 17:28:19 you get rid of init.txt, defaults are set in the source file and documented in options_guide.txt, and you set options in .crawlrc or something 17:28:22 03naesten * rb6f769276c60 10/crawl-ref/source/main.cc: Allow listing highscores for non-default game types from command line. 17:29:09 Hmm, I don't think getting rid of init.txt is a good idea 17:29:10 galehar: it might be better to have such a file provided, just with everything commented out 17:29:37 galehar: it's options actually set there what's a problem 17:29:48 is that even a real problem? 17:29:56 kilobyte: you could have a command line to generate a commented out init.txt sure 17:30:40 SamB: for example lua includes. Or force_mores. Or runrest_stops. You won't get improvements unless you merge them manually. 17:30:53 we have to keep compatibility with the current system for stuff like autopickup_exceptions.txt 17:30:57 Well, yes 17:31:07 which is why user init files shouldn't be copies of init.txt 17:31:09 but a good part of it would benefit from being generated 17:31:16 galehar: I'd more all the defaults internally 17:31:23 instead, they should contain "include = init.txt" 17:31:42 most of them are options that just happen to be implemented in lua not C++ 17:32:04 I supported sourcing out options, like with the autoexceptions. Is that bad now? I figured it'd make everything easier to understand and more readable. 17:32:43 kilobyte: I think defaults should be set in the csv, you generate the reset_options method out of it. 17:33:03 anyway, the way options work now, some of them need to be overriden *before* the default and some *after* 17:34:26 so I'm not sure how you can make the init-in-C++ thing work with that... 17:35:01 SamB: so all need to be changed to be overridden after 17:35:18 that's going to be complicated for some of the set-more-than-once type options 17:35:34 All the defaults are in initfile.cc but I think I miss something. 17:35:48 SamB: it doesn't have to be done in C++ or generated from some csv or otherwise. There might be a default init file, hidden somewhere and included silently. 17:36:07 SamB: just as long as everything can be overridden 17:36:28 hmm, I just don't see the need for the "silently" part ;-) 17:36:37 whether the defaults are done in C++, lua or a .txt file doesn't matter, the user doesn't see a difference anyway 17:37:11 SamB: there is -- without it, people will think they can change that include 17:37:30 you mean, alter the file? 17:37:34 yeah 17:37:44 ignoring the nasty warning? 17:37:50 defaults should be immutable but overrideable 17:38:14 yeah, warnings are not good enough, no one reads them :( 17:38:23 anyway... 17:38:43 how would one tell crawl *not* to use such-and-such lua script? 17:39:57 make it pointless to disable it? 17:40:02 I guess 17:40:38 like, autofight is harmless if you don't have it bound to a key 17:40:55 so there is no gain in not including that lua file 17:41:04 and if that isn't working, I suppose the user could always just rebind (unbind?) the defined names... 17:42:44 Anyway, I definitely support the idea of a freshly-minted .crawlrc being nearly empty 17:43:12 oh, I've an idea 17:44:04 embed the defaults (in .crawlrc format) in the binary, but allow the user to request them to be printed 17:50:21 printed in germany, and mailed to their residence, at their expense 17:51:31 not that kind of printed ;-P 17:55:28 "I'd like my options file printed on deckle edged paper, please." 18:06:08 03kilobyte * r3976fbe97f27 10/crawl-ref/source/fight.cc: Fix monster weapons not showing on hits not from afar. 18:06:34 ?/G and ?/B improvements (https://crawl.develz.org/mantis/view.php?id=4573) by dpeg 18:09:18 for the LRD discussion: should Shatter obey the same? The fix there is trivial. 18:10:04 no need for those complex "recursive processing of shock wave", it's a matter of distance_iterator + exists_ray() 18:10:04 kilobyte: "the same" meaning stuff in the same connected component as you (locally inside the range of the spell) after destroying walls? 18:10:45 elliptic: do you mean it should somehow bend around corners? 18:11:28 perhaps 18:11:28 isn't shatter an earthquake... 18:11:34 I'm not sure how shatter is supposed to work 18:11:38 LRD was clearer 18:12:00 shatter is not an explosion centered at you; as st_ says, it is more of an earthquake centered at you 18:13:43 hmm, if it's an earthquake, why do walls matter at all? 18:13:53 SamB: right, currently they don't 18:14:10 maybe it needs a new name/description 18:15:23 This spell causes a burst of concussive force around the caster, which will 18:15:24 damage most creatures, although those composed of stone, metal or crystal, or 18:15:24 otherwise brittle, will particularly suffer. The magic has been known to 18:15:24 adversely affect walls. 18:15:42 I don't really know what "burst of concussive force means", but maybe it isn't an earthquake 18:16:08 hmm, it makes me think of cuncussion missles... 18:16:14 From the spell description alone I'd assume that the damage lessens with distance. 18:16:30 true that 18:17:00 though that'd happen with earthquakes 18:17:56 * jle feels a bit guilty for concussing half the orb guardians in the chamber while experimenting to see whether he could break through with shatter. 18:18:05 anyway, what's not exactly clear is what walls do to *it* 18:19:03 -!- galehar has quit [Ping timeout: 260 seconds] 18:19:50 elliptic: it's not really an earthquake 18:21:25 -!- Guest94446 has quit [Read error: Connection reset by peer] 18:21:35 it's more like a very, very powerful burst of low frequency noise 18:22:12 hmm, are dart traps supposed to affect your minions in zotdef? 18:22:15 one could certainly justify making shatter only hit monsters/walls you could see... but that would be a huge nerf to the spell and make it less cool 18:22:57 yeah, that'd go through solid rock alright 18:23:45 it might be abusable as is, though... mainly with permanent rock or thick layers of metal, since everything else disappears quickly 18:24:51 -!- Twilight has joined ##crawl-dev 18:25:08 elliptic: i say make it hit things it can reach, after blowing up walls, regardless of whether you can see them 18:25:22 -!- Twilight is now known as Guest28783 18:26:23 where by reach you mean something like what I said earlier about being in the same connected component as you (locally in the area of effect of the spell)? 18:26:34 elliptic: make it not able to pass through anything it can't destroy 18:27:07 elliptic: different than the LRD rule that was being talked about, if that's what you mean 18:27:22 eronarn: what do you mean by reach, then? 18:29:12 elliptic: in my mind, if you had: @...#o, but no other #s, the o could be subject to shatter even if the wall weren't destroyed 18:29:36 whereas LRD cast on the # would not hit the o in that case 18:29:40 yes it would 18:29:43 with my proposed rule 18:30:24 then i guess i just misunderstood your proposal, i don't think lrd should be hitting through walls at all unless it actually blows them up 18:32:57 essentially the spell reaches around the wall that is being targeted and deconstructs in all directions 18:33:35 eronarn: the reason why I don't like your rule is the following: 18:33:40 ....#o 18:33:43 @..... 18:33:44 -!- Torokasi has joined ##crawl-dev 18:34:15 compare this situation... you still can't "see" the side of the wall next to the orc, but now you can see the orc... would you say that LRD should be able to hit the orc here? 18:34:53 wouldn't the new retreating behavior make it much harder to lrd things on the other side of walls? 18:35:04 Wensley: we don't really want to rely on that 18:35:14 elliptic: with your proposal it would, as it should be. 18:35:17 it seems pretty natural to me that you should be able to hit the orc here with LRD, and also that these two situations should be treated the same 18:35:23 What should happen is that LRD creates noise on both sides of the wall. 18:35:24 how would you even know they're there and in position, aside from scrying? 18:35:35 Wensley: antenna 18:35:50 dpeg: yeah, just explaining the motivation for my proposal to eronarn 18:35:54 alright 18:36:13 elliptic: I believe your concept could work just as well with Shatter, only that the local area would be much larger 18:36:23 Which is something you probably already said. 18:36:44 elliptic: i think in that scenario you should be able to hit the orc, but not something one tile north of the orc 18:37:10 dpeg: it gets complicated though because you might only be able to reach some walls because you destroyed others... so you sort of have to take things in stages probably 18:37:45 which I think kilobyte hinted at and seemed to be indicating that he thought it was simpler, but I'm not sure exactly what his proposal was unless he wanted to only hit things you can see 18:37:49 The LRD is a much more serious concern, as far as I understand. Why not simply implement elliptic's cool proposal and see how it works out? 18:39:11 eronarn: but you can see one tile north of the orc too... I think this will be pretty confusing 18:39:36 hmm, can you? 18:40:14 yeah, pillars aren't very good at blocking sight 18:40:36 i thought that the permissiveness was such that for any given @/# placement at least one #-adjacent would be blocked from LOS 18:40:44 but if that's wrong then yeah it won't work in the case of pillars 18:40:49 let me check 18:41:32 yeah, in this situation that pillar doesn't block sight of anything, in fact 18:42:23 huh, that strikes me as bad :P 18:42:50 getting reasonable LoS on a square grid is pretty hard 18:43:00 at least crawl's is symmetric :P 18:43:08 Crawl's LOS is very permissive. Not sure whether that's good or bad but it does have a number of agreeable properties. 18:43:18 (for a long time angband had asymmetric line-of-fire... incredibly abusable) 18:43:26 Full symmetry, for example, but there was more. Haran and rob knew a lot about it. 18:45:07 you mean... anyone would even consider a non-symmetric LOS in a roguelike? 18:45:27 maybe only accidentally... 18:45:48 elliptic: tome still has asymmetric LOS 18:45:55 that's an error so glaring it's quite unbelievable someone would do that on purpose 18:45:56 kilobyte: only if the monsters had the advantage, perhaps 18:46:02 it also has enemies that can oneshot you if you don't prepare 18:46:09 ~good times~ 18:46:21 beh, tome is still angband 18:46:23 kilobyte: it was pretty bad, yes 18:46:37 kilobyte: no way, it has unlocks now. totally not angband 18:47:05 I'm rather surprised tome still has this old angband flaw... didn't tome get rewritten from scratch or something? 18:47:40 @ .### 18:47:46 oops 18:47:51 @.### 18:47:54 ##.## 18:47:56 ###.# 18:47:59 ####D 18:48:19 (you can fire stuff at the dragon without fear of retaliation as it approaches) 18:48:34 this is how the asymmetry worked in case anyone is curious 18:50:03 Is it confirmed that the current TOME by Darkgod still suffers from this? 18:52:28 yep 18:52:49 they *just* rewrote the code, in this past release, to make FOV and LOS identical 18:52:56 but it's still not symmetric 18:53:25 Eronarn: but they are adding species and areas by the dozen... I wonder about their priorities. 18:53:59 dpeg: hey, we still have roundlos. glass houses, stones, etc. 18:54:12 (why don't we have a 'glass house' entry vault with a big pile of stones?) 18:54:18 If they rewrote it, 's probably why the flaw is still there 18:54:23 roundlos vs squarelos is a design choice 18:54:32 asymmetric LOS cannot be anything but a bug 18:54:34 Can't easily merge code from vanilla Angband 18:54:44 Where it's been fixed, I think 18:54:48 dtsund: sure 18:55:18 Eronarn: no "glass wall" support, I think 18:55:29 translucent walls work fine 18:55:36 SamB: when people say glass in crawl that's what they mean 18:55:49 ...really, you could make the one entry vault with tons of stones in it that 18:56:45 sure, it would be scriptable. but some people wouldn't work it out, and there would be forum posts... 19:00:05 number of insurance cards i have received thusfar: *four* 19:21:21 Mumra: I wasn't suggesting using rocks to break out of it, just make the walls of that vault translucent rock. 19:25:30 aha 19:26:02 -!- Twilight has joined ##crawl-dev 19:26:28 -!- Twilight is now known as Guest55239 19:28:00 -!- Guest28783 has quit [Ping timeout: 265 seconds] 19:32:04 -!- evilmike_ has joined ##crawl-dev 19:33:05 -!- evilmike has quit [Read error: Connection reset by peer] 19:34:01 -!- dpeg has quit [Quit: zzz] 19:35:17 thoughts on removing the gold values for inventory items upon winning? 19:35:34 since they really don't mean anything now 19:38:14 I thought they were already out? 19:38:20 I think I keep thinking this 19:38:51 I don't see an issue with it 19:41:56 -!- ainsophyao has joined ##crawl-dev 19:42:18 would score for gold remain in? 19:46:40 wensley: I should clarify... currently items in your pack do not affect your score, but their gold values are still displayed when you win 19:46:57 oh, I see 19:47:07 I don't know if this is intentional 19:47:24 right, seems pointless 19:48:00 also weird that the horn of geryon of sword of cerebov would even have a price... 20:04:43 the idea would've been to reward players with points (from the item price) for carrying out those items, as a trophy 20:13:54 the question becomes, why should they need to carry them out to be rewarded for them 20:22:39 -!- Guest55239 has quit [Read error: Connection reset by peer] 20:26:37 -!- Twilight-1 has joined ##crawl-dev 20:28:36 -!- RichardHawk has quit [Ping timeout: 252 seconds] 20:36:07 -!- Torokasi has quit [Ping timeout: 252 seconds] 20:58:39 yeah, people should get rewarded for stuff they had to dump somewhere, too, shouldn't they? 21:00:31 -!- evilmike_ is now known as evilmike 21:13:35 -!- Zaba has quit [Ping timeout: 258 seconds] 21:15:11 -!- Zaba has joined ##crawl-dev 21:15:32 -!- mikee_ has joined ##crawl-dev 21:42:45 Crash when my high-elf hunter had just killed a monster (https://crawl.develz.org/mantis/view.php?id=4574) by vonbrand 21:50:23 -!- varmin has quit [Ping timeout: 258 seconds] 21:51:55 -!- greensnark has quit [Ping timeout: 258 seconds] 21:53:43 -!- greensnark has joined ##crawl-dev 21:56:13 -!- varmin has joined ##crawl-dev 21:58:13 -!- Pingas has quit [Read error: Connection reset by peer] 22:06:40 03kilobyte * rd7df244994ed 10/crawl-ref/source/fight.cc: Fix diagonal monster attacks causing crashes. 22:11:54 -!- Mu_ has quit [Quit: Defecator, may everything turn out okay so that you can leave this place.] 23:22:56 -!- Wensley has quit [Ping timeout: 260 seconds] 23:38:57 -!- mikee_ has quit [Quit: HydraIRC -> http://www.hydrairc.com <- It'll be on slashdot one day...] 23:50:41 LexAckson (L17 TrGl) ASSERT(you.can_pass_through_feat(grd(p))) in 'player.cc' at line 409 failed. (Abyss) 23:55:40 -!- dtsund has quit [Quit: dtsund]