00:04 aosdict: currently eroded weapons effectively get an -x enchantment 00:04 a rusty +0 dagger is essentially a -1 dagger 00:13 -!- theRaisse has joined #hardfought 00:13 -!- mode/#hardfought [+v theRaisse] by ChanServ 00:25 nice patch idea: forges in the dungeon so that you can repair a weapon you attacked an acid blob with 00:25 jonadab, are you there? 00:29 -!- APic has quit [Ping timeout: 246 seconds] 00:39 -!- APic has joined #hardfought 01:21 -!- greqrg has quit [Quit: Disconnected] 01:22 [gh] Tangles (Hea Gno Mal Neu), 4591 points, T:5717, suffocated by a gelatinous cube 01:22 Didn't even see it. 03:27 -!- theRaisse has quit [Ping timeout: 246 seconds] 03:37 -!- theRaisse has joined #hardfought 03:37 -!- mode/#hardfought [+v theRaisse] by ChanServ 03:41 -!- LarienTelrunya has joined #hardfought 03:47 -!- theRaisse has quit [Ping timeout: 240 seconds] 03:55 Evil Patch idea: the sokoban stair is directly underneath the Oracle, she cannot be teleported away, and you can only reach the Elemental Planes by #chat-ting to her while holding the Amulet of Yendor. 04:17 -!- theRaisse has joined #hardfought 04:17 -!- mode/#hardfought [+v theRaisse] by ChanServ 04:18 -!- Crawldragon has joined #hardfought 04:23 -!- theRaisse has quit [Ping timeout: 252 seconds] 04:26 Evil Patch idea: using #enhance while having skills that can be enhanced will reduce your amount of skill slots by one even if you don't actually specify any skill to enhance. And once you enhance a skill, it can not be trained any further; to reach Expert, you have to use the skill without enhancing it for a while until you got enough training to reach Expert, but the game doesn't tell you when you've reached that amount 04:27 Amy don't you think your patches are already evil enough? 04:28 -!- elenmirie_ has joined #hardfought 04:29 -!- elenmirie has quit [Ping timeout: 252 seconds] 04:29 -!- elenmirie_ is now known as elenmirie 04:30 -!- mode/#hardfought [+v elenmirie] by ChanServ 04:31 Not all of my EPIs actually get implemented. EPIs are actually a graveyard for bad ideas that are never meant to be implemented; it just so happens that some of them (but not all) fit into a certain infamous NetHack variant. 04:31 I was doing a bunch of programming yesterday and I did something I think you'll find funny 04:32 I finally got Spelunk!'s inventory system working, so you have a bunch of item slots for helmets, cuirasses, weapons, etcetera 04:32 but I programmed it so that if you try to equip a weapon in the "helmet" slot the player character stabs himself in the head and dies 04:32 lol, really? without a warning? 04:33 is Spelunk! supposed to be "instagib the char for every silly mistake, no matter how minor"? 04:33 No, but it is meant to be an homage to old-school adventure games 04:33 Well the helmet slot is a ways down on the inventory screen and it normally doesn't allow you to equip inappropriate items in the wrong slot 04:33 So you can't wear a helmet as a glove, right, and you can't wear a weapon as a boot 04:34 I might disagree on that last one. :P 04:34 * LarienTelrunya wields a block-heeled combat boot and whacks Crawldragon over the head! 04:34 but in the grand tradition of these things I couldn't resist but add a funny death scene to the game 04:35 I had another idea where if you get the turn counter all the way up to INT_MAX the universe goes into heat death and the character is listed as being "killed by Newton's Second Law of Thermodynamics" 04:35 yeah, adventure games (specifically stuff like Leisure Suit Larry, where not checking the road before crossing it would lead to being run over by a car spawned out of nowhere) are what that helmet stabbing death immediately reminded me of 04:37 Spelunk! is going to be a very silly game when it's finished. One of the monsters I have planned is called a Hidebehind and it's interpreted from legends that American lumberjacks used to tell to scare the new guy. 04:49 okay I'll be off for today and try to finish the current slex version; tomorrow is (planned) update day! Woohoo! :) 04:49 Woooo! 04:50 you should play too once it's online! *bundlebundlebundle* :D 04:50 it will have new terrain types - sand, snow, tunnels, stalactites, nether mist and many more! 04:51 ooooh stalactites 04:51 yeah, some of those terrains are hazards for flying/levitating characters specifically, just like water and lava have always been hazards for characters without those properties 04:51 stalactites will obviously impale flying characters but are completely harmless if you're on the ground ;) 04:56 -!- LarienTelrunya has quit [Quit: You see a pentagram inscribed on the floor here. Use #invoke to draw on the magical energies.] 05:55 -!- Crawldragon has quit [Quit: If beer is liquid bread, what is liquid peanut butter?] 07:23 -!- Grasshopper has joined #hardfought 07:23 -!- mode/#hardfought [+v Grasshopper] by ChanServ 07:28 Hello from hospital in Thailand :C 07:39 eep what are you in hospital for, Grasshopper? 07:40 !tell crawldragon the name "spelunk" for a roguelike-inspired game is already taken 07:40 Will do, Grasshopper! 07:41 elenmirie: I had an accident on my bike while crossing the bridge between Laos and Thailand 07:41 oh no... 07:41 get well soon! 07:42 train tracks crossing the road... my bike's front wheel went into the gap, and the pedal ripped a 5cm long and deep gash in my ankle 07:42 ah 07:42 K2: Message from FIQ at 2017-09-02 13:17 EDT: fiqhack updated: now, objects merge regardless of ID knowledge, and acid res is now shown if you have it intrinsically in the C-x a menu 07:42 K2: Message from FIQ at 2017-09-02 14:25 EDT: pushed fiqhack updates: more wands are ID-ed on use, arc gets a searching bonus 07:42 K2: Message from FIQ at 2017-09-02 16:13 EDT: pushed a bug fix to fh 07:42 ouch ouch ouch! 07:43 just a flesh wound 07:43 http://i0.kym-cdn.com/entries/icons/facebook/000/012/877/3qafhk.jpg 07:43 K2: funny, I was wondering what Conan would think 07:44 he probably wouldn't pay it any notice 07:44 just a scratch, all in a day's work... 07:44 ;) 07:44 I had to ride about 1km with my right ankle bleeding everywhere until I got to the immigration station 07:45 then I asked for help 07:45 you got blood all over the bridge? 07:45 pedalling with one foot 07:45 yeah, they'll have a hell of a time cleaning it 07:46 or maybe they'll just wait until it rains 07:47 but on the bright side, I get a free night's accommodation and 8 free stitches! 07:47 cool! 07:47 (medical insurance) 07:48 Are you far from home? 07:48 3 hours drive 07:48 better I take a night in hospital and drive tomorrow 07:48 Ah, so you won't have to ride the bike back I hope 07:49 no, but driving will be hard enough 07:49 mmm, yeah, not an automatic then 07:50 no... but right foot so needed regardless :D 07:51 I love autos.... I can just relax my left foot but my right always gets jealous 07:53 but hey, the bright side is I can add playing Nethack from a hospital to my "extreme Nethacking" credentials 07:54 from a hospital in Thailand. that's worth extra 07:54 FIQHack updated 08:05 [nd] elenmirie (Mon Hum Fem Neu), 4868 points, T:3314, killed by a dwarvish spear 08:55 FIQ: yes, I meant whether erosion should be moved to a damage-only penalty, instead of counting towards both. 08:56 Grasshopper: ouch! I had the same type of accident several weeks ago, didn't get banged up as badly as you though 08:59 -!- theRaisse has joined #hardfought 08:59 -!- mode/#hardfought [+v theRaisse] by ChanServ 09:04 -!- theRaisse has quit [Quit: Raisse suddenly disappears!] 09:04 -!- theRaisse has joined #hardfought 09:04 -!- mode/#hardfought [+v theRaisse] by ChanServ 09:07 -!- theRaisse is now known as raisse 09:07 morning raisse 09:07 afternoon! 09:07 err afternoon for you i imagine 09:07 heh 09:07 3pm here 09:08 9am here 09:08 did not want to get up but my 5 yr old had other plans 09:08 we've just been into town only to find that the cookware shop was closed where we wanted to buy our youngest daughter a really good frying pan for her birthday 09:08 (which was yesterday) 09:09 she had a cheap one that she used so much that it wore down in 6 months 09:09 6 months hah 09:09 was she using it 10 times a day? 09:10 no, it was a really crappy cheap one and she used it about twice a day (breakfast pancakes, and dinner whatever) 09:10 ah 09:10 that sort of use calls for a 100 euro pan that will last until she can bequeath it to her granechildren 09:11 s/grane/grand/ 09:11 worth every penny 09:11 yes, exactly 09:11 that's why we thought of it as a birthday present and she was all in favour 09:11 many shops here are open on the 1st Sunday of the month but not that one 09:13 my parents gave me various pieces of cookware for birthday gifts until I had to tell them to stop 09:13 i have an old iron skillet i've had for years, use it for camping out. it'll last forever 09:14 you can only usefully have a certain amount of cookware until you actually have a big kitchen all to yourself, which I do not 09:14 ours is overflowing with stuff we never use hah 09:15 we're actually organizing and consolidating items this weekend. cant believe it 09:15 I think I still have a ceramic garlic holder shaped like a garlic bulb. such a marginal use 09:17 [gh] k2 (Mon Hum Mal Neu) had knife bestowed upon him by Chih Sung-tzu, on T:2961 09:18 -!- raisse has quit [Quit: Raisse suddenly disappears!] 09:19 -!- raisse has joined #hardfought 09:19 we have a terracotta pot with holes in the sides, inscribed "garlic cellar", and it actually does keep garlic fresher than anywhere else 09:19 -!- mode/#hardfought [+v raisse] by ChanServ 09:19 -!- raisse has quit [Read error: Connection reset by peer] 09:19 hmm 09:20 i should make my wife aware of that object 09:20 'garlic cellar' 09:20 we love garlic 09:20 -!- theRaisse has joined #hardfought 09:20 -!- mode/#hardfought [+v theRaisse] by ChanServ 09:20 huh, so putting garlic in an enclosed space with a few airholes is actually useful? 09:20 we didnt grow garlic this year though :( 09:20 and we use *a lot*of garlic 09:21 we don't usually grow garlic either, just buy it in the store 09:21 -!- theRaisse is now known as raisse\splat 09:21 -!- raisse\splat is now known as theRaisse 09:21 -!- theRaisse is now known as raisse 09:21 many shops here are open on the 1st Sunday of the month but not that one 09:21 we grew some last year but the dogs dug up most of it 09:21 sorry for repeat, network problems 09:21 no worries 09:21 isn't garlic toxic to dogs? 09:21 well, at least none of us are vampires 09:21 oh they just dug it up, they didn't eat it 09:21 right 09:21 I know that potatoes and onions can be harmful to dogs. 09:21 If onions are, garlic probably is too 09:22 raw potatoes are harmful for everybody 09:22 (at least all mammals afaik) 09:22 mmm nightshade-family plants 09:22 we're building more raised and enclosed beds this year in preparation for next season 09:22 ah, harder to reach for dogs 09:23 and the rabbits who moved into our yard this year ;) 09:23 they sneak in when the dogs are napping 09:25 -!- raisse has quit [Quit: Raisse suddenly disappears!] 09:25 -!- theRaisse has joined #hardfought 09:25 -!- mode/#hardfought [+v theRaisse] by ChanServ 09:30 -!- theRaisse has quit [Read error: No route to host] 09:31 -!- theRaisse has joined #hardfought 09:31 -!- mode/#hardfought [+v theRaisse] by ChanServ 09:36 -!- raisse has joined #hardfought 09:36 -!- mode/#hardfought [+v raisse] by ChanServ 09:36 -!- theRaisse has quit [Quit: Raisse suddenly disappears!] 09:36 I think that only the leaves/stems of nightshades are not safe to eat 09:37 Grasshopper: and the flowers/seeds 09:37 Aren't green potatoes supposed to be at least mildly poisonous? 09:37 yes 09:39 green potatoes yes 09:45 [slex] dolores (Cav Red Fem Law), 14914 points, T:2553, burned by a fireball 09:45 i ate lots of raw potato as a kid. never bothered me. 09:45 when my mom was making chips i'd eat raw ones. i liked them 09:45 worst raw potato will do is give you a mildy upset tummy 09:46 now that's not to say they aren't hermful somehoe 09:46 er 09:46 harmful 09:46 on other recent topics, I never actually use a garlic press - I just crush the cloves under a flat-bladed chef's knife 09:47 then pull the skin out (it is easy once the cloves are crushed) and chop up the rest with the knife 09:47 to aosdict's point, both garlic and onion are poisonous to dogs 09:47 In Thailand, cloves of garlic are like a vegetable 09:47 so they don't even shop them up much 09:48 chop 09:48 a monk can eat a clove of garlic without guilt, that's all i'm sayin' 09:48 yeah, but who's going to kiss that monk afterwards? 09:48 a foocubus of course 09:49 ... a lucky person, as garlic enhances your "sexual health" 09:49 you should know this stuff by now 09:49 which is why my wife is always force-feeding it to me 09:49 :D 09:50 garlic breath turns me on 09:50 ... 09:50 get a room you two 09:50 oooh did i type that out loud? 09:54 is game replay an nh4 thing? 09:55 i really wish fiqhack's ui could be ported to 3.6 09:57 K2: I've got a room. With a fridge, TV, lounge and a great bed 09:57 All I had to do to get it was to bleed all over the place 09:58 Grasshopper: that's how I clean garlic too 10:03 -!- raisse has quit [Ping timeout: 240 seconds] 10:08 -!- noty has joined #hardfought 10:08 raisse / irina|log I think that garlic might possibly be my favourite vegetable 10:12 Grasshopper: same thing regarding using a knife instead of a garlic press. garlic presses are very annoying to clean and they always lose some of the garlic 10:13 heh 10:13 except I usually just twist and bend the clove till the skin comes off 10:13 or, for maximum noise, put the cloves in a hard metal pot and shake them around for a couple minutes 10:18 aosdict: my feelings exactly about garlic presses - waste of garlic and waste of time to clean 10:19 might make some potato leek soup later, complete with delicious garlic of course 10:20 pretty cold and rainy today, the perfect time to make a hearty soup 10:32 yani: the base Pw cost for a spell should be specified per-spell, so you can have low-level spells that are easy to cast but expensive, and high-level spells that are hard to cast and extremely expensive. 10:38 aosdict: YUM YUM 10:40 [gh] Tangles (Ran Orc Mal Cha), 1154 points, T:1512, killed by a giant beetle 10:45 hothraxxa: Goto turn only exist in fiqhack and dynahack (it existed in nitrohack but was removed in NH4 for whatever reason). Other than that, replay mode is a Nitro/NH4 thing 10:46 thanks. i like being able to remind myself what i was doing 10:46 One thing I miss about nitro replaymode is speed 10:46 NH4 replaymode is *slow* 10:47 i'm slow 10:47 compared to that, nitrohack replaymode isn't slow at all 10:47 i am slowing down 10:47 nooo 10:47 here eat this lizard 10:47 going forwards and backwards is completely seamless 10:47 yes. i turned 70 last week 10:47 happy birthday. i think 10:47 ty 10:48 [nd] Grasshopper (Cav Dwa Fem Law) killed Orcus, on T:59360 10:48 my dad turns 78 on the 16th 10:48 he's like 'ugh' 10:48 'bleh' 10:48 the number doesn't matter. how you feel is what counts 10:48 'aaaaaargh' 10:48 true 10:49 as longas as you can still fuck. that's all that matters 10:49 i'm at least 10 years younger than my calendar age 10:49 ... 10:49 hothraxxa: appropriate time to read Mark Twain's seventieth birthday speech? 10:49 the day that's no longer possible is the day i'll be ready to depart this place 10:49 i did not know he wrote one 10:50 to ascend as it were 10:53 [nd] Grasshopper (Cav Dwa Fem Law) killed Baalzebub, on T:59390 10:58 that twain guy had a way with words. he should try writing 10:59 FIQ: in fh, god gifts depend only on the number of god gifts previously given, right? I feel like that should still account for the number of wished artifacts 10:59 but not the other way around 10:59 hothraxxa: Happy birthday 10:59 artiwish odds depend only on the number of wished artifacts, but god gifts depend on the number of wished AND gifted artifacts 10:59 hey, don't make him change that yet! 10:59 for last week 11:00 grasshoppe: ty 11:00 I love that you are a septuagenarian 11:02 bwahahaha, stupid bitch! she could have had my orb of detection, but no... The water nymph steals a floating eye corpse from Sirius Fudd! 11:03 in the absence of evidence to the contrary, i intend to be the oldest to ascend 11:03 nymphs never were that intelligent 11:03 a title i likely already hold 11:03 hothraxxa: in 20+ years I'll supplant your title 11:03 you lucky bastard 11:04 assuming I on't get killed due to stupidity before that 11:04 i've been killed by stupidity lots of times 11:04 like letting my front wheel get caught in the railway tracks 11:05 ah yes. was that bicycle then? 11:05 mind you, I blame the relevant authorities... the Thai paramedics told me that my accident has happened several times before 11:05 4 lamps in izchak's shop. all oil 11:05 hothraxxa: yes bicycle 11:06 I am the oldest & fattest bicycle rider in Thailand .... I think 11:06 [gh] k2 (Mon Hum Mal Neu), 6054 points, T:6545, killed by a dwarvish watchman 11:06 unlikely 11:07 there are people here that are older and fatter... but I don't see them riding bikes as seriously as I do 11:08 and apparently my blood pressure has been dropping (nurses here check it every few hours, and it is lower than it was years ago) 11:10 I'm feeling better than I did years ago too... :D 11:11 hothraxxa: at this rate, by the time I'm your age I'll be Superman (better get my costume dry-cleaned, just in case) 11:14 !lotnoty Grasshopper 11:15 [nd] munterplant (Val Hum Fem Law) had Mjollnir bestowed upon her by Tyr, on T:4011 11:16 fiq: my fully id'ed blessed +o thirsty daggers of fire have stopped stacking. any idea why? 11:18 [nd] Grasshopper (Cav Dwa Fem Law) killed Asmodeus, on T:59476 11:19 !tell tangles how about a !lot command to take the pressure off my luck? 11:19 Will do, Grasshopper! 11:22 -!- NeroOneTrueKing_ has joined #hardfought 11:25 -!- NeroOneTrueKing has quit [Ping timeout: 240 seconds] 11:34 good idea Grasshopper 11:57 -!- Tarmunora__ has quit [Read error: Connection reset by peer] 11:58 -!- Tarmunora__ has joined #hardfought 11:59 -!- greqrg has joined #hardfought 12:03 * aosdict sautees leeks with garlic and butter. You smell a delicious smell. 12:12 -!- NeroOneTrueKing has joined #hardfought 12:15 -!- NeroOneTrueKing_ has quit [Ping timeout: 246 seconds] 12:24 fiq: that save corruption might be connected with my iron golem's poisonous cloud breath 12:27 yasi: spinach monster, which becomes a tin of spinach when tinned 12:28 Shouldn't spinach be a vegetable? 12:29 it is a vegetable, unless you mean it should be a % 12:32 in the "virtually no permafood is actually permafood" patch I'm contemplating, perhaps you should be able to tin fruits and veggies 12:33 then you need to increase the nutrition value for tins 12:33 why? 12:34 because most tinned food is hardly worth eating 12:34 certainly not worth the weight 12:35 that's kind of the point 12:35 if you want true permafood and have a tinning kit, you have to carry around a bunch of 50 nutrition tins 12:37 otherwise you have to stick to corpses and eating other comestibles before they go bad 12:37 which would be the default play, since tinning kits aren't that easy to come across 12:39 hothraxxa: a homemade tin has the same nutrition as an apple. 12:39 Well more ethan a kelp frond. 12:39 you know what apples are good for? horses 12:39 Granted, nowhere near as much as spinach. 12:40 Personally, I think the type (preparation method) should be deterministic so that a stack of tins are all actually the same. 12:40 So that if you have a stack of three tins, and eat one and it doesn't make your fingers greasy, the otherw two won't either. 12:40 Granted, that's not an issue for homemade tins. 12:45 That feeling when you find a +4 crystal plate mail lying on the floor. 12:46 [slex] dolores (Scr Ang Fem Cha), 14386 points, T:5766, killed by a blaster bolt 12:47 [dnh] NeroOneTrueKing (Ana Dwa Mal Cha) entered Minetown for the first time, on T:1525 12:57 -!- bug_sniper has quit [Ping timeout: 248 seconds] 13:01 Water gushes forth from the overflowing fountain! The Oracle drowns. 13:02 sorry 13:07 oracle should know to wear floaties 13:09 hothraxxa: possible 13:10 i noticed it coincided the last two times 13:10 hardly definitive 13:10 ok 13:10 but something to consider 13:13 Oracle should come to me for swimming lessons 13:18 fiq: any idea why my daggers stopped stacking? 13:24 yani: the advanced forms of spells are castable at Basic, but they up the failure rate. 13:25 Alternatively, do away with advanced spells completely, and have the spellbooks teach you two spells at once. (i.e. the spellbook of fireball teaches you both "fire bolt" and "fireball") 13:25 [dnh] NeroOneTrueKing (Ana Dwa Mal Cha) completed Sokoban, on T:4745 13:27 or "fireball" and "fire storm" 13:27 "cone of cold" and "glaciate" 13:27 assuming the actual spells aren't changed, fire bolt seems appropriate for a ray 13:28 cone of cold also seems appropriate for a ray, since cone spells aren't a thing 13:28 but Basic fireball is a ray that explodes on hitting the first target? 13:28 Is it the first target, or the first nonreflecting target? 13:30 I think I like fire blast and frost blast as names for the advanced spells, though fire storm and glaciate sound cool and powerful 13:31 -!- raisse has joined #hardfought 13:31 -!- mode/#hardfought [+v raisse] by ChanServ 13:32 It appears to explode regardless of whether or not the target has reflection 13:33 I like the two-separate-spells idea because it lets you balance them separately 13:33 hothraxxa: hm 13:34 probably me fucking up the new change to merging 13:34 Fireball could be dropped to level 3 and fire blast/storm could be raised to 5 or 6 13:34 And cast them more easily, too. (Z) -> (letter) instead of (Z) -> (letter) -> (y/n) 13:34 fiq: started to happen after k2 pushed the latest version, so probably 13:34 NeroOneTrueKing: yes 13:34 I like it 13:35 i was looking forward to enchanting these daggers 13:35 only problem is how to define two spells and have them both learnable from one spellbook. Sounds like a SMOP. 13:35 hothraxxa: fixed 13:35 although they hardly ever have the fire effect. how does that work? does it have only one of the effects? 13:35 yeah 13:36 ty 13:36 FIQ: i' 13:36 sorry about the inconvenience (probably nothing merges atm) 13:36 i'm around if you put out a fix soon 13:36 possibly could add in some other non-generated spellbooks for those spells 13:36 I did 13:36 K2: fh updated 13:36 k 13:36 FIQhack updated 13:36 aosdict: never-generated spellbooks of "fire blast" and "frost blast" 13:36 NeroOneTrueKing has played dcss? 13:37 a little 13:37 judging by your name suggestions :P 13:37 hothraxxa: save close and restart your game please 13:37 already did 13:37 and daggers merged 13:37 NeroOneTrueKing: fire bolt/frost bolt/fire storm/glaciate 13:37 thanks all 13:37 They are cool names, but nothing beats "Lee's Rapid Deconstruction" :D 13:37 haha 13:38 NeroOneTrueKing: ADOM's name for what nethack calls "dig" is kind of lame though 13:38 it's called "Magic Shovel" 13:38 gnite folks, time to sleep and let my foot rest (is gettting swollen, need some sleep) 13:38 aosdict: and add to the "read spellbook" code some extras to give the player both spells when reading the randomly generated spellbook 13:40 of course, then you could get a level 5/6 spell from a level 3 spellbook without risking the higher failure rate, but that's probably marginal 13:41 hothraxxa: the idea was to merge items regardless of if you know enchantment, erodeproofness, etc, or not 13:42 A more SMOP-ish method would be to check if the player knows the basic spell, and only if they are also Skilled+ give an option in the Z menu for the advanced spell. 13:42 doable, certainly 13:43 dnethack does something of the sort with artifacts that let you cast a particular spell while wearing/wielding the artifact 13:43 fiq: so pretty much the opposite of what happened 13:44 -!- Tarmunora_ has joined #hardfought 13:44 -!- NeroOneTrueKing_ has joined #hardfought 13:45 well, the reason I want to do this is so the advanced spell is available to people who can't get skilled in attack spells 13:47 Ah okay 13:48 -!- Tarmunora__ has quit [Ping timeout: 240 seconds] 13:48 -!- NeroOneTrueKing has quit [Ping timeout: 248 seconds] 13:48 [gh] k2 (Mon Hum Mal Neu), 1494 points, T:1968, killed by a gnomish watch captain 13:50 -!- Tarmunora__ has joined #hardfought 13:54 -!- Tarmunora_ has quit [Ping timeout: 248 seconds] 13:57 -!- raisse has quit [Ping timeout: 240 seconds] 14:03 [nd] simplejim (Wiz Gno Fem Neu), 2442 points, T:2907, quit 14:31 -!- raisse has joined #hardfought 14:31 -!- mode/#hardfought [+v raisse] by ChanServ 14:39 -!- raisse has quit [Quit: Raisse suddenly disappears!] 14:39 -!- raisse has joined #hardfought 14:39 -!- mode/#hardfought [+v raisse] by ChanServ 15:12 [gh] k2 (Mon Hum Mal Neu), 164 points, T:393, killed by a dwarf 15:12 -!- NeroOneTrueKing has joined #hardfought 15:15 -!- NeroOneTrueKing_ has quit [Ping timeout: 248 seconds] 15:15 -!- NeroOneTrueKing_ has joined #hardfought 15:16 -!- NeroOneTrueKing has quit [Ping timeout: 240 seconds] 15:22 [nd] simplejim (Rog Orc Fem Cha) rejected atheism with a prayer, on T:5092 15:24 [nd] simplejim (Rog Orc Fem Cha), 3786 points, T:5240, killed by a plains centaur 15:28 [gh] k2 (Arc Dwa Mal Law), 1140 points, T:1224, killed by a newt 15:35 [nd] Raisse (Cav Dwa Fem Law) became literate by engraving "gkjd", on T:4931 15:37 [nd] Raisse (Cav Dwa Fem Law) consumed animal products for the first time, by eating a goblin corpse, on T:5824 15:41 [nd] Raisse (Cav Dwa Fem Law) averted death, on T:6640 15:44 [gh] k2 (Arc Dwa Mal Law), 349 points, T:1054, killed by a kitten 15:48 [nd] Raisse (Cav Dwa Fem Law), 2676 points, T:7721, killed by a Woodland-elf 15:51 [fh] hothraxxa (Arc Hum Mal Neu) wished for "blessed +3 grayswandir", on T:36797 15:52 [gh] k2 (Arc Dwa Mal Law), 1359 points, T:834, killed by a gnomish wererat, while disrobing 15:56 -!- zombifier has joined #hardfought 15:58 -!- raisse has quit [Ping timeout: 240 seconds] 15:59 hothraxxa: more or less 16:08 [fh] hothraxxa (Arc Hum Fem Neu) wished for "blessed fixed historic +3 boots of speed named red ball zoomies", on T:36849 16:10 wtf? i got water walkers 16:10 fiq: are there changes to wishing that i should be aware of? 16:12 knowing FIQ, probably 16:18 [gh] k2 (Arc Dwa Mal Law), 847 points, T:2664, zombified by a dwarvish zombie 16:18 compared to? 16:18 you can't wish for object properties 16:18 artiwishing is 100%, 50%, 33%, 25%, 20%, etc since now only artiwishing affects wish rate 16:19 that is the changes to wishing 16:19 I think "boots of speed" doesn't convert to speed boots 16:19 you need to have said "speed boots" 16:19 hm 16:19 if that's the case, there's a bug 16:20 hm 16:20 yep you're right 16:21 did you change speed boots like grunthack has them? 16:21 where spped boots technically dont exist in gh 16:21 just the property 16:23 water walkers are useful to me right now so it's not a disaster 16:23 they do exist in grunthack 16:23 hothraxxa: sorry 16:23 np 16:24 FIQ no they dont 16:24 K2: they do 16:24 BOOTS("speed boots", "combat boots", 16:24 0, 1, FAST, 12, 2, 20, 50, 9, 0, LEATHER, HI_LEATHER), 16:24 speed boots dont exist in gh. you can find boots with speed property tho 16:24 they do 16:24 fake news 16:24 you are fake news 16:24 you are alternative facts 16:24 lol 16:25 all i know is, you cant wish for speed boots in gh 16:25 but yeah aosdict is right 16:25 what happened was 16:25 cuz they dont exist 16:25 it interpreted it as an object property 16:25 -!- ais523 has joined #hardfought 16:25 K2: can you read 16:25 I literally just pasted from grunt 16:25 :P 16:26 ais523: hi, I was going to ask something 16:26 is there some struct or whatever for holding the entire gamestate? 16:26 "don't ask to ask, just ask" 16:26 ais523: Message from jonadab at 2017-09-01 15:43 EDT: I like the idea of search always succeeding if it doesn't get interrupted, but taking multiple turns to do so. So A) you'd know when you've searched enough, and B) you don't have to use numeric prefixes. Better UI. 16:26 basically I wanted to figure out a way to implement replay checkpointing 16:27 i think you just made that up FIQ 16:27 and the answer is a) in regular NetHack, no; b) in NetHack 4, not yet but progress has been made in that direction 16:27 it'd clearly be a useful thing to have 16:27 ais523: well I was thinking 16:27 store some static var (or something else if possible) in logreplay.c as a checkpoint 16:27 for, say, 50 turns back 16:27 to speed up replaymode rewinding 16:28 I thought the game already had such a var? 16:28 FIQ: so you're saying the same bug you just found in FIQhack is also in gh (and prob in dynahack too) 16:28 Just wasn't sure how the save system worked so was going to ask you about it 16:28 K2: in grunthack, the bug does exist yes 16:28 because the dynahack author mentions it 16:28 less sure with dynahack, but possible 16:29 ais523: i.e. I figured the game used a var to store gamestate stuff when saving/restoring (isn't this what mf is for?) 16:29 so basically, store one of these for 50 turns back 16:29 and quickjump from there and go forwards, when rewinding 16:30 -!- ais523 has quit [Remote host closed the connection] 16:30 bye 16:32 -!- ais523 has joined #hardfought 16:32 FIQ: I read that as l ogre play 16:32 wat 16:32 you read what as l ogre play? 16:32 logreplay.c 16:32 oh 16:33 lol 16:33 ais523: not sure how much you missed, but yeah, isn't mf basically what I want? 16:33 hm no that isn't it 16:33 K2: https://github.com/NHTangles/GruntHack/blob/master/src/objects.c#L467 if you want a nice link from your own source repository 16:33 it's just a file descriptor which is useless for keeping track of gamestste 16:33 FIQ: the gamestate is stored in a bunch of global variables in vanilla 16:33 thx aosdict 16:34 in nh4 there's a struct gamestate which is meant to collect them 16:34 I am talking about NH4 context here, just to clarify 16:34 but I don't think they're all there yet 16:34 also, another thing I wanted to do 16:34 hmm this whole time i thought they didnt exist 16:34 but whenever I tried, I would get weird desync errors 16:34 i blame FIQ 16:35 was to make the forward thingy ignore diffs and playback using the commands 16:35 just like "forwards 1 turn" does 16:35 just as a perf test 16:36 that would break whenever options were changed 16:36 as that just records the result of the option change in the diff, it doesn't actually record the nature of the change 16:36 ais523: it would break in more cases than that 16:37 but it was meant as a performance test 16:37 to see what actually is faster 16:37 basically to see if it's worthwhile to try to use commands whenever possible in replaymode 16:38 how do you know if it's possible? 16:38 in the rare case where dynahack *doesn't* break the commands horribly (not really sure why it does that, nitrohack doesn't), it speeds up replaymode by one order of magnitude 16:38 this is why I wanted to test it 16:40 ais523: the game notices 16:40 "Warning: desync in (whatever), expected '~'" or whatever it says 16:40 it wouldn't necessarily notice immediately 16:40 only when a command took a different number of parameters 16:41 (e.g. because it's successful in one state but errors out before asking for parameters in the other) 16:41 but it doesn't take long in my experience 16:41 the longest I've seen it take until it notices is like 10 turns 16:41 basically the idea is 16:42 try to playback using the commands until desync, checkpointing every 25 turns or so 16:42 figure out where the desync happened 16:42 and playback those actions specifically using the diffs 16:43 but first I want to see if this is a worthwhile exercise 16:43 by figuring out what the difference, speed-wise, is, between replaying using commands, and using diffs 16:43 but when I tried to do that 16:43 -!- ais523 has quit [] 16:43 the game would desync in mysterious ways or just crash outright 16:43 and I feel that I was doing something wrong 16:44 -!- ais523 has joined #hardfought 16:44 bye 16:44 basically the idea is 16:44 try to playback using the commands until desync, checkpointing every 25 turns or so 16:44 figure out where the desync happened 16:44 and playback those actions specifically using the diffs 16:44 but first I want to see if this is a worthwhile exercise 16:44 by figuring out what the difference, speed-wise, is, between replaying using commands, and using diffs 16:44 that's why I 16:44 but when I tried to do that 16:44 the game would desync in mysterious ways or just crash outright 16:44 and I feel that I was doing something wrong 16:44 bye 16:44 that is what you missed 16:44 (bye was @ quit) 16:45 this connection keeps getting a bizarre problem where it can send but not receive 16:45 reconnecting fixes it 16:45 but it can take me a while to notice it's happened 16:45 (ignore stray "that's why I" line, I was "replaying" using commandbuffer) 16:47 [nd] munterplant (Val Hum Fem Law), 9899 points, T:5657, killed by an energy vortex 16:48 ais523: oh, btw, not sure if I told you before, but I noticed what made travel slow for the player -- there's some point, I didn't look exactly into what, that queries your properties a lot during travel 16:48 by caching the property results, I fixed the travel speed 16:49 so it's pretty much instant with animations disabled, unlike NH4 where it takes 1-2s 16:49 in places like bigroom/etc 16:49 oh good 16:49 I already caught one place where that happened and fixed iti 16:49 but there must have been another 16:50 and IIRC the 2nd most queried thing is strength for the purpose of carrycap stuff, I think fiqhack might have slowed that down due to calculating attributes on demand rather than caching them 16:50 but it didn't impede performance nearly as much, so I didn't bother changing it 16:52 -!- ais523 has quit [] 16:52 -!- ais523 has joined #hardfought 16:53 and IIRC the 2nd most queried thing is strength for the purpose of carrycap stuff, I think fiqhack might have slowed that down due to calculating attributes on demand rather than caching them 16:53 but it didn't impede performance nearly as much, so I didn't bother changing it 16:54 hmm 16:54 urgh, wish parser is annoying 16:54 hmm 16:57 boots of speed is parsed as boots with the speed property 16:57 urgh 16:57 how do I fix this sanely 16:58 why does the wish parser care about properties if you can't wish for them anyway? 16:58 wizmode 16:58 also it used to be possible to wish for properties briefly 16:58 but I figured it was a bad idea 16:59 so I disabled it 16:59 ah 17:02 [gh] k2 (Arc Dwa Mal Law), 187 points, T:616, killed by a gnomish wererat 17:09 ok, fixed it 17:12 !tell K2 updated fiqhack: Make boots of speed give speed boots, and boots of fumbling give fumble boots 17:12 Will do, FIQ! 17:12 [gh] k2 (Arc Dwa Mal Law), 8487 points, T:1114, killed by an iron piercer 17:12 ok 17:12 K2: Message from FIQ at 2017-09-03 17:12 EDT: updated fiqhack: Make boots of speed give speed boots, and boots of fumbling give fumble boots 17:12 Sorry, didn't mean to pierce you that hard 17:12 will that same fix apply to gh and dyna? 17:13 no need to apologize 17:13 i'm used to you ;) 17:13 FIQhack updated 17:14 K2: yeah I think so 17:14 -!- ais523 has quit [] 17:14 well, not straight off, of course 17:15 but similar 17:15 -!- ais523 has joined #hardfought 17:15 cool 17:17 ah 17:17 almost to 30 " FIQhack updated"s 17:17 heh 17:17 -!- NeroOneTrueKing has joined #hardfought 17:18 oh damn 17:18 i forgot 17:19 GruntHack updated 17:19 its in the MotD tho 17:19 what did you do 17:19 from a few days ago 17:19 Tangles added dungeon overview patch 17:19 ah 17:19 that's what bumped it up to 0.2.2 17:20 -!- NeroOneTrueKing_ has quit [Ping timeout: 248 seconds] 17:20 heh 17:20 this is more a coincidence than anything, but my menucolors is kinda of the inverse of how grunthack colors items 17:20 cursed is red in both 17:21 but my menucolor config colors blessed cyan and uncursed green 17:21 ADOM does the reverse 17:21 ah 17:21 er 17:21 s/grunthack/ADOM/ 17:23 K2: so when does grunthack get autoexplore :P 17:23 heh 17:23 i've thought about it, i like the feature 17:24 but it kind of takes the evilness away from all those pools of water and lava w/o paranoid setting for them 17:24 ais523: did you plan on porting the NH4 interface as a windowport for 3.7 or something else? 17:24 [gh] k2 (Arc Dwa Mal Law), 91 points, T:577, killed by a boulder 17:24 why not base it on the existing curses UI and go from there? 17:25 FIQ: I plan to rewrite the NH4 interface into an NH3 windowport, hopefully aimed at 3.7 17:25 but it's slow going 17:25 it's not all that different 17:25 atm I'm mostly working on termtest, which I've made a lot of progress on 17:25 ok 17:25 cool 17:25 best to figure out what's possible in advance of trying to do it 17:25 I see 17:25 alright out for dinner. see you later 17:26 have you figured out a way to do 256color or even truecolor whilst providing a fallback for 16color terms? 17:27 I believe so 17:28 actually, there is a reliable test that can be carried out remotely to see if a terminal supports "paletted 256color" (i.e. 256 simultaneous colours but you can choose what they are) 17:28 it has a few false negatives (e.g. Konsole) but a false positive would be very surprising 17:28 and it gets true positives on at least xterm and gnome-terminal 17:28 so no 256color fun for konsole? D: 17:29 konsole is amazingly broken in several different ways 17:29 they keep fixing them but breaking other things in the process 17:29 it is my terminal of choice, mostly because it does what I want 17:30 and I never bothered changing 17:30 is there a reason to 17:30 and if so, to what? 17:30 incidentally, it's unclear to me what the API for paletted 256 should be; should I just let the program specify truecolor and then reduce it to 256 myself? 17:30 the major issue I have with konsole is the weird rendering errors in NH4 from time to time 17:30 but as I understand it, it isn't unique to konsole 17:31 ais523: that sounds misleading 17:31 I've only ever seen them in Konsole, and I've also never managed to reproduce them outside NH4 17:31 yeah I can't repro outside libuncursed 17:31 FIQ: note that this is literally what xterm does if you send it truecolor codes 17:31 (iirc fiqrogue did the same thing, it used libuncursed too) 17:31 except it doesn't try to optimize the palette, it just uses a default palette 17:31 the test for that in termtest was fun to write 17:33 hm 17:33 actually I should look into that again 17:33 trying to repro term rendering errors in fiqrogue 17:33 due to how barebones it is 17:33 but using libuncursed 17:33 do it through ttyrec 17:34 so that if it happens we at least know what bytes were sent 17:34 I suspect that the exact timing might be relevant, though 17:34 or even things like the exact system calls used to send the bytes 17:34 actually, might be worthwhile to have an strace recording 17:55 -!- zombifier has quit [Remote host closed the connection] 17:56 [dnh] NeroOneTrueKing (Ana Dwa Mal Cha) killed Cerberus, on T:7391 18:03 [dnh] NeroOneTrueKing (Ana Dwa Mal Cha) destroyed Vlad the Impaler, on T:7623 18:05 [dnh] NeroOneTrueKing (Ana Dwa Mal Cha) wished for "blessed amulet of unchanging", on T:7643 18:06 -!- noty has quit [Ping timeout: 240 seconds] 18:11 -!- zombifier has joined #hardfought 18:12 -!- noty has joined #hardfought 18:26 fiq: now i'm sure that it's my iron golem's breath that causes the corrupted save file 18:39 -!- NeroOneTrueKing_ has joined #hardfought 18:42 -!- NeroOneTrueKing has quit [Ping timeout: 248 seconds] 18:42 -!- NeroOneTrueKing has joined #hardfought 18:44 -!- NeroOneTrueKing_ has quit [Ping timeout: 240 seconds] 18:46 ais523: Konsole supports 24-bit color. So you could very easily simulate palettized-256 on it, just by storing the palette yourself. 18:47 jonadab: it supports palettized 256 too (not that you'd care to use it when you have a better option) 18:47 it just doesn't tell you it does when you ask 18:47 -!- bug_sniper has joined #hardfought 18:47 Ah. 18:47 I see. 18:48 Are you still thinking about libuncursed that can do 8-bit and 24-bit color? 18:48 incidentally, I believe truecolor may be possible on the Linux console via exploiting a bug 18:48 and yes 18:49 Cool. 18:51 Lately I've been putting chicanery in my .bashrc that changes $TERM from foo to foo-256color for certain values of foo. The fact that this improves my experience with things like Brogue is indicative of the fact that $TERM is fundamentally broken. 18:52 actually, since I started testing terminals 18:52 I've discovered that many of them have become much better feature-wise 18:52 it's making my terminal testsuite hard to itself test because sometimes I can't find a terminal to fail the test 18:53 -!- zombifier has quit [Remote host closed the connection] 18:53 In general, modern terminals are significantly more capable than termcap/terminfo thinks they are. 18:53 termcap/terminfo can't even express half the things modern terminals do 18:54 They can almost all do 8-bit color. 18:54 Though I'm sure there's an exception somewhere, probably wincon. 18:55 I was disappointed that KiTTY doesn't do 24-bit color, though. 18:55 IPBT is worse 18:55 it understands the higher colour depths but doesn't know how to render them 18:55 so they just come out as white 18:55 Isn't that a multiplexer? 18:56 no, it's a ttyrec player 18:56 Oh, that. 18:56 Ok. 18:56 which i guess is a very specific special case of a multiplexer? 18:56 Yeah, most of the ttyrec players are pretty terrible. 18:56 I'm planning to write a new one 18:56 ttyplay itself is substantially the best one I could find. 18:56 along with libuncursed2 18:57 The worst thing about ttyplay is no rewind at all. 18:57 probably also a terminal, the two would be pretty similar code-wise 18:57 Oooh. 18:57 yes, this is basically because ttyplay doesn't parse the ECMA-48 code in the replay at all 18:57 it just copies it to the terminal with no idea what it means 18:57 so it works great for playing back a recording on the terminal that made it, as long as you don't rewind 18:57 Ah. 18:58 (btw, I've now moved to "ECMA-48" not "VT100" because it's better to be talking about a superset of modern terminals than a subset) 18:58 Indeed. 18:58 Maybe modern terminals will all implement it all. 18:58 Like browsers eventually did with CSS. 18:59 I've started adding ECMA-48 features to termtest even when no terminal I'm aware of supports them, if it's clear what they'd do 19:00 Interesting. 19:00 there are some really weird things in there, both the standard and the terminals 19:00 Is termtest at a point where it would make sense to start talking to the maintainers of various terminals about it? 19:01 for example, ECMA-48 optionally supports a model where you're effectively writing characters and formatting commands to a buffer; and then that buffer is used as input to the actual terminal, writing formatting commands appropriately 19:01 so you can, say, write "abcde", then overwrite the c with a newline, and end up with "ab" on one line and "de" on the next 19:01 I'm really confused as to how it works exactly, though; a brief description like that leaves more questions than answers 19:02 anyway, this means that there are two sets of cursor movement commands, one which moves the cursor within the buffer, and the other of which is just copied to the buffer itself and takes effect on the second half 19:02 I can see an API working that way, but I'd expect such an API to extract the terminal itself away. 19:02 terminals support all the first set, and an apparently arbitrary subset of the second set 19:03 both of which just do the cursor motion directly (which is the other option allowed in ECMA-48) 19:03 so I decided to test the rest of the latter subset as I can see no reason to leave it out 19:03 I see. 19:04 example: in the second set, the terminals I tested support "move right" but not "move left" 19:04 Ooookay, that's rather odd. 19:04 right 19:04 to start with I was just following terminal docs 19:05 it's only fairly recently that I started to look for reasons behind the patterns 19:05 Is the second set _useful_ for anything that you can't simulate by keeping state in your program initially and then sending it to the terminal when ready using the first set of commands? 19:05 and AFAICT, terminal designers are just copying from each other to make the programs work, rather than sitting down and designing things 19:05 s/commands/codes/ 19:05 jonadab: I got the feeling that they were intended for situations where computers weren't involved 19:06 just some sort of network of terminals 19:06 but don't know why you'd do that 19:06 Ah. 19:06 In the past (the rather distant past if we think in computer history terms), I can sort of see that. 19:06 A dumb terminal hooked up to a printer, or something. 19:07 there are a few other situations where the difference matters, e.g. c | would produce | (because the terminal would overwrite the c in memory before it did any printing), but c | could produce ¢ if the terminal decided to just send it all as-is to a printer 19:07 Hmm. 19:08 That seems like it would've been a lot more useful in the days before Unicode, than now. 19:08 ECMA-48 has an "intended" way to overlay characters too 19:08 hothraxxa: hmm ok 19:08 Well, not just unicode, also modern printer control languages. 19:08 but it's very much designed not as "you must implement all this", but as "if you want to implement feature X, here's the code you should use for it" 19:09 Ah. 19:09 like, it's a guide for how to express terminal codes, rather than how to make a terminal 19:09 ais523: ooh 19:09 speaking of term stuff 19:09 The only part that should've been mandatory from the get go is a "do you support such-and-such a code, yes or no" query. 19:10 you should look into mobaxterm for windows 19:10 for testing stuff 19:10 jonadab: xterm has one for SGRs, but it's slightly buggy 19:10 I haven't seen one for most other codes 19:10 it's the terminal MTF (discord person) recommends to people new to ssh stuff 19:10 and what he uses to play fh with 19:10 although gnome-terminal echoes anything it doesn't understand and you can pick that up indirectly via asking about the cursor position 19:10 so there's that 19:11 as for the question asked about termtest, it's only partially complete 19:11 although I've reached a stage where I can often add new codes very quickly 19:11 I feel I approximately doubled the number of supported codes today, haven't counted though 19:11 Interesting. 19:11 I also wrote a gamma test (for white and also red, blue, green individually), that could give some interesting results 19:12 IMO, it wouldn't hurt to have something that is for terminals what the "Acid Test" was for web browser ssh support. 19:12 ooh, SGR 3 really /is/ italic, I thought it was but couldn't find a source 19:12 ERr, css, not ssh. 19:12 and 6 is "fast blink", I wasn't curious about that in the past but perhaps I should have been 19:13 Italic is one of those things that has a long history of being so poorly supported a lot of people don't even know what correct support for it looks like. 19:14 well, it's surely fairly obvious what correct support for italic would look like 19:14 You'd be surprised how many people think it means oblique. 19:15 what is the difference? 19:16 ^ See what I mean? 19:16 * FIQ slaps jonadab around a bit with a large FIQ-bot 19:16 [dnh] NeroOneTrueKing (Ana Dwa Mal Cha) wished for "blessed wand of polymorph", on T:10388 19:16 FIQ: See the illustration here: https://en.wikipedia.org/wiki/Oblique_type 19:17 Oblique is just slanted, basically. 19:17 and that is different from italic how? 19:17 ais523: ^ See? 19:18 FIQ: Italic uses distinct letter forms. 19:18 A proper italic face is more stylized than a roman typeface. 19:18 Even within the same family. 19:22 But yes, if you hit the "italic" button in your word processor, and the font family you're using doesn't have an italic face, oblique is what you get when the word processor or the system's font renderer automatically slants the roman typeface. 19:23 This has been going on for as long as people have been using computers to do word processing, so it's become so entrenched in the public mindset that a lot of people don't actually know that italic is more than that. 19:38 jonadab: come to think of it, "oblique" actually makes more sense in context in the standard, but it does say "italic" 19:41 wow there are some silly/interesting ideas in this standard 19:42 not only does it have an "end of string" in the sense of allowing string parameters to commands, it also has an "end of string" in the sense of "OK, this is the end of the sequence of terminal codes", allowing a sequence of terminal codes to be embedded in a document that's about something else 19:43 Interesting. 19:43 (the latter is ESC d, in case you were wondering) 19:43 And how would a _terminal_ be expected to render that? 19:43 Stop recognizing escape codes from there out? 19:43 That seems like a bad idea. 19:43 The sort of thing that makes users have to close terminal windows and open new ones. 19:44 jonadab: I'm guessing it wouldn't 19:44 but it would make a lot of sense for, say, a standards-compliant ttyrec format 19:44 Ah. 19:44 Sure. 19:44 That kind of makes sense. 19:45 still, I'm not convinced the existing ttyrec(1) format will be at all easy to displace 19:45 especially if the gain is as small as that 19:46 [dnh] Tarmunora (Ana Dro Fem Cha) wished for "The blessed fixed +3 Ogresmasher", on T:8047 19:46 (jettyplay supports some backwards-compatible extensions to the format for expressing things like the terminal size, but changing from a length-prefixed to delimited format would be backwards-incompatible, and length-prefixed formats are easier to optimize the reading of anyway) 20:06 hmm, so there's a code to ask the terminal's model and version number, and gnome-terminal and xterm both implement it 20:06 my gnome-terminal claims to be a VT220 with firmware version 4402 20:06 and xterm claims to be a VT420 with firmware version 324 20:07 huh, konsole replies too 20:07 jonadab: btw, you can just do reset 20:07 no need to close and reopen 20:07 it's a VT100 with firmware version 115 20:07 FIQ: how would you send the reset? 20:07 well 20:07 kill the program that messed up your terminal 20:08 enter reset 20:08 and press enter 20:08 324 appears to be the actual version number of the xterm I have 20:08 FIQ: "stop recognising escape codes" would appear to make it impossible to do even that 20:08 hm 20:08 lovely 20:08 and reset works by sending escape codes, largely… 20:09 I can't see a pattern behind the firmware version reported by gnome-terminal or konsole 20:09 perhaps git blame the source for the terminals? 20:09 -!- remirol is now known as lorimer 20:09 see if the numbers appear somewhere 20:10 hmm, ESC o doesn't seem worth implementing; it purpose is to define validation criteria for locally-echoed input 20:10 e.g. allowing people to type only digits in a particular region of the terminal 20:10 err, I mean CSI o 20:11 ais523: what do you recommend over konsole? 20:11 that'd be useful in the days where any network round trip was slow… 20:11 FIQ: I personally use gnome-terminal; it used not to be great, but has really improved a lot recently 20:11 iirc I had an issue with gnome-terminal playing nethack4 20:11 its cursor looked nearly identical to corridors 20:12 xterm has the best console code parsing of any terminal I know, but it's limited to paletted-256 color depth and its UI sucks 20:12 oh yes, gnome-terminal acts annoyingly with dithered block drawing characters 20:12 ais523: gnome-terminal hasn't been usable since 2.22.0 20:12 jonadab: what does it do wrong? 20:12 They removed important configurability. 20:12 From the terminal profiles. 20:13 I do like konsole's configurability but some of its bugs are just bizarre 20:13 [651][fiq@fiq-desktop ~/fiqrogue_compiled]$ LANG=C gnome-terminal --version 20:13 GNOME Terminal 3.24.2 Using VTE version 0.48.3 +GNUTLS 20:13 a bit later than 2.20 :P 20:13 Specifically, different profiles can't have a different icon in modern gnome-terminal. 20:13 So you end up with 20+ identical icons on your task list. Unacceptable. 20:13 e.g. at one point, it reported mouse clicks off-by-one but only if you clicked on a particular part of the character 20:13 jonadab: oh, I use Unity, which doesn't take the icon from the program anyway 20:14 ais523: Mouse clicks? In a terminal? Why is that even a thing? 20:14 which is actually really annoying, and something I tried to patch myself before discovering that the code underlying Unity is a huge buggy mess 20:14 jonadab: weechat supports it 20:14 jonadab: why wouldn't you want to support the mouse? most terminals do it, even linux fbcon can support the mouse 20:14 FIQ: So does NH4. I disable it, because it causes problems. 20:14 I don't think it's wise to force people to use it, but I don't see why you wouldn't support it as least as an option 20:15 you can still do things like selecting text simply by holding down shift 20:15 ais523: the login form for NH4 should come with pallet changing and mouse disabled 20:15 and yes, NH4 has an option to disable use of the mouse because some people dislike it 20:15 because it's sufficiently terminal-disruptive 20:15 and useless for the login form regardless 20:15 I can see an argument for that 20:16 Because it's too late to NOT change the palette after you read the options. 20:16 yes 20:16 IF you already changed it. 20:17 Though I do wish Konsole had a "don't let applications change this palette" checkbox in the profile options. 20:17 this is just a result of terminals with buggy palette reset codes :-( 20:17 and those who find mouse annoying will already have to deal with its issues in login form, for no benefit 20:17 but yes 20:17 jonadab: incidentally, I actually partially fixed this by contacting a terminal maintainer and convincing them to fix their palette reset code 20:17 jonadab: if terminals had better console parsing in general, it should probably be possible to toggle some of the options 20:17 since some can be used actively maliciously 20:17 FIQ: Mouse is annoying even when somebody _else_ is using it if you watch them e.g. using dgl. Because it can cause moving _your_ mouse to kick you out of dgl and off the server and fill your terminal with garbage. 20:18 FIQ: xterm already lets you compile out many of the more malicious options 20:18 ais523: yeah but it's much more helpful if you don't have to recompile the entire program for stuff like that 20:18 yep 20:19 jonadab: I think by the time dgl watchmode is invoked, mouse codes has already been sent? 20:19 and thus has no effect on the viewer 20:19 not sure though 20:19 (unless the host changes the option mid-viewing, obviously) 20:19 incidentally, I found a security bug in xterm during all this: https://bugs.launchpad.net/ubuntu/+source/xterm/+bug/1629587 20:20 kept it secret for months (except from Ubuntu's security team and xterm's maintainer), but it's public now 20:20 ais523: something I've considered, for dgl purposes, is to look into if it would be possible to, when "viewing" NH4 games, invoke the game with the host save on watchmode 20:20 FIQ: This has happened to me on hdf, when watching other players playing NH4 variants. If they didn't bother to turn off mouse support in their options (WHY is off not the default?), which happens a lot during junethack because they're new to the variant and haven't set any options, MY terminal gets messed up if I watch them. 20:20 you can use it to keep files secret from xterm users by making it enter an infinite loop whenever it tries to read them 20:20 because the NH4 watchmode is much better than dgl's 20:20 jonadab: mouselook is something I use all the time when I play 20:21 it's so much faster than ;-farlook 20:21 just hover something to see what it is 20:21 this would also allow you to use your own interface configs 20:21 ... taking your hand off the keyboard is faster? 20:21 which solves issues like jonadab's 20:21 jonadab: when you're looking at a number of things, yes 20:21 and it solves the silliness that is having to press s a bunch of times to get the codepoints right 20:21 FIQ: I think dgl should at least strip "how to input" codes out of its watch mode 20:21 for people who have vs people who lack unicode support 20:21 configuring things like mouse input for watchers is obviously wrong 20:22 ais523: Possibly. With that said, I want a real NH4 process when watching regardless 20:22 it's so much nicer, but too bad it's only used on N4O where hardly anyone plays D: 20:23 FIQ: in libuncursed2 (actually also libuncursed1 but I never implemented) I'm planning to have libuncursed write a ttyrec as an option, but formatted for a generic terminal rather than the one that's actually in use 20:23 as for myself, turning off mouse is the first thing I do when configuring a new NH4 setup 20:23 I wasn't aware that mouselook was so unpopular 20:23 and I never use it 20:23 ais523: rumflump, the one who initially "recruited" a group of discord people to play NH4+variants, loves it 20:23 so you're not alone :P 20:24 ais523: The problem is that having mouse enabled causes moving the mouse at the wrong time to make it start sending tons of garbage characters to your computer every time the mouse moves. 20:24 This is bad. 20:24 jonadab: on my current laptop, incidentally, the trackpad is easier to use than some keyboard keys are 20:24 `, for instance, is harder to press than moving the mouse is 20:24 This sounds like a REALLY bad keyboard. 20:24 also, that's actually a shell bug, technically speaking 20:24 Because trackpads are one of the worst input devices ever designed. 20:25 ais523: Like, in bash? 20:25 right, these input sequences are specially designed to be differentiable from normal keyboard input 20:25 and a good shell should swallow them 20:25 I see. 20:25 bash is widely regarded as a pretty decent shell 20:25 ais523: NH4 itself has issues with mouse input 20:26 I don't remember details, but rumflump mentioned weird bugs with inputs in inventory windows 20:26 that turned out to originate from having mouse enabled 20:26 jonadab: if you send bash a cursor position report, for example, it just echoes it 20:26 terminals are allowed to say "the cursor is here" unsolicited if they want to 20:26 I see. 20:26 I guess it's just that no modern terminals do, because it would break bash 20:26 jonadab: I use mouse-mode in weechat on my OP to jump to channels faster 20:27 because it's touch 20:27 so the screen is much more easily reachable than my mouse generally is 20:27 on the desktop 20:28 jonadab: so I guess the fundamental problem here is that out of all the things that terminals are allowed to send unsolicited, mouse movement is one of the few that can actually be spammy in practice, just due to its nature 20:29 libuncursed does turn the mouse off when it exits, but it doesn't get a chance to do so after, say, sigkill 20:29 ais523: quitting from dgl watchmode also doesn't turn it off 20:29 so if you watch someone with mouse enabled, gg 20:29 that's because dgl watchmode is broken :-P 20:29 is there a simple fix? 20:30 I mean, it does already mess with the tty stream in watching 20:30 filter input configuration codes out of the dgl output stream 20:30 could be done serverside or clientside 20:30 so if it's just some code it has to ignore 20:30 it seems simple 20:30 right 20:32 I'd argue that most of CSI … h and CSI … l should be disabled in dgl watch output, that's all terminal configuration stuff 20:32 hm 20:33 you could make a case for enabling cursor show/hide (CSI ? 25 h/l), although there's an alternative code for those (CSI … SPACE q) 20:33 not sure which programs use in practice 20:33 well if NH4 uses CSI ? 25 h/l 20:33 s/NH4/NH/ 20:33 then it should remain enabled 20:33 FIQ: NH doesn't use any specific codes, it asks termcap/terminfo 20:34 hm 20:34 so I want to make it possible to send a save to nethack commandline-wise to have it watch it 20:34 $ tput civis | cat -v 20:34 ^[[?25l 20:34 seems simple enough 20:35 I guess NetHack would use CSI ? 25 l to hide the cursor at least on my configuration (with my terminal, $TERM, terminfo database…) 20:35 ais523: was there a specific reason you removed this btw? "REPLAY action 0/32232" 20:36 although, does NetHack ever hide the cursor at all? 20:36 FIQ: I don't believe the NH4 save format is aware of the action number 20:36 other than by scanning the entire file counting the actions 20:36 (which is what NitroHack does) 20:36 yeah that's not exactly feasible 20:36 hm 20:36 I guess it is then 20:36 because it takes <1s 20:37 oh that reminds me 20:37 I forgot to send a patch to K2 disabling the performance test option in dynahack 20:37 ais523: I don't remember if it does or doesn't, but I remember something coming up somewhere sometime about blind players, and NetHack doing something better than other roguelikes that had to do with the cursor. 20:38 The details are fuzzy in my mind. 20:38 ais523: btw, you mentioned nitrohack consuming a lot of memory due to caching 20:38 This is presumably "a lot" of memory by NetHack standards. 20:38 it caches every 1000 turns 20:39 when I changed this to every 1 turn to test memory footprint 20:39 it still didn't consume that much in this 34629-move game 20:39 try it on a pudding farm game :-D 20:39 s/turn/action/ 20:39 hmm yeah I could see that taking more 20:39 if it had >20mil actions 20:40 jonadab: on the subject of "control sequences that make the terminal unusable": ESC ` disables "manual input" entirely 20:40 I mean, I think someone was complaining that aimake of all things uses "a lot of memory", and it barely even uses a gigabyte, IME. 20:40 I assume that that'd make it not respond to any sort of keypress 20:40 ais523: Eep. 20:40 jonadab: it took a lot of optimization to get it under 1GB 20:41 also I should try disabling that checkpointing thing alltogether in nitrohack context 20:41 ais523: Ah, so old versions used more? 20:41 yes 20:41 see how speedy (or not) replaymode becomes 20:41 about that point I decided that I'd probably need to rewrite the memory-intensive bits in another language 20:41 since it'd have to play from turn 1 all the time 20:42 also my laptop at the time only had 3GB physical memory, so I was sort-of running close to the limit 20:43 . 20:43 hi 20:43 Hello K2, Welcome to #hardfought 20:44 when i were a lad, we never had no gigabytes 20:44 \o 20:44 Tangles: Message from Grasshopper at 2017-09-03 11:19 EDT: how about a !lot command to take the pressure off my luck? 20:44 all we had were kilobytes and we were happy to have that 20:44 ok, so it takes 1s to go backwards 1 turn on action 33433 with checkpointing "disabled" (I set it to do so every 100k actions) 20:44 640KiB should be enough for anyone? :-D 20:44 Ah those were the days. 20:44 still somewhat fast, allthough noticeably slower, of course 20:45 you had 640kb? boy you were lucky 20:45 even vanilla NetHack takes more than that nowadays 20:45 we had to work with 8kb and we were lucky to get that 20:45 hothraxxa: i remember upgrading my amd486 dx-100 system from 16 to 24MB of ram 20:45 hothraxxa: I once programmed a processor with 42 bytes of memory 20:45 cost me 300 usd to do it 20:45 obviously, this wasn't intended for general-purpose use 20:46 i've told the story before about upgrading our s360 model 40 from 256k to 384k in 1969 20:46 hah 20:46 it cost $1000000 20:46 ais523: so yeah, if replaying from commands in NH4 is comparable, it's more than 2 order of magnitudes faster 20:46 our first home computer had 64k 20:46 that's a million 1969 dollars 20:46 trs-80 20:46 hothraxxa: Yes, my first computer that I owned had a 20MB hard drive. All the computers I saw as a kid didn't have hard drives, only floppy diskettes. 20:46 but thats still 11 yrs ahead of your time hothraxxa 20:47 so I should really try to look into forcing commands and test stuff 20:47 FIQ: would be interesting to know what it's doing differently (other than running from the optimized-for-386s NetHack internals rather than the clearer-but-slower NH4 internals) 20:47 hi Tangles! lol 20:47 it was an rpq from ibm and took over 6 months to k nit 20:47 hothraxxa: was this back in the days of magnetic core memory? 20:47 rpq meaning it was made special order 20:48 for the record - i love seeing these discussions here. all of them 20:48 aye. 20:48 ais523: The fact that it takes just a second to replay the entire gamestate for 30k+ actions is remarkable 20:48 hothraxxa is immortal 20:48 But that's what you get with a game optimized to run on toasters :P 20:48 his true name is connor mcloed of clan mcleod 20:48 I can see how it'd cost a million dollars to add another 128 KiB if you have to assemble each individual bit by hand 20:49 ais523: Cores? You had cores? Back in the day people were happy to have drum memory... 20:49 actually, that seems pretty cheap… 20:49 ais523: not only that, the architecture for the model 40 did not allow for expansion 20:50 * jonadab doesn't actually know what kind of RAM computers used before drum memory. Something more expensive probably. 20:50 i once generated an os/360 system that ran multi-tasking and spooled output on a 32kb system 20:50 my brilliant and totally not evil bilious patch for 3.6.x hasn't been approved yet :( 20:51 jonadab: delay line, I think? 20:51 basically you have a long, thin pool of mercury 20:51 and store data as vibrations in it 20:51 basically, place a wave for a 1 bit, electromechanically 20:52 Wow. Would that even WORK? 20:52 when the wave reaches the other end, you re-introduce it back at the starrt 20:52 Ah. 20:52 I see. 20:52 it didn't have a huge capacity, and it was as slow as drum memory to access 20:52 So yeah, it would work, but. 20:52 but it did work 20:53 you don't have to worry about the reflections as they'd be too faint to measure after bouncing off the ends 3 times 20:53 Well, it would work, until you get seismic activity in your area, or a truck driving by on the street. 20:53 I have a book on how computers work, which was written when magnetic core memory had been only recently invented and was the new state of the art 20:54 it contains a fairly large portion on analog computers; both analog and digital computers were used at the time, digital hadn't "won" yet 20:54 they're blowing up the bay in a couple of minutes. i'm going to take pictures 20:56 Wait, they're blowing up the Bay Area? Where's dtype gonna live? 20:56 hothraxxa: it looks like you said "they're blowing up the bay..." 20:56 -!- zombifier has joined #hardfought 20:56 drum memory 20:56 every time I see that mention I think of the story of mel 20:56 *mentioned 20:57 Heh. Yes, it features in that story. 20:57 In fact its technical details have plot significance. 20:57 yeah 20:57 heh 20:58 I come back and see a very long discussion going back 3 hours. 20:58 aosdict: it went offtopic quite a bit 20:58 That's also one of the more interesting tech stories I've encountered. Right up there with the 500 mile email. 20:58 was any of this about nethack design or can I ignore it? 20:59 none of it is related to gameplay 20:59 aosdict: Do terminal codes count as NetHack design? 20:59 some of it was nethack related 20:59 but not gameplay 20:59 some amounts of it are related to very low-level details of how NetHack communicates with the terminal 20:59 but most NetHack designers don't care about that 20:59 yeah that's what it seems like 20:59 and half the rest let me handle the details for them 21:00 ooh speaking of nethack design 21:00 ais523: did you see #H5955 and pat's reply to it? what do you think? 21:01 there's some interesting ideas in there IMO 21:01 dont forget we have this - https://www.hardfought.org/irclog/ 21:01 in case you dosconnect and miss stuff 21:01 * jonadab concurs, there are interesting ideas in that. 21:01 *disconnect 21:01 FIQ: it's interesting but not something I've thought of much 21:01 *thought about 21:02 ais523: It's relevant, IMO, to the fact that most NetHack players consider #twoweapon strictly superior to two-handed weapons and shields, for all roles that can do it at all, even at basic skill. 21:02 some of his ideas is some I've thought about myself, some I disagree with, some I think is much better than similar ideas I've seen elsewhere 21:02 also, I see every email sent to devteam@nethack.org (except those deleted by the spam filter), and every bug report sent via the website (again except those deleted by the spam filter) 21:02 including my own 21:02 I don't necessarily see replies but I'm normally copied in unless there's a good reason not to 21:03 jonadab: I'm less down on weapon+shield than most people are 21:03 the basic problem here is that shields are only a negligible part of your overall AC/Def (any single item is), whereas a second weapon is doubling your damage output 21:03 FIQ: is that the one you have linked in on your devteam wiki page? 21:03 ais523: generally people's argument is, twoweaponing allows you to kill things twice as faste 21:03 *fast 21:03 while only providing a small AC penalty 21:03 FIQ: yes 21:03 when factoring enchantment 21:03 aosdict: yes 21:04 shields can have better extrinsic than offhand weapons (notably, reflection) 21:04 which is their only redeeming feature atm 21:04 I am working on a proposal to address this 21:04 aosdict: I intend to try to log my devteam reports and interesting replies there from now on 21:04 hackedhead thinks shields are a good value in Fourk, for roles that can get expert in shields. 21:05 I actually think NetHack could do with better shields, so that offhand weapons have more competition 21:05 effectively, shields do not contribute to AC but are very effective at reducing hit chance 21:05 But I think aosdict's Armor/EV split notion is even better for shields. 21:05 FIQ: right; we can't really blanket-make emails to the devteam public as the original sender might not want us to 21:05 although, hmm, perhaps there should be a checkbox on the bug report form "I'm happy for this report to be made public" 21:05 also, two-handed weapons _ought_ to be the best for sheer damage, with the benefit of twoweapon being versatility (like if you want a saber for silver damage) 21:05 yes, please 21:06 a public bugtracker would be so much more useful 21:06 for everyone 21:06 Bugzilla, please. 21:06 that said, even minor changes to the bug report form can take a while 21:06 Not Trac or some other horrid abomination. 21:06 trac isn't bad, from a functional POV 21:06 NH4 devteam ran into weird issues with it 21:06 but afaik nobody outside NH4 managed to reproduce it 21:06 FIQ: Trac is what you get when people who haven't used Bugzilla design a bug tracking system. 21:07 trac is great UI-wise, IMO 21:07 but it does weird things to servers 21:07 yes 21:07 everyone who tried to host it ended up giving up 21:07 I don't mind the UI at all 21:07 it worked fine 21:07 as an user 21:07 but apparently it caused great pain to server admins 21:07 mediawiki's like that too 21:09 hm 21:12 ais523: how can I call a random number without affecting the gamestate? 21:12 there is a display RNG, right 21:12 (I want to test something) 21:12 FIQ: in the NH4 engine? use rn2_on_rng() and give it the display RNG as the parameter 21:12 I forget what the constant for it is called 21:12 might be just rng_display 21:12 rng_display apparently 21:13 thanks 21:13 [gh] k2 (Arc Dwa Mal Law), 581 points, T:1903, killed by an iguana 21:23 well they blowed her up real good 21:29 \o/ 21:29 KABLAAAM! 21:34 ok so this save can play all the way up to turn 10.2k without a desync using commands only 21:34 it will do nicely for testing purposes 21:36 ...once I can figure out how to kill these mores/tabs in replaymode 21:38 whoops, I have a feeling that's my fault somehow :-( 21:39 also I think nitrohack used a more minimal run of the gamestate when replaying to avoid dealing with animations and whatnot 21:39 I should look into if I can replicate that 21:39 for now I just disabled animations 21:42 -!- NeroOneTrueKing_ has joined #hardfought 21:45 -!- NeroOneTrueKing has quit [Ping timeout: 240 seconds] 21:46 [gh] k2 (Arc Dwa Mal Law), 1149 points, T:8201, killed by an iguana, while fainted 21:46 ok so it takes 33 seconds to playback turn 1->10405 using commands only in NH4 21:46 hm I should look into the callgrind thing 21:47 for this test 21:47 to see what takes time 21:47 it's worth mentioning that allbeit this is slower than what it takes to get to the final turn normally, this is done using binary save checkpoints, which I am not using at all for this test 21:51 ais523: how does a binary checkpoint look like in the save? 21:51 * as the first character on the line 21:51 the save I am using take atrociously long time to load which is rather annoying so I wanted to truncate the end 21:51 thanks 21:51 although it'll only be used in save file loading if the binary checkpoint index is up to date 21:52 wow 21:52 no wonder it took forever 21:52 there are numbers after the *s, each is the index of another checkpoint, connected round in a loop 21:52 for some reason it has only made 5 checkpoints? 21:52 first one links to last, each other one links to the one before 21:52 this is a very long game... 21:52 and not checkpointing enough is a common NH4 issue 21:52 I think what happened is that the algo used to decide when to checkpoint ended up being sensitive to compression ratio 21:52 *003ba407 $2159461$eNrsvQtg(...) 21:52 https://imgur.com/a/AQdVn 21:52 and the better the compression got, the fewer checkpoints were made 21:52 it looks like this? 21:52 yep 21:52 thanks 21:53 it made 9 checkpoints in total in that game 21:53 number after the * is the index to find the previous/last checkpoint, number between $s is uncompressed size, the rest is just compressed data 21:53 it's 43599 turns... 21:54 you might want to tweak the when-to-checkpoint settings 21:54 well I guess that is another room for optimization 21:54 checkpointing more often than "basically never" :P 21:54 Those are some nice photos hothraxxa! 21:55 ok so truncating the end 21:55 it took me to turn 36800 21:55 thanks tangles 21:55 so there was no checkpoint for almost 10k turns 21:55 no wonder it took forever to load 21:57 FIQ: if you truncate away the last checkpoint 21:57 then the index will be broken and it won't use any of them 21:57 I truncated everything after the checkpoint 21:57 hothraxxa: #'s 4 and 6 are my favs 21:57 ah, OK 21:57 ==24960== brk segment overflow in thread #1: can't grow to 0x4 21:57 Why did this happen when I activated replaymode in callgrind? 21:58 hothraxxa: why foreworks on sept 3rd? is that a canadian holiday? 21:58 *fireworks 21:58 no this is a city festival 21:58 illuminate barrie 21:58 also I just realized that callgrinding something that takes 33s *outside* was probably a bad idea 21:58 huh 21:58 should probably have cut it down somewhat 21:58 just cuz 21:58 they were set to music if you tuned in the right station 21:58 that stuff isnt cheap 21:59 hothraxxa: Ah, are you in Barrie? I've been there. 21:59 nope. this is a festival town and that's the summer end festival 21:59 FIQ: that's a bizarre error message, the most obvious cause is an attempt to malloc enough memory that the address space wrapped, but mallocs that large don't go through brk on glibc 21:59 jonadab: recently? 21:59 neat 21:59 hothraxxa: Umm, June 2016, I believe. 22:00 so close 22:00 so the waterfront would have been about half done 22:00 Stayed in Barie rather than Toronto proper for affordability reasons. 22:00 I didn't get up to the Barie waterfront. 22:00 heh. that's a long commute though 22:00 Well, we were driving in from Owen Sound. 22:01 After taking the ferrie from South Baymouth. 22:01 oh i see, a waystop 22:01 *ferry 22:01 This was on the trip where I saw all five of the great lakes. 22:01 cool! 22:01 Two of them for the first time. 22:02 And Ontario for the first time from the ground (I'd previously only seen it from the air). 22:02 my best count is four. didn't get lake michigan 22:02 I'd seen lake Michigan before, both from western Michigan and also from downtown Chicago. 22:02 i mean i've seen lake michigan, but not on the trip where i saw the four others 22:02 Ah. 22:03 Yeah, we crossed the bridge at Mackinack, so we saw Lake Michigan there. 22:03 it was a motorcycle trip back in in 1981 22:03 We only saw one little corner of a bay of Lake Superior, but enough to count it :-) 22:03 i rode all along the north shore of lake superior. amazing scenery 22:04 But we saw all kinds of Lake Huron on that trip, and that was one of the two I'd not seen before (Superior being the other). 22:04 I mean, we spent three nights on Manitoulin Island. 22:04 And visited Mackinack Island also. 22:04 And took the ferry across to the Bruce Penninsula. 22:04 So yeah. 22:04 Lots of Lake Huron. 22:05 Lots of _fog_ on Lake Huron. 22:05 ais523: it consistently happens when I invoke replaymode 22:05 not sure what's going on 22:05 Lake Erie, of course, I've seen scads of times, because I've lived in Ohio for decades. 22:05 gnite fols. dbl shift tomorrow bleh 22:06 K2: Sounds like "fun". 22:06 uh huh 22:06 FIQ: neither am I 22:06 try regular valgrind to see if it knows what's going on? 22:08 843,648,851 /home/fiq/fiqhack/libuncursed/src/libuncursed.c:copywin [/home/fiq/fiqhackdir/fiqhack] 22:08 403,992,000 /home/fiq/fiqhack/libnethack/src/memfile.c:mwrite [/home/fiq/fiqhackdir/fiqhack] 22:08 305,459,964 /home/fiq/fiqhack/libuncursed/src/libuncursed.c:wclrtobot [/home/fiq/fiqhackdir/fiqhack] 22:08 so apparently the windowport is the biggest offender in replaymode like this 22:08 which I suppose makes sense 22:09 this is in a playback of 500 turns, because when playing back just 50, it blamed mread/mwrite which made no sense because those aren't really used 22:11 not sure if it's even feasible to dismiss the windowport during playback 22:11 hmm 22:11 [gh] Tangles (Val Hum Fem Law) had silver dagger bestowed upon her by Tyr, on T:1182 22:12 Oh that'll go nicely against all the were-things. 22:12 ais523: valgrind didn't have anything to say when doing the same process: http://home.fiq.se/valgrind.log 22:13 FIQ: that's even more bizarre, as callgrind and valgrind use the same engine 22:13 IIRC nitrohack uses a dummy windowport during playback 22:15 ais523: just looked, seems you're right 22:16 static struct nh_window_procs replay_windowprocs, orig_windowprocs; 22:16 with a bunch of commands 22:16 some dummy ones, etc 22:26 [gh] Tangles (Val Hum Fem Law), 1716 points, T:1903, killed by a winter wolf 22:53 K2: found your arc bones (killed by an iron piercer) twice. 22:53 Well techically they were my bones the second time. 22:54 You had a fair bit of loot for that early in the game. 23:34 -!- NeroOneTrueKing has joined #hardfought 23:36 -!- NeroOneTrueKing_ has quit [Ping timeout: 240 seconds] 23:44 polytraps in minetown... grunthack really is the variant that beats you up and takes your lunch money. 23:55 Tangles: but minetown can have polytraps 23:55 in vanilla too 23:55 it's uncommon 23:55 but can happen 23:56 EPI: an actual patch that turns variants of nethack into slashem extended 23:56 diff