00:03 [hdf-us] [dyn] Jane (Blue) (Sam Hum Fem Law), 1419 points, T:1239, killed by a rothe, while frozen by a monster's gaze 00:10 -!- Grasshopper has joined #hardfought 00:10 -!- mode/#hardfought [+v Grasshopper] by ChanServ 00:17 [hdf-us] [xnh] Luxidream (Ran Orc Mal Cha) became literate by reading a scroll, on T:9627 00:22 -!- Grasshopper has quit [Ping timeout: 260 seconds] 00:31 [hdf-us] [nd] redshirt (Rog Orc Fem Cha), 20592 points, T:15003, killed by a giant zombie 00:36 -!- raisse has joined #hardfought 00:36 -!- mode/#hardfought [+v raisse] by ChanServ 00:41 -!- Menche has joined #hardfought 01:06 -!- Grasshopper has joined #hardfought 01:06 -!- mode/#hardfought [+v Grasshopper] by ChanServ 01:18 -!- bug_sniper_ has quit [Ping timeout: 264 seconds] 01:30 -!- Tarmunora has joined #hardfought 01:32 -!- Tarmunora_ has quit [Ping timeout: 264 seconds] 01:37 [hdf-us] [nd] respite (Ran Elf Fem Cha) chose an elven dagger to be named "Sting", on T:1 01:41 -!- bug_sniper_ has joined #hardfought 01:49 -!- LarienTelrunya has joined #hardfought 01:49 . 01:49 LarienTelrunya: Message from K2 at 2018-04-08 16:45 EDT: I didnt update slex on the eu server yet, for hopefully obvious reasons. so its just the us server that we're dealing with at the moment 01:52 !tell K2 anything enlightening on the recover problem? I don't really see why the new version wouldn't work, since the only change I made is that it should correctly handle levels with index numbers >256 now; no clue why that would make it stop working 01:52 Will do, LarienTelrunya! 01:58 !tell K2 just tested on hdf-us: autorecover works fine for me 01:58 Will do, LarienTelrunya! 02:01 !tell K2 just fwiw: I did test that recover change locally, it worked fine on windows, and then I tested it on the esm server where it also worked fine, so its not working on hdf came as a surprise to me 02:01 Will do, LarienTelrunya! 02:02 !tell K2 but I guess you're right, I'm planning to push the savebreak version soon, maybe this week or next week, and then declare it save-stable and focus only on bug fixes until junethack 02:02 Will do, LarienTelrunya! 02:15 -!- raisse has quit [Ping timeout: 264 seconds] 02:19 -!- raisse has joined #hardfought 02:19 -!- mode/#hardfought [+v raisse] by ChanServ 02:39 -!- raisse has quit [Ping timeout: 260 seconds] 02:56 [hdf-us] [dyn] Matti (Robin) (Mon Hum Mal Neu), 0 points, T:86, killed by kicking a wall 03:03 [hdf-us] [nh4] Matti (Robin) (Rog Hum Mal Cha), 9213 points, T:671, quit 03:06 -!- raisse has joined #hardfought 03:06 -!- mode/#hardfought [+v raisse] by ChanServ 03:25 -!- deadnoob_ has joined #hardfought 03:28 -!- tubs has joined #hardfought 03:29 -!- deadnoob has quit [Ping timeout: 245 seconds] 03:33 -!- tubs has quit [Ping timeout: 260 seconds] 03:50 -!- noty has quit [Quit: WeeChat 1.9.1] 03:54 -!- raisse has quit [Ping timeout: 255 seconds] 03:59 yasi: in the spirit of nethack, slex's rings of discount action should raise your chr 04:02 the ring of discount action was inspired by https://nhqdb.alt.org/?1854 by the way 04:04 scroll of literacy 04:05 !who 04:05 FIQ: [hdf-us] No current players 04:05 FIQ: [hdf-eu] No current players 04:05 ok 04:06 FIQ: you said something about slex's recover problem when I wasn't here? 04:07 -!- tubs has joined #hardfought 04:12 so if you're hungry for a hot dog but can't eat it because you're out of ketchup, you should eat a warm dog? 04:12 -!- tubs has quit [Ping timeout: 276 seconds] 04:14 -!- tubs has joined #hardfought 04:14 lol :D 04:14 good night 04:16 LarienTelrunya: Yes, I stated that it needs to be fixed 04:16 broken recover is what made kerio stop hosting grunthack 04:17 I also told K2 to try to tell you to fix your shit so that people don't have to pray that their games don't crash 04:17 or lose saves/progress 04:17 ... but I tested the autorecover function on hdf-us just a few hours ago and it worked fine for me 04:18 in fact, the autorecover function was also working fine for Tariru, the problem was that the game itself crashed again upon restoring 04:18 Did you test on a game the size of Tariru? 04:18 Well the crash implied that level files were missing 04:18 Meaning it didn't properly recover at all 04:18 really? what was the message again? 04:18 There was a stack trace 04:18 It was trying to read a NULL level 04:18 -!- bug_sniper_ has quit [Ping timeout: 240 seconds] 04:19 WTF 04:19 you mean the one at http://termbin.com/1g23 ? 04:19 one sec 04:20 -!- tubs has quit [Ping timeout: 260 seconds] 04:20 [hdf-us] [nd] Grasshopper (Sam Hum Mal Law) made his first wish - "blessed fixed magic mirror of merlin", on T:10876 04:20 -!- tubs has joined #hardfought 04:23 Your code is impossible to reaad 04:24 But yes I meant that one 04:24 *read 04:24 huh? where does that read a NULL level? 04:25 #4 04:26 Caveat: I don't know the NH3 save system very well 04:26 The way it is called there might be correct for all I know 04:26 But it surely doesn't look right to me 04:27 well, the thing is, recover.c originally assumed that all level index numbers are smaller than 256, that's why I patched it in the first place, and I have no idea why that somehow made it stop working on hdf 04:27 in fact, this is the change I made: https://github.com/SLASHEM-Extended/SLASHEM-Extended/commit/184ffb50bf49c0995989752495101411d2be545a 04:27 Also I suggest you don't make ridiculously long lines 04:28 It makes code much easier to debug 04:28 *Anything* in that line might be done wrong in this context 04:28 yes, I noticed that too, and split that line into many lines in the next commit 04:28 But what? Who knows! There's like 50 statements 04:28 sorry 04:28 Slex code looks like that all over the place 04:28 Not my problem if your code is impossible to debug 04:29 hey, I'm trying to make it better! relax! 04:29 I will however reiterate what I said months ago, before it had bitten you in the ass -- long lines like that make for bad code 04:30 -!- tubs has quit [Ping timeout: 268 seconds] 04:30 LarienTelrunya: oh so you replaced one dubious assumption with another 04:30 will you remember that 5k level limit? 04:30 also that comment was wrong even before your edit 04:31 how would you do it? even if that amount of levels doesn't exist, it will still go through all of them, so setting it to one million or something would make it take forever 04:31 an xchar can only hold 0-127 04:31 LarienTelrunya: I would define a macro somewhere, MAXLEVELS, MAX_LEDGER_NO or whatever 04:31 for the entire source tree 04:32 that way if you do try to exceed the limit, the game whines rather than silentily breaking recover 04:32 -!- raisse has joined #hardfought 04:32 -!- mode/#hardfought [+v raisse] by ChanServ 04:32 LarienTelrunya: either that or fix recover's file reader completely, but that's more involved 04:32 yeah but that requires having a way for the game to figure out how many levels there are 04:33 I am pretty sure the game knows how many levels the game will have the moment you start playing 04:33 it needs to 04:35 okay, just tested on my ubuntu VM with a wizard mode savegame that has level files of the yendorian tower (index 277+) - produced a segfault, let the game use autorecover, works 04:35 now I'll try manual recovery 04:35 LarienTelrunya: there is other 256 assumptions in that file 04:35 for instance, lock 04:36 hmm 04:36 actually no, that's a string 04:36 a string with 256 rather than BUFSZ size for whatever reason 04:36 wtf I can't find the recover binary?! 04:36 should be fine 04:37 does your local build of slex have a recover binary? mine somehow doesn't 04:37 I don't know 04:37 My local slex build is ancient 04:37 and I don't remember where it is 04:37 ah 04:38 I only ever touched it for curses you know 04:38 after Elronnd complained I did something weird 04:38 turns out it was a slash'em-specific issue 04:39 well, thing is, for me the recovery works alright, so I can't really tell why it failed for Tariru; of course it may be that I fucked it up somehow, but it might also be an unrelated other bug I introduced somewhere that caused it to fail only in this particular case 04:39 Recover effectively dials back time until last save or level change 04:39 -!- tubs has joined #hardfought 04:39 [Demo] loves this feature 04:39 but besides that 04:40 it means that the load crash is at that point, not the same point as the original crash 04:40 Tariru's original crash happened while he tried to load his save normally, though 04:40 i.e. he saved the game, tried to load it back up and bam segfault 04:40 I don't know what is wrong here, and I really don't feel like trying to digest slex code to find out. All I know is that the obvious reason seem like one of 2 things 04:41 1: your stuff is doing incorreect assumptions regarding load. Try to save the game with a were-whatever and restore after recover. Also try it with PFSC 04:42 Both with the were in and out of view 04:42 2: The recover utility is somehow broken. Who knows what is wrong 04:42 yeah, I suspect a weresomething to be the culprit 04:42 If what you say is true, then it's almost certainly #1 04:43 Careful here -- if the game crashes *on load*, that is the one case where the game can generally not recover 04:43 the level Tariru was on was the one completely filled with monsters, and I guess at 49,8 there's a bugged werecreature 04:43 Crashing in the load/save process is NH3's weak point regarding save integrity 04:43 if only I could get the save to load on my VM, I could probably diagnose it better 04:44 Why can't you get the save to load? 04:44 but for that I need to find out the exact configuration K2 uses for hdf, so I don't get "configuration incompatibility" upon trying ot load it 04:44 If you mean you're simply just getting the same crash 04:44 run it in gdb 04:44 or valgrind 04:44 hdf-slex is probably compiled with some compile flags that I didn't set locally 04:44 Debug symbols 04:44 oh, those cause configuration incompatibility? 04:44 Oh, no 04:45 i.e. a save with debug symbols won't load in a version compiled without them? 04:45 it will 04:45 Something else then. Ask K2 about it 04:45 yeah 04:46 meanwhile I'll test shapechangers a bit 04:46 Can you just remove the offending monster from the save file? 04:46 tubs: NH3's save system is rather rudimentary 04:46 It is a very nontrivial process to do 04:47 Even NitroHack saves are easier to recover 04:47 NH3's system is just a raw memory dump 04:47 with pointers changed to something that persists across runs 04:48 Monsters are in a ll? 04:48 ll? Linked list? 04:48 Yes 04:48 No way to snip it out using a hex-editor? 04:49 I'm not saying what you're asking is impossible 04:49 It isn't 04:49 It's just a nontrivial process 04:49 so 04:49 You need to find the offending monster 04:49 It's the type of thing I am forced to do day-to-day for work :-( 04:49 And make sure everything adds up 04:49 I've experimented a bit with werecreatures but they don't seem to crash for me 04:49 tubs: Sounds like fun 04:49 FIQ: it actually is! Super satisfying when you get stuff working. 04:50 tubs: I mean, you need to figure out where monsters even are in the save file 04:50 if only I could load Tariru's save, I could then experiment all day long to find out where exactly it crashes 04:50 Also how your processor dumps structs (this is generally nonportable) 04:50 FIQ: Ah, I thought LarienTelrunya knew roughly where/which creature it was 04:51 LarienTelrunya: stupid question but you're definitely the same 32/64 bit-ness? 04:51 not sure; my VM is 32-bit ubuntu 04:51 tubs: LarienTelrunya said there was a configuration incompatibility 04:52 Which implies that the game at least tried to read the save 04:52 So I would assume so 04:52 Just a guess though... 04:52 -!- raisse has quit [Ping timeout: 264 seconds] 04:52 Actually I thought hdf was on 64bit.. 04:52 I'd assume hdf was 64bit! 04:53 LarienTelrunya: see 04:54 gah, so it won't work at all? :( 04:54 this is why I and others encourage people to use NH4's save system :-) 04:54 Ubuntu 16.04.4 LTS 4.14.4-041404-generic doesn't tell us much alas 04:54 FIQ: that was one of my happiest moments reading the nh3->nh4 changelist :) 04:54 tubs: picture these 2 scenarios 04:54 LarienTelrunya: it should be easy enough to set up a 64bit VM no? 04:55 #1: You are a NH3 variant developer. Someone finds a bug. Now you need to try to reproduce it. However, the save is suddenly at some previous point in time (last level change/save), and reproducing the bug is basically impossible because it was something with a particular monster that generated 04:56 #2: You are aa NH4 variant developer. Someone finds a bug. You tell the server maintainer to hand you the save, and you fire it up in gdb/valgrind. Now you know what the bug was so you can fix it right away 04:56 Which one is easiest to debug= ;p 04:56 *? 04:56 * tubs wonders if this is a trick question 04:57 Clearly the easiest one is #1 because #2 wasn't even a challenge /s 04:57 -!- deadnoob has joined #hardfought 04:57 :D 04:57 This is one reason as to why my bug fixes tend to be rapid 04:57 ACtually is there any reason you couldn't get slex compiling under WSL? 04:58 someone finds a bug, if I'm around most bugs are fixed within a few minutes 04:58 "funnily", my VM refuses to install gdb :( if I "sudo apt-get install gdb", it gives a bunch of 404 Not Found errors 04:58 tubs: Should work. However as said, NH3 saves are nonportable 04:58 It might create yet another save "type" 04:59 LarienTelrunya: apt-get update 04:59 first 05:00 oh, apparently I need to run those commands as sudo? 05:00 -!- deadnoob_ has quit [Ping timeout: 260 seconds] 05:00 Yes, it can't update its records on package files and stuff without root 05:01 tubs: in my very first foray into linux 05:01 when I was to update my system, the game complained that I wasn't root 05:01 I didn't know about sudo back then 05:02 so I just logged in as root and id it there instead 05:02 with the GUI 05:02 I think I used KDE 05:02 i.e. I logged into KDE as root 05:02 I'm sure this was a great idea 05:03 Err http://de.archive.ubuntu.com/ubuntu/ utopic/main libpython3.4 i386 3.4.2-1 404 Not Found [IP: ] 05:03 that's what happens if I try to install gdb 05:03 I told you to update first 05:03 but I did! still happens! 05:03 try to upgrade the entire system 05:03 and then try again 05:03 apt-get upgrade, that is 05:04 ugh, if I try that I get hundreds of pages of 404's 05:04 if it still doesn't work, ask #ubuntu 05:04 maybe it tries to apt-get stuff from a server that no longer exists 05:05 LarienTelrunya: oh, utopic is 14.10 05:05 I think there's no support for that version anymore 05:05 using LTS is preferred 05:05 for this reason 05:06 so yes, the 14.10 packages are probably gone 05:06 I'd give WSL a shot tbh, it should be identical to an ubuntu system for the purposes of slex 05:07 -!- elenmirie has joined #hardfought 05:07 -!- mode/#hardfought [+v elenmirie] by ChanServ 05:07 LarienTelrunya: Ubuntu 14.10's support was dropped in July 05:07 2015 05:07 wow that is one short support period 05:08 wtf 05:08 oh god I've been sitting on this toilet for so long my legs have died 05:08 now I can;'t get up 05:12 -!- elenmirie_ has joined #hardfought 05:13 -!- elenmirie has quit [Ping timeout: 268 seconds] 05:13 -!- elenmirie_ is now known as elenmirie 05:17 -!- raisse has joined #hardfought 05:17 -!- mode/#hardfought [+v raisse] by ChanServ 05:19 [hdf-us] [nd] Grasshopper (Sam Hum Mal Law) completed Sokoban, on T:12302 05:26 okay I think I've figured it out: etc/apt/sources.list has to be edited to change archive.ubuntu.com to old-releases.ubuntu.com and then it should work 05:27 that will hopefully allow me to install gdb on ubuntu 14.10 :) 05:39 woohoo it works! 05:41 :) 05:43 now I'd only need to find a way to load Tariru's save so I can debug it with gdb 05:43 *grumble grumble* WSL 64bit *grumble* 05:44 well it was enough of a pain in the butt to set up this virtual machine... 05:46 you know, in windows, I can load slex savegames from a 32-bit build (made on Windows XP) in a 64-bit build (made on Windows 7) and in reverse! why wouldn't the same be true for unix? 05:49 Likely the windows compiler makes stronger guarantees 05:49 Wait, you're not on Windows 10? :x 05:50 LarienTelrunya: likely the windows 7 build you have is 32bit 05:51 System type: 64-bit operating system 05:51 LarienTelrunya: I meant the nethack build 05:51 oh 05:52 is there a way to check that? 05:52 Probably 05:52 <[Demo]> file? 05:52 Not ingame though 05:52 [Demo]: LarienTelrunya is on windows 05:55 <[Demo]> upload it to like gcp cloud shell or whatever the free shell that google offers is called then 05:55 Google has a shell? 05:55 huh I guess it does 05:56 <[Demo]> https://console.cloud.google.com/cloudshell/editor?showFTMessage=false&pli=1&shellonly=true log into a google acc and go here, yeah 05:57 <[Demo]> and theres even like a direct file upload button i think 05:57 come to think of it: [Demo]: you have a unix system too on which you're testing your dnethack variant, right? is it 64-bit? 05:57 <[Demo]> yeah 05:58 -!- ProzacElf has quit [Ping timeout: 256 seconds] 05:58 maybe you can help me out with troubleshooting this, by building slex-2.1.7, and then I send you Tariru's save and hopefully it will load for you? 05:59 -!- elenmirie has quit [Ping timeout: 265 seconds] 05:59 <[Demo]> im about to head out but maybe later? 06:00 :) 06:00 important: if you do, please make sure you build from the slex-2.1.7 branch, not the master branch as that has had a savebreak and will not load Tariru's save 06:06 <[Demo]> ok 06:06 <[Demo]> i have a great role idea 06:06 <[Demo]> analchemist 06:07 <[Demo]> noooooooooo 06:07 <[Demo]> crap 06:07 lol :P 06:07 <[Demo]> ok so i hadnt written the name out yet but ur an anarchist alchemist 06:07 <[Demo]> not a chemist with a test tube in its asss 06:07 <[Demo]> im really sad that the name doesnt work 06:09 -!- raisse has quit [Ping timeout: 264 seconds] 06:10 you're an alchemist :D 06:10 <[Demo]> but the idea is the role gets special drops from monsters 06:10 <[Demo]> like newts drop eyes 06:11 <[Demo]> bats drop shit idk 06:11 <[Demo]> and you can make some regular potions but also some other special ones 06:11 <[Demo]> a lot of which cause great harm 06:11 <[Demo]> to shit you throw them at 06:13 FU@ASLKJALSKDJASD 06:14 high potion of disaster? (that exists in Elona and ruins whoever you throw it at :D) 06:14 tubs: why twice 06:14 tubs: was that a typo death? 06:14 caves of qud has something called neutron flux 06:14 No 06:14 Amulet of life saving 06:14 it creates a huge explosion and kills everything in its radius 06:14 I TOLD MYSELF "REMEMBER BLINDFOLD REMEMBER BLINDFOLD REMEMBER BLINDFOLD" 06:14 it is likely to also kill you 06:15 Then as I pressed "<" 06:15 I was like 06:15 oh shit 06:15 please 06:15 tubs: but you had another turn 06:15 what happened 06:15 Not sure, I am sure I couldn't do anything :s 06:15 I bet I've done something really stupid... 06:15 !lastgame tubs 06:15 tubs: [hdf-us] No last game for tubs on this server. 06:15 tubs: [hdf-eu] https://eu.hardfought.org/userdata/t/tubs/dnethack/dumplog/1523172511.dnh.txt 06:15 did you not have reflection? 06:15 <[Demo]> i did that in fiqhack earlier last week lol 06:15 No, I was planning on using a blindfold 06:16 [Demo]: in 2016 06:16 there was a bug 06:16 medusa would stone you only if you *did* use a blindfold 06:16 I didn't get an action, looks like she does her gaze at the start of the turn 06:16 LarienTelrunya: if you need help running something on a unixy system I should be able to 06:17 tubs: oh, right, you also have a local slex install! 64-bit I suppose? 06:17 tubs: LarienTelrunya wants you to run slex 06:17 enjoy 06:18 LarienTelrunya: btw Tariru is at 2/3 06:18 Yeah I built it myself on mac, but if that causes issues I will boot up my Linux machine 06:18 LarienTelrunya: which commit do you want me to build? 06:18 FIQ: yay! hopefully he finds the last staircase soon! 06:19 tubs: better use the linux one to maximize the chances that Tariru's save will load; commit 184ffb50 on the slex-2.1.7 branch (not the master branch since that one isn't save-compatible with the version on hdf) 06:19 yes ma'am 06:20 compile it with PUBLIC_SERVER defined (unfortunately I don't know if K2 uses any other compile flags that aren't defined by default) 06:20 tubs: Tariru's save is online. You may not reveal information he wouldn't have to him 06:20 FIQ: sure! 06:20 I am a good boy 06:20 LarienTelrunya: give me 5mins to set up the Linux machine, I only moved into this flat a couple of weeks ago, so not everything is unpacked! 06:20 FIQ: oh, thanks for the reminder, the same is obviously also valid for me 06:20 ok :) 06:20 <[Demo]> im not even sure k2 would be happy that its going beyond the variant dev 06:21 [Demo]: Yeah I'm not sure 06:21 I'm happy not to, whatever works and isn't naughty 06:21 it's for the greater good, if we manage to squash that stupid bug then everyone will benefit! 06:21 K2 once gave me a dynahack save, but that's because I was the only one at the time who had a remote chance of recovering it 06:21 But yeah, that was also K2 who gave it 06:21 now LarienTelrunya is spreading it 06:21 LarienTelrunya: Perhaps ask K2 first? 06:22 yeah, I guess I should 06:22 I mean I'm sure he wouldn't mind 06:22 But it's a good policy to ask either way 06:22 yes 06:22 or you're straining his trust 06:22 <[Demo]> send it to kerio 06:22 öpö 06:22 lol 06:23 what's a kerio? 06:23 !tell K2 hmm, hdf is 64-bit, right? I unfortunately only have a 32-bit ubuntu system on my virtual machine, so chances are it won't load for me no matter what. Would it be okay for you if I give Tariru's save to tubs, who has a 64-bit unix machine, so he can test whether it loads for him so we can troubleshoot? 06:23 Will do, LarienTelrunya! 06:24 Also wtf? I cooked 1kg of beef, 1.5kg of chicken, tzimmus, parsnips, potatoes, AND a cake last night, and now there is only half a cake left. I was banking of eating the leftovers today :-\ 06:24 <[Demo]> I still think it would just be easiest for Amy to do it on google free cloud shell 06:26 [Demo]: that would work too! 06:27 I'd offer to build the right version and give you ssh access here LarienTelrunya , but then you'd still have to trust me with the save being on my machine so it wouldn't actually help 06:28 yeah, I guess the best course of action would be to wait until K2 reappears 06:29 <[Demo]> nethacks current alchemy system is closer to bartending than alchemy 06:30 I was actually toying with the idea of complicated alchemy 06:31 But the question is does it just add complexity without more fun? 06:32 !tell K2 Do this to minimize DynaHack save load failures after a crash: in topten.c, there is a function claled tt_oname. Add a "return NULL;" on top of it 06:32 Will do, FIQ! 06:33 !tell K2 This will prevent DynaHack saves failing to load past Valley of the Dead in the event of a crash 06:33 Will do, FIQ! 06:39 -!- raisse has joined #hardfought 06:39 -!- mode/#hardfought [+v raisse] by ChanServ 06:41 -!- deadnoob_ has joined #hardfought 06:44 -!- deadnoob has quit [Ping timeout: 245 seconds] 06:56 !tell ais523 Why save bones but not options? In both cases this is a matter of "will we presume a mandatory/optional diff?". I.e. you don't need to save any of them if you just care about diffs, but if you assume diffs can be omitted, you need to save both 06:56 Will do, FIQ! 07:03 morning 07:03 K2: Message from LarienTelrunya at 2018-04-09 01:52 EDT: anything enlightening on the recover problem? I don't really see why the new version wouldn't work, since the only change I made is that it should correctly handle levels with index numbers >256 now; no clue why that would make it stop working 07:03 K2: Message from LarienTelrunya at 2018-04-09 01:58 EDT: just tested on hdf-us: autorecover works fine for me 07:03 K2: Message from LarienTelrunya at 2018-04-09 02:01 EDT: just fwiw: I did test that recover change locally, it worked fine on windows, and then I tested it on the esm server where it also worked fine, so its not working on hdf came as a surprise to me 07:03 K2: Message from LarienTelrunya at 2018-04-09 02:02 EDT: but I guess you're right, I'm planning to push the savebreak version soon, maybe this week or next week, and then declare it save-stable and focus only on bug fixes until junethack 07:03 K2: Message from LarienTelrunya at 2018-04-09 06:23 EDT: hmm, hdf is 64-bit, right? I unfortunately only have a 32-bit ubuntu system on my virtual machine, so chances are it won't load for me no matter what. Would it be okay for you if I give Tariru's save to tubs, who has a 64-bit unix machine, so he can test whether it loads for him so we can troubleshoot? 07:03 K2: Message from FIQ at 2018-04-09 06:32 EDT: Do this to minimize DynaHack save load failures after a crash: in topten.c, there is a function claled tt_oname. Add a "return NULL;" on top of it 07:03 K2: Message from FIQ at 2018-04-09 06:33 EDT: This will prevent DynaHack saves failing to load past Valley of the Dead in the event of a crash 07:04 LarienTelrunya: hdf servers are 64 bit. switch your VM's to use 64 bit versions of ubuntu. its free 07:04 hi K2, and sorry for the Amystorm 07:04 Beholder used Message Storm! It's super effective! 07:04 also, this is what I have noticed with your recover binary for slex - you cant specify a player 07:05 WTF 07:05 you have to point it to the folder (level in this case) where the running game files are 07:05 and then it will try to recover *everything* in that folder 07:05 K2: I encourage you to do what I suggested for Dyna. It should prevent issues like what jonathanhanes had a few months ago 07:05 FIQ I will thanks 07:05 ok 07:06 Sorry 07:06 so i was able to finally manually recover Tariru's game 07:06 urgh... that just sounds so very wrong... why on earth would the recover utility do that?! 07:06 but thank god there werernt other legit games ongoing at the same time 07:06 LarienTelrunya: is it possible that your setup is incorrect 07:06 but when i try to point it to a specific player file, it doesnt understand, its looking for a folder 07:06 this is slex, it is based on slash'em which originated before 3.4.3 07:07 is it possible that its recovery utility is dumber? 07:07 K2: did you do manual recoveries on slex before? and did it work at that time? 07:08 if other legit games was going on at the time of the recovery 07:08 those games would, alongside Tariru's, all have been destroyed 07:08 sounds like fun 07:09 on the esm server, the level files are in the same folder as the recover utility, so I just run "./recover 1003BSOD2" and it works 07:09 * LarienTelrunya is confused 07:09 LarienTelrunya: try to have 2 broken saves at the same time 07:09 LarienTelrunya no, i had never had to do a menual recover on slex before this 07:09 er, crashed 07:09 not broken 07:09 manual 07:09 K2: oh, so it was probably broken all the time :( 07:09 not that i recall 07:10 hmm ok one sec 07:10 so every other variant in existience you use ./recover -d 07:10 not every other variant 07:10 -d? 07:11 * FIQ whistles 07:11 what FIQ 07:11 not NH4 variants 07:11 variants that pertain to the conversation 07:12 sorry for being a pedant 07:12 slashem/slashem/games/slex-2.1.7/slexdir-5871# ./recover Usage: ./recover [ -d directory ] base1 [ base2 ... ] 07:12 I'll leave you 2 alone for now 07:12 -!- MiseryMyra has quit [Quit: two forces in balance \\ ancient endless balance \\ then nothing] 07:12 so I guess if you use -d then that is the "directory switch", so to speak, and tells it "recover everyting in that directory", which can obviously have fatal consequences 07:13 bah 07:13 K2: where are the level files for ongoing slex games stored on hdf-us? I just opened a game for testing 07:13 ok, i see what i did 07:13 gdammit 07:14 they go in the level subdirectory that your game creates 07:14 no wait 07:14 ah, so I guess there's 5BSOD2.1 and such now 07:14 that IS right.... hrm 07:15 yeah thats right 07:15 if you want, I can crash my current hdf-us game on purpose now so you can test whether it can be recovered, probably by using ./recover level/BSOD2 (replace "level" with the actual name of the directory that has the saves) 07:15 so its not me 07:16 but gdamn 07:16 sure 07:16 real quick i have to go soon 07:16 okay, crashed 07:16 why can you crash games on purpose 07:16 there should be level files left in the folder now 07:16 that sounsd like a security issue 07:16 *sounds 07:16 FIQ: that's for bug testing, but I removed it for the next development version because yeah, as you said, it's insecure 07:17 -!- raisse has quit [Quit: Suddenly, Raisse disappears out of sight.] 07:17 it means that people can savescum 07:17 -!- raisse has joined #hardfought 07:17 -!- mode/#hardfought [+v raisse] by ChanServ 07:17 how *do* you crash games on purpose? 07:17 Cannot open level 0 for 5BSOD2. 07:17 FIQ: nah, there's specific conditions that make it happen, it involves the missingno race which is borderline impossible to play in a real game anyway, so there's not going to be any demoscumming going on with that 07:17 are you saying that you added a way to crash the game on purpose into a server build intentionally without asking K2 about it? 07:18 ./recover -d level 5BSOD2 07:18 Cannot open level 0 for 5BSOD2. 07:18 and as I said, it's removed in the next version 07:18 K2: try this please: ./recover level/5BSOD2 07:18 maybe that will work 07:18 (without -d) 07:18 the directory is called "level", right? 07:18 ./recover level/5BSOD2 07:18 ./recover: cannot chdir to /slex-2.1.7/level/. 07:19 GAH 07:19 ok so 07:19 -!- deadnoob has joined #hardfought 07:19 recover binary for slex is broken 07:19 and 07:19 fwiw I looked at grunthack's source 07:19 its recovery utility also asks for directory 07:19 you added a way to purposely crash the game and didnt let me know about it 07:19