00:01 -!- Umbire has joined #tnnt 00:05 -!- fakepass has quit [Quit: Leaving] 00:09 -!- fakepass has joined #tnnt 00:19 oof. gotta get crackin' 00:24 -!- stagnate has quit [Ping timeout: 248 seconds] 00:37 -!- crunchbubba has quit [Quit: Ping timeout (120 seconds)] 00:39 $lastgame tangles 00:39 Furey: [AU] https://au.hardfought.org/userdata/T/Tangles/tnnt/dumplog/1697973759.tnnt.html 00:44 fyi, i loaded those tangles bones from australia, really neat how bones can migrate from au to us so smoothly. 01:05 -!- Furey has quit [Quit: Client closed] 01:19 -!- stagnate has joined #tnnt 01:33 -!- Furey has joined #tnnt 01:35 -!- crunchbubba has joined #tnnt 01:55 -!- Umbire has quit [Remote host closed the connection] 01:55 -!- Umbire has joined #tnnt 02:52 -!- crunchbubba has quit [Quit: Client closed] 02:54 -!- Furey has quit [Quit: Client closed] 03:00 -!- Furey has joined #tnnt 03:16 -!- crunchbubba has joined #tnnt 03:52 heh, Furey has so many heavy corpses in his BoH that he can't take them out, because he would be crushed under the load 03:52 yeah, i'm takiing this marie kondo thing too far :( 03:53 I don't understand what you are doing. I though at first you were taking them to a central ice box 03:53 i got triggered when i lost my unihorn, looking all over for it. 03:53 probably a nymph stole it and went into a closet somewhere. 03:53 I sent you ingame mail, did it show up? I didn't see it 03:54 no, i turned off mail a little while ago 03:54 do an #overview? 04:12 I tend to set down the bag before doing that, and then only when I'm lugging stuff to altars. 04:17 i just like making the levels clean, it doesn't advance the game any 04:24 -!- theRaisse has joined #tnnt 04:25 I do that when it gets chaotic but only by stacking them in one place.  I don't lug them around the map. 04:26 I just looked for the yesu ttyrec where he REALLY cleaned the entire dungeon 04:29 -!- crunchbubba has quit [Quit: Client closed] 04:30 -!- crunchbubba has joined #tnnt 04:47 if you were still watching, kicking all the trees on monk home 1 is a real strat, food for vegan monk 04:53 -!- Furey has quit [Quit: Client closed] 04:57 -!- raisse has joined #tnnt 04:58 -!- theRaisse has quit [Ping timeout: 252 seconds] 05:03 -!- raisse has quit [Ping timeout: 264 seconds] 05:10 -!- stagnate has quit [Ping timeout: 248 seconds] 05:32 -!- raisse has joined #tnnt 06:05 'course it is. 06:06 Not many trees around as it is in vanilla. 06:18 -!- fakepass has quit [Quit: Leaving] 06:22 -!- crunchbubba has quit [Quit: Client closed] 06:57 -!- elenmirie has joined #tnnt 08:44 -!- elenmirie has quit [Quit: Going offline, see ya! (www.adiirc.com)] 08:54 -!- engelson has joined #tnnt 10:23 -!- engelson has quit [Quit: Quit] 11:46 -!- Furey has joined #tnnt 12:31 -!- elenmirie has joined #tnnt 12:38 -!- Furey has quit [Quit: Client closed] 14:04 -!- elenmirie has quit [Quit: Going offline, see ya! (www.adiirc.com)] 14:12 -!- crunchbubba has joined #tnnt 14:45 -!- fakepass has joined #tnnt 14:52 -!- Furey has joined #tnnt 15:23 -!- crunchbubba has quit [Quit: Client closed] 15:53 -!- nurdglaw has joined #tnnt 15:59 -!- Furey has quit [Quit: Client closed] 16:00 -!- Furey has joined #tnnt 16:00 -!- theRaisse has joined #tnnt 16:03 -!- raisse has quit [Ping timeout: 248 seconds] 16:04 $time 16:04 K2: Message from Furey at 2023-10-22 02:43 UTC: check messages please 16:04 K2: 2023-10-23 16:04:47 UTC. 2023 Tournament begins in 8d 07:55:12 16:04 single digits! 16:13 -!- crunchbubba has joined #tnnt 16:24 exciting 16:24 qt: Message from K2 at 2023-10-21 20:06 UTC: <3 16:26 glad you sent that because I just noticed I missed some mentions of me (it didn't light up because I had this channel open while I left the computer on over the weekend) 16:26 also because it's nice to feel loved 16:28 I am bringing a land mine to valley of the dead to express my appreciation of all tnnt devs. 16:28 heh 16:36 (coughs) 16:36 * K2 hands crunchbubba a mask 16:37 or a unihorn 16:37 whichever 17:59 -!- theRaisse has quit [Ping timeout: 264 seconds] 18:17 -!- nurdglaw has quit [Quit: Client closed] 18:29 -!- Furey has quit [Quit: Client closed] 18:33 -!- Furey has joined #tnnt 18:38 -!- nurdglaw has joined #tnnt 19:09 -!- elenmirie has joined #tnnt 19:34 -!- Furey has quit [Quit: Client closed] 19:38 crunchbubba: you're still having that problem with the OSX default terminal emulator right? 19:38 yes 19:39 will you try export TERM=xterm-kitty and then connect again and run TNNT? 19:39 oddly that makes it work fine for me, but all other TERM values I've tried the issue persists 19:39 I want to see if it fixes it for you too 19:44 -!- Furey has joined #tnnt 19:45 I use an ancient shell, but setting TERM variable (or trying to) says The terminal database could not be opened. using dumb terminal settings. 19:45 xterm-kitty is not a setting that is available in the terminal gui 19:46 but ... the game looks pretty clean! 19:47 and reverting back to TERM xterm-256color manually in the shell and relaunching has the problem again! 19:48 so that's a fix in some sense 19:49 yay! 19:50 yeah I'm sure it's not in your local terminfo but it seems to fix it on hdf 19:50 but it seems like all the other xterm-foo termcap entries cause this problem on hdf 19:50 yes, for the 3 I tried 19:51 can you upload the terminfo file and I will look to see if anything jumps out at me? 19:55 you'd have to get it from K2 to have the remote file. I think any problems with the terminfo files are going to be with the ones on hardfought -- when I run TNNT locally I don't have any problems in the default terminal emulator 19:55 -!- aosdict has joined #tnnt 19:55 -!- mode/#tnnt [+o aosdict] by ChanServ 19:56 btw, another interesting thing is that if I playback the ttyrec of one of those sessions, the random repeated letters/numbers don't appear 19:56 instead it's as though they are blank spaces. so something that appeared as |...dddddddd| will appear as |...| (and everything after it will be shifted off to the left/misaligned with other lines) 19:57 that's true both in OSX Terminal and in other terminal emulators that don't exhibit this issue 19:58 K2 maybe you could try running `infocmp xterm-kitty xterm-256color` on hardfought and see what differences it flags 20:03 K2 was 3.7 built with a different libcurses version than TNNT? 20:03 because it's also weird that the problem only happens in TNNT (afaict) 20:03 https://termbin.com/3so4 20:04 no it should all be the same 20:04 oh 20:04 one sec 20:04 qt, try connecting to eu server and test same things 20:05 there are some very slight differences between libs on the OS vs chroot on hdf-us - i havent had a chance to update the chroot yet, because if i do when ppl are playing, it can cause a crash 20:05 on hdf-au/eu they match (OS and chroot) 20:06 it's happening to me on EU too 20:06 ok 20:06 i thought it might but was worth a try 20:07 $who 20:07 Furey: [EU] qt | [US] Furey allanb Bartlem | [AU] No current players 20:08 also still not happening in 3.7 on EU 20:08 hmm 20:12 export TERM=rxvt also seems to fix it 20:25 it does! and that's a choice I can pick locally 20:25 I hadn't tried that one 20:25 I' 20:26 I'm 99% sure it's the ansi "repeat last char" function that is broken 20:27 yeah, it does seem like it's something to do with repeating characters. I didn't know there was an ANSI repeat character function 20:28 like the upper left corner of the popup box is l in the alternate char set, and the flat line is q, so it sends "lq" and then "repeat last char 57 times" and then "qj" where j is the char for upper right hand corner 20:28 and I may have misstated l and j but very close to that 20:29 ansi/curses jump through a lot of hoops to try to minimize the number of chars sent across the wire 20:31 I'll be AFK for a while. drilling holes and running hoses to try to ensure that I have running water this winter. kind of important. I may wander by the computer but may not be able to type 20:32 if that's what's happening, it seems odd that 3.7 doesn't have the same issue 20:45 K2 when you have a chance will you tell me what output you get from `tput -Txterm-256color rep` and `tput -Txterm-kitty rep` on hdf? 21:08 -!- nurdglaw has quit [Quit: Client closed] 21:11 -!- Umbire has quit [Ping timeout: 260 seconds] 21:21 -!- Loggers_VIII has joined #tnnt 21:35 -!- Furey has quit [Quit: Client closed] 21:37 -!- crunchbubba has quit [Quit: Client closed] 21:38 -!- Furey has joined #tnnt 21:42 -!- Loggers_VIII has quit [Quit: Client closed] 21:43 -!- Loggers_VIII has joined #tnnt 21:50 -!- Furey has quit [Quit: Client closed] 21:53 -!- Furey has joined #tnnt 21:59 qt - https://termbin.com/rdwzd 21:59 ^ same for both 22:00 well in that case I don't see how it would be the cause of the problem, if it's outputting the same repeat escape sequence for both of them 22:00 thanks for checking btw 22:00 no prob 22:00 wish i could be more help 22:01 bhaak is pretty good with this sort of thing 22:08 qt is there anything online, like on any mac forums where users are seeing weird behavior in other terminal sessions? 22:17 -!- Furey has quit [Quit: Client closed] 22:22 K2: I don't know, when I search for things like "characters repeating macos terminal" I just get a bunch of stuff like "how to activate keyboard key repeating on macOS" 22:24 https://stackoverflow.com/questions/8138161/boxes-and-lines-in-ncurses-through-ssh this is the closest thing I found and it's not actually about the OSX terminal emulator (nor does it sound very close anyway) 22:24 :/ 22:24 could ask about it on stack overflow, idk what community on there would be the right place though 22:25 you could argue it fits in general programming, sysadmin, OSX (sort of) 22:27 -!- Loggers_VIII has quit [Quit: Client closed] 22:27 you gotta keep volunteer firefighter rainmen on standby for that kind of thing. i know huge tech companies do that but without the volunteer firefighter part. 22:28 -!- Loggers_VIII has joined #tnnt 22:34 -!- Furey has joined #tnnt 22:35 -!- elenmirie has quit [Quit: Going offline, see ya! (www.adiirc.com)] 22:38 that would probably be a fun job, puzzling out the most challenging bugs 22:39 too bad i can't do rain dances 22:43 take a look at this post and the subsequent discussions in the thread. 22:43 https://lists.gnu.org/archive/html/bug-ncurses/2017-07/msg00013.html 22:49 Furey: that is interesting and sounds like a similar issue, but the characters don't just get converted to the equivalent in the unshifted character set. the character that replaces window borders varies, and then on the map some of the ASCII characters just repeat across the room after 1 normal square 22:49 let me show you one sec 22:50 https://0x0.st/HJ0c.png one example 22:50 https://0x0.st/HJ0B.png another example 22:50 it's almost like it has the ASCII character in memory and uses it to fill the space because it doesn't know what it's supposed to be repeating 22:51 btw when viewing a ttyrec, the repeated characters aren't rendered at all (not replaced with the normal characters, just not rendered) 22:51 yeah definitely not identical issue, in your PNG's the first char is fine and then the repeats are broken. 22:51 yeah 22:52 crunchbubba said he had one TERM= that worked and another TERM= that did not work? 22:53 yeah I guess it's probably based on rep support, though the weird thing is that K2 checked tput rep for one of the working and non-working TERMs on hdf and apparently they both reported they supported it with the same escape sequence 22:53 rxvt, xterm-kitty, vt100 <-- do not exhibit the problem 22:53 xterm-(everything else I've tried) <-- does exhibit the problem 22:54 get a pair of working and non-working that are close and look at all the differences, not just one tput capability 22:54 like xterm-kitty and xterm-whatever 22:54 yeah we got that earlier, one sec 22:54 https://termbin.com/3so4 22:55 (also my curses/termcap/terminfo knowledge is years and years rusty!) 22:56 thanks for the disclaimer, but better rusty than near-nonexistent :) 22:57 can you make your own xterm-test terminals? start with xterm-kitty and change one thing (or a few things) at a time towards xterm-256color and see when it goes from noproblem to problem. 23:00 -!- Furey has quit [Quit: Client closed] 23:01 that's a smart idea 23:04 -!- Furey has joined #tnnt 23:05 I guess the way to do that would be `infocmp -u xterm-kitty xterm-256color` (which seems to incorporate xterm-256color by reference and override only the parts that differ), then take out one of the differences, compile it with tic(1), try connecting, etc? 23:09 yeah, either one at a time, or binary search on the 40 differences .. gonna be tedious 23:29 Furey: I think the significant difference may be that the xterm-kitty file on hdf indicates it has no rep support, which would make sense. K2 gave me the terminfo file from hdf and I can see that it says no rep support. but for some reason he sees output from running 'tput -Txterm-kitty rep' on hdf so idk 23:30 maybe he gave me the xterm-kitty file from /lib/terminfo and there's a second one in /etc/terminfo, or something like that 23:30 -!- crunchbubba has joined #tnnt 23:32 ok yeah I'm pretty sure the problem here is just that OSX Terminal doesn't support rep, but by default it reports TERM=xterm-256color and the hdf xterm-256color terminfo file says it does support rep 23:32 but I have no idea why vanilla nethack does not have a problem 23:33 I can repro it locally without ssh if I set TERM to one that does have rep support in my local terminfo and run local copy of TNNT 23:33 i think crunchbubba didnt have any issues using iterm2 on mac 23:33 (vs default terminal) 23:34 but 3.7 does not exhibit this problem. 3.6 does, so idk what changed but I guess it's some xterm incompatibility with the default OSX terminal emulator 23:35 perhaps vanilla nethack makes 40 calls to 'put one char' and the nethack version with a problem makes one call to 'put this char 40 times' (which degrades to 40 'put one car' if terminfo does not advertise repeat capability) 23:36 when I run my (very poor) bot, it gets sent repeat char ansi sequence, but no clue what terminal the system thinks it is 23:36 what TERM (if any) does it send? 23:36 I do not know. as fars as I can tell, none 23:37 let me look to see if I explicitly say in the negotiation 23:37 but that's how I was immediately pretty sure that the problem involves repeat last char (perhaps 'repeat last graphical char') 23:40 the bot uses telnet, and does not specify a terminal type 23:42 this is the desc for the repeat  CSI Ps b Repeat the preceding graphic character Ps times 23:43 and CSI is just ESC [ 23:44 and I think that's just ANSI for generic character movement on terminals 23:46 ahh 23:46 so it must be failing to recognize the shifted character set as "graphics characters" 23:46 perhaps 23:47 because it does seem to have rep support, it's just that when it's supposed to repeat a non-ASCII/line-drawing character, it instead repeats the previous ASCII character 23:48 my code I think does not differentiate between graphics char and char, and it works(worked?) because the repeat sequence was always immediately after the graphic char was sent 23:49 like my desc earlier of how a popup box borders are done. alt chars (which is lines) for corner and a flat line, then repeat graphic char 23:50 -!- elenmirie has joined #tnnt 23:52 yeah I think I missed this part of the conversation, it would probably have saved me a lot of confusion if I'd seen you talk about this in detail earlier :) 23:56 ha