01:27:29 -!- Amnesiac_ is now known as Amnesiac 01:35:32 -!- amalloy_ is now known as amalloy 01:37:10 advil: those commits look fine to me, thanks for the quick fixes 01:37:50 Unstable branch on crawl.develz.org updated to: 0.24-a0-702-gaeb99fb (34) 02:02:43 that linker error is quite peculiar; I can't see how the first failing commit would cause that 02:03:07 perhaps Travis changed their vm environment somehow 02:09:04 Windows builds of master branch on crawl.develz.org updated to: 0.24-a0-702-gaeb99fb 02:56:30 Banishment to abyss via distortion caused game to register as entering Depths:5 13https://crawl.develz.org/mantis/view.php?id=12078 by kitchen_ace 02:59:43 Monster database of master branch on crawl.develz.org updated to: 0.24-a0-443-g80245de 03:31:00 Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-2457-gf0d5a3770a 03:55:53 Fork (stoatsoup) on crawl.kelbi.org updated to: 0.22-s0-5-g842d55c91e 04:22:23 Game crashes after selecting background, only with backgrounds that select weapon 13https://crawl.develz.org/mantis/view.php?id=12079 by Goratrix 04:28:06 Good morning all! I'm having some problems gettin 0.16 through 0.18 to compile. It always fails on ability.o with this error: 04:28:09 https://pastebin.com/Mwu3sjaa 04:34:55 Stable (0.19) branch on crawl.s-z.org updated to: 0.19.5-1-g796a6487f8 04:46:42 -!- amalloy is now known as amalloy_ 05:51:20 Namanix: you're using gcc right? If so you need to apply this commit because modern gcc changed things since 0.18 05:51:24 https://github.com/crawl/crawl/commit/5e19f30c998b7066edc6c7e0037d5f9f703ec6f5 05:52:19 (I think that's the problem anyhow, I had the same issue a while ago) 05:52:20 How do I apply those changes to the old stables with the dgl script? 05:52:33 I'm not too great with git on the console :( 06:07:45 this should work (at least for 0.18): save this file, go to the source directory, run 'patch -p1 < patch.txt' 06:07:48 https://pastebin.com/raw/iYvLa1DR 06:08:01 it might work for the others too but i haven't tested it 06:09:18 Thank you! I'll try this :) 06:09:30 if i messed something up, sorry, i'm up way too late right now! :p 06:09:43 Haha :P 06:12:57 (to be clear: i'm not a dev nor an expert these things, but i did have that problem before, and *do* happen to be awake so i thought i'd take a crack at it) 06:23:09 Stable (0.18) branch on crawl.s-z.org updated to: 0.18.1-51-gd549091e01 06:23:51 Oh wait. I think those sizzell messages could be mine 06:24:03 I need to change the hostname lol :P 06:46:09 Stable (0.17) branch on crawl.s-z.org updated to: 0.17.2-2-g6415ff710d 06:46:37 Jep. Now I know for sure it's my server doing this haha 07:13:12 Stable (0.16) branch on crawl.s-z.org updated to: 0.16.2-14-g7133f4a77c 07:39:57 Stable (0.15) branch on crawl.s-z.org updated to: 0.15.2-1-g79e6ab2eea 07:58:05 Stable (0.14) branch on crawl.s-z.org updated to: 0.14.2-1-g46dc7f8a84 08:29:43 Unstable branch on crawl.s-z.org updated to: 0.24-a0-702-gaeb99fb3f2 (34) 08:30:56 Experimental (bcrawl) branch on crawl.s-z.org updated to: 0.23-a0-2457-gf0d5a3770a 08:40:27 Webtiles server stopped. 08:47:25 woah Sizzell is back????? 08:48:03 or is that namani's bot.. 08:58:11 That's my bot 08:58:14 I'm going to turn that af 08:58:17 I'm going to turn that of 09:05:16 Okay it's off. Will keep it that way :P 09:21:09 03advil02 {GitHub} 07* 0.24-a0-703-gd2de06d: tweak windows install instructions 10(10 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/d2de06d31dec 09:32:33 Unstable branch on crawl.kelbi.org updated to: 0.24-a0-703-gd2de06d31d (34) 09:36:35 03advil02 {GitHub} 07* 0.24-a0-704-g6ba659e: Tweak windows install instructions again 10(8 seconds ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/6ba659e1adfd 09:46:15 hm 09:47:14 aidanh fyi symbolic links won't work on windows 09:47:41 Unstable branch on crawl.kelbi.org updated to: 0.24-a0-704-g6ba659e1ad (34) 09:52:58 what fun 09:53:58 03advil02 07* 0.24-a0-705-g0f4390f: Replace some symbolic links with copies 10(45 seconds ago, 2 files, 0+ 0-) 13https://github.com/crawl/crawl/commit/0f4390f660a2 10:01:58 ah, are you building directly on windows? 10:02:34 Unstable branch on crawl.kelbi.org updated to: 0.24-a0-705-g0f4390f660 (34) 10:08:50 03advil02 {GitHub} 07* 0.24-a0-706-gcb3158e: just one more windows install instructions tweak 10(11 seconds ago, 1 file, 5+ 1-) 13https://github.com/crawl/crawl/commit/cb3158e78b19 10:09:24 yeah, trying to see if I can replicate this crash 10:09:51 on my own build that is 10:11:12 hm, I can't reproduce it on wine or on win7 with a cross-compiled crawl.exe 10:11:41 did you try the trunk build download? 10:11:44 or your own? 10:11:53 I can get it to crash reliably in win10 with the trunk download 10:11:58 in virtualbox 10:13:19 hm I wonder if there's a good reason for the install doc to be .txt as opposed to .md 10:13:46 The build is still failing. (master - d2de06d #11941 : Kyle Rawlins): https://travis-ci.org/crawl/crawl/builds/591099157 10:14:02 wtf 10:16:51 oh maybe that is failing on the crawl target, it's just doing it in parallel with monster 10:17:30 Unstable branch on crawl.kelbi.org updated to: 0.24-a0-706-gcb3158e78b (34) 10:18:45 looks like it's failing for both 10:20:35 and yes, the trunk build crashes for me under wine 10:20:52 no symbols in that exe unfortunately 10:26:18 yeah does not crash in my own build 10:26:50 side note but the box around a highlighted button in the main menu looks off on this build: https://www.dropbox.com/s/votfslx4jfto6ww/Screenshot%202019-09-29%2010.25.53.png?dl=0 10:28:25 it's a small screen and in fullscreen mode, could be factors 10:34:05 huh, thats odd 10:34:27 looks like I introduced a second crash around there recently, I think it's independent though 10:36:21 what commit / build is that screenshot from? I remember fixing an off-by-one error in line-drawing code about a year back 10:36:28 branch is_valid has bit me so many times, I don't know why I haven't scrapped it yet 10:37:21 that is from 0.24-a0-693 10:40:38 The build was fixed. (master - d2de06d #11941 : Kyle Rawlins): https://travis-ci.org/crawl/crawl/builds/591099157 10:40:52 what I did to fix that was delete the build cache 10:40:55 in travis 10:41:27 so I think it was just a transitory thing but kept repeating because the broken object file got into the build cache 10:42:00 03advil02 07* 0.24-a0-707-gb641180: Fix a crash with full pregen 10(2 minutes ago, 1 file, 3+ 2-) 13https://github.com/crawl/crawl/commit/b6411803c1cc 10:43:23 hmm, did something change that would make the pregen progress screen not center in console? 10:47:39 Unstable branch on crawl.kelbi.org updated to: 0.24-a0-707-gb6411803c1 (34) 10:59:15 that would likely be 77bbe519b1 11:01:26 %git 77bbe519b1 11:01:26 07Aidan Holm02 * 0.24-a0-677-g77bbe51: Respect popup child widget's preferred size 10(1 year, 1 month ago, 1 file, 5+ 0-) 13https://github.com/crawl/crawl/commit/77bbe519b127 11:01:56 but logically that looks correct; it seems the size requests for the box widget aren't correct 11:02:19 container probably needs expand_h = true 11:02:27 ah ok 11:04:14 for some reason expand_v = true does not seem to make it centred vertically 11:05:27 I think it wasn't centered vertically before, just horizontally 11:06:06 actually, expand_v might make the whole thing centered vertically but the text still appears at the top of the box? 11:06:29 yeah, right, the container needs justify_items = CENTER as well 11:09:07 good catch on the travis issue btw 11:27:12 Stable (0.23) branch on underhound.eu updated to: 0.23.1-89-g277498a001 13:05:53 Unstable branch on crawl.akrasiac.org updated to: 0.24-a0-707-gb641180 (34) 13:48:37 03Aidan Holm02 07* 0.24-a0-708-gaa951a3: Fix hiscores not showing in webtiles (12077, Ge0ff) 10(15 minutes ago, 2 files, 3+ 2-) 13https://github.com/crawl/crawl/commit/aa951a354a79 14:02:38 advil: any chance you know how the trunk build is built? 14:02:52 it's a standard crosscompile that is run on CDO 14:02:57 Unstable branch on crawl.kelbi.org updated to: 0.24-a0-708-gaa951a354a (34) 14:03:38 so one way I have tested things like this in the past is by doing the same cross-compile in a docker container (since I can't do it natively on mac) and then pulling the result into my VM...kind of a pain 14:05:01 hm, as in the same crosscompile that travis does? 14:05:17 yeah 14:05:25 "If a game type is chosen, the user expects // to start a new game." 14:05:32 not sure this logic is correct 14:05:38 well, it's not what I expected at least 14:06:26 that's pretty much what I did, with a xenial container, although I just cross-compiled the exe 14:06:32 I'll try building the full set 14:07:15 I guess it's not necessarily exactly what travis does in that CDO is probably older libs 14:12:02 hrm, yeah 14:13:41 I found that startup menu behaviour pretty unintuitive as well 14:15:19 it's also still possible to reuse an existing name when starting a sprint game, somehow 14:15:45 aidanh: `expand_h = true` does exactly what I want, but it looks bad on local tiles. Webtiles seems to ignore it, so using an #ifdef will work, but I'm guessing webtiles ignoring it is actually a non-implemented feature? 14:16:10 I think sprint saves are separate from regular game saves? 14:16:43 yeah, webtiles UIs are completely in HTML 14:17:34 well, CSS and JS too, of course, but the principle is that HTML already has better tools for this sort of thing 14:21:27 03advil02 07* 0.24-a0-709-g2bac2fb: Restore centering for the progress popup in console 10(2 minutes ago, 1 file, 7+ 0-) 13https://github.com/crawl/crawl/commit/2bac2fb1508e 14:21:27 03advil02 07* 0.24-a0-710-g5bd470a: Don't clear player name when selecting arena (12075) 10(2 minutes ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/5bd470a27781 14:22:30 I think what would be best for the other game types is for there to be a textbox to the right of the name that indicates whether selecting that game type will try to start a new game or load an existing save 14:22:53 could do something like grey out the name box for arena, I suppose 14:23:36 what about simply moving saves to a different screen? 14:24:17 it's possible 14:24:47 there would be pushback from offline players about that I suspect (because of Change) 14:25:01 but that would also make the main menu a lot more normal relative to 2019 14:25:09 yeah 14:25:39 don't forget about the case where a player always uses the same name and has like one save at a time 14:25:59 I think that's what the current menu is optimized for, and might be the normal use case 14:26:03 consolidating the pregen / seed options with game type could be nice as well 14:26:33 right 14:26:43 true; it's one keystroke to resume a game; pretty hard to beat 14:27:06 though I may make that menu accessible in character selection as a kind of hack around getting server owners to update their configs 14:27:13 have to make sure I know what the consequences are though 14:28:20 yeah, I think for my use case (where I have somewhere between 20-50 save games of various kinds, versions, and names kicking around at any given time) and usually test on console, a separate save browser would be better though ;-) 14:28:49 that's quite something 14:29:17 maybe there's some way to have a separate save browser but also preserve one button starts, like a "quick load last game" option 14:29:27 I tend to accumulate around 15-20 saves with names like foo and askjfdl, and purge them every few weeks 14:29:48 I get those and also test games from bug reports and players 14:29:54 not very good at purging them 14:30:05 that's probably the best compromise: [ Resume last game ] [ New game ] [ ...other menu items] 14:31:24 heh yeah I have 41 in my current saves folder and another 53 or so kicking around in various folders 14:31:52 the latter mostly real games, which are useful to have for testing 14:33:25 Unstable branch on crawl.kelbi.org updated to: 0.24-a0-710-g5bd470a277 (34) 14:37:00 oof 14:38:00 btw, how difficult would it be to repurpose the pregeneration code to generate temporary level layouts before a game starts? 14:38:36 I had in mind using randomly generated levels as a background to the main menu, at least on local tiles 14:38:51 oh interesting 14:39:07 that should be doable, it might need some tweaking to keep it from being too slow 14:39:33 but for example temple layouts should be quick enough 14:40:20 ah I see now why this bug report happened -- the arena type + player name interaction didn't actually happen on mouseover before, just keyboard scrolling 14:43:13 oh that reminds me; would it be possible to generate the temple before the dungeon? I always have to wait a few seconds whenever I enter temple 14:43:22 The build is still failing. (master - aa951a3 #11944 : Aidan Holm): https://travis-ci.org/crawl/crawl/builds/591174327 14:43:45 yeah; I'm not actually entirely sure what the logic was with the arena entry clearing the player name 14:43:45 that is true of some of the other cases where the name gets cleared, as well, looks like 14:44:31 yes, that would be possible...shouldn't be as much as a few seconds I think? Or do you mean because pregen is triggered? 14:46:02 I believe crawl finishes generating the rest of the dungeon before temple is generated and entered, yeah 14:46:07 actually mouseovers just did nothing to the name box for the old menus, so we may get more "bug" reports because the new UI makes selection consistent 14:47:57 yeah, although I guess it's easier to mouse over and trigger the weirdness now 14:49:43 apparently my code is robust enough that just moving temple earlier in the pregen order works, though there would be some further tweaking needed 14:51:15 if it is to be moved earlier I guess an even better option might be to just generate temple at the time that the temple map is picked 14:51:51 hi, another issue with the windows local tiles, I put it as note in the previous mantis issue, but might as well paste here while you are fixing :) : when the 'Loading complete, press any key to start.' intro screen is shown, if instead of pressing a key you click the mouse, it registers that mouse click as if it was done on the next screen, i.e. 14:51:52 if you click on the loading screen in the place where the 'Hints mode' is going to be displayed, it immediately jumps into the Hints mode menu after showing the next screen. 14:52:08 oh yeah I've been noticing that one 14:52:21 it's not platform-specific 14:52:21 oh ok :) 14:52:36 well I hadn't actually got to characterizing what the issue was, so thanks :-) 14:53:58 haha oops I closed the wrong bug 14:53:58 interesting, I can't reproduce that 14:54:00 on linux local tiles that is 14:55:27 oh, might have something to do with most recent game being auto focused; the mouse-up event is acted upon 14:55:40 yeah, it was loading a save for me 14:56:57 well, it is loading whatever save is under the mouse cursor 14:57:21 if I click somewhere on the screen that will not have UI, nothing happens 14:58:33 I guess the screen should transition on mouseup? might be more robust to add some sort of event filtering between layout transitions though 14:59:21 yeah...did you change the opening popup at all? 14:59:31 well, it's been 'widgified' 14:59:36 heh 14:59:44 one thought I've had would be to modify the pregen progress popup to deal with that ui to, which would assimilate it more to the rest of the code 14:59:50 *too 15:01:10 other than that, it just waits for a keypress, and mouseclicks are currently sent through again as CK_MOUSE_CLICK keypresses if they're otherwise unhandled 15:01:15 for compatibility 15:01:22 gotta rip that out at some point 15:01:42 that could be nice, yeah 15:10:55 huh apparently CDO is using mingw32 15:11:02 so it's *not* a standard cross-compile 15:16:10 aha 15:17:24 would there be any particular reason for that? mingw64 should produce 32-bit exe files, after all 15:18:17 I'm not sure the "32" there means what it seems like it should mean 15:19:00 The build is still failing. (master - 5bd470a #11945 : advil): https://travis-ci.org/crawl/crawl/builds/591182887 15:19:39 the compiler is i686-w64-mingw32.static-g++, so it might actually be building 64bit binaries 15:21:49 doesn't it hae to be? I thought nobody had been able to build 32 bit windows, which implies CDO hadn't been. (32 there is as opposed to real-mode Windows and is historical) 15:23:02 mingw32 is the old platform... according to the wiki there was some disagreement when mingw-w64 was forked 15:23:37 msys2 is built around mingw-w64 15:23:53 specifically it seems to be using this: https://github.com/mxe/mxe 15:24:03 the exe I cross-compiled certainly seems to be 32-bit 15:24:03 from 2014 15:24:18 for release we normally build 32 bit windows only via MSYS2 cross-compiling 15:25:11 that's also what travis does (because I set it up based on the release procedure) 15:26:45 the repository that builds the trunk dailies is at this commit, fwiw: https://github.com/mxe/mxe/commit/b2fcc7a486 15:26:55 -!- amalloy_ is now known as amalloy 15:27:01 I don't have permissions to mess with this repo, just view it 15:28:59 I'm amazed this build works tbh 15:30:37 that.. sounds pretty rickety yeah 15:31:34 er I should say via mingw-w64 cross-compiling, sorry 15:32:20 right 15:32:35 thanks for the info; i'll give this funky mxe build a shot 15:34:33 btw I do have the capability to trigger the CDO build at will 15:34:50 well, I shouldn't abuse nap.kin's cpu 15:37:12 03advil02 07* 0.24-a0-711-g964def8: Fix a compiler warning (shadowed variable) 10(72 seconds ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/964def8651a8 15:37:12 03advil02 07* 0.24-a0-712-gf68d918: checkwhite 10(21 seconds ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/f68d91829209 15:48:13 Unstable branch on crawl.kelbi.org updated to: 0.24-a0-712-gf68d918292 (34) 15:48:30 I've made a mistake configuring dgamelaunch and now it crashes on ssh login. Where can I find log files for dgamelaunch? I can't find a log location in the config 15:52:56 namanix: When I had this problem I ended up straceing it, there aren't always satisfactory logs 15:53:35 straceing? 15:53:35 "Crashes on ssh login" is usually a permissions problem 15:53:35 Running it under "strace" 15:54:01 Okay I'll try that :) 15:55:08 eg (as user "crawl") doing "strace -f -o foo dgamelaunch" and looking in "foo" afterwards, a pastebin of which might be interesting 16:29:48 The build is still failing. (master - 964def8 #11946 : advil): https://travis-ci.org/crawl/crawl/builds/591201898 16:34:10 oh heh, I guess I did do skip ci on the checkwhite commit 16:46:31 03advil02 07* 0.24-a0-713-gbfea814: Apply some brute-force output debugging to the problem 10(31 seconds ago, 1 file, 33+ 0-) 13https://github.com/crawl/crawl/commit/bfea814871e4 16:47:59 usually no one has to see me do this kind of stuff ;-) 16:48:22 but needs must 17:02:54 Unstable branch on crawl.kelbi.org updated to: 0.24-a0-713-gbfea814871 (34) 17:06:26 aidanh: well, the crash is in the playerdoll code looks like 17:11:22 03advil02 07* 0.24-a0-714-g40886de: More public bisection 10(2 minutes ago, 1 file, 6+ 30-) 13https://github.com/crawl/crawl/commit/40886de7c53b 17:17:48 Unstable branch on crawl.kelbi.org updated to: 0.24-a0-714-g40886de7c5 (34) 17:32:14 specifically in fill_doll_for_newgame 17:33:23 I have to admit I was a bit worried about unwinding the entire player like that 17:34:13 it's such a heavy object and its initialization is historically a source of bugs 17:34:37 though I haven't done the work to know if that's actually the problem here 17:38:30 The build was fixed. (master - bfea814 #11947 : advil): https://travis-ci.org/crawl/crawl/builds/591220395 17:48:28 03advil02 07* 0.24-a0-715-g0b739ee: One more batch of debug code 10(34 seconds ago, 2 files, 10+ 9-) 13https://github.com/crawl/crawl/commit/0b739ee41d2d 18:03:10 Unstable branch on crawl.kelbi.org updated to: 0.24-a0-715-g0b739ee41d (34) 18:09:23 The build was fixed. (master - 40886de #11948 : advil): https://travis-ci.org/crawl/crawl/builds/591226770 18:22:45 Unstable branch on underhound.eu updated to: 0.24-a0-715-g0b739ee41d (34) 19:12:33 aidanh: nice stuff for the ui rework landing! Do you have any migration notes (for forks) other than "check the commits which rewrote the menu you are interested in"? 19:49:49 03advil02 07* 0.24-a0-716-g2d0fce6: Don't generate unknown weapons for weapon selection 10(2 minutes ago, 2 files, 3+ 11-) 13https://github.com/crawl/crawl/commit/2d0fce64a113 20:00:14 well, that seems to have fixed the crash 20:00:50 I still can't figure out what the actual mechanism was for the crash, though, or why it was OS-specific 20:01:56 I stepped through the entire relevant function (give_items_skills) in lldb on my mac, and while a weird item with subtype WPN_UNKNOWN is definitely there, it wasn't clear to me why it should crash any of the relevant code, it seemed like tile picking handled that case correctly 20:03:06 Unstable branch on crawl.kelbi.org updated to: 0.24-a0-716-g2d0fce64a1 (34) 20:26:20 interesting 22:10:35 I'm thinking of merging this before the release (maybe) https://github.com/crawl/crawl/pull/1040 22:10:47 I like the revised version 22:11:13 tho maybe without the +4 SH 22:38:26 is it too early to suggest removing the spell Battlesphere if we merge this? I want to be the first 22:41:33 at first I found the name "staff of Battle" confusing since this is an item you never melee with, but I guess that's not actually confusing 22:41:58 battlesphere is a battle thing that isn't doing melee 23:53:46 Battlestaff when