00:46:49 -!- amalloy_ is now known as amalloy 02:48:52 Monster database of master branch on crawl.develz.org updated to: 0.24-a0-443-g80245de 03:30:58 Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-2820-gdc0aeed4c6 03:32:21 Fork (bcadrencrawl) on crawl.kelbi.org updated to: 0.22.1-1800-ga33b126f4d 03:54:55 Rubinko (L16 NaVM) Crash caused by signal #11: Segmentation fault (Spider:2) 10:23:15 !messages 10:23:16 (1/2) ebering said (5d 23h 27m 13s ago): I don't even remember what I was merging 37 weeks ago.... maybe the summoner highlighting? 10:23:20 !messages 10:23:20 (1/1) advil said (1d 16h 26m 50s ago): there's some race-condition-like issues with ui state syncing; I patched the most immediate issue for seed selection in particular in b773885, but I don't think that commit is the full story & it's something your input would be helpful on even if you can't contribute code per se (not sure of the legalities though) 10:24:24 !git b773885 10:24:24 %git b773885 10:24:24 07advil02 * 0.25-a0-288-gb773885: Fix (?) a race condition with ui state syncing 10(2 days ago, 1 file, 8+ 0-) 13https://github.com/crawl/crawl/commit/b77388520f1f 10:29:53 advil: sounds like the specific issue there is that webtiles is syncing the empty textbox's state to the server; that's only supposed to happen when the player types something 10:34:49 iirc the server should also propagate the client->server update back to all clients (so spectators see changes), but i think in that case the player client will drop the message 10:36:42 right, at https://github.com/crawl/crawl/blob/4a1327ee3aad6193e6b19f244245cd75b0287210/crawl-ref/source/webserver/game_data/static/ui.js#L393 10:41:50 so re the general possibility of messages crossing paths, it should only occur within a popup; popups have "generation" ids, and sync messages have an id; if there's a mismatch, the message is dropped 10:45:59 within a given popup, it'd depend on how the UI works, but I imagine changing widget A in response to input on widget B on the server could lead to issues if the user changes A and B in quick succession on the client 10:46:39 it's not inconceivable that that could be useful; e.g. enabling/disabling a "Start" button based on whether you've entered a valid seed 11:23:42 Stable (0.23) branch on underhound.eu updated to: 0.23.1-91-gf373564dc4 14:44:42 I like the idea of new Conjure Flame spell, but not the implementation. I don't know if casting it with normal speed species would be better experience, but my CeFE has hard time. Very often monsters step onto CJ spot and flames don't start. Is it intended behaviour (should players cast it when monsters aren't nearby)? 14:45:33 s/CJ/CF 14:46:47 If not, then I have a suggestion: casting CF gives you a status (like digging for Fo), and the next time you move, you leave cloud behind. 15:24:33 notes.cc:382:17: warning: format string is not a string literal (potentially insecure) [-Wformat-security] 15:24:35 die(make_stringf("Unknown note type %d", type).c_str()); 15:24:37 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 15:24:57 can someone explain this error and how I fix it? 15:26:48 die apparently doesn't need make_stringf since it wants a format string, so: die("Unknown note type %d", type) 15:27:08 also that's a warning, not an error, unless you use -Werror for some reason 15:31:15 aha, thanks 16:17:11 CabbageHands (L19 MfGl) ASSERT(!invalid_monster(&mons)) in 'mon-death.cc' at line 1716 failed. (Vaults:1) 16:32:18 alexjurkiewicz the general issue is that the first argument to a printf-like function is not supposed to be anything except a string literal for security reasons, because (especially if the string is provided by a potential adversary) that opens up to all sorts of attacks or just confusing errors 16:32:28 where `die` is a printf-like function 16:33:56 !tell aidanh aha thanks! Syncing from webtiles only when there's user action makes a lot of sense, and I didn't realize was the intended behavior. So the bug then is that for some reason the textbox sends state on creation (?), which is a bit different than what I thought. 16:33:56 advil: OK, I'll let aidanh know. 16:34:22 ah right, so it's differentiating a user provided string from one hardcoded in source 16:35:57 yeah 16:36:38 in this case there's no obvious attack (unless make_stringf has a problem ofc) but the warning is not smart enough to know that 16:37:07 right, I didn't explain that aspect but that's what it doesn't like about computed format strings 16:37:08 this is why you often see `printf_like("%s", s.c_str());` too 16:37:33 as opposed to `printf_like(s.c_str());` 16:37:56 this used to be a common issue in C source code, people would do printf(some_string); and some_string came from a user who could drop a % in it and inspect the stack that way 16:38:19 or use locale stuff to inspect various other things 16:42:34 Yermak: my understanding is that you're not supposed to be able to use conjure flame to create a gap when the monster is currently adjacent 16:42:51 this is why monsters stepping on the flame stops it from igniting 16:44:54 New branch created: pull/1230 (1 commit) 13https://github.com/crawl/crawl/pull/1230 16:44:54 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1230 * 0.25-a0-291-ge6ba257: Increase Quicksilver Dragon Armour ER to 8 (was 7) 10(8 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/e6ba25703259 16:45:28 If it's so, then I understand. But sometimes I am successful at it. 16:46:12 At creating flame cloud despite adjacent monster. 16:47:24 as a centaur, i can imagine that 16:47:40 actually, how does player move speed affect conjure flame? i'm not sure 17:02:31 I think ebering committed something to make it invariant to move speed, but I might not have fully understood 17:03:24 !crashlog Rubinko 17:03:53 25. Rubinko, XL16 NaVM, T:33605 (milestone): http://crawl.akrasiac.org/rawdata/Rubinko/crash-Rubinko-20191214-085454.txt 17:09:10 %git 56a5b49892 17:09:10 07advil02 * 0.25-a0-287-g56a5b49: Don't crash when checking if fedhas protects player from igniting 10(3 days ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/56a5b4989272 17:10:04 03advil02 07* 0.25-a0-291-g84510d0: Fix a missing ! 10(33 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/84510d0c19a9 17:11:28 hm, there's a lot of this *act->as_monster() going on 17:11:42 which is quite unsafe, as it's guaranteed to give an invalid reference for the player 17:19:59 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-291-g84510d0c19 (34) 17:36:44 03advil02 07* 0.25-a0-292-gaa31f04: Make fedhas_protects less likely to crash if player is in the mix 10(4 minutes ago, 9 files, 23+ 21-) 13https://github.com/crawl/crawl/commit/aa31f046c41c 17:47:40 advil: I had sort of thought a function like that would always be better designed to take a ref. I guess typical usage cases dictate pointer vs ref rather than a rule about what the function does? 17:49:32 <|amethyst> also, instead of target ? foo : false, target && foo might be more idiomatic 17:51:52 <|amethyst> in this case, it might be better to make it take a const actor&, and just get rid of the ->as_monster()s 17:52:16 <|amethyst> assuming the actors can't actually be null 17:53:03 <|amethyst> (and, of course, put the as_monster and corresponding null check inside the function) 18:09:04 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-292-gaa31f046c4 (34) 18:10:48 !seen floraline 18:10:49 I last saw floraline at Thu Dec 5 10:05:03 2019 UTC (1w 2d 13h 5m 46s ago) joining the channel. 18:15:51 yeah, that makes sense, re: the fedhas_protects function 18:21:30 how can I print debug info during new game menus? 18:21:47 dprf() doesn't appear in the log once the game start 18:23:30 Unstable branch on underhound.eu updated to: 0.25-a0-292-gaa31f046c4 (34) 18:23:45 alexjurkiewicz: not sure how aidanh does it, but you can always just use printf and check output in your terminal 18:24:00 does require that you run the app from terminal even if it's tiles, of course 18:25:05 that works, thanks for the idea 18:36:27 hmm,the output only appears when i exit crawl 18:36:37 i guess i need to flush something 18:38:32 fflush(stdout); 18:39:45 !tell aidanh something about ui code changed with your updates a couple of weeks ago so that the bloatcrawl game modifier code receives two events every time you press a key 18:39:45 alexjurkiewicz: OK, I'll let aidanh know. 18:40:00 !tell aidanh specifically the on_keydown_event 18:40:00 alexjurkiewicz: OK, I'll let aidanh know. 18:43:59 !tell aidanh Changing to on_hotkey_event seems to fix the issue, but you mentioned this works differently so it might not be a good fix...? 18:44:00 alexjurkiewicz: OK, I'll let aidanh know. 21:06:38 -!- Tiobot is now known as Guest75278 22:27:47 -!- amalloy is now known as amalloy_