00:00:46 Stable (0.13) branch on crawl.s-z.org updated to: 0.13.0-7-g3df9ff1 00:01:59 Unstable branch on crawl.s-z.org updated to: 0.14-a0-266-g978670c (34) 00:02:25 -!- Nex has quit [Quit: Page closed] 00:02:45 -!- Roarke has quit [Ping timeout: 250 seconds] 00:05:59 Unstable branch on rl.heh.fi updated to: 0.14-a0-266-g978670c (34) 00:08:31 -!- thened has quit [Quit: thened] 00:09:19 -!- emeraldemon has quit [Quit: emeraldemon] 00:10:18 -!- _dd has joined ##crawl-dev 00:10:53 Stable branch on crawl.develz.org updated to: 0.13.0-7-g3df9ff1 (34) 00:11:02 -!- Deckard_Pain has quit [Ping timeout: 256 seconds] 00:11:16 -!- VolteccerJack has quit [Quit: HydraIRC -> http://www.hydrairc.com <- Wibbly Wobbly IRC] 00:14:29 -!- eith has quit [Ping timeout: 272 seconds] 00:15:18 -!- eith has joined ##crawl-dev 00:18:04 Unstable branch on crawl.develz.org updated to: 0.14-a0-266-g978670c (34) 00:21:10 -!- hyperbolic has joined ##crawl-dev 00:23:59 -!- elliptic has quit [Ping timeout: 272 seconds] 00:31:26 -!- nonethousand_ has joined ##crawl-dev 00:35:45 -!- nonethousand has quit [Ping timeout: 272 seconds] 00:35:47 -!- Kaput has quit [Read error: Operation timed out] 00:36:53 -!- slifty has quit [Quit: Leaving...] 00:36:55 -!- hyperbolic is now known as elliptic 00:38:48 -!- Pacra has quit [Ping timeout: 256 seconds] 00:40:21 -!- Sgun has quit [Read error: Connection reset by peer] 00:42:13 -!- Taraiph has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] 00:43:42 -!- gammafunk_ has joined ##crawl-dev 00:44:34 -!- jameyd has quit [Quit: jameyd] 00:47:22 Windows builds of master branch on crawl.develz.org updated to: 0.14-a0-266-g978670c 00:51:08 -!- ystael has quit [Ping timeout: 240 seconds] 00:52:10 -!- Bcadren has quit [Quit: Friends help you move. Real friends help you move bodies.] 00:57:11 -!- dtsund has joined ##crawl-dev 00:57:21 Windows builds of pre-release branch on crawl.develz.org updated to: 0.13.0-7-g3df9ff1 01:02:37 -!- Quashie has quit [Ping timeout: 272 seconds] 01:06:38 -!- Roarke has quit [Quit: Page closed] 01:15:12 -!- Giomancer has quit [Quit: Leaving] 01:21:24 -!- bh has quit [Quit: Page closed] 01:21:58 ontoclasm: do you want to join baddev? "# TEAMCAPTAIN neil" and tell |amethyst if you do; note that you've played about as many games as bh, and gotten higher scores too 01:30:59 -!- PepeRC2 has quit [Ping timeout: 260 seconds] 01:32:16 hah 01:32:29 i suppose 01:32:56 !tell |amethyst I guess I'll join baddev if you don't mind me dragging everyone down 01:32:57 ontoclasm: OK, I'll let |amethyst know. 01:37:00 -!- Nex has quit [Quit: Page closed] 01:43:19 -!- Kalir has quit [Quit: CADENZA-CLASS CLOCKWORK KNIGHT--OPERATION TERMINATED] 01:50:09 -!- sildraith has quit [Quit: Leaving] 02:01:09 -!- omnirizon has quit [Ping timeout: 272 seconds] 02:03:17 -!- magistern has quit [Quit: This computer has gone to sleep] 02:04:26 -!- absolutego has joined ##crawl-dev 02:10:04 -!- st_ has joined ##crawl-dev 02:12:57 -!- wheals has quit [Ping timeout: 252 seconds] 02:15:05 -!- Vandal has quit [] 02:16:15 -!- SkaryMonk1 has joined ##crawl-dev 02:16:51 -!- Hailley has quit [Quit: Petra is closed!] 02:17:47 -!- randart has quit [Ping timeout: 248 seconds] 02:22:42 -!- wheals has joined ##crawl-dev 02:23:17 -!- ChongLi has quit [Read error: Connection reset by peer] 02:26:15 -!- randart has quit [Ping timeout: 245 seconds] 02:29:55 -!- minced has quit [Quit: Page closed] 02:31:07 -!- mong has quit [Quit: Leaving] 02:31:39 -!- magicpoints has quit [Ping timeout: 252 seconds] 02:34:54 -!- indspenceable has quit [Remote host closed the connection] 02:35:17 -!- kilobyte has joined ##crawl-dev 02:35:19 -!- randart has quit [Ping timeout: 265 seconds] 02:35:20 -!- indspenceable has joined ##crawl-dev 02:37:32 -!- magicpoints has quit [Ping timeout: 240 seconds] 02:39:35 -!- indspenceable has quit [Ping timeout: 245 seconds] 02:40:09 -!- ussdefiant has quit [Ping timeout: 272 seconds] 02:40:16 -!- mint_ has joined ##crawl-dev 02:41:13 -!- mint_ is now known as Thyme 02:42:23 -!- Basil has quit [Ping timeout: 260 seconds] 02:45:26 -!- minqmay has quit [Quit: Leaving] 02:47:03 -!- ystael has quit [Ping timeout: 252 seconds] 02:56:50 -!- wheals has quit [Quit: Leaving] 03:11:40 ontoclasm: Does 'dragging everyone down' mean that you'll win? 03:13:03 -!- ainsophyao has quit [*.net *.split] 03:13:03 -!- MakMorn has quit [*.net *.split] 03:13:53 -!- MakMorn has joined ##crawl-dev 03:15:28 -!- djanatyn has quit [Read error: Operation timed out] 03:18:13 -!- ainsophyao has joined ##crawl-dev 03:18:28 -!- sumguy has quit [Quit: Page closed] 03:21:48 -!- Xenobreeder has quit [Quit: Page closed] 03:23:53 -!- nonethousand_ is now known as nonethousand 03:24:24 -!- gammafunk_ has quit [Quit: gammafunk_] 03:25:13 -!- SkaryMonk2 has joined ##crawl-dev 03:26:04 -!- SkaryMonk1 has quit [Ping timeout: 265 seconds] 03:35:37 -!- buppy has joined ##crawl-dev 03:36:04 -!- simmarine has quit [Quit: Leaving] 03:36:30 is runed dragon armour a bug? 03:38:35 -!- cptwinky has quit [Remote host closed the connection] 03:38:44 depends how you got it 03:38:59 did it come off the corpse runed or was it just lying on the floor 03:39:25 it was on the floor and cursed 03:39:50 with no ego obviously 03:39:59 that's fine 03:40:29 dragon armour generates rarely but it's still just armour and can be cursed and enchanted like normal 03:40:43 afaik 03:40:58 but can it get egos? 03:41:49 i don't believe it can 03:42:05 but don't quote me on that 03:45:59 -!- indspenceable has joined ##crawl-dev 03:50:43 -!- indspenceable has quit [Ping timeout: 272 seconds] 03:51:29 -!- xFleury has joined ##crawl-dev 03:58:17 -!- ontoclasm has quit [Quit: Leaving.] 03:59:03 -!- yalue has quit [Ping timeout: 272 seconds] 04:02:24 -!- kait has quit [Ping timeout: 252 seconds] 04:12:42 -!- dtsund has quit [Quit: dtsund] 04:16:22 -!- _D_ has quit [Client Quit] 04:17:48 -!- kait has quit [Ping timeout: 240 seconds] 04:22:38 buppy: no egos on dragon armour 04:22:54 buppy: other than FDA I guess 04:23:07 buppy: where F is for Fairy 04:24:17 alefury: faerie doesn't get brands either 04:28:04 -!- Stelpa has quit [Remote host closed the connection] 04:37:53 -!- elliptic has quit [Read error: Operation timed out] 04:38:27 So I took another stab at fixing the MSVC build in 0.13. No luck. 04:38:28 xFleury: You have 1 message. Use !messages to read it. 04:38:43 Unfortunately the compilation warnings provide no hints either. :( 04:38:46 !messages 04:38:47 (1/1) kilobyte said (17h 46m 3s ago): crawl actually does use threads, but only while reading the database (and I'd like to do so for maps as well). This has support for Windows, although it can be disabled without big ill effects. 04:39:36 -!- odiv has quit [Ping timeout: 245 seconds] 04:41:48 kilobyte: crawl might be compiling SQLite with multithreaded support, but are you sure it actually uses it? :| 04:42:37 |amethyst: Some commands autofight sends are apparently getting lost here https://crawl.develz.org/tavern/viewtopic.php?f=9&t=9574 04:43:37 |amethyst: it's on CLAN. I don't know whether that's just a network problem on the user end or if something is wrong with crawl. 04:43:56 -!- elliptic has joined ##crawl-dev 04:44:10 oh, nice 04:44:19 elliptic: Some commands autofight sends are apparently getting lost here https://crawl.develz.org/tavern/viewtopic.php?f=9&t=9574 04:44:53 elliptic: I don't know whether it's just a network problem on the user end, or if something is wrong with how crawl sends keys 04:45:16 yeah, I've seen that happen in my own games occasionally when lag is high and I'm using tab a lot 04:45:49 I don't fully understand why it happens though :/ 04:46:03 it's kind of bad when it wastes scrolls 04:46:18 which can happen due to fr or vr being sent 04:51:23 -!- Blazinghand has quit [Read error: Connection reset by peer] 04:54:07 -!- kait has quit [Read error: Connection reset by peer] 04:55:51 -!- indspenceable has joined ##crawl-dev 05:00:00 -!- indspenceable has quit [Ping timeout: 245 seconds] 05:05:11 -!- Somefellow has quit [Ping timeout: 272 seconds] 05:23:48 -!- SkaryMonk2 has quit [Ping timeout: 240 seconds] 05:24:54 -!- SkaryMonk1 has joined ##crawl-dev 05:25:29 -!- Zen_ has quit [Client Quit] 05:30:24 -!- indspenceable has joined ##crawl-dev 05:34:28 -!- indspenceable has quit [Ping timeout: 240 seconds] 05:39:51 Napkin: more spam: https://crawl.develz.org/wiki/doku.php?id=dcss:issue:7633 05:51:53 hmm, how do I delete a page in the wiki? 05:52:48 -!- elliptic has quit [Quit: Leaving] 05:53:17 just removing the content, it seems 05:53:24 thanks for pointing it out, alefury 05:54:37 -!- Thyme is now known as Basil 05:55:29 -!- dead_ has quit [Ping timeout: 250 seconds] 05:58:42 -!- indspenceable has joined ##crawl-dev 05:59:05 -!- Bloax has quit [Quit: Error: Bloax not found] 05:59:55 -!- Roarke has quit [Quit: Page closed] 06:03:24 -!- indspenceable has quit [Ping timeout: 252 seconds] 06:15:40 -!- odiv has quit [Ping timeout: 260 seconds] 06:21:42 -!- Tarquinn has quit [Quit: Leaving] 06:23:49 -!- Ladykiller69 is now known as Drahbeg 06:28:02 -!- elliptic has joined ##crawl-dev 06:28:19 -!- indspenceable has joined ##crawl-dev 06:31:36 -!- JaGGedTK has quit [] 06:32:43 -!- indspenceable has quit [Ping timeout: 248 seconds] 06:33:33 xFleury: yes, and it's a part that uses SQL so without multithreading you get nice crashes 06:34:50 k ty 06:35:09 -!- Ragdoll has joined ##crawl-dev 06:38:11 -!- rossi has quit [Quit: WeeChat 0.4.2] 06:41:33 -!- Sgeo has quit [Read error: Connection reset by peer] 06:50:32 -!- C7ty has quit [Ping timeout: 256 seconds] 06:53:51 -!- jeanjacques has joined ##crawl-dev 06:54:20 -!- C7ty1 has quit [] 06:54:26 -!- indspenceable has joined ##crawl-dev 06:56:52 -!- John__ has quit [Quit: Page closed] 06:58:48 -!- indspenceable has quit [Ping timeout: 240 seconds] 07:09:48 -!- Basil has quit [Ping timeout: 256 seconds] 07:14:08 -!- Kilahine has quit [Ping timeout: 240 seconds] 07:14:48 -!- SkaryMonk has quit [Ping timeout: 240 seconds] 07:14:54 -!- SkaryMonk2 has joined ##crawl-dev 07:17:22 -!- indspenceable has joined ##crawl-dev 07:21:42 -!- indspenceable has quit [Ping timeout: 256 seconds] 07:23:53 -!- Quashie has quit [Ping timeout: 272 seconds] 07:24:19 -!- thened has quit [Quit: thened] 07:24:27 -!- SkaryMonk1 has quit [Ping timeout: 248 seconds] 07:24:57 -!- SkaryMonk has joined ##crawl-dev 07:26:48 -!- Basil has quit [Ping timeout: 240 seconds] 07:27:15 -!- syllogism has joined ##crawl-dev 07:36:58 |amethyst, was reading scrollback but I'm not flush with time, with regard to what you were talking about earlier (11 hours ago or so) as it applies to melee_attack, were you talking about the organization / execution path of the code, or the specific operation of some piece of code? 07:37:48 -!- magicpoints has quit [Ping timeout: 240 seconds] 07:38:53 -!- indspenceable has joined ##crawl-dev 07:40:43 -!- Vandal has joined ##crawl-dev 07:43:20 -!- indspenceable has quit [Ping timeout: 245 seconds] 07:44:35 -!- ystael has quit [Ping timeout: 245 seconds] 07:45:36 no news on rhf getting 0.13 right? 07:45:52 <|amethyst> Cryp71c: mostly the fact that things still happen in a different order for monsters and players, and things are subtly dependent on those orders 07:45:52 |amethyst: You have 1 message. Use !messages to read it. 07:46:17 <|amethyst> Cryp71c: so that nearly every time we try to change something, we introduce a new bug 07:47:07 <|amethyst> Cryp71c: in other words, nothing you introduced... and it's better than it was before, where none of the code was common 07:47:22 |amethyst, ah..yeah. I tried to fix what I could when I wrote it but most remained in the same order they were in (or roughly so) as pre-melee_attack and I've been lazy in fixing the remaining issues in a timely manner. 07:47:42 I'll get a checkout of the code and try and get back into it...see what more I can do. 07:54:36 -!- Quashie_ has quit [Quit: Leaving] 08:00:04 -!- newbie is now known as Guest82088 08:00:54 -!- indspenceable has joined ##crawl-dev 08:03:25 -!- trckry has quit [Ping timeout: 246 seconds] 08:05:02 <|amethyst> !seen joosa 08:05:03 I last saw joosa at Wed Oct 2 04:22:24 2013 UTC (2w 8h 42m 39s ago) saying '|amethyst: I'll try to take a look today' on ##crawl-dev. 08:05:14 <|amethyst> absolutego: nope 08:05:27 -!- indspenceable has quit [Ping timeout: 265 seconds] 08:08:12 cln is rather spiky 08:08:21 -!- PepeRC2 has quit [Ping timeout: 265 seconds] 08:09:36 <|amethyst> should have TZer0 watch the system while that's happening 08:09:45 <|amethyst> not sure if it's RAM, CPU, network, or what 08:10:07 <|amethyst> it's odd if it's happening right now, since there aren't that many users on 08:10:48 <|amethyst> cszo has been quite loaded, but it also has the tournament scripts 08:11:33 |amethyst: Any chance you could walk me through the changes you made the map parser since 0.12, so that I might able to get MSVC build working again? I dunno if I have time to poke in the dark. :S 08:12:21 <|amethyst> That I made? I added the ORDER: directive, and rebuilt with bison 2.5 (it had been 2.4.1 when mumra built it) 08:12:46 Could you link me the commit for that? And do if anyone else made any changes? 08:13:27 <|amethyst> %git :/ORDER 08:13:30 07|amethyst02 * 0.13-a0-2949-g26c9fab: Document ORDER: 10(6 weeks ago, 1 file, 7+ 0-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=26c9fab5300f 08:13:37 <|amethyst> hm 08:14:35 |amethyst: any progress on luajit? 08:14:35 <|amethyst> %git f94d332 08:14:36 07|amethyst02 * 0.13-a0-2946-gf94d332: Add a new "ORDER:" map header for sorting game-mode maps. 10(6 weeks ago, 9 files, 41+ 7-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=f94d332ad4a3 08:14:40 <|amethyst> that's the one that added it 08:14:55 <|amethyst> %git 7837b113 08:14:55 07|amethyst02 * 0.13-a0-2948-g7837b11: Regenerate prebuilt yaccage. 10(6 weeks ago, 3 files, 1994+ 1904-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=7837b113668f 08:15:07 <|amethyst> that's the one that rebuilt the prebuilt .cc file 08:15:45 <|amethyst> I would recommend trying to regenerate that with different versions of bison to see if it changes anything 08:15:49 -!- jeffro has quit [Ping timeout: 272 seconds] 08:15:51 <|amethyst> what was the exact error you got? 08:16:03 -!- orionstein_away is now known as orionstein 08:16:17 -!- slifty has quit [Quit: Leaving...] 08:16:41 <|amethyst> hasufell: haven't had a chance to do anything else with it, and I'd like to discuss it with a few more devs 08:16:45 -!- rapierx has quit [Ping timeout: 250 seconds] 08:16:48 "input flex scanner failed" failed during parsing of atlas.des (iirc) 08:17:03 <|amethyst> xFleury: oh, right, without any more of an error message 08:17:08 -!- nonethousand has quit [Read error: Connection reset by peer] 08:17:12 in release mode, the game runs, but there's bits of corruption on the map 08:17:22 tiles missing their graphic 08:18:10 and mumra said he had a few crashes when moving around in wizmode 08:19:50 <|amethyst> hasufell: could you post a report on mantis, with the stuff we talked about earlier (the suggestion to prevent dgamelaunch builds, at least 64-bit ones, from using luajit) 08:20:16 <|amethyst> hasufell: that way when I do have time, I'll have something I can point other people at 08:20:43 |amethyst: yeah, but I am lazy right now. I'll wait until I am not lazy anymore 08:20:53 <|amethyst> :) 08:21:14 <|amethyst> I know that feeling 08:21:28 |amethyst: also, any chance the tests get fixed? 08:21:44 <|amethyst> tests? 08:22:03 -!- indspenceable has joined ##crawl-dev 08:22:18 ugh, just thought of something; maybe the "input flex scanner failed" is just the Working Directory being wrong; there's a directory tree mirroring "source/msvc/des" that mirrors "source/dat/des", except it has no files in it. 08:22:34 |amethyst: https://crawl.develz.org/mantis/view.php?id=6121 08:22:50 <|amethyst> xFleury: and altar/altar.des would be alphabetically first.. 08:22:55 actually, I am not sure if it is still present anymore 08:23:58 <|amethyst> hasufell: I haven't tried make test with a tiles build 08:24:04 <|amethyst> hasufell: so that part may still be broken 08:24:53 <|amethyst> hasufell: the errors mentioned in your comment have probably been fixed---kilobyte runs the tests fairly frequently I believe (definitely the stress test, maybe others) 08:25:08 <|amethyst> s/maybe/probably/ 08:25:25 |amethyst: but no one commented on the bug 08:25:54 since I am still lazy I am not sure if I'm going to run another test 08:26:24 -!- indspenceable has quit [Ping timeout: 252 seconds] 08:28:35 <|amethyst> I can confirm 08:30:52 <|amethyst> There :) 08:32:21 -!- araganzar has quit [Ping timeout: 250 seconds] 08:32:22 <|amethyst> !lg * t br.enter=bailey 08:32:22 Unknown field: br.enter 08:32:29 <|amethyst> !lm * t br.enter=bailey 08:32:29 208. [2013-10-16 13:28:43] jeanjacques the Covert (L11 SpNe) entered a Bailey on turn 12031. (Orc:3) 08:32:41 <|amethyst> !lg * t place=bailey s=map 08:32:42 22 games for * (t place=bailey): 5x bailey_polearm_1, 5x bailey_axe_2, 3x bailey_polearm_4, 3x bailey_polearm_2, 3x bailey_polearm_5, bailey_axe_3, bailey_axe_4, bailey_axe_minmay_hex_keep 08:32:45 <|amethyst> hm 08:33:00 <|amethyst> console crawl -test gives me /home/neil/src/crawl/crawl-ref/source/dat/des/portals/bailey.des:371: Couldn't find a map for bailey_axe 08:33:29 -!- absolutego has quit [Ping timeout: 265 seconds] 08:33:51 -!- absolutego has joined ##crawl-dev 08:34:45 <|amethyst> oh, never mind 08:35:02 <|amethyst> re-running crawl without -test fixed that 08:35:12 <|amethyst> I guess it was something about my des cache 08:35:47 <|amethyst> but I did confirm that tiles crashes on -test 08:37:02 -!- Kilahine has quit [Ping timeout: 241 seconds] 08:43:09 -!- indspenceable has joined ##crawl-dev 08:44:10 -!- Bloax has joined ##crawl-dev 08:44:24 |amethyst: 1. I never run -test in tiles (does it even support that?), 2. the test harness runs --builddb first because of a known but hard to fix race condition 08:44:52 this race bit us just before the tourney, too 08:45:55 because of 2., I don't catch errors when building the db with special flags 08:47:31 -!- indspenceable has quit [Ping timeout: 246 seconds] 08:48:58 <|amethyst> kilobyte: 1. it looks like no, it doesn't, so I retitled the bug 08:49:03 <|amethyst> kilobyte: it tries, but crashes 08:49:29 <|amethyst> for 2. that explains why my first attempt failed 08:53:51 Hiermare the Practitioner (L9 DrHe) (D:9) 08:54:13 <|amethyst> hm, CAO wbeitles freeze 08:54:22 <|amethyst> s/beit/ebtit/ 08:55:20 Zermako the Insei (L5 DrMo) (D:3) 08:55:42 <|amethyst> oh, it's in state D 08:55:50 -!- Kilahine has quit [Ping timeout: 245 seconds] 08:55:56 <|amethyst> so can't even strace it to see what's happening 08:56:04 <|amethyst> ("it" = webtiles server process) 08:56:17 <|amethyst> err, what? 08:56:21 <|amethyst> now it's state T???? 08:56:49 -!- brian_ has quit [Quit: leaving] 08:56:55 ... how can a process spontaneously go into T? 08:57:02 <|amethyst> it was the one I tried to strace 08:57:10 <|amethyst> but I killed strace when I saw it was hanging 08:57:14 <|amethyst> let me try reattaching 08:57:29 ah, strace might use STOP for attaching 08:58:55 <|amethyst> not quite, but ps says 08:58:57 <|amethyst> T stopped, either by a job control signal or because it is 08:58:57 <|amethyst> being traced 08:59:05 <|amethyst> I hadn't realised the latter part 09:00:07 <|amethyst> accesses to the chroot fs are going rather slow 09:00:19 <|amethyst> but memory isn't quite full 09:01:40 Webtiles server stopped. 09:01:56 -!- indspenceable has joined ##crawl-dev 09:02:05 Webtiles server started. 09:05:21 doh, right, I'm an idiot. Forgot that ptrace stops the process every time it notifies its supervising process, so that the latter has a chance to handle the notification and possibly mess with the tracee's state 09:06:03 -!- indspenceable has quit [Ping timeout: 240 seconds] 09:11:47 -!- xiayu has quit [Quit: Leaving] 09:12:44 -!- LexAckson has joined ##crawl-dev 09:18:43 -!- gowby has quit [Ping timeout: 250 seconds] 09:18:48 -!- mizu_no_oto has quit [Quit: ["Textual IRC Client: www.textualapp.com"]] 09:20:27 -!- rapierx has quit [Ping timeout: 250 seconds] 09:21:28 -!- indspenceable has joined ##crawl-dev 09:22:27 -!- jeffro has joined ##crawl-dev 09:23:41 "Cannot find data file 'defaults/autopickup_exceptions.txt' anywhere, aborting" 09:23:51 -!- SkaryMonk1 has joined ##crawl-dev 09:23:54 That path should be 'dat/defaults/autopickup_exceptions.txt' 09:24:12 Shouldn't it? I mean, Crawl.exe isn't running within 'dat' folder. 09:24:13 -!- MaxFrost has quit [Ping timeout: 272 seconds] 09:24:40 -!- SkaryMonk has quit [Ping timeout: 256 seconds] 09:25:48 -!- indspenceable has quit [Ping timeout: 240 seconds] 09:25:48 -!- RiotInferno has quit [Ping timeout: 240 seconds] 09:28:57 Oh nevermind, I see how this works. 09:33:21 -!- omnirizon has quit [Ping timeout: 272 seconds] 09:38:00 -!- RiotInferno has quit [Quit: Leaving.] 09:39:12 -!- indspenceable has joined ##crawl-dev 09:40:48 -!- Basil has quit [Ping timeout: 240 seconds] 09:42:26 Could one of the devs here do me a favor and examine "cheibriados_altar_2" and tell me if that file is wrongly formatted at all? It's crashing on MSVC on the 'shuffle_array' function, something to do with the map_lines. 09:42:35 map_lines, in dgn_map_add_transform 09:43:39 -!- indspenceable has quit [Ping timeout: 248 seconds] 09:44:20 Anyone develop in windows? 09:45:06 It's in overflow.des, line 143 09:45:25 The NSUBST or SUBST tags seem to be where the crash is. 09:45:38 it's happening a lot that the terminal doesn't refresh until a press a key (and then with some lag), could this be a local issue? 09:45:45 SUBST: 1 : 1:50 2:30 3:10 4:10 09:45:47 NSUBST: 4 = 1:1 / 1:2 / 1:3 09:46:12 maybe it's just an illusion and it's good old lag 09:46:17 -!- ontoclasm has joined ##crawl-dev 09:46:58 What do those functions do? 09:50:00 substitute things 09:50:44 the specifics are in the vault syntax documentation 09:51:59 they can be used to randomize vaults 09:56:17 -!- indspenceable has joined ##crawl-dev 09:56:38 A small percentage of the uses of SUBST,NSUBST are throwing std::vector out of range exceptions on MSVC. 09:57:06 fedhas_altar_fruit_tree, demons_altar, nooodl_cocytus_jagged_cave, etc 09:58:23 These vaults were likely edited between 0.12 and 0.13, and I think whoever did the editing didn't use those functions correctly. 10:01:03 -!- indspenceable has quit [Ping timeout: 260 seconds] 10:01:07 -!- ackack has joined ##crawl-dev 10:02:28 -!- bmfx has quit [Ping timeout: 240 seconds] 10:05:28 -!- Kilahine has quit [Ping timeout: 256 seconds] 10:06:35 -!- Valarioth has quit [Quit: Page closed] 10:06:46 -!- Kilahine has quit [Max SendQ exceeded] 10:09:26 -!- bmfx_ has quit [Ping timeout: 256 seconds] 10:10:08 -!- ackack has quit [Quit: Leaving] 10:12:22 -!- Blazinghand has joined ##crawl-dev 10:12:52 -!- indspenceable has joined ##crawl-dev 10:13:36 Or, more accurately, they used them in a way that *shouldn't* have caused a crash; have a look at random.h line 118 10:14:13 shuffle_array(vector &vec), that shouldn't crash if the vector is empty, but it does; because: 10:14:19 shuffle_array(&vec[0], vec.size()); 10:14:24 is its 1st line, lol 10:17:24 -!- indspenceable has quit [Ping timeout: 265 seconds] 10:19:05 -!- Blazinghand has quit [Quit: Leaving] 10:19:37 SWEEEEEEETTTTT I've fixed it! MSVC is once again working. :D 10:20:14 This shuffle nonsense, and the rarity of vaults that relied on it not crashing, also explains the random crashes while exploring around in wizmode that mumra brought up. 10:21:08 -!- Kilahine has quit [Ping timeout: 240 seconds] 10:22:35 -!- jeffro has quit [Ping timeout: 248 seconds] 10:23:43 The server is shutting down. 10:23:43 Bloax: You have 1 message. Use !messages to read it. 10:23:45 what an orb run 10:23:49 !messages 10:23:51 (1/1) bmfx said (4h 55m 27s ago): you left a rather scary ghost in my volcano 10:24:17 !tell |amethyst Can you add "if(vec.size() == 0) return;" to random.h line 118 to fix an out-of-bounds exception occuring in void shuffle_array(vector &vec) when vec is empty? 10:24:18 xFleury: OK, I'll let |amethyst know. 10:26:31 !tell mumra I think I've found your random crash culprit; line 118 of random.h assumes `vec` of void shuffle_array(vector &vec) is never empty, when several vaults in 0.13 will cause that function be called with an empty `vec`. 10:26:32 xFleury: OK, I'll let mumra know. 10:27:24 -!- Morg0th has quit [Quit: leaving] 10:28:55 -!- Mandevil has quit [Quit: leaving] 10:30:05 -!- the_glow has quit [Ping timeout: 272 seconds] 10:30:25 -!- indspenceable has joined ##crawl-dev 10:31:02 -!- magistern has quit [Quit: This computer has gone to sleep] 10:33:53 -!- demi_ has quit [Client Quit] 10:34:35 -!- indspenceable has quit [Ping timeout: 245 seconds] 10:34:54 Chaos javelins always mulch after hit. by Ragdoll 10:40:43 -!- radinms has quit [] 10:45:48 wontolla (L27 GrWz) ERROR: range check error (-1 / 80) (Zot:2) 10:45:49 -!- LordLovebone has quit [Ping timeout: 250 seconds] 10:46:18 <|amethyst> xFleury: &vec[0] is failing for a size-zero vector? 10:46:18 |amethyst: You have 2 messages. Use !messages to read them. 10:47:25 <|amethyst> !tell ontoclasm You could only drag us down by winning (we might have to kick kilobyte off the team next tournament for being too good :) 10:47:26 |amethyst: OK, I'll let ontoclasm know. 10:47:48 <|amethyst> !tell ontoclasm Added 10:47:49 |amethyst: Yes, vec[0] throws an exception if vec is empty. 10:47:49 |amethyst: OK, I'll let ontoclasm know. 10:48:04 thanks :D 10:48:04 ontoclasm: You have 2 messages. Use !messages to read them. 10:48:04 |amethyst: And in Visual Studio 2005 and 2008 already do bounds-checking on operator[] by default, in both debug and release builds. 10:48:09 <|amethyst> xFleury: ah, yeah, I see 10:48:25 |amethyst: So this will crash on MSVC regardless of Debug or Release. 10:48:36 <|amethyst> xFleury: vec.data() would work but is C++11 only, so I'll do the change you suggest 10:49:32 -!- scummos^ has quit [Remote host closed the connection] 10:49:54 |amethyst: btw, it looks like Flex/Bison wasn't the problem, it's a path issue that's causing parsing to fail. I've got a setup that's working for me at the moment, but I haven't narrowed down teh precise change to the path that fixed it. 10:50:25 -!- cptwinky has quit [Ping timeout: 245 seconds] 10:52:10 -!- indspenceable has joined ##crawl-dev 10:52:23 |amethyst: I managed to win once, and only because of a RNG bug 10:52:47 if you're at 20 hp, next to 6 acid blobs 1 azure jelly, you're not supposed to be alive 10:53:01 <|amethyst> xFleury: btw, line numbers aren't useful unless you tell us which version :) 10:53:19 kilobyte: sometimes Xom is amused to let you live in otherwise certain death scenarios. 10:53:28 -!- P1x3lF1r3 has quit [Ping timeout: 240 seconds] 10:54:01 |amethyst: It's from the public release source of 0.13.0. 10:54:07 template 10:54:08 void shuffle_array(vector &vec) 10:54:10 { 10:54:11 <|amethyst> ah, that's why it didn't line up for me 10:54:12 if(vec.size() == 0) return; 10:54:13 shuffle_array(&vec[0], vec.size()); 10:54:15 } 10:54:21 This is the function, and it includes my fix. 10:54:34 <|amethyst> Yeah, I found the line you were talking about, but it's 123 in trunk 10:54:54 <|amethyst> also, don't use tabs :) 10:55:23 -!- indspenc_ has joined ##crawl-dev 10:56:25 03|amethyst02 07* 0.14-a0-267-gf2b8b55: Don't crash when shuffling length-0 vectors (xFleury) 10(5 minutes ago, 1 file, 3+ 1-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=f2b8b55cca63 10:56:40 -!- indspenceable has quit [Read error: Connection reset by peer] 10:56:52 Oh oops, yah, I can set MSVC to use spaces instead of tabs, and have done that in the past. But haven't since upgrading to MSVC 2012. 10:57:03 thx for reminder :P 10:57:39 Cherry-picked 1 commit into stone_soup-0.13 10:57:41 IIRC the standard says you can request such pointers, it's only dereferencing them what's an undefined behaviour 10:57:49 Flun (L14 TrCK) ERROR in 'libutil.cc' at line 1020: screen write out of bounds: (1,8) into (80,7) (Lair:3) 10:57:52 Flun (L14 TrCK) ERROR in 'libutil.cc' at line 1020: screen write out of bounds: (1,8) into (80,7) (Lair:3) 10:57:54 <|amethyst> kilobyte: the problem is you can't write &vec[0] without writing vec[0] 10:57:57 kilobyte, well, that becomes questionable with operator overloading 10:58:00 Flun (L14 TrCK) ERROR in 'libutil.cc' at line 1020: screen write out of bounds: (1,8) into (80,7) (Lair:3) 10:58:03 <|amethyst> kilobyte: and vectors are allowed to do bounds checking 10:58:05 !lm flun crash -log 10:58:06 1. Flun, XL21 HaHu, T:48089 (milestone): http://dobrazupa.org/morgue/Flun/crash-Flun-20130516-202724.txt 10:58:12 because &a[i] is no longer just (a + i) 10:58:17 ...not that one. 10:58:22 Flun (L14 TrCK) ERROR in 'libutil.cc' at line 1020: screen write out of bounds: (1,8) into (80,7) (Lair:3) 10:58:28 http://dobrazupa.org/morgue/Flun/crash-Flun-20131016-155759.txt 10:58:33 <|amethyst> Zaba: well, vectors except for vector guarantee contiguous allocation 10:58:35 Flun (L14 TrCK) ERROR in 'libutil.cc' at line 1020: screen write out of bounds: (1,8) into (80,7) (Lair:3) 10:59:34 <|amethyst> Zaba: so if v[0] exists then &v[0] is a pointer to an array of size v.capacity() 11:00:12 Flun (L14 TrCK) ERROR in 'libutil.cc' at line 1020: screen write out of bounds: (1,8) into (80,7) (Lair:3) 11:00:17 Flun (L14 TrCK) ERROR in 'libutil.cc' at line 1020: screen write out of bounds: (1,8) into (80,7) (Lair:3) 11:00:19 <|amethyst> hmm 11:00:30 Flun (L14 TrCK) ERROR in 'libutil.cc' at line 1020: screen write out of bounds: (1,8) into (80,7) (Lair:3) 11:00:32 -!- SkaryMonk1 has left ##crawl-dev 11:00:32 <|amethyst> that's the wrapcprintf stuff I changed 11:00:38 -!- Flun has joined ##crawl-dev 11:00:46 <|amethyst> this happened with Turgon too, but in his case it was crashing before that 11:00:48 I'm having some sort of bug where I can't annotate levels 11:01:00 I press '!' and it brings me to the game menu 11:01:22 we noticed :P 11:01:31 Sizzell Flun (L14 TrCK) ERROR in 'libutil.cc' at line 1020: screen write out of bounds: (1,8) into (80,7) (Lair:3) 11:01:34 oh? 11:01:37 ah 11:01:47 for future reference, "brings me to the game menu" means it crashed :P 11:01:56 i see 11:01:58 <|amethyst> Flun: Could you file a bug? Turgon has the same problem earlier with long annotations, but I can't reproduce it 11:02:06 <|amethyst> s/has/had/ 11:02:11 -!- alefury has quit [Quit: ChatZilla 0.9.90.1 [Iceweasel 24.0/20130918041159]] 11:02:13 sure 11:02:30 <|amethyst> Flun: also, include your config 11:02:40 -!- gregunderscorem has quit [Quit: Page closed] 11:02:42 ok 11:02:43 <|amethyst> and let me get a save backup while you're not playing 11:03:01 <|amethyst> !lm Flun crash x=v 11:03:02 9. [2013-10-16 16:00:30] [v=0.13.0] Flun the Grappler (L14 TrCK) ERROR in 'libutil.cc' at line 1020: screen write out of bounds: (1,8) into (80,7) (Lair:3) 11:03:19 -!- lobf has quit [Quit: lobf] 11:04:01 -!- UrQuan has quit [Ping timeout: 250 seconds] 11:04:31 <|amethyst> Flun: http://dobrazupa.org/saves/Flun-crash-20131016.cs is the link to you save (you won't be able to access it though)---include that in the bug report 11:04:36 <|amethyst> s/you save/your save/ 11:04:39 thanks neil 11:05:13 <|amethyst> Let me find the link to the tavern thread 11:05:25 <|amethyst> https://crawl.develz.org/tavern/viewtopic.php?f=9&t=9429 11:05:31 okay, i'll put those links in 11:07:03 -!- ussdefiant has joined ##crawl-dev 11:07:39 03elliptic02 07* 0.14-a0-268-gffff52d: Fix vampire knights coming with water elementals 3/4 of the time (N78291). 10(2 minutes ago, 1 file, 1+ 1-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=ffff52d87250 11:07:41 -!- xnavy has quit [Read error: Operation timed out] 11:08:09 Cherry-picked 1 commit into stone_soup-0.13 11:09:30 Good bug. 11:10:04 -!- Beohg has quit [Quit: Page closed] 11:10:22 -!- P1x3lF1r3 has joined ##crawl-dev 11:11:31 sounds thematic 11:11:38 hm I was wondering why they came with water elementals 11:13:13 Annotating a level crashes game by flun 11:13:52 -!- elliptic has quit [Quit: Leaving] 11:15:20 did anybody see my ridiculous scythe crash 11:16:46 ontoclasm: what did you do, attempt to add scythe cleaving? 11:16:47 <_< >_> 11:18:47 ??git 11:18:47 git[1/2]: Readonly: git clone git://gitorious.org/crawl/crawl.git Read/Write: git@gitorious.org:crawl/crawl.git Web interface: http://git.develz.org/?p=crawl.git or http://gitorious.org/crawl/crawl 11:20:17 ??git[2] 11:20:17 git[2/2]: https://crawl.develz.org/wiki/doku.php?id=dcss:help:git 11:20:50 -!- alefury has joined ##crawl-dev 11:21:20 Cryp71c: see also http://s-z.org/neil/git/?p=crawl.git;a=tree;f=crawl-ref/docs/develop/git;h=49f7f830b8774bc9e812de8e5cbd7a69176cfe54;hb=HEAD 11:27:17 -!- ketsa has quit [Quit: ChatZilla 0.9.90 [Pale Moon 24.0.2/20130925021446]] 11:28:23 -!- Kilahine has quit [Quit: Quitte] 11:29:57 Grunt: no, i once again managed to find a bug involving removing hydra heads 11:30:03 i guess this is like, my specialty 11:30:33 The 27-headed ontoclasm hydra comes into view. 11:33:28 if you remove all of my heads i just respawn, boris style, with one more head the next time 11:36:35 -!- Kalir has quit [Changing host] 11:37:49 -!- Garhauk has quit [Ping timeout: 250 seconds] 11:37:49 -!- timbw_ has quit [Ping timeout: 250 seconds] 11:38:07 -!- Vizer_ has quit [Ping timeout: 265 seconds] 11:38:59 lol, you're making me want a unique monster named after myself; "Fleury - the unique ice wraith". 11:39:27 screw, Fennar, if you have low rC, it's me you need to worry about 11:40:46 ??Freezing wraith 11:40:46 freezing wraith[1/1]: Hits you. Freezes you. Doesn't drain you. Slow. 11:46:59 -!- ussdefiant has quit [Ping timeout: 260 seconds] 11:50:15 -!- the_glow has quit [Ping timeout: 272 seconds] 11:50:50 -!- elliptic has joined ##crawl-dev 11:52:29 uh oh, looks like unrand tile offset are screwed up 11:52:59 ontoclasm: do you happen to know what happens if the tiles get reordered in a stable release? 11:53:02 -!- Stormphoenix has quit [Quit: Page closed] 11:54:09 -!- Cryp71c_ has joined ##crawl-dev 11:54:30 -!- Cryp71c__ has joined ##crawl-dev 11:55:23 -!- ainsophyao has quit [Remote host closed the connection] 11:55:35 -!- VolteccerJack has quit [Ping timeout: 272 seconds] 11:55:36 -!- Pulseman is now known as VolteccerJack 11:58:26 -!- Cryp71c has quit [Ping timeout: 264 seconds] 11:58:44 * TZer0 watches clan 11:59:07 -!- Cryp71c_ has quit [Ping timeout: 272 seconds] 11:59:14 Hmm... I'm actually having some issues in regards to CPU-usage.. 11:59:22 %Cpu(s): 87.0 us, 5.6 sy, 0.0 ni, 6.6 id, 0.3 wa, 0.0 hi, 0.3 si, 0.0 st 11:59:26 this might explain it.. 11:59:58 absolutego: I think this is because the tournament is running 12:00:27 -!- ackack has joined ##crawl-dev 12:00:42 it's possible that something's wrong on my end too because i haven't seen anything this bad 12:00:53 -!- ackack has quit [Client Quit] 12:00:54 Quick question.. 12:00:55 just 28 games running is using 87% of the cpu? 12:01:08 why is there no one playing on the finnish server? 12:01:15 TZer0: it doesn't have 0.13 12:01:23 ahhhh 12:01:35 I think it was compression that was happening 12:01:35 (so you can't play tournament games there) 12:01:37 hmm 12:02:02 TZer0: i tried logging out then logging back in and as i was typing my username i got a bunch of letters missing... is it possible packets are getting lost? you can also notice this in-game (some actions do not register) 12:02:37 hold on, I noticed that you complained around the hour-change last time 12:02:50 I'll try moving the compression of replays to the middle of the night 12:03:16 Usage has gone down to about 50% 12:03:31 but generally less 12:03:37 bzip was eating like 40%+ 12:04:40 and I can safely assume that very few people will be playing in the middle of the night to notice bzip eating the CPU 12:05:40 Doesn't generating enormous directory pages cause strain on the webserver? 12:06:00 It seems like I could easily DoS a crawl web-server by requesting the directory of the morgue on a few seperate threads. 12:06:04 -!- Xenobreeder has quit [Quit: Page closed] 12:06:22 -!- blabber has joined ##crawl-dev 12:07:50 Hmmm 12:08:00 indeed.. 12:13:23 -!- blabber has quit [Quit: leaving] 12:13:41 -!- blabber has joined ##crawl-dev 12:14:04 On just about every crawl webserver I've ever played on, browsing the top-level morgue directory is useless anyways; it's too large to browse 12:14:25 I think it should be disabled, and only allow the individual user directories to have directory browsing. 12:15:17 xFleury: what's the problem there? It works for me... 12:15:57 -!- dead_ has quit [Ping timeout: 250 seconds] 12:16:01 kilobyte: Doesn't generate the page http://crawl.akrasiac.org/rawdata/ cause significant load on the web-server? 12:16:35 -!- rast- has joined ##crawl-dev 12:16:39 kilobyte: And what purpose is there to browse that *entire* directory? Arn't users just have a particular subdirectory anyways? 12:16:54 -!- indspenc_ has quit [Remote host closed the connection] 12:17:08 -!- rast has quit [Ping timeout: 240 seconds] 12:17:13 disk IO; it stats a gazillion dirs 12:17:15 -!- rast- is now known as rast 12:17:20 -!- indspenceable has joined ##crawl-dev 12:17:26 hmm, indeed it takes a while 12:17:52 useful if you want to find how an user is named or get some statistics 12:18:26 at least I grabbed that whole data via a script a few times 12:19:45 idea: what if we shown just the name rather than the date of last write? 12:19:46 If disabling it is not option, maybe just cache it by at least an hour, preferably a 24+ hours. 12:19:51 Stable (0.13) branch on crawl.akrasiac.org updated to: 0.13.0-9-gbc74161 12:20:20 (not sure what filesystem is in use, though) 12:21:55 -!- indspenceable has quit [Ping timeout: 272 seconds] 12:28:17 -!- wolfram has quit [Quit: Page closed] 12:35:59 Unstable branch on crawl.akrasiac.org updated to: 0.14-a0-268-gffff52d (34) 12:36:41 -!- Gotham has joined ##crawl-dev 12:36:46 -!- imantor has quit [Quit: Page closed] 12:41:40 -!- randart has quit [Ping timeout: 246 seconds] 12:47:13 -!- orionstein is now known as orionstein_away 12:47:29 -!- Drahbeg has joined ##crawl-dev 12:47:37 Why not rename asmodeus 12:47:42 to Asmodeus Mordzart 12:48:37 and give him mesmerize 12:48:41 -!- heteroy has quit [Ping timeout: 265 seconds] 12:48:46 with special text 12:49:41 Wolfgang Asmodeus 12:50:00 allright mesmerize and hellhounds 12:50:19 -!- tkappleton has quit [Ping timeout: 248 seconds] 12:50:28 -!- gregunderscorem has quit [Client Quit] 12:51:59 lantea seems to shut down a lot 12:52:08 and change his sceptre to a conductors baton 12:52:35 -!- Leafsnail has quit [Ping timeout: 272 seconds] 12:52:41 -!- indspenceable has joined ##crawl-dev 12:57:39 03kilobyte02 07* 0.14-a0-269-g4ceef45: A quote for SpAs. 10(4 days ago, 1 file, 5+ 0-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=4ceef457b6db 12:57:39 03kilobyte02 07* 0.14-a0-270-gfc2aeb4: Fix unrand tile mismatches between architectures. 10(44 minutes ago, 1 file, 1+ 1-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=fc2aeb4d5915 13:02:28 <|amethyst> kilobyte: re not statting things in the directory list, didn't Napkin have a CGI for that? 13:02:45 <|amethyst> maybe we could drop that on on the other servers 13:04:06 well, it's simple 13:04:25 -!- ainsophyao has joined ##crawl-dev 13:04:47 CDO shows names and timestamps, which stats everything 13:05:10 hrm, this data expired from CAO's page cache already :/ 13:05:26 no, no timestamps 13:05:52 http://crawl.develz.org/cgi-bin/ttyrecs?Napkin 13:05:58 https://crawl.develz.org/morgues/trunk/ 13:06:15 <|amethyst> oh, right, it was the automated ttyrecs fetching that was the big problem 13:06:27 <|amethyst> since there are so many more of those than players 13:06:33 because it was listing all ttyrecs first 13:06:45 http://crawl.develz.org/cgi-bin/ttyrecs?Napkin - very quick 13:07:09 http://crawl.develz.org/ttyrecs.cgi-bin.txt <- that's all 13:07:25 CAO takes ages to produce http://crawl.akrasiac.org/rawdata/; CDO has it split between versions and is a good deal faster, but still takes a while from cold cache 13:08:09 -!- mizu_no_oto has quit [Quit: ["Textual IRC Client: www.textualapp.com"]] 13:08:17 heh, CAO takes ages even with hot cache somehow... 13:09:35 probably cause a few of us clicked those links ;) 13:09:43 and brought the server to a halt xD 13:09:56 -!- Blazinghand has joined ##crawl-dev 13:10:16 we should use a external directory indexer with cache 13:10:27 <|amethyst> possibly the scoring scripts are wrecking the cache 13:11:59 -!- lobf has quit [Client Quit] 13:15:39 <|amethyst> anyway, I must go again. Midterms season 13:15:48 good luck 13:15:59 poor students, poorer teachers 13:17:01 kilobyte: maybe it keeps statting? 13:17:35 CDO uses ext4, which stores names in directory blocks, the inode is read only for the timestamp 13:18:07 SamB: well, inodes should be in the page cache after the first of us made it stat the whole lot 13:18:09 <|amethyst> hmm 13:18:12 Has there ever been any attempts to profile the web-server? 13:18:20 <|amethyst> CAO's chroot is ext4 as well 13:18:41 kilobyte: point 13:18:48 <|amethyst> but I note it doesn't have diratime or relatime 13:19:22 CDO is sane, with noatime 13:19:43 but stat doesn't bump atime, and how much resources will the atime on the dir(s) take or do I have no idea what that's for 13:19:46 * kilobyte doesn't understand why {,rel}atime is still a thing; even worse, relatime is currently the default. 13:20:09 kilobyte: I think there was a patch to do atime only once a day or so 13:20:16 SamB: right, true 13:20:21 SamB: yes, it's called "relatime" 13:20:22 <|amethyst> Oh, I forgot relatime is default now 13:20:36 cdo has only ext3 13:20:56 oh noes 13:21:02 it's ext4, oops 13:21:07 SamB: which still means writing several gigs or tens of gigs per day for no reason 13:21:22 <|amethyst> and CAO's chroot is relatime, I just didn't notice. And barrier=1 13:21:59 -!- Gotham has quit [Ping timeout: 272 seconds] 13:22:18 -!- Snufkin has joined ##crawl-dev 13:23:06 <|amethyst> okay, going for real now :) 13:23:10 Amehtyst 13:23:12 Wait! 13:23:16 Amethyst* 13:23:18 <|amethyst> yes? 13:23:19 SamB: without snapshots it "merely" adds some churn that can kill a SSD by itself in 10-20 years; with btrfs or LVM snapshots you lose that much of disk space per day 13:23:45 <|amethyst> Snufkin: make it quick 13:24:00 I wanted to voice a question: do you know anything about being locked out on webtiles due to save game in progress? 13:24:57 -!- st_ has quit [Read error: Connection reset by peer] 13:25:02 <|amethyst> you got disconnected but your crawl process hasn't noticed yet. Usually it resolves itself after 5-10 minutes but I just killed your process early 13:25:23 Thanks much, I'm sorry to inconvenience you! 13:25:26 <|amethyst> Assuming this was CAO you were talking about 13:25:30 Yep! 13:25:39 <|amethyst> cool beans; later 13:25:42 pce 13:25:52 -!- st_ has joined ##crawl-dev 13:25:53 -!- Flun has quit [Quit: Page closed] 13:33:05 -!- ainsophyao has quit [Remote host closed the connection] 13:34:02 -!- Gotham has joined ##crawl-dev 13:39:26 -!- VolteccerJack has quit [Ping timeout: 240 seconds] 13:39:26 -!- Pulseman is now known as VolteccerJack 13:40:05 -!- magicpoints has quit [Ping timeout: 272 seconds] 13:40:14 Quick question.. how do I avoid walking straight into zot traps as a Djinni? 13:42:19 * xFleury ponders if he was the only one who shrugged at relying on specific properties of the file-system for performance design decisions. 13:42:31 Debian builds of the trunk branch on crawl.develz.org updated to: 0.14-a0-270-gfc2aeb4 13:42:50 TZer0: 1. play stable, the tourney is going on, 2. same as any other race: get XP, pray 13:43:15 you could spam butterflies ever few steps but that's pretty grindy 13:43:53 And a quick way to get to the Abyss.. 13:44:18 alternatively, play any version between: 13:44:21 %git 0.13-a0-2461-g0deee0c 13:44:32 07kilobyte02 * 0.13-a0-2461-g0deee0c: Don't place boring traps randomly. 10(3 months ago, 1 file, 9+ 0-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=0deee0c0e46a 13:44:32 spoiler: I don't really care about the tournament since I'm still not *that* good 13:44:32 %git 0.13-a0-2468-g78e597d 13:44:32 07MarvinPA02 * 0.13-a0-2468-g78e597d: Do place random Zot traps 10(3 months ago, 1 file, 2+ 2-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=78e597d2ebe8 13:45:01 I'd love to play competitively on the tournaments, but I just can't bare the latency. 13:45:24 xFleury: where are you located? 13:45:32 Ontario, Cnada. 13:45:35 ah. 13:45:36 Canada* 13:45:50 Not quite far from where the server is hosted, TBH. 13:45:57 bear * 13:47:22 Perhaps once I finish up restoring MSVC support on Crawl 0.13.0 I'll spend a bit of time playing around with the webserver build. 13:49:13 -!- simmarine has joined ##crawl-dev 13:51:01 -!- eb has quit [Quit: I quit] 13:55:35 kilobyte: fwiw, I agree zot-traps should be left as a sorta "flavor" trap for select branches rather than just a random trap to be placed 13:56:51 -!- Quashie has quit [Ping timeout: 245 seconds] 13:58:56 -!- Cryp71c__ is now known as Cryp71c 13:59:29 -!- Snufkin has quit [Quit: Page closed] 14:00:07 Then again though, if it was left up to me, I'd bring back Nausea and try to making starving to death a real thing again. 14:03:09 "again"? 14:03:18 that implies that it was a real thing once 14:06:36 xFleury: starving to death is raaaaare 14:06:47 there are certain class which are capable of doing so 14:07:08 but only if you go to the Tomb or other places without enemies which can be eaten as a troll or similar 14:07:12 without enough backup food 14:07:47 -!- Stelpa has quit [Changing host] 14:08:09 As I recall, going berserk as Trog, suddenly "Very Hungry", munch down a contaminated chunk, "Nausea". No Cure Potions. Sit it out, "Starving". Eat food, now "Very Hungry" and Nausea again. 14:08:16 It was like a bad situation slowly getting worse, lol. 14:09:28 the problem there is "Sit it out," 14:09:51 there was no reason to stand still and waste turns and nutrition when you had nausea 14:10:39 Did you not regain HP while under Nausea? 14:10:47 I can't remember. 14:11:00 you regained HP under nausea... 14:11:10 you are getting it confused with sickness 14:11:29 Well, therin is the reason for standing still wasting turns. :P 14:11:49 Not always 100% after a berserk battle ;D 14:11:53 except nausea wasn't sickness 14:16:24 -!- Naruni has joined ##crawl-dev 14:23:48 -!- Morg0th has quit [Ping timeout: 245 seconds] 14:24:53 -!- Blazinghand has quit [Ping timeout: 268 seconds] 14:35:09 -!- Nightbeer has joined ##crawl-dev 14:39:15 -!- Roarke has quit [Quit: Page closed] 14:39:15 -!- Drahbeg has quit [Ping timeout: 272 seconds] 14:45:53 -!- OCTOTROG has quit [Remote host closed the connection] 14:50:48 -!- OCTOTROG has quit [Remote host closed the connection] 14:51:09 is there any reason why casting invisibility or haste don't warn you if your magic contamination will overload? 14:51:39 because it is usually not a sure thing one way or the other 14:51:57 -!- dead_ has quit [Ping timeout: 250 seconds] 14:52:10 and giving warnings when it is would leak information about exactly how much contam you have 14:52:23 -!- LordLovebone has quit [Ping timeout: 250 seconds] 14:54:51 is that really that bad? 14:55:22 95% of me getting overloaded is because I think I wait enough after casting invis, but it turns out I didn't. 14:55:58 indspenceable: you realize that there is already an indicator on the screen telling you how much contam you have, right? 14:56:30 if you have grey contam, then it's risky to take actions that give you more contam because it might become yellow 14:56:39 there is, it starts saying something when you're at least 50% certain to get malmutated 14:57:43 -!- wheals has joined ##crawl-dev 15:00:15 -!- xnavy has quit [Ping timeout: 240 seconds] 15:00:30 -!- Kellhus has quit [Quit: Page closed] 15:03:19 -!- randart has quit [Ping timeout: 272 seconds] 15:05:00 alright, thanks. the contam indicator is unreliable IMO because it's not graduated before the dangerous part (it doesn't matter that there are like 6 levels of contamination because getting even 1 of those levels is Very Bad). Didn't realize there was ever a prompt cause I hadn't gotten it. 15:05:01 anyway, nvm me. 15:05:15 a wind drake just blew me through a stone wall 15:13:11 -!- yalue has quit [Quit: Leaving] 15:22:10 -!- GuraKKa has quit [Quit: Leaving.] 15:23:52 -!- syllogism has quit [] 15:26:49 -!- Burer has quit [Quit: Truly, the end of days.] 15:26:58 -!- Tux[Qyou] has quit [Quit: Excess flood] 15:32:51 -!- ystael has quit [Ping timeout: 249 seconds] 15:33:26 -!- Cryp71c has quit [Quit: Leaving] 15:39:29 -!- _D_ has quit [Quit: Lost terminal] 15:44:17 -!- VolteccerJack has quit [Quit: HydraIRC -> http://www.hydrairc.com <- The professional IRC Client :D] 15:47:19 -!- kingbuddyboy has quit [Quit: Page closed] 15:48:48 -!- Pisano1 has quit [Quit: Leaving.] 15:50:04 -!- heteroy has quit [Read error: Connection reset by peer] 15:51:47 -!- _dd has quit [Remote host closed the connection] 15:57:47 kilobyte: Ah hah! It was you: https://gitorious.org/crawl/crawl/commit/30670662abe8d1661a589e493ca44220dc92b4d9 15:58:02 -!- dsn has quit [Quit: Page closed] 15:58:58 kilobyte: The culprit behind my MSVC woes: http://git.develz.org/?p=crawl.git;a=commit;h=f2b8b55cca63307837bfdcffa1c0545458b444a4 15:59:49 "whoops" 16:03:11 -!- Lawman0 has quit [Ping timeout: 260 seconds] 16:03:21 -!- absolutego has quit [Quit: leaving] 16:03:55 -!- mizu_no_oto has quit [Quit: ["Textual IRC Client: www.textualapp.com"]] 16:03:57 -!- Roarke has quit [Quit: Page closed] 16:05:10 -!- dtsund has joined ##crawl-dev 16:05:51 -!- Shazbot has quit [Quit: Leaving] 16:11:36 -!- kunwon1 has quit [Excess Flood] 16:13:12 -!- kunwon1 has joined ##crawl-dev 16:13:25 -!- Xenobreeder has quit [Ping timeout: 250 seconds] 16:14:26 -!- Morg0th has quit [Quit: leaving] 16:19:38 why "whoops"? It appears to be a defect in MSVC to me. 16:20:29 also, we use &vector[0] in 29 other places 16:20:59 -!- PunchRockgroin_ has quit [Quit: Page closed] 16:22:35 -!- heteroy has quit [Ping timeout: 272 seconds] 16:23:58 -!- Nightbeer has quit [Read error: Connection reset by peer] 16:25:08 -!- Wolfram has quit [Ping timeout: 250 seconds] 16:26:48 You can use &vector[0], just not if the vector is empty. 16:27:34 -!- blabber has quit [Quit: leaving] 16:28:28 -!- dtsund has quit [Ping timeout: 245 seconds] 16:28:37 xFleury: that's the difference between vector[] and vector.at() 16:28:55 <|amethyst> kilobyte: the C++ standards gives those two the same operational semantics 16:29:17 <|amethyst> both a[n] and a.at(n) are defined as *(a.begin() + n) for string, array, deque, and vector 16:30:12 |amethyst: yeah, and iterators are explicitely guaranteed to have a single "past the end" value 16:30:22 <|amethyst> but you can't dereference it 16:30:51 -!- Vizer has quit [Ping timeout: 245 seconds] 16:31:24 24.1.5 16:33:14 -!- Nerem has quit [Ping timeout: 264 seconds] 16:33:45 <|amethyst> "The library never assumes that past-the-end values are dereferenceable" 16:34:29 <|amethyst> not that that says it doesn't make its own past-the-end values dereferenceable, but I don't see anything that would guarantee that 16:38:32 -!- dtsund has joined ##crawl-dev 16:38:33 -!- zzxc has quit [Ping timeout: 250 seconds] 16:40:00 -!- buppy has quit [Remote host closed the connection] 16:43:03 -!- dtsund has quit [Ping timeout: 245 seconds] 16:45:10 -!- Ragdoll has quit [Quit: Houdoe] 16:46:06 hrm, the standard is damn hard to read when it comes to operator[]. It does contrast it to at() (bounds checked), but the wording is not totally clear. 16:46:56 I find other documentation that specifies operator[] explicitely as not bounds checked, but I failed so far to find this in ISO/IEC 14882:2011 16:46:56 <|amethyst> I've always understood it as, at() is bounds-checked, while out-of-bounds access with operator[] is undefined behaviour 16:46:57 In MSVC, there's almost no difference between [], and at(). They're both bounds checked by default, unless you use some specific pre-processor directives to disable bounds-checking. 16:47:13 In both Debug AND Release. 16:48:05 -!- dead_ has quit [Ping timeout: 250 seconds] 16:48:21 I seem to recall that you can take the address of that past-the-end element, just not dereference it, but I can't seem to find that in the standard. 16:49:20 <|amethyst> kilobyte: you can have a pointer to the past-the-end element. AIUI the problem is that vec[0] is considered a dereference even if you're only taking the address of the resulting lvalue 16:50:41 -!- Quashie has quit [Read error: Connection reset by peer] 16:50:43 <|amethyst> kilobyte: with C++11, data() would be such a past-the-end pointer; and in a non-empty vector &vec[0] + vec.size() would be fine (since it only dereferences vec[0] and not vec[size]) 16:51:53 I found so far that &array[past_the_end] and &*nullptr are valid. 16:52:13 <|amethyst> they are? 16:52:56 <|amethyst> I guess &* are collapsed for pointers? 16:53:28 -!- gammafunk has quit [Quit: gammafunk] 16:54:49 -!- C7ty has quit [Quit: If at first you don't succeed, skydiving is not for you] 16:55:01 -!- Datul has quit [Ping timeout: 250 seconds] 16:55:41 -!- indspenceable has quit [Remote host closed the connection] 16:56:28 _C_ (not C++) 6.5.3.2 16:56:38 lemme try to find this in C++ 16:57:08 -!- Pacra has joined ##crawl-dev 16:57:34 <|amethyst> I don't think C++ has that rule (looking at 5.3.1 expr.unary.op 16:57:56 <|amethyst> it does say 16:58:02 <|amethyst> Note: a pointer to an incomplete type (other than cv void) can be dereferenced. The lvalue thus obtained 16:58:05 <|amethyst> can be used in limited ways (to initialize a reference, for example); this lvalue must not be converted to a 16:58:09 <|amethyst> prvalue, see 4.1. 16:58:50 <|amethyst> so there are at least some situations where you're allowed to dereference a bad pointer as long as you don't actually use the value 16:59:02 <|amethyst> but that doesn't apply to the current situation 17:00:39 -!- thened has quit [Quit: thened] 17:01:31 -!- ystael_ has quit [Ping timeout: 268 seconds] 17:02:03 -!- SteampunkDuck has quit [Ping timeout: 260 seconds] 17:03:24 Sigh.. my SSD probably hates me right now. I've gone through about 50 GB of deletions trying to narrow down the precise change to the MSVC solution files that fixes the path issue. 17:03:56 xFleury: did it regress? 17:05:05 -!- dtsund has joined ##crawl-dev 17:06:07 Did it regress? It was working in 0.12, then broke sometime before 0.13. I have it working again on my machine, but I've trial&error'd soo much that it's a nightmare to narrow it down. 17:06:25 I'm resorting to, changing one thing, testing if it was the thing that fixed it. If not, wipe the folder, and try again. 17:07:29 hrm, the standard contrasts operator[] with at(), but every time it does so, it is clear about at() but not []. 17:07:35 -!- Garhauk has quit [Ping timeout: 250 seconds] 17:08:10 I'm giving up... even though this makes no sense to me otherwise, even if it's a bug in MSVC, this doesn't change anything as it still needs to be worked around on our side 17:08:57 -!- jameyd has quit [Quit: jameyd] 17:09:31 <|amethyst> kilobyte: many of the other occurrences are things that won't be empty, but there are at least a few where an empty file could cause a crash in VS 17:10:06 -!- fufumann has quit [Read error: Connection reset by peer] 17:10:50 -!- Vbitz has quit [Ping timeout: 245 seconds] 17:11:07 :S 17:11:37 in other words: let's build all official releases with gcc for foreseable future :p 17:11:55 because I guess there's more pitfalls like this 17:12:39 It's apparent that a considerable amount of effort has been put into the code for cross-platform compilation support. 17:12:45 -!- hurdos has quit [Ping timeout: 252 seconds] 17:12:54 Would be a waste to give up over a &vector[0] issue. 17:13:25 s/cross-platform/cross-compiler 17:14:33 0.12 ran very smoothly on MSVC, so I'm not expected there to be many more issues in the 0.13 code. 17:14:53 <|amethyst> if VS2012 has vector::data() and string::data() we could do wrapper inlines/macros 17:15:05 -!- xnavy has quit [Ping timeout: 268 seconds] 17:15:14 xFleury: having more compilers means more warnings and flow analyzers that can point out real bugs 17:15:43 <|amethyst> e.g. #define APTR(x) (&(x)[0]) most places and ((x).data()) on MSVC 17:16:15 |amethyst: It's difficult already to get people to use PURE, I can't imagine trying to correct all the missing APTR. 17:16:16 -!- FDahlmeyer has quit [Quit: FDahlmeyer] 17:16:21 xFleury: but that's for development; if there are two compilers for a platform, one very well tested and one known to be flaky, it's obvious which one to choose for releases 17:16:22 <|amethyst> and when we can finally drop C++03 support it could just use data() both ways 17:16:38 <|amethyst> xFleury: &foo[0] is pretty easy to search for, though 17:16:44 "flaky" at least with regards to testing on our side 17:18:03 * kilobyte wonders how clang fares on Windows. 17:18:40 its newest releases are quite good. They produce slower code than gcc, but not by much. 17:18:54 and clang has some major flow analyzing goodies 17:19:15 which would be good for finding bugs 17:19:34 <|amethyst> kilobyte: yeah, we got some good finds from SLi's search for dead assignments 17:20:07 <|amethyst> %git a926585 17:20:08 07|amethyst02 * 0.12-a0-2371-ga926585: Fix ENCH_LIFE_TIMER duration (SLi). 10(8 months ago, 1 file, 1+ 0-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=a9265854a996 17:20:21 <|amethyst> %git d43d9222 17:20:21 07|amethyst02 * 0.12-a0-2370-gd43d922: Fix incorrect Zot point costs (SLi). 10(8 months ago, 1 file, 2+ 0-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=d43d92223efb 17:20:29 <|amethyst> %git aa31237 17:20:31 07|amethyst02 * 0.12-a0-2375-gaa31237: Don't prompt for swapping in the same amulet (#4134). 10(8 months ago, 1 file, 10+ 6-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=aa312377cad6 17:25:34 Wow, that looks like a great analysis inded. In the "fix incorrect zot points" one, did it raise a red flag because "scale20" was being assigned twice, with the previous value having no effect? 17:26:56 <|amethyst> xFleury: exactly 17:28:58 <|amethyst> gcc has dead-store elimination in the optimizer, but no way to turn that into a warning 17:29:47 <|amethyst> (and it might not be as thorough as clang's, since it's intended for optimisation not debugging) 17:30:04 clang produces some obviously bogus warnings so I have some doubts as for code it produces, though 17:30:24 I silenced most of them, one can't really be: 17:30:26 ./menu.h:541:9: warning: private field 'ncols' is not used [-Wunused-private-field] 17:30:31 |amethyst: would you mind taking a loot at https://crawl.develz.org/mantis/view.php?id=2532 and telling me what you think about it? 17:32:46 <|amethyst> I don't like the wording of the message 17:33:11 <|amethyst> if it's going to print a message, it should be a message that actually tells you the stuff you learned 17:33:34 <|amethyst> so you don't have to x over them to figure out "what they told you" 17:34:06 <|amethyst> I'm also not sure whether or not is_follower is the right check, because I'm not familiar with that part of the code 17:35:07 <|amethyst> On a less technical note, I worry that it would encourage tedious behaviour as you try to get your orcs to ID distortion brand for you 17:35:26 <|amethyst> I'll add a comment 17:35:44 <|amethyst> kilobyte: err, actually it's correct 17:36:05 |amethyst: well, trying to turn it into a warning would probably lead to it being a bit more thorough 17:36:10 <|amethyst> kilobyte: I removed ncols from the class definition and the constructor initializer list and it compiles just fine 17:36:18 -!- Blazinghand has joined ##crawl-dev 17:36:33 though, for debugging, it's better if the dead stores AREN'T actually eliminated 17:36:56 <|amethyst> SamB: yeah, that's what I meant by "optimisation not debugging") 17:36:59 <|amethyst> s/)// 17:37:27 <|amethyst> I guess "debugging" wasn't the right word, but "quality assurance" 17:37:31 <|amethyst> proactive debugging 17:37:31 "Xom laughs as you force your Orcs to equip and unequip distortion branded weapons." 17:37:37 -!- LogicNinja has quit [Client Quit] 17:37:38 hell I'd laugh too xD 17:37:45 if kiku hates unliving things, and will excommunicate you for turning yourself into a statue 17:37:49 why can gargoyles worship him 17:37:54 |amethyst: it seems quite common to adapt optimization algorithms for such use 17:38:29 atomicthumbs: because you have your gods mixed up 17:38:31 like "hey, we're building a data structure that tells us about this stuff; why not do some warnings based on it?" 17:38:35 oh whoops 17:38:37 kiku doesn't care, yred does 17:38:48 I thought kiku was death and yred was necromancy 17:38:51 -!- Quashie has quit [Ping timeout: 260 seconds] 17:39:01 * SamB wonders if it happens the other way round, too 17:39:03 or are they 17:39:06 brb reading manual 17:39:51 yeah yred doesn't like me 17:39:59 "you're a statue, fuck off" 17:40:41 which is weird since you can still die in statue form 17:40:51 <|amethyst> kilobyte: column_composer::ncols is unused because the code that would use it just has columns.size() instead 17:40:53 make him dislike ddoor 17:41:19 <|amethyst> wheals: you can't die, you can be destroyed 17:41:28 I think the reasoning is that statues aren't alive or dead for the duration 17:41:33 they just... are 17:41:42 I wonder if yred refuses djinni 17:41:51 <|amethyst> no, they're alive 17:42:05 <|amethyst> not like humans, but like demons or angels 17:42:15 but they're made of magic 17:42:24 aren't they just about as alive as a fire elemental 17:42:28 so do statue form deaths leave ghosts? 17:42:43 |amethyst: i agree, the flavor text sucks... thanks 17:42:57 they don't need to eat. to determine whether djinni are truly alive we need to know if they can reproduce 17:43:02 i trust someone will get right on this 17:43:08 This seems odd. I edited `crawl-ref.sln` and a few other files, then ran "git commit -a" yet it doesn't pickup the changes to `crawl-ref.sln`. o.O 17:44:05 `crawl-ref.sln` should be a "staged" file or what not, considering it's the MSVC equilvalent of a makefile. 17:44:19 atomicthumbs: to quote the movie predator: "if it bleeds we can kill it" 17:44:51 <|amethyst> atomicthumbs: angels also don't need to eat and probably don't reproduce, yet yred accepts them 17:45:00 xFleury: a tracked file 17:45:26 <|amethyst> also, yred doesn't hate all unliving things 17:45:31 <|amethyst> just "artificial" ones 17:45:52 so, no golems, then 17:45:53 <|amethyst> which isn't really well-defined, except to give a list of what is artificial :) 17:45:55 :V 17:46:02 did I miss an angel mutantrace 17:46:07 not mutantrace 17:46:08 atomicthumbs: do angels have the naughty bits? 17:46:10 that is another game enitrley 17:46:21 at least succubi and incubi do, at least 17:46:31 <|amethyst> atomicthumbs: well, I say yred accepts them because of profane servitors 17:46:37 -!- LexAckson has quit [Read error: Connection reset by peer] 17:46:39 <|amethyst> atomicthumbs: who are fallen angels turned by/to yred 17:46:45 ah 17:46:51 well, they're profane 17:46:57 undead dead angels actually, so this doesn't say much 17:47:09 <|amethyst> true 17:47:22 I'd allow them for Yred, though 17:48:20 Oh, nevermind, it picked up the changes, but didn't seem to tell me on the console. I see the changes now after pushing it to github. :| 17:48:36 xFleury: and it IS a tracked file; you're positive you changed something in that file itself? 17:49:00 hmm, I was too busy investigating to read the intervening chat 17:49:11 or well some of that time was talking to my sister ... 17:49:21 lol, sorry 17:49:40 no problem, I didn't actually spend much time investigating 17:49:47 03|amethyst02 07* 0.14-a0-271-g6aec31b: Remove an unused member variable (kilobyte) 10(69 seconds ago, 2 files, 2+ 2-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=6aec31b66f90 17:49:57 I just checked if I had such a file, since I'd not have one if it wasn't tracked ... 17:50:08 -!- Aryth has quit [Ping timeout: 256 seconds] 17:52:51 -!- Soadreqm has quit [Read error: Operation timed out] 17:52:51 -!- Aryth1 has quit [Ping timeout: 253 seconds] 17:53:07 -!- indspenceable has joined ##crawl-dev 17:53:14 -!- heteroy has joined ##crawl-dev 17:53:46 any idea what's the problem with changing tile offsets in stable? 17:54:11 hasn't the cached tilesheet problem been fixed already? 17:54:51 at least I quite can't see anything wrong in console (duh) and local tiles (item enums don't get saved except as names) 17:55:14 kilobyte: it's a web thing 17:55:33 (#7634 -- tilesheets differing according to the version of perl and architecture) 17:56:32 if we made webtiles preload the tiles, and put them in a location whose name was based on the commit, maybe we could stop worrying about this 17:57:10 actually we wouldn't even need to move them; we could just abuse the query string 17:57:40 isn't the browser/server supposed to check timestamps? 17:57:51 -!- coffeemonster has quit [Ping timeout: 250 seconds] 17:58:02 only if we didn't tell it that the data we already sent it would be good for some time 17:58:24 If perl is only used binsort the tiles, I'd rather it'd just be rewritten in C++. perl is a major ball-weight on Windows. The thing has more dependencies than anything else I've ever seen. 17:58:43 If-Modified-Since 17:58:53 kilobyte: requires a request though 17:59:15 I don't think hash order is even stable across runs? 17:59:41 <|amethyst> probably we could fix most of that by turning caching back to off 17:59:42 since that "predictable hash => DoS" bug got fixed in all the scripting languages 17:59:54 SamB: ah, that 17:59:57 |amethyst: yes, that would be even simpler 18:00:11 <|amethyst> since bandwidth doesn't seem to be the biggest limitation 18:00:12 xFleury: it was a bug in the script that just happened to work in the past 18:00:34 would be a bit more sluggish starting with a hot cache; does webtiles get restarted when you die and start again? 18:00:55 xFleury: there's "only" 49 perl scripts used by Crawl, and I'd migrate more stuff to perl rather than from it 18:00:57 Still though, even if it can be fixed simply. I'm still against using perl. Does anyone know the extent of perla's use? 18:01:01 |amethyst: I think the point was latency, not bandwidth 18:01:25 xFleury: as unlike python or ruby, it is modular, doesn't have that many dependencies, is present on all platforms but Windows by default 18:01:34 though I don't actually know what the headers weigh in at these days 18:01:59 doesn't perl come with msys anyway 18:02:06 SamB: it does 18:02:08 unlike, say, python 18:02:25 which you'd have to install separately and then somehow point our build system at 18:02:26 <|amethyst> I'm not sure when exactly the webtiles client side requests the tile sheet 18:02:43 |amethyst: there's one way to find out 18:03:02 there are only 8 tilesheets anyway 18:03:03 turn on request monitoring and fire it up 18:03:25 so, worst case that's about 4 RTTs? 18:03:39 -!- Kalir has quit [Quit: CADENZA-CLASS CLOCKWORK KNIGHT--OPERATION TERMINATED] 18:03:56 can be pulled in parallel, too 18:03:59 or, wait, how many RTTs to set up a connection. not that that actually happens? 18:04:17 actual pipelining would do better of course 18:04:26 yeah 18:04:49 but you can't have pipelining due to ancient broken implementations 18:04:49 the 4 figure included the 2 connections actually permitted by the RFCs 18:05:01 drat 18:05:16 no idea why anyone would still bother about those 18:05:22 so for pipelining you need SPDY? 18:05:23 <|amethyst> SamB: so what Cache-Control: should I use for game data files? 18:05:50 * SamB doesn't remember how Cache-Control works 18:06:08 <|amethyst> SamB: webtiles does either no header (the current behaviour on the webtiles servers) or Cache-Control: no-cache (the default settings) 18:07:09 hmm, this w3 copy of RFC 2616 is nicely formatted 18:07:16 <|amethyst> SamB: so turning that to default in the config would mean all the tilesheets are loaded at least at every visit, possibly every spectation/game 18:07:17 The HTTP meta "Cache-Control"? I believe that's ignored by modern browsers. 18:08:43 well the defaults (for no header) are supposed to be defined in http://www.w3.org/Protocols/rfc2616/rfc2616-sec13.html#sec13.4 18:11:17 would be good to fix this for the future, but right now I'm more concerned what to do with this bug in stable 18:12:15 looks like no-cache should be fine 18:12:20 <|amethyst> The issue being that different versions of stable have different tile indices? 18:12:55 -!- Leafsnail has quit [Quit: Why is the alphabet in that order? Is it because of that song?] 18:12:55 |amethyst: the issue being if this is a problem 18:13:06 solution 1: deal with webtiles somehow (moving or renaming the files a bit?), solution 2: make tiles data arch-dep in 0.12 and 0.13, solution 3: carry a patch in the Debian packaging and recommend other distros to do this as well 18:13:25 -!- Zaba has quit [Ping timeout: 272 seconds] 18:13:27 http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html#sec14.9.1 has verbiage that indicates that no-cache doesn't actually preclude caching, just precludes re-use without re-validation 18:13:55 |amethyst: distributions (at least Debian and something RPMy I vaguely remember) split the packaging into arch-dep and arch-indep parts 18:14:06 kilobyte: can't backport the patch ? 18:14:09 <|amethyst> kilobyte: oh, and the hash order thing 18:14:26 SamB: can, but that would break webtiles 18:14:28 I mean officially 18:14:30 <|amethyst> I don't think webtiles is the only issue 18:14:38 really??? 18:14:51 <|amethyst> see jpeg's section in docs/develop/save_compatibility.txt 18:14:52 %git fc2aeb4d5915807a4ab8b62ff5adf3d1ead3c01d 18:14:52 07kilobyte02 * 0.14-a0-270-gfc2aeb4: Fix unrand tile mismatches between architectures. 10(6 hours ago, 1 file, 1+ 1-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=fc2aeb4d5915 18:15:16 |amethyst: are there any tiles still getting put into the save file other than as names? 18:15:16 <|amethyst> tile indices for walls and floors apparently are stored in the level files 18:15:20 <|amethyst> oh 18:15:47 <|amethyst> not sure how to check easily 18:16:46 this commit changes only main.png, which is used for in-inventory/on-ground items 18:16:52 I thought this did not happen anymore 18:17:23 kilobyte: I think they get put in as numbers, but the names are stored as well and the numbers recalculated based on that? 18:17:45 SamB: there's an array of names, the numbers stored are indices into that array 18:18:01 I must go, my sister is demanding that I play magic with her 18:18:05 kilobyte: exactly! 18:18:12 that's for walls and floors, the file that changes here is neither of those 18:18:26 -!- Zaba has joined ##crawl-dev 18:18:35 The first time Crawl.exe is ran, does it build anything on the file-system? Like a compilation of all the DES files or something? 18:18:38 SamB: how many Schahrazads do you have in your deck? 18:18:58 the decks are both hers actually 18:19:00 I can't keep deleting my entire /crawl folder, lol, my SSD is gonna die. :S 18:19:09 the one I've been using is her splinter deck 18:19:21 xFleury: directories: saves/ and morgue/, plus the cache in saves/des and saves/db 18:19:22 xFleury: How big is the /crawl folder? 18:19:23 xFleury: yes 18:19:34 xFleury: it does 18:19:57 xFleury: they should work well enough wrt going back and forth nowadays 18:20:16 if something is going wrong, you should probably trace it with that sysinternals tool I forgot to find for you before 18:20:28 xFleury: at least unless you run multiple Crawls in parallel and/or update the files while Crawl is running 18:20:48 xFleury: procmon.exe, looks like 18:21:00 kilobyte: hmm, that should be safe for the sqlite stuff right? 18:21:16 SamB: _should_ but is not :( 18:21:29 our locking is buggy 18:21:53 we don't build the DB in one transaction? 18:22:27 Considering I don't have a saves/ or /morgue folder in my build, there's gotta be another place that gets written to. :S 18:22:31 what about the des cache? do we do the rename trick there? 18:22:35 The behavior of a first-run is very different than a second-run. 18:22:40 a seemingly guaranteed reproducer: remove the dependency on builddb in like 1664 of the Makefile, run "make [debug] -jX test" 18:22:46 xFleury: Procmon.exe! get it! now! 18:22:50 Maybe it's about time I take up SamB's advice and use a sysinternals tools 18:23:02 (I've been ignoring his advice for days..) 18:23:16 SamB: the des cache is a nightmare you really don't want to look into, except as something to nuke from the orbit and rewrite 18:23:53 SamB: it blindly reads it, and if a _detected_ inconstency happens, an exception makes it rebuild the cache again then retry 18:24:24 the files are not kept open (can't really, there's too many of them), so renaming wouldn't help 18:25:21 462 files under saves/des, from 125 sources 18:29:10 nuke from orbit... best concept ever 18:33:47 ah hah! ty Procman.exe! 18:34:03 The saves/morgue/etc was going in "source/MSVC" instead of "source". 18:34:26 xFleury: you can mess with the defines, the Makefile has some documentation 18:34:57 -!- scummos^ has quit [Ping timeout: 272 seconds] 18:39:02 even letting your debug builds use that Windows' equivalent of $HOME should be safe enough 18:40:16 the installer uses SAVEDIR='~/crawl' 18:41:31 xFleury: If it's not too huge you could consider messing around with it on a RAM drive. 18:41:52 Which ought to at least save your SSD from a lot of wear and tear. 18:42:21 I kind of fail to see the reason to avoid rebuilds. They are incremental after all... 18:42:23 -!- gammafunk has joined ##crawl-dev 18:43:25 every compilation you have ~330MB of .o files, even if the des cache got fully rebuilt every time (it's not), that's mere 3MB. 18:52:14 -!- sildraith has quit [Quit: Leaving] 18:52:53 -!- dead_ has quit [Ping timeout: 250 seconds] 18:55:15 -!- Hailley has quit [Quit: Petra is closed!] 18:56:08 -!- duckroller has quit [Quit: pokemonz] 18:57:06 -!- the_glow has quit [Ping timeout: 245 seconds] 18:59:47 So I've set the environment variable of crawl so that everything "seems" to be loading from the correct path, yet the pathing issue remains. :S 19:00:27 -!- OCTOTROG has quit [Remote host closed the connection] 19:01:27 or, hmm, maybe I've missed a couple more folders. 19:02:14 Basically, what "worked" was shoving everything into a single massive folder, and running that. 19:02:33 And right now, I'm trying to submit a fix where that is not done, lol. 19:05:29 -!- indspenceable has quit [Read error: Connection reset by peer] 19:06:17 -!- indspenceable has joined ##crawl-dev 19:06:40 -!- indspenceable has quit [Read error: Connection reset by peer] 19:07:42 -!- indspenceable has joined ##crawl-dev 19:09:57 fr: charges used is inscribed on a sack of spiders or box of beasts similar to wand charge count 19:09:58 -!- ldf has quit [Quit: ldf] 19:12:05 -!- Gotham has quit [Read error: Connection reset by peer] 19:12:22 -!- medgno2 has quit [Quit: WeeChat 0.4.1] 19:12:28 -!- jeanjacques has quit [Ping timeout: 264 seconds] 19:13:26 -!- jeanjacques has joined ##crawl-dev 19:20:48 -!- Camicio has quit [Remote host closed the connection] 19:20:54 -!- Azzkikr has quit [Ping timeout: 264 seconds] 19:24:52 -!- pythonsnake has quit [Ping timeout: 246 seconds] 19:25:36 -!- wheals_ has joined ##crawl-dev 19:28:04 -!- ShadeTornado has quit [Quit: Leaving] 19:28:07 -!- blackflare has quit [Ping timeout: 260 seconds] 19:29:03 -!- wheals has quit [Ping timeout: 272 seconds] 19:32:42 -!- Tarquinn has joined ##crawl-dev 19:32:52 -!- wheals_ is now known as wheals 19:35:52 http://i.imgur.com/RGg2XXg.png 19:36:12 sigh... finally got it working again, and the directory is clean 19:37:33 -!- TAS-2012v has quit [Ping timeout: 272 seconds] 19:38:16 -!- Taraiph has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] 19:38:36 -!- dtsund has quit [Quit: dtsund] 19:41:50 -!- Quashie_ has quit [Ping timeout: 240 seconds] 19:41:54 -!- Zrachosh has quit [Quit: Page closed] 19:42:37 -!- pelotr0n has quit [Ping timeout: 272 seconds] 19:56:17 -!- alefury has quit [Read error: Connection reset by peer] 19:58:13 Is there any reason why /subsystem:console would be required if FULLDEBUG is defined? o.O 19:58:33 My "incredible" fix is incredibly simple: https://github.com/xFleury/crawl-0.13.0-fairplay/commit/9756e02e7d925f7073c5aa7896b3e29e3aac6dc4 19:59:04 -!- Daekdroom has quit [Quit: Saindo] 20:00:21 -!- Basil has quit [Ping timeout: 268 seconds] 20:00:26 uhm, that's what debug builds are for... 20:01:38 -!- heteroy has quit [Quit: ChatZilla 0.9.90.1 [Firefox 24.0/20130910160258]] 20:02:07 -!- indspenceable has quit [Read error: Connection reset by peer] 20:02:15 Bison/Flex fails to parse the DES files if FULLDEBUG is defined but the subsystem is /subsystem:windows. 20:02:26 -!- indspenceable has joined ##crawl-dev 20:02:29 Removing FULLDEBUG, it works. 20:02:37 Alternatively, /subsystem:console, and it works. 20:02:53 -!- rossi has quit [Ping timeout: 272 seconds] 20:03:12 -!- tsn has quit [Client Quit] 20:03:16 "fails"... how? 20:03:42 Fails with a highly ambigous error message "input to flex input scanner" failed or something along those lines 20:05:24 "#define FULLDEBUG" results in DEBUG_DIAGONOSTICS, DEBUG_ITEM_SCAN, DEBUG_MONS_SCAN, and DEBUG_BONES getting defined. 20:05:31 -!- st_ has quit [Ping timeout: 263 seconds] 20:05:36 Right now, I have no idea what any of those are. 20:05:40 ie, an actual debug build 20:06:14 -!- morik___ has quit [Ping timeout: 240 seconds] 20:06:35 DEBUG_ITEM_SCAN is sanity checks for items every turn, DEBUG_MONS_SCAN same for monsters, DEBUG_BONES is some minor stuff for ghosts, DEBUG_DIAGNOSTICS is almost everything else that there is to debug builds 20:07:00 So... nothing that would mess with PARSING the map? 20:07:13 -!- Zermako has quit [] 20:07:19 I imagine that there are parser diagnostics 20:07:49 what I don't know is if doing debug log output (presumably to stderr) causes an exception if there's no console subsystem 20:09:02 I highly doubt it, but that would explain it. The Flex/Bison stuff is loaded with diagnostic output stuff. 20:10:22 -!- jeanjacques has quit [Quit: quit] 20:15:11 -!- indspenceable has quit [Remote host closed the connection] 20:18:06 kilobyte: why so many files generated ? 20:18:41 SamB: every element, plus a lockfile, goes into a separate file 20:18:49 xFleury: is that the only difference? 20:19:18 Seems like it, I'm still testing it to see if it's coincidence. 20:19:39 maybe if we have no stderr we should open /dev/null 20:20:13 -!- pelotron_ has quit [Read error: Connection reset by peer] 20:21:11 -!- obelus has quit [Quit: Page closed] 20:21:19 SamB: every vault produces two halves: headers and body, headers are in memory all the time, the body gets loaded then freed when the vault is instantiated. Plus, there are per-file "lua preludes". 20:21:52 it's probably helpful to list the operations: 20:22:53 -!- P1x3lF1r3 has quit [Ping timeout: 272 seconds] 20:23:47 1. "load all headers of all vaults, in bulk". 2. "load all file preludes and execute them one by one". 3. "load the body of vault X". 4. (during initialization) "delete and replace all vaults+prelude that comes from file X" 20:25:59 hi guys 20:29:31 -!- Lawman0 has quit [Read error: Connection reset by peer] 20:29:38 what's 4 for 20:30:04 (and how do you undo having run that prelude?) 20:30:08 to build the cache in the first place, and to update it 20:30:26 no, 4. happens first 20:30:29 oh 20:30:36 see, that's a confusing numbering ;-) 20:30:42 yeah, sorry 20:33:51 do we clean out dead files from the cache? 20:35:01 (shouldn't du have an --iec flag?) 20:35:36 -!- P1x3lF1r3 has joined ##crawl-dev 20:36:00 kilobyte: so, do we create all of these files at one time? 20:37:44 so my sister tells me there is a card that basically lets you say "no u" when told to mill 20:39:46 Playing DCSS in MSVC in Debug build feels like playing webtiles.. >.> 20:40:55 SamB: yes 20:41:19 SamB: technically, a couple of test targets load additional vaults later, but I see no reason to do that 20:41:46 so I'm unclear why we don't put everything but .lk in one file, or possibly just put everything in one file 20:42:03 it's not DOS so the file itself can be locked 20:42:24 -!- Valarioth has quit [Quit: Page closed] 20:43:10 issues to remember: 1. refreshing .des files takes time, so it's quite important to not reload everything when just one file has been changed 20:43:35 2. multiple crawls can be started in parallel, and notice the need to rebuild simultaneously (thus locking) 20:43:46 or the rename trick could be used, the point being to avoid intermingled files, or one crawl trying to read from another instance's in-progress write ... 20:44:12 when I said "everything" I meant everything for one .des file 20:44:19 3. a crawl can start, _then_ the source .des get updated, then another crawl starts. Currently, we change the cache under old processes which leads to breakage. 20:44:30 yeah, a .des file is an unit 20:45:08 problem 3. could be avoided by keeping the cache open and moving the new version over it 20:45:37 except on windows we'd have to give up instead 20:46:13 on POSIX systems it's a matter of renaming, on Windows you'd need tricks like moving it to a temp name and requesting a delayed deletion, but that's doable. Only DOS/Win98 disallowed messing with an open file. 20:47:01 well, there MIGHT be a way to open it in the first place such that we can do that link-replacement dealy 20:48:35 -!- minced has quit [Ping timeout: 250 seconds] 20:52:29 -!- wya has quit [Quit: Page closed] 20:54:25 -!- wheals has quit [Quit: Leaving] 20:56:21 1. open-and-lock tmpfile, 2. process and write the cache into it, 3. rm the_cache || mv the_cache some_random_name, 4. mv tmpfile the_cache 20:56:46 <|amethyst> is there any situation where we could self-assign a monster_info ? 20:58:15 who knows? 20:58:34 (add an assert and find out?) 20:58:54 <|amethyst> eh, might as well just fix it then :) 20:59:24 FILE_FLAG_DELETE_ON_CLOSE affects not just you but also all other users 21:00:23 Bah, MSVC still having some issues: dry fountains still showing "?", and if you try to auto-path near them, you get an out-of-bounds error: http://i.imgur.com/KfY5q28.png 21:00:36 so it looks like 3. could be replaced with: open(, FILE_FLAG_DELETE_ON_CLOSE), close, mv 21:01:10 Anyone familiar with how dry fountains are generated? 21:01:38 -!- FDahlmeyer has quit [Quit: FDahlmeyer] 21:04:22 -!- ainsophyao has joined ##crawl-dev 21:04:36 -!- ainsophyao has quit [Remote host closed the connection] 21:05:19 -!- Kaput has quit [Ping timeout: 246 seconds] 21:07:03 * SamB would "grep -i fountain *.cc" from the source dir 21:07:10 -!- Bloax has quit [Quit: Error: Bloax not found] 21:07:22 (or use MSDEV's search tool) 21:07:46 <|amethyst> xFleury: there were some recent changes related to those 21:08:39 <|amethyst> %git 5f5ef2de 21:08:40 07kilobyte02 * 0.14-a0-27-g5f5ef2d: Squash four types of dry fountains into one. Reflowing is gone... 10(3 weeks ago, 9 files, 15+ 53-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=5f5ef2decf95 21:09:49 by kilobyte.... 21:09:59 lol j/k xD 21:10:03 <|amethyst> but the other kinds should have been fixed up to the one remaining kind on game load 21:10:45 -!- heteroy has quit [Ping timeout: 272 seconds] 21:10:59 -!- heteroy_ is now known as heteroy 21:11:29 <|amethyst> xFleury: that's because most commits these days (well, a plurality) are by kilobyte: http://s-z.org/crawl-stats/authors.html 21:11:51 <|amethyst> xFleury: so if you want to know "who touched this last", kilobyte is a good guess 21:12:46 <|amethyst> If you want to know "who introduced a bug", but kilobyte and me are good guesses :) 21:12:51 <|amethyst> s/but/both/ 21:12:56 What does #define TAG_MAJOR_VERSION 34 do? 21:13:43 well, if we bump that a lot of code will be #if'd out 21:13:52 old compatability code 21:13:56 well, some not so old 21:14:06 <|amethyst> see docs/develop/save_compatibility.txt in the source and also http://crawl.develz.org/wordpress/save-compatibility-in-dcss-2 21:14:14 and the file format will be incompatible 21:14:53 lol, the code no longer compiles if I remove that #define xD 21:15:19 <|amethyst> yes, crawl needs that version number :) 21:15:46 <|amethyst> read those docs and that blog post before messing around with anything involving save compatibility 21:15:49 -!- heteroy has quit [Ping timeout: 272 seconds] 21:16:46 <|amethyst> Probably the blog post is better starting points, since it's aimed at the curious player rather than experienced crawl devs 21:17:05 <|amethyst> s/better/a &/; s/nts/nt/; 21:19:36 -!- Basil has quit [Ping timeout: 245 seconds] 21:20:43 |amethyst: you're doing mostly bug fixing, so blaming you for being a source of new bugs is rather unfair 21:20:59 having fewer active people is quite a concern, too 21:21:15 <|amethyst> kilobyte: my bug fixes aren't always the most careful :) 21:23:27 Anyone know off-hand the GIT address of trunk/master? 21:23:30 <|amethyst> I'm all for recruiting more devs. We have several frequent contributors 21:23:33 (for a git clone) 21:23:52 <|amethyst> git://gitorious.org/crawl/crawl.git 21:23:59 are we looking at http://www.ohloh.net/p/stonesoup again? 21:24:11 ??git 21:24:12 git[1/2]: Readonly: git clone git://gitorious.org/crawl/crawl.git Read/Write: git@gitorious.org:crawl/crawl.git Web interface: http://git.develz.org/?p=crawl.git or http://gitorious.org/crawl/crawl 21:24:25 <|amethyst> SamB: I was looking at my gitstats page 21:24:30 I actually built that URL myself just now for a google search though 21:24:35 |amethyst: gitstats? 21:24:41 -!- KLANG has quit [Quit: Page closed] 21:24:50 <|amethyst> ??stats 21:24:50 starting stats[1/3]: A character's starting stats are equal to species contribution + job contribution. See {starting stats[2]} for the species contributions and {starting stats[3]} for the job contributions. 21:24:53 <|amethyst> ??gitstats 21:24:54 gitstats[1/3]: http://www.ohloh.net/p/stonesoup 21:24:58 <|amethyst> hm 21:25:04 <|amethyst> ??gitstats[2] 21:25:05 gitstats[2/3]: http://s-z.org/crawl-stats/ — updated infrequently 21:25:08 <|amethyst> that one 21:25:47 <|amethyst> haven't updated in a few weeks, guess I should 21:26:19 -!- kait has quit [Read error: Connection reset by peer] 21:26:20 oh and here I could have found the URL I really wanted just by doing ??gitstats 21:26:20 DracoOmega had quite better shit-done/time ratio than me, it would be great to get him back 21:27:24 |amethyst: Is there a way to pull that with the dependencies? 21:27:52 "git clone git://gitorious.org/crawl/crawl.git" didn't let me compile the contribs folder. 21:28:10 well, actually, nevermind, I'll just copy the contribs over from another folder 21:28:12 <|amethyst> xFleury: after you get it do git submodule update --init in the top-level directory 21:28:18 <|amethyst> don't just copy the contribs 21:28:21 kk 21:28:31 <|amethyst> you'll end up getting out-of-sync that way 21:29:24 03|amethyst02 07* 0.14-a0-272-g0850502: Fix some assignment operators. 10(26 minutes ago, 3 files, 12+ 5-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=085050246da2 21:29:34 0.13 still has the bug where battlespheres dissipate when you take stairs 21:29:52 enemy ones, you mean? 21:33:53 hmm, what happened between Sep 2009 and Oct 2009 ... 21:34:24 hmm, it seems we need to write more comments 21:38:55 03MarvinPA02 07* 0.14-a0-273-gfbf0dc1: Remove some trapped loot vaults 10(2 days ago, 1 file, 0+ 29-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=fbf0dc1ee176 21:40:06 <|amethyst> SamB: the massive drop in the number of lines? I think that's when contribs were moved to submodules 21:40:18 |amethyst: ah 21:40:26 -!- Dalvant has quit [Ping timeout: 264 seconds] 21:41:39 Debian builds of the trunk branch on crawl.develz.org updated to: 0.14-a0-272-g0850502 21:41:51 hrm.. the MSVC solution files is looking for a missing file called "rng.cc". Is there a new file I should be included to replace that file? 21:42:16 -!- bh has joined ##crawl-dev 21:42:21 xFleury: yes, hash.cc 21:46:07 * xFleury switches /subsystem:console. 21:46:17 And look at that, suddenly Bison/Flex is parsing the map correctly. 21:46:54 hmm, so we SHOULD open /dev/null if stdout/stderr doesn't go anywhere? 21:47:03 (yes I know NT spells that some other way) 21:47:39 -!- gammafunk has quit [Quit: gammafunk] 21:47:56 -!- Crehl has quit [Quit: Leaving] 21:48:37 -!- Taraiph has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] 21:48:42 -!- Tarayinda is now known as Taraiph 21:51:23 -!- BlastHardcheese has quit [Ping timeout: 260 seconds] 21:52:31 -!- qoala has quit [Client Quit] 21:52:51 -!- qoala has joined ##crawl-dev 21:52:54 <|amethyst> SamB: btw, did you have an ohloh account? 21:53:05 <|amethyst> SamB: someone should add MarvinPA as another alias for MarvinPA 21:53:18 <|amethyst> so it doesn't claim he's one of the newest contributors :) 21:53:19 I do have one, yes 21:53:34 <|amethyst> http://www.ohloh.net/p/stonesoup/aliases 21:53:51 oh yeah, i accidentally pushed some stuff as that when i set up cygwin i guess 21:54:27 -!- Roarke has quit [Ping timeout: 250 seconds] 21:54:27 and didn't notice because that's what chei says anyway :P 21:54:45 <|amethyst> yeah, chei goes by email address then mailmaps it 21:54:47 don't we have mailmap for that kind of mistakes? 21:54:59 SamB: I think if /subsystem:windows, stderr, stdout, already goto the windows equilvalent of /dev/null. 21:55:15 right, chei didn't get confused i don't think since i still had the usual email set 21:55:18 <|amethyst> SamB: ohloh seems to go by name (plus its own alias table), not email address 21:55:23 xFleury: they are not open at all, which is different 21:55:33 |amethyst: it doesn't honour mailmap? 21:55:52 <|amethyst> SamB: we don't have a mailmap in the repository 21:55:59 -!- kait has quit [Ping timeout: 260 seconds] 21:56:19 we COULD for mapping old/broken names/emails to new ones 21:57:09 <|amethyst> maybe we should add that and remove crawl-ref-cia 21:57:31 -!- kait_ has quit [Ping timeout: 245 seconds] 21:57:38 hmm? 21:57:59 the one we use for Cheibriados, of course, maps names to IRC nicks 21:59:25 |amethyst: On that git repo you linked me, I still see "dry fountain" as a "?", and it still crashes when I hit "tab" near one. :S 21:59:52 <|amethyst> oh, you were doing 0.13 before... 22:00:08 yah, the general public release of the 0.13 source 22:00:10 <|amethyst> it's definitely not related to kilobyte's change then, since that was 0.14 22:02:00 -!- VolteccerJack has quit [Quit: HydraIRC -> http://www.hydrairc.com <- The alternative IRC client] 22:05:04 <|amethyst> maybe I should remove roguelike.git from s-z.org's git before I embarrass myself 22:05:14 <|amethyst> it was mostly my "learning python" project 22:06:05 did anyone ever check it out? 22:06:07 -!- WalkerBoh has quit [Remote host closed the connection] 22:06:18 or look at it? 22:06:20 <|amethyst> no clue, I don't watch my logs :) 22:06:30 besides searchbots I mean 22:08:41 SamB: I did a bit of research on what happens if you write to `stdout` or `stderr` without a console. It looks like you're right; we need to redirect it to a /dev/null, else expect a crash. 22:09:21 -!- heteroy has quit [Max SendQ exceeded] 22:10:13 03kilobyte02 07* 0.14-a0-274-g5554bc6: Inline a function. 10(3 days ago, 2 files, 3+ 9-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=5554bc657cdc 22:10:13 03kilobyte02 07* 0.14-a0-275-g9841f3f: Staticalize and simplify. 10(3 days ago, 2 files, 9+ 19-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=9841f3fb7365 22:10:13 03kilobyte02 07* 0.14-a0-276-g3c6a286: Remove a test case for old-style portal vaults. 10(56 minutes ago, 1 file, 0+ 14-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=3c6a2862f587 22:10:13 03kilobyte02 07* 0.14-a0-277-g660d43f: Move maps used by the test suite to dat/des/test 10(21 minutes ago, 10 files, 308+ 545-) 13http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=660d43f70c72 22:10:17 -!- Moanerette has quit [Quit: Page closed] 22:10:35 -!- ackack has joined ##crawl-dev 22:16:51 -!- wheals has joined ##crawl-dev 22:20:42 -!- Zifmia has quit [Ping timeout: 272 seconds] 22:23:13 -!- kait has quit [Ping timeout: 272 seconds] 22:23:43 -!- mizu_no_oto has quit [Quit: Computer has gone to sleep.] 22:25:30 -!- BlastHardcheese has joined ##crawl-dev 22:25:36 -!- BlastHardcheese has quit [Changing host] 22:25:36 -!- BlastHardcheese has joined ##crawl-dev 22:28:58 -!- ussdefiant has joined ##crawl-dev 22:30:30 does anyone have thoughts on how hard it would be to generalize webtiles into an API for roguelikes? 22:31:44 webtiles isn't even an API for crawl yet 22:32:28 fair... 22:32:32 I mean, it would be hard to use it for e.g. an Emacs-based client because of all the enums 22:33:10 Brian would love webtiles for Brogue, but he doesn't think he could pull it off from scratch 22:33:17 does it normally send keystrokes to the server? 22:33:53 well, it might be reasonably easy to adapt some of the code? 22:35:12 unsure. I haven't looked at the code, 'cause it's AGPL 22:35:25 it is? 22:36:42 yep 22:36:43 -!- magistern has quit [Quit: This computer has gone to sleep] 22:36:52 -!- Guest62734 has joined ##crawl-dev 22:37:23 where is this written? 22:37:28 -!- SwissStopwatch has quit [Ping timeout: 240 seconds] 22:37:59 -!- lobf has quit [Quit: lobf] 22:39:34 Brogue is AGPL, not webtiles 22:40:09 http://brogue.createforumhosting.com/post13307.html 22:40:13 -!- SwissStopwatch has joined ##crawl-dev 22:40:13 ah 22:41:02 -!- Guest62734 has quit [Ping timeout: 240 seconds] 22:41:03 so does brogue come with instructions to set up the banner telling you where to get the sources for the version in question? 22:41:38 probably not. That clause only matters if your'e running it as a service 22:41:40 This fountain bug is getting under my skin, I don't know enough about map generation to track it down. 22:41:58 Like, I don't know where to set breakpoints, or where even to look. 22:42:29 what happens if you stand on it? 22:42:33 or look at it 22:42:38 standing on it is fine 22:42:53 If you hit "tab" to do an auto-search, it seems to try and get details about that tile. 22:42:57 what happens if you hover the mouse over it? 22:43:03 The tile returns some generic ID of "-128" which is used as in an index of an array. 22:43:09 And so it throws an out-of-bounds. 22:43:34 hmm, are we using "char" somewhere dumb? 22:43:35 I'll try and find another dry-fountain and mouse over it. 22:44:04 hrm.. I didn't think of that one. But wouldn't that have been -127? 22:44:19 <|amethyst> -128 is 0xf0 22:44:29 you mean 0x80 ? 22:44:33 <|amethyst> err, yeah 22:45:34 Stable (0.13) branch on crawl.lantea.net updated to: 0.13.0-9-gbc74161 22:46:48 -!- Mandevil has quit [Quit: Leaving] 22:48:08 <|amethyst> xFleury: when you say -128, which number is that? 22:50:03 <|amethyst> If you're in wizard mode, hovering over the square should give a list of a bunch of things (tile indices and flavours, etc) 22:50:46 -128 is the value of 'dungeon_feature_type' 22:51:12 It's an enum, containing no negative values. 22:51:34 return from something like "dungeon_feature_type feat = env.map_knowledge(p).feat();" 22:51:37 <|amethyst> yeah, DNGN_DRY_FOUNTAIN is 128, so this sounds like a bad use of a char or sign extension 22:51:39 where "p" is the coord_def 22:52:20 <|amethyst> oh 22:52:41 <|amethyst> struct map_cell has 22:52:43 <|amethyst> dungeon_feature_type _feat:8; 22:52:47 whoops 22:52:55 <|amethyst> likewise trap_type 22:53:06 -!- Zifmia has quit [] 22:53:13 I thought I remembered MSVC having char signed ... 22:53:48 oh wait bitfields 22:53:55 how do you fix that? 22:54:10 add a bit? 22:54:12 I honestly don't know what the ":8" does, lol. 22:54:26 Constrain that int to 8 bits? 22:54:47 int would have been signed everywhere 22:55:08 -!- Rebthor has quit [Ping timeout: 240 seconds] 22:55:10 -!- ophanim has quit [Read error: Connection reset by peer] 22:55:49 <|amethyst> SamB: dungeon_feature_type probably has a signed base type 22:55:54 <|amethyst> which means that bit field is signed 22:56:07 yeah, ouch 22:57:31 there should be a warning about such fields :-( 22:57:40 There was no warning. 22:58:06 could be that there is one that's not on by default, I suppose 22:58:13 in MSVC I mean 22:59:43 -!- Taraiph has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] 23:00:19 -!- heteroy has quit [Ping timeout: 272 seconds] 23:00:33 So would the fix be to remove ":8" and redefine that enum as something like "enum dungeon_feature_type : UINT8"? 23:01:17 not really, no 23:01:24 :| 23:01:46 * SamB goes to look for helpful compiler flags 23:03:23 <|amethyst> If the value of an enumerator is 23:03:23 <|amethyst> stored into a bit-field of the same enumeration type and the number of bits in the bit-field is large enough 23:03:27 <|amethyst> to hold all the values of that enumeration type (7.2), the original enumerator value and the value of the 23:03:30 <|amethyst> bit-field shall compare equal. 23:03:45 standards violation then? 23:03:51 The Word According to Ken 23:04:44 <|amethyst> yeah, 7.2 makes it sound like 8 bits is enough for something where the enumerators range from 0 to 255 23:05:52 -!- jeffro has joined ##crawl-dev 23:06:02 hmm, C11 seems to have much larger print than C++11 ... 23:06:24 <|amethyst> xFleury: what if you change, in map_knowledge.h, the implementation of dungeon_feature_type feat() const: 23:06:53 -!- shmup has quit [Remote host closed the connection] 23:07:22 <|amethyst> to: return dungeon_feature_type(uint8_t(_feat)); 23:07:34 huh, why do they require enumerators be representable by int ... 23:07:36 <|amethyst> not that I think that's a good solution 23:07:49 -!- Silurio has quit [Ping timeout: 268 seconds] 23:08:03 I wonder what clang-cl would do 23:08:08 Haskell got enums right 23:08:50 |amethyst: Ok, it's compiling now. Gonna take a few minutes to find a dry fountain though. 23:09:19 <|amethyst> use wizmode 23:09:30 <|amethyst> &(dry_fountain 23:09:58 then move 23:10:14 -!- Pisano2 has quit [Ping timeout: 240 seconds] 23:10:55 As this is 0.13, "permadry_fountain"? 23:10:59 -!- MaxFrost has quit [Ping timeout: 272 seconds] 23:11:01 <|amethyst> http://software.intel.com/en-us/articles/intel-c-compiler-attributes-signedness-to-enum-bit-fields 23:11:25 <|amethyst> mentions the behaviour of MSVC 23:11:27 So far so good; "permadry_fountain" is rendering correctly. 23:11:53 <|amethyst> xFleury: save, recompile without that fix, and see if it crashes 23:12:05 <|amethyst> just to make sure it wasn't wizmode doing something funny 23:12:16 |amethyst: I was going to try and find one in the wild, but yah I can do that too. 23:13:41 <|amethyst> this is one of the few places that uses bit fields in Crawl 23:14:32 <|amethyst> the others are item_def::base_type and item_def::mon_type 23:14:33 Remind me never to use bitfields in my own C++ projects, lol. 23:14:52 -!- minced has quit [Quit: Page closed] 23:15:02 so M is 8 right? 23:15:34 |amethyst: Reverting your "fix" and recompiling results in wizmode spawning "&(permadry_fountain" showing a "?". 23:15:41 but item_def is safe because, you know, we are never going to have that many base types 23:15:42 |amethyst: In otherwords, it looks like your fix fixed it. 23:16:07 :D 23:16:13 <|amethyst> SamB: yeah, M is 8 here since every enumerator is <= 255, but not all are <= 127 23:16:27 <|amethyst> SamB: and K is irrelevant since e_min is zero 23:17:00 <|amethyst> okay, so how to fix it properly, because that double-casting is a bad hack 23:17:20 ninka MSVC 23:17:25 er, ninja 23:18:21 Wrap it up with #ifdef TARGET_COMPILER_VC 23:18:23 ? 23:19:21 <|amethyst> that doesn't make it any less ugly :) 23:20:12 From the little I've read on bitfields, it sounds like their behavior is compiler specific. So even if it's ugly, it's probably justified. 23:21:10 <|amethyst> if we're going to use that fix I'd rather do it unconditionally 23:21:33 xFleury: me and |amethyst just read it in the standard that this is perfectly acceptable 23:21:37 <|amethyst> since MSVC might not be the only compiler or compiler version that does that 23:21:40 or should I say I 23:21:53 what WE do, I mean, not what the compiler does 23:22:36 <|amethyst> SamB: what's the proper way to cite a paragraph? 7.2 (7) ? 23:22:54 what, no paragaph symbol? 23:23:14 * SamB is kind of surprised they never cite paragraphs in the standard 23:23:32 -!- kingbuddyboy has quit [Quit: Page closed] 23:24:06 -!- bh has quit [Quit: Page closed] 23:26:34 -!- N78291 has quit [Quit: null] 23:31:53 -!- Xenobreeder has quit [Quit: Page closed] 23:33:02 <|amethyst> hm 23:33:14 <|amethyst> I guess another fix would be to change that field to a uint8_t 23:33:31 -!- ystael has quit [Ping timeout: 245 seconds] 23:33:33 <|amethyst> it would require inserting some casts, but we have to insert casts either way 23:33:34 not as nice in gdb though 23:33:41 -!- rapierx has quit [Ping timeout: 250 seconds] 23:33:43 <|amethyst> hm, true 23:34:29 -!- darktwinge has joined ##crawl-dev 23:35:36 -!- bh has quit [Changing host] 23:35:45 -!- Kalir has quit [Changing host] 23:36:59 why is "asm" a declaration and not a statement ... 23:37:23 and why is the production called "asm-definition" ... 23:37:35 -!- Twinge has quit [Ping timeout: 272 seconds] 23:39:45 hmm, is it an actual requirement of C++ that there be a C compiler with a compatible ABI? 23:42:42 <|amethyst> SamB: probably a declaration so that it's possible to use it at file scope 23:42:56 hmm 23:42:59 -!- bh has quit [Quit: Page closed] 23:43:20 <|amethyst> in case an implementation wanted to use that for defining entire functions in assembly, for example 23:43:56 now that you mention it, there are other things one could do with it at that scope. 23:44:38 (things having to do with data, perhaps) 23:44:58 <|amethyst> yeah 23:45:02 <|amethyst> or sections 23:45:17 <|amethyst> e.g. "everything that follows goes into section .foo" 23:45:27 yes, you'd likely want to put that data in sections anyway 23:45:41 -!- qoala has quit [] 23:46:15 the "everything that follows" probably wouldn't work too well as I expect the compiler often changes sections explicitly 23:46:28 -!- eb has quit [Read error: Connection reset by peer] 23:46:42 Unstable branch on crawl.lantea.net updated to: 0.14-a0-277-g660d43f (34) 23:48:16 -!- Flun has quit [Quit: Page closed] 23:48:58 <|amethyst> true 23:49:10 <|amethyst> I was thinking code sections, not data 23:51:28 -!- ontoclasm has quit [Quit: Leaving.] 23:54:06 -!- indspenceable has joined ##crawl-dev 23:54:44 GCC has a better way to do that 23:55:28 -!- Stelpa has quit [Remote host closed the connection] 23:55:29 -!- heteroy has quit [Read error: Connection reset by peer] 23:55:47 -!- zardo has quit [Quit: Page closed]