01:28:05 -!- amalloy is now known as amalloy_
01:46:57 <Kelbell> Fork (gooncrawl) on crawl.kelbi.org updated to: 0.22-a0-227-geda511d7be
02:05:39 <Kelbell> Webtiles server stopped.
02:05:47 <Kelbell> Webtiles server started.
10:36:24 -!- MarvinPA_ is now known as MarvinPA
13:03:37 -!- amalloy_ is now known as amalloy
13:15:33 <Henzell> Stable (0.20) branch on crawl.akrasiac.org updated to: 0.20.1-2-g1830e61
13:23:21 <Cheibriados> 03kitchen-ace02 07https://github.com/crawl/crawl/pull/995 * 0.24-a0-140-gcf7bdf0: Merge Piercer into the storm bow 10(3 weeks ago, 3 files, 8+ 7-) 13https://github.com/crawl/crawl/commit/cf7bdf0f6669
13:39:05 <Cheibriados> sdl2-mixer in contrib directory fails to build in Arch Linux 13https://crawl.develz.org/mantis/view.php?id=11931 by kitchen_ace
16:03:28 <floraline> chat bots shouldn't increment the spectator count in the lobby
16:14:08 <amalloy> floraline: you'd have to teach the webtiles code who's a chat bot
16:14:18 <floraline> yeah i'm wondering what the best way to do that is
16:14:22 <amalloy> and anyone can always write another if they want
16:14:26 <floraline> maintain a blacklist?
16:14:41 <floraline> or chat users can just choose not to increment the counter?
16:14:54 <amalloy> what is the goal of this? why does it matter that a chat bot is a "spectator"?
16:15:08 <floraline> i think it's misleading and looks bad
16:15:54 <floraline> a spattering of users in the lobby have one spectator, but it's not really a human there
16:16:10 <floraline> it teaches users that the spectator count does not really indicate that there is a discussion happening or other people to watch with
16:16:17 <floraline> and ultimately they will just ignore it
16:16:42 <Kelbell> Fork (gooncrawl) on crawl.kelbi.org updated to: 0.22-a0-814-g8c131ce026
16:16:46 <geekosaur> ...but this is always true to some extent.
16:16:50 <amalloy> i imagine there are often humans watching a game who don't talk
16:17:14 <floraline> well sure but that is still different from a bot
16:17:34 <floraline> if you know humans are watching someone it might suggest that you maybe want to see it too
16:17:34 <geekosaur> how do you tell the difference?
16:17:36 <amalloy> is it? if you want users to believe that #spectators means "someone to talk to', it's the same
16:18:45 <amalloy> tbh if there's a game going on with exactly one spectator, and that spectator is a chat bot, it means that whoever's playing is knowledgeable enough to ask to be spectated, and that they probably actively use the bot - it's probably a decent game to watch
16:19:28 <floraline> that's confusing to someone who doesn't know all of the context
16:19:45 <floraline> a new user might assume it is a person, and they click through to learn about spectating and chatting
16:20:52 <floraline> but almost every game they click on would have the same spectator who doesn't talk, maybe they would conclude it's a bot or maybe they wouldn't
16:21:03 <floraline> either way i think it's just not the best experience for new users
16:21:27 <floraline> i guess it comes down to whether a chat bot is considered a spectator
16:21:42 <floraline> and what users expect when they see the spectator count
16:21:58 <floraline> they have to guess what it means and i think it should mean the thing they're most likely to guess
16:54:12 <advil> I mean, having accounts marked as bots in some formal way would probably be a useful thing in general, so something like this would be good
16:54:33 <advil> I'm not sure how easy it would be to do, though
16:55:32 <advil> something hacky might not be that bad
16:56:48 <advil> how do we do it for the tournament scripts? I assume at some point someone just made a list of bot accounts?
16:57:16 <advil> (obv crawl-playing bots and chatbots aren't quite the same thing, but they both use the same kind of user account)
17:18:45 <ebering> the tournaments don't pick up beem because beem doesn't play stable games to generate milestones
17:21:32 <geekosaur> and they do pick up qw, etc., they're just not generally good enough to show much
17:26:47 <geekosaur> (also for some reason people want to know if humans have been overthrown by the bots yet)
17:45:55 <alexjurk1> floraline: i agree, this would be a good change
17:46:15 <alexjurk1> floraline: there are only a few bots, so a manual blacklist (based on a text file that can be edited) would probably be a great way to start
17:47:03 <alexjurk1> floraline: agree on your points about spectator count being currently misleading and about it being qualitatively different to know you are being watched by a human vs nobody/bots
18:00:48 <gammafunk> yeah maybe you could add a special spectator count blacklist
18:00:56 <gammafunk> bots have to be hard-coded into the tournament scripts too, re: earlier question
18:01:05 <gammafunk> we don't have an elegant system for that either
18:08:39 -!- MarvinPA_ is now known as MarvinPA
18:10:17 <gammafunk> floraline: and yeah, to be clear, the only way this could work would be if you had list to track them
18:12:46 <gammafunk> aside from having some special bot api like services such as discord use
18:17:24 <advil> ebering yeah, I was talking about qw-type bots, but in some ways it's exactly the same thing
18:28:15 <Napkin> gammafunk: hihihi@last wordpress post ;-)
18:32:24 <gammafunk> Napkin: wait, which one?
18:32:48 <gammafunk> or maybe I don't know what you mean
18:33:18 <gammafunk> ooh
18:33:29 <gammafunk> and what's this, I get to read advil's draft of an upcoming blog post
18:33:38 <floraline> hey that's cheating
18:34:09 <gammafunk> too bad, devs cheat all the time!!
18:42:59 <gammafunk> Napkin: if you mean the one about thanking your for hosting scripts on CDO, no problem. It worked just fine in practice, so thanks again for making that available
18:43:28 <Napkin> nah, the bugfix 0.23.1 post
18:44:14 <Napkin> ja? glad to hear the scripts had enough cpu & disk speed
18:44:29 <Napkin> believe me, no btrfs for the next server...
18:51:59 <Napkin> anyways, great posting, nice writing :)
19:00:51 <gammafunk> glad you enjoyed
19:19:50 <advil> heh yeah I'll probably post a version of that around when 0.23.2 is released
19:22:14 <advil> still needs some work
20:13:15 -!- amalloy is now known as amalloy_
20:38:15 <alexjurk1> how do I invert this test? if (item.flags & ISFLAG_RANDART)
20:49:02 <Yermak> if (!(item.flags & ISFLAG_RANDART)) ?
20:51:29 <alexjurk1> yeah. I just assume there is some crazy bitflag way to do it which is better
20:52:03 <Yermak> Which is better than 2 bit operations, i.e. one?
20:53:21 <Yermak> Or better in aesthetic way?
20:53:26 <alexjurk1> the latter
20:54:09 <Yermak> How about  if (~item.flags & ISFLAG_RANDART) ?
21:23:29 <gammafunk> advil: one small suggestion I had looking at the draft was to explain in a touch more detail what seeded play actually is; I think some readers might not be familiar
21:23:38 <gammafunk> and your description might be a bit too brief for them
21:24:47 <gammafunk> looks like it'll be an interesting writeup; I think we haven't had one like that since neil last talked about how webtiles works and how that horrible save compat bug got fixed before that
22:30:30 <alexjurk1> when running make, there's a big delay between the line "Configured with:..." and the build actually starting (well, a delay of a few secs). What code is running in this time?
22:30:42 <alexjurk1> I can't see the string 'Configured with' in the codebase
22:31:31 <alexjurk1> actually, I see it in a lot of submodules
22:34:10 <gammafunk> alexjurk1: I don't get any such line when I run make
22:34:31 <gammafunk> is that part of your build script somehow?
22:36:04 <gammafunk> hrm, also
22:36:16 <gammafunk> In file included from acquire.cc:6:0:
22:36:16 <gammafunk> AppHdr.h:342:0: warning: "DEBUG_STATISTICS" redefined
22:36:16 <gammafunk> #define DEBUG_STATISTICS
22:36:24 <gammafunk> when I try to build tiles with debug build
22:37:08 <gammafunk> using nice make debug -j3 TILES=y USE_PCRE=y EXTERNAL_FLAGS_L=-DDEBUG_STATISTICS
22:37:28 <gammafunk> are we enabling debug stats by default now?
22:38:24 <gammafunk> looks like it's enabled whenever debug diagnostics is
22:45:38 <alexjurk1> $ nice make -j8
22:45:39 <gammafunk> hrm, maybe I just copied this alias wrong
22:45:40 <alexjurk1> Configured with: --prefix=/Applications/Xcode.app/Contents/Developer/usr --with-gxx-include-dir=/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.14.sdk/usr/include/c++/4.2.1
22:45:42 <alexjurk1> /Applications/Xcode.app/Contents/Developer/usr/bin/make -C rltiles all ARCH=x86_64-apple-darwin18.2.0 NO_PKGCONFIG=Yes TILES=
22:45:44 <alexjurk1> make[1]: `levcomp.lex.cc' is up to date.
22:45:46 <alexjurk1> this is what I see when building on osx
22:45:56 <gammafunk> that might be a clang thing?
22:46:12 <gammafunk> or something about the os x build, not sure
22:46:35 <alexjurk1> it really slows down builds, it seems to happen before make actually starts
22:46:47 <gammafunk> yeah, it's something os-x specific it seems
22:46:56 <gammafunk> probably not clang as that's called via make
22:47:11 <gammafunk> I'd look at what the os x build is doing
22:48:47 <alexjurk1> yeah. might need to pull out the macos equivalent of strace
23:05:22 <advil> I see the "configured with" line but I don't get a particularly noticeable delay there
23:07:01 <floraline> when's the next dcss convention
23:10:41 <alexjurk1> advil: how long does a no-op `time make` take? it's about 4.7secs on my system when run repeatedly
23:11:18 <advil> about 3s
23:12:15 <advil> when I do it that way, it does seem like most of that time is between "Configured with" and the output of the build parameters, but it's probably doing all sorts of stuff there
23:27:39 <alexjurk1> yeah. That's the part I am trying to speed up
23:27:47 <alexjurk1> but not even sure what code is executing
23:32:05 -!- amalloy_ is now known as amalloy
23:34:18 <geekosaur> that looks like something did gcc -V, capturing stdout, but that line goes to stderr?
23:34:50 <geekosaur> (note that the Makefile does probe what "gcc" is installed, even if you tell it to use clang)
23:36:19 <geekosaur> gcc -v, not -V
23:36:59 <geekosaur> but the invocations I see have 2>&1 on them
23:42:07 <geekosaur> also the Makeifle does a lot of other $(shell ...) stuff before anything else
23:42:23 <geekosaur> and stat-ing every source and object file