00:28:43 <Eksell> Unstable branch on CRAWL.XTAHUA.COM updated to: 0.21-a0-372-gd0ff781 (34)
01:20:40 <Gretell> Unstable branch on crawl.develz.org updated to: 0.21-a0-372-gd0ff781 (34)
02:00:22 <Gretell> Windows builds of master branch on crawl.develz.org updated to: 0.21-a0-372-gd0ff781
02:55:08 <Gretell> Monster database of master branch on crawl.develz.org updated to: 0.21-a0-372-gd0ff781
03:21:29 <Rotatell> Unstable branch on crawl.beRotato.org updated to: 0.21-a0-372-gd0ff781 (34)
03:22:33 <aidanh> Is the gold_turn item in tile_layout_priority completely useless? As far as I can see, it increases the HUD height by one line, but I don't see any more text..
03:39:02 <geekosaur> I am under the (possibly mistaken) impression the functionality was removed but the entry was left so as to not completely break people's games through rc file incompatibility
03:42:32 <aidanh> Makes sense; is there any mechanism for warning the user about outdated config?
03:43:41 <aidanh> I'll remove what's left of the functionality and update the options guide then
03:46:00 <geekosaur> for  unix you get a message which will disappear as the game starts but come back when you exit it. this doesn't work on windows except in console mode because windows console is a bit weird, and it doesn't work at all for servers because the message gets erased too fast
04:00:50 <aidanh> What about using mpr or something like that?
04:04:14 <Cheibriados> 03Aidan Holm02 07https://github.com/crawl/crawl/pull/620 * 0.21-a0-369-g045f200: Remove useless visual separator 10(15 hours ago, 1 file, 5+ 13-) 13https://github.com/crawl/crawl/commit/045f20046539
04:04:14 <Cheibriados> 03Aidan Holm02 07https://github.com/crawl/crawl/pull/620 * 0.21-a0-370-gf60cf24: Ensure TabbedRegion is given enough height to render tab icons 10(60 minutes ago, 3 files, 14+ 3-) 13https://github.com/crawl/crawl/commit/f60cf2481fea
04:04:14 <Cheibriados> 03Aidan Holm02 07https://github.com/crawl/crawl/pull/620 * 0.21-a0-371-g52fd115: Remove dead code, update docs for tile_layout_priority 10(8 minutes ago, 4 files, 4+ 20-) 13https://github.com/crawl/crawl/commit/52fd1154bd01
05:51:33 -!- amalloy is now known as amalloy_
05:53:46 -!- amalloy_ is now known as amalloy
06:41:45 -!- nosratheno_ is now known as nosratheno
13:23:55 <Henzell> Unstable branch on crawl.akrasiac.org updated to: 0.21-a0-372-gd0ff781 (34)
17:39:43 <Cheibriados> 03Aidan Holm02 07https://github.com/crawl/crawl/pull/620 * 0.21-a0-373-g02fdb0d: Add GL render clipping support for local tiles 10(4 days ago, 3 files, 18+ 0-) 13https://github.com/crawl/crawl/commit/02fdb0d27041
17:39:43 <Cheibriados> 03Aidan Holm02 07https://github.com/crawl/crawl/pull/620 * 0.21-a0-374-g2fda903: Fix sidebar monster/inv background flicker on resize 10(4 days ago, 4 files, 4+ 23-) 13https://github.com/crawl/crawl/commit/2fda90317f96
17:39:43 <Cheibriados> 03Aidan Holm02 07https://github.com/crawl/crawl/pull/620 * 0.21-a0-375-g8a7175e: Fix tiles minimap viewport indicator leaving the sidebar 10(4 days ago, 1 file, 2+ 0-) 13https://github.com/crawl/crawl/commit/8a7175ed40fd
17:39:43 <Cheibriados> 03Aidan Holm02 07https://github.com/crawl/crawl/pull/620 * 0.21-a0-376-g70398a4: Improve layout+spacing of tiles main dungeon view 10(4 days ago, 4 files, 38+ 7-) 13https://github.com/crawl/crawl/commit/70398a4624a6
17:39:43 <Cheibriados> 03Aidan Holm02 07https://github.com/crawl/crawl/pull/620 * 0.21-a0-377-g9c746a7: Fix a few tiles rendering glitches 10(2 days ago, 1 file, 24+ 6-) 13https://github.com/crawl/crawl/commit/9c746a7f6e35
17:39:43 <Cheibriados> 03Aidan Holm02 07https://github.com/crawl/crawl/pull/620 * 0.21-a0-378-g7b4f9c3: Clean unused code, add assertion 10(29 hours ago, 1 file, 1+ 9-) 13https://github.com/crawl/crawl/commit/7b4f9c3e87f0
17:39:43 <Cheibriados> 03Aidan Holm02 07https://github.com/crawl/crawl/pull/620 * 0.21-a0-379-g581a512: Fix menu rendering when window size is reduced 10(2 days ago, 2 files, 28+ 22-) 13https://github.com/crawl/crawl/commit/581a512b3834
17:39:43 <Cheibriados> 03Aidan Holm02 07https://github.com/crawl/crawl/pull/620 * 0.21-a0-380-gea4b691: Stop drawing current menu page after a MEL_TITLE entry 10(29 hours ago, 2 files, 19+ 10-) 13https://github.com/crawl/crawl/commit/ea4b691112d8
17:39:43 <Cheibriados> 03Aidan Holm02 07https://github.com/crawl/crawl/pull/620 * 0.21-a0-381-g5fb22d6: Remove useless visual separator 10(29 hours ago, 1 file, 5+ 13-) 13https://github.com/crawl/crawl/commit/5fb22d6c1880
17:39:43 <Cheibriados> 03Aidan Holm02 07https://github.com/crawl/crawl/pull/620 * 0.21-a0-382-g1e22c19: Ensure TabbedRegion is given enough height to render tab icons 10(15 hours ago, 3 files, 14+ 3-) 13https://github.com/crawl/crawl/commit/1e22c190e9fb
17:39:43 <Cheibriados> ... and 4 more commits
17:50:39 <gammafunk> commit storm
17:53:04 <geekosaur> at least they're committed to it...
17:53:17 <aidanh> groan
18:03:12 <Lantell> Unstable branch on underhound.eu updated to: 0.21-a0-372-gd0ff781 (34)
18:29:22 <alexjurkiewicz> reasons why hellcrawl is better than vanilla crawl #37: f9a9639 hellmonk New background: Torpor Knight.
18:32:36 <gammafunk> reasons why hellcrawl is worse than vanilla crawl #37: f9a9639 hellmonk New background: Torpor Knight.
18:33:15 <wheals_> why have playable torpor knights when you could have playable torpor snails
18:33:18 <gammafunk> !lg chequers chei won
18:33:19 <Sequell> No games for chequers (chei won).
18:33:24 <gammafunk> gee, I'm not surprised
18:33:44 <gammafunk> that's true, use a wand of enslavement on them
18:33:49 <gammafunk> too bad you don't get the passive slow
18:36:48 <amalloy> i assume the global winrate for TK is 100%
18:37:03 <gammafunk> obv
18:37:40 <gammafunk> hrm
18:37:44 <gammafunk> !lg . chei / won
18:37:45 <Sequell> 2/8 games for gammafunk (chei): N=2/8 (25.00%)
18:37:49 <gammafunk> !lg amalloy chei / won
18:37:50 <Sequell> 7/15 games for amalloy (chei): N=7/15 (46.67%)
18:37:55 <gammafunk> ok, that's good at least
18:38:06 <gammafunk> as the only dev who likes chei (and who can win games of crawl)
18:38:08 <amalloy> !lg . s=god/won
18:38:09 <Sequell> 65/418 games for amalloy: 7/8x Trog [87.50%], 7/15x Cheibriados [46.67%], 5/19x Okawaru [26.32%], 4/13x Ru [30.77%], 3/5x Gozag [60.00%], 3/4x Jiyva [75.00%], 3/18x Ashenzari [16.67%], 3/19x Lugonu [15.79%], 3/6x Fedhas [50.00%], 3/9x Makhleb [33.33%], 3/7x Zin [42.86%], 3/6x Vehumet [50.00%], 2/15x Qazlal [13.33%], 2/6x Dithmenos [33.33%], 2/2x Hepliaklqana [100.00%], 2/9x Kikubaaqudgha [22.22%],...
18:38:13 <gammafunk> at least you represent a decent che winrate
18:38:31 <amalloy> !lg . s=god/won o=%
18:38:32 <Sequell> 65/418 games for amalloy: 1/1x Beogh [100.00%], 2/2x Hepliaklqana [100.00%], 7/8x Trog [87.50%], 3/4x Jiyva [75.00%], 3/5x Gozag [60.00%], 3/6x Fedhas [50.00%], 1/2x Uskayaw [50.00%], 3/6x Vehumet [50.00%], 7/15x Cheibriados [46.67%], 3/7x Zin [42.86%], 2/6x Dithmenos [33.33%], 1/3x Pakellas [33.33%], 3/9x Makhleb [33.33%], 4/13x Ru [30.77%], 5/19x Okawaru [26.32%], 2/9x Kikubaaqudgha [22.22%], 1/...
18:38:41 <gammafunk> the thing is
18:38:45 <gammafunk> !lg amalloy chei s=crace
18:38:46 <Sequell> 15 games for amalloy (chei): 3x Octopode, 3x Formicid, 2x Naga, Gargoyle, Deep Elf, Troll, Ghoul, Minotaur, Demonspawn, Vine Stalker
18:39:10 <amalloy> !lg amalloy chei s=crace / won o=%
18:39:11 <gammafunk> that's a decent mix, still a lot of races with strong early game
18:39:11 <Sequell> 7/15 games for amalloy (chei): 1/1x Vine Stalker [100.00%], 1/1x Deep Elf [100.00%], 1/1x Minotaur [100.00%], 2/3x Formicid [66.67%], 1/2x Naga [50.00%], 1/3x Octopode [33.33%], 0/1x Ghoul [0.00%], 0/1x Troll [0.00%], 0/1x Demonspawn [0.00%], 0/1x Gargoyle [0.00%]
18:39:39 <gammafunk> good 0% tr of chei winrate
18:39:46 <amalloy> that was like my second online game
18:39:57 <amalloy> !lg . tr chei
18:39:58 <Sequell> 1. hyperamalloy the Brawler (L16 TrVM of Cheibriados), slain by the 27-headed Lernaean hydra (kmap: swamp_old_school) on Swamp:4 on 2015-12-11 02:34:41, with 146402 points after 34494 turns and 2:52:33.
18:40:04 <amalloy> rip, i guess not
18:40:05 <gammafunk> it was a hyper account!
18:40:12 <amalloy> i played Tr^Chei offline very early
18:40:20 <gammafunk> My first chei game was ghee
18:40:33 <gammafunk> and excellent character to teach you how awful chei is
18:40:53 <amalloy> i wonder why i let myself get killed by TLH
18:41:03 <gammafunk> lerny deaths are kind of funny that way
18:41:04 <amalloy> i vaguely remember trying to empty some wands into him
18:41:24 <gammafunk> !lg * cikiller=the_lernaean_hydra
18:41:26 <Sequell> 1958. astranger the Fanfare of Bloodshed (L16 HOSu of Makhleb), slain by the 27-headed Lernaean hydra on Swamp:4 (grunt_swamp_rune_castle) on 2017-10-16 01:26:25, with 99535 points after 31054 turns and 3:36:34.
18:41:28 <gammafunk> !lg . cikiller=the_lernaean_hydra
18:41:29 <amalloy> !lg . tr chei -tv
18:41:30 <Sequell> No games for gammafunk (cikiller=the_lernaean_hydra).
18:41:31 <Sequell> 1. hyperamalloy, XL16 TrVM, T:34494 requested for FooTV: telnet://termcast.develz.org or http://termcast.develz.org.
18:41:38 <gammafunk> oh I guess I haven't died to it save on speedruns
18:42:33 <gammafunk> anyhow that change was dumb, because he's also removing AK, which is more interesting than TK (albeit not much more)
18:42:35 <amalloy> wow what awful play
19:39:06 <Cheibriados> 03advil02 07* 0.21-a0-373-gcbc6281: Further tweak restart behavior and options (Zibokwe) 10(18 minutes ago, 12 files, 142+ 69-) 13https://github.com/crawl/crawl/commit/cbc62810bc49
19:39:16 <advil> working on game end/restart/startup stuff is a bit unnerving
19:39:38 <advil> hopefully I didn't completely screw anything up
19:40:24 <alexjurkiewicz> hellmonk: good snail knight
19:40:28 <alexjurkiewicz> hellmonk: release soon?
19:40:35 <hellmonk> probably this weekend
19:41:58 <alexjurkiewicz> 😎
19:44:20 <Cheibriados> 03advil02 07* 0.21-a0-374-gaf2b63a: Delete an obsolete comment that I missed 10(25 seconds ago, 1 file, 0+ 2-) 13https://github.com/crawl/crawl/commit/af2b63a13872
19:49:08 <|amethyst> advil: +        return ret = ret && Options.restart_after_save;
19:49:43 <|amethyst> why both the "ret =" and the "return" when the next line is "return ret"?
19:49:43 <advil> ugh
19:49:43 <advil> that's just a typo
19:49:55 <advil> I guess it works by accident
19:50:13 <advil> hm, maybe leaving should have the same behavior as winning
19:50:13 <|amethyst> could just make that
19:50:38 <advil> you're going to do the || thing I had before I bet
19:50:44 <|amethyst> oh, I guess it's ugly ugly isn't it
19:50:59 <|amethyst> ah, maybe not so bad:
19:51:47 <|amethyst> never mind
19:51:51 <advil> heh
19:51:52 <|amethyst> it's even uglier with ?:
19:52:07 <advil> I thought this read more clearly than the one-line || version, basically
19:53:35 <advil> though I have this one bool calculation in there in startup.cc that makes me want to punch myself
19:54:24 <|amethyst> advil:  can't you drop the && crawl_state.last_game_exit != GAME_EXIT_ABORT ?
19:54:34 <|amethyst> advil:  since that would pass the ||
19:55:22 <|amethyst> err, wouldn't pass
19:55:56 <advil> ah true
19:56:14 <|amethyst> putting that || into a function might help
19:56:43 <|amethyst> also the ( ) around the last && is unnecessary
19:56:50 <|amethyst> since && is associative
19:57:34 <advil> it went through several iterations with various negations & applications of demorgan's law, that's probably what happened there
20:00:15 <|amethyst> also, I'd consider making that an enum class, though that does mean you'd probably want to drop type from the name for conciseness
20:00:38 <|amethyst> enum class game_exit { unknown, won, ... };   then   game_exit::unknown when you want to use it
20:01:53 <|amethyst> the main advantage is that it doesn't implicitly convert to integer
20:02:01 <|amethyst> but also less uppercase :)
20:02:14 <advil> I did briefly think about that, but didn't do it
20:02:26 <|amethyst> I'm slowly working on converting all the ones where we don't need to do arithmetic
20:02:33 <advil> I see
20:02:39 <|amethyst> so I'll get to it eventually if you don't, it just might be a few years :)
20:02:43 <advil> heh
20:04:06 <advil> btw I discovered one more thing NORETURN does
20:04:41 <advil> oh, nevermind
20:05:35 <amalloy> well now i'm curious what you thought
20:05:54 <advil> I thought it prevent warnings about fallthroughs, but I think we don't have those on at all
20:05:59 <advil> in switches
20:06:13 <gammafunk> btw, will anyone be able to look over aidanh's nice Tiles PR?
20:06:27 <advil> I'm planning on looking over it, I've already glanced at the first parts
20:06:31 <gammafunk> cool, thanks
20:06:40 <advil> haven't run any of it
20:06:52 <gammafunk> I know...running Tiles feels so...wrong...
20:07:07 <advil> and I apparently decided to rewrite how crawl does restarts
20:07:15 <gammafunk> nice
20:07:31 <gammafunk> I added an option to force a map in mapstat/objstat as an aside when working on wizlab maps
20:07:31 <advil> startup is so incredibly complicated in this game
20:07:36 <gammafunk> yes
20:08:12 <gammafunk> I actually really don't like how crawl always exits you out of the program aggressively
20:08:19 <gammafunk> like you can't really get back to the main menu in some situations
20:08:27 <advil> I think I fixed a bunch of that!
20:08:30 <gammafunk> nice
20:08:42 <advil> all the cases where crawl used game_ended to exit out of a menu should now always restart
20:08:54 <gammafunk> that'll be great
20:09:00 <advil> (except for dgl, and webtiles under certain situations)
20:09:16 <gammafunk> quite the opposite of the 90s games dcss is suppose to be inspired by!
20:09:30 <gammafunk> no "Are you sure you don't want to kill more demons for Makhleb?" prompt or anything
20:11:18 <advil> also I changed the default for restart_after_game to a version of true
20:11:26 <advil> where before it was false for console
20:20:45 <Jorgrell> Unstable branch on crawl.jorgrun.rocks updated to: 0.21-a0-374-gaf2b63a (34)
20:28:24 <amalloy> fwiw i have restart_after_game set to true for online play, but i like it false locally since i mostly just do quick wizmode testing of stuff
20:29:15 <gammafunk> does that preserve your webtiles session I wonder
20:29:17 <amalloy> or iterative stuff when i'm changing the source and i want it to be quick to quit out and recompile a new version. obviously the stuff i like doesn't have to be the defaults, just i enjoy offering opinions
20:32:12 <advil> does it do anything online?
20:32:15 <advil> it shouldn't
20:33:14 <advil> both before and after my changes, having DGAMELAUNCH prevents it from being set in the rc
20:33:22 <advil> yeah, I think the default should be what players will expect, it's easy enough to override in an rc for us
20:34:27 <advil> well, what players will expect if they aren't used to the current quirky defaults
20:45:23 <Cheibriados> 03advil02 07* 0.21-a0-375-g7cd14ec: Various cleanups, including enum classing the game exit type (|amethyst) 10(5 minutes ago, 12 files, 78+ 74-) 13https://github.com/crawl/crawl/commit/7cd14ec42818
21:00:29 <amalloy> advil: doesn't it? i thought online it closed the ssh connection after game end if you didn't specify otherwise
21:00:47 <amalloy> otoh i assume you have tested recently it and know better than i do
21:01:13 <advil> maybe that once happened?
21:01:14 <advil> if there's a non-dgl server it might do that
21:04:38 <advil> aidanh: one picky thing that someone will give you a hard time about sooner or later is the bracketing conventions, which have {} only ever on a line by themselves
21:04:48 <advil> [someone such as me, right now :)]
21:05:14 <advil> full description here: https://github.com/crawl/crawl/blob/master/crawl-ref/docs/develop/coding_conventions.txt , though that's a bit over-elaborate
21:06:11 <geekosaur> amalloy, it umps you to the dgl menu, you only disconnect with 'q' there
21:07:04 <geekosaur> but in so doing, it bypasses the game's main menu (which is why the dgl menu duplicates it, multiple times for each game variant and each supported version)
21:10:12 <advil> !source command.cc::809
21:10:13 <Sequell> Can't find command.cc::809.
21:10:16 <advil> !source command.cc:809
21:10:16 <Sequell> https://github.com/crawl/crawl/blob/master/crawl-ref/source/command.cc#L809
21:10:40 <advil> what is supposed to be going on with 'first' there?
21:11:01 <advil> (one of aidanh's commits cleans that up but I can't even figure out what the original code is doing)
21:13:24 <advil> maybe it ensures that at least one command_type is passed?
21:14:39 <amalloy> advil: i think so, yes
21:14:51 <geekosaur> no, it's required for stdarg to work
21:15:24 <amalloy> oh. right. so you know what type the ...s will have?
21:15:30 <geekosaur> on some architectures, the only way to reference arguments requires that you be able to refer to the first of them. which is what va_start is doing
21:15:53 <amalloy> no, i guess not that
21:16:22 <geekosaur> on x86 va_start doesn;t so anything that couldn;t be done some other way. but there are architectures on which that's the only way to get to *any* of the stdarg/varargs parameters
21:16:54 <geekosaur> *do anything
21:17:51 <geekosaur> (also yes, in theory you could start that at desc instead, but that forces desc into the varargs part. which is a deoptimization on any platform that allows register passing)
21:18:04 <advil> I thought va_start was called with the last non-variable parameter?
21:18:45 <geekosaur> how do you index from a register to the stack?
21:19:11 <advil> I have no idea I'm just reading some documentation :)
21:19:35 <geekosaur> (also, uh, there are platforms where *stack* doesn;t work that way; stack is a relatively recent concept, see algol 60 and how it slipped recursion into the languagew)
21:20:49 <geekosaur> on those platforms, varargs has to be done via some kind of descriptor block and you need a pointer to that block, so the parameter used by stdarg to reach that block necessarily resides in that block and not where the fixed parameters go
21:21:38 <geekosaur> so yes, you don't necessarily *need* first, but it's cleaner and there are optimization considerations (varargs can be *much* slower than fixed)
21:27:21 <advil> hmm
21:27:28 <advil> I think the intent here is to ensure that there's at least one argument
21:28:47 <geekosaur> if that was the intent, it is flawed. stdarg only *works* if there is at least one varargs parameter
21:29:51 <geekosaur> because the anchor point you pass to va_start is itself forced to be varargs
21:30:22 <advil> I guess I'm a bit confused, isn't printf a counterexample?
21:30:39 <geekosaur> also note printf --- whether there are any actual variable parameters is determined by the content of the format string. you *cannot* determine it by making it the anchor, nor by va_start / va_arg
21:31:22 <advil> in these functions it is supposed to be determined by hitting an argument that is a 0
21:31:30 <geekosaur> as I already said, the only difference here between using desc vs. using first is whether you take a significant hit by pushing desc into the varargs block
21:31:34 <advil> so it always needs one
21:32:02 <advil> I know I'm very inexperienced with varargs but what you're saying is just not what the docs for <cstdarg> say :)
21:32:08 <geekosaur> because using itwith va_start forces it into the varargs block
21:32:35 <geekosaur> no, cstdarg gives the viewpoint for a C/C++ programmer's use of it
21:33:00 <geekosaur> I'm talking from a low level implementation view, because that is ultimately *why* it works the way it does
21:33:40 <geekosaur> cstdarg doesn;t want to do that because doing so would limit implementation choices --- some of which I've mentioned
21:34:11 <geekosaur> (and you need those choices because usually it's the cpu architecture that limits how you implement it)
21:34:43 <|amethyst> FR: change that so it doesn't use varargs
21:34:55 <|amethyst> you'd just need to add one set of {}s to the calls
21:34:56 <advil> this varargs thing it is doing is definitely bad
21:35:10 <advil> if you don't give a 0 somewhere in the sequence, who knows what it'll do
21:35:14 <|amethyst> C varargs is generally bad, yes
21:35:18 <advil> memory corruption I assume
21:35:24 <|amethyst> from a type-safety point of view
21:35:39 <|amethyst> probably not memory *corruption* since it just reads from them
21:35:45 <|amethyst> just a crash
21:35:52 <advil> ah right
21:36:13 <advil> I guess that's why whoever wrote this has a hard cap of 10 varargs
21:38:09 <advil> what aidanh noticed is that if the first argument to that function is 0, it still starts collecting varargs
21:38:35 <geekosaur> yeh, the usual failure mode here is a call just keeps going until it hit unallocated memory and segfaults
21:38:39 <advil> the `first` argument, not the first argument
21:39:08 <gammafunk> gotta love argument names like "first"
21:39:16 <gammafunk> I guess if "first" is the first argument it's fine
21:40:09 <|amethyst> _add_insert_commands(column_composer &cols, int column, const string &desc, vector<command_type> cmds, int space_to_colon)
21:40:45 <|amethyst> then you can also combine the two overloads since you can just put the optional space-to-colon number at the end after the list
21:40:55 <|amethyst> I guess there's probably something better than vector<> there
21:41:45 <|amethyst> I guess initializer_list<command_type>
21:41:47 <advil> There's some other differences between those functions that don't make any obvious sense
21:42:21 <advil> one calls add_formatted, the other calls something else
21:43:16 <|amethyst> oh wow
21:43:39 <|amethyst> the first one (with the space_to_colon argument) doesn't put the first command onto the vector, but instead adds it last...
21:43:47 <advil> yeah
21:44:19 <advil> both will have this va bug, which should actually trigger a bunch
21:44:25 <advil> how do we even have a working help menu?
21:44:36 <|amethyst> ?
21:44:43 <|amethyst> when do we pass zero as the first command?
21:44:57 <advil> _add_insert_commands(cols, 0, "                  \\|/        \\|/", 0);
21:45:01 <advil> for example
21:45:09 <|amethyst> oh wow
21:45:12 <|amethyst> why is that even
21:45:24 <advil> ikr
21:45:27 <|amethyst> yeah, that's a potential problem
21:46:13 <|amethyst> it probably happens to work because one of those slots on the stack happens to be zero from the previous call
21:46:19 <advil> heh
21:46:36 <advil> it's possible aidanh did notice a crash from this, I couldn't tell from the commit message
21:48:42 <|amethyst> oh
21:48:49 <|amethyst> I see why the first argument is different
21:48:59 <|amethyst> that one gets stuck at the very beginning
21:49:06 <|amethyst> subsequent ones get pasted in for %s
21:49:21 <|amethyst> o
21:49:22 <|amethyst> r
21:49:24 <|amethyst> hm
21:50:21 <|amethyst> in the version with the extra argument only...
21:50:53 <|amethyst> that is extremely
21:51:20 <|amethyst> I don't feel like opening twitter, and it's hard to pick out just a few lines anyway
21:51:32 <|amethyst> but I think there's definitely some @crawlcode here
21:51:32 <advil> haha
21:51:50 <advil> this might be the most @crawlcode thing I've encountered
21:51:55 <advil> but yeah, hard to summarize
21:52:31 <advil> I think aidanh's fix would allow 0-length argument lists that would start va_args
21:52:45 <advil> I might just quick commit a different fix, unless you want to do the vector thing now
21:53:35 <|amethyst> *I* don't want to do the vector thing now, along with all the other refactoring that needs to happen
21:53:40 <advil> haha
21:53:42 <advil> exactly
21:53:43 <|amethyst> I want *someone* to do that now though :)
21:53:57 <gammafunk> is qw available?
21:54:08 <gammafunk> maybe elliptic could a code_plan
21:54:09 * geekosaur considers a comment about tech debt
21:54:14 <|amethyst> advil: which commit was that btw?
21:54:18 <|amethyst> advil:  aidanh's fix
21:54:25 <advil> https://github.com/crawl/crawl/pull/620/commits/fdb80c3ddb85423d69267d13c5b048c359b4117a
21:54:40 <advil> that will fix the case I pasted earlier
21:55:09 <gammafunk> all this is good, it's going to distract people from any irregularities in my forthcoming --force-map argument
21:55:11 <advil> but there's no independent way to know when to stop the varargs, so it would allow a call with none at all and start varargs
21:55:21 <|amethyst> that doesn't fix the other function though, since it does need to treat the first differentl
21:55:26 <advil> yeah
21:55:30 <advil> I will fix both
21:55:56 <|amethyst> changing it to take a vector is actually not too bad to do right now
21:56:02 <|amethyst> changing both of the functions
21:56:17 <|amethyst> I'm not sure if initializer_list can be used here, but if so that's better than vector
21:56:31 <advil> although the 0-length arg thing is actually no different than if the caller just doesn't terminate the sequence in a 0
21:56:59 <advil> I wonder if nargs was added at some point because this was crashing
21:57:14 <|amethyst> probably more as a defense against it
21:57:20 <|amethyst> proactively
21:58:10 <|amethyst> hm
21:58:14 <advil> whoa til that the help menu has a faq section
21:58:29 <|amethyst> why doesn't insert_commands take its vector by reference I wonder
21:59:56 <gammafunk> the FAQ is pretty old
22:00:00 <gammafunk> we updated it some time ago
22:01:18 <advil> |amethyst: are you doing the vector thing?
22:02:22 <|amethyst> yeah
22:02:47 <advil> ok, I won't commit anything
22:03:16 <advil> aidanh: this is a very good find, btw
22:06:27 <|amethyst> I'm keeping `first` separate for the version with space_to_colon
22:06:39 <gammafunk> that code was so bad, neil couldn't let it continue to exist
22:06:41 <|amethyst> since it works differently and can't be omitted
22:06:58 <|amethyst> I'm tempted to move it ahead of the desc
22:07:06 <|amethyst> since it appears there
22:07:12 <|amethyst> I think I will, actually
22:07:47 <gammafunk> hrm, this is not really related, but I wonder if in objstat_generate_stat()
22:07:53 <gammafunk> if I really need an exception for Forest
22:08:05 <gammafunk> I don't condition on the other branches that still have enum values
22:09:39 <gammafunk> I guess it still exists in that live games might transfer with it
22:10:54 <advil> ??forest
22:10:54 <Sequell> forest[1/1]: Wood walls are unique to this branch. As the name tells, spriggans keep the verdant rune of Zot here. Don't try to get it unless you are _very_ careful, or you will die in 10 turns. Deprecated branch, no longer appears.
22:10:55 <gammafunk> right, I don't do this for blade
22:11:06 <gammafunk> which if it's valid to do for forest, it's valid to do for blade
22:11:11 <gammafunk> probably I just don't need this condition
22:12:27 <gammafunk> I need this iteration thing for somethign else, so I'll just clean it up and move it to a function
22:12:56 <advil> huh, just a little bit after the last major tag increase
22:13:13 <|amethyst> oh, is this all duplicated for hints???
22:13:30 <advil> lol
22:13:53 <advil> I haven't yet dug into the blame for this file
22:14:24 <gammafunk> aidanh really uncovered a rock with some good critters under it
22:15:26 <advil> jpeg huh
22:16:00 <advil> in 2010
22:16:17 <advil> is how long that varargs bug has been there
22:18:51 <|amethyst> well, at least C++11 didn't exist yet
22:20:08 <advil> yeah, I'm sure back then this would have been really annoying to do however you're doing it
22:21:08 <|amethyst> hm, in an old version of gcc, when compiling with no optimisation, I get a warning from exit_type_to_string
22:21:56 <|amethyst> because with -O0 it doesn't do enough flow analysis to know that those are all the possibilities
22:22:11 <advil> oh, that's annoying
22:22:15 <advil> well, I can add the default
22:22:25 <|amethyst> to avoid that, and also for safety in case someone does perversely cast an integer, I'd add  return "eggplant"; at the end
22:22:26 <advil> this is with the enum class version?
22:22:32 <|amethyst> no, I didn't change that
22:22:42 <advil> I switched it to an enum class in a later commit
22:22:48 <|amethyst> I think it's better to have that as an extra return after the switch, rather than a default
22:23:23 <|amethyst> that way you *do* still get the analysis that tells you if you missed a branch, if you're compiling with enough optimisation for it to be able to do that
22:23:31 <|amethyst> s/missed/forgot/
22:23:36 <|amethyst> hmm
22:24:34 <|amethyst> assuming a return at the end doesn't defeat that...
22:25:19 <advil> my compiler will check the switch for a default
22:26:29 <|amethyst> I mean that some compilers (with some options) will check a defaultless switch on an enum, and give you a warning if there's a missing case
22:26:42 <advil> yeah, that's what I meant
22:26:51 <advil> not sure why I wrote default
22:27:02 <|amethyst> hm
22:27:04 <|amethyst> oh
22:27:10 <|amethyst> I get an error in gcc
22:27:13 <|amethyst> from your changes
22:27:17 <|amethyst> but easy enough to fix
22:27:46 <advil> ah so does travis
22:30:05 <|amethyst> oh, wait, I wasn't even building without optimisation; I was using -Og
22:30:45 <advil> you are talking about the collision on game_exit in the exception class?
22:30:52 <|amethyst> yeah
22:31:01 <|amethyst> renaming it to exit_reason
22:31:20 <|amethyst> I guess just "reason" would probably be sufficient, but I already committed so nyah
22:31:28 <advil> ah ok
22:45:06 <advil> hmm did you commit and push?
22:45:19 <|amethyst> no, I didn't push yet
22:45:46 <advil> ok...just don't want my compile error hanging around too long :)
22:46:05 <advil> only reason travis isn't emailing me all is that one job is hung
22:46:10 <advil> *us all
22:54:16 <|amethyst> I kind of want to change this to return a const char * instead, but without messing with TilesFramework::send_exit_reason and write_message_escaped I guess there's no point
22:54:50 <|amethyst> I am going to make it static though, since it's not used outside of end.cc
23:06:59 <Cheibriados> 03|amethyst02 07* 0.21-a0-376-gab89da6: Don't have a member with the same name as its type. 10(38 minutes ago, 2 files, 4+ 4-) 13https://github.com/crawl/crawl/commit/ab89da687e0f
23:06:59 <Cheibriados> 03|amethyst02 07* 0.21-a0-377-g5e037d9: Take an argument by const reference. 10(49 minutes ago, 2 files, 2+ 2-) 13https://github.com/crawl/crawl/commit/5e037d904f15
23:06:59 <Cheibriados> 03|amethyst02 07* 0.21-a0-378-g022a3cd: Fix an _add_insert_commands varargs overrun (aidanh, advil) 10(50 minutes ago, 1 file, 76+ 105-) 13https://github.com/crawl/crawl/commit/022a3cd6c2ec
23:06:59 <Cheibriados> 03|amethyst02 07* 0.21-a0-379-gdf9fdda: Avoid a warning from gcc. 10(19 minutes ago, 1 file, 1+ 0-) 13https://github.com/crawl/crawl/commit/df9fdda5c3f3
23:06:59 <Cheibriados> 03|amethyst02 07* 0.21-a0-380-gd65dc97: Staticalise a few functions. 10(11 minutes ago, 2 files, 9+ 8-) 13https://github.com/crawl/crawl/commit/d65dc97583ad