00:27:08 <Eksell> Unstable branch on CRAWL.XTAHUA.COM updated to: 0.24-a0-429-g4608c8fa30 (34)
00:28:33 <Eksell> Stable (0.23) branch on CRAWL.XTAHUA.COM updated to: 0.23.1-79-g0e89c35649
00:56:09 <alexjurkiewicz> there is an array in describe.cc `const char* const size_adj[] =`
00:56:29 <alexjurkiewicz> how do I write a run-time assert that my index is not larger than the array?
00:56:40 <alexjurkiewicz> I can check it doesn't underflow with ASSERT(static_cast<int>(size) > 0);
01:13:08 -!- tehgeek is now known as MoonMoon
01:13:24 -!- MoonMoon is now known as SunMoon
01:14:42 <Pinkbeast> alexjurkiewicz: Isn't this just a job for ARRAYSZ?
01:14:51 -!- SunMoon is now known as Psymania
01:45:32 -!- amalloy_ is now known as amalloy
02:04:27 <|amethyst> ASSERT_RANGE(size, 0, ARRAYSZ(size_adj));
02:05:37 <|amethyst> (assuming 0 is a valid size)
02:05:57 <|amethyst> (and that the size of the array is not a valid index)
02:06:42 <|amethyst> s/valid size/valid index/
03:17:14 <Rotatell> Unstable branch on crawl.beRotato.org updated to: 0.24-a0-429-g4608c8f (34)
03:28:39 <alexjurkiewicz> nice
03:34:26 <Cheibriados> New branch created: pull/1084 (1 commit) 13https://github.com/crawl/crawl/pull/1084
03:34:26 <Cheibriados> 03alexjurkiewicz02 07https://github.com/crawl/crawl/pull/1084 * 0.24-a0-430-g6632800: Guard against out of bounds array access 10(5 minutes ago, 1 file, 1+ 0-) 13https://github.com/crawl/crawl/commit/6632800931ab
03:51:14 <Kelbell> Fork (stoatsoup) on crawl.kelbi.org updated to: 0.21.2-987-g5d2755fedf
05:07:59 -!- amalloy is now known as amalloy_
09:38:59 <advil> alexjurkiewicz: was there a specific case that came up for that assert? FWIW that constraint should already be partially enforced by the use of the enum type + the compile time check on the array, it's really just size < NUM_SIZE_LEVELS that's missing
09:45:43 <advil> well, I suppose it's probably a bloatcrawl thing
09:46:03 <Cheibriados> 03advil02 07* 0.24-a0-430-g69c6a51: Fix signed/unsigned comparison (gammafunk) 10(3 minutes ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/69c6a515889a
10:02:30 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.24-a0-430-g69c6a51588 (34)
11:23:58 <Lantell> Stable (0.23) branch on underhound.eu updated to: 0.23.1-79-g0e89c35649
11:58:04 <Kelbell> Fork (stoatsoup) on crawl.kelbi.org updated to: 0.21.2-988-gd4dae30470
12:51:30 <advil> %git 38f3c1a516ae676
12:51:30 <Cheibriados> 07advil02 * 0.24-a0-290-g38f3c1a: Make ds, wn, and dswn starts more deterministic by seed 10(4 weeks ago, 4 files, 17+ 6-) 13https://github.com/crawl/crawl/commit/38f3c1a516ae
13:14:01 <Henzell> Unstable branch on crawl.akrasiac.org updated to: 0.24-a0-430-g69c6a51 (34)
16:00:51 -!- amalloy_ is now known as amalloy
16:36:48 <Cheibriados> 03advil02 07* 0.24-a0-431-g80c1907: Use unordered_set for tags, instead of set 10(5 hours ago, 7 files, 51+ 39-) 13https://github.com/crawl/crawl/commit/80c1907ebbf6
16:36:48 <Cheibriados> 03advil02 07* 0.24-a0-432-g90a0943: Cache the top three vault tags as bools 10(3 hours ago, 6 files, 90+ 11-) 13https://github.com/crawl/crawl/commit/90a09432193d
16:36:48 <Cheibriados> 03advil02 07* 0.24-a0-433-gbe6fa4f: Faster starts_with implementation 10(2 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/be6fa4f6cc9d
16:37:40 <advil> that's pretty much the bottom of the barrel for really obvious/easy levelgen speedups
16:39:55 <advil> I think all told I've gotten about 20-30% off the total time for a non-optimized build, which is not too bad
16:44:24 <advil> well, there's still 4s total time spent in `has_tag`, but I'm not sure how to optimize that further without changing something less low level, the problem is just that it's called millions of times
17:08:07 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.24-a0-433-gbe6fa4f6cc (34)
17:45:08 <advil> MarvinPA: would you read that last post from a github random in the "they" PR as transphobic? I can't tell if I'm overreacting, so I'm not sure what the best thing to do with it is
17:46:11 <advil> (in any case, as a professional linguist I do think changing to singular "they" is absolutely the right thing to do)
18:01:39 <MarvinPA> oh ugh, it does read that way yeah
18:03:41 <MarvinPA> i guess i can just comment that it definitely won't be optional, but happy to block/ban people if needed too
18:19:56 <advil> I wasn't sure about how to respond
18:20:53 <advil> I guess blocking maybe depends on what they do after one of us comments, github block is very extreme
18:22:30 <Lantell> Unstable branch on underhound.eu updated to: 0.24-a0-433-gbe6fa4f6cc (34)
18:41:24 <alexjurkiewicz> advil: yeah it happened while fork developing but it seemed like a good thing to catch for normal crawl since the results seemed pretty bad
18:42:02 <alexjurkiewicz> about the github comment I think it is going to be, but I figured if you get clarification then you ("the devs") can make a clear comment that it's Not OK
18:42:29 <advil> yeah, I saw that you responded
18:51:42 <advil> oh great, lua string comparison is locale sensitive
20:43:24 -!- amalloy is now known as amalloy_
21:03:00 -!- amalloy_ is now known as amalloy
21:13:48 <Pinkbeast> floraline tells me anyone with voice here can run a build on CKO; please could someone rebuild Stoat Soup? I think I have fixed the horrible starvation bug.
21:18:14 <gammafunk> running now
21:18:41 <Kelbell> Fork (stoatsoup) on crawl.kelbi.org updated to: 0.21.2-989-g6bb62fb1ba
21:18:45 <Pinkbeast> Thanks
21:18:46 <gammafunk> ...that was unusually fast...
21:19:28 <Pinkbeast> I guess it uses ccache; I only changed two files, no headers. Now, did it work?
21:20:20 <Pinkbeast> Yes! I'm not starving to death. :-/
21:20:28 <Pinkbeast> Thanks again.
21:51:39 <advil> man, this lua string issue is really annoying. A seed divergence issue that just came up was happening because capitalized unique names were being sorted differently on different systems, apparently because the default locale used by lua is different (and capital letters are collated differently)
21:52:56 <advil> the most obvious solution is to force "C" locale I guess?
21:54:49 <advil> what locale even collates capital with lowercase letters
21:55:27 <|amethyst> en_US does on my system
22:04:30 <advil> I think also on mac for some reason it is not inheriting the locale from the main process
22:06:41 <gammafunk> advil: was that the divergence I saw?
22:08:55 <advil> it's not what I was just chasing down but it is conceivable if you have yet another locale, I don't know how far this extends
22:10:26 <advil> it's coming up on current trunk seed 1 for hanged_man_spin_cycle which at the absdepth corresponding to Depths:5 (where it places in that seed) builds a monster weight table that includes both uniques (capitalized) and non-uniques at the same weight, so sorts them
22:10:56 <advil> *hangedman
22:10:56 <advil> which I can easily fix by changing the capitalization but this is kind of a worrying issue
22:11:14 <advil> there are all sorts of locales that will change sort orders in minor or major ways, not just capitalization issues
22:12:56 <gammafunk> mine is just en_US
22:13:32 <|amethyst> also, I don't know if we ever convert strings to numbers in Lua, but apparently that's locale-dependent too
22:13:45 <advil> haha of course it is
22:14:15 <advil> we do do that sometimes, not sure how much
22:14:43 <|amethyst> probably most relevant for numbers with decimal points
22:14:50 <|amethyst> but not entirely
22:14:57 <|amethyst> s/entirely/only/
22:19:40 <advil> ugh lua doesn't use locale in any particularly compartmentalized way, so I'm not sure their os.setlocale is a viable option
22:20:27 <|amethyst> Yeah, from the name, I would figure os.setlocale would just call the C setlocale function
22:20:34 <advil> yes
22:21:07 <gammafunk> !source describe_spell
22:21:07 <Sequell> 1/1. https://github.com/crawl/crawl/blob/master/crawl-ref/source/describe.cc#L3141
22:21:41 <|amethyst> hm, and I don't know if there's a particularly portable choice to just set LC_COLLATE
22:21:55 <|amethyst> I guess "C"
22:23:11 <gammafunk> in that function, I don't quite understand how the more variable can be accessed from 3229
22:24:06 <gammafunk> oh
22:24:08 <|amethyst> gammafunk:  that's the more() function, not the variable
22:24:09 <gammafunk> it's not
22:24:11 <gammafunk> dammit
22:24:21 <|amethyst> more() is maybe not the best name for a global function
22:24:30 <gammafunk> yeah I guess I got foiled by a bad variable name
22:24:31 <|amethyst> concise though it may be
22:24:39 <gammafunk> or right, a bad metho name
22:25:13 <advil> oh heh
22:25:27 <advil> yeah the `auto more` is local to its own scope
22:25:47 <|amethyst> even message_window::more isn't quite so bad
22:27:04 <|amethyst> compared to the free function more() on message.cc:1774
22:27:22 <|amethyst> IMO rename it to prompt_more or something
22:27:35 <advil> or put it in a namespace
22:30:39 <gammafunk> allow me to change the subject from me doing renaming work to one of getting clarification on how this menu code works
22:31:07 <gammafunk> currently learn_spell()  creates MemoriseMenu via _choose_mem_spell()
22:31:22 <gammafunk> yet in viewing the spell description from said menu
22:31:41 <gammafunk> you can directly hit M to (M)emorise the given spell
22:32:01 <gammafunk> _choose_mem_spell() gets the spell to memorise from this menu
22:32:27 <gammafunk> so in the case where the player uses that spell description hotkey
22:32:49 <gammafunk> and where this is calling learn_spell(spell) (another function signature)
22:33:28 <gammafunk> how does it doing that not lead to the parent caller of learn_spell() (no argument) calling learn_spell(spell) a second time?
22:33:58 <gammafunk> oh, maybe it just does call it a second time?
22:34:32 <gammafunk> no, I don't think it does call it a second time, since apparently that'd print a message
22:34:55 <gammafunk> ahh
22:35:02 <gammafunk> I think it does call it a second time, perhaps
22:35:04 <|amethyst> already_learning_spell
22:35:06 <gammafunk> damn
22:35:11 <gammafunk> that seems kind of a bad design
22:35:47 <gammafunk> for my sif casting thing I'll have a similar problem
22:36:32 <gammafunk> except I'll have no queued delay to check or anything
22:37:49 <gammafunk> since I have to pass a new `sif` variable to various spell description functions in order to get descriptions to show correct info already, and this is kind of invasive
22:38:17 <gammafunk> I'm tempted to set some kind of global state like how we do for spell memorisation searches
22:38:40 <gammafunk> but I'm worried that this state wouldn't get properly unset given all ways to exit this menu, somehow
22:45:06 <gammafunk> |amethyst: not sure if you saw my musing there, but I wonder if I should set up some global state like you.trying_to_sif_cast (with a better name) since I apparently need that to properly handle this menu state and because spell description functions currently have to be modified to have a "trying to sif cast" variable to show correct info
22:45:34 <gammafunk> just not sure if I can put resetting that state in e.g. my menu classe's deconstructor and have it actually work
22:46:32 <gammafunk> well, it's also the spell casting code that needs this new "sif cast" variable, maybe those functions could use the argument and the menu/desc code could use the state
22:52:15 <gammafunk> heh
22:52:25 <gammafunk> what do I see as a browse player.h but
22:52:36 <gammafunk> // Skill menu states
22:52:36 <gammafunk> skill_menu_state skill_menu_do;
22:52:36 <gammafunk> skill_menu_state skill_menu_view;
22:56:21 <gammafunk> oh, I suppose I could reliably reset the state in my function that creates the menu
22:56:36 <gammafunk> so this shouldn't be too hard after all