00:28:23 <travis-ci> The build passed. (master - 3ca0c31 #11274 : kitchen-ace): https://travis-ci.org/crawl/crawl/builds/512820575
00:29:53 <alexjurk1> !seen ontoclasm
00:29:54 <Sequell> I last saw ontoclasm at Sun May 14 21:54:24 2017 UTC (about 1y 45w 4d 6h 35m 29s ago) parting ##crawl-dev, saying 'chanpart'.
00:30:02 <alexjurk1> woah. really two years?
01:01:06 <Eksell> Unstable branch on CRAWL.XTAHUA.COM updated to: 0.24-a0-217-g1152973 (34)
01:02:22 <Eksell> Stable (0.23) branch on CRAWL.XTAHUA.COM updated to: 0.23.1-41-g7e51ed3
01:24:32 <travis-ci> The build passed. (master - 1152973 #11275 : Umer Shaikh): https://travis-ci.org/crawl/crawl/builds/512829696
02:07:50 <Gretell> Stable branch on crawl.develz.org updated to: 0.23.1-41-g7e51ed3 (34)
02:37:54 <Gretell> Unstable branch on crawl.develz.org updated to: 0.24-a0-217-g1152973 (34)
03:08:42 <Gretell> Windows builds of master branch on crawl.develz.org updated to: 0.24-a0-217-g1152973
03:30:56 <Kelbell> Stable (0.23) branch on crawl.kelbi.org updated to: 0.23.1-41-g7e51ed32d8
03:42:05 <Kelbell> Fork (gooncrawl) on crawl.kelbi.org updated to: 0.22-a0-816-g14fc994bcc
03:59:48 <Gretell> Monster database of master branch on crawl.develz.org updated to: 0.24-a0-217-g1152973
04:20:13 <Cheibriados> Some items not auto-picked up 13https://crawl.develz.org/mantis/view.php?id=11939 by iliketurtles
08:04:30 <Cheibriados> New branch created: pull/1022 (1 commit) 13https://github.com/crawl/crawl/pull/1022
08:04:30 <Cheibriados> 03Nikolai Lavsky02 07https://github.com/crawl/crawl/pull/1022 * 0.24-a0-218-g36cce10: Don't try to generate any removed branches in mapstat 10(14 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/36cce107a6cd
11:13:34 <fourba> n/exit
11:13:37 <fourba> augh
12:22:58 <Lantell> Stable (0.23) branch on underhound.eu updated to: 0.23.1-41-g7e51ed32d8
13:02:52 <Henzell> Stable (0.23) branch on crawl.akrasiac.org updated to: 0.23.1-41-g7e51ed3
13:08:37 <Henzell> Unstable branch on crawl.akrasiac.org updated to: 0.24-a0-217-g1152973 (34)
13:57:04 <advil> !tell gammafunk did you still want to remove that abyss transporter vault?  Also, perhaps related (though also possibly just me not understanding the code), while working on c779bbd I noticed I couldn't figure out how transporters do get placed in abyss via normal abyss builder stuff
13:57:04 <Sequell> advil: OK, I'll let gammafunk know.
17:16:53 <gammafunk> advil: re: abyss placing transporters, doesn't it happen during the same phase as for normal levels
17:17:25 <advil> yeah, that much I got, I just never figured out where it did happen
17:19:05 <gammafunk> !source dungeon.cc:2450
17:19:05 <Sequell> https://github.com/crawl/crawl/blob/master/crawl-ref/source/dungeon.cc#L2450
17:19:19 <gammafunk> that's where the markers are turned into the features
17:19:23 <advil> right, that's for normal levels though
17:19:33 <advil> that could path didn't seem to get called for abyss? unless I missed something
17:19:55 <advil> *code
17:20:06 <advil> I did find code to handle when transporters get shifted
17:20:26 <gammafunk> yeah, that was fun to write
17:20:32 <gammafunk> let me see re: abyss layout
17:22:26 <gammafunk> ...
17:24:53 <gammafunk> that's not the source of this transporter problem somehow, is it?
17:25:19 <advil> heh it could be!
17:25:28 <gammafunk> but how is it placeing sometimes?!
17:25:40 <gammafunk> I'm about to flip some crawlcode
17:25:40 <advil> well wizmode code *does* call that code path
17:25:52 <gammafunk> god
17:26:02 <advil> so if you've seen it in the wild then this hypothesis is wrong
17:26:07 <advil> I haven't myself
17:26:12 <gammafunk> well
17:26:19 <gammafunk> people can generate inside the vault
17:26:19 <advil> but I don't go in abyss much
17:26:34 <gammafunk> I suppose it's possible that it actually never generates with the transporter
17:26:52 <gammafunk> and the only way people have done it is via wall breaking (which is common in abyss) and being generated inside it
17:26:59 <gammafunk> but if that's what this bug is, I'm really going to hate myself
17:27:03 <advil> lol
17:27:15 <advil> well who was to know, I mean I stared at this bug for a lot of hours
17:27:23 <advil> testing all the time with wizmode placement only of course
17:27:29 <advil> I guess we could test by changing the weights rather than wizmode placement
17:27:37 <gammafunk> yeah, how do I test this
17:27:40 <gammafunk> so I could still be in wizmode
17:27:43 <gammafunk> but just upweight it
17:27:45 <advil> yeah
17:27:47 <gammafunk> and go to abyss:3
17:28:08 <advil> maybe tweak some code in case there are other conditions on subvault placement
17:28:36 <gammafunk> subvault placement?
17:28:49 <gammafunk> what does that refer to specifically? this vault doesn't use subvaults
17:28:55 <advil> sorry, just vault placement
17:29:52 <advil> btw did you see the bug I fixed the other day? it was kind of good: a fallback abyss portal placed on top of a transporter marker (since fallback vaults can overwrite floor features in already placed vaults), and then the transporter overwrote the abyss portal
17:30:09 <gammafunk> I did see without looking into the logistics of how that could happen
17:30:35 <gammafunk> I guess that happened after the transporter placement in the normal builder?
17:30:35 <advil> yeah, the transporter vault in that case placed first
17:30:58 <advil> I was wondering if something similar could be happening in the abyss, was what led me to look for where the transporter placement happened there
17:35:55 <gammafunk> interesting, I can't seem to get my vault to place in abyss
17:36:19 <gammafunk> set weight high and others got selected, removed all other rune vaults, now I get a solo rune just lying on the map every time
17:37:54 <advil> !flip abyss
17:37:54 <Sequell> (╯°□°)╯︵ ssʎqɐ
17:49:17 <advil> yeah, I'm having similar troubles
17:49:28 <advil> trying to get x_factor to place
17:49:44 <gammafunk> looking back at the builder code
17:49:48 <advil> I got it to happen once with ctrl-r, and there were transporters, but that code path might be different
17:50:08 <gammafunk> yeah, that's wizmode, and wizmode does specifically do the transporter placement call
17:50:10 <advil> cannot get it to happen with just level changing, even though I've forced that to be the only possible exit vault
17:50:23 <gammafunk> it does seem to me that the abyss builder goes through the same codepath though
17:50:49 <gammafunk> _builder_by_type() -> generate_abyss() -> ..oh
17:50:51 <gammafunk> never mind
17:51:03 <gammafunk> I'm apparently misreading that return over and over again
17:51:28 <gammafunk> ok so that seems pretty clear that the transporter placement will never happen, but I'm not seeing any sort of veto messages either
17:51:55 <gammafunk> there's this comment in generate_abyss()
17:52:01 <gammafunk> // Generate the initial abyss without vaults. Vaults are horrifying.
17:52:01 <gammafunk> _abyss_generate_new_area();
17:52:33 <gammafunk> but _abyss_generate_new_area() calls _generate_area()
17:52:42 <gammafunk> which does in fact call _abyss_place_vaults()
17:56:05 <advil> man I just can't get this vault placement code to do anything
17:58:34 <advil> huh
18:02:36 <advil> oh, I think I'm being messed up by these chance tags for the dummy exit
18:02:39 <advil> does chance have a default?
18:04:02 <advil> or maybe I'm just not understanding something here
18:07:15 <advil> something about the chance values makes exit vault placement extremely rare, even though I have it set so that exits place a lot
18:18:52 <gammafunk> interesting, so if I get a 30 hd boggart to banish me
18:19:05 <gammafunk> the level on abyss:5 generates with the rune vault
18:19:12 <gammafunk> on a new game
18:19:20 <gammafunk> the transporters place and work
18:19:50 <gammafunk> ah, and yes, now I see how I had confused myself before
18:20:06 <gammafunk> I unconfused myself then confused myself and now I'm unconfused again
18:20:38 <amalloy> the abyss does that to people
18:20:42 <gammafunk> _build_dungeon_level() -> _builder_by_type() -> generate_abyss() followed by return true
18:20:57 <gammafunk> ad that point in _build_dungeon_level() place_vaults is true
18:21:02 <gammafunk> so normal transporter placement happens
18:21:29 <gammafunk> hence baring this weird issue where it doesn't seem to subsequently want to place the rune vault
18:21:36 <gammafunk> for future level generation, that is
18:21:48 <gammafunk> it does seem that the transporters work and the vault places normally for non-wizmode
18:22:26 <advil> ok, yeah, I finally did figure out how to get one of these to place
18:22:28 <gammafunk> the placement of the rune vault happens in generate_abyss() via a call to _generate_area()
18:23:57 <advil> I still have no idea what is up with these chance values on the dummy vault though, I effectively couldn't get exit vaults to place until I removed those
18:24:03 <advil> even though the exit places fine, just vaultless
18:24:26 <advil> also part of my trouble is that x_factor doesn't have allow_dup and so couldn't re-place
18:24:37 <advil> that might apply to testing the rune vault
18:24:53 <gammafunk> ah
18:24:55 <gammafunk> hrm
18:25:13 <gammafunk> so is this just somehow a thing for abyssal rune vaults? how do they allow multiple placement
18:25:56 <advil> can you see the same rune vault more than once?
18:26:13 <gammafunk> let me try
18:27:42 <advil> well, sorry for the red herring
18:29:17 <gammafunk> yeah, I can
18:29:17 <gammafunk> tso outpost is placing multiple times
18:29:30 <gammafunk> had a monster banish me, got that vault, then reading teleport scroll and resting to generate a new area
18:29:36 <gammafunk> did that multiple times and got the output vault
18:29:41 <gammafunk> *outpost
18:30:13 <gammafunk> I guess that's another mystery though, why my vault doesn't seem to want to place more than once
18:30:41 <gammafunk> it's not tagged differently in a way that would matter for this
18:31:36 <advil> I've gotten it to reliably place over and over again by adding allow_dup (and I've messed with the tags)
18:33:22 <gammafunk> weird
18:34:38 <gammafunk> advil: do you have a timeframe for making the point release, e.g. you're trying to get it done today?
18:34:38 <advil> today-ish
18:34:47 <advil> I probably won't be building release versions until tomorrow
18:34:48 <gammafunk> I guess I can just quickly rework the vault to use normal doors then
18:35:09 <gammafunk> later perhaps I can have more time to play with this transporters-in-abyss issue again
18:35:19 <advil> I might spend a few more minutes looking at this, now that I've figured out what it takes to generate this vault on demand
18:35:26 <gammafunk> yeah, feel free
18:35:55 <gammafunk> that comment in the abyss gen suggesting that vaults shouldn't place in initial areas yet it proceeding to place vaults
18:36:28 <gammafunk> hrm
18:36:29 <advil> how does CHANCE work when it isn't set explicitly?
18:37:12 <gammafunk> looks to me like it might be ignored
18:37:26 <gammafunk> dng_chance() does nothing if the arg isn't a number
18:37:43 <advil> and then when the map selector does a chance roll?
18:38:05 <gammafunk> well I see there are two chance functions
18:38:13 <gammafunk> dgn_depth_chance() as well
18:38:24 <advil> now that I think about it, have I ever even seen an abyss exit vault?
18:39:05 <gammafunk> but to answer that question, I think if no chance gets set, the vault selector for chance vaults would presumably find no matches?
18:43:53 <gammafunk> so it looks to me like abyss might do two rounds of vault placement
18:44:22 <gammafunk> oh I see
18:44:42 <gammafunk> it only does that for the minivaults/chance vaults/extra vaults
18:45:59 <gammafunk> ug
18:47:05 <gammafunk> in _abyss_place_vaults()
18:47:13 <gammafunk> while (vaults_placed < maxvaults)
18:47:13 <gammafunk> {
18:47:13 <gammafunk> const map_def *map = random_map_in_depth(level_id::current(), extra);
18:47:26 <gammafunk> yet the function's signature is
18:47:27 <gammafunk> const map_def *random_map_in_depth(const level_id &place, bool want_minivault,
18:47:30 <gammafunk> maybe_bool extra)
18:47:47 <gammafunk> this is not filling me with vast amounts of confidence
18:48:28 <gammafunk> after one round of that, extra is true if the map wasn't tagged as extra
18:48:46 <gammafunk> but that argument is actually controlling whether a minivault should be chosen
19:08:01 <advil> the reason I was wondering about chance is not chance vaults per se, but because the placement call for all tagged vaults in abyss incl abyss_exit vaults has check_chance=true, presumably because of the dummy maps, and I couldn't get a single abyss exit vault to place until I set that param to false
19:09:00 <advil> granted the CHANCE values on the dummy exit maps are like 95%, so maybe they're supposed to be extremely rare, but changing those to very low percentages didn't help (commenting them out did)
19:18:11 <gammafunk> advil: looks like map_chance::valid() would return false since map_chance::map_chance() initializes map_chance::chance to -1
19:18:41 <advil> ah
19:18:47 <gammafunk> which would cause any vault not to get added to the vector of chance vaults
19:19:13 <advil> ok, so seeing what that vector looks like would be a good test
19:19:42 <gammafunk> yeah
19:19:50 <gammafunk> in _random_map_in_list()
19:23:01 <Lantell> Unstable branch on underhound.eu updated to: 0.24-a0-217-g1152973720 (34)
19:26:42 <Floodkiller> what's the make command for a WebTiles build? want to make sure I didn't miss something
19:26:59 <advil> just for yourself?
19:27:15 <Floodkiller> yeah, just to compile test before I push and break auto-updates
19:27:59 <advil> `make WEBTILES=y` for local use, probably `make WEBTILES=y USE_DGAMELAUNCH` if you want to mimic a server setup
19:28:19 <Floodkiller> alright, thanks
19:28:29 <advil> er
19:28:38 <advil> `make WEBTILES=y USE_DGAMELAUNCH=y`
19:28:42 <advil> sorry
19:29:41 <Floodkiller> oh, it seems to be missing server specific files, guess that won't work
19:30:00 <advil> hm neither of those should need anything special
19:30:36 <Floodkiller> it's probably because it can't build WebTiles on Windows
19:30:40 <Floodkiller> from the look of it
19:30:41 <advil> oh heh
19:30:41 <advil> yeah
19:30:50 <Floodkiller> oh well, I guess I'll just push and pray
19:30:59 <advil> can you run docker on windows these days?
19:31:05 <floraline> that's the spirit!
19:31:28 <Floodkiller> probably, but I'm not going through that effort just to add a real_mp_max field to the WebTiles GUI
19:31:47 <Floodkiller> I think I've changed everything that's needed for it
19:31:58 <Floodkiller> and if not, there's a weekend to fix/revert it
19:32:13 <advil> what could go wrong
19:32:36 <Floodkiller> game crash on launch from missing value/field
19:32:38 <floraline> literally nothing, probably
19:32:48 <Floodkiller> is probably worst case
19:32:57 <Floodkiller> or bad compile
19:33:20 <advil> do you have travis set up on your repo?
19:33:28 <Floodkiller> yes, but I haven't fixed/updated it in forever so it's kind of worthless
19:33:46 <advil> if you use something similar to the mainline travis settings and push to a branch, then it will at least build with both of those build configs I mentioned above
19:34:18 <Floodkiller> as in, I've changed enough stuff from mainline at this point that ttravis fails on all builds
19:34:35 <advil> haha ok
19:37:26 <advil> well, you could still test the build, it looks like your failures are all in the tests (e.g. arena crashes, something is wrong with some egos)
19:38:19 <advil> something wrong with miscast effects, maybe torment specifically
19:40:36 <Floodkiller> I'll take a look at travis then after it finishes the latest commit and see if webtiles fails before the arena tests then
19:50:01 <advil> aha
19:50:22 <advil> gammafunk: that code path works but is only used when initially entering an abyss level
19:50:39 <advil> so teleporting never does call the transporter placement
19:51:02 <gammafunk> yeah, that would make sense
19:51:17 <gammafunk> I guess I didn't logically follow up my understanding of how it is being called right the first time
19:52:08 <gammafunk> the vault placement code for abyss doesn't have the transporter placement call
19:52:21 <gammafunk> but I still don't understand why I can't get the rune vault to place subsequent times
19:52:43 <gammafunk> I guess one question is where that transporter placement call should go
19:52:56 <advil> yeah, for some reason it can't go too late in that sequence
19:53:06 <gammafunk> _abyss_postvault_fixup()?
19:53:07 <advil> maybe in _abyss_fixup_vault?
19:53:24 <gammafunk> maybe, I see postvault actually activating the markers though
19:53:27 <gammafunk> maybe that needs to happen first
19:53:28 <advil> the postvault thing was too late for some reason that I haven't discerned (well, I tried right after that)
19:53:30 <advil> ah
19:53:49 <gammafunk> not sure if it does need to happen for these transporter markers though
19:53:56 <advil> yeah maybe before that call
19:53:58 <advil> let me test
19:54:49 <advil> yes, that worked
19:55:19 <advil> btw, were all the reports of the form, both transporter dest and landing are missing?
19:55:28 <advil> or were there some cases where one or the other was there?
19:55:34 <advil> I have a vague memory that there were
19:55:43 <gammafunk> hrm, good question; I want to say that both were but I'm not really sure
19:55:55 <gammafunk> transporters definitely missing, but not 100% sure on landing sites
19:57:37 <advil> one thing I was wondering was, what would happen with shifting transporters if the iterator got to the landing site first? I *think* it might wipe it
19:57:43 <advil> haven't been able to test that though
20:01:53 <advil> maybe not
20:05:26 <advil> ok, nope, that was just my overactive imagination, that case works
20:09:39 -!- amalloy is now known as amalloy_
20:10:20 <advil> hm, I wonder is there some way that transporter placement could be subsumed into marker activation?
20:11:23 <Cheibriados> 03advil02 07* 0.24-a0-218-ga595e75: Place transporters consistently in abyss 10(4 minutes ago, 1 file, 1+ 0-) 13https://github.com/crawl/crawl/commit/a595e75330f5
20:12:01 <Cheibriados> 03advil02 07[stone_soup-0.23] * 0.23.1-42-g8abd8fc: Place transporters consistently in abyss 10(5 minutes ago, 1 file, 1+ 0-) 13https://github.com/crawl/crawl/commit/8abd8fc9a213
20:17:26 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.24-a0-218-ga595e75330 (34)
20:21:32 <gammafunk> interesting, so activate_all() isn't necessary for transporters?
20:21:55 <gammafunk> seems that the concept of marker activation isn't what I think it is
20:34:10 <gammafunk> ah I see
20:34:19 <gammafunk> marker activation normally happens after level is fully generated
20:35:59 -!- amalloy_ is now known as amalloy