00:13:00 <Cbrotell> Unstable branch on cbro.berotato.org updated to: 0.33-a0-644-g64fd3f8b57 (34)
00:56:36 <Gretell> Monster database of master branch on crawl.develz.org updated to: 0.33-a0-644-g64fd3f8b57
02:27:15 <ahyangyi> Oh, turns out the bolt::harmless_to_player isn't even called(), the new targeter has harmful_to_player() always returning true.
02:27:53 <ahyangyi> No, looking at wrong parts of code again.
02:48:43 <NotCheibriados> New branch created: pull/4208 (1 commit) 13https://github.com/crawl/crawl/pull/4208
02:48:44 <NotCheibriados> 03Yi Yang @ Anteros02 07https://github.com/crawl/crawl/pull/4208 * 0.33-a0-644-g7e1a7e8189: Let beam targeter to take whether it's harmless into account. 10(5 minutes ago, 2 files, 2+ 2-) 13https://github.com/crawl/crawl/commit/7e1a7e8189de
04:34:10 <Lantell> Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5249-g4a8afe7061
05:07:04 <Henzell> Unstable branch on crawl.akrasiac.org updated to: 0.33-a0-644-g64fd3f8 (34)
09:39:31 <Lantell> Wyvric (L26 DrCj) ERROR in 'mon-util.cc' at line 694: bogus mc (no monster data): invalid monster_type 1000 (1000) (Zot:3)
14:01:36 <NotCheibriados> 03Implojin02 07* 0.33-a0-645-g2dfc2d5bbc: Revert "Fix drop_disables_autopickup artefact interactions" 10(57 minutes ago, 1 file, 1+ 2-) 13https://github.com/crawl/crawl/commit/2dfc2d5bbcbd
14:01:36 <NotCheibriados> 03Implojin02 07* 0.33-a0-646-g5978acde03: Exclude artefacts from drop_disables_autopickup (DemifiendCruithne) 10(35 minutes ago, 1 file, 5+ 1-) 13https://github.com/crawl/crawl/commit/5978acde034e
15:04:19 <cgettys> I've gotten one of my sticks of RAM to behave at JEDEC speeds, though it's clearly failing as the XMP profile results in memory errors... so I'm back
15:04:43 <cgettys> Continuing the discussion from Discord the other day RE: glu... is there a setting somewhere to turn off hover tooltips? trying to work out if mipmaps actually are even being used
15:05:35 <cgettys> Ah wait, I missed a comment
15:05:45 <cgettys> // If the window size is less than the view height, the textures will    // have to be shrunk. If this isn't the case, then don't create mipmaps,    // as this appears to make things blurry on some users machines.    bool need_mips = (m_windowsz.y < 32 * VIEW_MIN_HEIGHT);
15:06:20 <cgettys> so first I have to figure out how to trigger that, then my question becomes interesting
15:09:13 <Discord|> <03i​mplojin> suspect the solution to the glu issues is either to remove the glu calls (looked like we already had an alternate codepath there) or to update them to modern opengl, whatever that entails
15:09:54 <cgettys> Right, the alt code path is "just don't generate mipmaps", which we also do if on e.g. android where it's OpenGL ES instead of full OpenGL
15:10:16 <cgettys> but even vaguely modern OpenGL ES versions also support mipmap generation using a different function,  just like modern OpenGL versions
15:11:06 <cgettys> I sort of suspect that phones would benefit more from mipmapping than most desktops these days, but graphics isn't my forte :D
15:11:27 <cgettys> But, the first step to deciding if it's worth keeping, would be to see "what does it look like with/without"
15:11:34 <Discord|> <03i​mplojin> sure
15:12:11 <cgettys> guess I can just hard-code the conditional and see if it's noticeable that way
15:12:22 <cgettys> But the annoying part is getting the screenshots to match perfectly
15:12:31 <cgettys> will figur eout how to hack off the hover text locally
15:12:47 <geekosaur> .oO { write once, debug everywhere }
15:12:48 <cgettys> other than that screenshot looks reproducible to me just by loading same game state
15:19:44 <cgettys> ah-ha
15:19:45 <cgettys> tile_tooltip_ms = 0
15:36:49 <cgettys> getting somewhere... gonna add a new maybebool option to allow forcing enabling or disabling mipmapping to save me some rebuilds and give users more control
15:46:12 <cgettys> Ok, so either I've done something dumb, or mipmapping only is relevant when window is too small or zoomed in or out
15:46:14 <cgettys> which would make sense I think :D
15:51:36 <cgettys> Yeah, ok, getting somewhere
15:51:53 <cgettys> if I zoom in to 1.5x, then I get a very subtle difference
16:31:53 <cgettys> Ok, that's fine, but something doesn't make sense here
16:32:00 <cgettys> unless there's another code path I'm missing
16:32:10 <cgettys> void OGLStateManager::load_texture
16:32:15 <cgettys> takes an xoffset and yoffset
16:32:49 <cgettys> if (xoffset >= 0 && yoffset >= 0)
16:32:59 <cgettys> This seems like it should be an or, offseting in one axis seems valid
16:33:28 <cgettys> And secondarily, if MIPMAP_CREATE is the mip_opt, we totally ignore the offsets entirely?
16:38:33 <cgettys> https://github.com/crawl/crawl/blob/5978acde034e9e1247c07d9454760a8d3a5455e1/crawl-ref/source/glwrapper-ogl.cc#L363
16:40:05 <Lantell> Unstable branch on underhound.eu updated to: 0.33-a0-646-g5978acde03 (34)
16:53:23 <cgettys> So part of it makes sense now. If offsetting in 1 axis, the other will be zero
16:53:33 <cgettys> if not offsetting at all, both will be negative -1
16:54:14 <cgettys> This seems a bit silly, when instead it could just do if xoffset == 0 && yoffset == 0, use glTexImage2D, otherwise, use glTexSubImage2D
16:54:21 <cgettys> avoiding magic numbers entirely
18:09:49 <cgettys> may be wrong on that, those functions are somewhat non-trival to understand
18:10:00 <cgettys> I understand what the code is doing now kinda, just not why
18:10:35 <cgettys> There are really two pretty much entirely distinct code paths going on here, one of htem is only used for fonts, the other is used by both
19:01:31 <NotCheibriados> 03hellmonk02 07[evokerstacking] * 0.33-a0-437-gd597c63895: savecompat 10(3 minutes ago, 2 files, 3+ 3-) 13https://github.com/crawl/crawl/commit/d597c6389545
19:01:31 <NotCheibriados> 03hellmonk02 07[evokerstacking] * 0.33-a0-438-g8679872f8b: correct comment 10(75 seconds ago, 1 file, 1+ 2-) 13https://github.com/crawl/crawl/commit/8679872f8b50
19:01:46 <Discord|> <09h​ellmonk> if someone could do some testing on this I'd appreciate it
19:06:13 <Discord|> <04d​racoomega> I can probably take a look at it later today
19:10:38 <dcss-ci> 04Build failed for 08evokerstacking @ 8679872f 06https://github.com/crawl/crawl/actions/runs/12625138085
19:15:42 <Discord|> <04d​racoomega> Currently trying to figure out a strange issue where if you're wearing no jewelry, and have armour that is entire melded, pressing R instead brings up the 'take off armour' menu, just like if you'd pressed T instead. >.>
19:16:08 <Discord|> <04d​racoomega> It acts normally if you have any jewelry or no armour at all or any unmelded armour
19:16:16 <Discord|> <04d​racoomega> So weirdly specific
19:18:41 <Discord|> <09g​ammafunk> that CI failure of the header-build-tests is weird; I get the same failure locally
19:19:36 <Discord|> <09g​ammafunk> I'm not familiar with this test but it not seeing the definition of uint64_t seems quite odd
19:20:13 <Discord|> <09g​ammafunk> and I see the same failure with gcc 13.3.0 on my system
19:22:59 <geekosaur> are you running ubuntu 24.04?
19:23:50 <geekosaur> although, I haven't been seeing build problems on my system which is 24.04
19:24:36 <Discord|> <09g​ammafunk> yes, 24.04
19:24:54 <Discord|> <09g​ammafunk> this is specifically for the header-build-tests target, mind you
19:24:58 <Discord|> <09g​ammafunk> not for the regular build
19:25:21 <geekosaur> right, CI started choking on them when github switched ubuntu-latest to 24.04
19:25:32 <geekosaur> I haven't tried the header-build tests locally
19:25:54 <geekosaur> but I wonder if maybe they would work because I upgraded from 22.04, so I have extra packages
19:26:08 <Discord|> <09g​ammafunk> it's saying we just need to include <cstdint> in an appropriate place
19:26:47 <geekosaur> hm, getting a lot of warnings about #pragma once
19:27:37 <geekosaur> yeh, just hit it
19:28:31 <Discord|> <09g​ammafunk> yeah, adding two includes for <cstdint> in two headers that gcc recommended did fix the test failure, at least
19:28:51 <Discord|> <09g​ammafunk> but yeah, still tons of warnings about those pragmas
19:29:49 <Discord|> <09g​ammafunk> this is also mildly worring:  conduct-type.h:30:5: warning: "TAG_MAJOR_VERSION" is not defined, evaluates to 0 [-Wundef]
19:44:42 <NotCheibriados> New branch created: pull/4210 (11 commits) 13https://github.com/crawl/crawl/pull/4210
19:44:53 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4210 * 0.33-a0-637-g64f2ed0c76: Use existing enum rather than less discoverable/comprehensible bool at ImageManager layer 10(5 hours ago, 3 files, 5+ 7-) 13https://github.com/crawl/crawl/commit/64f2ed0c76fb
19:44:53 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4210 * 0.33-a0-638-g3408d8194d: Add option to force mipmapping on or off 10(4 hours ago, 4 files, 26+ 4-) 13https://github.com/crawl/crawl/commit/3408d8194da7
19:44:53 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4210 * 0.33-a0-639-g4a854c7ccb: Address bad smell with load_texture arguments and understand the two distinct callers 10(2 hours ago, 7 files, 101+ 37-) 13https://github.com/crawl/crawl/commit/4a854c7ccb73
19:44:53 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4210 * 0.33-a0-640-g07c80f122f: Something is broken... hmmm 10(2 hours ago, 2 files, 18+ 4-) 13https://github.com/crawl/crawl/commit/07c80f122f7b
19:44:53 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4210 * 0.33-a0-641-g287215702f: Still broken, but in a more fun way. text renders as nonsense instead of not at all now. 10(85 minutes ago, 3 files, 9+ 7-) 13https://github.com/crawl/crawl/commit/287215702f98
19:44:53 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4210 * 0.33-a0-642-g522a3deda4: Fix mistakes, now everything works like it did before 10(78 minutes ago, 2 files, 8+ 6-) 13https://github.com/crawl/crawl/commit/522a3deda41a
19:44:53 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4210 * 0.33-a0-643-g2895b1579c: Still works 10(65 minutes ago, 2 files, 19+ 14-) 13https://github.com/crawl/crawl/commit/2895b1579cf1
19:44:53 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4210 * 0.33-a0-644-gafa92a7971: use a vector and resize instead of reallocating pointlessly all the time. 10(31 minutes ago, 2 files, 10+ 9-) 13https://github.com/crawl/crawl/commit/afa92a7971b9
19:44:53 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4210 * 0.33-a0-645-gd02760db0c: fix: actually build a block from (0, 0) to (m_max_advance.x, m_max_advance.y) and fix row-ordering 10(9 minutes ago, 1 file, 5+ 5-) 13https://github.com/crawl/crawl/commit/d02760db0cac
19:44:53 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4210 * 0.33-a0-646-gb619cd7f26: improvement: just make the whole block white, rather than only the max advance? 10(6 minutes ago, 1 file, 1+ 12-) 13https://github.com/crawl/crawl/commit/b619cd7f26c6
19:44:53 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4210 * 0.33-a0-647-ga28778a398: Remove debug statement 10(82 seconds ago, 1 file, 0+ 2-) 13https://github.com/crawl/crawl/commit/a28778a3985c
19:45:20 <cgettys> Sorry for the slight spam :D
19:45:59 <cgettys> Feedback is welcome. I haven't gotten as far as actually removing the GLU dependency yet, just got as far as actually sort of comprehending what this code does and improving it a little hopefully :D
19:46:33 <cgettys> It sounds like include ordering issues
19:47:24 <cgettys> RE: Yeah
19:48:08 <cgettys> Yeah, conduct-type.h should include tag-version.h explicitly
19:48:57 <cgettys> This may be helpful: https://clang.llvm.org/extra/clang-tidy/checks/misc/header-include-cycle.html
19:56:26 <cgettys> Ok, I can explain what's going on a bit I think
19:56:34 <cgettys> header-build-tests is compiling header files
19:56:48 <cgettys> i assume the goal is to ensure the headers are valid code?
19:57:06 <cgettys> But that causes gcc to get whiny, is probably some way to disabled it
19:57:10 <cgettys> that's the pragma once bit
19:57:19 <cgettys> So that's effectively noise
19:57:30 <cgettys> game-chapter.h:5:5: warning: "TAG_MAJOR_VERSION" is not defined, evaluates to 0 [-Wundef]
19:57:56 <cgettys> These are telling you that your include order is wrong and the main build maybe gets away with it because all headers that includes that header include the other header it needs first
19:58:22 <cgettys> and the cstdint ones are the same basic idea except compilation failures
20:14:17 <cgettys> The hacky solution to the pragma once one is probalby something like this? https://stackoverflow.com/a/21740973
20:14:31 <cgettys> Or -Wno-pragma-once-outside-header
20:16:33 <geekosaur> the -W one seems more right given that you're deliberately treating headers as C/C++ files
20:17:21 <cgettys> Except it doesn't seem to actually work on the gcc version I have on24.04 locally... :D
20:19:23 <NotCheibriados> New branch created: pull/4211 (2 commits) 13https://github.com/crawl/crawl/pull/4211
20:19:25 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4211 * 0.33-a0-647-g4aa29df5c6: fix: missing include 10(9 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/4aa29df5c61d
20:19:25 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4211 * 0.33-a0-648-gf0162a75a3: fix: pedantic / unnecessary semicolon 10(9 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/f0162a75a3f1
20:19:48 <cgettys> I'm assuming gammafunk already pushed or is about to push the cstdint include addition
20:20:00 <cgettys> If not happy to add another commit to that PR I just published
20:24:02 <geekosaur> okay, SO is telling me that -W is clang-only and gcc has no plans to implement it
20:24:21 <cgettys> I swear GCC has some equivalent magic invocation
20:24:58 <cgettys> https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html
20:25:18 <cgettys> -Wno-pragma-once-outside-header
20:25:21 <cgettys> is listed
20:25:32 <cgettys> it just doesn't work on the gcc version in 24.04 it seems
20:26:22 <cgettys> Ah, it's going to make gcc whatever is next, 15 or whatever
20:26:42 <cgettys> whereas Ubuntu 24.04 is running 13.3.0
20:26:44 <cgettys> Of course :)
20:31:54 <cgettys> Clang seems less whiny
20:32:15 <cgettys> of course, ideally CI would compile header-build-tests for both clang and GCC for better coverage :D
20:32:27 <cgettys> make header-build-tests -j15 TILES=y GXX=clang CXXFLAGS='-Wno-pragma-once-outside-header' <- works at surpressing warning on clang
20:34:37 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4211 * 0.33-a0-649-gabe262dd7c: fix: add missing cstdint includes 10(22 seconds ago, 2 files, 2+ 0-) 13https://github.com/crawl/crawl/commit/abe262dd7c1d
20:37:38 <cgettys> Added a third commit adding the missing cstdint includes too
20:37:38 <cgettys> should be enough to reduce it just down to the pragma once complaint
20:37:38 <cgettys> Unrelated - DracoOmega, any plans for a Staff of Forgecraft?
20:38:02 <cgettys> Also - another dumb question, how does the discord mirroring work for crawl-dev? I don't see a crawl-dev channel on the RogueLike Discord server
20:39:41 <Discord|> <04d​racoomega> There hasn't been a staff of summoning for a lot of years, so forgecraft not having one felt consistent with that
20:41:45 <cgettys> Ah, I hadn't noticed that. Fair, and if it'd throw off balance then not worth it. Is summoning the only other school w/o a staff?
20:41:52 <cgettys> actually I'll just check myself :D
20:43:32 <cgettys> There  isn't? STAFF_SUMMONING exists in code, but only for backward compat?
20:43:52 <Discord|> <04d​racoomega> No hexes or translocations, either - the latter of which never existed at any point, to my knowledge
20:44:08 <Discord|> <04d​racoomega> And yeah, it's been removed since 0.26
20:44:43 <Discord|> <04d​racoomega> A lot of enums are kept around for save compat reasons
20:45:05 <cgettys> Yeah, that's not an easy problem
20:45:07 <Discord|> <04d​racoomega> But even wizmode won't let you generate one
20:45:12 <cgettys> either you keep no-op entries around forever
20:45:19 <cgettys> or you need some save upgrade process
20:45:25 <cgettys> (which of course would be 1-way)
20:45:34 <cgettys> Your approach probably is more forgiving to players
20:45:57 <cgettys> (or of course you could just not have save compat, but that'd suck :D)
20:45:59 <Discord|> <03i​mplojin> the dev discord is private because nobody has time to deal with the moderation issues that opening it would bring, conversation here is bridged to the crawl-dev irc channel, invites are occasionally extended to trusted community members on a vibes basis
20:46:09 <Discord|> <04d​racoomega> We do have a save upgrade process, but a lot of item properties are basically stored as an enun index (ie: "What type of staff is this?") so the order needs to stay the same.
20:46:18 <cgettys> Makes sense. I'm too new around here for "vibes" :D
20:46:25 <cgettys> Ah, gotcha
20:46:34 <cgettys> then another approach would be to explicitly number the enum variants
20:47:05 <cgettys> But I'm sure you've thought of that
20:47:13 <cgettys> and it introduces problems w/ rollback and the like
20:47:37 <Discord|> <04d​racoomega> (Basically, if you ever look at a list of enums and see #if TAG_MAJOR_VERSION == 34 that's generally something that is removed but kept around for save compat reasons)
20:47:57 <cgettys> Ah, I had seen that and came across reference
20:47:59 <cgettys> But I got it backwards
20:48:04 <Discord|> <04d​racoomega> The degree to which the tagged things still function varies a lot
20:48:08 <cgettys> I assumed it was to gate new things
20:48:22 <Discord|> <04d​racoomega> A lot of removed species are generally relatively functional. A lot of removed items do nothing.
20:48:25 <cgettys> But that makes more sense
20:48:43 <cgettys> I'm coming at this from a "person who works on databases for a living" background :D
20:48:47 <Discord|> <04d​racoomega> There may or may not be save upgrade code specifically converting some old thing into a new thing (but this is on a case-by-case basis)
20:48:58 <cgettys> Which has a lot of parallels with what you're doing, in some ways
20:49:33 <Discord|> <04d​racoomega> Dummied out monsters generally cease to exist on upgrade, I believe, for instance
20:50:06 <Discord|> <04d​racoomega> But some people have actually deliberately dragged long-removed species forward many versions by keeping their save file around
20:50:08 <cgettys> Except the user data can have any schema they want, so we don't have that problem. On the other hand, we can't really remove features that have any significant usage
20:50:40 <cgettys> But same problem of binary format, compat for years to decades, you know
20:51:06 <Discord|> <04d​racoomega> (We don't generally make any guarantee that truly ancient save files won't end up buggy in some way over the years, but recent ones should almost always upgrade smoothly)
20:51:22 <cgettys> Makes sense
20:51:59 <Discord|> <04d​racoomega> (The amount of save games people actually want to return to after literal years is vanishingly few, after all, and at some point it's fine to say "Your dungeon is just going to be a little wierd)
20:52:08 <cgettys> Another random curiosity question - I see references to C++11 long being supported, but how modern are you guys? I see some constexpr around I think
20:52:50 <cgettys> And, if the answer is at least C++20, would PRs to move enums to being scoped enums be welcome/useful? they have better type safety, but can be slightly annoying to use for indexing (need a static_cast to size_t).
20:53:24 <Discord|> <04d​racoomega> I think we may just be on C++11, actually? (But I'm not really the one to ask)
20:54:03 <cgettys> It lets you write Staff::Wizardry instead of STAFF_WIZARDRY, without having the classic unscoped enum problem of being globally namesapced
20:54:38 <cgettys> Oh, I could have sworn constexpr was C++17, d'oh
20:54:42 <cgettys> parts of it are C++11
20:55:22 <cgettys> consteval is c++20 and you don't have any of that afaik
20:55:43 <cgettys> Ah wait I'm being dumb
20:55:48 <cgettys> STDFLAG = -std=c++11
20:55:52 <cgettys> yeah, C++11 only
20:57:26 <cgettys> off, no optional, variant, tuple, string_view
20:58:30 <cgettys> Might be willing to help modernize if there's interest, but the question is what the oldest compilers you guys want to support is, and I can't make that call obviously
21:00:31 <Discord|> <04d​racoomega> CAO is still running gcc 4.7, I think
21:01:05 <Discord|> <03i​mplojin> yeah, that
21:01:08 <cgettys> Ooof...
21:01:15 <cgettys> that's literally a decade old
21:01:33 <Discord|> <03i​mplojin> we're limited by CAO being ancient and also one of the most active webservers
21:04:37 <cgettys> I'm assuming with enough spare time one could install a newer gcc package side by side, or run a docker container, or something, but that takes work and it's not like anyone is paying to do it
21:06:26 <cgettys> But given that I don't exactly have a server lying around that I can offer to host, who am I to judge
21:08:40 <cgettys> And a docker container would potentialy add some perf overhead that maybe can't be afforded
21:10:57 <cgettys> Another solution would be to not compile on the server itself, assuming dcss is self contained enough
21:11:32 <cgettys> But again, I'm armchair quarterbacking, woul dhave to make every commit to master produce github artifacts and overhaul the webtile code to download said artifacts instead of compiling themselves
21:11:52 <cgettys> I'm assuming that that's what's happening, or what version it's running wouldn't necessarily matter so much
21:12:27 <cgettys> Frankly, I'm amazed that you guys have this compiling for everything from Ubuntu 24.04 back to whatever had gcc 4.7
21:12:46 <cgettys> 12.04?
21:23:45 <ahyangyi> Since we can build AppImage these days, can we ask CAO to run that version?
21:29:20 <cgettys> Question is how far back does AppImage still support, I'd guess
21:30:06 <cgettys> And also teh code to handle running multiple versions of the AppImages side by side, et cetera
21:31:21 <cgettys> I'm not sure how the multiplexing of versions is done, dgame-launch tool?
21:41:44 <NotCheibriados> 03regret-index02 07* 0.33-a0-647-gb60e247eb4: Don't generate substantially more amulets than rings (various) 10(59 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/b60e247eb4da
23:35:17 <Gretell> Unstable branch on crawl.develz.org updated to: 0.33-a0-647-gb60e247eb4 (34)
23:57:56 <Gretell> Windows builds of master branch on crawl.develz.org updated to: 0.33-a0-647-gb60e247eb4