00:28:43 <09gammafunk> lobf_: I did 00:28:43 oh, rip 00:54:16 Monster database of master branch on crawl.develz.org updated to: 0.29-a0-2-g12fcac58f8 00:56:39 Unstable branch on cbro.berotato.org updated to: 0.29-a0-2-g12fcac58f8 (34) 01:37:02 Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-4556-g06e78ff023 01:54:23 Fork (bcadrencrawl) on crawl.kelbi.org updated to: 0.03-738-gc9788ef0a2 02:17:35 -!- allbery_b is now known as geekosaur 04:22:45 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-4556-g06e78ff023 05:21:39 -!- stella is now known as BoostedStella 07:06:15 the other thing about save editing is that the file is compressed, that's part of the point of the cli chunk extraction tools, that it gives you an uncompressed version 07:07:17 so just trying to go at the full thing with a hex editor is not very productive 07:08:22 I wonder what's going on on freenode these days 07:09:07 <09gammafunk> oh, good point re: the compression, I'd totally forgotten about that 07:09:49 <09gammafunk> last time I had to fix a save, I did have to modify it using a hex editor, since it needed a god change, which was in the player section 07:16:34 yeah, I've done a bunch of hex editing of chunks, it's doable as long as the info is in an identifiable location 07:16:45 which often takes reading the code to find out 07:17:29 though sometimes you can just search for the relevant value in the chunk and try changing some of the instances until you get it 07:18:56 https://hexfiend.com/ is good on mac 07:19:56 <06advil> @ZureaL sorry to report that there still seem to be gateway issues on cbr2 07:20:10 <12ZureaL> Hrmmm.... 07:21:07 <06advil> actually https://cbro.berotato.org:8443/ loads fine, but https://cbro.berotato.org/ doesn't, and someone PM'd me that morgues are giving a bad gateway 07:22:05 <12ZureaL> that's weird, the whole system was restored properly, so it really should've just came back up the same, but I'll take a look when I can 07:25:48 hm. I just opened oneyeh, I just got 504 gateway timeout 07:26:14 (sorry for confusion at the start of that, I checked cbro instead of cbr2 at first) 07:43:19 03Alex Jurkiewicz02 {advil} 07* 0.29-a0-3-g53e8333: Use append for lists, not add 10(12 hours ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/53e83337bcd3 07:49:32 Unstable branch on crawl.kelbi.org updated to: 0.29-a0-3-g53e83337bc (34) 07:50:56 ooo thanks for the hexfiend tip 07:57:31 <12ZureaL> @advil fixed. It ended up being an internal DNS screw up on my part. 07:58:56 <12ZureaL> I actually messed it up a few days ago, but since the ngnix server hadn't been restarted, the dns query remained cached so it wasn't until I rebooted did that problem become apparent. All good now though 08:27:05 <06advil> thanks for fixing it! 08:27:09 <06advil> sounds fun and exciting 08:41:44 <06advil> I'm pretty sure https://github.com/crawl/crawl/issues/2376 ("The function start_delay does not actually start the delay. It just pushes it into the queue, and it gets picked up and started on the next turn.") is intentional, right? 08:54:37 <12ebering> yes 08:55:04 <12ebering> I guess the passwall start message could be moved 08:55:54 Are there any start messages which shouldn't be moved? 09:02:01 <06advil> I found an ancient commit that moved the start messages into the delay to fix a bug, but I don't even recognize the bug number format 09:02:20 <06advil> I would guess there are potential timing issues with messaging before the delay is started 09:02:30 <06advil> e.g. I don't know whether the interrupt message will happen on that initial turn? 09:02:39 <06advil> %git 64707c2d565008 09:02:40 <04Cerebot> haranp * 0.4-a0-20-g64707c2: Delays now give their startup messages when they start, rather than when they are pushed onto the queue. Fixes 1797499. (14 years ago, 3 files, 28+ 24-) https://github.com/crawl/crawl/commit/64707c2d5650 09:03:01 <06advil> 1797499 is a lot of bugs 09:05:23 <12ebering> must be sourceforge 09:05:43 <06advil> ah maybe 09:07:24 <12ebering> rip 09:07:24 yeh, I think that star after the committer means sourceforge instead of github 09:07:33 <12ebering> no that's just chei/notchei's format 09:07:43 <12ebering> the bug number is not a mantis nor github bug nmber 09:08:18 yeh, just verified that elsewhere (ugh, notchei is slow) 09:09:00 the oddly sized hash is a valid clue though 09:10:18 <12ebering> ? 09:10:37 <12ebering> the big giveaway is the git-svn-id in the commit message 😉 09:11:27 <12ebering> https://sourceforge.net/p/crawl-ref/bugs/559/ 09:11:30 <12ebering> found it! 09:11:59 <06advil> lol, that bug is no longer relevant 09:12:09 <06advil> it does appear that all of the delay ::start() calls are just messaging 09:13:48 <06advil> I guess what happened in that bug was that one delay queued another, and if that can ever still happen there could be a similar issue 09:13:51 <12ebering> I'm not sure it can happen currently but the delay queue system was set up to allow it 09:14:04 <12ebering> I think the best thing to do would be to call start() in start delay if the queue is empty 09:14:10 <06advil> yeah, I don't see any current cases of this 09:14:19 <06advil> seems reasonable 09:14:46 <06advil> a bit worried about subtle consequences though, re making this kind of change right now 09:15:30 <12ebering> yeah 09:15:34 <12ebering> might hold off 09:15:47 <12ebering> hm yeah calling start alone isn't correct 09:15:47 <06advil> I can respond in the issue 09:18:37 <12ebering> yeah ok this is a bit subtle 09:20:09 <12ebering> some delays get started by commands, and so in the main loop don't do anything to trigger end of turn, so the main loop goes to the beginning and sees that you're now delayed and calls handle_delay and off you go 09:20:09 <12ebering> (things like run and autotravel and macros) 09:21:07 <12ebering> but passwall and shaft self for example go through ability/spellcasting paths 09:22:15 <06advil> possibly those spells could just message via regular means immediately, rather than through ::start? 09:23:21 <12ebering> yes, and need to double check their delay timer values 09:30:22 <06advil> I guess it might be possible also to keep passwall (and similar) from setting you.turn_is_over 09:30:28 <06advil> also sounds tricky 09:34:16 <08wormsofcan> clearly, meliai should be worshipping beeogh 09:35:26 <12ebering> Yeah. It’s subtle enough that passwall can be shorted a turn for this release, better than breaking something worse 09:36:19 <06advil> do you think passwall is actually incorrect? (I.e. different from intended?) Wasn't so clear to me from the report 09:39:25 <03wheals> it's been like this forever, right? so if we change how it works, i assume we'd make passwall take a turn longer 09:45:53 <06advil> huh I suspect it may actually be not what was intended? It's written as 1 + wall count, and definitely takes 2 + wall count, which is certainly not a crazy amount to choose, it just seems more likely that it's actually supposed to be 1 + count 09:45:54 <09hellmonk> Make it 1 + wall count but bring back getting stuck in the wall and instantly dying 09:48:40 <06advil> the current delay time goes back to 2008, following a brief window of 2 + wall count, which before that was 1 + wall count * 2 09:48:48 <06advil> also a different off by one error was fixed in 2009 or so 09:50:29 <06advil> !git 2eadd5f7cdf3ce 09:50:31 <04Cerebot> greensnark * 0.5-a0-102-g2eadd5f: Restore passwall to something closer to its former glory. (14 years ago, 2 files, 2+ 13-) https://github.com/crawl/crawl/commit/2eadd5f7cdf3 09:52:30 <06advil> so in terms of original intent, it depends on whether this commit was fixing a bug where 2 + howdeep was the intent but didn't work, or whether it intended to reduce the delay time from 2 + howdeep 09:52:30 <06advil> (anyways, 1 + walls seems like an easily understandable delay, regardless of what someone intended 14 years ago) 10:11:04 ??passwall 10:11:04 passwall[1/4]: Allows passing through diggable wall/statues with a free turn (useful for guaranteed stabbing) on the other side. Delay before passing is 2+distance turns and is not stopped if attacked. Range is 1-13 based on earth/spell power; guaranteed minimum is 1+earth/8 tiles (max of 4, at skill 24). In 0.15, range is based purely on spellpower. 10:30:25 Unstable branch on crawl.akrasiac.org updated to: 0.29-a0-3-g53e8333 (34) 10:30:30 <12ebering> I think it was intended as 1+wall count, and with how delays are handled it shows that value in time taken 10:31:02 <12ebering> Also, shaft self probably has similar issues! 10:44:29 <06advil> I see 2 + wall count in time taken? Or maybe I'm misunderstanding 10:47:26 03advil02 07* 0.29-a0-4-gb0e57e5: fix: explicitly set UI_NORMAL before load_level in webtiles 10(13 minutes ago, 1 file, 5+ 0-) 13https://github.com/crawl/crawl/commit/b0e57e52b49a 10:48:41 Unstable branch on crawl.kelbi.org updated to: 0.29-a0-4-gb0e57e52b4 (34) 10:55:39 <12ebering> Hm. I must have misremembered 10:55:45 <12ebering> bit distracted with work atm 10:56:35 04Build failed for 08master @ b0e57e52 06https://github.com/crawl/crawl/actions/runs/1752213705 10:58:37 <06advil> well that's new 10:58:42 <06advil> mingw-ld segfaulted 11:01:22 <06advil> re passwall, when I passwall a 1 width wall, Time shows 3.0 11:01:22 <06advil> is what I mean 11:01:39 <12ebering> ah, good! I somehow remember it displaying 2.0 but maybe that's just because I'd read the start_delay call recently and wanted to believe 11:25:02 04Build failed for 08master @ b0e57e52 06https://github.com/crawl/crawl/actions/runs/1752213705 11:33:36 <03wheals> ld still angry 11:36:22 <06advil> yeah I tried rerunning it 12:18:57 -!- kiedtl is now known as cot 14:01:23 New branch created: pull/2383 (1 commit) 13https://github.com/crawl/crawl/pull/2383 14:01:23 03Alex Jurkiewicz02 {GitHub} 07https://github.com/crawl/crawl/pull/2383 * 0.29-a0-5-ge459ec6: Change return signature for set_rc 10(4 minutes ago, 1 file, 11+ 3-) 13https://github.com/crawl/crawl/commit/e459ec68fa94 14:20:01 03Alex Jurkiewicz02 {advil} 07* 0.29-a0-5-gb8ed39c: Change return signature for set_rc 10(23 minutes ago, 1 file, 11+ 3-) 13https://github.com/crawl/crawl/commit/b8ed39c57d95 14:24:25 Stable (0.28) branch on underhound.eu updated to: 0.28-b1-1-g3c77fb2ed1 14:32:31 Unstable branch on crawl.kelbi.org updated to: 0.29-a0-5-gb8ed39c57d (34) 14:34:15 04Build failed for 08master @ b8ed39c5 06https://github.com/crawl/crawl/actions/runs/1753067770 14:40:38 <06advil> is anyone able to try running the mingw crosscompile locally? I'm not currently set up for it 14:58:44 03advil02 07* 0.29-a0-6-ga859fcf: ci: reset cache for mingw builds 10(2 minutes ago, 1 file, 3+ 3-) 13https://github.com/crawl/crawl/commit/a859fcf985be 15:02:40 Unstable branch on crawl.kelbi.org updated to: 0.29-a0-6-ga859fcf985 (34) 15:15:59 -!- BoostedStella is now known as stella 15:25:55 that commit seems to have fixed things, guess it was just a ccache glitch 16:23:59 Unstable branch on underhound.eu updated to: 0.29-a0-6-ga859fcf985 (34) 16:44:33 -!- Guest038 is now known as twelwe 16:49:25 I made a small Inner Flame post here https://crawl.develz.org/tavern/viewtopic.php?f=8&t=27250 its nice unlike usual! 16:50:29 The TL:DR is I believe the risks of Inner Flame are completely mitigated by Foxfire which which it probably already synergizes too well 20:01:26 <10PleasingFungus> i love the bit about anguish + inner flame 20:01:26 <10PleasingFungus> that's big tech 20:19:08 pleasin: i missed it! and i love (new) inner flame. it actually delivers on the arcane marksman concept 20:20:42 also, should this command maybe be deleted? perhaps it's a very funny joke but there's a lotta lotta clutter in !cmd 20:20:42 !getcasualclass 20:20:42 The RNG chooses: 闘士. 20:20:58 on my screen it's a box and a +-, maybe one of em's a discord emoji 20:29:53 <10PleasingFungus> it's the japanese word for 'gladiator' 20:30:08 <10PleasingFungus> it's an unfortunate corruption in the sequell db 20:30:12 oh! 20:30:12 <10PleasingFungus> from this game 20:30:17 <10PleasingFungus> !lg * cls=闘士 20:30:20 <04Cerebot> 1. houki the 強打する者 (L7 DsGl), レイスに殺された in Ossuary (ossuary_due_cavern) on 2020-10-24 17:16:54, with 585 points after 4057 turns and 0:08:11. 20:30:40 <10PleasingFungus> someone was running something they shouldn't have 20:30:47 <10PleasingFungus> and now we have this nonsense floating around 20:30:55 wild 20:34:59 where was the post about anguish + inner? 20:53:58 <12ebering> tavern 20:56:19 <09hellmonk> its cool that inner flame has gone from being bad to being good 21:09:23 <06advil> today in lines I'm surprised I got right on the first try: const spell_type hovered_spell = last_hovered >= 0 && items[last_hovered]->data ? *static_cast(items[last_hovered]->data) : SPELL_NO_SPELL; 21:09:28 <06advil> also lines that are bad 21:10:33 03advil02 07* 0.29-a0-7-g068bf27: fix: reset scroll on spell search 10(10 minutes ago, 1 file, 12+ 0-) 13https://github.com/crawl/crawl/commit/068bf27d475f 21:10:33 03advil02 07* 0.29-a0-8-g0446c0d: fix: use menu api instead of deleteAll in a few menus 10(4 minutes ago, 3 files, 3+ 3-) 13https://github.com/crawl/crawl/commit/0446c0dae4f8 21:18:17 Unstable branch on crawl.kelbi.org updated to: 0.29-a0-8-g0446c0dae4 (34)