00:13:25 Unstable branch on cbro.berotato.org updated to: 0.35-a0-105-g08978299a1 (34) 00:55:31 Monster database of master branch on crawl.develz.org updated to: 0.35-a0-105-g08978299a1 02:08:39 0.35-a0-105-g08978299a1 (windows installer) crashes on startup after loading screen, as did yesterday's version. I'm on windows 10 02:08:49 Sorry for not going through the chain of command but I'm short on time 02:09:33 Last version that doesn't explode is: crawl_setup-0.35-a0-71-g137d089076.exe (2026-02-22) 04:19:05 <07wizardike> I'm getting a crash in yyparse called from _parse_maps in tiles but not console in a fresh clone after https://github.com/crawl/crawl/commit/e745ec412119c44a6a836a1cf2074c5c88379668 when trying to parse "vaults_rooms_hard.des". Running tiles after running console succeeds though as the result is cached. I have no understanding of yyparse so I was unable to fix this 04:35:13 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5261-gd9800d219b 05:11:24 Unstable branch on crawl.akrasiac.org updated to: 0.35-a0-105-g0897829 (34) 10:38:37 <09gammafunk> I'll check out the crash 10:44:11 <04dracoomega> I can't seem to reproduce the crash when building myself (even with the des cache cleaned), though the windows tiles version from cdo does crash for me also. 10:46:31 <09gammafunk> linux tiles doesn't seem to crash, one very odd thing: 10:46:44 <09gammafunk> Regenerating des: /home/gammafunk/Games/dcss/crawl-dev/crawl-ref/source/dat/des/branches/vaults_rooms_hard.des %Regenerating des: /home/gammafunk/Games/dcss/crawl-dev/crawl-ref/source/dat/des/branches/vaults_rooms_standard.des 10:47:11 <09gammafunk> after or as it generates the vaults_rooms_hard.des file wizardike mentioned, it outputs a lone % sign 10:47:19 <09gammafunk> it doesn't do this for any other file 10:47:32 <04dracoomega> Weird 10:47:48 <09gammafunk> makes me suspect that ITEMS: % line 10:48:16 <09gammafunk> maybe the lex/yacc parser just has a terrible form of error reporting 10:48:39 <09gammafunk> I guess I should try the installer build in wine 10:51:25 <09gammafunk> hrm, wine tiles of that build run right after installation does not crash for me 10:55:35 <04dracoomega> You do seem to be right that the prebuilt windows version stops crashing once that ITEM: % line is deleted. 10:56:53 <04dracoomega> If it's so malformed, it's rather surprising that the other versions don't complain in any way, though 10:57:27 <09gammafunk> it's possible that item statements aren't smart like kitem and somehow can't translate the 1st item slot to any item 10:57:36 <09gammafunk> which I believe % is equivalent to 10:57:51 <04dracoomega> Oh, is that what's 'supposed' to happen here? 10:58:55 <04dracoomega> Hmmm... the vault docs themselves say A lone "%" may be used as an alias for "any". in the ITEM: entry 10:59:54 <09gammafunk> yeah, r-i is just assigning any to the d glyph, effectively 11:01:11 <09gammafunk> I do actually see both other uses of ITEM assigning % to a slot and other cases where the vaults rooms are using ITEM. The vaults rooms are special in that they're loaded into the parent vault in a special way 11:01:43 <04dracoomega> Actually, the other ITEM using % I can see have something else on the line, too 11:01:47 <04dracoomega> Like ITEM: % no_pickup 11:01:57 <04dracoomega> And changing the vaults_hard line to that also seems to let the windows version work 11:02:12 <04dracoomega> Given that this is a crash at the raw parser level, it might be something very specific? 11:02:24 <09gammafunk> it's very unusual, yes 11:02:43 <09gammafunk> perhaps we're seen some abberant parsing at the lex/yacc level somehow due to that "lone" % 11:02:54 <04dracoomega> That's kind of what it's feeling like 11:03:21 <09gammafunk> it would be easier to see what's going wrong specifically if we could recreate the crash, but for now you could simply add a dummy weight probably 11:03:32 <09gammafunk> e.g. w:100 which would be a no-op 11:03:39 <09gammafunk> since there are no alternatives 11:04:38 <09gammafunk> er, I meant if I could recreate the crash, since I could step through the parser with gdb 11:06:08 <04dracoomega> The lex/yacc stuff is definitely over my own head 11:09:46 <09gammafunk> I guess we're not seeing any kind of CI crash on the windows runners? 11:10:00 <09gammafunk> maybe we don't make test on those 11:22:02 <04dracoomega> Out of curiousity, I had it output the raw lua that it was parsing the .des file into (ie: print the string of every dlua_chunk::add() call to console, and the item line for that vault seems to be omitted entirely 11:22:15 <04dracoomega> Everything before and after it is fine 11:23:45 <04dracoomega> Some somehow the parser just isn't bothering to try and add that line because of (presumably) however it is misunderstanding what it's supposed to do with it 11:26:53 <04dracoomega> In fact, replacing the % with almost any character seems to result in it being skipped. As soon as there is any second character, it does get parsed and added. 11:29:37 <04dracoomega> It's extremely opaque to me why this would be happening or what about the syntax definition file could cause this, but it does appear to be the case 11:30:53 <09gammafunk> what if you leave the line as it originally was, but add another line directly above it that does some some kind of multi-char statement, like changing it to SUBST: 0 = 0 ITEM: % 11:31:56 <04dracoomega> Still gets ignored. SUBST: 0 = 0 ITEM: % SUBST: 0 : 0000z outputs subst("0 = 0") subst("0 : 0000z") 11:32:06 <09gammafunk> oh, one last thing 11:32:14 <09gammafunk> : local d = you.depth() + 1 : mons("centaur warrior draugr w:" .. 9 - d .. : " / vault warden draugr w:" .. d) : mons("death knight w:" .. 7 - d .. " / deep elf death mage w:" .. d) : mons("phantasmal warrior w:" .. 7 - d .. " / flayed ghost w:" .. d) 11:32:31 <09gammafunk> I just noticed that this happens to be setting a local variable d 11:32:39 <09gammafunk> which is the same as the slot name ITEM is using 11:33:09 <09gammafunk> which should not matter, but I wonder if making that not the case somehow, e.g. changing d to D or any other character 11:33:15 <09gammafunk> in said lua code, if that would fix 11:34:10 <04dracoomega> No difference. Like, this is all happening at a level before any lua can actually happening. It's just turning it into the lines of lua code that can be called later. 11:34:31 <09gammafunk> sure, just ruling out levels of weirdness 11:34:56 <04dracoomega> (Like, if I add any arbitrary invalid character to that ITEM: line, it parses that and then Crawl immediately complains downstream that this is an invalid item statement. But currently it's not an item statement at all) 11:34:57 <09gammafunk> I was suspicious of the transition between an inline lua chunk 11:35:13 <04dracoomega> Which doesn't crash at all for me in WSL, but apparently crashes in the windows version? 11:35:16 <09gammafunk> and a DES chunk 11:35:30 <09gammafunk> but that doesn't seem to be a problem here 11:35:45 <09gammafunk> because said inline lua is actually interacting with the parser, of course 11:36:07 <09gammafunk> but it's true that the ITEM slot logic happens downstream (in terms of using the d glyph) 11:36:19 <04dracoomega> Like, I sort of assume the problem lies in items : ITEM {} | ITEM item_specifiers {} ; in a way that it causing it to ignore single-character arguments, but I can't really read lex/yacc 11:37:08 <09gammafunk> hrm 11:37:11 <09gammafunk> well on that note 11:37:37 <09gammafunk> $ git grep '^ITEM: *.$' branches/vaults_rooms_hard.des:ITEM: % 11:37:58 <09gammafunk> this is the only instance in all of DES of an item statement having a single char as an argument 11:38:35 <09gammafunk> so I guess right now the fix is simply to replace % with any (maybe any item also works) 11:39:21 <09gammafunk> (that regex I used says ITEM: followed by zero or more whitespace followed by a single character followed by end of line) 11:39:21 <04dracoomega> (Crawl seems to complain about 'any item' but not 'any', for the record) 11:39:40 <09gammafunk> right any item is probably just invalid, didn't remember the exact syntax 11:40:10 <04dracoomega> (And yeah, it's possible this is one of those deeply fiddly problems we may never have to actually solve >.>) 11:40:33 <04dracoomega> Even if that line is supposed to be acceptable syntax 11:50:54 03DracoOmega02 07* 0.35-a0-106-ge802e43057: Fix a cryptic offline windows tiles crash 10(33 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/e802e4305757 12:07:18 <09gammafunk> thank goodness it wasn't an online windows tiles crash 12:08:16 <09gammafunk> call me a unix greybeard, but having to administer an windows-based tiles server sounds awful 12:12:01 <04dracoomega> Haha 12:12:28 <04dracoomega> I guess that was technically redundant, yes 12:18:00 03RypoFalem02 07https://github.com/crawl/crawl/pull/5072 * 0.35-a0-54-g6d8695de07: add the new options to the options list 10(22 minutes ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/6d8695de079f 12:18:00 03RypoFalem02 07https://github.com/crawl/crawl/pull/5072 * 0.35-a0-55-g7ea2b11f0c: change longwalk default to constant 15 10(6 minutes ago, 2 files, 4+ 4-) 13https://github.com/crawl/crawl/commit/7ea2b11f0cc6 12:31:00 03DracoOmega02 07* 0.35-a0-107-gf104df3b86: Fix being able to stampede out of mesmerism/fear (Odds) 10(12 minutes ago, 1 file, 22+ 16-) 13https://github.com/crawl/crawl/commit/f104df3b86bb 12:31:00 03DracoOmega02 07* 0.35-a0-108-g77b1fc6fcf: Fix interaction between Enkindle and Divine Exegesis (canofworms) 10(33 seconds ago, 2 files, 5+ 2-) 13https://github.com/crawl/crawl/commit/77b1fc6fcfa1 12:31:01 03DracoOmega02 07[stone_soup-0.34] * 0.34.0-26-gae8d228adc: Fix interaction between Enkindle and Divine Exegesis (canofworms) 10(35 seconds ago, 2 files, 5+ 2-) 13https://github.com/crawl/crawl/commit/ae8d228adce9 12:38:39 03WizardIke02 07* 0.35-a0-109-gd07bcf4d7f: Refactor moving areas in the abyss 10(8 hours ago, 1 file, 3+ 14-) 13https://github.com/crawl/crawl/commit/d07bcf4d7f1a 12:38:39 03WizardIke02 07* 0.35-a0-110-g0e3de29d76: Fix Abyssal rune detection nuking map knowledge 10(5 hours ago, 2 files, 29+ 21-) 13https://github.com/crawl/crawl/commit/0e3de29d76c3 12:38:39 03WizardIke02 07* 0.35-a0-111-gadea0f3bd7: Fix the minimap missing out of view squares after Abyss shift 10(4 hours ago, 1 file, 2+ 5-) 13https://github.com/crawl/crawl/commit/adea0f3bd702 12:38:39 03WizardIke02 07* 0.35-a0-112-g4107091cb8: Remove unused function declaration 10(4 hours ago, 1 file, 0+ 2-) 13https://github.com/crawl/crawl/commit/4107091cb892 12:38:39 03WizardIke02 07* 0.35-a0-113-gfccfe28bf2: Fix Xom moving an altar braking travel to closest altar 10(53 minutes ago, 1 file, 3+ 0-) 13https://github.com/crawl/crawl/commit/fccfe28bf227 12:44:04 03Dossy Shiobara02 {GitHub} 07* 0.35-a0-114-g0c39c3cb1f: fix: correct crawl.grammar re: qual vs. qualname (#5053) 10(33 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/0c39c3cb1fcf 16:43:25 Unstable branch on underhound.eu updated to: 0.35-a0-114-g0c39c3cb1f (34) 18:13:36 <09gammafunk> Tournament results post: https://crawl.develz.org/wordpress/0-34-tournament-results 23:35:51 Unstable branch on crawl.develz.org updated to: 0.35-a0-114-g0c39c3cb1f (34) 23:58:54 Windows builds of master branch on crawl.develz.org updated to: 0.35-a0-114-g0c39c3cb1f