00:22:18 Depths:3 prison 13https://crawl.develz.org/mantis/view.php?id=11444 by SirSquish 00:48:02 -!- amalloy is now known as amalloy_ 02:46:21 Monster database of master branch on crawl.develz.org updated to: 0.22-a0-348-gd16d139 09:59:12 I think that is minmay_the_grid_ultimate, right? 09:59:55 in 11444 10:15:30 oh hm this whole map can probably rotate can't it 10:22:53 looking at the des, it appears as if the lower right corner can be entirely blocked off if all "s convert to xs, I may be misunderstanding how this works though 10:23:01 presumably this should get vetoed, but I guess it doesn't 10:25:19 If I'm right about it, that would exactly generate the (large) closet seen in 11444, so I suspect that is what's happening 10:26:31 it'd be nice if static analysis on vaults were possible 10:28:18 the_grid vaults are not going to be ones that make that easy 10:28:25 !vault minmay_the_grid_ultimate 10:28:25 1/1. https://github.com/crawl/crawl/blob/master/crawl-ref/source/dat/des/variable/the_grid.des#L1495 10:29:32 ...right, yeah, that probably nips that bright idea in the bud 10:29:44 heh 10:31:14 so SUBVAULT can rotate a vault to fit into a region? as long as the silhouettes match up? 10:31:37 yes, I think everything there can rotate or invert around either axis if the shapes match 10:32:04 including the whole vault itself (since the closet in the bug report is inverted around the x axis from the thing I think might be going wrong) 10:33:17 well, it'd still be possible to do some sort of reachability analysis on each part, surely 10:33:39 yeah...in fact there are documented conditions (at the beginning of that file) that subvaults are supposed to meet 10:33:51 but the problem here doesn't seem to be substituting subvaults 10:34:09 it's with `e.subst('" : x...')` 10:36:09 which does what... substitute " (which are placeholders, I'm guessing) with walls some of the time? and once in a blue moon they trap the player? 10:36:37 I *think* it substitues "s for walls vs floors with 1/4 chance of walls 10:37:17 right, makes sense 10:37:42 there's supposed to be a guaranteed path, so probably there's one too many " in the .des, but minmay will have to look at it 10:41:53 there's something like 6 crucial "s and if certain combinations of 3 of those go to x, then the closet appears...so it's pretty rare 10:42:14 er 4 of those 10:42:34 and that flaw could be in one of the subvaults not satisfying the constraints, yeah? 10:42:59 the flaw causing this bug, that is 10:43:06 I don't think so 10:43:15 in that I checked the subvaults and they look fine 10:43:36 the "s are all in the template map 10:43:39 then it's a simple matter of s/"/x/, s/[abc...\.]/x/, and look for the rooms which are disconnected 10:43:56 now if only vim had floodfill... 10:44:49 er... s/[abc..\.]/ / i meaan 10:44:57 yes...if you're trying to figure out how to fix it, that's easy enough, but I'm not sure which fix is the "right" fix, which is why minmay's input is needed 10:45:34 how to prevent it in general, yeah, once the substitutions are all done, it should be detectable 10:47:31 yeah, there are three different disconnected regions, if you get unlucky 10:50:00 I only see that one (changed the des) 10:50:41 the diagonal paths are harder to detect visually from reading the .des 10:51:10 (changed the des to always give walls, that is) 10:51:23 er, four 10:52:22 if you change that line to e.subst('" : x') and then use &P to generate a level, you can use X to see reachability 10:52:29 I'm pretty sure there's only one disconnected region 10:52:48 https://a.pomf.space/ubnphkjhhgqq.png 10:53:07 there are a bunch of diagonals there :) 10:53:17 green and blue are connected 10:53:31 cool visualization though 10:53:40 doh! 10:54:22 ok, yeah, it's just the yellow region which is tricky 10:56:36 well, by symmetry it looks like the top of the yellow triangle (lines 1593-4) should be . 10:56:46 yeah, seems like it 11:12:07 are the original splash screen files (without the logo) still available anywhere? 11:12:48 it'd be nice to separate them, so that the splash screen can be scaled to fit without affecting the logo 11:13:42 not a huge priority, of course 12:03:28 -!- Vaevictu1 is now known as Vaevictus 13:40:06 -!- amalloy_ is now known as amalloy 14:13:12 Dargar (L1 DrGl) Crash caused by signal #15: Terminated (D:1) 16:28:37 XP seems odd in linecrawl (possibly in Sprint in general) 13https://crawl.develz.org/mantis/view.php?id=11445 by damerell 16:28:37 -!- flgr_ is now known as flgr 16:28:41 -!- atrodo_ is now known as atrodo 17:01:12 -!- Krakhan|2 is now known as Krakhan 17:17:19 Still on https://crawl.develz.org/wiki/doku.php?id=dcss:server:setting_up_dgamelaunch_and_webtiles - "Ensure that user crawl has write access to /var/mail so that console messages can be sent". Done that, but "You fall into the water! You sink like a rock. (Couldn't open newrobin's spool file. Aborting.)" Any hints before I go off strace-ing it, please? 17:20:40 !learn add THE_WIZARD'S_CASTLE https://archive.org/details/msdos_Wizards_Castle_The_1981 17:20:41 THE WIZARD'S CASTLE[6/6]: https://archive.org/details/msdos_Wizards_Castle_The_1981 17:22:34 if you're on a system with selinux, make sure the label on /var/mail allows user crawl? 17:22:52 or use permissive 17:23:10 I'm not on a system with selinux. crawl is in group "mail". One might ask "in the chroot or not", but I tried both 17:34:13 -!- scion is now known as Guest40266 17:54:31 open("/var/mail//newrobin", O_WRONLY|O_CREAT|O_APPEND, 0666) = -1 EACCES ( 17:54:32 Permission denied) 17:54:37 sorry for miswrap 17:55:02 Which is odd because if I as "crawl" do: perl -we 'use Fcntl; $foo="/var/mail//newrobin/"; sysopen FOO, $f 17:55:05 oo, O_WRONLY, 0666; print FOO "foo"; close FOO;' 17:55:12 it just works, and sorry for miswrap again, what is going on? 17:56:03 Also, gah, paste it with the right O_WHATEVER mode :-( 18:00:53 <|amethyst> Pinkbeast: maybe it's working outside of the chroot but not inside? 18:01:27 |amethyst: It doesn't seem to be working at all (and I did the perl equivalent in the chroot) 18:02:02 <|amethyst> hm 18:04:11 <|amethyst> oh, you said that 18:04:11 And I've checked I'm trying to send messages from a dgamelaunch started after crawl was added to group "mail" 18:04:55 <|amethyst> hm 18:05:25 <|amethyst> Oh, I bet dgamelaunch is just switching to user crawl, group crawl, without the supplemental groups 18:06:02 <|amethyst> on CSZO and CAO the /var/mail directory is owned by the crawl user 18:06:27 ... and since it's in the chroot I don't mind doing that, so let's try that 18:06:51 <|amethyst> (probably owned by the crawl group, or whatever is the shed_gid in /etc/dgamelaunch.conf, would work too) 18:07:06 Thanks! That worked. 18:07:27 <|amethyst> ah, yeah 18:07:30 <|amethyst> dgamelaunch does 18:07:35 <|amethyst> dgamelaunch.c: if (setgroups (1, &globalconfig.shed_gid) == -1) 18:07:45 <|amethyst> so it only switches to the one group 18:08:53 <|amethyst> (whereas when you're running perl, I bet you're logging in or using sudo or su to become crawl, and those programs do load the full supplemental group list) 18:09:41 Quite. Thanks. I forgot that could happen. 18:22:55 03advil02 07* 0.22-a0-349-g873ca8c: Fix a possible disconnected region in minmay_the_grid_ultimate (11444) 10(4 minutes ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/873ca8c3a950 18:48:55 03advil02 07* 0.22-a0-350-g910c0f9: Remove a swastika-ish arrival vault (11428) 10(3 minutes ago, 1 file, 0+ 17-) 13https://github.com/crawl/crawl/commit/910c0f972a9f 19:09:49 Unstable branch on crawl.jorgrun.rocks updated to: 0.22-a0-350-g910c0f972a (34) 20:10:13 -!- amalloy is now known as amalloy_ 20:37:13 poll: should unlinked temporary items get fixed up on load? 20:38:13 right now (as of 9 years ago), they are supposed, but amusingly it only works for FULLDEBUG builds, not for normal builds (because a relevant #define is poorly placed and affected by compilation order) 20:38:19 *supposed to 20:39:07 they rationale for the 9yo change is "otherwise they just get on your nerves" 20:39:11 which is fair 20:39:43 if they get fixed up we'll get less bug reports (not that bug reports in any amount of detail have ever helped with this set of bugs) 21:16:31 -!- H is now known as Guest3928 21:26:59 -!- amalloy_ is now known as amalloy 21:49:44 03advil02 07* 0.22-a0-351-g29664b4: Make the behavior of DEBUG_ITEM_SCAN more consistent 10(9 minutes ago, 5 files, 60+ 49-) 13https://github.com/crawl/crawl/commit/29664b41f07d 22:09:57 Unstable branch on crawl.jorgrun.rocks updated to: 0.22-a0-351-g29664b41f0 (34) 22:57:04 -!- mikee__ is now known as mikee_