00:38:30 <ebering> good idea or best idea: all shafts are 3 floor shafts
00:44:34 <gammafunk> ebering: why only 3? Many branches have more than 3 levels...
01:01:30 -!- Pinstripelli is now known as Bcadren
01:13:35 -!- Rast- is now known as rast
01:50:37 <gammafunk> |amethyst: https://stackoverflow.com/questions/8234779/how-to-remove-from-a-map-while-iterating-it ; is that top answer correct in your view?
01:50:46 <gammafunk> specifically the recommendation for C++11 of     m.erase(it++);    // or "it = m.erase(it)" since C++11
01:51:02 <gammafunk> namely this: it = m.erase(it);
02:03:52 -!- amalloy_ is now known as amalloy
02:06:11 <amalloy> CanOfWorms: someone reported that as a bug already btw, sorta. the save/load tech
02:06:39 <amalloy> if you shoot arrows at something and it dies, apport doesn't necessarily work because the arrows aren't always on top of stack
02:06:47 <amalloy> but if you save/quit then it does because A is for Arrow
02:08:26 <amalloy> |amethyst, CanOfWorms: re the question of xv interface for a stack, what about leaving it as is for xv but adding a new key to x that's like "cycle through the (known) items in the stack i'm pointing at"? so you could do like x';;;v to glance at the first three items in a stack and then peer intently at the fourth
02:10:18 <CanOfWorms> I think at that point the key could just bring up the list
02:10:39 <amalloy> maybe that's better, yes
02:15:42 <Rotatell> ziggurnaut (L3 DgIE)  (D:3)
02:16:49 <gammafunk> amalloy: one of the methods that you pointed out should be simplified and split into two was the cause of the travis crash
02:17:13 <amalloy> nice. i don't remember what it was
02:17:13 <gammafunk> I added a "ignore this type" condition and tried to only clearl the map if it was empty
02:17:27 <gammafunk> but I forgot to actually remove entries from the map
02:17:32 <gammafunk> stop_constricting_all()
02:17:36 <gammafunk> actor::
02:17:53 <gammafunk> so I'm splitting it into two, since the unconditional one is very short and simple
02:18:27 <gammafunk> you can probably read it and see why it would crash
02:18:44 <gammafunk> latter an assert expected the constricting pointer to be empty
03:59:46 -!- amalloy is now known as amalloy_
04:02:22 -!- amalloy_ is now known as amalloy
04:31:30 <aidanh> |amethyst: are you in charge of s-z.org? on /crawl-dev/ over https, loading jQuery is blocked because it has a http URL
04:56:04 -!- amalloy is now known as amalloy_
07:44:43 <Yermak> CanOfWorms: https://crawl.develz.org/mantis/view.php?id=10337
08:45:29 <Cheibriados> New branch created: pull/681 (1 commit) 13https://github.com/crawl/crawl/pull/681
08:45:29 <Cheibriados> 03Aidan Holm02 07https://github.com/crawl/crawl/pull/681 * 0.21-a0-585-gc445f98: Remake god description UI with formatted_scroller 10(2 days ago, 5 files, 222+ 214-) 13https://github.com/crawl/crawl/commit/c445f9820121
09:43:10 <Cheibriados> 03advil02 07* 0.21-a0-585-g8bd9a9d: Don't create buggy ghosts in wizmode 10(4 days ago, 1 file, 2+ 0-) 13https://github.com/crawl/crawl/commit/8bd9a9d9c5cb
09:43:10 <Cheibriados> 03advil02 07* 0.21-a0-586-g907874f: Refactor ghost load/save to avoid global state 10(4 days ago, 6 files, 148+ 114-) 13https://github.com/crawl/crawl/commit/907874f4939b
09:43:10 <Cheibriados> 03advil02 07* 0.21-a0-587-g49f8b85: Move some initialization inside ghost_init 10(3 days ago, 3 files, 3+ 5-) 13https://github.com/crawl/crawl/commit/49f8b8591922
09:43:10 <Cheibriados> 03advil02 07* 0.21-a0-588-g79d79c4: Allow saving arbitrary vectors of ghosts 10(3 days ago, 8 files, 13+ 14-) 13https://github.com/crawl/crawl/commit/79d79c4c6a2a
09:43:10 <Cheibriados> 03advil02 07* 0.21-a0-589-geac747f: Allow partial loading of bones files, and enforce level limits on load 10(3 days ago, 5 files, 32+ 31-) 13https://github.com/crawl/crawl/commit/eac747fae7cb
09:46:52 <advil> johnstein: I just changed some of the technical details of how bones files work (in hopefully relatively subtle ways), if you happen to spot or hear of any oddities on cbro with bones files over the next few days, lmk
09:56:56 <Cheibriados> 03Aidan Holm02 {advil} 07* 0.21-a0-590-ge1710fc: Add uncollected rune/orb tiles for } menu 10(2 weeks ago, 24 files, 63+ 7-) 13https://github.com/crawl/crawl/commit/e1710fce0b25
09:58:29 <aidanh> oh, damn, was just about to fix up the demonic rune tile's extra pixel
09:58:35 <advil> heh oops
09:58:50 <advil> if you do it I'll merge the PR right away
10:01:19 <advil> yeah, people just add lots of linebreaks in complicated ? : statements
10:02:26 <advil> I think the basic idea for a scenario like that would be that it's easier to verify the (lack of) fallthroughs quickly if there's just a single return statement for each
10:02:46 <advil> I'm still adapting to this convention myself, it's not very natural coming from python
10:07:17 <aidanh> yup, makes sense
10:07:43 <|amethyst> gammafunk:  for std::map that seems good (likewise for std::list, and std::unordered_map).  It won't work with vector
10:08:26 <gammafunk> |amethyst: I tried to use it with the for loop initialization I had and got template errors
10:08:43 <gammafunk> so I switched to putting the items I wanted to clear into a vector like some other actor methods do
10:08:47 <gammafunk> and did a follow-up loop
10:09:11 <|amethyst> template errors?
10:09:21 <gammafunk> let me see if I can find
10:09:37 <Jorgrell> Unstable branch on crawl.jorgrun.rocks updated to: 0.21-a0-590-ge1710fce0b (34)
10:11:21 <advil> hmm #661 has some submodule weirdness in it
10:11:29 <advil> wonder how to deal with that
10:11:39 <|amethyst> aidanh:  thanks for spotting the log-viewer problem; I changed the jquery URL to https
10:13:00 <|amethyst> advil:  cd to top level, git submodule update, commit to the PR, and probably squash
10:13:10 <|amethyst> advil:  probably they committed it with the older sdl2
10:13:18 <advil> yeah, looks like it
10:13:19 <advil> thanks
10:13:24 <|amethyst> advil:  because they didn't do a submodule update, but did git add -a or something
10:13:58 <|amethyst> advil:  btw, have you gotten many bounces from that mail?
10:14:10 <advil> I got 0 anything from that email
10:14:21 <advil> I wonder if even the bounces got filtered
10:14:27 <|amethyst> I did receive it at least
10:14:33 <advil> ah ok was just about to ask you
10:14:37 <|amethyst> other people have gotten bounces from some addresses when sending
10:15:08 <advil> I should have gotten it back myself, right?
10:15:08 <|amethyst> you didn't receive a copy of your own mail, though?
10:15:10 <|amethyst> yeah
10:15:10 <advil> no
10:15:36 <advil> though sometimes gmail is weird about that
10:16:45 <advil> i.e. it collapses the sent/received version sometimes for mailing lists
10:17:35 <advil> gammafunk: did you get a sekret email? you were using gmail also for that iirc
10:19:00 <gammafunk> advil: I don't seem to have
10:19:07 <advil> hm
10:19:10 <gammafunk> oh I just got it sorry
10:19:11 <|amethyst> I just forwarded it
10:19:43 <|amethyst> gammafunk:  you got the one that was From: advil rather than me?
10:19:58 <gammafunk> oh, yeah, got the one from |amethyst
10:20:26 <|amethyst> and now everyone on the list has the full list, since it's in the To:
10:20:34 <gammafunk> none from advil nor any message from him in my spam folder
10:20:49 <|amethyst> it's entirely possible the bounces haven't happened yet
10:21:09 <advil> thanks for forwarding
10:21:10 <|amethyst> greylist it first, *then* bounce it when that times out
10:21:26 <advil> ah, yeah maybe
10:21:30 <advil> heh I hope this doesn't get my gmail address on some kind of bad list :-P
10:21:51 <gammafunk> can we just switch to snail mail? TIA
10:21:58 <gammafunk> one person is designated letter writer
10:23:34 <gammafunk> Sequell, no!
10:24:59 <|amethyst> advil:  it's more likely to get dobrazupa on a bad list
10:25:19 <|amethyst> advil:  since otherwise you could easily Joe-job enemies onto said list
10:25:33 <advil> heh
10:26:41 <aidanh> advil: oh, never mind, looks like I fixed that already
10:29:53 <advil> |amethyst: here's another possible variant of what we were discussing before that has minimal effort on all ends, I set up a one-click (well, it would be somewhere between 3 and 10) closed/private mailman list using dreamhost (could be crawl-private@dcss.xyz) and you have the old alias forward to that
10:32:38 <|amethyst> advil:  assuming dreamhost doesn't filter those forwards out, at least
10:32:49 <|amethyst> but, with some testing, it sounds reasonable
10:34:08 <|amethyst> advil:  depending on how much ability to configure things you have, I could even point dobrazupa.org's MX to whatever
10:34:31 <|amethyst> advil:  since, other than spam, the secret list is literally the only mail that goes through dobrazupa.org
10:34:45 <travis-ci> The build has errored. (master - eac747f #9073 : advil): https://travis-ci.org/crawl/crawl/builds/319736271
10:34:46 <|amethyst> advil:  I guess we'd need to figure out what to do with abuse@ and so on
10:34:55 <advil> right, I did some minimal googling the other day that made me think that the forwarding should be possible. to test it I could set up the email list with only us on it and you could try out the forward
10:35:05 <|amethyst> doh, it's later than I thought; I must be going for a while, later
10:35:09 <advil> later
10:35:17 <|amethyst> advil:  might want to find one of the hotmail (etc) users to test too
10:35:29 <|amethyst> though if it goes through to gmail it's probably fine
10:37:23 <aidanh> |amethyst: great, thanks
10:42:01 <aidanh> !tell |amethyst 1) probably better to use // not https:// -- it'll choose the one that matches the page protocol, and 2) the html is broken (starts with <script> tag) (not sure if it was broken before)
10:42:02 <Sequell> aidanh: OK, I'll let |amethyst know.
11:02:46 -!- jolamb_ is now known as jolamb
11:44:25 <Cheibriados> 03Yermak02 {advil} 07* 0.21-a0-591-g2314aa1: Require a single keypress to abort ?amnesia 10(13 days ago, 1 file, 8+ 34-) 13https://github.com/crawl/crawl/commit/2314aa112c07
11:45:25 <advil> I couldn't get submodule update to do what I wanted for some reason, I ended up just manually editing a patch
11:46:52 <advil> "You can't reactivate evo flight while you're drain-hovering." anyone know if this is by design?
11:47:32 <advil> I guess it makes sense
11:47:43 <advil> if the drain penalty is to mean anything
12:09:35 <Jorgrell> Unstable branch on crawl.jorgrun.rocks updated to: 0.21-a0-591-g2314aa112c (34)
12:21:58 <|amethyst> advil:  yeah, otherwise it might as well just be boot flying
12:21:58 <Sequell> |amethyst: You have 2 messages. Use !messages to read them.
12:22:18 <advil> yeah, I just hadn't thought it through
12:23:07 <|amethyst> aidanh:  re //, I actually think I prefer https anyway, so that at least that part won't be MITMed
12:23:31 <|amethyst> hm, wonder what happened to the <head>
12:24:46 <aidanh> any http part means the whole thing is insecure, really, because the biggest threat is probably the ISP
12:24:55 <|amethyst> aidanh:  I fixed 2
12:24:58 <aidanh> as long as it works with https I'm happy though :)
12:25:03 <aidanh> great!
12:25:55 <Henzell> Unstable branch on crawl.akrasiac.org updated to: 0.21-a0-591-g2314aa1 (34)
12:25:57 <|amethyst> advil:  I just got my welcome to the list, thanks!
12:26:55 <advil> ok, good
12:26:57 <advil> it's just the two of us right now
12:27:32 <|amethyst> aidanh:  I was actually a little surprised that I had HTTPS at all, I guess s-z's host added that
12:31:54 <|amethyst> advil:  just sent two mails, one from neil@s-z.org and one from neil@cs.uky.edu
12:32:12 <|amethyst> advil:  looks like both got through without any moderation
12:32:25 <advil> hm I don't see the uky one (it is supposed to be set private)
12:32:32 <|amethyst> advil:  oh, no, sorry
12:32:42 <|amethyst> advil:  I was seeing the mail from BCC: neil
12:32:48 <advil> ah
12:32:49 <|amethyst> advil:  yeah, only one came through from the list
12:33:02 <|amethyst> and now I get the "requires approval" on the other one
12:33:04 <|amethyst> great!
12:33:34 <|amethyst> ah, so we're both listed as admins?
12:33:41 <advil> yeah, I put you as an owner
12:33:52 <advil> just need to get you the admin password somehow
12:33:52 <|amethyst> so you got the "requires approval" too now?
12:34:31 <advil> well google spamfiltered it, but yes
12:34:51 <|amethyst> advil:  you can just drop that message if you want, or approve it, either way
12:35:08 <advil> ok...it's set to no archive at the moment so it doesn't really matter
12:35:29 <|amethyst> yeah, no archive is good I think
12:36:59 <advil> right
12:37:29 <advil> since most of it is discussions about future list members
12:37:57 <advil> (btw let me know if you don't want to be a list admin)
12:39:44 <|amethyst> I don't care one way or the other
12:40:07 <|amethyst> probably is good to have multiple list admins in the case of emergencies, but I don't see myself doing much adminning
12:40:41 <|amethyst> let's see
12:42:11 <Cheibriados> 03gammafunk02 07[notzombiehands] * 0.21-a0-591-gb5e88c4: Refactor constriction code (amalloy, PR#678) 10(31 minutes ago, 11 files, 137+ 69-) 13https://github.com/crawl/crawl/commit/b5e88c436fbd
12:42:11 <Cheibriados> 03gammafunk02 07[notzombiehands] * 0.21-a0-592-geba8ca7: Require solid ground for Borgnjor's Vile Clutch (amalloy, PR#678) 10(31 minutes ago, 1 file, 8+ 3-) 13https://github.com/crawl/crawl/commit/eba8ca756031
12:42:11 <Cheibriados> 03gammafunk02 07[notzombiehands] * 0.21-a0-593-g0ab4757: Fix spacing (amalloy, PR#678) 10(63 minutes ago, 1 file, 3+ 4-) 13https://github.com/crawl/crawl/commit/0ab4757fa821
12:42:11 <Cheibriados> 03gammafunk02 07[notzombiehands] * 0.21-a0-594-gbb6e1b2: Shorten a spell noise check (amalloy, PR#678) 10(62 minutes ago, 1 file, 4+ 3-) 13https://github.com/crawl/crawl/commit/bb6e1b24a7fc
12:42:11 <Cheibriados> 03gammafunk02 07[notzombiehands] * 0.21-a0-595-g63f7c27: Satisfy grammarians (amalloy, PR#678) 10(61 minutes ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/63f7c2777d62
12:42:11 <Cheibriados> 03gammafunk02 07[notzombiehands] * 0.21-a0-596-g2ca2f0c: Fix logic in a constriction damage check (amalloy, PR#678) 10(60 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/2ca2f0cec8ce
12:42:11 <Cheibriados> 03gammafunk02 07[notzombiehands] * 0.21-a0-597-g3caa6c9: Tweak Kiku's book gifts and remove a double entry (amalloy, PR#678) 10(59 minutes ago, 1 file, 4+ 5-) 13https://github.com/crawl/crawl/commit/3caa6c98a40f
12:42:11 <Cheibriados> 03gammafunk02 07[notzombiehands] * 0.21-a0-598-g31419a2: Allow Vile Clutch to work under Sacrifice Love (amalloy, PR#678) 10(59 minutes ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/31419a2a8d00
12:42:18 <Cheibriados> 03gammafunk02 07https://github.com/crawl/crawl/pull/678 * 0.21-a0-591-gb5e88c4: Refactor constriction code (amalloy, PR#678) 10(31 minutes ago, 11 files, 137+ 69-) 13https://github.com/crawl/crawl/commit/b5e88c436fbd
12:42:18 <Cheibriados> 03gammafunk02 07https://github.com/crawl/crawl/pull/678 * 0.21-a0-592-geba8ca7: Require solid ground for Borgnjor's Vile Clutch (amalloy, PR#678) 10(31 minutes ago, 1 file, 8+ 3-) 13https://github.com/crawl/crawl/commit/eba8ca756031
12:42:18 <Cheibriados> 03gammafunk02 07https://github.com/crawl/crawl/pull/678 * 0.21-a0-593-g0ab4757: Fix spacing (amalloy, PR#678) 10(63 minutes ago, 1 file, 3+ 4-) 13https://github.com/crawl/crawl/commit/0ab4757fa821
12:42:18 <Cheibriados> 03gammafunk02 07https://github.com/crawl/crawl/pull/678 * 0.21-a0-594-gbb6e1b2: Shorten a spell noise check (amalloy, PR#678) 10(62 minutes ago, 1 file, 4+ 3-) 13https://github.com/crawl/crawl/commit/bb6e1b24a7fc
12:42:18 <Cheibriados> 03gammafunk02 07https://github.com/crawl/crawl/pull/678 * 0.21-a0-595-g63f7c27: Satisfy grammarians (amalloy, PR#678) 10(61 minutes ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/63f7c2777d62
12:42:18 <Cheibriados> 03gammafunk02 07https://github.com/crawl/crawl/pull/678 * 0.21-a0-596-g2ca2f0c: Fix logic in a constriction damage check (amalloy, PR#678) 10(60 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/2ca2f0cec8ce
12:42:18 <Cheibriados> 03gammafunk02 07https://github.com/crawl/crawl/pull/678 * 0.21-a0-597-g3caa6c9: Tweak Kiku's book gifts and remove a double entry (amalloy, PR#678) 10(60 minutes ago, 1 file, 4+ 5-) 13https://github.com/crawl/crawl/commit/3caa6c98a40f
12:42:18 <Cheibriados> 03gammafunk02 07https://github.com/crawl/crawl/pull/678 * 0.21-a0-598-g31419a2: Allow Vile Clutch to work under Sacrifice Love (amalloy, PR#678) 10(59 minutes ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/31419a2a8d00
12:42:58 <|amethyst> advil:  did you get the "Testing forwarding..." message I just sent?
12:43:16 -!- amalloy_ is now known as amalloy
12:43:26 <|amethyst> advil:  I haven't received it yet, but I *just* sent it
12:43:43 <gammafunk> !tell amalloy I updated the PR/branch based on your feedback; let me know if you see anything in that amiss (if you have time)
12:43:44 <Sequell> gammafunk: OK, I'll let amalloy know.
12:43:45 <advil> yes, but it was spamfiltered again
12:43:54 <|amethyst> hmmm
12:43:56 <advil> wonder if .xyz domains are suspect
12:44:00 <|amethyst> okay, it came through for me
12:44:25 <|amethyst> advil:  huh
12:44:45 <amalloy> |amethyst: if you sent it to me i didn't get it at all, not even spam
12:44:45 <Sequell> amalloy: You have 1 message. Use !messages to read it.
12:44:48 <advil> I haven't used this domain for emails at all before
12:44:55 <|amethyst> amalloy:  no, you're not on the new list yet
12:44:58 <advil> amalloy: we are testing a new list
12:45:00 <amalloy> wow, snubbed
12:45:03 <advil> heh
12:45:08 <|amethyst> advil:  possible... spamassassin let it through, but google is kind of picky
12:45:08 <|amethyst> X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00,RCVD_IN_DNSWL_NONE autolearn=ham autolearn_force=no version=3.4.1
12:45:57 <|amethyst> oh, but of course my SA has been trained to not think I'm spam
12:46:30 <|amethyst> the *other* spamassassin, running on s-z.org, gave me
12:46:54 <|amethyst> well, a bunch of thing, 3.1 points, with 5 points needed to mark it as spam
12:47:08 <|amethyst> 2.1 for coming from an .xyz address
12:47:13 <advil> heh
12:47:39 <|amethyst> and 1.0 for KAM_LAZY_DOMAIN_SECURITY Sending domain does not have any anti-forgery methods
12:47:39 <|amethyst> which is probably my fault, since s-z.org doesn't have SPF
12:48:05 <|amethyst> (also 0.5 for being "linkbait", and -0.5 from the bayesian filter)
12:48:30 <advil> does seem like there are a lot of bad actors using .xyz, I guess you get what you pay for (I bought it on sale)
12:48:49 <|amethyst> heh
12:51:07 <advil> old c-p is currently set to forward?
12:51:11 <|amethyst> yeah
12:51:35 <|amethyst> so you can add people whenever, I guess
12:54:45 <|amethyst> advil:  hm, no, the "needs approval" message even gets KAM_LAZY_DOMAIN_SECURITY, so I guess that's on your end
12:55:02 <|amethyst> advil:  (the one I just got from your test)
12:55:18 <|amethyst> well, "you" meaning the domain the list is running on
12:56:44 <|amethyst> advil:  hm, and the approved copy of that message didn't even get that report header; not sure what's up there
12:57:07 <|amethyst> advil:  also, nice having an email address even shorter than mine
12:57:26 <advil> heh, I was glad to get that one
12:57:52 <advil> that one at least wasn't spamfiltered by google
12:58:13 <advil> well I guess a mailing list where things get spamfiltered is still better than one where things just disappear
12:58:33 <|amethyst> and a mailing list is better than a sendmail alias :)
12:58:55 <advil> should I just mass add the list? I feel slightly weird doing that
12:59:24 <|amethyst> advil:  I think so... at least this way people can unsubscribe themselves if they want to
12:59:49 <advil> maybe you could send an email to the alias explaining what we're doing, since many more of the people on there know you than me?
13:00:14 <|amethyst> advil:  the alias won't work, but I'll send that as a reply to dpeg's email
13:00:30 <advil> yeah, or just to the raw list
13:00:44 <|amethyst> oh, that one just went to us
13:01:12 <|amethyst> oh, no, it went to the old list too, so most people didn't see it, but some did
13:01:29 <advil> I think dpeg might have replied to the alias actually
13:09:04 <advil> just to check, did you send it?
13:10:08 <|amethyst> I *just* did
13:10:22 <|amethyst> to the raw list of addresses
13:11:12 <advil> ok, thanks
13:11:19 <advil> I don't suppose you know how to write spf records?
13:11:26 <|amethyst> nope
13:11:44 <|amethyst> https://tools.ietf.org/html/rfc7208  does I bet
13:11:57 <|amethyst> though maybe not in a practical way :)
13:12:40 <|amethyst> I was saddened when they rebackronymed SPF
13:13:10 <|amethyst> I mean, I guess "Sender Policy Framework" encompasses more than "Sender Permitted From", but it doesn't have the zing
13:13:19 <amalloy> |amethyst: reviewing gammafunk's PR makes me wonder: why is actor::constricting a *map<mid_t, int> that gets allocated and freed as it becomes filled/empty respectively, rather than just being a map<mid_t, int> that might be empty?
13:14:05 <|amethyst> amalloy:  to save some space in the monster struct/array
13:14:05 <|amethyst> amalloy:  I think it was kilobyte's idea
13:14:20 <amalloy> i mean, sure, that seems like the obvious reason. i guess my question is, is that actually a good reason
13:14:57 <|amethyst> probably not, unless someone's planning on porting it to a 16 MB machine or something
13:15:12 <amalloy> what's the overhead of an empty std::map anyway?
13:15:25 <|amethyst> let's see
13:15:35 <amalloy> probably more than a single pointer, i imagine, so there is a nonzero space difference
13:16:02 <amalloy> but not likely to be worth the dev-time overhead of having to handle the constricting list so delicately
13:18:03 <|amethyst> amalloy:  looks like the sizeof() is 24, but I don't know if that is the full story for an empty list
13:18:24 <|amethyst> well, I should say, with GNU libstdc++ on a 32-bit system
13:19:22 <|amethyst> so not quite half a cache line
13:20:08 <|amethyst> yeah, I would not be opposed at all to making it inline
13:20:31 <amalloy> okay. well, something to do after gammafunk's thing is merged. i'm not quite sadistic enough to give him that merge conflict
13:24:52 <gammafunk> bring it you're nothing but a grammarian
13:25:50 <|amethyst> (as this case illustrates, kilobyte's nick is most appropriate, since he cares about every kilobyte.  I imagine he started programming on 8-bit systems)
13:29:56 <advil> ok, well I think I've created a spf record, no idea if it'll help or hurt
13:34:42 <amalloy> trying very hard not to get sunburned, eh?
13:36:44 <gammafunk> you gotta watch out for those sun demons
13:39:57 <advil> ugh still seeing spf neutral
13:40:01 <advil> I have no idea what I'm doing
13:41:33 -!- crate_ is now known as crate
13:53:12 <aidanh> are ghost/illusion dolls supposed to rotate through each possible tile?
14:19:42 <advil> ok, mass subscribe finally done
14:19:57 <travis-ci> The build was fixed. (notzombiehands - 31419a2 #9077 : gammafunk): https://travis-ci.org/crawl/crawl/builds/319824593
14:20:03 <advil> I *think* I got the spf record right but who knows
14:20:43 <|amethyst> advil:  I just got five more subscription notifications
14:20:56 <advil> yeah, you probably get all those bounce notifications too
14:21:11 <|amethyst> oh, I see, these aren't for me
14:21:28 <|amethyst> I was a bit worried that I had been subscribed six times :)
14:21:31 <advil> heh one immediate unsubscribe
14:28:47 <advil> |amethyst: were you able to get the admin password somehow?  should I send it to you via some probably not perfect means? (keybase? signal? other?)
14:39:44 <|amethyst> what are these things?
14:39:55 <|amethyst> I assume the admin password is something separate from my list password
14:40:26 <advil> yeah
14:41:06 <advil> I can just send it to you plaintext too, it's not as if this is super s3cure
14:41:23 <|amethyst> yeah, just /msg it to me
14:41:48 <|amethyst> I guess you could ctcp it if you're worried about freenode admins :)
18:17:59 <Lantell> Unstable branch on underhound.eu updated to: 0.21-a0-591-g2314aa112c (34)
18:48:45 -!- Amnesiac_ is now known as Amnesiac
19:10:24 -!- amalloy is now known as amalloy_
19:33:16 -!- amalloy_ is now known as amalloy
20:14:34 -!- Electric is now known as reaverb
21:33:23 <advil> huh, crazy...placing a player ghost in a vault works, and it just mirrors you
21:33:27 <advil> it must be the illusion code
21:33:35 <advil> already works, that is
21:35:13 <advil> of course now I need to figure out where on earth that happens
21:35:24 <hellmonk> is there any chance of pghost vaults before the tournament
21:35:48 <hellmonk> or is it strictly a next version thing
21:36:15 <advil> I was hoping so, but time is passing and it's hard to imagine they won't need some testing / discussion
21:36:15 <hellmonk> :(
21:36:33 <advil> I'm hoping to have a draft before xmas though
21:36:57 <advil> the code is fighting me at every step, I have to say
21:40:44 <advil> hellmonk: do it in hellcrawl so then I can steal the code
21:40:52 <advil> although you must've removed ghosts or something
21:42:10 <hellmonk> advil: yeah I removed ghosts a long time ago sorry
21:42:17 <hellmonk> like the first version of hellcrawl actually
21:42:53 <advil> heh
21:45:00 <hellmonk> it was a one line change
21:52:06 <advil> yeah, they are very separate from levelgen, which was sort of a pleasant surprise
22:04:49 <gammafunk> advil: I was thinking that would be a tough change to really get finalized before, say, 29th
22:05:04 <advil> probably
22:05:17 <gammafunk> If you think you can get there at least by then, but I hope to start the freeze around then
22:05:21 <gammafunk> as it is, we'll only have a week
22:05:27 <gammafunk> before the release/tourney
22:05:58 <gammafunk> oh and that's an fyi to anyone else in -dev with changes they're working on that wouldn't be appropriate during a freeze
22:06:43 <gammafunk> as always, bug fixes and minor tweaks are fine in that last week, just consider that some tournament servers install a stable version once
22:06:53 <gammafunk> and it can be difficult to get them to update it afterwards
22:07:22 <gammafunk> CWZ has been particularly challenging but I think hong is more available via email these days
22:07:37 <gammafunk> I haven't seen alex around so he may not have a lot of time either
22:07:57 <advil> I saw him say he's unavailable until next year
22:08:03 <advil> wouldn't even update hellcrawl
22:08:04 <gammafunk> ok, hopefully he can at least get 0.21 on cpo for tourney
22:08:17 <advil> my basic todo list is (i) ghost vault draft, (ii) maybe a new targeter for wall jump (I hate working on targeters), (iii) merge some of the bugfix PRs
22:08:43 <advil> do we need any more WJC work?
22:08:52 <gammafunk> I'm going to do a WJC game hopefully
22:08:58 <gammafunk> more to look at possible balance issues
22:09:17 <gammafunk> just need to get this branch pushed tonight, but I would say the god is "probably fine" for release
22:09:33 <gammafunk> it doesn't have any show-stopping issues in terms of basic design right now
22:09:33 <advil> yeah, I'm ok with releasing it, though I want to play more too
22:10:14 <advil> there's probably other UI work
22:10:40 <advil> there was a tavern thread where people seemed fairly confused about what the abilities do
22:10:45 <advil> so that could be communicated better
22:11:03 <|amethyst> ASCII art
22:11:20 <advil> well the ascii art is pretty good for what it conveys, but it doesn't convey the details
22:11:43 <advil> the thread may be kind of useful for tweaking info actually: https://crawl.develz.org/tavern/viewtopic.php?f=5&t=24932
22:12:05 <|amethyst> BTW, the wall jump ascii art should probably have a wall
22:12:06 <gammafunk> oh
22:12:09 <gammafunk> we need a WJC banner
22:12:24 <gammafunk> we still have time for that but I need to get the scripts rolling maybe
22:12:30 <gammafunk> there may be code changes for that
22:14:08 <advil> gammafunk: any thoughts on what a ghost vault should do if there are no bones? options, (i) don't place, (ii) place something else: (ii-a) place a mirror of the player (this is oddly the path of least resistance given the current code), (ii-b) place something ood to some degree
22:14:18 <advil> this is on the premise that there's a good item in there
22:14:42 <gammafunk> advil: well, part of what we had discussed was to make it so "ghost stuffing" or "ghost scumming" isn't a thing
22:14:45 <advil> yeah
22:14:55 <gammafunk> so it should basically not depend on rate of player death ideally
22:15:01 <advil> so that rules out (i) pretty much
22:15:43 <|amethyst> (ii-c) have some pre-stuffed bones files that don't get deleted (or hard code said ghosts)
22:15:44 <gammafunk> mirror of the player seems ok, although probably ideally it would be a randomized ghost
22:15:52 <gammafunk> yeah, something like that
22:16:33 <gammafunk> it'd be sort of nice if a ghost came from an actual game so that we weren't making "fake ghosts"
22:16:55 <advil> that might be harder for offline versions
22:16:56 <gammafunk> but I don't think it's too much a problem if we have to make a randomized one or some kind of mirror ghost
22:17:07 <gammafunk> true
22:17:17 <|amethyst> we can capture a few bones files from the previous version for each new version, and include that in the distribution
22:17:18 <advil> well, I guess we could just pull CAO's bones files right now or something
22:17:25 <gammafunk> heh
22:17:26 <advil> heh
22:17:35 <gammafunk> you just maybe wanna filter what names you use...
22:17:46 <|amethyst> I would change the names actually
22:17:47 <gammafunk> I think for the time being you can take the easiest path though
22:17:50 <|amethyst> or at least get permission
22:17:58 <advil> we could use dev names
22:18:07 <gammafunk> Fake Ghosts!
22:18:14 <|amethyst> advil: but, but...
22:18:31 <gammafunk> (say that as if you would say Fake News!)
22:18:31 <advil> haha
22:18:31 <|amethyst> advil:  that was part of my post-orbrun pipe dream
22:18:38 <gammafunk> seriously neil has plenty of ghosts to supply
22:18:49 <gammafunk> honestly we could probably just make every ghost used online be a neil ghost
22:18:59 <advil> wonder how hard a morgue->ghost converter would be
22:19:18 <|amethyst> advil:  it kind of exists, in two separate parts
22:19:23 <gammafunk> seems kinda a tough, but I guess we can already convert most of what we need
22:19:32 <|amethyst> advil:  you can load a morgue into your current character, and you can die and leave a ghost
22:19:41 <|amethyst> though the former is lacking quite a bit
22:20:05 <advil> ok well this all makes it easy to do ii-a as a placeholder
22:20:06 <gammafunk> updated to a new kernel as part of ubuntu updates
22:20:11 <gammafunk> running two parallel makes
22:20:23 <gammafunk> and ld gold of crawl crashes due to bus error
22:20:28 <|amethyst> hm, that's going to be my new way of saying "die in a fire"
22:20:28 <gammafunk> never seen a crash like that
22:20:35 <|amethyst> "die and leave a ghost"
22:20:47 <gammafunk> "die in a firestorm ghost"
22:20:59 <gammafunk> or perhaps die in a ghost firestorm
22:21:02 <|amethyst> gammafunk:  hm, that's the kind of thing I've seen when I have a bad bit in memory
22:21:05 <|amethyst> s/in/of/
22:21:13 <gammafunk> oh don't say thattttttt
22:21:25 <gammafunk> I'm just suspicious of this since it was a very recent kernel update
22:21:46 <gammafunk> so far no other kinds of crashes in any way
22:21:58 <gammafunk> wouldn't I see a kernel error for something like that? I suppose not
22:22:00 <|amethyst> did this happen more than once?
22:22:04 <gammafunk> no, just now
22:22:27 <|amethyst> check dmesg, there could be an error from the kernel
22:22:38 <gammafunk> appears to be none
22:22:58 <|amethyst> also, if it is just one bit of RAM, I believe it's possible to tell the kernel to just avoid that chunk of RAM
22:23:13 <|amethyst> or maybe it was a cosmic ray
22:23:31 <|amethyst> (alpha decay in the chip packaging is more likely, I hear)
22:23:48 <gammafunk> mfw a black hole merge breaks my crawl build
22:26:43 <advil> I think it's gonna turn out that that cleanup PR was more useful for getting me to understand the ghost code than for what it actually takes to implement this
22:27:04 <geekosaur> ld.gold does have bugs
22:27:06 <advil> well, some of it was useful
22:27:17 <gammafunk> the ghost demon stuff has always been a nasty area of the code base
22:27:26 <gammafunk> and I think many devs have improved it somewhat over time
22:27:35 <gammafunk> it used to be even more horrendous
22:28:00 <geekosaur> someone working on the llvmng backend in ghc HEAD was triggering one fairly regularly a month ago
22:28:24 <advil> you can sort of tell how many problems it has had from the absurd ASSERT scaffolding in load_ghosts
22:28:26 <gammafunk> hrm, don't think I had any updates to the compilation toolchain
22:28:39 <|amethyst> geekosaur:  yeah, but I don't think I've seen a bus error on Linux come from a software problem
22:28:46 <|amethyst> geekosaur:  other than perhaps a kernel bug
22:29:45 <geekosaur> memory reference to a segment not present in the process's page map iirc?
22:29:57 <|amethyst> wouldn't that just be a segfault
22:30:10 <|amethyst> I thought bus error on x86 came from a bad *physical* address
22:30:30 <geekosaur> x86 doesn;t really have bad physical addresses
22:30:42 <geekosaur> on other architectures it means an unaligned access
22:30:43 <|amethyst> I mean, a physical address without a stick of ram in it
22:31:06 <geekosaur> kernel bus errors, yes.
22:31:13 <geekosaur> that should never, ever be relevant to a user program
22:31:31 <|amethyst> I figured if it happened, the kernel would pass it along to userspace
22:31:43 <|amethyst> a flipped bit in a kernel page table for example
22:31:44 <geekosaur> userspace should never be able to trigger it
22:31:46 <|amethyst> hm
22:31:53 <geekosaur> it's a kernel panic condition
22:32:02 <gammafunk> I should have saved the crash
22:32:16 <|amethyst> ah, didn't realise it would panic in that situation, but I guess that makes sense
22:32:22 <geekosaur> means the kernel's own memory accounting is out of whack, and the system is no longer safe to operate
22:33:59 <|amethyst> hm, I know I have (though it was maybe 15 years ago) seen bus faults in userspace on x86
22:34:41 <geekosaur> linux 0.x/1.x didn't support SIGBUS at all (it was mapped to SIGUSR2 iirc). later versions map it to something but I don't recall what
22:35:15 <geekosaur> I think it has to do with completely void sections of the process page table vs. occupied segments with nothing mapped to them (segfault)
22:35:25 <geekosaur> since page tables are two levels on x86
22:36:41 <|amethyst> ah, found a SO answer that gives a few ways to get SIGBUS on Linux x86
22:36:54 <|amethyst> https://stackoverflow.com/a/39592322
22:37:02 <CanOfWorms> gammafunk: https://i.imgur.com/9dljN7B.png
22:37:20 <gammafunk> CanOfWorms: that's rad, thanks fo much
22:37:23 -!- aditya1 is now known as aditya
22:37:23 <gammafunk> *so
22:37:26 <hellmonk> z o m b i e h a n d s
22:37:33 <gammafunk> sorry, what?
22:37:35 <gammafunk> what are you talking about
22:37:45 <hellmonk> Z O M B I E H A N D S
22:38:26 <CanOfWorms> ??zombie hands
22:38:26 <Sequell> Borgnjor's Vile Clutch[1/1]: Not actually a nasty purse used by necromancers but the 'formal' name for the zombie hands player spell - L5 earth/necromancy, smite targeted constriction in a 3x3 area.
22:38:59 <geekosaur> second answer is what I've been talking about btw
22:39:40 <geekosaur> each mmapped file is a segment in a page table, if you go beyond the end of the file then you are usually outside of any segment and get SIGBUS
22:40:42 <geekosaur> the bad stack pointer thing looks like it should be the same, but stack handling is a bit screwy anyway (and there was a recent CVE due to various OSes getting the screwy parts wrong on x86)
22:40:50 <hellmonk> when the actual purple stairs implementation is three lines of code and the rest is just a matter of changing descriptions and such
22:40:57 <gammafunk> I do have these weird cpu messages that have been happening under ubuntu 17
22:41:28 <gammafunk> [23686.811273] perf: interrupt took too long (4033 > 3958), lowering kernel.perf_event_max_sample_rate to 49500
22:41:31 <gammafunk> ones like this
22:41:51 <gammafunk> hadn't seen those from dmesg before the move to 17
22:41:56 <gammafunk> at least I don't recall
22:42:05 <gammafunk> if only I understood anything about computers
22:42:48 <|amethyst> I understood things about computers back in the DOS days
22:42:53 <|amethyst> it was so much simpler then :)
22:44:58 * geekosaur knows what that is but not why ubuntu would have that enabled
22:45:20 <geekosaur> also I would call that kernel bug but not a critical one
23:48:23 -!- amalloy is now known as amalloy_