02:46:47 <Gretell> Monster database of master branch on crawl.develz.org updated to: 0.22-a0-136-g83a10e5
03:24:53 -!- amalloy_ is now known as amalloy
04:09:07 <gammafunk> advil: Finally made all the posts/announcements for 0.21.1, so let me know when you have those mac builds, and I'll upload/update
06:15:41 <Cheibriados> Autopickup turns off when a non-threatening entity (e.g. plant) turns invisible 13https://crawl.develz.org/mantis/view.php?id=11377 by Leszczynek
06:59:47 -!- amalloy is now known as amalloy_
07:34:19 -!- amalloy_ is now known as amalloy
09:29:18 <aidanh> !tell wheals 13168bc2512 seems to interact oddly with the abyss; the explore horizon isn't cleared after an abyss region change
09:29:18 <Sequell> aidanh: OK, I'll let wheals know.
10:00:16 -!- amalloy is now known as amalloy_
10:21:06 <advil> gammafunk: OS X builds for 0.21.1 are in https://www.dropbox.com/sh/mjk2xoklkhmyg2n/AAAfljuzOz4nC3-wYZ4qdZMEa?dl=0
10:28:49 <advil> !crashlog dylex
10:28:52 <Sequell> 2. dylex, XL20 GnWn, T:58455 (milestone): http://crawl.berotato.org/crawl/morgue/dylex/crash-dylex-20180203-021516.txt
10:55:18 <espais> so going back to the cygwin build thing, i've tried including <ios>, <cstdio>, <string>, even "unicode.h" and nothing seems to owrk
10:55:19 <Sequell> espais: You have 2 messages. Use !messages to read them.
10:55:36 <espais> i've also noticed that there are a few git projects for special builds of wcwidth
10:55:39 <espais> !messages
10:55:40 <Sequell> (1/2) Floodkiller said (13h 34m 27s ago): if you are interested in sso, I made this design a while ago https://crawl.develz.org/tavern/viewtopic.php?f=19&t=20673
10:55:49 <espais> !messages
10:55:49 <Sequell> (1/1) Floodkiller said (13h 34m 4s ago): might not fit a true sso, but could be useful for a migration plan if you are working on a true version
10:57:23 <advil> espais: did you try <cwchar>? that's the one that should explicitly include it, the others might transitively on some stdlib versions
10:58:00 <geekosaur> espais, ^ we were trying to figure out why the other files worked, that is not an answer to the file that is failing.
10:58:15 <espais> sure
10:58:20 <espais> yea i tried that, also doesn't work
10:58:25 <espais> would the compile log be helpful?
10:58:25 <advil> I think those suggestions came from the fact that wcwidth is used in a bunch of other files that apparently didn't lead to an error on your build
10:58:50 <espais> i can dump it to a pastebin, but yea i agree that it is bizarre
10:58:58 <geekosaur> and my analysis also tried to point out that I had tracked that on non-cygwin and the result might be different there
10:59:43 <espais> sure, you did it with msys i think?
10:59:45 <geekosaur> i.e. I said why it *does* work *on linux*. since it didn't work on cygwin, deductively things are different there
10:59:50 <advil> well, it's not bizarre in that cio.cc is early alphabetically and doesn't include much stdlib stuff
11:00:18 <advil> so somewhere between cio.cc and initfile.cc (maybe in initfile.cc) something gets included that pulls it in, I suspect
11:00:18 <espais> advil: that's kind of what i was wondering tbh, but i didn't dig into any of the pre-cio files to see if they made any calls
11:00:26 <geekosaur> (I have linux, and I have experience building on os x. I do not have experience with either msys2 or cygwin asa build environment)
11:00:42 <advil> I've built with msys2
11:01:16 <advil> the next ones after cio.cc that use wcwidth are fontwrapper.cc and initfile.cc, the font one is only relevant if you're building tiles
11:01:36 <espais> so the thing is though, if i'm including the 'correct' headers it should be compiling anyway i'd think, but i can not get beyond cio
11:01:36 <espais> ...i am building tiles
11:01:38 <espais> make TILES=y is my make call
11:02:38 <geekosaur> welp
11:02:53 <espais> unless if i'm missing something...i have been out of the c++ loop for a while
11:02:57 * geekosaur just consulted Dr. Google. consensus is that wcwidth() was there, is broken, is now hidden
11:03:22 <geekosaur> cygwin build will have to force use of the one crawl provides for Windows
11:03:26 <espais> i did come across this when i was poking around  yesterday: https://github.com/rinrin-/cygwin-alternative/blob/master/newlib/libc/string/wcwidth.c
11:05:01 <geekosaur> we *have* a wcwidth.
11:05:08 <geekosaur> we just have to force its use on cygwin
11:05:12 <espais> i see
11:06:53 <geekosaur> it's nott even in the contribs, so it should be relatively easy to make unicode.h and wcwidth.cc do the right thing on cygwin
11:06:57 <geekosaur> provided we have a way to reliably identify a build on cygwin
11:07:58 <espais> i'll give this a try, but according to Papa SO this should work: #if defined(_WIN32) || defined(__CYGWIN__)     // Windows (x86 or x64)     // ...
11:08:04 <advil> hm
11:08:06 <espais> for mingw
11:08:18 <advil> it might be simpler than all this...try #include "unicode.h"
11:08:40 <advil> oh, that's already there
11:08:58 <advil> oh maybe I see
11:09:31 <espais> did do that, doesn't work either
11:12:11 <advil> ok, try changing line 47 of unicode.h to "#if !defined(UNIX) || defined(__CYGWIN__)"
11:12:19 <advil> it's currently #ifndef UNIX
11:13:28 <advil> actually in crawl you can just use TARGET_OS_CYGWIN
11:15:02 <espais> heh, i had just gotten to that point in unicode.h
11:15:42 <espais> so with your first statement it gets beyond cio
11:15:51 <espais> trying with the TARGET define now
11:16:07 <advil> I'm still confused about why MSYS2 works though, from googling it looks like it should have a superset of the cygwin defines
11:16:30 <espais> that process i'm not really familiar with
11:17:06 <advil> although, it could be as simple as, msys2 actually provides a wcwidth
11:17:06 <espais> so in terms of best practices, should i be checking !defined(UNIX) || defined(TARGET_OS_CYGWIN)
11:17:29 <advil> yeah
11:17:45 <advil> both of those are more general defines that are internal to crawl, and that cygwin define checks multiple cases
11:17:52 <espais> fair enough
11:17:58 <espais> ok, time to sit through a few build processes
11:18:21 <espais> unfortunately i don't have a way to check if this breaks other builds atm, should i just create a pull req?
11:18:41 <advil> yeah, someone will need to test it in msys2 probably
11:18:54 <advil> (TARGET_OS_CYGWIN is #if defined (__CYGWIN__) || defined (__CYGWIN32__))
11:19:05 <espais> gotcha
11:19:22 <espais> i think i'm going to try to get more into bug fixing for you guys at least at a tertiary level to get back into c++
11:19:29 <espais> i've been pure python since about 2012
11:19:40 <espais> digging through the code is like....blehhhhhhhhhh....how does this all work again
11:20:04 <espais> (i used to develop for embedded ansi-c in my prior life)
11:20:55 <geekosaur> msys2 actually makes an effort to act like linux, so I would expect it to have a proper wcwidth in the same way linux does
11:21:25 <espais> heh, and see i thought that was the purpose of cygwin
11:21:26 <espais> TIL
11:21:48 <espais> just to see it again
11:21:52 <espais> !flip compatibility
11:21:52 <Sequell> (╯°□°)╯︵ʎʇᴉlᴉqᴉʇɐdɯ°ɔ
11:21:59 <advil> looks like mingw provides it, so msys2 should inherit that
11:22:13 <espais> gotten to god-abil.o so far
11:22:17 <espais> looks promising
11:22:46 <geekosaur> cygwin tries to chart a course halfway between windows and unix. msys2 aims specifically for linux portability
11:22:49 <espais> ahh i see
11:23:01 <espais> see i've never really used cyg for anything other than an SSH terminal until now
11:23:15 <advil> yeah I'm not super conversant with windows but these days it seems like msys2 is the most robust compilation environment for crawl. (But some aspects of it are *really* slow for unclear reasons, e.g. linking).
11:24:25 <Rotatell> Beargit (L22 DrNe) ASSERT(!invalid_monster(&mons)) in 'mon-death.cc' at line 1910 failed. (Depths:5)
11:24:29 <advil> it's also the one that gets tested the most by devs I think
11:24:39 <advil> there's also the new WSL thing, I think maybe wheals uses that or has used that?
11:24:44 <advil> !crashlog
11:24:50 <espais> is that the ubuntu on windows thing?
11:24:51 <Sequell> 18764. Beargit, XL22 DrNe, T:54839 (milestone): http://crawl.berotato.org/crawl/morgue/Beargit/crash-Beargit-20180203-162416.txt
11:25:07 <geekosaur> yes
11:25:11 <espais> i've used it too
11:25:13 <espais> i hate it
11:25:18 <geekosaur> it's still a bit incomplete
11:25:23 <espais> let me rephrase, i loved it while it worked
11:25:26 <espais> but
11:25:32 <geekosaur> and, I think sdl will never work right in it
11:25:47 <geekosaur> it emulates linux, it does not emulate x11
11:25:49 <espais> no x windows + terrible crashes + the fact that you need to constantly pull the windows insider builds made it annoying as hell
11:25:58 <advil> huh, I thought we fixed that exact constriction bug
11:26:10 <advil> maybe I just stared at it for a while and that's what I'm remembering
11:26:12 <espais> honestly the killer for me was needing to sign up for preview builds of windows ... had some really bad crashes that just wasn't worth it
11:26:12 <geekosaur> you may have. rotatell.
11:26:35 <geekosaur> iirc there's no updates on cbro because out of space
11:26:48 <advil> !git
11:26:49 <Sequell> %git
11:26:49 <Cheibriados> 07|amethyst02 * 0.22-a0-136-g83a10e5: Fix grammar in Gnoll manual entry (Ge0ff) 10(2 days ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/83a10e5898da
11:26:53 <geekosaur> johnstein was talking abotu wanting to force everyone to always on latest
11:27:06 <advil> it looks current
11:30:28 <espais> coffee break time, i'll be back in a bit to report on hopeful build success
11:35:39 -!- amalloy_ is now known as amalloy
11:49:40 <espais> new errors!
11:49:57 <espais> libunix.cc:30:18: fatal error: term.h: No such file or directory
11:51:39 <espais> looks like the make file needs to be updated too, will check out that compiler strap
11:54:32 <geekosaur> term.h is ncurses
11:54:54 <geekosaur> ...depending on how cygwin chose to configure ncurses, you may need a terminfo-dev package of some kind
12:08:32 -!- MarvinPA__ is now known as MarvinPA
12:12:23 <espais> so thiiiis is interesting
12:12:33 <espais> somehow i was missing the libncurses-devel package
12:12:51 <espais> compiling again after installing that gave me ambiguous entries for wcwidth
12:13:04 <advil> aha
12:13:09 <espais> so...removing that cygwin target...with libncurses installed....seems to be working
12:13:12 <espais> halfway through the build
12:13:15 <espais> christ
12:13:22 <espais> !flip espais
12:13:22 <Sequell> (╯°□°)╯︵sᴉɐdsǝ
12:14:45 <espais> i wonder if i tried installing the wrong libncurses package before, because i could have sworn i did
12:15:43 <espais> !tell FloodKiller I had similar thoughts, it's just the overall implementation of dealing with user response seems prohibitive
12:15:43 <Sequell> espais: OK, I'll let floodkiller know.
12:27:15 <espais> SIGH
12:27:15 <espais> make without tiles went flawlessly
12:27:17 <espais> building tiles now
12:27:29 <espais> so, user error, not dev error
12:27:44 <espais> !flip never doubt the devs
12:27:44 <Sequell> (╯°□°)╯︵sʌǝp ǝɥʇ ʇqn°p ɹǝʌǝu
12:34:29 <advil> haha
12:34:49 <advil> well like I said, no one has been building with plain cygwin lately so issues wouldn't be all that surprising
12:35:07 <advil> our official builds are cross-compiled, too, so they don't really test windows building at all
12:35:21 <espais> gotcha
12:35:35 <espais> so here's a bit of an open q that i had
12:35:43 <espais> i'm not really familiar with lua but python i'm much more familiar with
12:36:01 <espais> is there any way to do some kind of bot scripting (like qw) but rather using python to control it instead of lua?
12:36:20 <advil> probably not
12:36:25 <advil> crawl embeds a lua interpreter
12:36:32 <advil> so that's why lua works
12:36:33 <espais> probably would involve hijacking keyboard commands or something messy
12:36:55 <advil> well in principle you could do bots with anything if you build a way of interfacing
12:37:12 <advil> but lua has a huge head start because there's an API
12:37:18 <espais> yea
12:37:41 <espais> there's just so much i could do in python that you can't in lua as far as i know
12:38:05 <espais> i think i'd seen somebody interested (on the subreddit) that was going to use opencv
12:39:41 <geekosaur> there's a bot that uses the webtiles json interface (although it's written in haskell)
12:40:21 <espais> a json interface you say....
12:40:29 <espais> can you point me towards that bot?
12:40:40 <geekosaur> ??rw
12:40:40 <Sequell> rw[1/2]: <clouded_> rw is like watching a baby grow
12:40:43 <geekosaur> ??rw[2
12:40:43 <Sequell> rw[2/2]: https://github.com/rwbarton/rw
12:40:58 <espais> that is the best readme file ever
12:41:10 <geekosaur> although reverse engineering it from haskell might not be the preferred way to go
12:41:49 <geekosaur> the problem is, I don;t think there's any real documentation for the webtiles json interface
12:42:13 <geekosaur> there is rw and there is the webtiles js client
12:42:13 <espais> that's why i was interested in that interface
12:42:13 <geekosaur> but code kinda makes terrible documentation >.>
12:45:10 <advil> if you are running the bot locally, I'm sure you could modify the way lua works in crawl so that it can get outside whatever sandboxing we do; there appear to be lua-python bridges
12:55:41 <espais> that's true, though i was kind of playing with the idea of having a bot interact with webtiles
12:56:01 <espais> i thought it'd be fun to spin up a little botnet of 20 or so instances all playing concurrently
12:56:09 <espais> so if there's a json interface i could pretty easily hook that into python
13:18:17 <johnstein> cbro should be current, but yes, I want to do a big force update so I can delete as many past versions of trunk that I can
13:39:37 -!- Bammboobies is now known as Bammboo
13:59:08 -!- ProzacElf_ is now known as ProzacElf
15:18:52 -!- amalloy is now known as amalloy_
15:41:42 -!- Bammboobies is now known as Bammboo
16:14:28 -!- ProzacElf_ is now known as ProzacElf
17:40:27 -!- Bammboobies is now known as Bammboo
18:06:46 <hellmonk> johnstein: any chance of getting hellcrawl back up this weekend?
18:08:18 <johnstein> that's my plan. hoping tonight.
18:09:14 <hellmonk> cool
18:18:02 <rumflump> I tried playing a titan last week; I think I like regular demigods better
18:18:40 <rumflump> no gods and no armor :O
18:55:37 <wheals> !tell espais no idea what you mean by "sign up for preview builds of windows," which means i either did this without knowing, or it's no longer a requirement; for x windows, i download and run Xming, but i admit that is a pain
18:55:37 <Sequell> wheals: You have 1 message. Use !messages to read it.
18:55:38 <Sequell> wheals: OK, I'll let espais know.
18:55:44 <wheals> !messages
18:55:45 <Sequell> (1/1) aidanh said (9h 26m 26s ago): 13168bc2512 seems to interact oddly with the abyss; the explore horizon isn't cleared after an abyss region change
18:56:02 <wheals> ah, i tested it on abyss and it was ok, but i didn't wait around long enough to see a region change
18:56:16 <wheals> so i'll add that to the list of things to fix with it
19:14:44 <aidanh> Labyrinth has issues as well; forgetting parts of the map should probably clear the horizon for those parts
19:17:11 <aidanh> the passive mapping mut interacts strangely too; I think it'd work better if is_explore_horizon only counts neighbour cells that have been seen
19:34:36 <wheals> right, that was another thing
19:34:43 <wheals> but labyrinth worked fine in my testing
19:35:30 <wheals> once a square was forgotten, it no longer counted for having explore horizon around it
20:13:57 <aidanh> wheals: even without a region change there's problems if you keep walking in one direction
20:15:18 <aidanh> i.e. the way the abyss warps you and your local chunk back to the map centre as you near the edge (I'm guessing)
20:18:35 <gammafunk> advil: downloads page updated with OS X builds, thanks
20:19:56 <wheals> hmm
20:20:07 <wheals> maybe it just needs to be disabled there, though that is a little unsatisfying
20:21:53 <aidanh> I think that'd be an acceptable solution; afaik it's not really useful there
20:22:56 <aidanh> wrt only considering seen squares in is_explore_horizon(), I think it looks a lot better: https://a.pomf.space/sboarrhtuhhu.png
20:24:43 <aidanh> I checked lab again and it's not as bad as I thought
20:25:47 <aidanh> but perhaps a better method in lab/abyss would be to floodfill from the player's current position, and then find the border of that?
20:27:25 <wheals> maybe
21:27:01 -!- Vaevictu1 is now known as Vaevictus
23:05:47 -!- amalloy_ is now known as amalloy
23:07:10 -!- eb__ is now known as eb
23:39:00 <gammafunk> !seen ZiBuDo
23:39:00 <Sequell> I last saw ZiBuDo at Sat Feb  3 18:37:40 2018 UTC (10h 1m 19s ago) joining the channel.
23:39:23 <gammafunk> ZiBuDo: you around? I sent you an email about how to fix CJR's trunk logfile/milestones files. snark did have the backups we need
23:39:56 <ZiBuDo> oh
23:40:30 <ZiBuDo> i dont know why never saw the email
23:40:34 <ZiBuDo> ill try to do it tomorrow
23:40:35 <ZiBuDo> my bad
23:40:41 <gammafunk> oh I literally just sent it
23:40:57 <gammafunk> but saw you in channel so wanted to mention since wasn't sure if you got
23:41:04 <gammafunk> but cool, let me know if there are any issues
23:44:18 <gammafunk> well well well
23:44:22 <gammafunk> !lg *map~~its_a_trap
23:44:23 <Sequell> Broken query near '~~its_a_trap'
23:44:25 <gammafunk> !lg * map~~its_a_trap
23:44:26 <Sequell> 103. Lasty the Severer (L16 MuGl of Nemelex Xobeh), demolished by a swamp worm (called by a spriggan druid) on Swamp:1 (gammafunk_its_a_trap) on 2018-02-04 02:10:38, with 108073 points after 40789 turns and 1:06:48.
23:45:18 <gammafunk> Lasty: your irc client seems to be acting up a lot, doing lots of joins and quits