01:57:08 <Gretell> Windows builds of master branch on crawl.develz.org updated to: 0.21-a0-50-g941c862
02:52:25 <Gretell> Monster database of master branch on crawl.develz.org updated to: 0.21-a0-50-g941c862
03:12:27 <Rotatell> Unstable branch on crawl.beRotato.org updated to: 0.21-a0-50-g941c862 (34)
03:35:23 -!- amalloy is now known as amalloy_
03:36:58 -!- amalloy_ is now known as amalloy
04:45:38 -!- amalloy is now known as amalloy_
05:28:31 <ProzacElf> !tell gammafunk if you enter a disconnected part of a level and finish autoexploring that part, you get a message about an unvisited transporter even if you haven't seen the portion of the level it's on.  not sure if that's information that's supposed to be leaked or not
05:28:32 <Sequell> ProzacElf: OK, I'll let gammafunk know.
05:31:31 <gammafunk> oh, yeah, someone said there was a bug like this
05:31:31 <Sequell> gammafunk: You have 1 message. Use !messages to read it.
05:31:56 <gammafunk> perhaps I'm missing a "seen this square" check
05:31:59 <gammafunk> thanks, I'll look into that
07:24:36 <Pinkbeast> Not to ask a hardy perennial, but do we know what's going on with CDO getting 0.20 ?
07:35:15 <gammafunk> yes, the server admin was busy before the release and not able to get it in time for tournament
07:35:37 <gammafunk> he'll be able to get 0.20 installed on CDO soon, but after the tourney
07:35:47 <Pinkbeast> Oh, brilliant, thanks.
07:36:35 <Pinkbeast> I'll just have to get as much mileage out of Confusing Touch as I can over the next few days
10:01:30 <advil> hmm, can't log in to cbro ssh, is it just me?
10:01:54 <gammafunk> advil: I can just now
10:02:11 <gammafunk> authenticated in dgl as well
10:02:21 <advil> webtiles is ok for me, ssh just hangs
10:03:46 <gammafunk> yeah ssh is fine to me for cbro
10:03:52 <nikheizen> works on my machine
10:07:42 <advil> hm it's just this one computer, my laptop works
10:08:24 <Pinkbeast> advil: Pastebinning the output of ssh -v might be informative
10:09:29 <advil> well, it works now
11:02:36 <Cheibriados> New branch created: pull/548 (1 commit) 13https://github.com/crawl/crawl/pull/548
11:02:36 <Cheibriados> 03Elan Morin Tedronai02 07https://github.com/crawl/crawl/pull/548 * 0.21-a0-51-ga92a685: Transporter vault: Trog's Sanctum. 10(36 minutes ago, 1 file, 84+ 0-) 13https://github.com/crawl/crawl/commit/a92a68589ba8
11:07:47 <Cheibriados> New branch created: pull/549 (2 commits) 13https://github.com/crawl/crawl/pull/549
11:07:47 <Cheibriados> 03Elan Morin Tedronai02 07https://github.com/crawl/crawl/pull/549 * 0.21-a0-50-gb7b5ded: Altar overflow vault: Okawaru's Leap of Faith. 10(2 days ago, 1 file, 27+ 0-) 13https://github.com/crawl/crawl/commit/b7b5ded21542
11:07:47 <Cheibriados> 03Elan Morin Tedronai02 07https://github.com/crawl/crawl/pull/549 * 0.21-a0-51-g4cce568: okawaru_leap_of_faith adjustments. 10(6 hours ago, 1 file, 10+ 10-) 13https://github.com/crawl/crawl/commit/4cce568fad7a
12:34:46 -!- amalloy_ is now known as amalloy
13:04:18 <Henzell> Unstable branch on crawl.akrasiac.org updated to: 0.21-a0-50-g941c862 (34)
15:23:56 <Warrigal> Hey, I see that tile.cc seems to be declaring an array whose length is a variable. Line 461 says "png_bytep row_pointers[h];
15:23:56 <Warrigal> "
15:24:26 <Warrigal> ...where h is a png_uint_32 declared on line 416 and (presumably) populated on line 418.
15:24:51 <Warrigal> That's not standard C++, is it? I'm trying to build this in Visual Studio and it's erroring.
15:28:33 <Pinkbeast> It's a G++ extension because legal in C
15:28:59 -!- Warrigal is now known as t_swett
15:30:26 <geekosaur> then again visual c is infamous for implementingh only half of a standard and then doing the rest years later
15:56:27 <t_swett> I see there's this log of someone known as svendre working through all the same issues I'm looking at: https://s-z.org/crawl-dev/%23%23crawl-dev-20160429.lg
15:57:02 * bh pokes Sequell
16:05:43 <elliptic> !tstats
16:06:05 <Sequell> Stats after 16 days (t): 2997 players, 1062 runers, 543 winners, 1685 wins, 73356 games, winrate 2.30%, total player time 4y+187d+2:02:23.
16:06:08 <elliptic> !tstats t0.19
16:06:13 <Sequell> Stats after t0.19 days (t): 0 players, 0 runers, 0 winners, 0 wins, 0 games, winrate NaN, total player time 0.
16:06:18 <elliptic> !tstats 16 t0.19
16:07:49 <Sequell> 90s limit exceeded: killed !tstats 16 t0.19
16:07:52 <elliptic> come on
16:07:53 <elliptic> !tstats 16 t0.19
16:09:01 <Sequell> Stats after 16 days (t0.19): 3159 players, 1014 runers, 528 winners, 1423 wins, 79391 games, winrate 1.79%, total player time 4y+129d+8:36:16.
16:10:09 <Medar> elliptic: can you push the tournament reminder revert?
16:10:18 <elliptic> oh, sure
16:10:25 <elliptic> thanks for reminding me
16:12:08 <Cheibriados> 03elliptic02 07* 0.21-a0-51-gea492e9: Revert "0.20 tournament reminder." 10(54 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/ea492e930d02
16:12:28 <Medar> I can run the rebuilds
16:12:30 <Medar> ??rebuild
16:12:31 <Sequell> rebuild[1/2]: http://crawl.akrasiac.org/rebuild/ http://underhound.eu:81/rebuild/ http://crawl.berotato.org/crawl/rebuild/ http://crawl.xtahua.com/rebuild/ https://crawl.jorgrun.rocks/rebuild/ Bug |amethyst or Nap.Kin for CDO. Use your powers wisely.
16:13:51 <advil> t_swett: imo what |amethyst said in that log is right, it would be much better as `vector<png_bytep> row_pointers(h)`.  I guess no one has done that yet.
16:14:55 <t_swett> There's a lot about C++ I don't know. Does a vector automatically delete its storage when it goes out of scope or something?
16:15:11 <Medar> oh, I don't have admin on CJR so someone else had to do that one
16:15:51 <Lantell> Unstable branch on underhound.eu updated to: 0.21-a0-51-gea492e9 (34)
16:16:21 <Henzell> Unstable branch on crawl.akrasiac.org updated to: 0.21-a0-51-gea492e9 (34)
16:16:37 <Pinkbeast> t_swett: It may be more down to meaning a runtime-defined length is fine
16:17:04 <advil> t_swett: they both should do that, but (at least relative to my poor c++ knowledge) <vector> is a much better way to do length at runtime
16:17:26 <advil> *arbitrary length
16:17:39 <Eksell> Unstable branch on CRAWL.XTAHUA.COM updated to: 0.21-a0-51-gea492e9 (34)
16:17:41 <t_swett> *nod*
16:17:43 <advil> it would probably work in msvc too
16:17:56 <t_swett> I'm used to C and C#.
16:18:22 <t_swett> I know how totally manual memory management works. I also know how garbage collection works. I don't know about the stuff in between.
16:19:57 <amalloy> t_swett: most C++ memory management is done through destructors. when a vector<int>'s lifetime ends (eg because it has gone out of scope), its destructor is called, which is responsible for cleaning up any memory it allocated
16:20:37 <advil> yeah, explicit memory management is fairly rare in crawlcode, it mostly puts things on the stack
16:21:03 <Medar> !tell TZer0 FYI rebuilding CUE doesn't show messages in real time, because it's using gzip content-encoding. Doesn't really matter in practice I guess.
16:21:04 <Sequell> Medar: OK, I'll let tzer0 know.
16:21:10 <amalloy> whether it actually allocated any memory or not is not actually of interest to us, because the destructor promises to deallocate any memory, and we know the destructor will be called when we're done with it
16:21:59 <advil> here's a pithy SO answer summarizing: https://stackoverflow.com/questions/76796/general-guidelines-to-avoid-memory-leaks-in-c/78019#78019
16:23:37 <advil> hm, I don't have CJR admin either
16:25:53 <Medar> ??cjr
16:25:54 <Sequell> cjr[1/2]: https://crawl.jorgrun.rocks:8081/ Hosted in Montreal by Zibudo (admin@jorgrun.rocks) and Gramm. SSH login info: https://crawl.jorgrun.rocks/console.html.
16:27:01 <Rotatell> Unstable branch on crawl.beRotato.org updated to: 0.21-a0-51-gea492e9 (34)
16:35:55 <t_swett> . o ( why am I trying to get Crawl to work, when I could be doing my actual job instead--it's literally the same thing )
16:47:55 -!- amalloy is now known as amalloy_
17:09:17 <Cheibriados> Letter adjustment mis-reports c, i 13https://crawl.develz.org/mantis/view.php?id=11097 by Mattlistener
17:09:50 <Jorgrell> Unstable branch on crawl.jorgrun.rocks updated to: 0.21-a0-51-gea492e9 (34)
17:23:22 <elliptic> gammafunk: I killed the tourney script, in case you are wondering why it is off
18:11:32 -!- amalloy_ is now known as amalloy
18:35:15 <gammafunk> elliptic: thanks
19:23:52 -!- amalloy is now known as amalloy_
20:28:27 -!- amalloy_ is now known as amalloy
23:23:34 <Lightli> ??plan
23:23:35 <Sequell> plan[1/1]: https://crawl.develz.org/wiki/doku.php?id=dcss:planning:release_plans
23:25:30 <gammafunk> wow, so soon Lightli?
23:26:07 <gammafunk> we might have to get to some of those important items under "General todo"
23:27:51 <gammafunk> also, rip steam
23:28:08 <nikheizen> wasn't chequers going to do that
23:28:37 <gammafunk> well stream greenlight is gone as of June 6
23:28:54 <gammafunk> now you have to pay to get listed
23:29:08 <hellmonk> I have a suggestion for 0.21, change the airstrike damage formula to something that makes sense so that we can quit mocking it in irc and move on to mocking some other formula instead
23:29:26 <nikheizen> oh right yeah
23:29:33 <nikheizen> (finally tbh)
23:29:34 <gammafunk> at this point, hellmonk, we're keeping the formula specifically to troll hellcrawl dev
23:29:45 <hellmonk> lol
23:30:06 <gammafunk> For steam there would have to be some kind of legit organization that had valid tax info even if the money issue were addressed
23:32:13 <gammafunk> hellmonk: grats on running a great clan btw, you guys really had my teammates gunning for it close to the end
23:32:33 <nikheizen> gammafunk the puppet master
23:33:02 <hellmonk> ty, in the end we could not overcome the demise + manman + ultraviolent power combo
23:33:09 <gammafunk> no it was more UV4 coordinating points with help from ManMan/Demise, and those three earning a lot
23:33:11 <hellmonk> but it was fun
23:33:38 <WalkerBoh> i'm pissed because team awbw had our best tourney ever
23:33:40 <WalkerBoh> and ended up 5th
23:33:46 <WalkerBoh> it was nuts at the top this year
23:33:53 <gammafunk> I did my share and more than the designated "meme slot" would be expected (esp given we had a teammate who found they couldn't really play much after t started)
23:33:53 <WalkerBoh> i thought 0.20 was supposed to be harder
23:34:10 <hellmonk> yeah did you guys ever figure out why winrate was up
23:34:26 <gammafunk> but I only set out to get 4 wins; I do wish I'd gone for 5 but I did a lot of tournament twitch streams on top of all the other stuff
23:34:33 <gammafunk> we did not really
23:34:36 <gammafunk> only a few leads
23:34:46 <hellmonk> I cant think of any earlygame change that should have a ~25% increase on winrate or w/e it was
23:35:07 <gammafunk> but it's deffo up about 25%, even a bit higher if you look at non-goodplayers defined respective to 0.19 and 0.20
23:35:08 <nikheizen> WalkerBoh: sorry for stealing your mfcj nchoice points
23:35:12 <hellmonk> was the tournament lairratio up as well
23:35:17 <WalkerBoh> haha nikheizen, you bastard!
23:35:21 <gammafunk> winrate up like 35% for them too
23:35:24 <WalkerBoh> i came home from work and was very upset with you!
23:35:30 <nikheizen> :3c
23:35:34 <gammafunk> well elliptic found that the winrate increase was concentrated I think between xl9 and xl18?
23:36:07 <WalkerBoh> how did he determine that?
23:36:09 <hellmonk> weird
23:36:23 <WalkerBoh> looking at death ratios for different milestones?
23:36:55 <nikheizen> i guess that's when you have the most consumables but before you can become diesel?
23:36:59 <gammafunk> well it's a pretty simple query I think, although perhaps he didn't do it this way
23:37:03 <nikheizen> and consumables are *really* good now
23:37:19 <gammafunk> !lg !bot t / xl>=9
23:37:26 <Sequell> 20989/78336 games for bot (t): N=20989/78336 (26.79%)
23:37:32 <gammafunk> !lg !bot t0.19 / xl>=9
23:37:49 <Sequell> 21485/85406 games for bot (t0.19): N=21485/85406 (25.16%)
23:37:57 <gammafunk> yeah note how those are pretty similar
23:38:27 <gammafunk> !lg !bot t xl>=9 / xl>=18
23:38:31 <Sequell> 3850/20989 games for bot (t xl>=9): N=3850/20989 (18.34%)
23:38:33 <gammafunk> !lg !bot t0.19 xl>=9 / xl>=18
23:39:16 <Sequell> 3358/21485 games for bot (t0.19 xl>=9): N=3358/21485 (15.63%)
23:39:29 <gammafunk> !calc 18.34 / 15.63
23:39:30 <Sequell> 1.17
23:39:43 <gammafunk> !lg !bot t xl>=18 / won
23:39:43 <nikheizen> that's % of those which reached xl 9 which also reached xl 18?
23:39:45 <Sequell> 1684/3850 games for bot (t xl>=18): N=1684/3850 (43.74%)
23:39:49 <gammafunk> yes
23:39:59 <gammafunk> !lg !bot t0.19 xl>=18 / won
23:40:02 <Sequell> 1420/3358 games for bot (t0.19 xl>=18): N=1420/3358 (42.29%)
23:40:17 <gammafunk> but see how those percentages are not as different
23:40:32 <gammafunk> so it does seem to be a bit more concentrated in "mid game"
23:40:41 <gammafunk> maybe elliptic had a more sophisticated query than what I did there though
23:41:14 <WalkerBoh> hm
23:41:14 <gammafunk> the global difference is like 25% so even the xl9-xl18 thing I did doesn't seem to capture it all really
23:41:33 <hellmonk> not really sure what would cause that, did you guys accidentally re-add all the removed lair exp and then add even more lair exp or something
23:41:35 <WalkerBoh> [20:39:27]	WalkerBoh 	!lg !bot tall recentish xl>=9 s=cv / xl>=18 o=cv
23:41:36 <WalkerBoh> [20:39:41]	Sequell	13591/78174 games for bot (tall recentish xl>=9): 3738/20257x 0.20 [18.45%], 112/732x 0.20-a [15.30%], 3321/21279x 0.19 [15.61%], 37/206x 0.19-a [17.96%], 3197/17964x 0.18 [17.80%], 3186/17736x 0.17 [17.96%]
23:41:53 <WalkerBoh> seems like 0.19 is the outlier in that regard
23:42:01 <gammafunk> you have to be careful there WalkerBoh
23:42:10 <gammafunk> you need to include -a versions in many of those tournaments
23:42:19 <gammafunk> since some servers were running those during tourneys
23:42:23 <gammafunk> and which server did changed
23:42:32 <WalkerBoh> what does -a do?
23:42:44 <gammafunk> -a means 0.19-a 0.20-a etc
23:42:46 <gammafunk> like this
23:42:47 <WalkerBoh> oh i see
23:42:53 <gammafunk> !lg * t s=src,cv
23:42:55 <Sequell> 80365 games for * (t): 23417x cao (21333x 0.20, 2084x 0.20-a), 22822x cbro (22822x 0.20), 10105x cxc (10105x 0.20), 9958x cue (9958x 0.20), 7246x cjr (7246x 0.20), 3268x cwz (3268x 0.20), 2840x cpo (2754x 0.20, 86x 0.20-a), 709x lld (376x 0.20-a, 333x 0.20)
23:42:57 <nikheizen> !lg . won 1 x=cv
23:42:58 <Sequell> 1/41. [cv=0.16] nikheizen the Conqueror (L27 OgHu of Fedhas), escaped with the Orb and 4 runes on 2015-03-15 01:15:36, with 2396766 points after 69254 turns and 13:43:41.
23:43:32 <gammafunk> like cwz and lld see a lot more deaths due to how players on those servers tend to play (going for 15 runes a lot being one factor)
23:43:48 <gammafunk> we often due something like s=regexp_replace(cv, "-a", "")
23:44:03 <nikheizen> 0.16 was meleebug right
23:44:08 <gammafunk> yes, and there's that
23:44:17 <gammafunk> so 0.16 will look weird
23:44:23 <nikheizen> doesn't get counted in recentish at least
23:44:28 <elliptic> doesn't look like -a games make a big difference at this point
23:44:35 <WalkerBoh> that's what i'm thinking too
23:44:52 <gammafunk> that's cool, just something to be aware of when looking at these queries
23:45:10 <WalkerBoh> yeah thanks, i wasn't aware of that
23:45:27 <gammafunk> oh, and if you look at games outside of the tournament, be careful since cjr has a ton of duplicated games :(
23:45:33 <gammafunk> !lg . won s=gid
23:45:34 <Sequell> 80 games for gammafunk (won): 2x gammafunk:cjr:20160919060927S, 2x gammafunk:cjr:20160808042537S, 2x gammafunk:cjr:20160702012553S, 2x gammafunk:cjr:20160616004628S, 2x gammafunk:cjr:20160605010515S, 2x gammafunk:cjr:20160427010803S, gammafunk:cszo:20150107233802S, gammafunk:cbro:20160327043629S, gammafunk:cbro:20160401023904S, gammafunk:cbro:20160408191516S, gammafunk:cbro:20160907002915S, gammaf...
23:45:39 <WalkerBoh> i noticed that on cao too
23:45:41 <elliptic> WalkerBoh: I was just comparing 0.19 and 0.20 tourneys when I decided that the bulk of the difference was between XL 9 and XL 18
23:45:51 <WalkerBoh> gammafunk: oh never mind, maybe my duplicate games are on cjr
23:45:59 <gammafunk> yeah I'm not aware of duplication from cao
23:46:05 <elliptic> 0.18 not entirely relevant if we are just looking at differences between 0.19 and 0.20
23:46:25 <WalkerBoh> elliptic: i agree, i was just trying to get a sense of the normal variation
23:46:38 <gammafunk> !lg !bot t xl>=9 / xl>=13
23:46:42 <Sequell> 8642/20989 games for bot (t xl>=9): N=8642/20989 (41.17%)
23:46:42 <gammafunk> !lg !bot t0.19 xl>=9 / xl>=13
23:46:51 <WalkerBoh> elliptic: because i'm not aware of any major shifts in win rate outside of 0.16
23:46:55 <Sequell> 8141/21485 games for bot (t0.19 xl>=9): N=8141/21485 (37.89%)
23:47:13 <gammafunk> !lg !bot t xl>=13 / xl>=18
23:47:17 <gammafunk> !lg !bot t0.19 xl>=13 / xl>=18
23:47:19 <Sequell> 3850/8642 games for bot (t xl>=13): N=3850/8642 (44.55%)
23:47:26 <Sequell> 3358/8141 games for bot (t0.19 xl>=13): N=3358/8141 (41.25%)
23:47:58 <elliptic> btw should probably put !boring on all these queries
23:48:08 <gammafunk> yeah
23:48:11 <elliptic> doesn't make a big difference but we traditionally do that in case Sebi played
23:48:41 <WalkerBoh> is it possible that the mix of players in this tournament is just more skilled?
23:48:50 <WalkerBoh> like, % of games played by "vets" is higher?
23:48:52 <gammafunk> well we did some queries for non-goodplayers
23:48:53 <elliptic> WalkerBoh: that would be a pretty large shift
23:49:10 <gammafunk> based defining those before respective tourneys
23:49:12 <WalkerBoh> agreed
23:49:29 <gammafunk> !lg !bot !boring !@goodplayerspre20 / won
23:49:33 <gammafunk> er
23:49:36 <gammafunk> !lg !bot !boring !@goodplayerspre20 t / won
23:49:42 <WalkerBoh> in fairness, there were a similar number of players winning games this tournament compared to 0.19 - 529 vs. 544
23:49:45 <gammafunk> !lg !bot !boring !@goodplayerspre19 t0.19 / won
23:49:55 <WalkerBoh> but the # of won games shot from 1428 to 1710
23:50:01 <WalkerBoh> so wins / player went way up
23:50:05 <WalkerBoh> but # of players winning did not
23:50:07 <gammafunk> and the winrate difference is even higher for those groups between 0.20 and 0.19
23:50:35 <gammafunk> however I think we say that 0.20-a didn't really show a winrate shift until possibly quite late in the dev cycle
23:50:55 <gammafunk> and the shift didn't seem too dramatic, so it's possible that it was simply "more competitiveness"
23:50:59 <WalkerBoh> (when all the top guys starting training for the tourney?)
23:51:00 <Sequell> 90s limit exceeded: killed !lg !bot !boring !@goodplayerspre20 / won
23:51:05 <gammafunk> sequell no
23:51:14 <gammafunk> oh that was the bad query anyhow
23:51:26 <Sequell> 570/59088 games for bot (!boring !@goodplayerspre20 t): N=570/59088 (0.96%)
23:52:08 <Sequell> 441/62953 games for bot (!boring !@goodplayerspre19 t0.19): N=441/62953 (0.70%)
23:52:22 <elliptic> WalkerBoh: I guess I'm just skeptical of that sort of explanation because nothing so drastic has been visible in any past tourney
23:52:41 <elliptic> and yeah, gammafunk's query shows the winrate effect is still there for players who had fewer than 10 wins pre-tourney
23:52:51 <WalkerBoh> yeah, that's a fairly compelling stat
23:52:55 <WalkerBoh> goodplayer is 10 wins?
23:52:59 <gammafunk> yeah
23:53:18 <gammafunk> you could look at smaller numbers of pre-tournament wins but the numerator is getting pretty small already
23:53:46 <elliptic> it's also a bit weird given that looking at the changelog it feels like things should have gotten a bit harder
23:53:47 <elliptic> I guess you said that earlier
23:54:22 <WalkerBoh> qualitatively, i think the game is harder now than previously - i had more trouble with weaker characters than i thought i should
23:54:39 <WalkerBoh> but it also seemed to me that the top 50 players was way more competitive than previously
23:55:49 <elliptic> 50th place was only about 100 more tourney points than last year
23:56:02 <elliptic> 3151 vs 3043
23:56:25 <WalkerBoh> hm okay
23:56:34 <elliptic> and actually if we throw away bots it was 3108 vs 3043 :P
23:56:39 <WalkerBoh> so maybe it's much more top heavy than top 50
23:56:45 <WalkerBoh> i bet you're really proud of that!
23:56:48 <WalkerBoh> qw is getting good
23:57:16 <elliptic> still sad qw didn't beat minmay, that was my goal for most of the tourney
23:57:25 <nikheizen> at least he beat me
23:57:26 <elliptic> mainly because he was usually only a place or two ahead
23:57:28 <WalkerBoh> that would've been great
23:57:39 <nikheizen> by one spot, i think?
23:57:51 <elliptic> yes qw is sandwiched between you and minmay
23:57:55 <WalkerBoh> nikheizen: if you guys had had a team that was 6 deep you would've beat us for sure
23:58:06 <nikheizen> i should have played less and got 27th!
23:58:27 <nikheizen> WalkerBoh: yeah, i'm still happy that we could field a top ten team this year
23:58:57 <nikheizen> could have gotten top 3 if everyone had worked at least as hard as implo
23:58:57 <WalkerBoh> you guys gave us a bit of pressure, although we were focused on top 4 most of the time
23:59:57 <nikheizen> yeah you guys had a much more balanced team