Time Nick Message 00:18 paramat #2011 is ready for review/merge. this needed a lot of work 00:18 ShadowBot https://github.com/minetest/minetest/issues/2011 -- Conf.example: State biome noises in positional format. Increase indentation throughout file. Correct mistakes. Shorten lines by paramat 00:20 paramat ^ that includes Wayward_One 's pull #2010 but places the line in the correct place 00:20 ShadowBot https://github.com/minetest/minetest/issues/2010 -- Fix redundancy in a comment by Wayward1 00:28 paramat turns out the line fixed by 2010 was a few lines above where it should be. perhaps 2010 could be closed? 00:59 Zeno` Does it conflict with #1977? 00:59 ShadowBot https://github.com/minetest/minetest/issues/1977 -- Rewrite lua_api.txt into Markdown format by Wuzzy2 00:59 Zeno` err 00:59 Zeno` I'll leave it as a trick question 01:07 paramat hopefully not =) 01:13 Wayward_One oh, it does? should I close it? 01:13 Wayward_One *is 01:17 paramat 2011 contains your 2010 fix, and it seems that line you fixed is in the wrong place, perhaps you agree/ 01:18 paramat *? 01:18 Zeno` paramat, can you merge 2010 into your PR? 01:18 paramat its in there already with the author's permission 01:19 paramat so IMO 2010 should be closed and 2011 merged 01:19 Zeno` Ok, I'll do that no 01:19 Zeno` Ok, I'll do that now 01:21 kaeza 2015 is coming... both year an pull request number :) 01:21 kaeza hi 01:23 paramat to clarify, see logs for this channel. 2010 was correct to fix that line but while editing conf.example i saw that line was also in the wrong place and corrected that 01:25 paramat the file needs a little more work but the worst stuff is fixed, hopefuly clearer to read now 01:25 Zeno` yeah 01:30 paramat thanks, busy now but back later O/ 03:30 Tesseract sfan5: How is the Windows building going? 03:31 Tesseract Also, someone needs to get that one reaining blocker in GDB or something and figure out what's locking up. 03:33 Tesseract Seems GDB runs on Windows, so that should work. 03:36 Tesseract kaeza: You've been able to reproduce it, can you debug it? 03:36 gregorycu Tesseract: It's being discussed in #minetest 03:36 gregorycu We are not having luck 03:37 Tesseract Oh. 03:37 gregorycu That's kaeza, Zeno` and myself 03:37 gregorycu Zeno has installed the mingw toolchain, but can't repo 03:40 Tesseract repro! repo suggests repository to me. 03:40 Tesseract But it sounds like you did find an issue. 03:41 gregorycu Well, bad Lua crashes it 03:41 gregorycu But we don't have bad Lua here 03:41 gregorycu As far as I can tell 03:41 Tesseract You could just add a mutex and add mutex auto locks to every function in that file for a quick and dirty theory test. 03:43 gregorycu I can't reproduce the issue 03:43 gregorycu I have verified that different threads access those functions 03:44 Zeno` It's not JuaJIT is it? 03:45 Tesseract gregorycu: Well, fix that bug and see if it fixes the other bug. 03:46 Tesseract (or if they're the same bug rather) 03:46 gregorycu I can't repro, so I can't tell if it makes a difference 03:48 Tesseract gregorycu: Give the diff to kaeza and he can check it. 03:48 gregorycu I'm not sure where the fellow is, I need him to do a breakin and stackdump 03:48 Zeno` Tesseract, I think there are two (minor and low-priority) bugs related to #2006 now that I can see it on Windows instead of just looking at code 03:48 ShadowBot https://github.com/minetest/minetest/issues/2006 -- Error when deleting a world 03:50 Tesseract Zeno`: Huh? Just seems like it's trying to get directory lists on files. 03:51 Zeno` FindFirstFile error doesn't come from there though 03:51 Tesseract Zeno`: Your patch is mostly good (shouldn't ignore that error), but it doesn't solve the real problem. 03:51 Tesseract Zeno`: Oh... 03:51 Zeno` yeah, I think they should *both* be fixed :) 03:51 Tesseract Zeno`: Where then? 03:52 Zeno` gosh... gotta work out how to navigate in Windows 03:53 Zeno` getAvailableWorlds() 03:53 Zeno` line 232 of subgame.cpp 03:56 Zeno` Tesseract, just double checked with gdb and that is where fs::GetDirListing() is being called from (for that particular bug) 03:59 Tesseract Zeno`: That means that PATH_USER/worlds is a file? 04:00 Tesseract Those infostream lines should become dstream, too. 04:03 Tesseract Zeno`: That seems really bizarre. 04:03 Zeno` It is... doing more tests 04:05 Zeno` wth 04:07 Zeno` I can't even create a world hehe 04:12 Zeno` I'd better install git for Windows somehow 04:28 Tesseract Zeno`: Is my patch good to merge? 04:28 Zeno` It looks good, I haven't tested 04:30 paramat my commit messages for the last 2 merges were not good, sorry. if anyone has time and is able to edit them they should start with 'conf.example:' 04:30 Tesseract Neither have I :-/. Better wait just in case. 04:31 Tesseract paramat: Too late now, limit is 10 minutes. 04:32 paramat okay, apologies 04:49 gregorycu I just made minetest hang with a 5 second sleep 04:53 gregorycu Who wants to know where I did it? 04:56 gregorycu Ahh crap, no I didn't 06:21 Zeno` brb 07:54 * Zeno` has gone crazy 07:55 gregorycu Thats good 07:55 Zeno` How did the checkbox thing go? 07:56 gregorycu I'd prefer not to talk about it 07:56 Zeno` that bad? 07:56 gregorycu No idea still 07:56 gregorycu It's happening before the game starts 09:43 sfan5 Tesseract: haven't tried building yet 10:17 Zeno` people are asking about #2000. Yay or Nay? 10:17 ShadowBot https://github.com/minetest/minetest/issues/2000 -- Large increase in performance by Zeno- 10:22 Krock More like "Yay". 10:22 Krock But I don't like this hack where new_data is initialized 10:23 Zeno` *shrug*; I don't like it either but it makes a huge difference ;) 10:23 Krock if you say so... how big is that difference? +/-? 10:24 Zeno` in the actual function, 40-50% increase. Overall, 5-7%. (very roughly) 10:24 Zeno` actually that's not right. The actual function is up to 2x as fast 10:24 Zeno` so that's 200% 10:25 Zeno` overall is still 5-7% though 10:25 Krock nice. then a +1 from my side..but that doesn't count. 10:27 Zeno` I still have no idea about #1959 :( 10:27 ShadowBot https://github.com/minetest/minetest/issues/1959 -- Checking public serverlist checkbox causes minetest to freeze. 10:33 Zeno` sfan5, are you able to make a 32-bit build? 10:33 sfan5 Zeno`: mingw or msvc 10:33 Zeno` mingw 10:33 sfan5 yes 10:34 Zeno` ok, that'd be good. To rule certain things out I guess 10:34 sfan5 of current HEAD or something else? 10:34 Zeno` current HEAD (looking at 1959) 10:34 sfan5 the -rc1 build is from a145a6a 10:35 Zeno` just trying to reproduce... if rc1 exhibits it (32-bit) then at least there is 2nd confirmation 10:35 sfan5 there is already a win32 -rc1 10:37 sfan5 Zeno`: do you need a new build or not? 10:39 Zeno` sfan5, does the current mingw-32 build have the problem? 10:40 Zeno` and who makes the official mingw-32 builds? 10:40 sfan5 Zeno`: i dunno; me 10:40 Krock eeks. http://pastebin.com/5wyXbsFB 10:40 Krock but it compiled fine. 10:40 Zeno` if it is not in the official builds then I think it should be marked as unconfirmed and obscure (and the blocker removed) 10:40 sfan5 wat 10:41 sfan5 the RC builds are official 10:41 Zeno` so the RC build has the bug? 10:41 sfan5 thats not what i said 10:41 sfan5 i dunno 10:41 sfan5 haven't tested the win32 ver. yet 10:41 Zeno` What *I* am saying is that nobody can reproduce it :( 10:41 Zeno` nobody who has tried* 10:42 Zeno` and if it's not in the RC build then it's "obscure" and/or "unofficial-build" 10:42 Zeno` and should, therefore, not be a blocker but a "normal" bug 10:53 Zeno` ugh, it could be anything :( 11:19 kilbith someone can cares of the changelog please ? http://dev.minetest.net/Changelog 11:31 gregorycu We failed :( 11:32 gregorycu um... so, what we probably need is better diagnostics 11:32 gregorycu tracing, if you will 11:45 gregorycu_ I know it's the best idea of all time 11:51 sfan5 PilzAdam: couldn't this be fixed because 733 was already merged last year? https://github.com/minetest/minetest/issues/802 12:18 gregorycu #2000 12:18 ShadowBot https://github.com/minetest/minetest/issues/2000 -- Large increase in performance by Zeno- 12:18 Zeno` hmm? 12:18 gregorycu Looks good bruf 12:18 Zeno` no problem on Windows? 12:20 Zeno` gregorycu ^^ 12:21 gregorycu Looks good from a theoretical perspective 12:21 gregorycu I'll profile if you like ;) 12:21 Zeno` we'll have no theory here! 12:21 gregorycu Fuck, I have so many changes outstanding 12:21 Zeno` but yeah, so long as it works and profile improves (it will) 12:23 gregorycu Zeno, we really need tracing 12:23 Zeno` gregorycu, verbose gives a fair amount already 12:23 Zeno` but maybe it needs more 12:33 gregorycu By the way Zeno`, you know you can replace operator[] for a class 12:33 gregorycu Actually... hmm... 12:35 Zeno` ? 12:36 gregorycu With regards to allocating a whole bunch of MapBlock 12:36 gregorycu You can override its operator[] to make it fast 12:37 gregorycu If you so choose 12:37 Zeno` Yes. But the default constructor of Map*Node* is broken 12:37 gregorycu err... sorry, MapNode 12:38 gregorycu Because CONTENT_AIR is the default? 12:38 gregorycu Or rather, because it has a default? 12:38 Zeno` because it has a default 12:39 Zeno` there should be a constructor with no default 12:39 Zeno` but... 12:39 Zeno` there's not 12:40 Zeno` so when new[] is called it calls MapNode(content_t content = CONTENT_AIR, u8 a_param1=0, u8 a_param2=0) for no reason at all 12:40 Zeno` (taking up 40-50% of the function's time) 12:41 Zeno` calls it for each element* 12:41 gregorycu I'll take a look, see how it went 12:42 gregorycu So, it works 12:46 gregorycu_ It doesn't look good Zeno 12:47 Zeno` In what way? 12:47 gregorycu_ Performance wise 12:47 gregorycu_ It looks excellent 12:47 gregorycu_ See what I did there 12:48 Zeno` :/ 12:48 Zeno` you're crazier than I am 12:48 Zeno` which is saying something! 12:48 exio4 Zeno`, easier to rewrite 12:48 exio4 er 12:49 Zeno` exio4, got any numbers? 12:49 Zeno` err 12:49 Zeno` gregorycu, got any numbers? 12:50 gregorycu My data indicates it's twice as fast 12:50 gregorycu From 5.41% of the running time, down to 2.91% 12:50 gregorycu Well, almost twice as fast 12:50 exio4 uh, what I said here was meant to be sent in #minetest, by the way 12:51 Zeno` ok 12:51 Zeno` any objections to me merging #2000? It seems like a no-brainer to me 12:51 ShadowBot https://github.com/minetest/minetest/issues/2000 -- Large increase in performance by Zeno- 12:52 Zeno` those numbers are similar to mine btw 12:53 gregorycu I haven't looked at the code too deeply 12:53 gregorycu All the data is properly initialised later down? 12:55 Zeno` there is nothing to initialise... it's all done after the mem is allocated 12:56 Zeno` Tesseract, your patch is OK to merge 12:56 Zeno` (tested... finally) 12:56 gregorycu Where? 12:57 Zeno` gregorycu, where what? 12:57 gregorycu Where do the new elements in new_data get initialised? 12:57 gregorycu I assume by the caller? 12:57 Zeno` line 189 12:58 Zeno` and then 191 onwards 12:58 gregorycu That just initialises the existing area 12:58 gregorycu Not the new area 12:58 gregorycu Right? 12:58 Zeno` 191-203 copies from old to new 12:59 gregorycu But old is a subset of new 12:59 Zeno` no, new is a subset (possibly) of old 13:00 Zeno` actually, not even that... // Cancel if m_area already contains the requested area 13:01 gregorycu_ Screw this connection 13:01 gregorycu_ Are you saying that addArea removes area? 13:01 Zeno` nope 13:04 gregorycu *sigh* 13:05 gregorycu Ok, if you are sure 13:05 Zeno` not having much luck are you :P 13:06 gregorycu Yeah, it's bullshit, wtf is going on 13:06 gregorycu Anyway, yeah, ok 13:06 gregorycu Merry Christmas all 13:06 Zeno` Still an hour to go! 13:06 Krock gregorycu, too early. 13:06 gregorycu That's what she said 13:10 Zeno` gregorycu, flags play a part as well 13:10 Zeno` I see what you're suggesting now but it's not a problem 13:10 Zeno` All the flags are set to VOXELFLAG_NO_DATA 13:11 kilbith Zeno`, Jesus isn't born in Australia ;) 13:11 kilbith 13:11 Zeno` kilbith, tell that to my cousin 13:18 gregorycu Zeno, I just got another 2% speed up 13:19 Zeno` Tesseract, you may apply your own patch. I can't be bothered removing the line numbers hehe 13:20 gregorycu You wanna test this 13:20 Zeno` ype, make a PR 13:20 gregorycu Make VoxelManipulator::getNodeNoEx take a reference type, not a value type 13:22 Zeno` where? 13:23 gregorycu voxel.h, line 409 13:24 Zeno` Yeah I did this a few weeks ago and nobody believed me, hehe 13:25 gregorycu Great minds 13:27 gregorycu So you did get a perf improvement then? 13:27 gregorycu Not massive, but it's low hanging fruit 13:27 gregorycu It's 2%, it's not trivial either 13:41 sfan5 ok, so I got MSVC (VS13) working in a VM, it compiles so far, but most of the libs i have do not work with VS (i have .dll.a instead .lib) 13:42 Krock sfan5, I've got the libs but I'm not sure if they're VS13 compatible 13:43 sfan5 well.. i need 64bit libs too 13:43 Krock well then :/ 13:43 sfan5 basically all the libs used by the buildbot 13:43 sfan5 in 32bit and 64bit 13:43 Zeno` gregorycu, I did 13:43 Zeno` there is a closed PR somewhere which does it 13:49 gregorycu I made getNodeNoEx go from 9% to 7.15% 13:49 gregorycu Just by making it return a reference instead of a value 14:14 Zeno` return a reference? 14:15 gregorycu Yes 14:15 gregorycu getNodeNoEx currently returns a copy 14:15 Zeno` yeah, because of threads 14:16 Zeno` What is the reference referencing? 14:16 gregorycu But getNodeNoEx has no locing 14:16 gregorycu locking 14:16 celeron55 that's actually a valid optimization; it could be made much faster by actually storing the result in a reference too 14:17 celeron55 also, it's not thread-safe so zeno's point is invalid 14:17 celeron55 gregorycu: are you returning a statically allocated CONTENT_IGNORE node when not found? 14:17 Zeno` sheesh, just pass a pointer and store the result in that 14:17 gregorycu Yes 14:17 gregorycu Well, I named it differently 14:18 celeron55 Zeno`: lol, that would make an absolutely awful interface 14:18 gregorycu I'll submit a PR after christmas 14:18 gregorycu How frequent are releases? 14:19 Zeno` yeah, like a 3000 line main() was a great "interface" 14:19 Zeno` anyway, it's Christmas Day. I'm not going to argue 14:20 gregorycu You guys 14:20 gregorycu Anyway, catch ya later 14:21 PilzAdam Zeno`, why have you merged #2000 ? 14:21 ShadowBot https://github.com/minetest/minetest/issues/2000 -- Large increase in performance by Zeno- 14:21 PilzAdam it's not a bugfix 14:21 Zeno` PilzAdam, it's not a feature 14:22 PilzAdam only bugfixes are allowed during feature freeze 14:22 Zeno` I don't think I added a feature 14:22 PilzAdam thats not relevant 14:22 PilzAdam only bugfixes are allowed 14:22 Zeno` ? 14:22 Zeno` "feature freeze" 14:22 PilzAdam optimization isn't a bugfix 14:23 Zeno` and it's not a feature 14:23 PilzAdam "feature freeze" doesn't apply to features only 14:24 Zeno` Anyway I asked 3 days ago (maybe 2?) and it was for the most part approved 14:24 PilzAdam it applies to anything except bugfixes 14:24 Zeno` no, that's a "code freeze" 14:25 PilzAdam the name isn't that relevant 14:25 PilzAdam the point is that we want to find and fix bugs 14:26 Zeno` yes it is. "feature freeze" and "code freeze" are two different things 14:26 PilzAdam anything that isn't a bugfix just increases the risk of adding new bugs 14:26 Zeno` anyway I've been the only one even looking at the bugs until today :/ 14:28 Zeno` meh, revert it if you like 14:30 Zeno` seriously, what's the problem? 14:32 PilzAdam it may introduce bugs and we don't have enough time to test it before the release 14:32 Zeno` ok, well... look at this: https://github.com/minetest/minetest/commit/565959db860ddfbfdb996a6d4e796abd5cf08d22 14:33 PilzAdam the point of a "feature freeze" / "code freeze" ( / whatever you want to call it) is to minimize the risk of adding new bugs in the testing phase by not allowing anything except bugfixes 14:33 Zeno` that HIDES for no reason at all VALID warnings 14:33 PilzAdam that isn't good then 14:33 Zeno` so any time in future there is another bug (there is already one) that it hides nobody can see it 14:34 PilzAdam but how is that relevant now? 14:34 Zeno` because that is worse than adding something that is can be tested 14:34 Zeno` and it was added without even a PR 14:35 Zeno` there is at least one bug that it hides 14:35 PilzAdam talk to hmmmm then and revert it 14:35 Zeno` I have. He doesn't like the warnings that he calls bogus (even though they're not) 14:36 PilzAdam ask other devs what they think, then 14:36 Zeno` They've had the opportunity to say something 14:37 PilzAdam if "a lot" of devs think that the warnings should be taken seriously then tell hmmmm that he needs to take them serious 14:37 Zeno` What I'm saying is there is things in the last 24-36 hours a LOT worse than a tested patch 14:38 PilzAdam yes, that might be true 14:38 PilzAdam that doesn't make your patch better, or does it? 14:38 PilzAdam anyway, I guess I think that #2000 should not be reverted 14:38 ShadowBot https://github.com/minetest/minetest/issues/2000 -- Large increase in performance by Zeno- 14:39 Zeno` Of course it doesn't. I'm saying that 2000 was tested and is fine 14:39 PilzAdam just for future PRs, only merge bugfixes during "feature freezes" (despite it's name) 14:39 Zeno` ok, point taken 14:40 PilzAdam also it's christmas, so be happy :D 14:41 Zeno` Never said I wasn't happy :P 12:40AM 25th Dec 2014 here 14:41 Zeno` And I just saw Santa 14:43 shadowzone Zeno`, lucky he never comes to my house 14:43 shadowzone stupid brother 14:43 shadowzone +s 14:44 Zeno` Ignoring 2000... I'm more worried that I cannot create a new world with my Windows build 14:44 Zeno` I really hope it's something I did wrong building it and setting it up because I'm unfamiliar with Win7 rather than an actual bug 14:45 Zeno` Nobody else has the problem so I assume it's something I was doing wrong :/ 14:46 Zeno` Also #9159 14:46 ShadowBot Zeno`: Error: Delemiter not found in "HTTP Error 404: Not Found" 14:47 Zeno` err #1959 14:47 ShadowBot https://github.com/minetest/minetest/issues/1959 -- Checking public serverlist checkbox causes minetest to freeze. 17:21 travis-ci neoascetic/minetest#4 (travis-osx - 27a7c8f : Pavel Puchkin): The build passed. 17:21 travis-ci Change view : https://github.com/neoascetic/minetest/compare/545416a1ea3e...27a7c8f05f38 17:21 travis-ci Build details : http://travis-ci.org/neoascetic/minetest/builds/45043848 17:22 neoascetic Nice! :) 17:22 neoascetic Would you like to see this messages from travis if master build fail for OSX? 18:22 hmmmm you guys reclassified PR 1959? 18:23 hmmmm i agree, most people can't reproduce even with the affected builds 18:31 shadowzone When is 0.4.11 release? 18:49 hmmmm right now i'd think, but i guess others are waiting for things 20:00 travis-ci Latest master build for OSX: The build was fixed. 20:00 travis-ci Build details : http://travis-ci.org/neoascetic/minetest/builds/45052394 20:02 VanessaE don't make travis-ci join/part all the time. if it's gonna inform the channel of build events, just let it inform. joins/parts add excess traffic. 20:06 shadowzone VanessaE, got it, want me to inform him next time he joins? 20:07 VanessaE well whoever's controlling the travis-ci bot, yeah 20:09 sfan5 wait what 20:09 VanessaE ? 20:09 sfan5 why does this bot announce something for "neoascetic/minetest" in here 20:09 neoascetic Because there is no other way to test master on OSX 20:10 neoascetic travis-ci currently does not allow multi-os builds 20:10 neoascetic so I simply mirror master branch of minetest 20:10 sfan5 i know that travis offers osx, but why does the bot announce it in here 20:11 sfan5 i mean, it would make more sense if it would monitor minetest/minetest build status 20:11 sfan5 but it's still nice to have osx build status 20:11 neoascetic in fact, this trick with mirror is just a workaround 20:12 neoascetic when travis allows multi-os builds, we may integrate it with minetest in native way 20:32 VanessaE on a different note: I just pulled and compiled and I'm now getting 30 fps at 75m view range on VE-Survival spawn 20:32 VanessaE nice improvement. 20:34 shadowzone VanessaE, is that recent git? 20:35 VanessaE that's current HEAD as of half an hour ago. 20:35 shadowzone I have a question, would git pulling and compiling update MT to 0.4.11, or do I have to wait till the official release? 20:36 VanessaE you'll get 0.4.11 if you pull right after the release 20:36 VanessaE (or if you fetch the release tag, which I don't know how to do) 20:36 shadowzone Okay 20:36 VanessaE but it doesn't matter, the code is the same either way 20:36 shadowzone Which when the official release is...I dunno. 20:37 VanessaE [12-24 13:31] When is 0.4.11 release? 20:37 VanessaE [12-24 13:50] right now i'd think, but i guess others are waiting for things 21:55 hmmmm well 21:55 hmmmm i dunno about you guys but I'm going to a christmas eve party 21:56 hmmmm if you're actually releasing tonight, somebody else is gonna need to do it 21:56 sfan5 i'll not be building all of my libs again using msvc tonight 21:56 sfan5 if we want to release we won't have msvc builds (or just 32 bit if Krock does the build) 21:57 VanessaE what's stopping you from releasing right now? 21:58 sfan5 msvc 21:58 VanessaE besides that 21:58 sfan5 nothing 21:58 VanessaE then it makes sense, to my mind anyway, to release now and do the MSVC build separately. 21:58 VanessaE so as not to miss the assumed deadline (again) 22:00 sfan5 VanessaE: IIRC we wanted to release around christmas 22:00 hmmmm around christmas does not necessarily == christmas eve 22:00 hmmmm =/ 22:00 VanessaE I know that 22:00 hmmmm sp what 22:00 VanessaE but releasing after christmas completely loses the point 22:00 hmmmm so what's wrong with the MSVC stuff 22:00 sfan5 BlockMen is not here 22:01 hmmmm efgh 22:01 hmmmm see 22:01 sfan5 and has not been here for quite some time 22:01 VanessaE you may as well just release on Jan 1 or something if you don't release today or tomorrow. 22:01 sfan5 i have MSVC set up in a VM 22:01 hmmmm this is why I wanted to release earlier than the 24th 22:01 sfan5 but MSVC doesn't like .dll.a or .a 22:01 hmmmm the release process should start 2 days before a scheduled release 22:01 hmmmm alright i'll figure it out later when i'm back home 22:01 hmmmm bbl 22:02 sfan5 according to MinetestBot BlockMen has been missing sine 2014-11-09 22:02 sfan5 VanessaE: actually.. is the changelog even updated? 22:02 VanessaE partly but it's incomplete 22:02 VanessaE Calinou usually handles that doesn't he? 22:03 sfan5 i dunno 22:03 VanessaE most recent update there was on the 19th 22:03 sfan5 didn't ShadowNinja do it 22:05 VanessaE A large part of it yeah 22:05 VanessaE seems to be mostly him, you, Calinou and PA 22:42 Tesseract I feel like Christmas is a bad time to release, since everyone's doing things with friends/family. 22:43 Tesseract Idea: Add low/med/hi prio labels instead of low/normal labels for bugs/enhancements. 23:56 kahrl Well, fuck 23:56 kahrl I was updating the changelog 23:56 kahrl when I accidentally hit Ctrl-W