Time Nick Message 00:08 est31 ShadowNinja: my comment was a reply to sfan5's review comment 00:58 paramat i'll test the new occlusion culling, since some report it too aggressive 01:27 kaeza 999 PRs closed 01:41 paramat possible blocker #4923 01:41 ShadowBot https://github.com/minetest/minetest/issues/4923 -- map_generation_limit trigger server crash 05:40 paramat #4930 +1 05:40 ShadowBot https://github.com/minetest/minetest/issues/4930 -- Fix occlusion culling, again by lhofhansl 05:44 hmmmmm mmm 05:44 hmmmmm i wish the commit message had been fixed 05:44 Zeno` not too late 05:44 hmmmmm not important enough 05:44 Zeno` :) 05:45 paramat oops missed that, it could have been better 05:51 paramat i'm thinking #4888 can perhaps wait until after release, see comment https://github.com/minetest/minetest/pull/4888#issuecomment-266977228 05:51 ShadowBot https://github.com/minetest/minetest/issues/4888 -- Handle multiple deserialization of a block's entities by Rogier-5 05:52 paramat other than that it's just the 2 mod security PRs 05:52 paramat (that are milestones) 05:52 garywhite paramat: I know I'm not a dev, but I think as long as all the milestones have been completed, then just the mod security should be fixed and then push to release 05:52 paramat well there's a bunch of issue milestones 05:53 paramat they're the problem 05:53 paramat the plan was that if the mod security PRs were not ready in time we would just turn mod security off 05:54 garywhite Well ik ShadowNinja was working on it earlier 05:55 paramat so mod security is no problem, but we have 4 issues 05:55 paramat 'broken languages' is labelled as blocker but is probably not 05:56 paramat so we have #4923 #4810 #4528 05:56 ShadowBot https://github.com/minetest/minetest/issues/4923 -- map_generation_limit trigger server crash 05:56 ShadowBot https://github.com/minetest/minetest/issues/4810 -- Unittests crash on Release build 05:56 ShadowBot https://github.com/minetest/minetest/issues/4528 -- minetest 0.4.14 64bit crashes at startup on Windows 10 Pro 64bit: debug.txt permissions/path problem 05:57 paramat and 2 days to release 05:57 paramat :] 05:57 garywhite Well, I can verify that #4528 is only a Win10-related issue, because it does nothing of the sort on Windows 7 Pro 64bit 05:57 ShadowBot https://github.com/minetest/minetest/issues/4528 -- minetest 0.4.14 64bit crashes at startup on Windows 10 Pro 64bit: debug.txt permissions/path problem 05:57 paramat good 05:58 garywhite So it's just something in the newest version of the WinNT kernel, my guess 05:58 Zeno` is that not an issue with the installer? 05:58 paramat no idea 05:59 garywhite Then again, I use the 64-bit portable 06:02 Zeno` where is the installer code anyway heheh? 06:37 est31 We still need to 1. find out which languages are broken and 2. find a way how to disable broken languages 06:37 est31 because https://github.com/minetest/minetest/issues/4637 06:38 est31 https://github.com/minetest/minetest/issues/4638 06:38 sfan5 just move the translation files somewhere else is easiest probably 06:40 Zeno` sfan5, do you know where the installer is? 06:40 sfan5 you mean the windows installer we had last release? 06:40 Zeno` yeah 06:41 est31 rubenwardy does them 06:41 sfan5 ^ 06:41 Zeno` hmm ok 06:41 Zeno` I knew that but thought there might be a repo 06:41 Zeno` I'll ask him when I see him 06:42 Zeno` I really think that #4528 is probably an installer issue rather than a game issue 06:42 ShadowBot https://github.com/minetest/minetest/issues/4528 -- minetest 0.4.14 64bit crashes at startup on Windows 10 Pro 64bit: debug.txt permissions/path problem 06:42 paramat good :] 06:42 Zeno` the installer should, surely, make sure the destination has write perms for the user 06:43 Zeno` I'll see what rubenwardy says. It'd be good if we could see the code (I assume he's using something like NSIS or one of the other scriptable install systems) 06:44 est31 Zeno`: https://gist.github.com/rubenwardy/5397343 06:45 est31 https://github.com/minetest/minetest.github.io/pull/52 06:46 Zeno` oh a gist. I didn't think to look there :D 06:46 Zeno` was looking at his repos. Thanks 06:47 nrzkt Zeno`, maybe we should add this to our repository instead of third party 06:47 Zeno` I think rubenwardy should add it, yes 06:47 Zeno` ok it's install shield 06:48 Zeno` so the directory permissions can be set in that 06:50 Zeno` it could probably go into util/ where buildbot/ etc are 06:51 Zeno` might need some tweaking 06:52 Zeno` e.g. InfoBeforeFile=install.txt will have to be changed etc, but no big deal 06:54 kaeza Zeno`, InnoSetup, not InstallShield 06:55 kaeza BTW doesn't CMake have NSIS as CPack target or something? 06:57 Zeno` ok 06:57 cheapie PR #4847 having been merged does appear to have fixed the issues that VanessaE's servers were having. 06:57 ShadowBot https://github.com/minetest/minetest/issues/4847 -- Fix unnecessary block loading and duplicating objects by juhdanad 06:58 Zeno` kaeza, it's been about 3 years since I used NSIS but probably 06:58 kaeza `cpack --help | grep NSIS` → NSIS = Null Soft Installer | NSIS64 = Null Soft Installer (64-bit) 06:58 kaeza not sure how usable/user-friendly it is 06:58 Zeno` NSIS? 06:58 Zeno` I found it very good 07:00 Zeno` I can fix the debug.txt issue but I suspect that if debug.txt cannot be written to then any other file that needs to be written to will have permission problems as well, so I think it should be classified as an installer issue 07:03 paramat ok could you add a comment? 07:03 kaeza Windows builds should probably start using C:\Users\user\AppData\Roaming\Minetest as $path_user 07:04 kaeza actually, why is debug.txt written to PWD and not $path_user? 07:04 Zeno` different worlds? 07:04 Zeno` hmm 07:05 Zeno` where are things like players stored on windows? I'll look after I cook dinner 07:06 Zeno` Now I'm not sure. Do we want lots of different debug.txt ? e.g. one for each world/game? 07:06 Zeno` maybe that's better, I don't know 07:12 Zeno` well not really 07:12 Zeno` because two instances of a server might use the same 07:15 Zeno` so it really should be in the directory that the server is run from (I think) 07:19 kaeza uh, but isn't there a switch to set the log file? 07:19 kaeza --logfile 07:21 kaeza or maybe it *is* written to $path_user, but $path_user is '.' in RUN_IN_PLACE :| 07:22 Zeno` does that installer create an icon? 07:22 Zeno` and can the icon specify what directory to run from? 07:22 Zeno` so PWD is somewhere that's writeable? 07:22 kaeza https://gist.github.com/rubenwardy/5397343#file-setup32-iss-L41 07:23 Zeno` any way to cd wherever/blah before launching? 07:23 kaeza http://www.jrsoftware.org/ishelp/index.php?topic=iconssection 07:23 kaeza should add `WorkingDir: "{app}"` somewhere on the icon line for minetest.exe 07:24 Zeno` that's probably the solution 07:24 kaeza err, `WorkingDir: "somewhere"` 07:25 Zeno` yeah 07:25 Zeno` just need to work out where somewhere should be now :) 07:25 kaeza probably {userappdata}\Minetest 07:25 Zeno` most likely yes 07:26 Zeno` people running multiple servers on Windows are rare, and they can fend for themselves 07:26 kaeza or {userdocs}\Minetest to be a bit more visible 07:28 Zeno` possibly yeah 07:28 Zeno` but that should work, right? 07:29 Zeno` seems reasonable to me but I don't know much about Windows 07:29 kaeza I'm not on Windows right now and don't have a VM handy to test 07:29 kaeza I wonder if it would work on ReactOS :P 07:29 Zeno` I guess as long as the WD is somewhere writeable it should be fine 07:29 Zeno` hmm 07:30 Zeno` could try with Wine :D 07:30 Zeno` I don't have a VM either *sigh* 07:31 Zeno` I used to... I should have kept it 07:32 red-001 I have windows 07:32 kaeza let me set up the test env 07:32 red-001 so test if minetest runs if the working dir is set to {userdocs}\Minetest ? 07:33 Zeno` red-001, yeah 07:34 Zeno` or {userappdata}\Minetest (which I now think is better) 07:34 Zeno` even though it's less visible 07:38 red-001 yeah it does put its log there 07:38 Zeno` confirmation #1 07:38 ShadowBot https://github.com/minetest/minetest/issues/1 -- GlowStone code by anonymousAwesome 07:38 Zeno` thanks red-001 07:38 Zeno` wow, the first issue 07:38 Zeno` lol 07:39 red-001 but it doesn't save the settings there 07:39 Zeno` where does it save them? 07:40 Zeno` should it save the settings there? 07:40 red-001 well it is a standard on windows 07:41 red-001 to save that sort of stuff in appdata 07:41 red-001 especial if you have an installer 07:42 Zeno` we'd have to change the code to do that (I guess in porting.cpp). Maybe it's good enough to make sure the app actually runs without crashing for this release (using the installer) and then fix the other issues next release (if they don't cause crashes because of permissions)? 07:43 Zeno` and I think that changing it in code is... well I think it's not trivial due to multiple server instances (but maybe we can add something like #ifdef TARGET_INSTALLER to make it easier) 07:43 Zeno` it will take some thought though 07:44 Zeno` but you have a point 07:45 Zeno` and I wonder if the location of settings will also cause issues on Windows10 using the installer 07:46 Zeno` "I can fix the debug.txt issue but I suspect that if debug.txt cannot be written to then any other file that needs to be written to will have permission problems as well, so I think this should be classified as an installer issue" 07:46 Zeno` related to that 07:47 Zeno` red-001, the issue is #4528 (in case you missed it) 07:47 ShadowBot https://github.com/minetest/minetest/issues/4528 -- minetest 0.4.14 64bit crashes at startup on Windows 10 Pro 64bit: debug.txt permissions/path problem 07:49 red-001 well I also have a windows 10 machine 07:49 red-001 should I test if this fixes it? 07:50 Zeno` if you don't mind 07:51 Zeno` it'd be appreciated for sure 08:00 red-001 yeah it fixes the issue 08:00 red-001 but you can't create worlds 08:01 red-001 so that means that you can't change settings either 08:01 kaeza what did you try? 08:02 red-001 I changed the run in path to %appdata%\minetest 08:02 red-001 start in path* 08:03 Zeno` so maybe the installer can give the user write permissions to wherever minetest is installed as well 08:03 Zeno` this is a headache 08:03 kaeza maybe the installer should install to %appdata%\minetest and just run from there for now 08:03 red-001 everything still works if you use c:/games as the install path 08:03 Zeno` kaeza, that would work 08:03 kaeza (instead of program files) 08:04 Zeno` very unusual but it'd work 08:04 red-001 thats what minecraft does iirc 08:04 kaeza we can sort editing the paths in the binary for .16 08:04 Zeno` yeah that's what I'm thinking 08:05 Zeno` otherwise it's gonna block the release for quite a while 08:05 kaeza this line: https://gist.github.com/rubenwardy/5397343#file-setup64-iss-L17 08:05 Zeno` yeah ok 08:05 kaeza use {userappdata}\Minetest 08:05 Zeno` I'd be happy with that for this release 08:05 kaeza or something to that effect 08:06 Zeno` red-001, that's interesting 08:06 Zeno` good enough for minecraft good enough for us this release :D 08:07 Zeno` red-001, btw thanks for taking the time to test things 08:08 Zeno` kaeza, I really need to cook dinner. Can you summarise the proposed solution in the Issue? :3 08:08 kaeza Zeno`, sure 08:08 Zeno` cheers :) 08:09 red-001 its %appdata% 08:21 kaeza red-001, {userappdata} is Inno Setup's way to specify %APPDATA% :) 08:23 red-001 huh didn't know that 08:44 kaeza #4931 08:44 ShadowBot https://github.com/minetest/minetest/issues/4931 -- Official releases should should use %APPDATA%. 08:54 Zeno` now we just need rubenwardy to add the installer script to utils/ 08:54 Zeno` would you agree to that kaeza? 08:54 Zeno` (since it's based on your original script) 08:55 Zeno` utils/ seems right since that's where buildbot/ is 08:55 kaeza I think I released it as WTFPL or BSD but can't access the forums 08:55 kaeza in any case, go ahead 08:55 Zeno` Ok, I'll ask him 08:55 Zeno` off to eat my dinner which is (finally) ready :) 09:10 kaeza I wonder if MS killed UAC virtualization for Windows 10 09:11 kaeza or whatever that VFS thing is called 09:12 kaeza programs should just be able to write to a read-only directory and it gets redirected to some VirtualStore directory somewhere 09:24 nrzkt Zeno`: agree should be in util 09:24 nrzkt kaeza: microsoft killed DHCP client too 12:38 Calinou I'm quite sure VirtualStore is still a thing, kaeza and nrzkt 12:39 Calinou %LOCALAPPDATA%\VirtualStore here 12:39 Calinou (Windows 10 build 14393) 13:04 Zeno` Calinou, that is not where config.txt is trying to be written to though 13:04 Zeno` err debug.txt 13:05 Zeno` debug.txt is written to the directory that the application is launched in 13:05 Zeno` so, the working directory (or the binary directory) needs to be changed 13:07 Calinou should be changed so that it's written in %APPDATA% on Windows in non-run-in-place build, maybe 13:07 Calinou maybe disable it entirely by default too? 13:07 Calinou (or keep it to errors only) 13:25 Zeno` it only affects build installed using the installer though 13:25 Zeno` and a "proper" fix can be done in the next release (it'll take time) 13:26 Zeno` the proper fix (changing paths and stuff) is not exactly trivial and will probably require a new cmake flag 14:56 sfan5 Calinou: non-RUN_IN_PLACE is entirely unsupported on windows 16:23 sofar Was there anything anyone needed from me for the release? 16:23 sofar I'm probably not going to check github this week 17:08 red-001 has the 0.4.15-RC been released? 17:09 garywhite idk red-001. There's still a few bugs to fox AFAIK 17:09 garywhite fix* 18:07 Zeno` maybe we should just close #4931 and https://github.com/minetest/minetest/issues/4528#issuecomment-267906193 18:07 ShadowBot https://github.com/minetest/minetest/issues/4931 -- Official releases should should use %APPDATA%. 18:08 Zeno` it's not as if the installer is really official anyway considering it not in the source tree or part of the minetest project (yet) 18:09 Zeno` the fix is easy but people seem to be opposed to it 18:11 Zeno` there is no way to create an "install wherever you want" and and "an installer friendly" version of the build without adding a whole heap of stuff to porting.cpp 20:46 garywhite est31, Zeno`, sfan5, sofar, ShadowNinja: Does anyone know what day 0.4.15 is planned to be released? Cause some people want to make a trailer for MT and considering trying to release at the same time as 0.4.15 20:47 sfan5 garywhite: 21st if i remember the date correctly 20:48 garywhite sfan5: Do you know if that'll be delayed at all? 20:48 sfan5 do i look like i can tell the future 20:48 garywhite I was just asking cause ik there are still issues that have to be corrected 20:56 garywhite Is there a specific release manager for 0.4.15? 22:30 ShadowNinja garywhite: AFAIK there is no release manager ATM. The release may be delayed. Can't the trailer just be released immediately after they see the release announcement? 22:31 garywhite TruYes 22:32 garywhite True* 22:32 octacian garywhite: first, this isn't a trailer for 0.4.15! it's just MT in general. 22:33 garywhite octacian: I know, but if it can get released at the same time as 0.4.15, it'd be much cooler, like it was made to accompany it 22:34 ShadowNinja My security patches are considered blocking, so those will have to be merged. There are two other blockers, but one looks mislabeled (the language issue is hardly new or critical). 22:34 garywhite ShadowNinja: If necessary, the language system could be implemented into a 0.4.15a update just for it 22:44 sfan5 ShadowNinja: the language issue is not new, but it's critical for release 22:45 garywhite again, what about a supplemental release for 0.4.15? Like 0.4.15a? 22:46 ShadowNinja sfan5: Why so? Can't users just avoid those languages like they've done for every previous release? The font might be fixable, but if RTL support is an Irrlicht issue it isn't doing to get fixed for quite a while. 22:50 garywhite sfan5, what languages are complete and ready to be packaged?