Time Nick Message 01:17 paramat will push 2448 soon. now tested, and i created the 'no_float' bool instead of running 'g_settings->getBool' a million times within the loop 01:17 paramat #2448 01:17 ShadowBot https://github.com/minetest/minetest/issues/2448 -- Dungeongen: Optionally set ignore to be untouchable to disable floating dungeons by paramat 01:40 paramat now pushing 2448 01:52 paramat done =) 03:23 devmarth any 0.4.12 fedora builds? 03:29 paramat #2451 03:29 ShadowBot https://github.com/minetest/minetest/issues/2451 -- Heightmaps: Generalise for all mapgens. findGroundLevel: Return -MAP_GENERATION_LIMIT if surface not found by paramat 03:31 Zeno` ShadowNinja, what was the main purpose of https://github.com/minetest/minetest/commit/9e2a9b55e185b92074b32d3df336920f33b29e5a 03:31 Zeno` just reduce the number of functions? Or were there other objectives? 03:34 Zeno` Because most files other than game.cpp still do delete[] the returned string 03:34 ShadowNinja Zeno`: To reduce things like wide_to_narrow(std::wstring(narrow_to_wide(gettext()).c_str())); 03:35 Zeno` (I'm saying that because at first I thought a secondary objective was to avoid requiring the calling function to delete[] the string) 03:35 Zeno` Well, if I revert just string.cpp and string.h everything still works so maybe the reductions are still there and it just needs cleaning up? 03:37 Zeno` The only real duplication is in gettext.h (inline wchar_t* chartowchar_t(const char *str)) but maybe that can be wrapped so calling functions don't need to delete[] the returned string 03:38 Zeno` I don't understand why there aren't multiple free errors right at the moment anyway :/ 03:38 Zeno` e.g. https://github.com/minetest/minetest/blob/master/src/client.cpp#L1568 03:40 Zeno` oh, it returns a Buffer object? 03:40 Zeno` no 03:41 Zeno` oh you do have to free it. Must have been my imagination that those were gone away 03:42 ShadowNinja Zeno`: I'm not sure what causes the Android issues. Just net me check any patches you come up with. 03:42 Zeno` ShadowNinja, you will not like the patches :) 03:43 Zeno` https://github.com/Zeno-/minetest/commit/06e774b100bc6d093700e7fa6001e72f39dd4093 03:43 devmarth heh 03:43 Zeno` you will immediately see why 03:43 est31 ShadowNinja, if you have time, could you find out what I did wrong in #2447? 03:43 ShadowBot https://github.com/minetest/minetest/issues/2447 -- Add remote inventory inspection feature by est31 03:44 Zeno` but they do fix the Android issue 03:44 est31 because somehow I can't trigger the assert(false) 03:44 ShadowNinja "Revert all revent narrow_to_wide changes" 03:44 ShadowNinja recent* 03:45 ShadowNinja Zeno`: Try to revert little bits of it until it works. 03:47 Zeno` ShadowNinja, not so easy with the dir changes (and that was my first approach) 03:48 Zeno` These changes are quite little anyway 03:48 Zeno` fugly though 03:48 Zeno` I'd like to apply that patch in any case because I am sick of this issue after fudging about for so long 03:50 ShadowNinja est31: Um, you ARE using a Debug build, right? 03:51 Zeno` ShadowNinja, I know you don't like it (and I don't particularly either), but unless you can fix the issue I am giving up :) 03:51 est31 ShadowNinja, yes and I've tried that in many variations, with returning false by default (using the pushboolean), or with logging to error stream (nothing printed) 03:52 Zeno` est31, add some couts 03:52 Zeno` oh 03:52 * Zeno` fixes the formatting while ShadowNinja ponders the evilness of the world 04:00 Zeno` The bug has been there since before 0.4.12 :/ 04:11 Zeno` ShadowNinja, can I merge this? There is no reason I can see why it cannot be "cleaned up" later 04:11 Zeno` At least it works :p 04:14 * Zeno` has been gazing at the code so long now it all looks the same haha 04:21 Zeno` https://github.com/minetest/minetest/compare/master...Zeno-:refactor_gettext 04:22 Zeno` that's as much as I'm doing :P I need to move on 04:27 Zeno` I'll change the name and commit message (and squash) of course (it's no longer a refactor) 04:36 * Zeno` pokes ShadowNinja to see if he's still awake 04:36 Zeno` maybe he fell over and is badly hurt :3 Hope he's not home alone and in need of help 04:37 Zeno` If anyone knows his address perhaps we should send an ambulance over there just in case :3 04:37 ShadowNinja Zeno`: Heh, no. 04:39 ShadowNinja Zeno`: The (()) in the macros is unnecessary. 04:39 Zeno` yeah, we've had that discussion before 04:39 Zeno` lots of times :) 04:39 Zeno` I can remove them... no longer care about that kind of stuff :p 04:40 Zeno` technically they _are_ necessary though 04:40 ShadowNinja Zeno`: Really? In what case? 04:41 Zeno` what if I decide for whatever bizarre reason to have gettext_noop(a + b) somewhere in the code? 04:42 ShadowNinja I don't like how _c uses MultiByteToWide on win, but w_to_n doesn't. 04:42 Zeno` but that's too simple to demonstrate... I'll remove them 04:43 ShadowNinja Zeno`: I mean gettext((x)) is unnecessary, not (x). 04:43 Zeno` yes, I'd agree with that... 04:43 ShadowNinja The () from the function handles it. 04:43 Zeno` umm 04:43 Zeno` is the w_to_n existing? 04:43 * Zeno` looks 04:44 Zeno` I dunno why that's the case... I didn't touch it really (just formatting) 04:44 ShadowNinja Zeno`: er, w-to-n. 04:44 ShadowNinja ... 04:44 ShadowNinja n-to-w 04:46 Zeno` Odd. But that's how it's always been isn't it? 04:46 devmarth whats the "pause menu" /src/ file called? 04:46 Zeno` devmarth, it's in game.cpp 04:47 devmarth thanks zeno 04:47 Zeno` the logical place :3 04:48 ShadowNinja Zeno`: Yes, but we want to make it better... 04:49 Zeno` ShadowNinja, I know and TBH I hate that whole file. But I want to get the bug fixed first. I still plan on doing what the branch name says... "refactor" (which this doesn't do) 04:50 Zeno` There are too many #if #else #end #blah for my liking 04:50 Zeno` *for a start 04:51 Zeno` Normally I wouldn't push for a branch that is not finished to be merged but Android has not been tested since before 0.4.12 was released 04:51 Zeno` because it hasn't worked since February. It'd be nice for people to start testing it 04:52 Zeno` Since there are so many changes since then 04:53 Zeno` err 0.4.11 04:53 devmarth Zeno, I have not been able to upload a directory to GitHub through git, I've had to use Windows GUI version but I don't want to do that, any solution? 04:53 Zeno` devmarth, ask in #minetest please 04:53 devmarth I'll try Google first actually 04:54 Zeno` ok :) 04:55 * Zeno` removed the superfluous parenthesis 04:58 Zeno` maybe Windows doesn't need that alternate method any more 04:58 Zeno` I dunno.. I have no way of testing 05:00 Zeno` maybe I should just inline that function 05:01 Zeno` hmm, that won't work 05:02 Zeno` hang on... that's how it is in master 05:03 Zeno` https://github.com/minetest/minetest/commit/d3026293927f560465b325280fef697cc3361daa 05:04 Zeno` I don't think I should remove that 05:16 devmarth hey zeno, you here? sorry to bother, but i have one more question 05:17 devmarth i tried logging into git, and stuff, and when i push to a repository it says i dont have correct access rights, when its my own repository 05:18 ShadowNinja Finally got #2402 to build. :-) 05:18 ShadowBot https://github.com/minetest/minetest/issues/2402 -- Clean up and tweak build system by ShadowNinja 05:19 devmarth ShadowNinja, can you help me since Zeno seems to be afk? 05:20 devmarth oh nevermind, i did a silly mistake 05:20 devmarth i was logged in but didnt type the ssh address right 05:22 Zeno` https://github.com/minetest/minetest/compare/master...Zeno-:refactor_gettext is updated 05:22 Zeno` ShadowNinja, what was the problem? 05:24 Zeno` I had a look at that clang thing last night just before I fell asleep and it seemed strange 05:40 ShadowNinja Zeno`: Clang doesn't work with FreeType apparently, and I enabled FreeType detection by default. I just disabled it for Clang. 05:40 ShadowNinja Zeno`: Can you check that PR and comment on it please? :-) 05:41 Zeno` that's... I don't understand that, but anyway. Good to know I guess 05:41 Zeno` ShadowNinja, just let me push update 05:42 Zeno` I'll just make it a PR 05:42 Zeno` oh, yours? 05:42 Zeno` ok 05:43 Zeno` #2452 ... I need to make sure I didn't break windows build anyway heh 05:43 ShadowBot https://github.com/minetest/minetest/issues/2452 -- Fix android gettext issue by Zeno- 05:44 Zeno` I'll trust travis. Looking at 2402 05:50 Zeno` Renamed server binary to minetest-server ... I'm sure that'll be popular 05:51 ShadowNinja Zeno`: Seems so -- except that it might break things... 05:52 Zeno` ShadowNinja, I can't comment too much without actually testing all the builds :( 05:52 ShadowNinja Of course, this is 0.5, and you shouldn't expect perfect compat on master. 05:52 Zeno` they look fine 05:52 Zeno` oh it's for 0.5? 05:52 Zeno` hmm, that does change things a little bit 05:53 ShadowNinja Zeno`: Yes, that's what we're working on now... 05:54 ShadowNinja Zeno`: Does the code seem good though? That's what I want checked. 05:55 ShadowNinja ~tell nrzkt There's actually already a MapSaveThread -- it's called EmergeThread. 05:55 ShadowBot ShadowNinja: O.K. 05:55 Zeno` still going through it 05:56 Zeno` yeah it looks ok... pretty minor code changes really 05:58 ShadowNinja Alright, good. 05:58 ShadowNinja Zeno`: Are you O.K. with the binary rename too? 05:58 Zeno` not really, but it's a small issue honestly 06:00 est31 its breaking scripts 06:01 est31 you should at least provide a symlink to minetestserver 06:01 est31 or from ^ 06:06 ShadowNinja est31: It's a one char fix and you have to deliberately pull latest master. 06:07 est31 I just dont see the bene-fit 06:07 est31 sometimes you write stuff to-ge-ther 06:07 ShadowNinja Anyway, I'm merging #2404 soon. 06:07 ShadowBot https://github.com/minetest/minetest/issues/2404 -- Respect game mapgen flags and save world noise params by ShadowNinja 06:07 est31 you see I'm not the greatest friend of that :) 06:08 ShadowNinja est31: There's no real benefit, it's just the *right* name for it. 06:09 ShadowNinja It's slightly more readable though. 06:11 est31 I understand when a breaking change is so complicated that you can't add a compatibility layer, but here its nearly trivial 06:11 est31 its just a symlink 06:26 Zeno` ShadowNinja, doth though agree that #2452 is ok now? 06:26 ShadowBot https://github.com/minetest/minetest/issues/2452 -- Fix android gettext issue by Zeno- 06:47 * Zeno` pokes ShadowNinja 06:47 Zeno` He's probably hurt himself again 06:47 ShadowNinja Zeno`: No, not right now. Ask me tomorrow when I can review it properly. 06:48 Zeno` I'll probably get someone else to review before that ;) 06:49 ShadowNinja Zeno`: And if you've been having performance problems with the Dummy backend ( ;-) ) try this, it speeds up loadBlock by at least 2x: http://sprunge.us/hieH 06:49 ShadowNinja (well, not counting witngs like caches) 06:49 ShadowNinja things* 06:50 ShadowNinja Night. o/ 06:50 Zeno` night 06:57 est31 night 08:00 est31 Zeno`, #2453 08:00 ShadowBot https://github.com/minetest/minetest/issues/2453 -- Let travis build android by est31 08:11 Zeno` it failed the travis build though! 08:12 est31 yup 08:12 Zeno` 17 minutes... yikes 08:12 est31 its highly unreliable 08:12 est31 https://travis-ci.org/est31/minetest/jobs/53436997 08:12 est31 this one worked 08:13 Zeno` # Download sdk and ndk no wonder it fails sometimes hehe 08:13 est31 the only difference is http://pastebin.com/RXQAr2AA 08:13 est31 the download isn't the problem 08:13 est31 its done after 3 minutes 08:14 est31 main time expensive parts are building of dependencies 08:14 est31 irrlicht clang openssl and so on 08:14 Zeno` hmm 08:15 est31 the problem is some weird internal compiler problem: https://travis-ci.org/minetest/minetest/jobs/53438080#L6884 08:18 est31 and I thought after I got build working, it will remain to work 08:19 Zeno` could it be running out of disk quota? (I wouldn't think so, MT + irrlicht isn't *that* big) 08:21 est31 I've read about disk quota, and they only write that its limited, and that the limit changes between builds 08:22 Zeno` probably because different servers host the builds 08:22 Zeno` but I wonder if Travis could handle compiling something like Qt 08:22 est31 yea that was my thought too. travis hosts projects that are far more larger 08:24 Zeno` I don't even like compiling Qt on my computer hehe 08:24 Zeno` (I disable webkit... makes it a bit more bearable) 09:04 est31 Zeno`, it might be RAM issue 09:05 est31 I'm not sure 09:06 est31 right now I'm running a ram report build 09:06 Zeno` hi 09:06 Zeno` maybe RAM, yes 09:06 est31 https://github.com/travis-ci/travis-ci/issues/1972 09:09 Zeno` well, there's not much that can be done about that :( 09:11 est31 we have root access. its easy to mount a swap 09:12 Zeno` worth a try then I guess 09:13 est31 First I'm finding out whether RAM is the problem 09:13 est31 I let it print out ram usage every 2 secs 09:13 est31 https://travis-ci.org/est31/minetest/jobs/53441245 09:14 est31 not visible yet, as its still installing stuff 09:14 est31 nvm 09:16 est31 ubuntu 12.04 has old free tool 09:16 est31 no -h option yet 09:17 Zeno` -m will do :) 09:17 Zeno` or -g I guess 09:18 Zeno` could sleep 5 && while [ 1 ]; do sleep 1; echo y; done be the problem? 09:21 Zeno` maybe not 09:25 est31 thats to automatically say yes to google sdk license 09:25 Zeno` I meant the sleep, but it's probably not the problem 09:31 est31 mem seems to be no problem: https://travis-ci.org/est31/minetest/jobs/53441523 09:32 Zeno` yikes 09:32 Zeno` yes, I think you're correct 09:33 est31 next try. disk space 09:33 est31 lets hope it knows df -h 09:55 Zeno` great 09:55 Zeno` bash: /usr/local/android-sdk/platform-tools/adb: /lib/ld-linux.so.2: bad ELF interpreter 09:55 Zeno` just great 09:56 Zeno` I've broken it again 10:01 sapier Hello, can someone explain to me why now even default menu size on LINUX seems to be broken? 10:02 Zeno` sapier, can you review https://github.com/minetest/minetest/pull/2452 please? 10:02 sapier I know we had issues for other os every now and then but at least linux 8000x600 did work most of time 10:02 Zeno` I'd like to have a working Android version 10:02 Zeno` and it hasn't been working since before the last release 10:02 kilbith sapier, screenshot incoming 10:03 Zeno` and yes, 800x600 is broken as I mentioned yesterday 10:03 kilbith sapier, https://lut.im/n9J3C070/ARkpB3Kq 10:04 kilbith (800X600 - 96 DPI) 10:04 Zeno` kilbith, and that's broken 10:04 Krock Zeno`, are you sure the "MultiByteToWideChar" function is no longer required? 10:04 Zeno` the menu does not fit 10:04 kilbith Zeno`, ofc 10:04 Zeno` Krock, I just moved it 10:04 sapier same for me kilbith 10:05 Zeno` Krock, I mean all I did was move it.. it's still there 10:05 Zeno` Krock, but if you want to compile and confirm then that would of course be appreciated 10:05 Krock oh I see. 10:05 sapier guys instead of messing around with font size values to fix windows why not just implement the missing dpi detection there 10:05 Zeno` I just want this non-working android build fixed because it's been broken since february 10:06 Zeno` nobody can test the android version because no text shows! 10:06 Zeno` so I need one dev to say "ok it works", merge 10:06 sapier it's insane to try to fix a missing implementation in one part by changing values somewhere else 10:06 kilbith sapier, for recall : 4d744cf8 10:06 Zeno` sapier, surely the fact that there is no working android build is more important right this moment 10:07 sapier link please I don't speak githash 10:07 sapier android doesn't build? 10:07 kilbith https://github.com/minetest/minetest/commit/4d744cf87ad1cd108d125c58559cef4316e80a70 10:07 Zeno` sapier, it builds but there is *no text displayed* 10:08 Zeno` which is why I'd like to merge 2452 10:08 sapier well there's another commit changing font size right next to it 10:08 kilbith sapier, yes 13 to 15 10:08 sapier Zeno if 2452 works merge it it's a little bit to tricky to see from reading code if it works or not 10:08 Zeno` Krock, if you can confirm that 2452 still works on windows that'd be great 10:09 Krock Zeno`, if you want to wait 10 minutes... sure 10:09 Zeno` Krock, cheers 10:09 sapier << checking android build 10:10 sapier actually I wanted to do some mobf development but based on this minetest version ... :-( 10:10 sapier I don't understand how a gui bug like this one can get merged without beeing fixed immediatly who did break this? 10:12 Zeno` If it works for both Krock and sapier I'll merge 10:12 Zeno` I can't check because somehow I fscked up my android build env (again) 10:12 Zeno` I think fedora updates are mucking with my paths 10:14 Zeno` if 2452 doesn't work on android I'll remove the second commit in that PR, because I know it was working with just the first commit 10:14 sapier let me check it on android first 10:15 Zeno` sapier, of course 10:15 sapier I need to know if base version works prior trying your patch ;-) 10:15 Zeno` the first commit in the PR was working before I broke my toolchain 10:15 Zeno` sure :) 10:16 * Zeno` will go outside for a bit to do some thai chi (or whatever it's called) 10:16 sapier ok base version is already broken so no need to check your patch zeno 10:17 Zeno` no need? you've checked the patch? 10:17 sapier do you try to fix the issue? 10:17 Zeno` yes, it should fix the issue. I'll just wait for krock to see make sure it doesn't also break MS build 10:18 Zeno` There are 2 commits in the PR... The first most definately fixes the bug I just hope I didn't break it again with the second 10:18 sapier for what I remember the issue happening is what happens if you try to use android multibyte functions .. I already had that issue. Guess someone wante to be smart and believed the ugly code there was just for fun 10:18 est31 Zeno`, your patch works 10:19 est31 (1f4f3f2a) 10:19 Zeno` ok I shall merge if Krock works as well 10:20 Zeno` sapier, the commit that broke it was called "reduce ugly functions" or something heh 10:20 sapier est31 works on which os? 10:20 est31 android 10:20 Krock string.cpp compiled without any error 10:20 sapier Zeno`: well so I my guess was correct without even knowing what was done 10:20 Krock let's see if the linker likes it too 10:20 Zeno` sapier, maybe :) 10:21 sapier maybe ppl should start thinking why things are done this way instead of believing they're always more smart then the original author 10:23 sapier and even mor silly why did the one doing this "fix" something without even trying it? 10:24 Zeno` I imagine they tried it but not on android 10:24 sapier the ugly code was android only so it'd have been obvious there's something special on android about that part 10:24 Zeno` for a few commits there android doesn't even build 10:24 Zeno` s/doesn't/didn't 10:24 sapier if something doesn't build I can't change things there 10:25 Zeno` which made bisecting a PITA (as est31 can also confirm) 10:25 est31 yup 10:25 Zeno` sapier, est31 is working on a Travis android thingy 10:25 sapier sorry if this seems harsh but it's inacceptable to change things without testing them 10:25 est31 #2453 10:25 ShadowBot https://github.com/minetest/minetest/issues/2453 -- Let travis build android by est31 10:25 Zeno` sapier, I think we agree with that 10:25 sapier of course you can't test everything 10:26 sapier but if something is android only you can only test it on android 10:26 sapier and if android doesn't build you can't merge it by any means 10:26 Zeno` Krock, how long does your linker take? lol 10:27 Krock Zeno`, finished a minute ago. seems to work good 10:27 Krock meh 10:27 Zeno` meh? 10:27 Krock but the special chars aren't shown 10:27 sapier zeno your patch works for me too ... anyone able to test it on windows? 10:27 est31 even on a failing build, disk usage is no issue. https://travis-ci.org/minetest/minetest/jobs/53444462#L7030 10:27 Zeno` let me look, Krock 10:27 Krock Kreativitätsmodus is now Kreativittsmodus 10:28 Krock should I upload the binary? 10:28 Zeno` ok, I'll fix that 10:28 Zeno` no it's ok 10:28 est31 Krock, how did you make it german? 10:28 Krock est31, native language? 10:29 est31 yes mine too also on phone wondering why it doesnt display 10:29 Zeno` Krock, can you see in your git history two commits added by the PR? 10:29 Krock Zeno`, nope, I used the diff file 10:30 Zeno` ok, I'll just revert the second commit 10:30 Krock est31, what isn't displaying? the complete text or öäü-like? 10:30 est31 Krock, ?? 10:30 est31 its english 10:30 sapier is anyone here who is able to test windows? 10:31 Krock sapier, I'm testing it whenever my PC starts. it's my OS 10:31 Zeno` Krock, I've pushed an update... I'll squash the commits when you confirm it now works 10:31 Krock est31, so it doesn't translate? 10:33 est31 Krock, no. 10:33 Zeno` right now I don't care if it's ugly (as long as it works) 10:33 Krock est31, so I don't get your problem if it's a problem 10:34 sapier Krock could you test some dpi detection code for me once it's ready? 10:35 est31 Krock, I just wondered why your build was translated and mine not 10:35 Krock sapier, sure but it takes a while to compile 10:35 Krock est31, that's gettext's fault. it translates the text when minetest is started 10:36 Zeno` Krock, let me know when you've tested again please? 10:36 Krock Zeno`, compiling with the current pull 2452 changes 10:36 est31 Krock, which OS did you test again? 10:36 sapier Krock: first I need to find out how to do it so no problem 10:36 Zeno` Krock, is windows through and through :P 10:37 Krock est31, I'm on Windows 5.1 and the test phase is running since .. some years 10:37 sapier hmmm 10:37 sapier well you should've stayed at 5.0 10:38 est31 Krock, ok. Can you please flush the buffer and paste the stderr? 10:38 Krock sapier, if you mean win 5.0, minetest doesn't run there 10:39 Krock est31, I dno't get what you're speaking of. there's no error 10:41 est31 there is nothing to get :P. Just wondered what you meant with "some years" and "test phase". :D 10:41 est31 was joke :D 10:41 sapier Krock why not? who did break 5.0 ? ;-) 10:42 Krock well, the part "test phase" wasn't meant serious but nvm 10:42 Krock sapier, no idea, I haven't worked on it for 4 years 10:43 est31 like what I said. I thought you tested it on android like me and then wondered that your build was translated but mine not 10:43 est31 but if you are on win, then its sth else 10:43 Krock hmm.. gettext doesn't work on android? 10:45 sapier translation on android? 10:45 sapier did they implement proper i18n recently 10:45 sapier ? 10:45 sapier Krock: android isn't even aware that there are different languages on this planet 10:46 Krock sapier, at least my android phone knows in which language it should display the settings 10:46 sapier krock https://github.com/sapier/minetest/commits/windows_dpi I don't have windows so I can't even test if this compiles 10:47 sapier androids java vm does know it but not the lower linux os 10:47 sapier but gettext ain't java ;-) 10:48 Krock Zeno`, the chars are ok again 10:48 Zeno` ok, squashing and merging 10:48 Zeno` thank you HEAPS to all who helped with this 10:48 Zeno` it's good to finally get it out of the way 10:48 Krock np 10:51 Zeno` done and closing 10:53 Zeno` finally I can work on something else 10:53 Zeno` :D 10:54 Krock compiling windows_dpi commit.. 10:57 sapier windows doesn't seem to know what dpi screen is, but there's a global font dpi setting on windows. I assume ppl who want bigger guis want them throughout windows 10:58 Zeno` I dunno how any OS knows what the screen DPI is :) 10:59 sapier android does 10:59 Krock sapier, porting.cpp(691): error C3861: "SetProcessDPIAware": unknown identifier 10:59 Zeno` yeah android does 10:59 sapier hmm guess some header is missing 11:00 Krock also porting.cpp(701): error C2589: '(': invalid Token on the right side of '::" 11:00 sapier ohhh 11:00 sapier nope it's windows 6.0+ 11:01 sapier krock your os is too old ;-) 11:01 Krock sapier, nah. find an alternative 11:02 sapier well not sure if there is any ;-) 11:03 Krock lol. you copied the codes from stackoverflow 11:04 sapier partial ;-) 11:04 * est31 hides 11:05 sapier well all of it is windows api there ain't much you can change 11:05 sapier if a function is called foo you can only use it by calling foo() 11:08 sapier well there doesn't seem to be a solution for winxp at least nothing I can find within reasonable time ... and as winxp is out of support I'll not bother spending a lot of time for it 11:11 Krock check if the windows version is > 0x0501 and then you can use that code. 11:11 Krock I still can use the settings 11:11 sapier oh you mean just leave winxp as it is? 11:11 sapier possible but we'd still need someone to test if this helps on vista+ 11:18 Zeno` anyone care to review #2362? 11:18 ShadowBot https://github.com/minetest/minetest/issues/2362 -- Change uses of assert that are meant to persist with a different macro by Zeno- 11:20 sapier Zeno`: for what reason? 11:20 est31 assert is used wrongly 11:20 Zeno` so I can merge ;) 11:21 Zeno` assert is meant to go away when NDEBUG is defined. Using it as a lazy way of error checking is... dumb 11:21 sapier est31 cleaning up the incorrect usage is ok but why add additional asserts? 11:21 Zeno` I haven't added any additional ones 11:21 sapier noone except zeno is gonna use it 11:21 Zeno` I made a new macro for the lazy... that's all 11:22 Zeno` I'm not going to spend 6 months doing it another way 11:22 est31 clang even warns at some places that variables are set but not read. those variables are read by assert. 11:23 Zeno` because I'd never be able to keep up with the required and constant rebasing for a start 11:23 sapier hmm seems those strange additional asserts have been defined before 11:24 sapier map.cpp L770 inconsistency 11:25 sapier map.cpp L2863 possible functional change 11:26 sapier ..L3371 why not fatal_error? 11:26 Krock Zeno`, https://github.com/Zeno-/minetest/commit/828f9dd7#diff-34f48ad91ac6c202ac60b0348ae90e30R711 -> FATAL_ERROR_IF() ? 11:26 sapier mapnode.cpp L 74 FATAL_ERROR? 11:27 Zeno` those asserts() (L770) are okay to go away when NDEBUG is defined 11:27 Zeno` no, that's not a fatal error 11:27 sapier ok 11:27 Zeno` L74 in mapnode is a correct use of assert 11:28 Zeno` Krock, I can change that 11:28 Zeno` the point is it's a whole lot better than redefining assert so it behaves in a manner that isn't even consistent with the manpage decscription 11:29 Zeno` The question I had to ask all through that exercise was "is this a correct usage of assert() or lazy error checking" 11:29 Zeno` and it's not fun 11:30 sapier maybe ;-) I usually do assert for things I believe never gonna happen at all 11:31 Zeno` line 770 changed 11:31 Zeno` sapier, and that's the correct way to use it.. for things that will never happen at all 11:31 Zeno` things that can't happen if the logic is correct 11:31 Zeno` it's not how assert() is used before the PR :p 11:31 sapier well some of them look like regular errors too 11:31 Zeno` they do... 11:32 Zeno` look like regular errors* 11:33 Zeno` can you tell the difference? That was what was so difficult. Did the original person mean assert (as meant to be used) or as lazy error checking 11:33 Zeno` because redefining assert() to behave in a way it's not meant to behave is... well, it's horrible :( 11:34 sapier your changes seem to be fine, I doubt they'll change anything in future but cleaning up every now and then ain't wrong for sure 11:34 Zeno` There may be some that I got wrong but I erred on the side of caution 11:35 Zeno` so there are some sanity_check(....) ones that are really meant to be assert 11:35 Zeno` but *shrug* I think it's better than before 11:36 sapier at least not within the code I wrote, I can't tell what others intended with their asserts 11:36 Zeno` they'll change in future because it will be added to the dev wiki "do not use assert for error checking" and all devs will be made aware 11:36 sapier well look at how commits have been added the last weeks do you believe anyone really follows the rules? 11:37 Zeno` I think most of your code used assert() properly... hard to tell (there were 900 or something that I had to check) 11:37 Zeno` well, if someone uses assert incorrectly it'll segfault soon enough in a release build 11:37 sapier nope most of my asserts would've been sanity checks 11:37 Zeno` and we can yell at them 11:37 Zeno` sanity checks... there's a macro for that 11:37 Zeno` lol 11:38 sapier to be honest I don't even care about beeing yelled at anymore ;-) if I did I couldn't participate here as you're allways yelled at for anything ;) 11:38 Zeno` they can probably be assert though... sanity checks usually are meant to be assert() 11:38 Zeno` it's really only for error checking that assert is not appropriate 11:38 Zeno` so sanity_check() is probably deprecated before it's even born 11:39 sapier wait 11:39 sapier TextureSource::generateImagePart( 11:39 sapier the android part those should really be asserts not sanity checks 11:40 sapier worst to happen is a missing texture we shouldn't crash minetest for this 11:40 Zeno` those are some where I erred on the side of caution 11:40 Zeno` assert() _does_ crash (abort) though 11:40 sapier yes but assert is removed in release 11:41 Zeno` even the re-implemented broken one did 11:41 Zeno` ok, if they're really assert I will change 11:41 Zeno` these are the ones where I erred on the side of caution 11:41 Zeno` which file? 11:41 sapier you should detect those things in debug builds but on release they'll not cause anything bad except of missing textures 11:41 Zeno` correct 11:41 sapier tile.cpp L1407 11:42 Zeno` but in release builds (currently) they will cause something bad ;) (using assert) 11:42 sapier on some android devices not even textures would be missing 11:42 Zeno` because assert was re-defined 11:42 Zeno` changing them to assert now 11:43 sapier well ... not what I intended but as long as the code fixing non npot2 textures works nothing s gonna happen 11:43 Zeno` and that is the whole reason for the PR... asserts currently persist in release builds 11:43 Zeno` when they shouldn't 11:43 sapier it's in there because some android devices can handle non npot2 textures some cant 11:44 Zeno` ok fixed 11:45 Krock the inventory is kinda broken with current HEAD 11:45 Krock when I press a button in the unified inventory, the formspec doesn't update 11:46 sapier unified inventory bug? 11:46 Krock nope, it works fine with a 5/6 days older client 11:46 sapier strange 11:49 Krock false alert. I timed out right after I opened the inventory :3 11:50 Zeno` anyway if I can't merge it soon we're going to be back to square one where we don't know if the programmer meant assert() or meant "this is an error" 11:51 Zeno` like: https://github.com/minetest/minetest/pull/2362/files#diff-4f2cf5d1d78044ecd5922286ee4d371bR2736 11:51 Zeno` at least that one was commented 11:56 Zeno` will do some more testing with the plan of merging soon(ish) 11:56 sapier https://github.com/minetest/minetest/pull/2132 can someone check this on android? it' getting rid of the button mess around android screen 12:05 Zeno` so 2363 ok to merge now? 12:05 Zeno` err #2362 12:05 ShadowBot https://github.com/minetest/minetest/issues/2362 -- Change uses of assert that are meant to persist with a different macro by Zeno- 12:06 Zeno` stupid keys are mixed up on my keyboard (honest) 12:23 nrzkt hello, i will push a little mathematic algo improvement on Cloud::render() 12:24 Zeno` may we see it? :) 12:25 nrzkt i noticed we have a bug when we are in a cloud, we don't see we are in, i think it miss a thing like be in water 12:25 Krock just increase the fog in those clouds :) 12:26 nrzkt maybe, if you know how, don't hesitate to provide a patch, i don't have time to look at this ^^ 12:30 Zeno` nrzkt, I fixed your comments relating to 2362 and also some that Krock and sapier raised. Ok to merge now? 12:30 nrzkt #2362 is ok 12:30 ShadowBot https://github.com/minetest/minetest/issues/2362 -- Change uses of assert that are meant to persist with a different macro by Zeno- 12:30 Zeno` ok I shall squash. Thanks 12:33 Zeno` squash and merge 12:33 Zeno` will wait for travis 12:34 Zeno` Travis is slow... est31 is probably flooding it 12:36 est31 lol 12:36 est31 no activity by me 12:36 est31 I made a large jam 12:37 est31 now all FOSS community is waiting at each other :D 12:38 Zeno` :P 12:40 Zeno` merging now 12:44 nrzkt ShadowNinja, i see then. But EmergeThread isn't right if it's goal is only save :p 13:18 kilbith [Alert] Minetest fails to run on current master : http://pastie.org/10007209 13:21 kilbith sapier : http://irc.minetest.ru/minetest-dev/2015-03-07#i_4175997 13:22 Krock there was no change in init.lua for 4 months 13:22 kilbith no the cause do not come from that file 13:23 kilbith that's only the consequence 13:32 sapier this looks like completely messed up search paths 13:33 sapier /../fonts/liberationsans.ttf ?? well most likely there ain't a directory below root ;-) 13:33 kilbith git pull, compile and see by yourself 13:34 sapier well I'm only 3 commits behind but I'll try to be sure ;-) 13:35 kilbith it was fine this morning 13:35 * kilbith looks dubiously at Zeno's commits 13:36 sapier latest --> worksforme 13:37 kilbith wtf, rm -rf the whole folder, git clone, etc. not works for me 13:39 sapier ok another check with fresh clone 13:39 celeron55 note that most likely there will be a difference when toggling RUN_IN_PLACE 13:39 kilbith sapier, when you're on /bin, it needs to ../ for access to fonts/ 13:40 celeron55 but it should work with both settings; there is enough autodetection for things for it to work 13:40 celeron55 assuming this is on linux 13:40 sapier I can only check RUN_IN_PLACE because I'm not messing up my distro with manual installed files 13:41 sapier ok doesn't work with fresh clone so it's something within cmake files 13:41 kilbith yep, thanks 13:42 sapier most likely this error is older because it won't be recognizeable without doing a fresh clone 13:42 kilbith it worked fine this morning with latest paramat's commit 13:43 sapier cmakefiles haven't been changed today 13:43 kilbith *shrug* 13:43 kilbith i just pulled this morning and it worked fine 13:43 kilbith this is all i can say 13:44 sapier well that'd be easy to check give me a few minutes 13:46 manax Just pulled and rebuilt, and I get similar error ERROR[main]: "/../fonts/lucida_sans" 13:47 manax And "cannot open /../builtin/init.lua: No such file or directory" 13:49 sapier b65c is still fine 13:50 kilbith ffdf8de too 13:50 kilbith at least on my side 13:52 sapier that's quite curious none of the remaining commits change anything for cmake 13:52 sapier maybe it's an old bug not beeing visible due to another bug 13:52 kilbith "Fix Android text bug" <- related with fonts 13:53 sapier not really it's fixing the multibyte coversion someone believed to fix without even testing her/his fix 13:53 sapier and the issue ain't fonts as init.lua isn't found too 13:54 sapier not meaning this commit isn't a candidate for the issue 13:54 sapier a603 works too 13:54 sapier seems like the assert fix is it 13:56 kilbith can't see how/where 13:57 sapier me too, and why does it only break for new clones 13:57 kilbith nope, on pulling too 13:57 kilbith ^ another report above 13:58 kilbith (manax) 13:58 manax yep 13:59 sapier not for me 13:59 sapier I'm on this version and it still works 13:59 sapier well despite the broken menu of course 14:01 manax I did a pull and cmake with run in place, and make, which rebuilt everything (it looks like). I was on b65c, previously (i think).. 14:03 kilbith sapier, you ca revert that commit until a fix ? 14:03 kilbith *can 14:03 sapier without knowing what's actual reason? see porting.cpp L 423 ... strange assumption done there 14:04 kilbith we don't know the exact reason but this commit is problematic for sure 14:05 sapier let's first try to find the issue prior removing something that doesn't actually look like beeing the real error 14:06 kilbith yes but i feel we'll have for a moment to find out the bug 14:06 kilbith ppl can't play at all atm 14:06 sapier dev is not meant for regular ppl to be used for playing 14:07 kilbith usually it's reverting when we're confronted to a real nasty bug 14:07 sapier since when? 14:07 kilbith nvm 14:07 sapier revert is only an option if it can't be fixed within reasonable time 14:08 kilbith it's simply common sense ... 14:08 sapier right now we don't know if fix is a one line change we'll detect in 5 min 14:08 sapier if we wouldn't have to discuss ;-P 14:09 sapier it's always better to fix something instead of reverting it 14:09 kilbith depends of the case 14:10 kilbith if you may fix that quickly or not 14:10 sapier I see no exception of this rule 14:10 sapier if you'd insist on dev always beeing correct we'd have to add a dev_dev branch .... till ppl use the dev_dev branch 14:11 kilbith that branch is meant to unstable, not completely unusable 14:11 kilbith to be* 14:12 sapier same thing just different interpretations of unstable 14:12 kilbith i'm not going to talk semantic either 14:13 sapier I still haven't found a way to break my existing clone so this seems to hurt only under some circumstances if you guys could find out what those are instead of discussing we'd be very close to a fix 14:13 manax It very much seems like the path_share, in porting.cpp. Might be platform specific. I'm linux. 14:13 sapier me too manax 14:14 sapier quick n dirty fix switch to debug build mode that's the difference 14:14 sapier guess something is done within code beeing blanked on release 14:15 sapier ok found 14:15 sapier assert(readlink("/proc/self/exe", buf, BUFSIZ-1) != -1); 14:16 sapier fixing in a few seconds 14:16 manax heh, that was my working guess too. 14:19 sapier I'm checking if there's another similar issue prior pushing the fix 14:19 manax But from what I could see, that wasn't changed in the assert checkin... 14:19 Zeno` Does Ubuntu 12 support C++11? 14:21 Zeno` well there ya go 14:23 sapier kilbith: pull latest it should work again 14:24 kilbith 'k 14:24 sapier Zeno`: we already decided we'll switch to C++11 once next debian version is released 14:25 manax confirmed fix worked. 14:25 kilbith sapier, it works now, thanks. 14:25 Zeno` oh, typo! 14:25 Zeno` it should have been != not == 14:25 sapier nope 14:26 sapier you did enable assert blanking in release build with your commit 14:27 Zeno` oh, I missed it? 14:27 Zeno` i see. 14:27 Zeno` well that's another example of assert() that was being used for lazy error checking 14:28 Zeno` the sooner they're all fixed the better 14:28 Zeno` thanks for that one, sapier 14:28 Zeno` we may as well get all the issues "out in the open" now 14:29 Zeno` because the state of things before my commit was, quite honestly, not good 14:30 Zeno` e.g. that assert() that sapier fixed was really meant as an error 14:30 Zeno` but.. well, assert is not used for error checking 14:30 Zeno` so, *shrug* 14:31 Zeno` another incorrect assert usage fixed! 14:31 Zeno` we should party 14:34 Zeno` like it's 1999! 14:35 Zeno` err, maybe 1999 has already occurred 14:40 kilbith sapier, you intend to add the dpi detection on Windows and remit and dynamic size on !Android then? 14:40 kilbith -and 14:40 Zeno` I think we need to fix the 800x600 not working on *any* build as well 14:42 Zeno` especially since it's the default window size :/ 14:42 kilbith even in fullscreen the tabs are slighty going over the tab 14:42 kilbith the header, sorry* 14:43 Zeno` yeah the header does 14:43 Zeno` the logo* 14:43 Zeno` header... whatever it's called 14:44 kilbith that dynamic sizing was good before setting it static 14:44 kilbith we just need to consider the dpi on Windows 14:45 Zeno` there has already been a HUGE performance regression (now fixed) due to DPI probing 14:46 Zeno` I hope there is not another introduced when trying to fix it 14:47 Zeno` maybe sapier would like to see https://github.com/minetest/minetest/commit/9a0dd47057cad18522e43d159a06c856dfdeeef4 :P 14:50 sapier Zeno not sure if this is right way to do it 14:51 sapier wait it's x11 only 14:52 sapier I'd suggest making it a single function, or do calcDisplaDensity as static inline ;-) 14:52 Zeno` sapier, it was closing and opening X every time getDisplayDensity() was called 14:52 Zeno` fps went from about 60 to about 20 14:52 sapier ok x11 internal issues 14:52 Zeno` no, minetest issues 14:53 sapier yes but this beeing so slow is an x11 issue 14:53 Zeno` you were calling XCloseDisplay(x11display); every call to getDisplayDensity() 14:53 sapier you could most likely use a similar fix on android 14:53 sapier but I don't know how long android calls take compared to x11 14:54 Zeno` anyway, I just point it out because I think that of course regressions will happen when major changes occur 14:54 Zeno` we shouldn't revert... we should just fix 14:54 sapier to be honest I didn't even know x11 opening and closing are that heavy calls ;-) 14:54 Zeno` oh, they're huge :) 14:55 Zeno` very expensive :( 14:55 Zeno` hmm 14:55 Zeno` does android use the same method to get the dpi? 14:55 Zeno` nope 14:55 Zeno` (sorry, answered my own question heh) 14:57 sapier no it doesn't but I'm quite sure it doesn't cache the result too 17:06 manax I've been playing with adding an on_touch callback to nodes. I've got it working for a player (kinda) by adding some code to ServerEnvironment::step() but want it for SAO instead, I think... 17:06 Warr1024 you mean like when an entity collides with a node? 17:07 manax yeah. 17:07 manax I'd really like mobs to be able to trigger the on_touch as well, which I think means SAO. 17:07 Warr1024 There have been times when I've wanted an event like that 17:08 Warr1024 I ended up resorting to polling, i.e. setting a node timer or abm and just scanning for entities within a radius 17:08 est31 there is too much polling inside minetest 17:08 manax Yeah, sounds like the TNT mod i looked at, which doesn't seem like the right way to do it... 17:10 manax My current change works, although on_touch gets called continuously, rather than once per node-touch. 17:10 Warr1024 that's not really all that bad...? 17:10 manax But as I said, I think it really belongs in sao... 17:10 Warr1024 I guess, I mean, a Certain Other Mining-Themed Game has the same callback and it works like that, IIRC. 17:11 Warr1024 I was actually disappointed to find that MT didn't provide an API for reading an entity's bounding box, to be able to do my own collision checks, either 17:11 Warr1024 ...though that was some time ago, so maybe it's been added since? 17:12 manax Dunno. I'm really new to MT. 17:12 Warr1024 for instance, if you calculate the damage from an AOE effect like an explosion, it's generally based on a certain "reference point" on the entity, which I think is like the bottom center. 17:12 ShadowNinja object:get_luaentity().collision_box 17:13 Warr1024 I wanted taller entities to be able to receive damage along the top too 17:13 Warr1024 ShadowNinja: IIRC that doesn't work with players, though...? 17:13 ShadowNinja Might be collisionbox. 17:13 Warr1024 I'll have to look into that 17:13 ShadowNinja Warr1024: Yes, that one's hardcoded. 17:13 Warr1024 maybe for players I can just guess the box, since it's not awfully common to change that...? 17:13 Warr1024 cool 17:13 Warr1024 thanks for the tip 17:14 ShadowNinja You're welcome. 17:28 manax If I have an SAO, and the player goes too far away, does it get deleted? If so, is there away to keep it around "permanently"? 17:30 ShadowNinja manax: No, it's saved into the mapblock. And you can use forceload_block. 17:32 manax Thanks. 17:32 manax I'm struggling a bit with division of responsibility between the client and server... :P 18:00 est31 ShadowNinja, what exactly is blocking #1606 right now? 18:00 ShadowBot https://github.com/minetest/minetest/issues/1606 -- Add mod security by ShadowNinja 18:03 ShadowNinja est31: Nothing that I know of, I've just got to merge it. 18:03 est31 that's fine 18:07 T4im hehe, can we have a minetest.mkdir() then? I think creating directories in the worldpath is quite a frequent use of os.execute in common mods, that could be avoided… :D 18:08 est31 isn't there one in that patch? 18:08 T4im tbh I have not finihsed reading that huge pr yet… 18:08 est31 yup: https://github.com/minetest/minetest/pull/1606/files#diff-e62796adc1d9474bfa497540cf203d47R333 18:08 T4im well.. awesome then :D 18:09 est31 haven't fully read it either 18:09 T4im hmm quite possible that the settings related issues might be fixed with it too then D: 18:09 T4im :D 18:10 est31 today was mostly dedicated towards android travis ... with the result that travis hardware isn't enough. 18:10 T4im O_o 18:10 est31 I *could* "fix" it by doing make || make || make || make || make or o 18:10 est31 or so* 18:10 est31 but that would be ultra ugly hack 18:11 est31 #2453 18:11 ShadowBot https://github.com/minetest/minetest/issues/2453 -- Let travis build android by est31 18:11 Warr1024 est31: for X in 1 2 3 4 5; do echo attemp $X; make && break; done # :-) 18:12 Warr1024 There's a "make || break" joke in there somewhere. 18:12 est31 yea lol :D 18:12 T4im hmm aren't there make drop in replacements for that or so? 18:12 Warr1024 any idea what the most memory-hungry part of the process is? 18:13 T4im something in the direction of distcc… but not quite 18:13 Warr1024 also, I wonder if you can offload parts of it 18:13 est31 read the log of one of the failing builds 18:14 T4im maybe you get ccache working with travis? 18:15 est31 ccache is mostly broken for android 18:15 est31 travis doesnt work that way either 18:15 est31 its a "hey this is a fresh box with some build tools and git installed, lets clone the project and apt-get all dependencies" approach 18:15 est31 very clean 18:16 T4im well all CI systems do that to some degree… but sometimes you can cache stuff outside of those clean instances… not sure how much travis offers though 18:16 est31 but problem ofc that builds need ages as its building dependencies that get shipped, including irrlicht curl and friends 18:16 est31 travis offers some way of caching yes 18:17 T4im hmm ccache… not yet: https://github.com/travis-ci/travis-ci/issues/3061 18:18 T4im but eventually might be added too 18:18 T4im oh that's already added 18:19 T4im :D 18:19 est31 paralellism of android build is horribly broken too 18:19 est31 submake instances don't report to central jobserver, as they all get -jN arguments fed 18:52 WarrTab it's kind of a hack right now 18:53 WarrTab It'd be kind of nice if they honored the -j option specified by the user instead of overriding it 18:56 T4im well only they know how many cores are available, if they are flexible about it though 18:58 WarrTab The makefile tries to use some shell magic to get the core count 18:59 WarrTab Still, I'd like to be able to manually specify, esp. make -j1, i.e make --actually-work --no-race-conditions-please 19:00 WarrTab I'd think that non-parallel build should be default, with instructions on how to do parallel... 19:01 WarrTab It is a bit odd how much the makefile for android tries to do... even mt itself requires more steps 19:43 ShadowNinja Ugh, 81 notifications. 19:44 ShadowNinja Just for the engine. 19:46 ShadowNinja Merging #2404. 19:46 ShadowBot https://github.com/minetest/minetest/issues/2404 -- Respect game mapgen flags and save world noise params by ShadowNinja 19:48 ShadowNinja Also, I'll start squashing and rebasing #1606 to merge. 19:49 ShadowBot https://github.com/minetest/minetest/issues/1606 -- Add mod security by ShadowNinja 22:43 ShadowNinja Mod security rebased and squashed here: https://github.com/ShadowNinja/minetest/tree/security-2 22:44 ShadowNinja I'll merge it in a day or two unless something comes up. 22:45 VanessaE ShadowNinja: see #minetest 22:45 VanessaE probably won't conflict, but better to consider the recent discussion there 22:49 ShadowNinja It doesn't touch and network stuff, although it does touch the std::list -> std::vector change. 22:50 VanessaE some of your other, recent commits are affected by the PR that he referenced in the discussion is more of why I mentioned it, really 23:44 ShadowNinja Merging #2458 and #2431. 23:44 ShadowBot https://github.com/minetest/minetest/issues/2458 -- Add more android build files to gitignore by est31 23:44 ShadowBot https://github.com/minetest/minetest/issues/2431 -- Update minetest.desktop by mahmutelmas06