Minetest logo

IRC log for #minetest-dev, 2016-07-08

| Channels | #minetest-dev index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
00:49 Void7 joined #minetest-dev
01:29 Grandolf joined #minetest-dev
01:30 halt_ joined #minetest-dev
01:35 STHGOM joined #minetest-dev
01:53 Lunatrius joined #minetest-dev
02:33 Void7 joined #minetest-dev
02:46 paramat joined #minetest-dev
03:34 ptv joined #minetest-dev
03:44 est31 joined #minetest-dev
04:02 ssieb joined #minetest-dev
04:20 Taoki[laptop] joined #minetest-dev
04:37 DI3HARD139 joined #minetest-dev
04:48 est31 https://github.com/est31/minetest/commit/dcd4dfb3f4538a0a6905c58ebbb3ab532ef69eef
04:48 est31 PTAL
04:48 est31 hmmmm, paramat sofar ^
04:49 est31 will fix #4137
04:49 ShadowBot https://github.com/minetest/minetest/issues/4137 -- gcc 6.1 warning on sqlite with c++11
04:52 hmmmm ogh
04:52 hmmmm throwing exceptions in destructors
04:52 hmmmm who thought that was a bright idea
04:52 hmmmm yes, approved, +1
04:52 RichardTheTurd joined #minetest-dev
04:53 hmmmm actually no nvm
04:53 est31 ?
04:53 hmmmm i would prefer it if you got rid of FINALIZE_STATEMENT
04:54 hmmmm who coded it like that with macros on top of macros?  that's not making code clearer
04:54 hmmmm christ
04:54 paramat left #minetest-dev
04:54 hmmmm god dammit sn
04:55 est31 well, I do think it makes stuff clearer
04:55 est31 helps to avoid repetition
04:55 hmmmm how is sqlite3_finalize(statement) too much work
04:55 hmmmm there's no repetition there if we don't want to throw an exception in the first place
04:55 hmmmm just do something like
04:55 est31 we still want to log an error
04:56 hmmmm bool success = false;
04:56 hmmmm err
04:56 hmmmm = true;
04:56 hmmmm success &= sqlite3_finalize(statement) != SQLITE_OK;
04:56 hmmmm so on and so forth
04:57 est31 i dont think the code is much clearer if the destructor uses a different setup than the rest of the file
04:57 hmmmm what do you mean?
04:57 est31 and for the rest of the file, throwing exceptions makes sense
04:57 est31 hmmmm, well if the destructor doesnt use the macros
04:57 hmmmm the destructor is literally the only place where FINALIZE_STATEMENT gets used
04:57 est31 ofc, the FINALIZE_STATEMENT macro is unique to the destructor
04:58 est31 but SQLOK, which FINALIZE_STATEMENT uses is not
04:58 hmmmm either way I don't see an issue with modifying the code so no exceptions ever get thrown
04:58 est31 plus if you use success bool, then you lose information about what exactly failed
04:58 hmmmm so then why not revert FINALIZE_STATEMENT to the way it was before ShadowNinja changed it?
04:59 hmmmm (or should i say "cleaned it up" :)
04:59 est31 okay
04:59 nrzkt joined #minetest-dev
05:00 hmmmm right
05:00 hmmmm https://github.com/est31/minetest/commit/708337dfc2b3871dc6de983e781e4a4a60a1881d#diff-eacf226572fc4ceb2bbdb0874bb8afe6L290
05:00 hmmmm i didn't think you cared about preserving information about which statement failed, because that's what your commit already does anyway, no?
05:00 est31 yes
05:00 est31 why?
05:01 est31 hmmmm, it should print e.what on an error case
05:01 est31 no?
05:01 hmmmm ahh nvm
05:01 hmmmm i see it concatenates the statement with the error text
05:01 hmmmm in sqlok
05:08 est31 okay what about this https://github.com/est31/minetest/commit/483cca78a1312aa3b9c01049b5ebf6842929ec7a
05:09 est31 hmmmm, ^
05:10 hmmmm yeah
05:10 hmmmm +1
05:14 est31 pushed
05:16 Hunterz joined #minetest-dev
05:17 est31 hmmmm, any ideas where this warning can come from: http://pasteboard.co/8SpYwmXL1.png
05:17 est31 it unfortunately doesnt print the line
05:19 hmmmm wtf
05:19 hmmmm hrmm
05:19 hmmmm it could be from an inlined call to index()
05:21 est31 damn the sqlite code was duplicated in the rollback manager code
05:24 est31 i only have gcc 6 in a virtual machine, not my main os
05:29 est31 #4300
05:29 ShadowBot https://github.com/minetest/minetest/issues/4300 -- GCC 6 warning in treegen
05:30 est31 hmmmm, PTAL https://github.com/est31/minetest/commit/9db78dc4c4130c1ef08696e5d381a157cb66e656
05:30 est31 sorry that I didnt do it in one commit, but didnt saw the rollback warnings
05:31 est31 (my vm is very slow)
05:31 hmmmm thumbs up
05:31 hmmmm what are you testing?
05:31 est31 gcc 6.1
05:31 est31 or minetest, compiled with that compiler
05:31 est31 didnt actually start minetest yet, the vm has no x running
05:32 nrzkt est31, minetest works well with gcc 6.1
05:34 est31 that's good
05:37 OldCoder Has anybody yet build MT for Android 6?
05:37 OldCoder And had it work?
05:37 OldCoder built *
05:38 est31 OldCoder, what do you mean
05:38 est31 it works great on android 6
05:38 est31 or do you mean including setting the target API
05:40 OldCoder est31, Hi. I spent most of the day yesterday on Android MT. Yes, target API.
05:40 OldCoder When you build it for Android 6, what API do you use? 4 or 5?
05:40 est31 4 and 5 are very old, no?
05:40 OldCoder See question
05:40 OldCoder What is it that you do?
05:40 OldCoder Build for 4? 5? 6?
05:40 est31 https://source.android.com/source/build-numbers.html
05:41 OldCoder R
05:41 OldCoder Hm? Nvm that
05:41 OldCoder See question
05:41 OldCoder What is it that you do?
05:41 OldCoder Build for 4? 5? 6?
05:41 * OldCoder has put some effort into network and other debugging
05:41 est31 targetSdkVersion 9
05:41 est31 means android 2.3
05:42 est31 I still dont really understand what this targetSdkVersion thing is all about
05:42 OldCoder Hm
05:42 * OldCoder used the default files which said that... but with Android 6 SDK
05:43 OldCoder Crashes in moments upon network connect
05:43 est31 that should make no difference
05:43 est31 does it?
05:43 OldCoder Same APK works, finally, on a cheap tablet
05:43 OldCoder Seems to
05:43 OldCoder Need to experiment further
05:43 OldCoder For cheap Androids
05:43 OldCoder Must set timeout far higher than usual
05:44 est31 hmm you are right it crashes
05:44 OldCoder Oh? Thanks for testing
05:44 OldCoder Which APK did you try?
05:44 OldCoder Or which build type?
05:44 est31 self compiled
05:44 OldCoder Which build type?
05:45 OldCoder SDK, etc.
05:45 * OldCoder should clarify; the tablet he tried was Android 5 and ran his APK... after timeout change
05:45 OldCoder But under Android 6... belly up for same APK
05:51 est31 hmm i am trying it again, and it stops crashing
05:51 est31 weird
05:52 OldCoder Which build type? SDK, etc. ?
05:52 est31 no idea
05:52 OldCoder You self-compiled, though?
05:52 est31 it should be the default type
05:52 est31 yes
05:52 OldCoder What SDK are you using?
05:52 OldCoder I.e. what is installed?
05:53 est31 but I installed the sdk one day and forgot about it
05:53 est31 including which version it is
05:53 OldCoder OK
05:53 OldCoder It appears that Android 6 no longer works
05:53 est31 can you get some meaningful output from logcat?
05:53 * OldCoder can try in the medium term
05:53 * OldCoder needs to put Android aside for a week or so
05:54 * OldCoder tried to boot 2 different Android 6 X86 ports to do tests; they did not work
05:55 est31 OldCoder, you know about adb?
05:58 nrzkt don't include targetAPI v24 because we need to handle permissions differently, limit it to 23 atm
05:59 nrzkt v24 change many permissions , they are asked at runtime not install time
05:59 est31 isnt it 23?
05:59 est31 not 24
05:59 nrzkt i don't remember, just do a warn :)
06:07 OldCoder est31, yes I used adb 2 weeks ago to reflash my new phone
06:07 est31 OldCoder, well adb should give you a way to get the logcat from the device
06:07 OldCoder It's 23 presently... but for some reason Android 6 does not work
06:07 est31 adb logcat
06:07 OldCoder est31 yes I may experiment with that
06:07 OldCoder Got caught up in it yesterday
06:08 OldCoder Just wanted to know if people had successfully built for Android 6
06:23 T4im joined #minetest-dev
06:24 lisac joined #minetest-dev
06:31 T4im est31: btw #4284 doesn't need the wip label anymore
06:31 ShadowBot https://github.com/minetest/minetest/issues/4284 -- Travis: Skip installation and compilation for non-C++ related changes by t4im
06:42 DI3HARD139 joined #minetest-dev
06:57 edgrey joined #minetest-dev
06:58 nrzkt joined #minetest-dev
07:16 Taoki[laptop] joined #minetest-dev
07:25 Krock joined #minetest-dev
08:26 davisonio joined #minetest-dev
08:47 Megal joined #minetest-dev
09:20 Darcidride joined #minetest-dev
09:24 davisonio joined #minetest-dev
09:31 davisonio joined #minetest-dev
10:15 tenplus1 joined #minetest-dev
10:16 tenplus1 hi folks... just updated pull game#965 to disable bones or drop as items, could a dev please review
10:16 ShadowBot https://github.com/minetest/minetest_game/issues/965 -- Ability to Disable bones, drop items or keep inventory. by tenplus1
10:22 davisonio joined #minetest-dev
10:25 Krock joined #minetest-dev
10:28 Darcidride_ joined #minetest-dev
10:34 tenplus1 left #minetest-dev
10:43 davisonio joined #minetest-dev
10:45 davisonio joined #minetest-dev
11:03 Zeno` joined #minetest-dev
11:07 AnotherBrick joined #minetest-dev
11:25 lisac joined #minetest-dev
11:34 Krock2 joined #minetest-dev
11:57 Fixer joined #minetest-dev
12:28 rubenwardy joined #minetest-dev
13:19 STHGOM joined #minetest-dev
13:21 Lunatrius` joined #minetest-dev
13:34 STHGOM joined #minetest-dev
14:12 Gundul joined #minetest-dev
14:29 est31 joined #minetest-dev
14:31 ElectronLibre joined #minetest-dev
14:44 rubenwardy #4294 #4295 #4297
14:44 ShadowBot https://github.com/minetest/minetest/issues/4294 -- Add container[] and container_end[] formspec elements by rubenwardy
14:44 ShadowBot https://github.com/minetest/minetest/issues/4295 -- Add enter_button[] to press button on enter key by rubenwardy
14:44 ShadowBot https://github.com/minetest/minetest/issues/4297 -- Add custom wield item animation using linear keyframes (WIP) by rubenwardy
14:44 rubenwardy for the last one, I'd like to know if the basic method is correct and worth continuing
14:45 rubenwardy first two are complete imo
14:45 nore rubenwardy: haven't looked at the pulls yet, but what the second implements is something I wanted to have since a long time
14:47 T4im joined #minetest-dev
15:00 Hunterz joined #minetest-dev
15:01 proller joined #minetest-dev
15:03 rubenwardy nore: yeah, it's something that really bugs me about the creative inventory
15:03 rubenwardy once game#1179 is merged, I'm going to improve the creative inventory further
15:03 ShadowBot https://github.com/minetest/minetest_game/issues/1179 -- Add simple fast inventory by rubenwardy
15:08 tenplus1 joined #minetest-dev
15:08 tenplus1 hi folks
15:09 Megal_ joined #minetest-dev
15:10 tenplus1 I've had reports that saplings arent growing in latest daily builds of minetest...  placed sapling on dirt, light source beside, waited 5+ days now and still nothing
15:12 est31 gonna merge this in 10 minutes: https://github.com/est31/minetest/commit/27905f1a4a87a284e95673eeee2440a806b15d16
15:13 tenplus1 also https://github.com/minetest/minetest_game/pull/1186
15:17 hmmmm joined #minetest-dev
15:18 tenplus1 hi hmmmm
15:20 blaze joined #minetest-dev
15:23 ptv joined #minetest-dev
15:23 whitephoenix joined #minetest-dev
15:25 Void7 joined #minetest-dev
15:25 rubenwardy joined #minetest-dev
15:32 damiel_ joined #minetest-dev
15:39 rubenwardy hmmmm, you commented on this issue a while ago on IRC, what do you think of this PR? #4295
15:39 ShadowBot https://github.com/minetest/minetest/issues/4295 -- Add enter_button[] to press button on enter key by rubenwardy
15:39 rubenwardy http://irc.minetest.ru/minetest-dev/2013-03-03#i_2910252
15:42 rubenwardy Problems: 1. Can't have different button presses for each text field  2. May be useful to submit enter press and have no button 3. Will it capture android enter press
15:59 tenplus1 left #minetest-dev
16:02 Void7 joined #minetest-dev
16:22 est31 what about adding a property to text fields that yield in a "virtual" button being pressed?
16:22 est31 it would solve 1 and 2
16:22 est31 (instead of enter_button[])
16:22 est31 rubenwardy, ^
16:25 T4im what about the "do _not_ use a element name starting with `key_`; those names are reserved to pass key press events to formspec!" in the api btw?
16:26 est31 they are reserved?
16:26 T4im reserved but not used?
16:26 est31 yes that exists
16:26 est31 its like berlin airport :)
16:26 T4im heh
16:28 Krock lol
16:29 burli joined #minetest-dev
16:29 burli Hi
16:29 burli I have a question about the LuaJIT
16:30 burli I noticed that the LuaJIT can run as interpreter. Is it possible that the LuaJIT in Minetest is running as interpreter and not as JIT?
16:31 est31 burli, maybe, I've heard this concern before from someone else
16:32 est31 any ideas how to confirm this concern?
16:32 est31 (that person wasnt really trustable)
16:32 burli Krock has in this thread much better values than I have https://forum.minetest.net/viewtopic.php?f=47&t=15104
16:33 burli He has always 0 ms at the dummy loop, but I have 0.3 ms and more
16:35 burli I have tested it with a self compiled version from github and with the daily build from Ubuntu PPA. Both with LuaJIT enabled
16:35 T4im USE_LUAJIT around cmake or the ingame escape menu?
16:35 burli Both
16:39 burli I make a test and compile minetest without LuaJIT
16:39 burli I hope ;-)
16:40 KaadmY joined #minetest-dev
16:40 burli Then I can see if there is a significant difference
16:41 T4im "local delta = os.clock() - start_clock - loop_time" <-- isn't that cheating? :P
16:42 T4im removing the delta from the second delta... I would imagine that's near 0, yea
16:43 T4im oh wait, nvm, he -1*ed that, sneaky
16:43 T4im no wait, I think I was right the first time
16:44 burli hm, can I clean up the cmake files like with "make clean"?
16:46 STHGOM joined #minetest-dev
16:49 T4im either way, the jit compiler is turned on here
16:49 T4im let print(jit.status()) run from within minetest
16:50 T4im first value should be true
16:50 T4im unless luajit is lying :D
16:51 burli I compiled Minetest without JIT. Much slower.
16:51 burli looptime is now between 30 and 40 ms
16:55 Krock T4im, is the delta correct now or not?!
16:56 Krock loop_time is just there to measure the time of looping and calling a function
16:58 T4im well you take the diff of an empty function, that's expected to be near 0ms cpu time (or like 1-2µs, 2-3µs real time with instrumentation :P), I was confused why you would subtract that from the second one, but if you wanted to get "how much slower" it is, I guess it might make sense, yea
16:59 * T4im wants to try to instrument the core functions later and see what is actually expensive during normal operations
16:59 T4im overall I mean
16:59 Krock yeah, it's just about measuring the time of minetest.get_node. But this difference of loop times irritates me
16:59 T4im yea
17:00 T4im me too
17:03 T4im especially since burli has a higher function overhead but a lower get_node overhead
17:03 T4im maybe he has faster harddisks?
17:03 burli SSD
17:03 T4im ah
17:04 T4im and I take it krock uses spinning hdds
17:04 Krock that doesn't matter. The mapblocks should be cached serverside
17:04 T4im that would explain the lower get_node time, yea
17:04 T4im well, not on the first load?
17:04 Gael-de-Sailly joined #minetest-dev
17:04 Krock T4im, but it calls all nodes around singleplayer's position, so it must be cached
17:04 T4im hm
17:04 burli And I have a AMD FX 6300 CPU, just to make it complete
17:06 Krock I don't notice any additional CPU load during loading the nodes
17:07 T4im maybe his compiler is just better optimizing :P
17:08 Krock or perhaps overall just faster memory access?
17:08 burli Got the same values with my self compiled and the daily build from the Ubuntu PPA
17:09 burli But why is my empty loop slower?
17:11 burli However, I am shure that the JIT is running as JIT and I think that a get_node cache worth it
17:11 Krock *sure
17:12 burli argh
17:12 Krock Assembler instruction: MOV burli,#minetest-de
17:13 burli already there
17:13 est31 lol
17:13 Krock oh
17:21 Fixer ooops
18:01 T4im "gpgkeys: key 92DE8183 not found on keyserver" maybe gpg keys used during travis build should be committed to git, too?
18:09 Grandolf joined #minetest-dev
18:10 halt_ joined #minetest-dev
18:14 proller joined #minetest-dev
18:22 Void7 joined #minetest-dev
18:55 proller joined #minetest-dev
18:59 ud1_ joined #minetest-dev
20:03 davisonio joined #minetest-dev
21:02 twoelk joined #minetest-dev
21:05 Void7 joined #minetest-dev
21:07 Darcidride joined #minetest-dev
22:02 troller joined #minetest-dev
22:15 DI3HARD139 joined #minetest-dev
22:21 davisonio joined #minetest-dev
22:49 Tmanyo joined #minetest-dev
23:25 Fixer joined #minetest-dev
23:44 Fixer_ joined #minetest-dev
23:48 sloanonlinux joined #minetest-dev
23:54 proller__ joined #minetest-dev

| Channels | #minetest-dev index | Today | | Google Search | Plaintext