Minetest logo

IRC log for #minetest-dev, 2015-02-09

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

All times shown according to UTC.

Time Nick Message
01:00 Zeno` joined #minetest-dev
01:18 Megal joined #minetest-dev
02:57 Miner_48er joined #minetest-dev
03:02 Zeno` Wayward_One, is the latest git still giving you problems? It seems ok to me now
03:04 Zeno` i.e. I am getting 5 fps under callgrind (singleplayer game) which is about the same as I get connecting to a local server
03:09 Zeno` although g_fontengine->getTextHeight() is kind of expensive
03:09 Zeno` hmm
03:20 hmmmm Zeno`, cache all of the things
03:21 Zeno` I was just looking through the source code to see how "cacheable" text height is
03:22 hmmmm seemingly a lot of things sapier coded are expensive computations and he completely disregards the fact that the entire loop has to complete in under 16 milliseconds
03:22 hmmmm whenever i code crap that goes into main render loops i put performance first
03:24 Zeno` yeah of course
03:25 Zeno` currently it's being called 3 times in one function as well (gcc and most other compilers will not 'cache' those calls away)
03:25 hmmmm yep :(
03:25 hmmmm obviously all this stuff needs a lot of work
03:26 hmmmm anyway, if you want to optimize that you can since I'd argue it's part of a bugfix
03:26 hmmmm the bug being horrendous regressions
03:26 Zeno` thinking about it they can't really... without looking at getTextHeight() and what it does (which all-of-program optimisation might do because it's part of the same program but if it's in a lib it will never 'cache' them)
03:26 hmmmm and I'm building ASAP
03:26 Zeno` ok... I'm just tracking down to see how/when font can change first
03:27 Zeno` Oh, there is already a font setting changed callback anyway
03:28 Zeno` oh, it's not as nice as I'd hoped heh
03:28 hmmmm all these darn settings callbacks...
03:48 Zeno` I wish I'd finished my callback manager
03:49 Zeno` it's next on my list anyway
03:49 Zeno` more annoying is right now is how many settings might affect font height :/
03:51 * Zeno` has to think
03:53 Zeno` ok, I know
03:53 Zeno` I'll use the hammer :D
04:00 Zeno` lol, there is a bug anyway.  /set font_size 10
04:00 Kalabasa joined #minetest-dev
04:07 Zeno` err ~CGUITTFont() drops the driver?
04:09 kahrl Zeno`: the constructor grabs it
04:10 Zeno` yeah, just saw that
04:11 Zeno` I'
04:11 Zeno` I'd better recompile... my bt seems odd
04:16 Zeno` kahrl, are you familiar with this area of the code?
04:16 kahrl nope
04:16 kahrl I mean, I added it back in the day, but I never really read how it works
04:17 Zeno` ok
04:17 Zeno` I'll start reading docs (making a sandwich first)
04:17 kahrl copy-paste coding ;) but it's okay since it's library-like
04:26 Zeno` Maybe I could pretend it doesn't crash and just go back to caching it. At least there is no danger of it changing xD
04:42 Hunterz joined #minetest-dev
04:46 Kalabasa joined #minetest-dev
04:55 hmmmm hrmmm
04:55 VanessaE uh oh
04:55 VanessaE he's thinking again
04:55 * VanessaE hides
04:56 hmmmm so:  irrlicht, sqlite, zlib, openal, freetype, gettext, luajit, am I missing anything?
04:56 VanessaE hmmmm: curl
04:56 hmmmm i never made an official build before
04:56 hmmmm ahh
04:56 kahrl vorbis?
04:56 hmmmm I didn't realize I needed to add that separate
04:56 hmmmm well the good news is, once I get this set up I won't ever have to do it again =]
04:56 VanessaE also boost + leveldb
04:56 hmmmm boost??
04:56 VanessaE if you wanna support that, that is
04:56 hmmmm why, does leveldb use boost?
04:57 VanessaE idk
04:57 VanessaE maybe that's since changed
04:57 hmmmm shrug
04:58 kahrl s/vorbis/ogg, vorbis, vorbisfile/
04:59 VanessaE you may also need an ssl lib, I *think*
04:59 hmmmm oh god
04:59 VanessaE (to make curl happy)
04:59 hmmmm and i need to recompile 64 bit variants for each
04:59 hmmmm this is insanity
04:59 VanessaE strike that, it IS curl I'm thinking of
05:00 VanessaE libcurl4-openssl-dev or libcurl4-nss-dev or libcurl4-gnutls-dev
05:00 VanessaE (well as they exist on debian derivs anyway)
05:01 hmmmm can a nossl build be used?
05:01 VanessaE I don't see why not
05:01 VanessaE the only reason we even use curl is for media fetching and the server list
05:01 VanessaE and the mod store
05:01 VanessaE which...might be ssl
05:02 VanessaE well the website is https anyway
05:02 VanessaE ( https://forum.minetest.net/mmdb/ )
05:02 hmmmm nvm i found ssl builds for each
05:03 VanessaE ok
05:03 Zeno` I think I'll write the callback manager first
05:04 VanessaE <offtopic>
05:04 VanessaE ~seen sapier
05:04 ShadowBot VanessaE: I saw sapier in #minetest-dev 6 days, 6 hours, 8 minutes, and 37 seconds ago saying "steam? isn't this the tool deleting users homedir?"
05:04 VanessaE </offtopic>
05:04 VanessaE fek.
05:06 Zeno` I am not writing 12 callbacks for font height cache
05:07 hmmmm AHHHHHH
05:07 VanessaE ?
05:07 Zeno` should be able to add a callback for *any* settings changed
05:07 hmmmm all the curl builds are of the command line utility
05:07 Zeno` instead of one callback per setting
05:08 hmmmm you're not able to register the same callback for multiple settings?
05:08 hmmmm from what I remember of the callback mechanism, it passes along the name of the setting that changed
05:08 hmmmm why would it pass that along if it were designed to be one callback per setting
05:10 Zeno` you're able to use the same one, yeah
05:10 Zeno` but still
05:10 Zeno` anyway have to fix the bug first
05:10 Zeno` otherwise the callback is useless anyway :)
05:11 Zeno` then I'll call registerChangedCallback() 12 times (or however many font-related settings there are) :/
05:13 Zeno` or others that *may* affect font size
05:15 hmmmm btw
05:16 Zeno` Anyway, I'm not sure if Wayward_One still has the singleplayer problem. I've just profiled HEAD twice and it seems similar to connecting to a local server now
05:16 hmmmm i wonder how android doesn't have an exception on startup immediately
05:16 Zeno` it probably does
05:16 Zeno` why do you ask?
05:16 hmmmm notice how screen_dpi is not set in #ifdef __ANDROID__
05:16 hmmmm for the defaults
05:17 Zeno` well does it compile anyway? lol
05:17 Zeno` oh... hmmm
05:17 hmmmm also android, to my knowledge, does not use Xorg and therefore would not use the X DPI query
05:17 hmmmm so it'd default to g_settings->getFloat("screen_dpi") or whatever
05:17 hmmmm and therefore blow a SettingsNotFoundException
05:18 hmmmm unless I'm totally mistaken somehow
05:18 hmmmm i can't do android builds so
05:19 Zeno` I don't think you're mistake
05:19 Zeno` wonder why it works :/
05:20 Zeno` (or why the exception doesn't work)
05:20 hmmmm maybe nobody tried a recent enough build
05:22 VanessaE is not http://www.fileconvoy.com/dfl.php?id=g9c63a08b33eb932899961019369bd677d164a74aa new enough?
05:22 hmmmm maybe the java executed on startup writes a screen_dpi to a pregenerated minetest.conf
05:22 VanessaE that's the most recent one sapier released anyway
05:22 VanessaE from roughly Jan 17 though
05:23 Zeno` grepping for screen_dpi doesn't show anything that appears specifically and only for android
05:23 hmmmm :|
05:23 chchjesus joined #minetest-dev
05:23 Zeno` maybe android disables exceptions
05:23 hmmmm this is a logistics nightmare
05:24 hmmmm how am I supposed to release knowing full well there might be a bug with android
05:24 hmmmm but nobody currently present can actually verify
05:24 Zeno` well there is also #2269
05:24 ShadowBot https://github.com/minetest/minetest/issues/2269 -- string.cpp error on android build make
05:24 chchjesus joined #minetest-dev
05:24 VanessaE hmmmm: is this helpful?  https://source.android.com/source/initializing.html`
05:24 VanessaE well, minus the typo
05:25 hmmmm no :)
05:25 VanessaE I tied :P
05:25 VanessaE tried*
05:25 hmmmm I am not going to set up an android build environment, sorry
05:25 Zeno` I had one set up
05:25 hmmmm i only do build things because i need to
05:25 Zeno` took me two days
05:25 hmmmm right
05:25 kahrl hmmmm: the getDisplayDensity porting_android.cpp does not use screen_dpi, it seems
05:25 kahrl hmmmm: it calls some java stuff
05:25 Zeno` and then I upgraded Fedora and kind of broke it
05:25 hmmmm kahrl:  ahh ok
05:26 kahrl getDisplayDensity in*
05:26 Zeno` I grepped the wrong string ;)
05:26 hmmmm i guess i never saw porting_android.cpp because my IDE project file is chronically out of date =]
05:27 hmmmm yep, got it
05:27 hmmmm ...on that note, shouldn't we cache porting_android.cpp's variant of getDisplayDensity too?
05:28 hmmmm as if that didn't murder performance on the PC enough, it'd be absolutely fatal on android
05:28 hmmmm same with getDisplaySize
05:28 kahrl that would make sense, wouldn't it
05:29 hmmmm too much sense.
05:29 kahrl although with getDisplaySize, I dunno if that changes if you hold the phone vertically instead of horizontally
05:29 kahrl I doubt irrlicht handles that properly anyway though
05:43 hmmmm good point...
05:46 hmmmm kahrl, http://xiph.org/downloads/ doesn't have vorbisfile
05:46 hmmmm I would assume if anywhere it'd be there
05:47 kahrl might be part of libvorbis?
05:55 hmmmm i'd think so
06:05 jin_xi joined #minetest-dev
06:40 Hunterz joined #minetest-dev
06:49 VanessaE bbl
06:59 nrzkt joined #minetest-dev
06:59 nerzhul_ joined #minetest-dev
07:03 hmmmm yeah nevermind about the getDisplayDensity android stuff.. there is a static value caching there
07:04 kahrl oh, right
07:05 hmmmm i don't think i'm going to be able to make the builds tonight.... i'm dead tired at the moment and leveldb is extremely uncooperative
07:06 kahrl is leveldb required?
07:07 hmmmm i would argue it is
07:07 hmmmm all the previous windows builds included it, right?
07:07 kahrl no idea
07:09 kahrl I don't see it in the msvc 32-bit build of 0.4.10
07:10 kahrl nor the 64-bit one
07:10 kahrl it's not in 0.4.9 either
07:11 hmmmm i suppose i can skip it here too
07:11 kahrl of course, it could just be compiled in statically so there is no dll
07:16 kahrl if you do want to include it, perhaps try the fork at https://github.com/bitcoin/bitcoin/tree/master/src/leveldb
07:17 kahrl (its WINDOWS.md is somewhat promising)
07:53 guest365 joined #minetest-dev
08:14 Anchakor_ joined #minetest-dev
08:19 ImQ009 joined #minetest-dev
08:25 proller joined #minetest-dev
08:28 kilbith joined #minetest-dev
08:38 nrzkt joined #minetest-dev
08:39 nrzkt hmmmm, release will happend today ? :)
08:40 kahrl nrzkt: if all goes well, probably
08:41 nrzkt that's cool, and the new branch will be created ?
08:41 kahrl what branch?
08:43 nrzkt https://forum.minetest.net/viewtopic.php?f=3&amp;t=11172
08:43 nrzkt one master branch and one release branch
08:44 kahrl well there won't be a release right after the release, I hope
08:44 kahrl so no need for a branch right now
08:46 nrzkt ofc, but this permit to re-open contributions. The development is stalled since 3 weeks
08:47 kahrl http://irc.minetest.ru/minetest-dev/2015-02-08#i_4146398
08:47 kilbith the -dev branch should be opened just before the freeze
08:48 nrzkt perfect
09:16 crazyR joined #minetest-dev
09:54 Amaz joined #minetest-dev
09:56 FR^2 joined #minetest-dev
10:24 Kalabasa joined #minetest-dev
10:49 gregorycu_ joined #minetest-dev
10:51 gregorycu__ joined #minetest-dev
11:00 selat joined #minetest-dev
11:05 Kalabasa joined #minetest-dev
11:35 PilzAdam joined #minetest-dev
12:26 gregorycu__ joined #minetest-dev
12:36 gregorycu__ How's the release looking?
12:37 gregorycu__ Not the best it appears
12:37 gregorycu__ Better late than never
12:52 celeron55 lol i just looked at the forum; looks like some spambot manages to create spam reports about random topics
12:52 celeron55 wait, no, i was mistaken
12:52 * celeron55 is the worst forum moderator
12:58 celeron55 so, this goes in action now  https://forum.minetest.net/viewtopic.php?f=3&amp;t=11172
12:59 celeron55 so if the release happens today, i won't go messing around with any branches for now
13:00 gregorycu__ That
13:00 gregorycu__ That's from 4 years ago?
13:01 gregorycu__ Oops, misread your join date as post date
13:01 gregorycu__ Nevermind me
13:01 celeron55 if the freeze hasn't ended tomorrow, the freeze will be moved to another branch then
13:01 nrzkt no news from hmmmm, he was trying to test windows build 12 hours ago
13:02 gregorycu__ Anything I can help with
13:02 gregorycu__ Nothing is listed as a blocker so I'm not sure what's going on...?
13:03 nrzkt hmmmm doesn't have time to finish the release yet
13:04 gregorycu__ Is it the perf issue I highlighted ages ago? With regards to fade in - out?
13:04 nrzkt celeron55, what is your local hour please ? :)
13:04 gregorycu__ celeron55: I found a somewhat significant issue with rendering on my machine
13:05 gregorycu__ (I have no idea who the graphics expert is, it seems to be you)
13:05 gregorycu__ nrzkt: He must be sleeping
13:12 shadowzone joined #minetest-dev
13:12 celeron55 nrzkt: EET
13:14 celeron55 (UTC+2 at winter, UTC+3 at summer)
13:16 nrzkt okay, then i could say we are on the same day
13:18 gregorycu__ I am 17 minutes into a new day
13:24 Zeno` joined #minetest-dev
13:29 Zeno` is there a 0.4.12 branch yet?
13:30 nrzkt no, celeron55 mention it but there isn't
14:08 shadowzone joined #minetest-dev
14:08 celeron55 http://dev.minetest.net/All_rules_regarding_to_development
14:09 celeron55 i updated the development wiki to actually have a common page for these things
14:09 celeron55 it's not perfect but at least i now had a place to add the new rules in 8-)
14:10 celeron55 i think that if you want to get onto some PR merging right now, create the freeze-0.4.12 branch and get on it
14:11 celeron55 there's not much that can go wrong i guess
14:12 celeron55 or actually
14:13 celeron55 i wonder if our existing build scripts can handle a release that was made in a different branch and isn't directly a commit from master's history?
14:13 celeron55 i guess if they just pull a tagged commit from the repo, it should work
14:17 celeron55 i did add this in there too: "If a core developer does not have time for doing anything, their tasks will be explicitly moved to be done by someone else."
14:18 celeron55 but that is a bit problematic due to us generally not having anyone to do the thing
14:20 celeron55 ---
14:22 celeron55 however, now my main worry is that as i am not willing to allocate time for evaluating every core developer's work and all comments in every pull request and issue and i have to rely on other people noticing important problems regarding to those and reporting them to me, how do i get people to do that
14:25 celeron55 maybe it's not an issue and the key here is just having those rules explicitly stated
14:27 nrzkt thanks celeron55
14:29 nrzkt the wiki modifications are good :)
14:34 Guest69228 joined #minetest-dev
14:42 celeron55 everyone: you can merge pull requests if you want, as long as you create an untouched freeze branch on github before doing it
14:43 celeron55 oh whatever, i'll do it myself
14:44 celeron55 it's now there
14:45 celeron55 (probably nobody will do anything and hmmmm will just remove it tomorrow, but whatever)
14:47 crazyR joined #minetest-dev
14:50 JakubVanek joined #minetest-dev
14:56 CraigyDavi joined #minetest-dev
14:56 nrzkt thanks celeron55, good news !
14:57 kilbith the right measures by the right man
15:03 hmmmm joined #minetest-dev
15:11 celeron55 https://forum.minetest.net/viewtopic.php?p=169771#p169771
15:12 celeron55 one more rule about a thing that has bothered me (and probably many others) for some time
15:15 celeron55 i'm not going to touch any PRs today due to not being around lately and this rule not strictly existing previously, but with this i might end up doing that more in the future
15:17 nrzkt you talk about old PRs ?
15:21 celeron55 well, relating to this, all of them are old
15:22 nrzkt yes, and they doesn't permit to have a global view of the project and the user needs
15:43 SopaXorzTaker joined #minetest-dev
15:50 shadowzone joined #minetest-dev
15:58 jin_xi joined #minetest-dev
15:59 SopaXT joined #minetest-dev
16:21 Calinou joined #minetest-dev
16:31 ImQ009 joined #minetest-dev
16:36 Guest59915 joined #minetest-dev
16:53 twoelk joined #minetest-dev
17:10 ElectronLibre joined #minetest-dev
17:11 DFeniks joined #minetest-dev
17:15 Hunterz joined #minetest-dev
17:23 Guest54920 joined #minetest-dev
17:33 leat joined #minetest-dev
17:33 Krock joined #minetest-dev
17:37 Krock Does anyone here have experience with zlib? Somehow, the special characters support is lost when I use 1.2.8 with the following change: https://github.com/SmallJoker/minetest/commit/1c5dd
17:39 Krock Well, maybe it's a problem with gettext because I had to edit that file to get rid off some errors
17:50 JakubVanek zlib is compression library afaik, so if zlib is broken then everything should be broken
17:52 JakubVanek gettext is more likely to cause it
17:53 Krock hmm okay
17:53 JakubVanek sorry for my poor grammar, english is not my native language
17:53 Krock still, I like any help I get
17:53 rubenwardy joined #minetest-dev
17:56 MinetestForFun joined #minetest-dev
18:05 roniz joined #minetest-dev
18:10 luizrpgluiz joined #minetest-dev
18:10 luizrpgluiz hi
18:12 luizrpgluiz someone please tell me, when you leave the new version of minetest?
18:12 JakubVanek Krock: latest Linux Ubuntu build, zlib 1.2.8, Czech language: http://s29.postimg.org/4imebu2nb/Screenshot_9_2_2015_19_10_38.png
18:13 JakubVanek Krock: that limits the problem to Windows
18:13 JakubVanek I don't know many differences between mingw and msvc build
18:13 JakubVanek but there can be some
18:14 JakubVanek bbl
18:22 JakubVanek joined #minetest-dev
18:22 luizrpgluiz celeron55: hi
18:26 Robert_Zenz joined #minetest-dev
18:27 OldCoder joined #minetest-dev
18:31 shadowzone joined #minetest-dev
18:37 rubenwardy rebased #1843
18:37 ShadowBot https://github.com/minetest/minetest/issues/1843 -- Add setting to customise stack max by rubenwardy
18:45 nrzkt joined #minetest-dev
18:49 dzho joined #minetest-dev
18:50 rubenwardy What metadata would you think may be useful in the future, or now?
18:50 rubenwardy Like #2201
18:50 ShadowBot https://github.com/minetest/minetest/issues/2201 -- Mod Meta
18:51 rubenwardy How should forks with the same modname be supported?
18:51 rubenwardy Should they be?
18:54 selat joined #minetest-dev
19:02 JakubVanek mod sites would be good, release count shouldn't be just number but semantic version
19:04 rubenwardy The whole point is that the release count is machine readabout
19:04 rubenwardy *able
19:04 rubenwardy The version is for semantic versions
19:07 Robert_Zenz rubenwardy, semantic versions are just as good machine-readable, it's just three numbers stuffed into a single object.
19:07 rubenwardy It makes it simpler to process this way.
19:07 rubenwardy We can see what the core team thinks
19:08 rubenwardy #2272
19:08 ShadowBot https://github.com/minetest/minetest/issues/2272 -- Environmental Sounds
19:23 FR^2 joined #minetest-dev
19:29 est31 joined #minetest-dev
19:31 est31 Hi, I think d902bd31introduced an error
19:31 est31 Hi, I think d902bd31 introduced an error*
19:31 kahrl it introduced a typo in the commit message? :P
19:31 est31 can be fixed for example by merging #2225
19:31 ShadowBot https://github.com/minetest/minetest/issues/2225 -- Fixes for minetest.get_(all_)craft_recipe(s) by gregorycu
19:31 est31 nono
19:32 est31 crafts that have multiple numbers of the stuff as output arent found anymore
19:32 est31 s/multiple numbers/a number > 1/
19:33 shadowzone joined #minetest-dev
19:36 kahrl dunno why that would happen, it checks for NUL or ' ', doesn't it?
19:36 Amaz joined #minetest-dev
19:36 est31 ah yes
19:36 luizrpgluiz left #minetest-dev
19:37 kahrl although I don't understand why only getCraftRecipe was changed and not getCraftRecipes
19:38 est31 because that doesnt need a fixc
19:38 est31 its already bypassed in the lua handling code
19:38 est31 and the method isnt used anywhere else
19:38 est31 s/bypassed/taken care of/
19:38 kahrl oh, ok
19:40 kahrl I don't think that is great code style but meh
19:42 est31 ok you are right the method seems to work
19:42 est31 but still I get one craft as not listed
19:43 est31 I'll test whether its the fault of the mod
19:54 est31 ok its very likely its the mod's fault
20:05 crazyR_ joined #minetest-dev
20:21 est31 so whats blocking for the release now
20:27 JakubVanek joined #minetest-dev
20:32 roniz_ joined #minetest-dev
20:33 JakubVanek i thought that hmmmm is preparing build
20:37 JakubVanek ~seen hmmmm
20:37 ShadowBot JakubVanek: I saw hmmmm in #minetest-dev 13 hours, 26 minutes, and 40 seconds ago saying "i suppose i can skip it here too"
20:42 roniz__ joined #minetest-dev
21:01 alexxs joined #minetest-dev
21:38 ElectronLibre left #minetest-dev
21:52 shadowzone joined #minetest-dev
22:26 Robert_Zenz joined #minetest-dev
22:59 ElectricSolstice joined #minetest-dev
23:24 luizrpgluiz joined #minetest-dev
23:24 luizrpgluiz hi all
23:27 fireglow hi luizrpgluiz
23:45 zat joined #minetest-dev
23:47 luizrpgluiz left #minetest-dev

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