Time  Nick        Message
05:32 hmmmm       Zeno`:  Hey I haven't read much of the backlog to answer your question, 0.4.12 can be released when that super nasty performance regression is fixed
05:32 Zeno`       ok, just wanted to make sure that was the major stumbling block
05:34 Zeno`       It's not the easiest regression in the world to find :/
05:35 est31       havent you found it
05:35 est31       that viewrange adjustment?
05:35 Zeno`       nah, it's not that
05:36 Zeno`       I am now suspecting the MeshUpdateThread
05:37 Zeno`       hmmmm, have you had any luck with it?
05:37 hmmmm       i haven't had any chance to look at it
05:37 Zeno`       est31, I asked Wayward_One to try something for me but I'm not sure if he did or not
05:38 est31       did the bisect you did fail?
05:39 Zeno`       it didn't go so well, no
05:42 Zeno`       Mainly because I have to artificially slow things down to actually experience the issue
06:07 * Zeno`     doesn't understand
06:21 Zeno`       #2252
06:21 ShadowBot   https://github.com/minetest/minetest/issues/2252 -- Framerate cut in half, HUD problem ?
06:22 hmmmm       whaat
06:22 est31       is this dpi calculation going on every draw?
06:23 hmmmm       oh shit..
06:23 Zeno`       when I am running in callgrind drawtime with hud disabled is ~35. with hud enabled drawtime is about 250
06:23 Zeno`       average 200
06:23 Zeno`       that's a pretty large difference
06:23 est31       can you try whether you get it with other formspec too
06:24 est31       or is HUD no formspec
06:25 Zeno`       inventory formspec open or closed makes not much difference (if any)
06:27 hmmmm       yuppers
06:27 hmmmm       try reverting e201620ee1051545ed7856fb18f805c67adccc85
06:28 hmmmm       rather not that one
06:30 hmmmm       0f1d33933d264fa84b2d97668258d066da0ca01c
06:39 Zeno`       +1
06:39 Zeno`       drawtime is now the same with or without HUD
06:40 Zeno`       ~~40.... rather than 200 :)
06:41 est31       as it seems its doing a call to XOpenDisplay and XCloseDisplay at every render loop
06:41 hmmmm       christ
06:41 hmmmm       thanks a lot sapier :)
06:42 hmmmm       thanks Zeno`, est31, for tracking this down
06:42 hmmmm       you guys are awesome
06:42 hmmmm       I feel bad for not doing anything to help :(
06:42 est31       Zeno` did the time intensive work
06:43 Zeno`       hmmmm, you can help by checking my conflict resolution ;)
06:43 Zeno`       will push in a second
06:43 hmmmm       well
06:43 hmmmm       obviously don't revert the commit
06:43 hmmmm       just cache the value
06:43 est31       ^
06:44 Zeno`       duh
06:44 * Zeno`     slaps himself. wake up
06:44 hmmmm       lol
06:44 hmmmm       what time is it there in aussieland
06:45 est31       lol its calling that 4 times
06:46 hmmmm       now I suppose the real question is why bisecting failed here
06:46 hmmmm       we want to learn from our mistakes
06:47 Zeno`       there was no real way for me to measure it :/
06:48 Zeno`       /ctcp time Zeno`
06:48 Zeno`       might not work
06:49 Zeno`       haha
06:49 hmmmm       lol
06:49 hmmmm       you inverted your arguments
06:49 Zeno`       4:50PM Thursday
06:49 hmmmm       you're not supposed to fall asleep BEFORE dinner
06:49 Zeno`       oh :)
06:50 Zeno`       why not?
06:50 hmmmm       because you get tired after you eat
06:51 * Zeno`     makes a note
06:51 hmmmm       I wish you could set color themes for windbg :(
06:51 hmmmm       white backgrounds == eyerape at night
06:51 Zeno`       windbg?
06:52 Zeno`       https://github.com/Zeno-/minetest/commit/970b75550a36b6812432c7c3cc6665e9c0dc639f
06:53 est31       I'd cache it inside porting.cpp
06:53 hmmmm       probably a better idea
06:53 hmmmm       i was just about to say that fontengine could use that too
06:53 Zeno`       ok
06:54 est31       https://github.com/minetest/minetest/search?p=1&q=getDisplayDensity&utf8=%E2%9C%93
06:55 est31       when the progress bar at loading is drawn its called too
06:56 est31       and at other places
06:59 Zeno`       ugh, it's a float
07:00 Zeno`       will have to add an is_cached bool I guess
07:00 est31       or cache it everywhere
07:01 Zeno`       nah I mean I was just going to set it to -1 if not initialised
07:01 Zeno`       but I don't like comparing floats; would rather add the bool
07:02 est31       ah you want to calculate it at the first draw?
07:02 Zeno`       yeah
07:02 Zeno`       https://github.com/minetest/minetest/pull/2260
07:03 Zeno`       where did you mean to cache it?
07:04 est31       when the program starts
07:04 Zeno`       add an initDisplayDensity() function or something?
07:04 est31       static float c_display_density = getDisplayDensity();
07:04 Zeno`       and have it as global?
07:04 Zeno`       oh
07:04 Zeno`       yeah
07:04 est31       the way you do it, it is regenerated every iteration
07:04 est31       :)
07:04 Zeno`       no it's not :p
07:05 est31       ahh ok youre right
07:08 Zeno`       this is an alternative: https://gist.github.com/Zeno-/1d01769ee969ebeec50c
07:08 Zeno`       I probably like that better
07:09 Zeno`       saves a conditional :3
07:09 Zeno`       lol
07:09 Zeno`       I leave it to you and hmmmm to vote
07:10 Zeno`       I'm going with the gist... updating PR
07:10 hmmmm       wait make that g_static_cached_display_density
07:11 Zeno`       it's not global
07:11 hmmmm       ohhh
07:11 hmmmm       sorry i misread that
07:11 hmmmm       yup sounds good
07:11 Zeno`       #2260
07:11 ShadowBot   https://github.com/minetest/minetest/issues/2260 -- Fix performance regression by Zeno-
07:11 est31       it was my idea, so why be against it? :)
07:11 Zeno`       merging soon then
07:11 hmmmm       LGTM
07:12 Zeno`       thanks for the help hmmmm and est31
07:12 hmmmm       the coding purists are probably rolling in their graves right now
07:12 hmmmm       OH MY GOD GLOBAL STATIC VARIABLES
07:12 est31       1 question: does this also get executed for the server?
07:12 hmmmm       WHY AREN'T YOU USING A SINGLETON DESIGN PATTERN WITH WIN-WIN SYNERGY AND REACTOR OBSERVER MVCS
07:12 est31       because then the other might seem better
07:12 Zeno`       heh
07:12 Tesseract   Zeno`: Wait.
07:13 * Zeno`     halts
07:13 hmmmm       est31, it doesn't get executed for the server unless XORG_USED is defined
07:13 Tesseract   Zeno`: Make it a static inside the function.  No global needed.
07:13 est31       that was the approach before
07:13 Zeno`       swap it back? lol
07:13 est31       (the alternative)
07:13 Tesseract   static x = init(); return x;
07:14 Tesseract   Zeno`: If that's what you had before.
07:14 Zeno`       oh. ok
07:14 hmmmm       thats' better
07:14 hmmmm       also it's probably a good idea to change getDisplayDensity_init to something like... calcDisplayDensity
07:15 hmmmm       and don't make it static so other people can call it if they explicitly want to
07:15 hmmmm       for whatever odd reason...
07:16 Zeno`       demands, demands, demands
07:16 Zeno`       changing now
07:17 * est31     is happy the issue got nailed down
07:17 hmmmm       me too =]
07:20 Zeno`       calcDisplayDensity() being public... kind of silly imo
07:20 Zeno`       I'll have to add versions for non xorg builds :/
07:20 hmmmm       oh
07:20 hmmmm       nevermind
07:21 Tesseract   Oh, I mentioned two bugfix pulls a few hours ago -- did you check those?
07:21 hmmmm       i forgot about that
07:21 * Tesseract looks for #s
07:21 Zeno`       good :)
07:21 Zeno`       didn't want to type toooooo much
07:21 Zeno`       heh
07:21 Tesseract   #2245 & #2256
07:21 ShadowBot   https://github.com/minetest/minetest/issues/2245 -- Fix dying of lava causes repeated death by gregorycu
07:21 ShadowBot   https://github.com/minetest/minetest/issues/2256 -- Give full breath after death by SmallJoker
07:23 Tesseract   Zeno`: Your patch LGTM now.
07:24 Zeno`       ok thank you
07:25 Tesseract   hmmmm: Those two O.K.?
07:26 Zeno`       2256 is fine; I looked at it last night
07:26 Zeno`       2245 makes sense as well
07:28 Tesseract   Zeno`: Great!  Can you also check #2176?  It's a bit bigger.
07:29 ShadowBot   https://github.com/minetest/minetest/issues/2176 -- Fix translation issues by ShadowNinja
07:30 Tesseract   Oh, and you can merge those two, or wait til tomorrow and I'll do it then.
07:30 Zeno`       Ok, I'll merge mine and those 2 in a few minutes
07:30 Zeno`       2176 may take me a while to look through
07:33 hmmmm       yeah those works
07:34 hmmmm       I would prefer that he used getS32NoEx() instead but it's fine
07:34 hmmmm       yeah those work*
07:36 Tesseract   hmmmm: What are you reffering to by getS32NoEx?  None of the last 4 patches mentioned uses it or a related function.
07:37 hmmmm       https://github.com/SmallJoker/minetest/commit/28e0c9ba5e9a31519f896dbca297249f0d2f627e  ????
07:37 hmmmm       am I going nuts
07:38 hmmmm       try { foo = args.getS32("foo"); } catch (SettingNotFoundException &e) { foo = FOO_DEFAULT; } can be replaced with  foo = FOO_DEFAULT; args.getS32NoEx("foo", foo);
07:39 Tesseract   Ah, yes.  He didn't write that, but it could be amended.
07:39 hmmmm       for some reason I thought that was gregorycu's commit
07:39 hmmmm       hmm
07:39 hmmmm       yea it's okay like i said, not a deal breaker by any means
07:40 Tesseract   Alright.  It would be great if you could check out my translation pull too hmmmm.  But G'Night now.  o/
07:41 hmmmm       good night
07:42 hmmmm       excellent
07:42 hmmmm       2176 looks great to me.  lots of nagging issues fixed
07:44 Zeno`       night
07:44 * Zeno`     looks outside. It's still light
07:46 hmmmm       the days pass by way too quickly
07:51 Zeno`       I'd prefer the comment // You must free the returned string! (const wchar_t *narrow_to_wide_c(const char *mbs)) to actually indicate that the mem was allocated with new, but that's being pedantic (i.e. 2176 looks fine to me)
08:11 paramat     hmmmm, if you have time #2229 is ready (good screenshots too)
08:11 ShadowBot   https://github.com/minetest/minetest/issues/2229 -- Mgv7: Bugfixes for 0.4.12 and update conf.example by paramat
08:13 Zeno`       The endless vertical walls has been the main reason I've avoided mgv7 so far
08:15 paramat     they only occur if you use custom noiseparams in a recent version, i removed a rangelim a few weeks ago :}
08:21 hmmmm       great stuff
08:21 hmmmm       paramat, do you have anything more you'd like me to merge while i'm at it?
08:21 paramat     wow nice fix, my latest build runs faster now
08:21 paramat     nothing else thanks
08:21 hmmmm       btw add an extra newline in between your first and second lines in the git commit
08:22 hmmmm       that way when it gets printed as a single line only the first line gets printed
08:22 paramat     okay, looks neater
08:25 paramat     recently, every time i started a new world and first turned around i noticed the terrain would take ages to appear, now fixed
08:25 paramat     by the HUD bugfix
08:25 hmmmm       amazing how such an inoccuous looking commit could have such a big effect
08:26 hmmmm       so glad we didn't release this without fixing the regression
08:26 hmmmm       in any case, 0.4.12 is supposed to last us a while before 1.0.0
08:27 paramat     im working on improved noiseparams for mgv7, the cliffs work if you boost the noise scale to 20+, i can get nice mgv6 type terrain now
08:27 hmmmm       but one question:
08:27 hmmmm       how did that have anything to do with singleplayer?
08:29 hmmmm       boost what noise scale to 20+?
08:29 paramat     ah i see mediacrush is closing down
08:30 paramat     mgv7_np_height_select = 0, 20, (500, 500, 500), 4213, 5, 0.7, 2.0
08:30 hmmmm       so it DID work
08:30 VanessaE    hmmmm: wait, the formspec scaling thing still needs fixed
08:30 hmmmm       VanessaE, ??  whathhahtah
08:31 hmmmm       I am so not getting involved with the formspec scaling issue
08:31 VanessaE    look at the pause menu at different window sizes.
08:31 VanessaE    also the HUD.
08:31 paramat     perhaps some donation maoney could fund image hosting for MT, it's always a problem for me, back to imgur
08:31 VanessaE    well that should be fixed before 0.4.12 goes out, it's a regression
08:31 nrzkt       VanessaE, approved
08:34 paramat     hmmmm, http://i.imgur.com/lLg0RES.png
08:34 nrzkt       wtf xD
08:35 hmmmm       well that's interesting terrain to say the least
08:35 paramat     next i'll tune the 3d mountain noise, i could even use it to create floatlands throughout the atmosphere if you guys approve
08:41 paramat     so i have a feel now for what you were attempting with v7 base terrain, my work has been to realise that, using the existing code if possible
08:46 paramat     so, in mgv7 rare and scattered 3d noise can be the floatlands but also the mountains where they intersect the ground. i have been getting requests for 'levels' in core mapgen. the floatland noiseparams could be an alternative set, not necessarily the default :)
08:46 paramat     http://i.imgur.com/jnZ15hj.png
08:46 VanessaE    O_o
08:47 VanessaE    looks like a technological ant colony
08:48 paramat     my 'projecs' mod. very neglected mod with huge potential for endless growing architecture
08:48 paramat     must work on it ..
11:08 puhfa       Calinou: hmm, lowered gain to .5 and max_hear_distance to 6, still hearing it full-force at 200 m
11:08 puhfa       is there a good working example in one of the mods?
11:09 VanessaE    puhfa: homedecor's door sounds.
11:09 puhfa       ty
11:09 VanessaE    pos, gain and max_hear_distance have to all be specified, and the sound has to be 1-channel (mono)
11:09 VanessaE    otherwise it won't behave as you expect it to
11:10 VanessaE    https://github.com/VanessaE/homedecor_modpack/blob/master/homedecor/doors_and_gates.lua#L414
11:10 VanessaE    example ^^^
11:11 puhfa       yes, just copied those parameters, still the same
11:11 puhfa       arent you using oggs there as well?
11:11 VanessaE    yep.
11:11 puhfa       weird
11:11 VanessaE    make sure your sound file is mono
11:11 puhfa       oh
11:14 puhfa       it should be
11:14 puhfa       well, it is
11:15 VanessaE    *shrug*
11:21 gregorycu   So, https://github.com/minetest/minetest/commit/9a0dd47057cad18522e43d159a06c856dfdeeef4 fixes perf issues?
11:25 gregorycu   Ahh linux only, can't really look at that
11:25 gregorycu   game time
12:57 paramat     hmmmm #2175 is ready for review and merge O/
12:57 ShadowBot   https://github.com/minetest/minetest/issues/2175 -- Mgv5: Remove blobgen, add large caves. De-link terrain level from water level. by paramat
13:53 Wayward_One latest git still gives me the massive lag in singleplayer :(
13:54 Zeno`       Wayward_One, but changing the thread priority (the patch I asked you to test last night) made a difference?
13:55 Wayward_One it seemed to, a little
13:55 Zeno`       sigh
13:56 Zeno`       can you try the patch again now? (on top of latest git)?
13:57 Zeno`       and maybe change the line (in client.cpp) porting::threadSetPriority(threadId(), 9);   to  porting::threadSetPriority(threadId(), 20);
13:58 Wayward_One ok
13:58 Zeno`       'tis driving me nuts
14:02 Wayward_One sorry, lost the link >.<
14:03 nrzkt       Zeno`, can you look at your PM ? :)
14:05 Zeno`       I don't have any PMs (?)
14:06 nrzkt       strange, i have a working conversation with you on IRC
14:06 nrzkt       pidgin doesn't tell me there is a problem
14:07 Zeno`       lol
14:36 sfan5       nrzkt: tip: pidgin is a bad irc client
15:24 nrzkt       sfan5 maybe but it's my choice and it works properly for my needs.
16:13 Calinou     translations don't work in latest Git
16:19 Tesseract   Calinou: Specify "don't work".
16:19 Calinou     the game just uses English
16:19 Calinou     no errors in log
16:19 Tesseract   I tested with language=es and it worked in spanish for me.
16:20 Calinou     I'll try manually specifying language
16:20 Tesseract   You've got gettext support enabled, translations compiled, and are using a supported laguage?
16:20 Calinou     yes (French)
16:21 Calinou     -- GetText enabled; locales found:
16:28 Zeno`       Wayward_One, do you experience the singleplayer lag issue with 0.4.11?
16:29 Zeno`       or 0.4.10?
16:33 Tesseract   Calinou: Manual language works or not?
16:35 Calinou     nope
16:42 Tesseract   Calinou: Found the issue.  *grumbles something about C's stupid preprocessing functionality*
16:50 Tesseract   Calinou: Fixed.
16:51 Zeno`       the c preprocessor is great!
16:52 Wayward_One Zeno`, just tested 0.4.10, lag seemed normal, let me try 0.4.11
16:55 Zeno`       the C preprocessor is the greatest invention since sliced mushrooms
16:57 Tesseract   Zeno`: Maybe, but it could stand some improvement, like "#if LOLNOTDEFINED" being an error.
16:58 Zeno`       you can do that with #error can't you?
16:59 Zeno`       e.g. #ifndef LOL #error "oops" #endif
16:59 Tesseract   Zeno`: No, I mean using undefined values should be an error, like "if (typodvarname)" is an error.
17:01 Tesseract   Zeno`: You could do that of course, but it adds three lines to every use of #if on a defined constant.
17:04 Zeno`       make a macro! :)
17:05 init        make a macro that adds a macro for you
17:05 init        s/adds/makes
17:05 Zeno`       j/k ... the thing is it failed and was easy to find. Stuff happens.
17:06 Zeno`       i.e. no big deal IMO :P
17:31 Wuzzy       Hi! Is the bug that the big subgame logo in the main menu is not shown for some resolutions already known?
17:33 VanessaE    as far as I can see it's only omitted if it won't fit
17:33 VanessaE    (just tried it)
17:33 Wuzzy       Yes, that's what I meant.
17:33 Wuzzy       But isn't the logo scaled anyways? So this behavior is inconsistent
17:34 VanessaE    it only scales with window witch
17:34 VanessaE    width*
17:34 VanessaE    ah
17:34 Wuzzy       Hmpf, this sucks. :-(
17:34 VanessaE    but the formspecs scale with window *height
17:35 VanessaE    therein lies the problem
17:35 Wuzzy       So: Should I report on GitHub? :P
17:36 Wuzzy       or is this not news to you?
17:36 VanessaE    I knew the scaling was undesirable, but I didn't know the two were scaling by different "keys".  definitely report it.
17:36 VanessaE    (scaling of formspecs is undesirable I mean)
17:39 VanessaE    I believe I made the suggestion once before, but I will say it again for the record:  if you're gonna scale something in a windowed setting like this, you need to scale based on the smaller of width vs. height of the window at any given moment, not just arbitrarily pick one or the other ahead of time.
17:40 VanessaE    and apply that idea to every element/group where there's window-size-based scaling, so stuff like your bug doesn't happen.
17:40 VanessaE    (nevermind that formspecs shouldn't scale to *begin* with)
17:41 Wuzzy       Reported. Issue #2262.
17:41 ShadowBot   https://github.com/minetest/minetest/issues/2262 -- Subgame banner is hidden for wide windows
17:41 est31       Wuzzy: add screenshot?
17:42 Wuzzy       if you insist ...
17:42 Wuzzy       but its REALLY easy to test IMO
17:42 Wuzzy       just start minetest and scale the window around and see what happens
17:58 VanessaE    bbl
18:06 sofar       VanessaE: do you happen to know if PilzAdam is still active? He's made a ton of mods that are useful but I don't see him doing much maintenance on them.
21:56 shadowzone  So, is there gonna be a release anytime soon?  Cause I think Zeno` assigned me to change logs.