Minetest logo

IRC log for #minetest-dev, 2014-06-28

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

All times shown according to UTC.

Time Nick Message
00:37 paramat joined #minetest-dev
00:46 paramat hi hmmmm, it seems that 0.4.10 feature freeze is at the end of this weekend, please could you consider and approve these mapgen commits? https://github.com/minetest/minetest/pull/1390 https://github.com/minetest/minetest/pull/1407
00:47 sapier last time we did remove some of prollers addon that resulted in a lot of discussions I don't see any reason to rush into this
00:53 Eater4 joined #minetest-dev
00:53 Gethiox joined #minetest-dev
01:12 hmmmm hmmm
01:12 hmmmm yeah, I'm not sure what the rush is, but then again I don't see why we should keep them either, and it's just a simple merge...
01:12 hmmmm honestly though, I don't want to commit things unless they're my own improvements
01:13 hmmmm I am sick of being in this rut where I can't get anything done
01:13 hmmmm it's not that I need motivation, I just need to start feeling better
01:14 hmmmm nevertheless i am disappointed in myself for doing exactly zilch for the past.... I don't even know how long
01:14 hmmmm I'm going to put the biome/mapgen improvements on hold again, because it's the testing crap that takes such a long time
01:14 hmmmm it's like I can never get anything that's nice enough
01:14 hmmmm really discouraging
01:14 paramat cool, no rush. i wouldn't want anything to delay 0.4.10 any longer
01:15 sapier hmmmm? maybe you're to perfectionistic sometimes ;-)
01:15 hmmmm haha doubtful
01:16 hmmmm I don't really like it, but in order to get some of the features I want in mapgen v7, i'm going to need to hard code them
01:16 hmmmm i'd rather have an all-math solution
01:16 hmmmm the biome thing is a wreck
01:17 sapier what about doing a good suggestion and ask for comments?
01:17 sapier sometimes explaining a solution to someone else makes you discover a even better one
01:21 hmmmm because talk is cheap
01:21 hmmmm see, when I was feeling better, I was demotivated and I didn't have as much time
01:22 hmmmm now that I am motivated and have more free time, there's some medical condition wrong with me and I am in near constant pain and can't focus on jack shit
01:22 hmmmm :(
01:22 hmmmm this thing has really been killing my productivity
01:22 Taoki joined #minetest-dev
01:23 sapier well in this case I suggest focusing on getting fit again, health is always way more important then any hobbys
01:23 hmmmm that's the thing though
01:23 hmmmm I can't really do anything about it on my own
01:24 hmmmm I am waiting on test results and doctors appointments and all this crap and it takes like months to make any progress at all
01:24 sapier that's crap true
01:24 hmmmm pretty sure I'm going insane because of this whole thing too
01:25 hmmmm hehe
01:25 hmmmm i'm sure it'll get fixed eventually, and then in a month I'm starting my new job and I'll have even more time since it's remote
01:27 sapier good :-) keep thinking positive it might help getting healthy again ... thinking the other way wont help for sure
01:27 hmmmm well if it turns out to be something like ulcerative colitis, I'm screwed since that's a chronic lifelong condition
01:28 sapier then thinking positive is even more important
01:28 hmmmm that's probably my biggest fear, is this turning out to be something i'm going to deal with for the rest of my life
01:33 sapier none of us knows wha't life's prepared for us ... thinking positive always helps
01:35 hmmmm we'll see......
01:35 hmmmm what have you been working on lately?
01:35 sapier android port is almost ready to merge
01:35 hmmmm by the way, great job on the performance improvements
01:35 hmmmm it's like no matter what we do, though, minetest is always slow
01:35 sapier there are some ugly things left in it but I think it's good enough to get it tested on a little bit more devices
01:36 sapier especially for imroving controls we need community feedback
01:36 hmmmm I wonder if I should take a break from mapgen things and work on improving graphics
01:36 hmmmm there are a lot of things that could be done
01:36 sapier true, there are some design decisions beeing quite strange and cause a lot of performance drop
01:37 hmmmm the irrlicht scene node setup especially comes to mind
01:37 hmmmm apparently there's no good reason why the entire thing is just one scene node
01:37 sapier at least on first glance ... I don't know that code good enough for a final judgement so I did focus on those parts where performance is wasted within relatively small parts of code
01:38 hmmmm if we make each MapBlock a scene node, we'd have occlusion culling
01:38 sapier sounds reasonable
01:38 hmmmm right
01:38 hmmmm so did I tell you this idea before?
01:39 sapier yes that idea was mentioned every now and then but noone seems to really have done work
01:39 hmmmm each MapBlock will have 3 scene nodes, 1 for solid materials, 1 for translucent, and 1 for liquids.. I originally said 2 but RBA wanted a third for liquids (not sure if that'll have any effect though)
01:39 hmmmm right, so that'll have to be what I'll work on
01:44 paramat interesting, will it fix translucency ordering between mapblocks? (but not within)
01:44 hmmmm no, within
01:44 hmmmm the idea is to have separate scene nodes for material types
01:45 hmmmm so we'd be fixing two problems at the same time
01:47 sapier left #minetest-dev
02:00 paramat also there's been interest in irrilicht particles, due to their speed, collision detection not being needed for some purposes
02:02 paramat (irrlicht)
02:15 kahrl hmmmm: but... say there are two semitransparent windows in the same mapblock. They would be in the same scene node (because they have the same material). How would this fix their draw order?
02:18 hmmmm there isn't a way to z-order?
02:18 hmmmm hmm no
02:19 hmmmm this is interesting
02:19 hmmmm so the clouds have their own scene node, but why do clouds interfere with water and other transparent materials on the map node?
02:22 kahrl I guess because the transparency sorting fails because the map scene node has a huge bounding box
02:22 kahrl s/because//
02:22 kahrl so irrlicht doesn't know whether to draw the clouds first or the map, so it just does it the same way around every time
02:24 hmmmm oh, that makes sense
02:24 kahrl this stuff should be fixable if every mapblock is a scene node (there would still be problems if the mapblock overlaps the clouds, but those a few)
02:24 kahrl are*
02:25 hmmmm well rather, every mapblock will *have* a scene node
02:25 kahrl yeah
02:25 hmmmm I am not going to extend a mapblock with ISceneNode.... ew..
02:25 kahrl hehe, that'd be fun on the server side as well
02:26 hmmmm oh yeah!
02:26 hmmmm so now do we have a ClientMapBlock with ISceneNode*s?
02:27 kahrl well, MapBlock already has a MapBlockMesh* member
02:27 kahrl only compiled in in the client
02:27 hmmmm mm yeah
02:27 hmmmm I'll follow suit then
02:39 hmmmm hmmmm
02:39 hmmmm sapier, why did you declare statenames[] in clientiface.h
02:39 hmmmm that gets compiled every single time clientiface.h is included
02:47 hmmmm dammit
02:47 hmmmm I wish enums were namespaced like they are in Java
02:48 hmmmm symbols like Invalid and Init are in the global namespace now
02:48 hmmmm symbols like Invalid and Init are in the global namespace now
02:48 hmmmm whoops
03:04 hmmmm damn
03:04 hmmmm C++11 has solved this problem with "enum class"
04:56 stormchaser3000 joined #minetest-dev
05:28 kaeza joined #minetest-dev
05:59 Endymion joined #minetest-dev
06:04 hmmmm https://github.com/kwolekr/minetest/commit/c91f8b1d251c81dcfc398c9ca2a4d79086212839
06:06 hmmmm fixes a metric crapton of warnings, prevents the same string table from being duplicated 500 times in each object file, removes common words like "Init" from the global namespace, fixes horrible style issues (e.g. foobar(BLAH,foo); - the parameters look (to me, anyway) like foo is a member of BLAH because my font size isn't big enough)
06:13 Exio4 joined #minetest-dev
06:19 Endymion left #minetest-dev
06:20 Hunterz joined #minetest-dev
06:28 asl joined #minetest-dev
06:57 Krock joined #minetest-dev
06:57 Krock left #minetest-dev
07:06 RealBadAngel joined #minetest-dev
07:44 restcoser joined #minetest-dev
07:55 sfan5 whoever added the "feature request" tag
07:55 sfan5 a pull adding a feature is not a "feature request" but an "(low-priority)? enchantment"
08:14 sapier joined #minetest-dev
08:21 sapier hmmmm that clientstate variable is an error I'm gonna fix it
08:31 RealBadAngel sapier, http://i.imgur.com/Gms0nLH.png
08:33 sapier can't open that file
08:33 RealBadAngel ^^ full special tiles support is already here, just changing constant from 2 to 6
08:41 VanessaE sfan5: an enchantment?  heh.  what is this, sorcery?
08:42 VanessaE honestly, I would have to say not to mark an enhancement as "low priority" if fully-working code is provided.
08:43 sapier that marking doesn't tell anything about code quality
08:44 VanessaE well let's assume for the moment that "fully working" implies good code.
08:45 VanessaE and indeed, it doesn't - good quality, fully-working code can be marked "low priority" just as easily as the opposite
08:45 sapier I did assume that but "enhancement" or "low-prio enhancement" still don't tell anything about it ;-)
08:46 VanessaE (I'm not saying such a case has happened recently, I'm just trying to avoid people using those labels as a way of avoiding something they just don't care about)
08:47 sapier I'd suggest using the milestones to flag things that are almost sure to be included in next version
08:47 sapier and I'd rename the double milestone to 0.4.11 ;-)
08:47 VanessaE heh
08:52 sapier hmm seems like kahrl already fixed that issue I'm working at right now ... maybe I should have a look at commit history first
08:53 LemonLake joined #minetest-dev
09:10 CraigyDavi joined #minetest-dev
09:11 LemonLake joined #minetest-dev
09:24 Calinou joined #minetest-dev
09:26 proller joined #minetest-dev
09:39 proller sooo..  i'm merging ? https://github.com/minetest/minetest/pull/1390 https://github.com/minetest/minetest/pull/1407 ?
09:39 sapier I disagree
09:39 proller but you never use it
09:40 sapier I don't wanna have any discussions about removed features 2 days prior feature freeze they wont harm and can be removed for 0.4.11 too ... no need to rush this
09:40 proller 1390 - nobody never use, exept me
09:41 sapier may or may not be true, I just can't verify it
09:42 sapier once 0.4.10 is out I'll not bother to stop you about merging them but if you wanna do this now get explicit agreement of some other coredev first
09:44 sapier ShadowNinja: https://github.com/minetest/minetest/issues/1423 I guess that one is best to be fixed by you. most likely you already know what's happening on reading the desc
09:46 sapier https://github.com/minetest/minetest/issues/1389 anyone to fix that issue?
09:48 ImQ009 joined #minetest-dev
10:17 LemonLake joined #minetest-dev
10:23 lemonlake_ joined #minetest-dev
10:25 PenguinDad joined #minetest-dev
10:26 Garmine joined #minetest-dev
10:35 PilzAdam joined #minetest-dev
10:49 Jordach joined #minetest-dev
10:58 ZOOOO joined #minetest-dev
11:16 Exio4 joined #minetest-dev
11:16 ShadowNinja joined #minetest-dev
11:19 Taoki[mobile] joined #minetest-dev
11:19 CraigyDavi_ joined #minetest-dev
11:42 Krock joined #minetest-dev
12:23 sapier joined #minetest-dev
13:00 vifino joined #minetest-dev
13:02 Krock the problem of issue #1421 must be there somewhere: https://github.com/minetest/minetest/blob/master/src/player.cpp#L287
13:10 Krock 15:08:12: INFO[main]: Failed to write E:\Programme\minetest\bin\..\worlds\next_test\players\singleplayer
13:11 Krock there seems to be a .close_file() function required after opening the fle
13:13 Krock https://github.com/minetest/minetest/blob/master/src/filesys.cpp#L690
13:14 Krock why is a temponary file required?
13:14 Krock celeron55!
13:16 sfan5 Kray: to make sure files are not corrupted when there is a power outage
13:17 Krock *Krock
13:17 sfan5 oh yeah
13:17 Krock well, something in that function return false and disallows from overriding singleplayer
13:18 Krock seems like a file"stream" has not been closed
13:18 sfan5 if it returns false it either means it could not open the file for writing or write the data to file
13:19 Krock or it couldn't delete & rename it
13:19 Krock that's the 2nd possibility
13:20 sfan5 https://github.com/minetest/minetest/blob/master/src/filesys.cpp#L345
13:20 Krock because, I have a up-to-date singleplayer.~mt and an old singleplayer file
13:20 sfan5 wtf
13:21 sfan5 you will never need to delete something in a path that is about 10000 chars
13:21 sfan5 or even 100
13:21 sfan5 1000*
13:21 sfan5 or even 250
13:22 sapier that's quite ugly as there's a define containing max path length for most os
13:43 Megaf joined #minetest-dev
13:50 Krock guys, SEIOUSLY?
13:51 Krock I fixed that player problem
13:51 Krock putting on th server on now
13:51 sapier create a pull request containing the fix ;-)
13:51 Krock would ned to update my fork and I never get that done :/
13:52 Krock git hates me
13:52 Krock 2 lines are missing
13:52 sapier you can do a completely new clone too
13:52 Krock src/environment.cpp line 464
13:52 Krock testplayer.deSerialize(is, path);
13:52 Krock is.close();
13:53 Krock src/player.cpp line 316
13:53 Krock testplayer.deSerialize(is, path);
13:53 Krock is.close();
13:53 Krock done
13:57 Rhy5 joined #minetest-dev
13:57 Rhy5 left #minetest-dev
13:59 ZOOOO joined #minetest-dev
14:10 tomreyn joined #minetest-dev
14:13 Krock profed 2 times, it actually fixed the problem
14:45 Taoki[mobile] joined #minetest-dev
15:05 Krock https://github.com/minetest/minetest/pull/1424
15:08 Krock sapier ^
15:28 iqualfragile joined #minetest-dev
15:56 cg72 joined #minetest-dev
16:11 stormchaser3000 joined #minetest-dev
16:11 hmmmm joined #minetest-dev
16:13 ImQ009 joined #minetest-dev
16:24 Megaf joined #minetest-dev
16:32 Calinou joined #minetest-dev
16:33 zat joined #minetest-dev
16:45 Calinou why do I have a short freeze when the game places/sets a node through an ABM, but not when I place a block by hand?
16:45 sapier VanessaE could you please review: https://github.com/sapier/minetest/blob/android_merge_branch_21/doc/README.android AND https://github.com/sapier/minetest/blob/android_merge_branch_21/src/game.cpp L1003+
16:46 sapier I don't think it's related to the placing calinou but to the abm handling itself
16:48 Calinou to change lots of nodes and handle plant growth, what is better: low interval and high chance, or the inverse?
16:48 sapier btw I intend to merge the android patch tonight (quite late) so please review
16:48 sapier if there are a lots of nodes there both is bad
16:48 sapier abm's actually are only usefull if they handle quite rare nodes
16:49 Calinou “A low interval and high chance spreads the load”, says leaf decay ABM, but is this applicable to plant growth?
16:49 sapier what node do you want to apply the abm to?
16:49 Calinou farming:soil, farming plants, saplings…
16:50 Calinou I use interval 2 for mob spawning already, works nicely (used on lots of nodes: stone with air nearby)
16:50 Calinou this ABM spawns entities, not nodes
16:50 sapier Calinou I really doubt that works fine ;-)
16:51 Calinou it does work fine, CPU usage is low
16:51 Calinou much better than using high interval, which causes freezes
16:51 sapier but I have to leave long intervals with high chances used to cause major spikes in cpu usage as all abms have been called same time
16:51 sapier but there have been ideas how to fix it not sure if they got merged
16:54 Calinou I'm attempting interval 2 and 1 for leaf decay
16:55 Calinou low interval and high chance works better
16:55 Calinou it's smoother too
17:00 Sokomine sapier: it's great if that gets merged. please provide it in the f-droid appstore. i think that one will fit nicely to minetest
17:02 Sokomine leaf decay sometimes takes too much time on crowded servers - saplings fall when the player has long moved on. perhaps we need a better solution for that anyway
17:04 rubenwardy joined #minetest-dev
17:18 rubenwardy_ joined #minetest-dev
17:27 Calinou provide links to .apk on minetest.net too
17:27 Calinou and don't forget to add a disclaimer that some copies of it are illegal
17:29 kaeza joined #minetest-dev
17:36 hmmmm sapier, ???
17:44 proller merge https://github.com/minetest/minetest/pull/1390 ?
17:49 lemonlake_ is this for freeminer?
17:50 proller just remove useless stuf
17:51 Amaz lemonlake_, not, it's for minetest...
17:51 lemonlake_ mhm I realised
17:55 sfan5 proller: https://github.com/minetest/minetest/issues/1422
17:56 proller its ShadowNinja's stuff
17:56 sfan5 >.>
17:56 sfan5 ShadowNinja: ^
17:57 sfan5 the person who will do the MSVC builds should note this here: https://github.com/minetest/minetest/releases/tag/untagged-c8b0fdef1d9454a338c5
18:01 smoke_fumus joined #minetest-dev
18:40 kaeza joined #minetest-dev
18:40 vifino joined #minetest-dev
18:44 ShadowNinja celeron55: It requires Python (3) and the Flask and APScheduler pypi packages.  sfan5: Have you tested the Python one?  (still reading backlog...)
18:49 sfan5 ShadowNinja: no
18:50 sfan5 +t yet
19:00 Jordach sfan5, 404 ;P
19:01 kaeza how is the player password hashed again?
19:09 kaeza nvm, found it
19:09 PilzAdam sfan5, add "FIx blocker bugs" to the list
19:19 Megaf folks, got a crash when I did //pos2
19:19 Megaf http://paste.debian.net/107203/
19:20 Krock what's the status of this one? https://github.com/minetest/minetest/pull/1424
19:23 ShadowNinja sfan5: You'll probably want to start it on a port like 8000 and have nginx proxy the correct address to it.  Also, you should enable X-Sendfile in that case.
19:27 werwerwer joined #minetest-dev
20:13 vifino joined #minetest-dev
20:34 nore joined #minetest-dev
20:42 sapier joined #minetest-dev
20:43 CraigyDavi_ joined #minetest-dev
20:59 sapier https://github.com/minetest/minetest/pull/1417 if there are no open issues I'm gonna merge this in a few minutes
21:00 CraigyDavi joined #minetest-dev
21:29 cg72 joined #minetest-dev
21:43 VanessaE joined #minetest-dev
21:54 VanessaE sapier: FAIL :)
21:55 sapier can you be a little bit more specific? ;-)
21:55 VanessaE #184
21:58 sapier hmm ok ... should've read the description ... my comment would've been "do I have to put a green sock to my nose too?" ;-)
21:58 VanessaE LOL
21:59 VanessaE "Start the game and create a new world using minetest_game." would have given you a clue ;)
22:00 sapier Why is this supposed to be a regression I'm almost sure part of engine responsible didn't ever change
22:00 VanessaE along with "Do not install any mods..."
22:00 sapier the basic issue is "have a lot of map blocks loaded
22:00 sapier "
22:01 VanessaE at the time I filed that bug, I used to be able to get 120+m of view range and ~30 fps in worlds like redcrab.suret.net which if you recall is very OLD
22:02 VanessaE now, I can't even get a third of that performance and that world does not change.  no new mods (and thus, no changes to the texture pack support), no new server engine
22:02 VanessaE at the time I filed that bug, I had began to notice a fall-off in my view range and fps.
22:02 sapier could you try recent version with those worlds? if this is really related to amount of loaded blocks my recent performance fixes could have a recoginzeable effect
22:02 VanessaE I have tried.
22:03 sapier what's been the result?
22:03 VanessaE markedly lower performance than it used to be previously.
22:03 sapier previously is supposed to be which version?
22:04 VanessaE I would guess, after your recent fixes, I'm back to about half of what I used to get back then
22:04 VanessaE previously would be early 0.4.x
22:05 sapier did you ever check this? ;-)
22:05 VanessaE what do you mean?
22:06 sapier last time I did compare 0.3 to 0.4 performance was even better at 0.4, and that's been previous my fixes ... so knowing which version was the fast one would help
22:06 * VanessaE tries redcrab...
22:06 VanessaE feh. network protocol is busted again.  grey screen.
22:07 sapier new client old server?
22:07 VanessaE (that server still, if you recall, runs 0.4.4)
22:07 VanessaE yeah
22:07 sapier strange
22:07 sapier thought that combination was checked
22:08 VanessaE I see all my stuff in my inventory/hotbar
22:08 VanessaE but the world does not load
22:08 sapier what's server ip?
22:08 VanessaE redcrab.suret.net 30401
22:10 sapier works for me
22:11 VanessaE my local copy is at commit 95c7f35a
22:11 PenguinDad Works here too
22:12 VanessaE lemme try updating.
22:15 VanessaE nope.avi
22:15 VanessaE deader than 4'oclock for me.
22:15 VanessaE git HEAD now.
22:16 sapier disable curl
22:16 sapier for what I know it's broken on that server
22:17 VanessaE um, that server should predate cURL entirely.
22:18 * VanessaE tries...
22:18 VanessaE enable_remote_media_server = false  in my .conf
22:18 VanessaE and nope, that doesn't help.
22:20 sapier are you sure that's really gonna work?
22:21 PenguinDad VanessaE: try deleting your cache
22:21 VanessaE matter of fact EVERYTHING is loading slowly today, even my totally idle Creative server.
22:21 sapier hmm ok didn't work second time
22:22 VanessaE as in, everything is failing to load today beyond a certain point
22:22 VanessaE I can't get more than one mapblock on my Creative server.
22:22 VanessaE PenguinDad: that ain't gonna do anything, that's just to receive the media, not to receive the map.
22:23 VanessaE and no, my Creative server is NOT locked up
22:23 VanessaE (it's only using about 2% CPU)
22:23 VanessaE this represents a big problem I think.
22:25 VanessaE same issue if I sign out/back in
22:25 sapier but he's right clearing cache will work around the issue
22:25 hmmmm sapier, was there something wrong with what I committed last night?
22:25 VanessaE I only get a couple of mapblocks and the server stops responding to my actions - I can't open doors, for example.
22:25 hmmmm i thought you mentioned something about a certain variable not being a state
22:26 sapier true there's been a change today ... wait I'll check more carefully
22:27 hmmmm and PLEASE don't pollute the global namespace with very common names like "Init"
22:27 hmmmm grr
22:27 sapier I assumed you did formal changes only  ... you're right about that ... now idea why I did that
22:29 VanessaE ok, I also do not see the sign-out message (via IRC) when I leave the dysfunctional server (I'm talking about my Creative server here).
22:30 sapier no those changes don't look like they could break anything
22:46 sapier I'm gonna merge this one https://github.com/minetest/minetest/pull/1424
22:46 CraigyDavi` joined #minetest-dev
22:53 ImQ009 joined #minetest-dev
23:21 cg72 joined #minetest-dev
23:21 cg72 left #minetest-dev
23:21 Sokomine i can't connect to redcrabs server with recent clients either - which is very sad as that's a good server with many valuable buildings
23:23 VanessaE redcrab needs to get off his duff and update the server :P
23:23 Sokomine that would be best. i'd love to see it on the public list
23:28 sapier Sokomine: works exactly one time after clearing cache
23:30 Sokomine oh. hm. that is odd
23:30 Sokomine if you know with so much precision: is there a workaround?
23:30 sapier yes it's a bug in httpfetch mechanism
23:31 sapier I don't think so it's a wrong assumption in that code, old servers don't behave that way
23:31 Sokomine if it could be fixed, that would be great. a great server could then be used again
23:31 Sokomine it's also a good sign of stability if such an old server still works :-)
23:31 sapier it's fixed in current but requires a change on both sides
23:32 VanessaE which means it isn't fixed - because redcrab will never make such a change.
23:32 VanessaE he's never around
23:32 Sokomine oh, well...next reply from redcrab might be next year if his current rate of reply allows any predictions
23:32 VanessaE oh you meant for current servers, derp
23:33 kaeza joined #minetest-dev
23:33 sapier yes Sokomine but it's either drop httpfetch completely or be buggy ... decide yourself
23:33 Sokomine and i think it's a good sign that a server might work with *that* little maintenanc and over such long time with diffrent versions
23:35 VanessaE sapier: aw come on, if it can work once, it can work repeatedly.,
23:35 * VanessaE pokes kahrl
23:36 sapier problem is httpfetch assumes to get some message but it doesn't
23:36 VanessaE then add a timeout.
23:37 sapier no
23:37 sapier I will not add a timeout causing intialization to be mixed up again once something takes more time then usuall
23:37 sapier y
23:57 SmugLeaf joined #minetest-dev

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