Minetest logo

IRC log for #minetest-dev, 2015-08-19

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

All times shown according to UTC.

Time Nick Message
00:01 paramat merge complete
00:02 RealBadAngel #3079 should be merged for it too
00:02 ShadowBot https://github.com/minetest/minetest/issues/3079 -- Remove use of engine sent texture tiling flags - theyre no longer needed by RealBadAngel
00:04 paramat yeah
00:12 paramat RealBadAngel, shall i go ahead and merge 642 and we can change tree tops later? or are you now updating that PR?
00:12 VanessaE NOOOOOO
00:12 VanessaE 642 is a no-go
00:13 paramat because of issues stated earlier?
00:13 VanessaE yes
00:13 VanessaE it will break nearly every existing texture pack and some mods as well
00:13 RealBadAngel it doesnt break anything
00:14 RealBadAngel texture packs makers shall update their TPs in order to use engine features
00:15 paramat i don't completely understand the arguments. texture packs without normals means just don't use normalmaps, seems reasonable
00:15 VanessaE no
00:15 VanessaE texture packs without normals means they'll end up using these new default normals
00:16 VanessaE (if the relevant shaders are enabled)
00:16 RealBadAngel VanessaE, your request means effectively disable parallax mapping
00:16 VanessaE where did I say that?
00:16 RealBadAngel it does mean that
00:16 VanessaE supply normals but don't load them if the matching diffuse/color texture is in a different folder
00:16 VanessaE what's so hard about that?
00:17 VanessaE same folder?  load them, by all means.
00:17 RealBadAngel ability to replace one map with the other
00:17 RealBadAngel you want to break another feature
00:17 VanessaE better to force those few people to supply copies of the matching texture than to force every texture author to supply 300+ blank normalmap images
00:18 paramat we already have some normalmaps so a few more in 642 doesn't change anything, whether the implementation is right or wrong
00:19 VanessaE then fix the implementation before 0.4.13 goes out
00:20 RealBadAngel fix the feature by disabling it?
00:20 VanessaE ...
00:20 RealBadAngel every tp out there is incomplete
00:21 RealBadAngel i should rather remove autogen from core - because its absurdly costly - than implement your request
00:24 paramat well normalmaps are an optional shader thing, many/most texture pack authors won't be adding normalmaps, using such a pack means players should not enable bumpmaps
00:25 RealBadAngel https://forum.minetest.net/v​iewtopic.php?f=4&t=8714
00:25 paramat certainly the implementation could possibly be improved, we need other's input on this
00:25 RealBadAngel VanessaE, your proposal would mean above impossible
00:26 VanessaE RealBadAngel: so?  special handling of certain corner cases is not a bad thing.
00:26 VanessaE that ^^^ is one such place.
00:28 paramat perhaps *ideally* if a texture pack does not supply a matching normalmap for a texture, the normalmap should be set to flat
00:28 VanessaE that'
00:28 VanessaE that's what I was proposing
00:28 VanessaE hence the "same folder" requirement
00:28 paramat yeah agreed
00:29 RealBadAngel while i can agree to use flat normal map i cant for "same folder"
00:31 est31 ok  #3081 is out
00:31 ShadowBot https://github.com/minetest/minetest/issues/3081 -- Fix inventory replace bug by est31
00:32 paramat sorting this out may have to wait for 0.4.14, how it is at the moment may not be perfect but also not broken
00:32 paramat (talking about normalmaps)
00:34 paramat i will merge 642 later since it is just extra normalmaps
00:40 paramat back later
00:40 paramat left #minetest-dev
00:42 Wayward_One est31: have you seen my latest comment yet?
00:44 est31 Wayward_One, yes
00:44 est31 I can however not reproduce the bug, with that apk
00:44 Wayward_One Hmm
00:44 RealBadAngel ive googled a bit for that
00:45 est31 seems we'll have to hotfix it.
00:45 RealBadAngel looks like its something about texturing
00:46 RealBadAngel http://irrlicht.sourceforge.ne​t/forum/viewtopic.php?t=43774
00:47 RealBadAngel this seems to be related
00:47 est31 VanessaE, can you check the new error message in 3058?
00:57 VanessaE I like the message, but shouldn't you store that in a variable somewhere and print *that* (plus "technical code" bit for the first call)?
00:58 est31 ?
00:59 est31 "that"
00:59 VanessaE you print the exact same message twice, with some add-on text in the first case.
00:59 VanessaE or rather, you supply it twice.
01:00 est31 Perhaps one can argue that as its two different sentences, all languages can separate
01:00 VanessaE shouldn't you do something like msg=""Oooops! Seems that mod list download failed. It might be that this is a problem with the website, caused by the descision to take the mod store offline, or a bug in the game." and then fgettext_ne(msg.." The technical code is: ")..result or fgettext_ne(msg)  ?
01:00 VanessaE something like that anyways
01:01 VanessaE ok, fair enough
01:02 est31 refresh
01:02 est31 better now?
01:02 VanessaE right, exactly that
01:03 VanessaE I'm good with that, then
01:03 est31 slightly adjusted the messag
01:03 est31 e
01:04 est31 now waiting for hmmmm to approve 3081 and 3058
01:04 VanessaE better.
01:04 est31 (or any other devs)
01:05 VanessaE he's still taking a shit :)
01:05 est31 long shit
01:05 VanessaE lincoln log ;)
01:35 hmmmm 3081 looks good to me
01:43 est31 seems I'll have to re-apply https://github.com/minetest/minetest/commi​t/60f31ad52348c53209701d0112f9f809018f7188
01:43 est31 @ least Wayward_One can reproduce a bug with the method
01:45 hmmmm okay then
01:45 hmmmm that for the blank menu screen bug?
01:46 est31 yes
01:47 hmmmm as for the error message text on 3058
01:47 hmmmm that's "friendly" but it doesn't exactly comfort users
01:47 hmmmm "oops, something went wrong and it could be any number of things"
01:47 hmmmm how about something like
01:48 hmmmm "Oops, something went wrong with getting the mod database list!"
01:48 hmmmm "Try again at some later time."
01:48 hmmmm "If the problem persists, it may be a bug.
01:48 est31 no
01:48 est31 operational issues may be a bug too
01:49 hmmmm do you want error reports up the wazoo
01:49 est31 but "try again at some later time sounds good"
01:49 hmmmm cause this is how you get them
01:49 hmmmm est any modstore PRs are your responsibility from now on
01:49 hmmmm problem reports
01:49 est31 no thank you
01:49 est31 the easiest way to not get error reports is by just not having that button
01:50 est31 the "mod store" button
01:50 VanessaE I tend to agree
01:50 hmmmm well I don't get it
01:50 hmmmm did celeron just decide to get rid of MMDB?
01:50 VanessaE didn't it get lost when xyz gave up hosting the forum?
01:50 est31 VanessaE, yes
01:51 est31 (AFAIK)
01:51 VanessaE plus let's face it, iqual was....rather stubborn about how mods are added and approved.
01:51 est31 iqual?
01:51 VanessaE iqualfragilew
01:51 VanessaE -w
01:51 hmmmm okay
01:51 VanessaE the guy who managed the actual website used for submitting mods to the store
01:51 hmmmm soooooo
01:52 est31 ok
01:52 hmmmm what do you two want to do
01:52 hmmmm are we not confident in the MMDB coming back
01:52 VanessaE cornernote is working on something that should be usable to replace it
01:52 est31 the future of the mmdb is something like that bower thingy
01:52 VanessaE but it's not ready right now
01:52 est31 ea
01:52 est31 yea*
01:52 hmmmm well
01:53 hmmmm disable the button if you want
01:53 hmmmm just remember the entire version is getting shipped like this
01:53 est31 but enabling untrusted mods needs security to be set on by default
01:57 est31 kahrl, sfan5, other devs, your opinions on removing mmdb support entirely? Should mmdb ever come back? VanessaE and me say no.
01:58 paramat joined #minetest-dev
02:02 est31 Wayward_One, can you try a commit for me?
02:02 Wayward_One Sure
02:03 est31 this one : https://github.com/est31/minetest/commit/​6452db11706a337124d6c485b4a66478bc8b42e9
02:03 Wayward_One Ok
02:09 paramat RealBadAngel i'll merge 642 in a moment
02:11 sgtbigman joined #minetest-dev
02:22 paramat now merging game 642
02:26 paramat merge complete
02:27 est31 and, I almost forgot, today's an important date
02:27 est31 the website competition ends
02:27 est31 The astounding number of 0 (zero) entries were officially submitted
02:28 est31 I only have two half-official entries
02:28 est31 one from Calinou
02:29 est31 one from srifqi
02:30 est31 they both lack the "post in topic" requirement
02:30 est31 but its a very formal one, so I guess both can count as submissions
02:39 paramat apparently 3079 can go in now that game 641 is in
02:40 VanessaE #3079
02:40 ShadowBot https://github.com/minetest/minetest/issues/3079 -- Remove use of engine sent texture tiling flags - theyre no longer needed by RealBadAngel
02:41 est31 game#641
02:41 ShadowBot https://github.com/minetes​t/minetest_game/issues/641 -- Add tiling info for grass nodes by RealBadAngel
02:41 est31 why does 3079 have a requirement for 641 to be merged?
02:42 paramat perhaps not, i don't understand the issues
02:44 est31 3079 is unrelated from how I see it
02:44 est31 they both cover tiling flags though
02:44 Hunterz joined #minetest-dev
02:44 paramat i misunderstood RBA then
02:45 VanessaE the way I read it, 3079 removes obsoleted code.
02:46 VanessaE and 641 adds in node-def-based flags to replace it.
02:46 est31 also, it seems that 3079 is a speedup
02:46 est31 however, I wonder whether we should merge it so short before a release
02:47 VanessaE may as well
02:47 VanessaE if release has to be put off a day or two as a result, that's not very harmful; it can always be reverted if push comes to shove.
02:49 est31 no
02:49 est31 reverting release commits is no good
03:09 paramat left #minetest-dev
03:20 sprinkl joined #minetest-dev
03:21 Miner_48er joined #minetest-dev
03:35 Hijiri joined #minetest-dev
03:37 sloantothebone joined #minetest-dev
03:37 Hunterz joined #minetest-dev
04:39 tuy joined #minetest-dev
05:11 est31 joined #minetest-dev
05:15 est31 ok plan for today: lets determine what website to chose from the two, then let make celeron55 do the switch, and wait for paramat/other game devs to label the 0.4.13 release
05:15 est31 also, determine whether to ship with mmdb, or whether to fix it, or what the error messages should contain, if not fixing mmdb
05:20 hmmmm i really would prefer for mmdb to stay in minetest
05:25 Hijiri joined #minetest-dev
05:27 est31 The current mmdb implementation is very bad.
05:27 est31 e.g. its fullscreen for some reason
05:28 est31 which means the tab buttons are right where the minetest version text is
05:28 est31 also, the mods in mmdb don't reflect the rich set of mods minetest has
05:29 est31 also, we should have a subgame store as well
05:29 est31 chosing mods is hard
05:29 est31 perhaps however the subgames should even be shipped with minetest
05:29 est31 but then people who count every single bit wont like it
05:32 sfan5 est31: mmdb needs a lot of work, but i'm not for removing support entirely
05:34 est31 I don't want to make a release with mmdb not being online.
05:34 est31 at the end of the day celeron55 says he doesnt want to host mmdb, and our users have a non functional button
05:35 est31 hmmmm even fears that we would drown in error messages.
05:36 hmmmm celeron actually said that he's not interested in continuing MMDB?
05:36 hmmmm ugh
05:36 hmmmm dammit
05:37 est31 dunno
05:39 est31 thats part of the things i want to find out
05:39 est31 and get cleared
05:39 nanepiwo who hosts the skins database?
05:39 sfan5 which skin database?
05:40 est31 whats the url again?
05:40 nanepiwo minetest.fensta.bplaced.net
05:41 est31 seems to be an user named "addi".
05:41 est31 you can contact them in the forum thread I guess: https://forum.minetest.net/viewtopic.php?id=4497
05:42 nanepiwo I see
05:47 hmmmm oh speaking of skins
05:47 hmmmm est31, you know vanessae's skin compatibility problem?
05:47 hmmmm is that a blocker or not?
05:48 est31 yes I know of that backface problem, if it is what you mean
05:48 est31 I dont think it is a blocker
05:48 hmmmm yeah
05:48 hmmmm ok fair enough
05:49 hmmmm i don't either tbh
05:49 hmmmm if you get everything in, we can release tomorrow or whatever
05:50 Hunterz joined #minetest-dev
05:50 est31 tomorrow for you
05:50 est31 I hope to release when paramat gets around next, and we have cleared the mmdb issue
05:51 est31 before you go offline: what would be an acceptable message for my fix PR?
05:52 est31 I remember you critizised that the message listed the number of possible reasons
05:53 est31 my motivation to add that was so that I both can say "it can be because we had disabled mmdb"
05:53 est31 and not say "forget mmmdb, we have turned it off"
05:53 hmmmm well
05:54 hmmmm so the main problem we're trying to solve is to let users know that the MMDB does exist, it might not be working *right now*, but try again sometime in the future before filing a bug report about it
05:54 hmmmm so it needs to be reassuring
05:54 hmmmm as well as provide data about real errors if they are real
05:55 est31 yes
05:55 hmmmm I honestly don't know
05:55 hmmmm I agree with your stance against basing the error message off of HTTP error codes now
05:55 hmmmm it could be a legitimate bug
05:56 hmmmm if the user gets a 404 how do we know that's caused by the webserver or a bug in minetest requesting the wrong page
05:56 hmmmm etc.
05:56 hmmmm dare we ask the users to troubleshoot on their own a little bit first before submitting an error report?
05:58 hmmmm "Oops!  Unfortunately we are having trouble retrieving the MMDB."
05:58 hmmmm <technical error message here>
05:58 hmmmm " - The MMDB could be down temporarily.  Try again soon."
05:59 hmmmm " - If you have not been able to load the MMDB after several attempts, check your computer's network connection."
05:59 hmmmm ??
05:59 hmmmm or something like that
06:01 kahrl how about inserting some comment like <!-- mmdb_disabled = true --> into the 404 page for .../mmdb/*
06:01 kahrl which is checked by the client and changes the error message
06:02 est31 thats a good proposal
06:03 hmmmm yeah but we don't control the website
06:04 kahrl we don't?
06:06 est31 celeron55 does
06:06 kahrl yeah, he's been around quite often lately
06:07 est31 we will have to not use mmdb if we want to run a moderation less mod repository
06:07 est31 otherwise people with old clients could install malware with a single click
06:08 kahrl good point
06:08 Hijiri joined #minetest-dev
06:09 est31 ok made the official announcement in the forum https://forum.minetest.net/viewtopic.ph​p?f=14&amp;t=12843&amp;p=187611#p187611
06:12 thatgraemeguy joined #minetest-dev
06:14 est31 vote your favourite, you have 7 days
06:19 RealBadAngel joined #minetest-dev
06:31 johnwayne1986 joined #minetest-dev
06:49 celeron55 i really can't get the point of srifqi's entry, but Calinou's is surprisingly competent; it's better than the current one in some ways and not awful in others
06:51 nore joined #minetest-dev
06:55 celeron55 hmmmm: i haven't said i am not interested in continuing mmdb; not immediately setting it up was a result of many things; but now that it has been away for some time and we are starting to see stuff like minetest-bower, i feel like trying to get mmdb going again is not worth it in the long term
06:55 hmmmm alright, fine
06:56 celeron55 or may not be; the way this is implemented needs to be reconsidered
06:56 hmmmm remove the button
06:57 hmmmm @ est31
06:58 rubenwardy joined #minetest-dev
06:59 rubenwardy mmdb has been offline for 8 months, i don't think it's coming back
07:00 celeron55 if it is decided that mmdb is indeed the best choice, i can set it up again, seeing that its authentication has been separated from the forum database and some of the other issues fixed
07:01 rubenwardy hmmmm: I'd like to see hotbar selection changes in client side lua. Also CAO control
07:02 hmmmm you'll be able to do some rudimentary CAO manipulations like rotation changes and slight movements
07:02 rubenwardy celeron55 in it's current state mmdb is not practical. It takes too long to add a mod, it takes to long to release a new version
07:03 rubenwardy1 joined #minetest-dev
07:04 rubenwardy1 I think that client side lua should focus on client side prediction for mods and visual effects.
07:04 Guest18502 It should be possible to make client only visual objects (like SceneNodes)
07:04 Guest18502 ffs
07:05 Guest18502 joined #minetest-dev
07:05 RealBadAngel with #3079  tiling flags are still here, buts its enough to set material flags in the engine, no point doing the same using other method in shaders
07:05 ShadowBot https://github.com/minetest/minetest/issues/3079 -- Remove use of engine sent texture tiling flags - theyre no longer needed by RealBadAngel
07:06 RealBadAngel but for the flags to be working, they have to be used in tiles that need that, like dirt with grass - 641 does that
07:07 Guest18502 Minetest is currently too limited as a game engine, really you can only make simple Minecraft clones. I want to see more diverse gameplay.
07:07 tuy joined #minetest-dev
07:09 celeron55 rubenwardy_: stuff like custom player controls should be possible at some point too
07:09 celeron55 and custom player physics
07:10 celeron55 i mean, i'm not requesting features; just saying this is important in order to support any other things than walking on the ground
07:10 hmmmm my ultimate goal is to make minetest into an FPS
07:12 hmmmm not different games, different genres altogether
07:12 hmmmm we can emulate 2d games by restricting camera angles
07:12 celeron55 emulating 2d games is dumb in my opinion; there's way too much overhead in all the 3d handling for that
07:13 hmmmm we can do RTSes by restricting the camera angle and Y position of the local player
07:13 celeron55 (but i guess it should be possible if somebody insists)
07:13 hmmmm and units can be selected by punching them :p
07:13 Darcidride joined #minetest-dev
07:14 celeron55 well, ideally "punching" shouldn't really be a non-removable engine thing
07:15 celeron55 however, by going that way, i feel like minetest will become a worse version of Buildat 8)
07:15 hmmmm maybe
07:16 celeron55 but buildat has way too many design issues as where i left it that using it doesn't make sense either
07:16 celeron55 i guess continuing minetest's iterative history will work out fine enough
07:16 hmmmm right
07:16 hmmmm but the point of the iterations are to work toward a goal
07:16 hmmmm my own personal goal is to have the ultimate flexibility where we can actully tell people it's a game engine with a straight face
07:17 hmmmm and not look stupid
07:17 celeron55 i think that's the only way minetest can stay relevant and interesting in the long term
07:18 celeron55 there's going to be a group of people that still care about punching trees, but it will get very small and be limited to practically only children (because punching trees will always be a new thing to new people)
07:19 celeron55 and big corporations are better at pushing things to children than us
07:19 celeron55 and children are bad contributors anyway, and MT can't live without contributors
07:22 celeron55 hmmmm: when are you planning to have an initial version of client-side lua?
07:22 hmmmm whenever it gets done
07:22 celeron55 because i'm pretty sure i'm going to end up contributing to it once it exists, but i might not necessarily be starting it
07:23 celeron55 or, well, i could start it, if nobody else is planning to
07:23 hmmmm i already started it
07:24 hmmmm i don't know if you read this anywhere yet but it's basically a separate thread that does nothing but runs lua
07:24 hmmmm events are asynchronously added to the event queue and then callbacks are run in lua
07:26 nrzkt joined #minetest-dev
07:26 celeron55 sounds interesting
07:26 hmmmm the way lua works on the server side is that callbacks can be used to filter events before they actually happen
07:27 hmmmm this is really nice and clean design but it's dog slow
07:27 hmmmm it's unacceptable for a real time application like a game rendering thread
07:27 hmmmm so you'll get an event such as CSEVT_PLAYER_PUNCH for example which has the pos
07:28 hmmmm and then in that event callback you'll create a particle spawner or whatever that spews out a bunch of sparks
07:29 hmmmm or if you get a CSEVT_INVENTORY_MOVE you'll have to be aware of the fact that this event is telling you that "something happened" rather than "this specific thing is happening"
07:29 hmmmm which means checking the current state of the inventory before modifying it
07:29 hmmmm I'm just using these as dumb contrieved examples
07:32 hmmmm while this does seem like a big problem i don't think it will be too bad in practice because the game rendering loop takes far longer than the lua loop should
07:33 hmmmm i don't doubt you're going to have 33ms screen updates and for the lua callbacks to finish way before then
07:34 hmmmm even if a couple frames do pass by, it's not likely that the object or node or whatever you intend to reference is going to magically disappear within that timeframe
07:34 hmmmm and if it does disappear, then so what
07:34 hmmmm just abort the operation
07:34 celeron55 well, just keep in mind that if you do notice that it will become somehow wonky, this is the time to redesign it
07:35 hmmmm nothing can be wonkier than the way modding currently is
07:35 hmmmm sorry
07:35 celeron55 sorry?
07:35 hmmmm I don't mean to offend you
07:36 celeron55 "sorry" offended me, but nothing prior to that didn't; why do you assume i take this as a personal thing?
07:36 hmmmm because the entire premise behind the way current scripting works is a bad idea
07:36 hmmmm you're relying on incompetent 15 year olds to write code that finishes execution within a finite amount of time
07:36 hmmmm or else the entire server gets hung up
07:37 celeron55 that's not a reason why i would take any comments about the current modding system personally
07:37 celeron55 the current system sucks; but also many other systems will suck
07:37 hmmmm i feel as if this is the least bad
07:38 celeron55 let's consider for example this: how are you planning to allow implementing custom player physics in it?
07:39 hmmmm i haven't considered physics at all
07:39 celeron55 stuff like that probably has to go to be actually invoked from the physics thread (which is the main thread currently)
07:40 celeron55 stuff in for example there can probably be completely separate from the other client-side scripts, or if they need to communicate, they could use the event system
07:41 celeron55 or maybe the global settings or something like that; just anything else than direct calls
07:41 rubenwardyy joined #minetest-dev
07:41 hmmmm physics would have to handle the way a CAO moves
07:41 celeron55 same thing for custom player controls; but that might be practically the same thing as physics
07:42 hmmmm the moment that CAO hits something, that generates an event that that client game scripting can handle
07:42 hmmmm I don't know, maybe the physics thread like you were suggesting can do soft real-time calculations
07:42 hmmmm every so many MS
07:43 hmmmm how about 50 ms
07:43 celeron55 what i'm saying is that there will need to be client-side scripts that have to be called immediately, getting a result from them immedidately; and they will probably have to be a separate thing assuming the bulk of the scripts are run (probably appropriately) in a separate threaad
07:43 hmmmm every 50 ms, the physics script executes for some type of CAO
07:43 celeron55 -a
07:44 hmmmm well what the client side scripts are mostly intended for are things like
07:44 hmmmm player punches something
07:44 hmmmm makes a sound event
07:44 hmmmm err, makes a punch event, which the script thread then starts a sound for
07:44 hmmmm or maybe spawns some particles there
07:44 hmmmm maybe adds a dynamic light
07:44 hmmmm etc.
07:45 celeron55 things like what i am describing will be kind of like the next step up from configuration; like configuration settings that are simple functions instead of static values
07:45 celeron55 maybe they don't have access to anything else than a very strict environment where they can't cause much delays or trouble; but custom player controls and physics probably need those
07:46 hmmmm physics is going to be difficult no matter what model we use
07:47 celeron55 well, frankly, not really
07:47 hmmmm the primary issue is that we're relying on an un-preemptable script that's already dog slow to work in a real time manner
07:47 celeron55 not much more difficult than anything else in this
07:48 celeron55 that script isn't doing any of the world generation or pathfinding that generally kills the minetest server timing
07:48 hmmmm it's okay if the lua scripts process subtracting 10 hp from a player 300 ms later
07:48 hmmmm but it's not okay if 300ms pass when the CAO was supposed to hit the wall and bounce back
07:49 celeron55 in any case, this is a thing that has to be implemnted and you can't just sweep it under the rug and pretend it doesn't exist
07:49 hmmmm it would be so much more palatable if this weren't in lua and we could rely on the physics algorithms to be somewhat fast
07:49 celeron55 we can provide a library of fast algorithms to lua
07:49 hmmmm sure, no problem
07:50 hmmmm the problem is the skill of the lua modders
07:50 celeron55 when something becomes popular, it will be optimized
07:50 hmmmm no matter how fast your actual functions that do the real work are, they're going to do something stupid and degrade the entire user experience
07:50 celeron55 those physics things will be rather simple and an easy target to optimization
07:50 celeron55 and an easy target for profiling
07:50 hmmmm well
07:50 hmmmm I noticed that a lot of popular mods are unoptimal right now
07:50 hmmmm because they work well enough
07:51 hmmmm enough of those good-enough mods get piled onto something like vanessae's server and they create something horrifying
07:51 celeron55 you're going to get that kind of problems with the separately threaded ones too; they will block each other from running and cause havoc that way
07:52 celeron55 you can't prevent stupid
07:52 hmmmm in any case
07:52 celeron55 stupid will happen; just make things nice for good modders
07:52 hmmmm we have to deal with the ramnifications of stupid
07:52 celeron55 yes, but don't make that your primary goal
07:52 celeron55 otherwise you'll make Java or something
07:54 hmmmm i feel like custom physics is an entirely different requirement to be honest
07:54 hmmmm it's out of scope for what i'm doing right now
07:54 hmmmm it's nice to keep in mind but if i try to make everything work all at once it's going to fail
07:55 celeron55 you can probably consider it out of scope
07:55 hmmmm any complex system that is in a working state evolved from a once unsophisticated system years ago
07:55 celeron55 just don't try to solve it at all with the separate thread and maybe i'll do it or something
07:55 hmmmm oh, you're interested in minetest again or something? :/
07:56 celeron55 am i not allowed to?
07:56 rubenwardy client side lua makes Minetest more interesting
07:56 hmmmm you don't develop things you don't use
07:56 hmmmm heh
07:56 celeron55 well what if i would use custom physics and player controls
07:57 hmmmm i'd be really happy for you to start developing once again but you can't flake out like everybody else does
07:57 hmmmm there seriously needs to be more than 1 consistently active developer
07:58 hmmmm else it's going to end up like sapier where half the job got done and then he lost interest and dropped off the face of minetest-dev
07:58 hmmmm also zeno was supposed to abstract away irrlicht and he disappeared
07:59 celeron55 well frankly abstracting away irrlicht is very tedious and difficult
07:59 celeron55 can't blame him
08:00 celeron55 probably almost stupidly tedious and difficult, but i'm not going to tell that to someone who wants to try
08:00 Yepoleb_ joined #minetest-dev
08:00 hmmmm everybody seems to follow a pattern where they start off with lots of energy and motivation and end up doing great things for minetest
08:00 hmmmm then they disappear
08:01 hmmmm then they come back and say "hey i'm going to get involved again!" which happens, for two weeks or so
08:02 celeron55 well, MT is kind of overwhelming for any of us
08:05 nrzkt shadowdevs :D
08:06 celeron55 i think one issue is that people are unable to team up with each other to work on a larger feature
08:06 celeron55 and then burn out when trying to get it done in any reasonable amount of time
08:06 celeron55 but that's hard to solve when few people even have enough skills to do anything at all
08:10 celeron55 but clearly it is not worth it on a personal level to contribute something big to minetest
08:13 celeron55 or it might be, but it's so long-term that nobody realizes it
08:13 sloantothebone joined #minetest-dev
08:14 hmmmm i think we might eventually overtake minecraft in popularity
08:14 hmmmm it's an entire genre, i feel, that's about as permanent as FPSes or RTSes
08:14 hmmmm remember when every FPS was a doom clone in 1993?
08:15 celeron55 i wasn't playing games when i was four years old but i do know that
08:15 celeron55 i'm wondering if people generally hated doom and its clones back then
08:15 celeron55 like they are starting to hate minecraft and its clones now
08:16 celeron55 or, well, years ago actually
08:17 hmmmm yeah
08:17 hmmmm but at the same time doom is considered to be like... the reference for FPSes
08:17 hmmmm and it being open sourced helps a lot
08:18 celeron55 well, people get fed up of something and they move on to something else, and then after 10 years they realize the thing they got fed up actually was good, they just had too much of it
08:18 celeron55 this happens everywhere
08:19 celeron55 altough, minecraft is still ridiculously popular; just not very much among older people
08:19 hmmmm when this second wave of popularity comes around, minetest should be well refined at that point
08:20 hmmmm although at its current trajectory it's more of an OS than emacs is
08:20 celeron55 the "being fed up" is affecting all of us too, of course
08:20 rubenwardy <hmmmm> i think we might eventually overtake minecraft in popularity
08:20 rubenwardy It'll take a lot of work
08:21 hmmmm might have to fight this urge to feature creep
08:21 rubenwardy Have you seen this thread? https://forum.minetest.net/v​iewtopic.php?f=6&amp;t=13061
08:22 hmmmm right
08:22 hmmmm less features, more refinement in general
08:22 hmmmm less emphasis on new features
08:23 rubenwardy we do need both
08:24 celeron55 one of the reasons i'm reluctant in touching minetest is that i'm still not sure what it's supposed to be; it diverged from my original plans so much in 2011 that i haven't been able to define it properly ever since
08:24 hmmmm it's an OS
08:24 hmmmm all we need to do is add preemption to lua threads and there you go
08:24 hmmmm preemption to lua to make threads*
08:25 rubenwardy It's a fully fledged OS, lacking only a good voxel game
08:25 rubenwardy content wise
08:26 celeron55 it's a voxel game and game engine that just lacks a good voxel game and a good game engine
08:26 celeron55 just a regular open source project, that is
08:26 hmmmm i think we're doing okay for a small handful of volunteers
08:27 hmmmm mm okay i can hardly form coherent sentences anymore, i need to go to sleep
08:28 celeron55 well, MT has to do better in order to stay relevant
08:28 celeron55 it's a fact
08:30 Gael-de-Sailly joined #minetest-dev
09:01 Darcidride joined #minetest-dev
09:01 johnwayne1986 hey, what about making a kickstarter (or sth) project for making minetest 5.0 with defined goals? ;)
09:15 Calinou joined #minetest-dev
09:18 Calinou https://github.com/minetes​t/minetest_game/issues/525
09:18 Calinou close maybe? I'm sure there's already an engine issue for it
10:06 ElectronLibre joined #minetest-dev
10:44 tuy joined #minetest-dev
11:08 blaze joined #minetest-dev
11:15 rubenwardy joined #minetest-dev
11:23 johnwayne1986 joined #minetest-dev
11:23 Gael-de-Sailly joined #minetest-dev
11:25 Zeitgeist_ joined #minetest-dev
11:39 sgtbigman joined #minetest-dev
11:51 srifqi joined #minetest-dev
12:15 proller joined #minetest-dev
12:43 RealBadAngel joined #minetest-dev
12:44 RealBadAngel msg NickServ identify angelxx1
12:44 RealBadAngel lmao
12:45 nrzkt epic fail :D
12:45 RealBadAngel indeed ;)
12:48 Donillo huh, you better be using longer passwords
12:48 Donillo :P
12:51 sgtbigman or ssl certs :p
13:14 nrzkt joined #minetest-dev
13:25 VanessaE RealBadAngel: your password sucks. :)
13:30 exio4 RealBadAngel: I hope you already changed your password
13:34 RealBadAngel i will change it, i have to ;)
13:34 RealBadAngel anyway i figured out growing of mese crystals
13:35 RealBadAngel first of all farms would work only at certain depth and below, like -1000
13:35 VanessaE I would say underwater, with a mese block as the base and "seed".
13:35 VanessaE and a LOT of heat
13:35 VanessaE such as a lava source under the mese
13:35 RealBadAngel "soil" will be obsidian, turneable to it only by using diamond hoe
13:36 VanessaE but that probably would conflict with the lavacooling code.
13:36 RealBadAngel and lotsa lava around instead of water
13:36 RealBadAngel no
13:36 RealBadAngel abm will be put on obsidian soil
13:37 VanessaE that makes no sense to me, really
13:37 RealBadAngel it will work, believe me :)
13:37 VanessaE why would obsidian generate anything other than silica or somthing?
13:37 VanessaE use a mese block for the seed.
13:37 VanessaE soil*
13:37 RealBadAngel thats about the seed
13:38 RealBadAngel not the ground where it could be planted
13:38 RealBadAngel also a special tool will be needed to cut off a piece of crystal grown
13:39 RealBadAngel pick will dig all the formation at the time
13:39 RealBadAngel same as for formation found grown naturally
13:40 RealBadAngel the tool will reduce growing stage by 1, givin back one crystal
13:40 RealBadAngel but wont remove minimal growing stage
13:42 RealBadAngel and the seed can be also made expensive
13:42 RealBadAngel like 8 mese crystals with obsidian shard in the middle
13:44 RealBadAngel thats gonna bring whole new level to the game. and kinda unique
13:46 Donillo what is mese? what is it's legend on generation? :)
13:47 Donillo pardon my engrish
13:48 VanessaE it's a crystalline artifact of alien origin, somewhat similar to a (very) hard plastic.
13:48 Calinou it's on old Minetest wiki...
13:48 VanessaE electrically conductive, usable as a power source when combined with other materials.
13:48 Calinou it's something related to a Finnish programming community
13:48 Calinou the 8 in 8) and 8D smileys are "Mese glasses"
13:49 VanessaE oh the name?  that's a Finnish shorthand for "MSN Messenger" or something
13:50 Calinou yeah
13:51 RealBadAngel https://imgrush.com/vxfrNtQojkvy.png
13:52 RealBadAngel they do glow, more grown, the brighter
13:52 Donillo aren't first two way too different?
13:52 Calinou we still need glowmap textures :(
13:52 RealBadAngel and comparing to mese_in_stone that looks really nice
13:53 RealBadAngel theyre a bit transparent
13:53 VanessaE Calinou: glowmap?
13:53 RealBadAngel the third map....
13:54 RealBadAngel you see, modern mapping techniques requitre not just one map
13:54 RealBadAngel but about 5 of them
13:54 Donillo I think crystalls alone can coexist with mese in stone
13:54 RealBadAngel but... that would kill the engine for sure ;)
13:54 Calinou VanessaE, a texture overlaid over the main one, that constantly glows, even in the dark
13:55 Calinou could be coloured optionally
13:55 VanessaE oh
13:55 Calinou RealBadAngel, https://github.com/kmkolasinski/AwesomeBump
13:55 RealBadAngel Calinou, thats the tool im using
13:55 Calinou Diffuse, Normal, Specular, Height, Occlusion, Roughness, Metallic
13:55 Calinou :)
13:56 RealBadAngel diffuse is regular texture
13:56 RealBadAngel occlusion i have no slightest idea what for
13:56 RealBadAngel rest of them are actually useable
13:57 RealBadAngel and normal + height are combined into one
13:57 RealBadAngel that makes 5
13:59 Calinou occlusion is rarely used, yeah
13:59 Calinou you also need a specular map most of the time
13:59 Calinou even though Minetest doesn't use any
13:59 Calinou it's used more often than height, in general
14:01 est31 joined #minetest-dev
14:04 RealBadAngel Donillo, i dont think that they can coexist
14:05 RealBadAngel ofc they could be made so, but why?
14:05 RealBadAngel mese is special
14:06 VanessaE it used to be.
14:07 VanessaE then diamonds were added, which sorta diminished that
14:25 RealBadAngel joined #minetest-dev
14:26 RealBadAngel diamonds, mese, metals....
14:26 RealBadAngel theyre not different atm
14:27 RealBadAngel just a different colour
14:27 rubenwardy RealBadAngel, are those diamonds smooth growing?
14:27 RealBadAngel not yer
14:27 rubenwardy are there different nodes? crystal_1, 2, etc
14:27 RealBadAngel yes
14:27 rubenwardy s/diamonds/crstals
14:28 RealBadAngel 4 stages are different nodes and drops
14:29 RealBadAngel stage one gives one crystal
14:29 RealBadAngel most rare stage 4 gives 4
14:29 Brains y
14:29 Brains (wrong window, sorry)
14:30 RealBadAngel crystalline bell will be able to take out one crystal without destroyin the seed
14:35 RealBadAngel you could be able to use it on natural ones, but they wont regrow
14:41 RealBadAngel i do have code almost done, propably will merge it tomorrow
14:41 RealBadAngel its up to mt to get it or not, imho its worth it
14:47 Player_2 joined #minetest-dev
14:47 proller joined #minetest-dev
15:32 Megaf joined #minetest-dev
15:43 nrzkt joined #minetest-dev
15:49 hmmmm joined #minetest-dev
15:50 Gael-de-Sailly joined #minetest-dev
16:02 est31 joined #minetest-dev
16:06 est31 okay I'll push a commit that removes the modstore button
16:07 nrzkt ps ax
16:07 nrzkt oops :p
16:07 rubenwardy missing - as well
16:08 rubenwardy nevermind
16:08 * est31 does ps -Al
16:09 Calinou I'm hardcore, ps a
16:15 rubenwardy ps -aux for me
16:15 nrzkt i don't need users then no u :D
16:16 rubenwardy I don't actually know what each does. I don't need users either
16:17 est31 Wayward_One, have you tested the commit?
16:21 nrzkt Jenkins Android builds will be offline for a moment because i'm migrating my server to a new machine and the Android VM needs some things to be migrated
16:22 est31 ok
16:28 Wayward_One est31: yes, it worked fine
16:37 est31 ok then gonna push the commit then
16:59 est31 okay, mmdb disable commit is pushed
17:05 Krock joined #minetest-dev
17:09 Hunterz joined #minetest-dev
17:11 crazyR joined #minetest-dev
17:21 MinetestForFun joined #minetest-dev
17:23 est31 ok, as we reach the release, we need somebody to write a changelog.
17:23 est31 any volunteers?
17:28 nrzkt Calinou go !
17:30 wischi joined #minetest-dev
17:31 * est31 looks at the issues with the blocker label... all fixed.
17:31 est31 seems we can make the release
17:31 sfan5 does this mean that i can make builds soon?
17:32 est31 we need to wait for paramat to make the game release
17:32 est31 last week he expressed his wish that he wants to do it
17:33 est31 but from the engine side everything is ready
17:34 nrzkt then tag the branch now
17:34 nrzkt and do the version modifications
17:35 est31 lets do it.
17:36 nrzkt est31
17:36 nrzkt before tagging
17:37 nrzkt https://github.com/minetest/minetest​/commit/5622fb2cee6d86f0220ea5773ec7​8f5f2f6714c3#commitcomment-12790406
17:37 nrzkt can you do this ?
17:37 nrzkt the translation choosen is very bad
17:39 est31 hrmmm...
17:39 est31 might conflict weblate, dunno
17:40 est31 lets try it
17:41 est31 can you push a fixing commit?
17:41 kilbith joined #minetest-dev
17:41 nrzkt est31 sorry but i'm at work, cannot :s
17:42 kilbith as i said, "JcJ" is a very unclear acronym
17:42 kilbith don't be so nitpicky
17:43 nrzkt JcJ is know by all RPG gamers
17:43 nrzkt known*
17:43 kilbith our players are kids, bear in mind
17:45 est31 lets push two other very trivial prs
17:46 est31 3077 and 3063
17:47 srifqi joined #minetest-dev
17:49 kilbith rubenwardy: can you write the changelog ?
17:49 Donillo what is JcJ?
17:50 kilbith PvP
17:50 kilbith the most efficient option for the changelog would be a selective grep of #minetest bot messages
17:51 est31 you can generate such logs with git too
17:51 est31 the point of a changelog is to filter the list for important changes
17:52 est31 not all commits should appear in changelog
17:52 kilbith ideally it should fill a .txt whenever the bot announces a commit
17:52 est31 also, the bot has been offline during some times, making commits not getting recorded
17:53 est31 and you'll have to filter jeja's mesecon commits out, and other repos that also log to the bot
17:54 kilbith right, and we have no way to sort by level of importance
17:54 est31 nice, the release commit will be the 4900th one :)
17:57 Robert_Zenz joined #minetest-dev
17:58 rubenwardy I was considering it, kilbith
17:58 rubenwardy I guess I could
17:59 est31 ok, lets push the button, there were no new blocker issues the last days, everbody had enough time to declare what they deem blocker.
17:59 est31 I'll have to do a clean clone build of minetest, to assure its working as intended
17:59 rubenwardy when was the last release? Feb?
17:59 est31 rubenwardy, yes
18:00 rubenwardy okay, I'll start the changelog
18:00 est31 builders: don't build yet, wait for mtgame to be released
18:03 kilbith est31: meanwhile can you blast that "0.5.0-dev" in releases ?
18:03 est31 ?
18:03 kilbith https://github.com/minetest/minetest/releases
18:04 est31 dunno why nrzkt did that release
18:04 nrzkt i didn't do it
18:04 nrzkt github make a release for each tag
18:04 nrzkt this is automated
18:05 est31 perhaps no tag then?
18:05 kilbith the 0.5 tag existed before nrzkt and there was no such release
18:06 kilbith unless i'm mistaking with flag
18:06 twoelk joined #minetest-dev
18:10 hmmmm oh wait
18:10 hmmmm don't forget to update the contributors
18:10 rubenwardy Are nerzhul's commit messages always too long?
18:11 hmmmm a commit message can never be too long, just too many columns in the same line
18:11 est31 hmmmm, good catch
18:11 est31 hmmmm, can you make a commit?
18:12 est31 otherwise I do it
18:12 hmmmm yeah i'll do it
18:13 est31 RealBadAngel, last chance to be not listed in "previous core developers"
18:13 hmmmm also last call for changes
18:22 OldCoder joined #minetest-dev
18:22 hmmmm oh wait a minute
18:22 hmmmm I realize this is probably a moot point because you can just look at the commit logs, but...
18:23 hmmmm do you think it might be a good idea to remove contributors' emails from the credits?
18:23 rubenwardy bored shitless, someone else can continue: https://gist.github.com/rube​nwardy/e9e832cbf6216635b610
18:24 hmmmm I mean, I don't think people who don't look at commit logs are going to email them anything intelligent
18:24 rubenwardy yeah, I say remove it
18:24 hmmmm hehe, my name is so absent from the changelog
18:24 est31 some people only get the binary, and not the source
18:25 est31 but I guess this is the internet age, everybody can find this out
18:25 hmmmm also i thought there were way many more changes than that
18:25 est31 so I'm ok with it
18:25 est31 its incomplete
18:25 hmmmm est31:  yeah, like I said.. people who get only the binary probably don't have business emailing anybody
18:26 hmmmm I mean what are they going to say
18:26 hmmmm "Uhhm I really like minetest but you need to add kangaroos to the game"
18:28 hmmmm when I originally re-did the credits tab, I added email addresses because that's what other credits tabs in FOSS do
18:29 rubenwardy hmmmm, most of your changes up to that point are reverts
18:29 hmmmm hah!
18:30 hmmmm serves them right for making bad commits
18:30 rubenwardy I got up to 16th of March
18:30 hmmmm makes sense then, I took a minetest break around that time
18:30 est31 ha lol
18:31 est31 every time hmmmm does a break, things go awry
18:31 rubenwardy true
18:31 hmmmm I remember the last time I returned everything was broken again
18:31 hmmmm FPS regressions, doesn't even compile, etc.
18:31 hmmmm maybe it's just a coincidence that it seems that way because I'm not aware of the current problems
18:34 rubenwardy surely there's an app which allows you to go through a commit list and click "feature", "bug fix", "discard" etc
18:34 rubenwardy "rephrase"
18:35 rubenwardy s/app/program/
18:36 hmmmm is sokomine a core dev?
18:36 est31 no
18:36 hmmmm hm okay
18:36 hmmmm I lost track
18:36 luizrpgluiz joined #minetest-dev
18:36 est31 https://github.com/orgs/minetest/people
18:37 hmmmm yeah the thing about that is you need to hover over everybody's names
18:37 hmmmm there's no plain list
18:37 hmmmm so i hate doing it
18:37 rubenwardy maybe I'll write an app for it
18:37 hmmmm lol
18:39 Amaz joined #minetest-dev
18:39 luizrpgluiz hey devs,I liked the news from minetest release candidate, will put more to this new version?
18:53 Gael-de-Sailly joined #minetest-dev
18:58 est31 joined #minetest-dev
19:18 est31 hmmmm, before we have asked every dev whether they wanted their email to be mentioned
19:20 est31 so those people might complain now
19:20 est31 e.g. blockmen or paramat
19:22 RealBadAngel my actual email is maciej.kasatkin@o2.pl
19:23 RealBadAngel just for the record ;)
19:23 hmmmm whoops
19:23 hmmmm should we get rid of it
19:23 hmmmm RealBadAngel:  that came from the commit log
19:24 hmmmm est, I'll remove blockmen and paramat
19:24 est31 and nrzkt
19:24 est31 & sapier
19:25 nrzkt you can let my mail
19:25 est31 TeTpaAka and gregorycu too
19:25 hmmmm oh
19:25 hmmmm gregorycu isn't around though
19:25 hmmmm we can't ask him so we'll have to assume that's a no
19:26 est31 well, he didn't actually object I think
19:26 nrzkt offtopic: zfs send|zfs recv is gorgeous for cloning a dataset with a bhyve VM on FreeBSD :D
19:27 est31 hrmm RealBadAngel wants to get #3073 fixed before release
19:27 ShadowBot https://github.com/minetest/minetest/issues/3073 -- tileable_{horizontal,vertical} flags cause fps decrease
19:27 luizrpgluiz left #minetest-dev
19:28 hmmmm alright
19:28 hmmmm if it fixes the problem though
19:28 rubenwardy changelog pretty much done: https://gist.github.com/rube​nwardy/0522aebc5265233ea45d
19:28 hmmmm there's so much mixed up information and misinformation
19:28 rubenwardy Just need to go through and check for mistakes
19:28 est31 we can wait a day or two for him to fix the bug, no?
19:28 hmmmm yeah what the hell.
19:29 hmmmm it is a pretty nasty bug.  i wouldn't want to release with that.
19:29 hmmmm RealBadAngel, do you want your email updated?
19:29 hmmmm er not updated, changed to the real one
19:32 rubenwardy I used this to sort the commits: http://pastebin.com/LRcUgWMX
19:38 rubenwardy okay, edited this a bit: https://gist.github.com/rube​nwardy/0522aebc5265233ea45d
19:38 RealBadAngel hmmmm, yes i would like it to be changed
19:39 rubenwardy How can I find out who jeanpatrick.guerrero@gmail.com is?
19:39 kilbith its me
19:40 rubenwardy thought so
19:40 Robert_Zenz RealBadAngel, how can I be of help?
19:40 RealBadAngel est31, hmmmm about 3073 im talkin to author of the issue right now
19:40 RealBadAngel Robert_Zenz, lets talk in private
19:41 rubenwardy was utf support actually added, or was just some commits done towards it?
19:46 hmmmm it was added, we use libiconv
19:46 rubenwardy for the record, these commits were discarded from the changelog: https://gist.github.com/rube​nwardy/07f6c105b13b35a3f8be
19:46 est31 rubenwardy, chat isn't done with utf-8 yet, but we added conversion routines
19:46 hmmmm the UTF-8 change commit deprecates wide characters in the chat messages
19:46 est31 also they are used
19:47 est31 not yet hmmmm, unfortunately.
19:47 est31 but some bugs got fixed with encoding issues
19:47 est31 also, the "narrow" encoding which has been system dependent, became utf-8
19:48 est31 proller wants his email to be removed as well
19:49 rubenwardy okay, the change log is done on my end, you may want to go through and remove / modify messages to what you think are relevant
19:50 Donillo about encodings, no more that crap when I can't see russian text from Windows players anymore?
19:50 est31 perhaps unify the "Loic Blot"/"Ner'zhul"
19:51 rubenwardy yeah
19:51 Donillo huh, very nice :)
19:52 rubenwardy done
19:52 est31 wow quite a list
19:53 rubenwardy these didn't make the cut, not sure if you agree with me https://gist.github.com/rube​nwardy/07f6c105b13b35a3f8be
19:53 H-H-H joined #minetest-dev
19:54 rubenwardy yeah, it's pretty big
19:54 rubenwardy I wrote a script to help me through it
19:55 rubenwardy I'm tired, time to go to sleep o/
19:55 est31 Perhaps translators should be attributed somehow too
19:55 est31 only with a list of their names
19:56 rubenwardy yeah
19:56 rubenwardy and minetest_game needs to be done
19:56 rubenwardy it's a good release when "bug fixes" is longer than "features"
19:56 hmmmm alright
19:56 hmmmm https://github.com/kwolekr/minetest/commit​/5009d31a333a8e1fff8c97bed69873164c72c600
19:57 hmmmm look good?  anybody else want their details in the credits tab modified?
19:57 est31 lgtm
19:59 RealBadAngel joined #minetest-dev
20:01 est31 okay, lets wait one or two days on the engine side for RealBadAngel to fix the fps bug with the tileable flags.
20:09 nore hmmmm: I'd like an update for me on the website if possible (I already updated the credits however)
20:10 hmmmm i can't change the website, sorry
20:10 nore nvm, it's not that urgent
20:11 nore since there is the website competition anyway, I guess it will be easy to change it after the new website is used
20:19 rubenwardy joined #minetest-dev
20:19 rubenwardy nore, I have access to the website
20:21 rubenwardy is that okay: http://www.minetest.net/contributors
20:21 nore yep, thanks!
20:35 paramat joined #minetest-dev
20:37 paramat hi nore sfan5 last thing for release game#640 shall we merge something like this?
20:37 ShadowBot https://github.com/minetes​t/minetest_game/issues/640 -- Fire: Slow down spread and reduce lua load by paramat
20:40 hmmmm i want to be able to use fire but it consumes the entire serverthread
20:40 paramat don't forget #3079
20:40 ShadowBot https://github.com/minetest/minetest/issues/3079 -- Remove use of engine sent texture tiling flags - theyre no longer needed by RealBadAngel
20:40 hmmmm would really like to time limit abm processing
20:42 sfan5 paramat: the pr should include a comment why that ABM was disabled
20:42 paramat yes ok
20:43 sfan5 seems good otherwise
20:45 paramat the way sounds are done may be making it heavy too, something to work on after release
20:46 nrzkt ~tell est31: Jenkins slave for android is back.
20:46 ShadowBot nrzkt: O.K.
20:56 RealBadAngel paramat, please hold on with 3079
20:57 paramat sure, i was leaving the merge to others anyway
20:57 RealBadAngel theres something wrong with clamping, im workin now with author of the issue https://github.com/minetest/minetest/issues/3073
20:57 RealBadAngel to find out whats wrong
21:11 paramat credits tab is fine for me thanks
21:17 DFeniks joined #minetest-dev
21:21 RealBadAngel paramat, i have found the reason, going to update 3079 soon
21:21 paramat ok
21:22 RealBadAngel i will test it for a while if everything is ok now
21:23 EvergreenTree joined #minetest-dev
21:31 sgtbigman joined #minetest-dev
21:31 ElectronLibre left #minetest-dev
21:47 RealBadAngel ok #3079 is updated
21:48 ShadowBot https://github.com/minetest/minetest/issues/3079 -- Remove use of engine sent texture tiling flags - theyre no longer needed by RealBadAngel
21:48 RealBadAngel issue 3073 can be closed
21:50 paramat cool
22:07 troller joined #minetest-dev
22:37 paramat left #minetest-dev
22:52 damiel joined #minetest-dev
23:09 Zeitgeist_ joined #minetest-dev
23:23 est31 joined #minetest-dev
23:23 est31 RealBadAngel, what did you change with #3079?
23:23 ShadowBot https://github.com/minetest/minetest/issues/3079 -- Remove use of engine sent texture tiling flags - theyre no longer needed by RealBadAngel
23:24 RealBadAngel i removed checking tileable flags for constructing fast faces
23:24 RealBadAngel that was a reason for the slowdown
23:25 est31 hrmmm, seems sb will have to test it...
23:25 RealBadAngel i did
23:25 RealBadAngel it is working ok now, theres no drop
23:25 Fritigern joined #minetest-dev
23:26 RealBadAngel but first to test will be propably Robert_Zenz, the one who reported it
23:26 est31 yea
23:27 RealBadAngel anyway this PR has to go in .13
23:27 RealBadAngel this is more bugfix than cleanup
23:27 est31 if it does the speedup, I'll merge it.
23:28 RealBadAngel it does by two things
23:28 RealBadAngel fixing a bug and removing obsolete code
23:29 RealBadAngel anyway i do have crystalline bell ready, testing it now :)
23:32 Fritigern joined #minetest-dev
23:33 est31 man the new grass is ugly without bumpmapping
23:34 est31 but for the sake of peace, it should be kept this way
23:34 RealBadAngel lol
23:35 RealBadAngel i do really love how grass looks like with bumps and displacement, especially the sides
23:36 RealBadAngel but i do agree its a bit to bright now
23:36 RealBadAngel before it was more juicy green
23:37 RealBadAngel i will propably revert that commit
23:37 est31 its not your fault
23:37 est31 its paramats
23:38 est31 he promised a compromise, what he took was the barf texture
23:38 est31 but lets not talk about it
23:39 RealBadAngel compromises are worth shit
23:40 RealBadAngel you will never fit all the tastes around
23:40 est31 ^
23:40 est31 joined #minetest-dev
23:40 est31 oops wrong button :)
23:41 est31 whats again the default for the tileable flags?
23:41 RealBadAngel better to have some1 with good taste to decide whats good lookin
23:41 RealBadAngel seamless is default
23:41 RealBadAngel only some nodes need that flag, like dirt with grass or snow
23:42 RealBadAngel sorry, tiles, not the nodes
23:42 est31 tileable_horizontal = true or false which is default?
23:42 RealBadAngel both true == texture seamless
23:42 RealBadAngel both true are default
23:42 est31 ok
23:43 RealBadAngel you can define true in tiledef but that makes no sence while theyre default
23:43 RealBadAngel so you should do that only with false case
23:43 est31 interesting, I cant reproduce what robert zenz can reproduce
23:43 est31 you know the displacement bug
23:44 est31 what do i have to turn on?
23:44 RealBadAngel parallax mapping
23:44 est31 in the gui whats the option there?
23:44 RealBadAngel parallax mapping ;)
23:44 est31 there is only bumpmapping and parallax occlusion
23:44 RealBadAngel ouch
23:44 RealBadAngel occlussion
23:45 RealBadAngel but thats the same
23:45 est31 ah now I see the bug
23:45 est31 ok
23:45 RealBadAngel its very nature is to move the texture
23:45 est31 around 30 fps
23:45 est31 err
23:45 RealBadAngel if its not seamless, then houston we have a problem ;)
23:45 est31 around 20
23:45 RealBadAngel apply now the patch
23:46 RealBadAngel with lotsa grass around your fps should double basically
23:46 est31 I see
23:46 est31 when the tileable flags are true, I get ~14 fps
23:46 est31 while I get ~30 fps with the tileable flags being false
23:46 RealBadAngel thats not the problem with flags themselves
23:46 est31 err
23:47 est31 other way round
23:47 est31 RealBadAngel, I know
23:47 est31 I only want to see the difference
23:47 RealBadAngel problem was i disabled fastfaces for grass with them
23:47 est31 because thats the bug
23:47 RealBadAngel and that was the mistake
23:47 est31 now lets try your patch
23:49 RealBadAngel btw, parallax occlusion is now rather incorrect
23:50 RealBadAngel we are using relief mapping
23:50 est31 okay your patch works
23:50 RealBadAngel we do have still code for parallax occlusion mapping but i think its obsolete. too low quality of the method
23:51 est31 I agree with you now, we can merge it IMO.
23:52 est31 perhaps also remove the part where the texture flags are sent to the shader?
23:52 RealBadAngel nononono
23:52 RealBadAngel no way :)
23:52 RealBadAngel i need that code
23:52 est31 for what
23:52 RealBadAngel to send to shaders anything
23:52 est31 well thats ok
23:52 RealBadAngel irrlicht lacks of setting vertex attributes
23:53 RealBadAngel ive spent months figuring out working way to do so
23:53 est31 just remove the part where you send the g and b values for the tileability flags
23:53 RealBadAngel this code is an absolute MUST HAVE
23:54 est31 its a very neat trick, but we should send only what we read at the other side
23:54 est31 I dont mean that you should remove all attribute sending code
23:54 est31 only the part that sends the tileability h and v flags
23:54 RealBadAngel so lets leave only R part
23:54 est31 yes
23:54 RealBadAngel i will save a copy of the code
23:55 est31 its in git, nothing is lost :)
23:55 RealBadAngel i prefer my own ways
23:55 sloantothebone joined #minetest-dev
23:56 est31 no problem with that
23:56 RealBadAngel shall i update the pr then>
23:56 RealBadAngel ?
23:56 est31 yes
23:57 wischi2 joined #minetest-dev
23:57 RealBadAngel gimme a few minutes then

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