Minetest logo

IRC log for #minetest-dev, 2013-02-10

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

All times shown according to UTC.

Time Nick Message
01:35 VanessaE joined #minetest-dev
01:36 VanessaE joined #minetest-dev
02:05 curiousNoob joined #minetest-dev
02:05 curiousNoob What is the PEER ID used for in communication, and when s it agreed upon/determined for a given connection?
02:09 ecube joined #minetest-dev
02:11 VanessaE joined #minetest-dev
02:12 curiousNoob Hi!
02:15 ecube joined #minetest-dev
02:22 hmmmm curiousNoob, the peer id is sent in the basic packet header after the server assigns it with the peer_id_new field in the CONTROLTYPE_SET_PEER_ID control packet
02:22 hmmmm it is the first application-level packet that the server sends
02:22 rarkenin_ What file is that in, in the source? I seem to have overlooked that.
02:23 hmmmm connection.cpp, line 832.
02:23 rarkenin_ Perfect, thanks!
02:24 rarkenin_ And that's sent once, right after the client sends something?
02:25 hmmmm yes, it is sent in response to the client's packet of type TYPE_ORIGINAL
02:26 rarkenin_ Sorry, but where is that thing about TYPE_ORIGINAL?
02:27 hmmmm connection.h, line 232
02:27 hmmmm that first packet has no actual payload
02:28 hmmmm you can read about this in doc/protocol.txt, by the way.
02:28 rarkenin_ Oh, didn;t see that file. Thanks!
02:29 rarkenin_ So sequence numbers are incremented with each packet?
02:29 hmmmm yes
02:30 rarkenin_ Why does it start so high? It's 16-bit unsigned, so it'll wrap around almost immediately.
02:31 hmmmm dunno
02:31 hmmmm it doesn't really matter.. the client and server just need to agree on what the initial number is
02:36 rarkenin_ How do they agree on an initial? Just the one sent in the firstpacket?
02:36 hmmmm they agree on it because it's defined as such in the code
02:36 rarkenin_ Ah, OK.
02:38 rarkenin_ So right after seqnum, the payload begins?
02:38 rarkenin_ And the payload starts with TOCLIENT_FOO or TOSERVER_BAR?
02:39 hmmmm the sequence number is sent in reliable packets, not sent with regular packets
02:39 rarkenin_ So the actual game-chat-and-entity stuff is sent in non-reliable packets?
02:40 rarkenin_ In what places are reliable used?
02:41 hmmmm seemingly everywhere
02:42 rarkenin_ So then I just stick my payload, TOSERVER_FOO/TOCLIENT_FOO and parameters right after the seqnum in that case?
03:16 ecube joined #minetest-dev
03:22 ecube joined #minetest-dev
03:29 hmmmm yes, well, your payload starts with TOSERVER_FOO/TOCLIENT_FOO
03:42 ecube joined #minetest-dev
05:54 ecube joined #minetest-dev
06:24 ecube_ joined #minetest-dev
06:55 proller joined #minetest-dev
07:35 proller joined #minetest-dev
07:43 proller hi all. here branch with like- waterplus mod finite liquids
07:43 proller https://github.com/proller/minetest/commits/liquid
07:43 ffoxin joined #minetest-dev
07:43 proller question: why LIQUID_LEVEL_MASK 0x07 ?
07:44 proller i want to change to 0x3f
09:13 ffoxin joined #minetest-dev
09:25 Taoki Hi. Is there a specific part in the code where the server reads settings from minetest.conf and sends their values to clients? Any examples of that?
09:32 iqualfragile joined #minetest-dev
11:16 PilzAdam joined #minetest-dev
11:21 Jordach joined #minetest-dev
11:21 Jordach joined #minetest-dev
11:45 thexyz Taoki: server.cpp:4299
11:46 Taoki thanks
11:47 PilzAdam thexyz, do you have anything against this: https://github.com/celeron55/minetest/pull/462 ?
11:48 thexyz it's good as long as it works
11:48 PilzAdam I just tested it
11:48 thexyz then merge it
11:58 Taoki Erm... you do see the each of the 2 planes from any direction right?
11:58 Taoki Just not the opposite one
11:59 Taoki Since if not, it means that looking at a plant from its front-right would show it, but looking from its back-left would cause it to not be seen
12:12 PilzAdam is there anything wrong in this: https://github.com/celeron55/minetest/pull/429
12:50 SpeedProg joined #minetest-dev
13:05 rarkenin_ left #minetest-dev
13:20 PilzAdam is there any good reason why this isnt merged yet? https://github.com/celeron55/minetest/pull/275
13:23 rubenwardy joined #minetest-dev
13:37 Taoki PilzAdam: Still not sure about that plantlike backface-culling change. What does it do exactly?
13:37 Taoki Plants have 2 planes in total in an X shape IIRC. If backface culling would be enabled, you'll only be able to see them from one direction, which would completely break them
13:37 PilzAdam previously the images were backface culled and 4 images were drawn
13:38 Taoki If it's just not rendering the back face of each plane however, when looking at it from one direction, then it's good
13:38 Taoki Yeah. But when drawing 2, does it make sure for each to show the one you can see from either side?
13:38 PilzAdam backface culling is disabled now
13:39 Taoki That means that at any given moment and from any angle, exactly 2 faces should be drawn for each plant
13:39 Taoki And visible
13:39 Taoki If it works like that then I agree it's a good change and should be merged without worrying
13:50 rubenwardy joined #minetest-dev
14:10 hmmmm joined #minetest-dev
14:23 rubenwardy joined #minetest-dev
14:25 rsiska joined #minetest-dev
14:41 Taoki My changes to physics are nearly done, just need to fix one more bug. To those interested, I recomment desting them on my GIT. https://github.com/MirceaKitsune/minetest/tree/movement
14:41 Taoki PilzAdam: ^ If you have nothing else to do in MineTest or otherwise at the moment
14:43 Taoki What my branch does: 1 - Turning nearly all player movement values into settings, configurable in minetest.conf (we can even have low gravity now). 2 - Tweaking a lot of physical properties to work correctly and smoothing... such as proper diving in water, correct acceleration in all circumstances and directions, low air acceleration by default so you can't move around in the air, etc.
14:45 Taoki Worst was air acceleration which allowed the player to walk freely in mid-air, now it's 1/4 of ground acceleration. Second worst was diving into water... when hitting water at high speed the player would have their speed instantly cut and slowly begin to sink in, now they will go in hard and slowly descent to water friction speed. Going up / down when swimming of flying also uses acceleration now.
14:45 PilzAdam building....
14:46 Taoki ok. The bug that's not fixed yet is the Y axis acceleration. The new function is used, but the formula is somehow wrong and it's still too fast, rather than the same as X and Z acceleration. Once this is fixed however, the code is ready as far as I'm concerned.
14:47 Jordach Taoki, its just getting the physics to a "T", isnt it?
14:47 Taoki Jordach: To a T?
14:47 Jordach its a term
14:47 PilzAdam Taoki, do you have decreased the Y fast movement when flying?
14:47 Taoki Not sure what that means :P But this should get physics working realistically and correctly
14:48 Taoki PilzAdam: No. When flying without fast movement, you have the normal ground speed and acceleration. When flying with fast movement, you have speed and acceleration for fast movement
14:49 Taoki Apart from the changes I listed, most should work like before
14:52 ffoxin joined #minetest-dev
14:52 PilzAdam its definetly slower than before
14:52 Taoki Ground speed shouldn't be. But that's all about the default values, I can raise them
14:52 PilzAdam flying at fast speed up and downwards is slower
14:53 Taoki BTW: Diving in water might not always be noticeable. To really see how it works, find a deep enough lake ( 10 or more blocks) and fly very high above it, then stop flying and drop into it very fast. You now sink to the bottom
14:53 Taoki PilzAdam: How much slower does it feel like as a percentage? I can turn up the defaults in that case
14:53 PilzAdam dunno, 70%?
14:54 Taoki Ok. Fast move feels pretty fast to me
14:54 Taoki Ground speed seems the same
14:54 Taoki Water movement is too slow however, will fix that
14:55 PilzAdam water movement is too fast
14:55 PilzAdam its like normal walking
14:55 Taoki :P
14:56 PilzAdam does it depend on the viscosity?
14:56 Taoki PilzAdam: Only when you have fast_move enabled
14:56 Taoki no. Turn off fast_move mode and try walking in water then
14:56 Taoki That's because when fast_move is on, you have more acceleration. This probably changes water friction
14:56 PilzAdam I have always_fly_fast = false
14:56 Taoki ok. Don't think that affects it
14:57 PilzAdam fast move should only effect flying and walking when holding E
14:57 Taoki Still, try walking in water without fast move
14:57 Taoki Yeah I can try to fix that
14:58 PilzAdam it should stay as before, fast_move shouldnt have any effect on moving in water
14:58 Taoki Currently fast_move also gives a higher acceleration all the time when enabled
14:58 PilzAdam (except when flying)
14:58 Taoki ok, will fix
14:59 PilzAdam so, now playing with gravity :-)
14:59 Taoki Try a value 4 times lower, you'll find it pretty fun :)
14:59 PilzAdam does the server send all the settings to the client?
14:59 Taoki Just commited code that does that, it does now
15:02 PilzAdam heh, with gravity = 100 you cant jump onto a slab, but walk on it
15:03 Taoki PilzAdam: Fixed the fast movement acceleration, update the branch again and try water now
15:05 Taoki Also, test flying and ground speed again, both with and without fast movement, and let me know if you think any of them should be faster or slower by default. Personally I find the current defaults good, but IIRC few should have changed from what they used to be
15:05 PilzAdam also rename water to liquid in the settings names
15:05 Taoki sure
15:06 PilzAdam now holding E while swiming gives you the fast speed but swiming up/down is still slow
15:06 Taoki Water speed is very low now, need to turn it up a little
15:07 Taoki right. Which would the correct behavior be?
15:07 Taoki Swimming down doesn't really exist, since you sink down regardless of what you do if you don't swim up
15:07 PilzAdam either now effect of fast move on swiming or increase the vertical speed when moving fast
15:07 PilzAdam s/now/no
15:08 Taoki I think fast underwater is good, so I'll try to make it change the sinking and swimming only
15:09 Taoki For swimming down it's hard however, since the physics do that and it's not a player control
15:09 PilzAdam maybe use shift to swim down?
15:09 PilzAdam (with and without fast)
15:09 PilzAdam since in RL you can swim down too
15:10 Taoki Yeah... and when not holding shift have the player stall in water
15:10 Taoki Anyway, need to go for now. Will fix these when I'm back
15:10 Taoki Still, I would appreciate it if you could look into something, if that's ok
15:11 PilzAdam another bug: when moving fast and realsing W and E you slide like 10m over the ground
15:12 Taoki With my code, look at player.cpp line 100 (void Player::accelerateY). That's the new function aimed to make Y acceleration work the same as X and Z one. But for some reason it's not working the same when given the same value as Player::accelerateY. I can't figure out why... if you have some time and like those changes and wish to help, maybe you can at least find a hint on what I should change
15:12 Taoki That one's normal, since now you have the normal acceleration when not holding down E any more
15:12 Taoki Acceleration also means deceleration
15:12 Taoki I hope that one can slip by since it doesn't do any harm, and I also think it's a bit more fun that way :)
15:13 Taoki BRB for now
15:14 Calinou joined #minetest-dev
16:05 rarkenin_ joined #minetest-dev
16:05 rarkenin_ Hi! Quick question about seqnum. It is incremented with each packet in either direction, or is a separate seqnum used for each direction?
16:40 iqualfragile joined #minetest-dev
17:06 Jordach_ joined #minetest-dev
17:06 Jordach_ joined #minetest-dev
17:37 VanessaE joined #minetest-dev
18:22 doserj joined #minetest-dev
18:37 celeron55 rarkenin_: separate for each peer
18:37 celeron55 umm... actually, separate for each side of a connection
18:54 blah_craft joined #minetest-dev
18:54 rarkenin_ celeron55: Sorry, had to go AFK. But, for the server and one client, a packet , for instance , that requests media might be 65521, and the next packet from the server would be 65522?
18:56 Taoki PilzAdam: Finished the water physics changes you suggested. Feel free to pull the branch again and test, let me know if that's good now
18:56 Taoki Note that up / down swimming seems a very little bit faster than horizontal swimming, due to the Y acceleration code which I haven't figured out how to fully fix yet
18:57 Taoki But other than that it should work as discussed. You now swim down with the Sneak key, up with the Jump key by default, and you only sink very slowly in time.
18:58 blah_craft Hi all - I'm getting a "generate_shader(): failed to generate" error when I enabled shaders. I'm pretty sure it's an OpenGL / graphics driver issue, but wondered if u knew any way to debug it?
19:00 Taoki celeron55: Feel free to test my branch as well if you aren't working on something else right now, and let me know if there's something you believe needs to be changed absolutely. https://github.com/MirceaKitsune/minetest/tree/movement
19:01 Taoki [16:41:44] <Taoki> What my branch does: 1 - Turning nearly all player movement values into settings, configurable in minetest.conf (we can even have low gravity now). 2 - Tweaking a lot of physical properties to work correctly and smoothing... such as proper diving in water, correct acceleration in all circumstances and directions, low air acceleration by default so you can't move around in the air, etc.
19:01 Taoki [16:43:46] <Taoki> Worst was air acceleration which allowed the player to walk freely in mid-air, now it's 1/4 of ground acceleration. Second worst was diving into water... when hitting water at high speed the player would have their speed instantly cut and slowly begin to sink in, now they will go in hard and slowly descent to water friction speed. Going up / down when swimming of flying also uses acceleration now.
19:02 F00 Diminished air control may make conventional (whole-block-only) stairs particularly tedious
19:03 Taoki Don't think so, but I can test to see how it works
19:03 Taoki BRB for now though. My water fixes are in the latest branch tho
19:08 blah_craft any guesses on that issue loading shaders?
19:08 celeron55 rarkenin_: no, the ids from different ends can be anything
19:10 celeron55 Taoki: not being able to control horizontally when falling is a good thing in games
19:10 celeron55 Taoki: it's one of the things where reality *really* sucks
19:10 celeron55 eh
19:10 celeron55 s/not //
19:14 rarkenin_ celeron55: So it just matters that each direction has continuous peer IDs, but each direction is indepedent of each other. Right?
19:30 doserj joined #minetest-dev
19:47 celeron55 rarkenin_: yes
19:48 rarkenin_ OK, thanls!
19:48 celeron55 ehm
19:48 rarkenin_ *thanks*
19:48 celeron55 peer id is a different thing
19:48 celeron55 you are talking about packet ids
19:48 rarkenin_ Yeah, sorry, I meant sequnums. I was just thinking one thing, typing another.
19:49 celeron55 what are you doing, if i may ask?
19:50 celeron55 blah_craft: you can usually see a more precise error in stderr/stdout as printed by irrlicht
19:50 rarkenin_ Actually, making a little utility to query extended stuff from the server, for automated detailing, like mapping, and activity levels.
19:50 celeron55 so you are implementing the network stack yourself? on what language?
19:51 rarkenin_ PHP, actually
19:53 rarkenin_ If I have time, I may as well extend it to be a bot that runs on PHP in CLI mode(as opposed to webserver SAPIs)
19:53 thexyz wow
19:53 thexyz but why php?
19:53 rarkenin_ The original plan was for this querynig to be done by webservers hosting server lists.
19:54 rarkenin_ Also, I like PHP, and I don't like wasting my time typing variables.
20:03 Taoki celeron55: Air speed can still be controlled with my patch, just very little. The old way felt very wrong and basic to me... but since the physics are settings now old air acceleration can be re-enabled too
20:03 Taoki Problem is how to fix the new Y acceleration, which still doesn't work right
20:03 rarkenin_ afk for 15 minutes.
20:03 Taoki Gonna keep looking into that for now, since it should be the last bug left
20:04 Taoki Really love how the new physics are working and feeling now though :)
20:32 doserj PilzAdam: I included your suggestion for pull-request https://github.com/celeron55/minetest/pull/439
20:36 rarkenin_ What file(s) explain(s) the NodeDef, MapBlock, and MapNode serialization formats?
20:37 rarkenin_ All I can find are the respective source files, which don't help me much.
20:37 PilzAdam doserj, nice
20:45 rarkenin_ Does Minetest assume a machine endianness?
20:52 hmmmm no
20:53 Taoki What's the exact difference between float and f32?
20:54 hmmmm actually come to think of it, it does... the endianness of the client and server must match
20:55 doserj joined #minetest-dev
20:58 hmmmm taoki, f32 means a floating point number that's 32 bits wide.. float on the other hand may be anything
20:58 Taoki ok. So float can store more decimals and be more accurate overall
20:59 hmmmm there are no restrictions on the size of floating point numbers except float must represent a subset of the values of double, which must represent a subset of the values of long double
21:33 rarkenin_ OK, another question. I'm looking at a wireshark dump of a TOCLIENT_CHAT_MESSAGE.
21:34 rarkenin_ I see my magic bytes, PeerId, channel, type, sequence number(since its reliable), then a really odd 01 00(hex) before the payload starts. What is that?
21:36 rarkenin_ Actually, it's just the 01 after the seqnum that puzzles me. I know where the 00 came from.
21:55 Taoki New physics are ready for testing and merging: https://github.com/celeron55/minetest/pull/483
21:56 Taoki VanessaE: Feel like switching your server to my branch and maybe someone else can join to test stuff?
21:58 VanessaE dare I ask what I'll be getting into ? :-)
21:58 VanessaE oh the water physics et.al?
21:59 Taoki Yes. See the points in the note on that pull request
21:59 VanessaE checking...
22:00 doserj joined #minetest-dev
22:01 VanessaE sure, what the hell :D
22:01 Taoki ok :) Just merged master into my branch, update again if you just got it
22:01 hmmmm alright
22:01 VanessaE I'm just about to clone it now.
22:02 hmmmm vanessa, let's +q $~a
22:02 hmmmm i'm getting sick of this
22:02 VanessaE hmmmm: that'll catch way too many false positives
22:02 hmmmm there's nothing else we can do
22:02 iqualfragile joined #minetest-dev
22:04 VanessaE hmmmm: there is - keep banning whole class C's, or even larger ranges, and forbid anyone from just nonchalantly clering the ban list.
22:04 VanessaE sooner or later there won't be any proxy addresses left to use.
22:04 hmmmm have you seen how many different class As he's had?
22:04 VanessaE and maybe yeah, ban webchat entirely.
22:04 VanessaE three by my count
22:05 hmmmm so far
22:05 hmmmm he stopped using the 189 which i guess was his ip without VPN
22:05 VanessaE actually he hasn't stopped using that one
22:05 VanessaE I've seen him just recently on that host
22:06 hmmmm not with the most recent spamming
22:06 VanessaE that wasn't LM, that was markv.  different person or so I thought.
22:07 hmmmm i guess not
22:08 * VanessaE sighs
22:09 hmmmm we'll give it a day and judge how much we actually need to do something about it
22:13 VanessaE ok
22:14 doserj rarkenin_: http://forum.minetest.net/viewtopic.php?pid=69521#p69521
22:28 Taoki Testing low gravity on VanessaE's server. If anyone is interested, compile my branch ( https://github.com/celeron55/minetest/pull/483 ) and connect :)
22:47 doserj joined #minetest-dev
22:50 Taoki VanessaE: Might be doing a patch to my code in 5 minutes, don't shut down yet
22:50 VanessaE bah, I just did haha
22:50 VanessaE ok, going back up with it
22:51 VanessaE (gravity set to 0.2 as before)
22:51 Taoki NP :)
22:52 proller joined #minetest-dev
22:57 VanessaE should I be worried about your next change? ;)
22:58 Taoki VanessaE: Update my branch again. I fixed a bug with being unable to sink into water when near the surface
22:58 Taoki Not at all :)
22:59 VanessaE ok, sec
22:59 Taoki Also, please set gravity to 1.5 or something around that, it seems to be good and much more usable (I had other values in mind with 0.2 :P )
22:59 Taoki Something between 1 and 2 is good
22:59 VanessaE I'll set it to 1.6, moon-like. :-)
23:00 Taoki yes :)
23:02 VanessaE waiting for the install to finish.
23:03 VanessaE ok, it's up
23:05 VanessaE I still can't swim down....
23:05 Taoki VanessaE: Might be another issue with the swim key, no idea in that case
23:06 Taoki VanessaE: Ahhh. Aux 1 descends means the E button is used to go down?
23:06 VanessaE yes
23:06 VanessaE and it is used to sprint when on land also
23:06 VanessaE shift is sneak in this case
23:07 Taoki Yes, it is my fault then sorry :P Hope everyone can bare another restart, gonna fix that too
23:07 VanessaE lol
23:07 Taoki VanessaE: When that option is enabled, is fast movement possible any longer?
23:07 VanessaE sure
23:08 Taoki Cuz in this case, it will mean fast movement also descends fast automatically
23:08 Taoki Don't restart yet tho, need to test locally first
23:08 VanessaE on land, E runs fast when fast is enabled, or descends ladders.  When flying, j turns fast on/off, and E descends
23:09 VanessaE maybe fast while swimming shouldn't be possible?
23:09 VanessaE just E or shift descends, and that's it?
23:09 Taoki Fast swimming is useful. I'm not sure how this is to be done sadly...
23:09 VanessaE hm
23:10 Taoki Since with this, E means going up fast automatically when fast movement is on
23:10 VanessaE then let j toggle fast on/off like while flying
23:10 VanessaE then E (or shift) descends.
23:10 VanessaE at least it's consistent then
23:11 VanessaE think of swimming as flying in water :-)
23:11 VanessaE then it makes sense
23:11 Taoki When fast movement is off this works fine. But when it's on things conflict
23:11 Taoki Since E either activates fast, either swims down
23:11 VanessaE when fast movement is on, make it always on.
23:11 VanessaE let E swim down only.
23:11 VanessaE in other words, make the controls behave just like flying through the air
23:12 Taoki need to think
23:13 Taoki I think I should do that, but only when aux1 descending is enabled
23:13 VanessaE yeah
23:28 doserj joined #minetest-dev
23:28 Taoki VanessaE: Coded and tested, you can update my branch. aux1_descends will work properly underwater now
23:30 VanessaE sure thing
23:30 VanessaE updating now
23:33 VanessaE ok, it's in
23:33 Taoki connecting
23:33 VanessaE perfect.
23:34 RealBadAngel was connected just in time
23:34 RealBadAngel to see server shutting down lol
23:35 Taoki It's back now :)
23:38 VanessaE Taoki: every so often when I'm swimming on top of the water, it'll jump out of the water several blocks up (in low gravity anyway)
23:38 Taoki VanessaE: Only in low gravity and because you're swimming up in Fast mode
23:39 Taoki Which throws the player upward more than normally. But on normal gravity it's not as visible
23:39 VanessaE right
23:39 VanessaE ok so that's fine then
23:39 VanessaE now here's a real challenge:  make the water visibly splash when the player enters/exits it :D
23:39 F00 Being able to toss yourself out of the water a bit ought to be desirable, I think.
23:39 VanessaE (think client-side particles as with digging)
23:41 F00 The most hilarious thing is seeing somebody die because they can't get themself out of lava due to a panel or something blocking them
23:42 VanessaE Taoki: possible oversight:  Should one be able to stop themselves short by switching from W+E to S+E?
23:42 Taoki VanessaE: Normally yes, but there's still some inertia
23:43 VanessaE ok good
23:43 Taoki That should work as expected now though
23:43 Taoki yes, works fine
23:43 Taoki When you have fast movement on and hold down E, acceleration is automatically higher too
23:45 VanessaE it seems to work as one would expect now.
23:45 VanessaE the inertia in fast mode takes some getting used to, but it's good
23:48 Taoki RealBadAngel: I see you online but not moving or saying anytihng
23:49 RealBadAngel sry, was busy, back now
23:49 Taoki ok., WB

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