Time  Nick         Message
15:20 RealBadAngel Hi, a quick question: when changing protocol version, protocol ID shall be changed too?
15:21 celeron55_   never
15:21 RealBadAngel thx
15:22 celeron55_   the protocol id is there only to distinguish minetest packets from random data or some other protocol
15:23 RealBadAngel k
15:23 celeron55_   also i don't think you'll need to increase the protocol version for the addition of the flags to the end of the single packet; it'll be increased once something actually incompatible is added upstream (like taoki's stuff)
15:24 PilzAdam     RealBadAngel, if you change the protocol version you can move this back to correct place: https://github.com/celeron55/minetest/commit/0709918bd82ab942445e12aca5bceb85c6ce05e0
15:24 RealBadAngel hmmm, increasing size of the packet wont cause problems?
15:24 celeron55_   PilzAdam: it will not be changed now
15:24 RealBadAngel if the client will send more data?
15:25 celeron55_   hmm... actually, you'd need to add a check for whether the packet is long enough in case an old client connects to a new server
15:25 celeron55_   the other way there is zero problems
15:27 RealBadAngel without new version checks should be on both sides, imho this would lead to just makin extra code
15:28 RealBadAngel and after other update like Taoki's one the checks will become obsolete
15:32 PilzAdam     it would be nice if this would be developed further: http://c55.me/blog/?p=275
15:37 RealBadAngel celeron55_, checks for the server alone could be done, but if the server will run the mods that will use the new feature that can lead to crashes or to incorrect behaviour at least
15:38 RealBadAngel PilzAdam, ive seen some1 made mod called "betterchest" out of your carts box
15:38 RealBadAngel wont it dissapear when used clearobjects?
15:39 PilzAdam     yep
15:39 RealBadAngel servers alone can force clearing objects so such mod is unusable then, right?
15:40 PilzAdam     i guess so
15:40 PilzAdam     btw: https://github.com/Uberi/minetest/commit/0f3bd365e1d7e7c6dfe99d862923e78acd5aa70e <- Ive tested it and it works
15:41 RealBadAngel so imho shall be added asap
15:42 RealBadAngel using that greifers can crash servers
15:42 Taoki        Just a note: My models branch already upates protocol version to 14. If that will come soon after RBA's patch, there might be no need to increment it twice
15:42 RealBadAngel how soon is soon? :)
15:42 RealBadAngel may happen some folks will download it just meanwhile and start to scream somethin is wrong
15:43 RealBadAngel numbers are infinite anyways
15:43 Taoki        RealBadAngel: After your key reading code is in, all that's needed is the script to handle animations and it will work for the most part
15:43 Taoki        The player script will need to be improved but that's the last remaining concern
15:44 RealBadAngel celeron55_ ?
15:44 RealBadAngel up to you
15:44 Taoki        All else works fine. Apart from an Irrlicht bug that causes bone positioning to break animations and the mesh so far... but Irrlicht will fix that in the future
15:44 celeron55_   i already mentioned taoki's stuff up thre
15:44 celeron55_   +e
15:44 Taoki        Also: The CODE part of my branch should be ok to merge as well. Not the script and player yet, but the code yes
15:45 RealBadAngel Taoki,  i think some1 posted here yesterday somethin bout this particular bug
15:45 Taoki        oh?
15:45 Taoki        Would be nice if someone could fix it, though it is not urgent. Just doesn't allow animation blending to work yet, and stuff like the head looking up / down
15:45 RealBadAngel Taoki i pasted you start of todays talk
15:45 Taoki        yeah
15:45 Taoki        will read again maybe i missed it
15:46 Taoki        Yeah, that's about the protocol version. Read it
15:47 Taoki        In my opinion you can leave protocol version as is, since after that my code should be ok to merge soon. Models code can probably go in today also
15:47 Taoki        There might still be small bugs, but normally minor things.
15:48 celeron55_   it is, of course, possible also to support multiple protocol versions in a single server version, but it generally is not worth the trouble
15:51 Taoki        Anyway: Once RBA's changes are in master, I'll update latest master, merge it in my branch, fix git conflicts if needed... and if no one knows any major bugs with the current models code, I can pull request it. Maybe it can even get in tonight... if small bugs exist they can be fixed on the way (surely no one will do major tests on the code until it's in, so it goes to large servers)
15:51 Taoki        Then I'll likely be finishing the scripts (will probably need help past some point) so the 3D player can be ready too
15:51 RealBadAngel i need half an hour or so to make last changes to the patch and test it
15:51 Taoki        sure
15:52 celeron55_   PilzAdam: re. http://c55.me/blog/?p=275: i might, but currently there is just too many things to worry about (and my other projects) in order to sink the required time to it
15:53 RealBadAngel but the fix that PilzAdam mentioned is rather urgent
15:53 PilzAdam     already merged
15:53 RealBadAngel afaik Landmine is using this to crash the servers
15:55 celeron55_   landmine is quite useful; he's harmless and finds out problems before any actual bad guys 8)
15:58 PilzAdam     https://github.com/celeron55/minetest_game/pull/40 <- this is a matter of taste so i suggest we vote or something like this
16:00 celeron55_   care to make screenshots of with and without?
16:01 Taoki        lol landmine. I seen him around here a few times. Some projects have "common trolls" which manage to stick around, and not do quite enough to get permanently bad
16:01 Taoki        Xonotic has one too, but he's way worse :P
16:05 PilzAdam     celeron55_, old: http://www.zimg.eu/i/1124599787 new: http://www.zimg.eu/i/2296078330
16:05 PilzAdam     its after ~1/2 of the animation
16:06 Taoki        BTW. There's something else I wanted to ask about. Not sure if I can do it, but IMO that would be really nice and needed
16:06 Taoki        With the current lighting system, would it be posisble to allow for colored lighting? Lighting can't be given a color currently... that would be a relally nice feature
16:07 celeron55_   not as-is, but somebody did once insist on implementing it, and did so
16:07 celeron55_   it requires three times the computing power
16:07 celeron55_   and roughly twice the space
16:07 Taoki        Nice. Is his code still around?
16:07 Taoki        Hmmm
16:08 Taoki        If it can be a client choice that should be fine
16:08 celeron55_   the server has to do it too
16:08 Taoki        So if it's disabled the player sees all lights white
16:08 Taoki        ok. That could have it as a feature too, and only enable it if the server has enough power
16:08 celeron55_   sounds complicated-ish, but anyway, umm... i wonder who did it
16:09 celeron55_   (oh and also, it looks kind of dull with the vertex color based lighting)
16:09 Taoki        My aim is still to get real irrlicht lighting working, but it's hard to imagine a way around sun shining in caves. So this one I likely won't be able to do alone
16:09 Taoki        What I will once the models is over, is to try add support for some nice parallax mapping :)
16:10 celeron55_   it was here, as a branch, but the guy has removed it https://github.com/Queatz/minetest/
16:10 Taoki        ahhh, ok
16:11 celeron55_   and i don't seem to have it pulled to my local repo
16:12 celeron55_   this video exists http://www.youtube.com/watch?v=1MWZxyCsbv4
16:12 celeron55_   somebody probably has it if you ask around
16:12 Taoki        Damn, that looks so beautiful :(
16:12 Taoki        Yeah, totally should
16:14 celeron55_   it's fairly trivial to implement though
16:14 Taoki        true... not the most important feature. But it will contribute to making the world even more beautiful and extended
16:14 celeron55_   you basically just triple the light bytes in nodes and make the lighting code loop through all three channels
16:16 Taoki        Right. Does it noticeably decrease performance on older machines however? Can anyone confirm they sense a performance decrease?
16:16 Taoki        Since that doesn't sound like a lot more harder task
16:20 celeron55_   i am not aware of how much of cpu time the lighting currently takes
16:20 celeron55_   at map generation time it is already a real PITA
16:20 celeron55_   but i don't know of the updates at other times
16:21 Taoki        Probably very little... calculating 1-3 integer values per node should be easy (I assume each lighting point is a value between 0 and 255)
16:21 celeron55_   i wouldn't make any assumptions like that
16:22 celeron55_   it's a recursive algorithm that generally has a range of 15 nodes to each direction
16:23 celeron55_   in 3d space that isn't exactly lightweight
16:23 Taoki        I see
16:26 celeron55_   also, there are already two values per node (day and light)
16:26 celeron55_   night*
16:26 celeron55_   or, actually not day and night, but rather "with sunlight" and "without sunlight"
16:29 RealBadAngel im actually going to upgrade my box by the end of the yr
16:29 RealBadAngel 3 yrs old hardware is a bit too old
16:30 RealBadAngel btw testing now the patch with changes
16:32 RealBadAngel already wrote a letter to santa ;)
16:32 RealBadAngel need SDD, some more RAM, better GFX and new MOBO
16:33 RealBadAngel and quad core instead of dual
16:35 RealBadAngel if santa will be kind enough maybe bigger lcd, and old one will go as secondary
16:38 Taoki        bbl, later
16:39 PilzAdam     is it ok to fix this two things in one commit? https://github.com/celeron55/minetest_game/pull/41
16:40 celeron55_   they're much related and small -> yes
16:41 RealBadAngel what about torches being able to place underwater?
16:41 PilzAdam     anything against the commit?
16:42 RealBadAngel its ok with me, shrubs on my cacti farm was makin  me nervous ;)
16:42 PilzAdam     RealBadAngel, this is part of this problem: Water should flush walkable/buildable nodes away
16:45 PilzAdam     i had several attempts to do this in Lua, but it is all too slow
16:46 RealBadAngel water spread code should be the place
16:48 celeron55_   i guess liquids flushing buildable_to nodes could be added
16:49 celeron55_   it's not even a group so making a fast check in C++ is like doing nothing
16:49 PilzAdam     this wouldnt effect torches; i would prefer walkable
16:49 celeron55_   you mean non-walkable?
16:49 PilzAdam     this walkable thing confuses me all the time ;-)
16:49 celeron55_   walkable is more like "collidable" or "solid" but whatever
16:49 celeron55_   there is a problem with that
16:49 PilzAdam     i always think walkable=you can walk into the node
16:50 celeron55_   for example snow
16:50 celeron55_   and other very small nodeboxes
16:50 PilzAdam     add a field water_resistent?
16:50 PilzAdam     *flushable; default = true
16:51 celeron55_   i guess making the check to be (buildable_to == true || walkable == false) could work
16:51 celeron55_   i don't think a new field is a good way at all
16:52 celeron55_   it should just work
16:53 PilzAdam     <celeron55_> and other very small nodeboxes <- what do you mean exactly?
16:53 celeron55_   snow mod.
16:53 celeron55_   put water on snow
16:54 PilzAdam     ah, okay
16:54 celeron55_   and now i'm off to something else, i'll seriously burn out if i try to answer even half of your stuff
16:54 celeron55_   ->
16:55 PilzAdam     heh
17:01 RealBadAngel celeron55_, im ready with testing the patch. now the question remains bout protocol change.
18:06 RealBadAngel celeron55_, are you there? what shall i do with protocol version?
18:08 PilzAdam     <celeron55_> and now i'm off to something else [...]
18:10 RealBadAngel and im here with finished patch, and Taoki expecting to get it today
18:10 RealBadAngel to merge 3D players
18:11 RealBadAngel i think adding extra checks, when next update make them obsolete is pointless
18:11 RealBadAngel so im adding protocol version change and posting the final patch