Time  Nick         Message
00:49 VanessaE     looking at sapier's latest menu stuff, the servers' flags need to be identified as such, or at least visibly divided from the descriptions.
00:49 VanessaE     it looks (literally) like garbage at first
00:51 VanessaE     even a simple vertical line would probably be enough.
08:26 sapier       I can add a simple line, I already tried that. due to the indention mess it doesn't really look good either. are flags really required in list or could they be moved to description too?
10:48 kahrl        oh wow, this actually seems to work
10:49 kahrl        I changed CONTENT_IGNORE and CONTENT_AIR back in my patch and added some code to nodedef.cpp to make the vector large enough initially
10:49 kahrl        http://paste.dy.fi/F99
10:50 kahrl        does anyone dare to test it?
10:50 kahrl        for me it works in singleplayer mode and multiplayer and should be compatible with old servers
10:51 kahrl        brb
11:44 PilzAdam     kahrl, seems to work fine for me
11:45 PilzAdam     tested singleplayer, old client new server, old server new client and new server and new client
11:46 kahrl        awesome
11:46 PilzAdam     (everything local, though)
11:47 kahrl        shouldn't matter
11:47 PilzAdam     and IPv6
12:41 Exio4        i would change those to 0, 1, 2 in a future when something needs to break though
12:42 PilzAdam     why not change it depending on the protocol version?
12:58 proller      any objections? https://github.com/proller/minetest/commit/97559798d79d74359566d43bb5f01571f7885414
13:55 RealBadAngel guys. this is how shaders work atm: http://i.imgur.com/cN7LRp9.jpg
14:16 Taoki        I love it :)
16:43 sapier       proller why do I need to add a node to transforming liquid engine? Isn't this a property of a node?
16:44 proller      its queue,
16:44 proller      in this branch already working falling liquids
16:45 proller      but too buggy falling code
16:45 sapier       so this shall replace the add function later?
17:02 VanessaE     kahrl: dare I try that patch? :)
17:03 kahrl        yes you do :P
17:05 VanessaE     couple of whitespace warnings, but it applied....
17:07 VanessaE     well, it compiles....
17:09 VanessaE     here goes nuthin'
17:10 VanessaE     seems to work so far
17:10 VanessaE     care to try connecting with 0.4.7-stable?
17:11 kahrl        it won't work in your case
17:11 VanessaE     oh.
17:12 kahrl        new servers are only compatible with old clients if the 4096 node limit is not exceeded
17:12 VanessaE     oh, well I don't exceed it with mods, just those unknown blocks were doing that.
17:12 kahrl        yeah, sadly old clients can't handle that even on a new server
17:12 VanessaE     looks like I'm *just* under the limit:  3:32:16: VERBOSE[main]: registerNode: registering content id "4092": name="worldedit:placeholder"
17:14 VanessaE     PilzAdam did say it seemed to be cross-compatible though
17:14 kahrl        yeah, since he probably stayed far below the limit
17:18 VanessaE     well at least the server doesn't crash anymore :P
17:18 kahrl        that's good :)
17:18 kahrl        so if everything seems fine I will push it
17:19 VanessaE     I seem to be able to dig/build okay
17:19 VanessaE     nothing looks out-of-the-ordinary
17:19 kahrl        or, since I have to go right now, if any dev agrees to push it go ahead
17:19 kahrl        this is the link again: http://paste.dy.fi/F99
17:20 kahrl        I'll be back</terminator> in an hour or two
17:22 VanessaE     the irony of all this effort is I'm having trouble finding any of those damned unknown nodes :)
17:24 VanessaE     hah, finally found one :D
17:48 VanessaE     there's still a problem it appears.
17:49 VanessaE     I think I just tried to exceed 4096, by accident, and am getting "C++ exception" at startup in a mod I'm not even messing with (I'd renamed a mod to avoid a conflict, since the half a dozen or so register_node() calls therein have colon prefixes, one of which redefines a node)
17:50 proller      preparing for leveled snow - https://github.com/proller/minetest/commit/2024a6c92d5710c8684c4dd0b48e8b93c5a49285
17:52 Taoki        nice nice
17:52 proller      VanessaE, and here ^^ is new function for getting level of liquid or..
17:53 VanessaE     proller: ok.  Once you guys are stable on that whole thing, I'll look into using it :)
17:54 hmmmm        actually that commit looks pretty good
17:54 hmmmm        lol, great abuse of the ternary operator
17:54 hmmmm        +     return getParam2() & LEVELED_MASK ?: f.leveled > LEVELED_MAX ? LEVELED_MAX : f.leveled;
17:55 hmmmm        that bit should be cleaned up, otherwise it looks alright
17:55 proller      but how without temp vars?
17:56 hmmmm        yes
17:56 celeron55    so i saw the discussion about 0.5.0
17:57 celeron55    i seriously doubt the mob thing will happen almost ever
17:57 celeron55    simply nobody is interested in doing it
17:57 hmmmm        everybody is
17:58 celeron55    it's a good goal, but something has to change for it to be reached
17:58 hmmmm        and it will happen simply because it needs to happen, it's the next level in minetest
17:58 Calinou      actual C++ AI would be best for mobs
17:58 celeron55    our problem isn't AI
17:58 Calinou      and better entity management
17:58 celeron55    our problem is almost everything else than AI
17:59 hmmmm        mobf seems to be pretty fast, and with helper functions for mob placement there should be no "lag" problems
17:59 Taoki        celeron55: Everyone wants mobs, and there are 2 existing mob systems which are almost ready (even if not quite)
18:00 Taoki        I agree with hmmmm on this one
18:00 Calinou      >mobf >fast
18:00 hmmmm        i am prepared to focus all my effort on making mobs a reality
18:00 Taoki        +1 for that :)
18:00 celeron55    hmmmm: well that makes me a bit more convinced
18:01 proller      hmmmm, how to return EXPR ?: OTHEREXP   without ?:  ?
18:02 celeron55    i'm worried because in the end the level of responsibility and other might need client-side scripts or a deeper C++ implementation of mobs
18:02 Taoki        hmmmm: You might want to take a look at my highlights in this post: http://forum.minetest.net/viewtopic.php?pid=97878#p97878
18:02 celeron55    eh
18:02 celeron55    responsivity*
18:02 hmmmm        honestly i'm not sure what x ?: y does without checking on my own
18:02 hmmmm        if i hadn't seen you use that sort of construct, i would've thought it was a malformed statement and would cause a compiler error
18:03 proller      a?:b  === a?a:b
18:04 hmmmm        figured as much
18:05 hmmmm        you're not going to be able to turn that into nice code without making it more than 5 lines, i'd say
18:05 hmmmm        i'd try something such as:
18:05 Taoki        hmmmm: My primary question is, why is it that we can't include SimpleMobs yet. What's still so wrong about it?
18:05 hmmmm        u8 level = getParam2() & LEVELED_MASK;
18:05 Taoki        Apart from maybe the fact that it's too simple. And of course the things I pointed out in that post and maybe more, but such can be fixed over time
18:05 hmmmm        if (level) \n\t return level;
18:05 proller      already, 4 lines 8(
18:06 hmmmm        return (f.leveled > LEVELED_MX) ? LEVELED_MAX : f.leveled;
18:06 hmmmm        proller, more lines of code are better
18:06 hmmmm        for something like this anyway
18:06 proller      ok.
18:06 proller      (i'm from perl 8-)
18:06 hmmmm        not sure about others here, but i'm pretty sure everybody else hates nested ternary operators without parentheses
18:07 celeron55    i wonder how many lines some java programmer would make from that
18:07 celeron55    like, i'm sure they never put more than a single variable on a "return" line for example 8)
18:07 hmmmm        well they'd need to make a Param2IteratorFactorySingleton class, which would be its own file
18:08 hmmmm        i think the GPL license is about 40-50 lines, along with all the necessary doxygen comments, my estimate would be 80 lines total
18:09 Calinou      general public license license :D
18:12 hmmmm        meant to say GPL comment text
18:12 celeron55    the gnu gpl public license
18:15 proller      where to make snow<->water melting abms? i prefer in c++
18:15 hmmmm        i'd leave it out for now at least
18:16 proller      leave in core?
18:17 hmmmm        people might not want snow to melt into water though
18:18 proller      PilzAdam - not, but others - yes
18:18 proller      near lava, torch, it must be, in other places - depend on temperature
18:19 hmmmm        well, i don't
18:19 hmmmm        it'll cause problems if finite liquids aren't enabled
18:19 proller      depend on it
18:19 hmmmm        maybe only add that abm if finite liquids are on
18:19 proller      +1
18:28 Anchakor1    RealBadAngel: do you have pics of the shaders with low-res textures? (perhaps low-res upscaled without anti-aliasing?)
18:29 RealBadAngel i do have with 32x
18:30 RealBadAngel http://i.imgur.com/D7lWBNL.jpg , default technic_game textures, shaders with filtering enabled
18:31 RealBadAngel for 16x theres no point in turnin shaders on
18:31 Calinou      there is
18:31 Calinou      better lighting, day-night cycle smoothing
18:31 Calinou      looks much better
18:32 RealBadAngel ah, yes
18:32 RealBadAngel but thats it. i meant bumpmapping, parallax occlusion and specular
18:32 Anchakor1    doesn't look good
18:33 RealBadAngel because such effects are not for low res textures
18:33 RealBadAngel imho 128x textures are absolute minimum to use bumpmapping
18:34 RealBadAngel and even then filtering has to be enabled
18:35 RealBadAngel http://i.imgur.com/cN7LRp9.jpg
19:05 Taoki        RealBadAngel: Is parallax close to merge ready state now? Really curious to be able to enable it ^^
19:06 RealBadAngel i need a few more days, changes are pretty complex
19:07 VanessaE     kahrl: as you figured, old clients can connect to the patched server, but receive this error:  http://i.imgur.com/1QLKOCa.png
19:08 VanessaE     kahrl: is this fixable at all?
19:09 Anchakor1    hmm all pics of the shaders I saw make things glossy and muddy
19:12 VanessaE     you mean the bumpmap stuff?
19:13 VanessaE     that's because the bumpmaps themselves are too strong.  easily fixed given enough time to tune them (see HDX, they're much more subtle there, though some additional tuning is needed)
19:14 Taoki        I like the current looks. Bumpmaps are just a bit useless without real lighting, but parallax is fixing that
19:25 Taoki        RealBadAngel: http://forum.minetest.net/viewtopic.php?pid=99492#p99492 This will prolly help with technic_game machines :)
19:29 Taoki        RealBadAngel: But it appears you might not be able to set chest inventories via meta:set_string. This really sucks and IMO is a weird system. Hopefully this will work to set any other meta for spawned structures
19:30 Taoki        eg: In case of machines, already spawn them pre-configured and functional
19:31 RealBadAngel that is nice
19:55 * VanessaE   pokes kahrl
19:56 proller      can we use "auto i = f()" ?
19:58 Exio4        auto? isn't that something of the new spec?
19:59 proller      yes
20:00 proller      c 11
20:15 hmmmm        no C11
20:37 kahrl        VanessaE: not fixable, at least not without major rework
20:37 VanessaE     damn.
20:37 kahrl        I assume there were other clients connected to the server before that one
20:37 VanessaE     eh?
20:38 kahrl        so the server already had to allocate dummy node IDs and the count went over 4096
20:38 VanessaE     oh, definitely so.
20:38 kahrl        (these are sent to new clients that connect later in the node definition packet)
20:38 kahrl        if it was the first client, there would have been no dummy node IDs yet
20:39 VanessaE     oh right, I got you now
20:39 kahrl        and old clients produce that error message in that case; there is a bug in the old deserialization code by the way
20:39 VanessaE     well I guess push that to git and at that point you may as well do whatever protocol revision you need to for other stuff and call it 0.4.8
20:40 kahrl        https://github.com/minetest/minetest/blob/master/src/nodedef.cpp#L796
20:40 VanessaE     (as long as it's gonna break, you may as well break everything all at once ;)  )
20:40 kahrl        if i > MAX_CONTENT, this code reads the node ID but doesn't read the ContentFeatures that follow
20:41 kahrl        my patch fixes that as well :)
20:42 kahrl        I'm not sure we need a new protocol version
20:42 VanessaE     what can old clients do to "clear out" those nodeid's?
20:43 VanessaE     or are we just stuck with another divide?
20:43 kahrl        most servers should not see any breakage
20:43 VanessaE     heh, it figures :P
20:44 kahrl        just put a note in your server topic that only dev versions can connect for now
20:44 VanessaE     fair enough
20:44 VanessaE     what about that C++ exception I mentioned?
20:44 kahrl        it's expected
20:44 VanessaE     ok
20:45 kahrl        caused by the bug I mentioned
20:45 VanessaE     ok, gotchya
20:46 VanessaE     I'll keep my mod set at its last good set then.
20:46 VanessaE     er...yeah
20:47 kahrl        brb
22:37 kahrl        any word on #822 and my patch?
22:38 VanessaE     you should *probably* link your patch to that issue :)
22:39 VanessaE     (#791 was also supposed to fix this also, but apparently that one doesn't work)
22:41 kahrl        I haven't looked at it in detail but I think #791 is at least missing changes to deSerialize()
22:42 kahrl        and doesn't handle unknown nodes properly
22:42 VanessaE     I haven't looked either, just happened to notice it the other day while I was thumbing through the issue list
22:42 kahrl        I linked it in #822, in fact :)
22:42 VanessaE     well that too :)
22:43 VanessaE     (but I knew about it before you linked :P )
22:44 VanessaE     I think if you merge that, some kind of announcement that the other server owners will likely see is warranted.
22:45 kahrl        Something in the news section?
22:45 kahrl        of the forum
22:45 VanessaE     mmm...yeah I guess that would make sense
22:48 proller      froozen https://github.com/proller/minetest/compare/weather
22:48 VanessaE     holy..
22:48 kahrl        proller, that's why you don't merge master into feature branches
22:48 VanessaE     oh, hah looked like a huge pull request at first :D
22:49 proller      ups
22:49 proller      https://github.com/proller/minetest/compare/weather#files_bucket
22:50 proller      kahrl, no problem with merge
22:57 kahrl        https://github.com/minetest/minetest/pull/827
22:59 proller      how to make good rain & snow