Time  Nick         Message
05:09 hmmmm        i am having a lot of trouble with the lua stuff
05:10 hmmmm        seemingly randomly -- it's based on coordinates though -- when i access a value in the array returned by my function it'll apparently be nil
05:10 hmmmm        is this a common problem, maybe?
05:12 hmmmm        http://pastebin.com/jJBNtu0r for the hell of it, i'll paste the problem here.  someone might be able to help.
05:13 hmmmm        i am not setting anything to nil - everything from the C++ side is exactly as expected
05:13 hmmmm        seemingly, anyway.
05:13 hmmmm        i modified WorldEdit 4 in order to test this
05:29 hmmmm        i have a feeling that the lua stack is getting screwed up
05:30 hmmmm        if i add print(noiseval) after local noiseval = ...
05:30 hmmmm        it'll print out on the first iteration and all, but on the second, it errors with "bad argument #1 to add_node, table expected, got nil"
05:34 VanessaE     the Lua code looks clean to me, on first glance
05:35 VanessaE     also, you don't need ,1 in your for loops.  just fyi,.
05:35 hmmmm        nevermind what i said previously, that was me being dumb and commenting out np.  really, it tries to call kill() when i pass a nil to print().  so anyway i tried print(type(noiseval) and i am getting interesting results.. every x = 2 and 3 are nil and the rest are numbers
05:35 hmmmm        every 2 and 3... .... ...WTF
05:39 hmmmm        if this were some sort of memory corruption i would be having way more problems everywhere else, no..?
05:39 VanessaE     depends on what's getting corrupted.
05:40 VanessaE     in the old days, I would have said you might be overflowing some memory area by just a few bytes - not enough to cause major issues, just a few wtf's
05:40 hmmmm        i'm clearly not though
05:41 VanessaE     right, in this day and age I guess it's more difficult to do that sort of overflow
05:41 hmmmm        it's not - but there's just nowhere for me to do that
05:43 hmmmm        somewhere in between the C++ to Lua code, those tables are getting really corrupted
05:44 hmmmm        could it be getting garbage collected?
05:45 VanessaE     no idea
05:45 hmmmm        you should be able to return a 2d array in lua
05:45 hmmmm        i don't see why not
06:07 hmmmm        i'm going to take a break from this bullshit
10:02 Progers_     Hi:D
10:02 Progers_     hi celeron55 )
10:03 Progers_     celeron55 you are on Ubuntu?
10:08 sfan5        i think he's busy right ow
10:08 sfan5        *now
10:12 RealBadAngel yeah, with snorring ;)
10:13 RealBadAngel its sunday mornin, almost no folks there
10:14 sfan5        morning? its 11:16
10:15 RealBadAngel well, since its sunday... ;)
10:16 RealBadAngel i also slept way longer than usual
12:33 thexyz       just as celeron55_ requested, all travis builds fail now ><
12:34 thexyz       apparently, that's because they updated clang to 3.2
12:37 RealBadAngel at least it worked for a few days ;)
13:11 sfan5        :D
13:43 RealBadAngel Calinou, you there?
13:45 RealBadAngel oops
17:37 Progers      Hi
20:26 Progers      hi
20:59 PilzAdam     thexyz, this should be done too: https://github.com/PilzAdam/minetest/commit/28f5b55643080e07a89af5a2e21ffadf3781be19
20:59 thexyz       PilzAdam: merge it then
21:00 thexyz       though I don't really like the whole idea of "on_rightclick"
21:00 thexyz       especially the fact that everything is done server-side
21:09 thexyz       celeron55: I rebased my leveldb pull request https://github.com/celeron55/minetest/pull/326
22:17 dsimon       hey all, i'm looking for some advice on what i should do next on my client_requests_blocks branch
22:17 dsimon       it's been a while since i worked on it, and c55 helped me out a bit with it
22:17 dsimon       but it never got to a point where it was ready to be merged with master
22:18 dsimon       in particular, has anything about the mapblock model or network infrastructure changed massively in the last 6 months or so that i need to adjust for?
22:19 dsimon       and also, what's the best way to stress test?
22:20 dsimon       the branch is part of c55/minetest github if anybody wants to poke at it
22:34 dsimon       brb
23:47 hmmmm        oh my god
23:47 hmmmm        i'm going to murder something
23:47 hmmmm        want to know what my lua problem was?
23:48 hmmmm        the numbers i was using as index arrays in Lua were apparently a bit off or something
23:48 hmmmm        something that didn't show up when i print()ed the coordinates
23:48 hmmmm        as a result i'd get nil instead of my number
23:48 hmmmm        so i have to use map[math.floor(x - x1) + 1] ...
23:49 hmmmm        why does lua have to be dynamically typed
23:49 hmmmm        dynamically typed languages are HORRIBLE.
23:52 dsimon       or at least dynamically typed languages with bad error reporting and overly "helpful" automatic type casting
23:52 hmmmm        oh hey wait you're the one doing the client_requests block branch
23:52 dsimon       yah
23:52 hmmmm        you went missing for a couple months there
23:52 dsimon       yeah, 5 :-)
23:52 dsimon       work got really busy
23:52 dsimon       but i'm back to having at least some free time now
23:53 hmmmm        no problemo
23:53 hmmmm        the mapblock model hasn't changed, neither has the network infrastructure... a couple packets have been changed and added but nothing that would break what you currently have
23:54 hmmmm        so what does it need to get finished, would you say?
23:54 dsimon       well, first off, there are some bugs
23:55 dsimon       sometimes blocks aren't being transferred
23:55 dsimon       which is not a good bug since that issue is the whole point of my changes :-)
23:55 dsimon       but i should be able to track it down; it's more likely to be an outright logic error than some subtle race condition thing
23:55 hmmmm        oh that reminds me, i forgot to tell you about the addition of TCP for bulk transfers
23:56 dsimon       secondly, i need to get it working under load
23:56 hmmmm        I believe right now it's only being used to transfer media, not blocks, so
23:56 dsimon       ah, ok
23:56 dsimon       that should be alright, my changes only involve blocks
23:57 dsimon       anyways, c55 tried out my changes and said that there were some performance issues
23:57 dsimon       he made a few improvements to address that, mostly tweaking various delay constants
23:57 hmmmm        i see
23:58 hmmmm        he kept telling me to take a look at it but i never really did
23:58 dsimon       but i'd like to see if i can't get it working better at core
23:58 hmmmm        i can't give too much input on it
23:58 dsimon       which is why i'm interested in stress testing
23:58 dsimon       ok
23:58 hmmmm        i've been working on a lot of other things related to the mapgen and noise
23:58 dsimon       the main point of it is like the name says: the client requests what blocks it wants, instead of the server providing the blocks that it predicts the client needs
23:58 hmmmm        well, i know that much
23:58 hmmmm        as far as specifics with your code though
23:59 dsimon       ok
23:59 dsimon       well, the data structure for mapblock changes in one way: there's a counter for the # of changes made, basically a version number
23:59 dsimon       the client requests blocks and specifies what version, if any, it already has
23:59 dsimon       that way the server does not have to retransmit blocks that are unchanged