| Time |
Nick |
Message |
| 00:51 |
|
bas080 joined #minetest-dev |
| 01:04 |
|
bas080 joined #minetest-dev |
| 01:26 |
|
OldCoder joined #minetest-dev |
| 01:40 |
|
NakedFury joined #minetest-dev |
| 01:45 |
|
deltib joined #minetest-dev |
| 01:53 |
|
deltib joined #minetest-dev |
| 02:05 |
|
bas080 joined #minetest-dev |
| 03:56 |
|
OldCoder joined #minetest-dev |
| 04:14 |
|
VargaD joined #minetest-dev |
| 04:32 |
|
AllegedlyDead joined #minetest-dev |
| 04:51 |
|
BrandonReese joined #minetest-dev |
| 05:01 |
|
kaeza joined #minetest-dev |
| 05:07 |
|
us_0gb joined #minetest-dev |
| 05:33 |
|
BrandonReese joined #minetest-dev |
| 06:18 |
|
Zeitgeist_ joined #minetest-dev |
| 06:50 |
|
BrandonReese joined #minetest-dev |
| 06:52 |
|
ImQ009 joined #minetest-dev |
| 08:12 |
|
werwerwer joined #minetest-dev |
| 08:26 |
|
ImQ009 joined #minetest-dev |
| 09:18 |
|
salamanderrake joined #minetest-dev |
| 09:46 |
|
jojoa1997 joined #minetest-dev |
| 10:37 |
|
proller joined #minetest-dev |
| 10:57 |
|
jojoa1997 joined #minetest-dev |
| 11:12 |
xyz |
wtf is this |
| 11:12 |
xyz |
https://github.com/minetest/minetest/blob/master/src/guiFormSpecMenu.cpp#L388 |
| 11:14 |
xyz |
why is label set to L"" in constructor and fdefault to label text and then flabel is assigned the label text |
| 11:15 |
|
proller joined #minetest-dev |
| 11:27 |
|
Zeitgeist_ joined #minetest-dev |
| 11:29 |
|
ImQ009 joined #minetest-dev |
| 12:23 |
|
ImQ009 joined #minetest-dev |
| 12:26 |
|
PilzAdam joined #minetest-dev |
| 12:46 |
xyz |
perhaps deSerializeLongString should at least check size of incoming buffer? |
| 12:46 |
xyz |
otherwise it's way too easy to DoS any server |
| 12:58 |
|
Zeitgeist_ joined #minetest-dev |
| 13:01 |
|
Zeitgeist_ joined #minetest-dev |
| 13:47 |
|
hmmmm joined #minetest-dev |
| 13:52 |
RealBadAngel |
hi hmmm |
| 13:52 |
RealBadAngel |
ive changed waypoints a bit: http://i.imgur.com/nPbRaU6.png |
| 14:12 |
|
OldCoder joined #minetest-dev |
| 14:16 |
|
tiemay joined #minetest-dev |
| 14:18 |
tiemay |
when installing minetest, should "minetest-minetest_game-1a9362a" be renamed to "minetest_game"? |
| 14:18 |
RealBadAngel |
yes |
| 14:18 |
tiemay |
thanks |
| 14:26 |
|
proller joined #minetest-dev |
| 14:43 |
|
zat joined #minetest-dev |
| 14:49 |
|
proller joined #minetest-dev |
| 14:50 |
|
iqualfragile joined #minetest-dev |
| 15:01 |
|
proller joined #minetest-dev |
| 15:02 |
|
rsiska joined #minetest-dev |
| 15:16 |
|
Yepoleb joined #minetest-dev |
| 15:22 |
|
rsiska joined #minetest-dev |
| 15:48 |
RealBadAngel |
anybody have anything more against https://github.com/minetest/minetest/pull/1112 ? |
| 15:48 |
RealBadAngel |
if not, its ready to merge |
| 15:49 |
xyz |
well if you're asking for nitpicking |
| 15:49 |
xyz |
you've added tab here https://github.com/minetest/minetest/pull/1112/files#diff-18513665750ef5adf42b5ec29e14162eL2368 |
| 15:49 |
xyz |
and there's no space between number and "nodes/meters" |
| 15:50 |
RealBadAngel |
i was checking this tab |
| 15:50 |
RealBadAngel |
in my pull it looks like its properly formatted |
| 15:50 |
xyz |
the "distance suffix" is so stupid i can't believe you actually implemented it |
| 15:50 |
RealBadAngel |
that was on demand |
| 15:52 |
xyz |
yea asking you to remove it now would seem like trolling |
| 15:52 |
xyz |
well other than that I have nothing to say, I think it's fine |
| 15:53 |
RealBadAngel |
on the other hand, if some1 decide to make space flight simulator out of the engine |
| 15:53 |
RealBadAngel |
that will be handy |
| 15:54 |
xyz |
hm? |
| 15:54 |
RealBadAngel |
ah, i can see whats wrong with that tab, will fix it |
| 15:55 |
RealBadAngel |
remove the mapgen, add only entities, move the player as if he were a ship |
| 15:55 |
RealBadAngel |
add some more hud elements |
| 15:55 |
RealBadAngel |
and voila, space 3d shooter game is here ;) |
| 15:56 |
RealBadAngel |
distance in meters in space would be just stupid :) |
| 15:58 |
xyz |
oh, that's deep |
| 16:00 |
|
bas080 joined #minetest-dev |
| 16:03 |
RealBadAngel |
;) not really, just adding more sense to that ;) |
| 16:03 |
proller |
space flight simulator - main target for minetest |
| 16:17 |
|
Jordach joined #minetest-dev |
| 16:28 |
|
Calinou joined #minetest-dev |
| 16:34 |
sfan5 |
can we merge #1110 already? |
| 16:34 |
ShadowBot |
https://github.com/minetest/minetest/issues/1110 -- Include system info in the HTTP user agent on Windows by sfan5 |
| 16:40 |
|
kaeza joined #minetest-dev |
| 16:56 |
kahrl |
I'm okay with it so yeah |
| 17:20 |
RealBadAngel |
ive fixed that tab in waypoints commit |
| 17:24 |
RealBadAngel |
https://github.com/minetest/minetest/pull/1112 |
| 17:38 |
xyz |
yay! |
| 17:39 |
xyz |
while we're at it |
| 17:39 |
xyz |
can you explain once again why using irrlicht's function to get 2d screen pos from 3d world pos won't work? |
| 17:43 |
|
VargaD joined #minetest-dev |
| 17:49 |
tiemay |
yay! chakra has built in Colemak support! I ran `setxkbmap us -variant colemak`. how can I set it back to qwerty? |
| 17:49 |
tiemay |
oops |
| 17:52 |
|
rubenwardy joined #minetest-dev |
| 18:00 |
RealBadAngel |
xyz, i need coords in range 0-1, not in ablos |
| 18:00 |
RealBadAngel |
oops, ... not in absolute screen coordinates |
| 18:02 |
RealBadAngel |
i will need them for another features |
| 18:08 |
|
exoplanet joined #minetest-dev |
| 18:20 |
|
rubenwardy_ joined #minetest-dev |
| 18:26 |
xyz |
RealBadAngel: in range 0-1 so they will be relative to what? |
| 18:27 |
|
AllegedlyDead joined #minetest-dev |
| 18:28 |
|
zat joined #minetest-dev |
| 18:31 |
|
grrk-bzzt joined #minetest-dev |
| 18:48 |
|
salamanderrake joined #minetest-dev |
| 18:58 |
|
EvergreenTree joined #minetest-dev |
| 18:58 |
|
EvergreenTree joined #minetest-dev |
| 19:15 |
|
salamanderrake joined #minetest-dev |
| 19:30 |
|
proller joined #minetest-dev |
| 19:49 |
|
xyz joined #minetest-dev |
| 19:53 |
|
Neological joined #minetest-dev |
| 20:14 |
|
sapier joined #minetest-dev |
| 20:14 |
sapier |
https://github.com/minetest/minetest/pull/1116 |
| 20:14 |
sapier |
ShadowNinja: could you try if this fixes your areamods issues? |
| 20:15 |
sapier |
Celeron55 joining players will be shown after completing download of media as of this pull |
| 20:16 |
PilzAdam |
"this will result in old clients failing to join on login without having completed download of all media in a previous session. Is this a big problem?" I guess no; we cant promise forward-compatibility |
| 20:16 |
sapier |
I haven't removed the conlock because it will be required if we separate scriptapi to it's own thread, or allow async api to access server functions |
| 20:18 |
troller |
again remixing of anything |
| 20:18 |
sapier |
we could fix this issue by adding evaluating client version and handle old clients in a different way. But if there isn't another issue then old clients have to join twice I don't think it's worth the additional code |
| 20:19 |
sapier |
troller first read then shout |
| 20:19 |
sapier |
If you had you'd have seen what you call "remix" is actually consolidating duplicated code to single functions |
| 20:20 |
xyz |
sapier: apparently fetching serverlist blocks game start, i.e. if servers.minetest.net is slow and you try to play a local game the client hangs for quite some time |
| 20:20 |
troller |
1,121 additions and 891 deletions. |
| 20:20 |
troller |
^ good remove of duplicated code |
| 20:20 |
sapier |
shouldn't serverlist be read in a async call? |
| 20:22 |
xyz |
dunno? |
| 20:22 |
sapier |
you're talking about a delay in mainmenu aren't you? |
| 20:22 |
xyz |
yes |
| 20:23 |
troller |
sapier, why it not splitted in separate 15 pulls? |
| 20:24 |
sapier |
because it's a single feature |
| 20:24 |
sapier |
that feature just fixes multiple issues at once |
| 20:25 |
troller |
again imagined issues? |
| 20:26 |
sapier |
xyz it should be read asynchronously ... hmm but I don't know if it's possible to abort a http request maybe it has to finish for mainmenu to close |
| 20:26 |
|
khonkhortisan joined #minetest-dev |
| 20:27 |
xyz |
indeed, but for some reason it waits and this is EXTREMELY annoying |
| 20:27 |
VanessaE |
"-Joining player standing in world long before player can move" \:D/ |
| 20:27 |
|
zat joined #minetest-dev |
| 20:28 |
sapier |
btw vanessae if this is reviewed and last bugs are out your servers shouldn't be vulnerable to out of memory attacks too |
| 20:28 |
sapier |
at least to a special variant |
| 20:28 |
VanessaE |
sapier: oh you mean that string-too-long bit that comes up every so often? |
| 20:29 |
sapier |
no I mean that queue all textures 20 times if 20 clients connect at same time |
| 20:29 |
sapier |
39MB * 20 clients ... 800mb ok I guess your hell of a machine could stand this for one of your servers |
| 20:30 |
sapier |
whats your player max? |
| 20:30 |
xyz |
hey i just wanted to talk about a memory attack |
| 20:30 |
sapier |
I added a max simultaneous connnect setting in this pull |
| 20:31 |
troller |
it must be = 1 ! |
| 20:31 |
sapier |
it's not best option but it'll fix the issue until we can add better variants |
| 20:31 |
|
tomasbrod joined #minetest-dev |
| 20:31 |
rubenwardy |
If players send chat messages to a joining player, has the issue been solved where the joining player does not recieve them (is not ready to recieve chat) |
| 20:31 |
rubenwardy |
? |
| 20:32 |
sapier |
yes ... somehow :-) other players won't see the joining player until he's ready to receive ... ok there's a very very small gap in between |
| 20:32 |
|
tomasbrod left #minetest-dev |
| 20:33 |
sapier |
hmm no there ain't a gap for time beeing, that gap might reoccur on async scriptapi ... but not for player <-> player chat but mod <-> broadcast only |
| 20:34 |
xyz |
I'm not exactly sure how bad it is |
| 20:34 |
sapier |
I guess you can't really expect a broadcast to be received by a joining player so I don't consider this to be a problem |
| 20:34 |
xyz |
well to know what I'm talking about just check the code of deSerializeLongString, it's really simple bug |
| 20:35 |
sapier |
I haven't fixed anything there this is most likely a different bug |
| 20:35 |
xyz |
I wasn't talking to you ;) |
| 20:36 |
sapier |
oh ... :-) |
| 20:36 |
VanessaE |
sapier: player max on my servers is 50. |
| 20:36 |
xyz |
this will just make server allocate 4gb of memory or so |
| 20:36 |
sapier |
Ahh another thing, I don't consider 1116 to be production ready, I could only do basic testing it most likely contains bugs |
| 20:37 |
xyz |
it doesn't check actual length of the buffer anywhere |
| 20:37 |
sapier |
ok this single server can allocate 2gb at once in worst case |
| 20:37 |
VanessaE |
at present, my "almost vanilla" server has 12 clients connected and actively playing, another couple apiece on the nostalgia, "survival", and lua mapgen test sites. creative has 1 more, and realtest is idle. |
| 20:37 |
VanessaE |
(at least, so says the master server lit) |
| 20:37 |
VanessaE |
(list*) |
| 20:38 |
sapier |
it's not a matter of connected clients, actually connected clients reduce risk ;-) |
| 20:38 |
VanessaE |
sapier: that was to proller's chagrin ;) |
| 20:40 |
xyz |
12 CLIENTS! JUST IMAGINE THIS NUMBER! |
| 20:40 |
sapier |
considering usual 2 clients per server this seems to be an improvement ... no more no less |
| 20:40 |
VanessaE |
xyz: it was double that earlier, but I restarted at the request of several users to install some skins and change a few settings. |
| 20:42 |
xyz |
that was usual until buildcraft came |
| 20:43 |
VanessaE |
whatever the reason, this is immaterial to sapier's subject. |
| 20:43 |
sapier |
talking about buildcraft is celeron working on a android version or did he just play with his android toy? |
| 20:45 |
xyz |
I don't remember him working on android version |
| 20:45 |
xyz |
he was working on sailfish (or something) version |
| 20:46 |
|
sapier1 joined #minetest-dev |
| 20:49 |
xyz |
so I guess nobody is working on android version right now |
| 20:50 |
* sfan5 |
isn't working on an android version either because irrlicht broken things |
| 20:50 |
sfan5 |
aka never update something that works |
| 20:50 |
sapier1 |
oh you're right sailfish not android |
| 20:50 |
xyz |
what things? |
| 20:51 |
ShadowNinja |
proller doesn't seem to have said anything helpfull within the past few days... |
| 20:51 |
sfan5 |
it doesn't show the mainmenu, just black screen |
| 20:51 |
sapier |
no error message at all? |
| 20:51 |
sfan5 |
dunno whether it crashes, haven't researches further yet (no time) |
| 20:51 |
sfan5 |
no |
| 20:51 |
sapier |
sad |
| 20:51 |
sfan5 |
researched* |
| 20:52 |
sfan5 |
it doesn't even create a debug.txt |
| 20:52 |
xyz |
oh cool, a lot of updates |
| 20:52 |
xyz |
what about gdb? |
| 20:52 |
ShadowNinja |
Mostly just criticising sapier's pulls without looking at them. |
| 20:52 |
xyz |
ShadowNinja: why do you think he's not looking? |
| 20:52 |
xyz |
how you can know? |
| 20:52 |
sfan5 |
I can try to debug it, I just didn't have any time yet |
| 20:53 |
ShadowNinja |
xyz: He makes generic criticisms like "it's overenginered" or "you're fixing a imagined issue". |
| 20:53 |
xyz |
is he not allowed to? |
| 20:54 |
ShadowNinja |
xyz: Not without a basis for it, no. |
| 20:54 |
xyz |
sfan5: I guess a good idea would be to maintain our irrlicht fork with some patches |
| 20:54 |
sfan5 |
yeah |
| 20:55 |
xyz |
ShadowNinja: basis? |
| 20:56 |
ShadowNinja |
xyz: http://en.wiktionary.org/wiki/basis |
| 20:56 |
sapier |
ShadowNinja: please this discussion isn't worth time spent on it. Could you try 1116 if this is enough to fix your aremod? |
| 20:56 |
xyz |
ShadowNinja: well and what you're doing is criticizing proller for criticizing sapier |
| 20:56 |
xyz |
and asking to remove comments you don't like |
| 20:58 |
ShadowNinja |
The difference being that I have a basis. :-) But perhaps sapier's right. I'll check 1116. |
| 21:00 |
xyz |
yeah you don't like his personality, that's a strong one |
| 21:01 |
sfan5 |
xyz: how is this for touchscreen controls: http://sfan5.duckdns.org/upload/userdata/1/vout.mkv |
| 21:01 |
|
kahrl joined #minetest-dev |
| 21:02 |
xyz |
is it so hard to upload it somewhere with decent dl speed? |
| 21:02 |
sfan5 |
well sorry |
| 21:02 |
sfan5 |
yes |
| 21:02 |
xyz |
are you browsing from potato? |
| 21:02 |
xyz |
it's like 10 kb/s |
| 21:02 |
sfan5 |
oh.. that is because I'm torrenting stuff |
| 21:02 |
sapier |
:-) guess you have to wait for some minutes till we can say anything about this controls ;-) |
| 21:03 |
sfan5 |
ok, turned off the torrents |
| 21:03 |
sapier |
still slow but way better |
| 21:03 |
xyz |
hey it's up to 20 kb/s! |
| 21:03 |
sapier |
hmm ... not exactly |
| 21:03 |
xyz |
twice as much speed! |
| 21:03 |
sfan5 |
any complaints for dl speed to telekom.de please :P |
| 21:04 |
sapier |
I had 10-15 before now I have about 15 ... not a big difference |
| 21:04 |
xyz |
uploading it somewhere would be a good idea |
| 21:04 |
xyz |
no, really, what the fuck |
| 21:04 |
|
khonkhortisan joined #minetest-dev |
| 21:04 |
xyz |
10 kb upload speed in 2014 |
| 21:04 |
xyz |
my phone's 3g is like 10 times faster or something |
| 21:04 |
* sfan5 |
haz 100 kb upload |
| 21:04 |
sapier |
I don't have much more here :-/ |
| 21:05 |
sfan5 |
pls stop dl'ing, I'm uploading to dropbox |
| 21:06 |
ShadowNinja |
Too late, I got it. :-) |
| 21:07 |
VanessaE |
sfan5: the movement is good, but the big, HUGE contention is digging, placing, and especially climbing up/down |
| 21:07 |
VanessaE |
based on the sample of users from my servers. |
| 21:07 |
ShadowNinja |
Or part of it? |
| 21:07 |
sfan5 |
well, I only finished the looking-around yet |
| 21:07 |
ShadowNinja |
^ |
| 21:07 |
sfan5 |
https://dl.dropboxusercontent.com/u/30267315/vout.mkv |
| 21:07 |
VanessaE |
I got enough of the video to see what you intended to show at least. |
| 21:08 |
VanessaE |
I saw the climbing up/down should be dead easy - two buttons at one side of the screen that are only visible if you're on something that is climbable. (or a jump button appears there otherwise() |
| 21:08 |
VanessaE |
s/saw/say/ |
| 21:08 |
sfan5 |
that's a good ida |
| 21:08 |
sfan5 |
idea* |
| 21:08 |
VanessaE |
I mean shit, does it have to be so user-unfriendly as buildcraft clearly is? |
| 21:08 |
sfan5 |
no |
| 21:09 |
sfan5 |
it should be better ofc, so people will use official MT instead of builtshit |
| 21:09 |
VanessaE |
yes, 10000000000% agree |
| 21:10 |
VanessaE |
this tablet client is so busted-ass that half the requests for help I get on my servers boil down to "HELP! KILL ME! I'M STUCK!" |
| 21:10 |
VanessaE |
because they can't climb down a stinkin' ladder |
| 21:10 |
VanessaE |
or they can't read a G*ddamn sign. |
| 21:10 |
VanessaE |
(no infotext? idk) |
| 21:11 |
sfan5 |
anyway, code is here: https://github.com/sfan5/minetest/commit/3c17ff21e4ede9e467d3ae8c891afa9b7050d03c |
| 21:11 |
xyz |
i think the idea is faulty |
| 21:12 |
xyz |
maybe it'd be somehow okay for devices with 2 hardware sticks |
| 21:12 |
xyz |
but emulating it on touchscreen? dunno |
| 21:13 |
ShadowNinja |
sfan5: You set touchscreen to both true and false by default. |
| 21:13 |
sfan5 |
how about you can switch between what you've been working on and my thing? |
| 21:13 |
troller |
mechwarriors returns! |
| 21:13 |
VanessaE |
xyz: well the problem isn't what kind of physical device is used for the input. the problem is how you interpret that input |
| 21:13 |
sfan5 |
ShadowNinja: don't tell me setDefaults just does .push_back on a list |
| 21:13 |
sfan5 |
s/Defaults/Default/ |
| 21:14 |
xyz |
VanessaE: what? |
| 21:15 |
ShadowNinja |
sfan5: Hmmm? The second overrides the first. But it should be set only once. |
| 21:16 |
VanessaE |
xyz: well, you're implying that the control system for a game like minetest/craft isn't really viable on a tablet, yet? |
| 21:16 |
VanessaE |
yes*? |
| 21:17 |
xyz |
VanessaE: the scheme i saw in sfan5's video won't work |
| 21:17 |
VanessaE |
xyz: oh, I think he was just trying to demonstrate the positioning of the pointer relative to center |
| 21:17 |
ShadowNinja |
MineCraft PE seems to pull it off fairly well. |
| 21:18 |
VanessaE |
the pointer could just as easily have been the person's finger and the positioning of the box could just as easily have been down in the far lower corner. |
| 21:18 |
VanessaE |
tablets are a lot more accurate in their detection of where you're touching than it seems like |
| 21:18 |
VanessaE |
it's just that your fingers are so big that you gotta SEE what you're touching most of the time, hence large buttons |
| 21:19 |
VanessaE |
but with a proportional interface like what sfan's got there, you only need to know where the center is. |
| 21:19 |
VanessaE |
as soon as you take your finger off, I'm sure it assumes "centered" and stops moving/turning, yes? |
| 21:20 |
sfan5 |
yes |
| 21:20 |
sfan5 |
anything else wouldn't make sense |
| 21:21 |
xyz |
well who knows, this video only show this thing |
| 21:21 |
VanessaE |
then all it needs after that is a few buttons along the left side of the screen for climb down (which is greyed out most of the time), jump/climb up, and inventory, and maybe a button for chat |
| 21:22 |
VanessaE |
anything else can be buried inside the regular menu button since those things wouldn't be used as much. Any formspec that comes up with have a big, huge, green button that says "OKAY" and another big red one that says "CANCEL", added to then by the client. |
| 21:23 |
VanessaE |
positioned at the top or bottom of the screen |
| 21:23 |
VanessaE |
apparently buildcrap users are having trouble exiting from formspecs. |
| 21:23 |
xyz |
oh now you're as bad as people who call microsoft m$ |
| 21:24 |
xyz |
sfan5: how do you plan to do digging, punching and right clicking? |
| 21:24 |
* sfan5 |
is merging pull 1110 now because <kahrl> I'm okay with it so yeah |
| 21:24 |
sfan5 |
well.. |
| 21:24 |
VanessaE |
xyz: I've played with the client, and I've fielded enough help requests from people who actively play on it that I'm pretty sure I know what I'm talking about. |
| 21:24 |
VanessaE |
it IS crap. |
| 21:24 |
xyz |
and placing too |
| 21:24 |
sfan5 |
like you said with the touch, hold, double touch |
| 21:24 |
sfan5 |
s/said/proposed/ |
| 21:25 |
VanessaE |
tap to punch, hold to dig, reverse pinch to open./ |
| 21:25 |
VanessaE |
makes more sense to me |
| 21:25 |
sfan5 |
what's a reverse pinch? |
| 21:25 |
xyz |
lol |
| 21:25 |
VanessaE |
you know, two fingers, then you spread them apart. |
| 21:25 |
VanessaE |
the reverse of pinching them together |
| 21:25 |
VanessaE |
usually used to zoom on a web page |
| 21:25 |
xyz |
there's no "open" in minetest |
| 21:25 |
VanessaE |
xyz: open a chest? |
| 21:25 |
VanessaE |
open a door? |
| 21:25 |
VanessaE |
open any formspec? |
| 21:25 |
sfan5 |
s/open/rightclick/ |
| 21:25 |
VanessaE |
you know, the equivalent of right-clicking. |
| 21:26 |
VanessaE |
but as this is a tablet, there is no right-ANYthing |
| 21:26 |
xyz |
sfan5: so without a cursor? like in my gui? |
| 21:26 |
xyz |
VanessaE: yea open an apple and stuff |
| 21:26 |
|
Gethiox3 joined #minetest-dev |
| 21:26 |
sfan5 |
no, with cursor (that may chanche if it turns out that is super crappy) |
| 21:26 |
VanessaE |
xyz: come on man, make an effort. |
| 21:26 |
VanessaE |
you know perfectly well what I meant by that |
| 21:26 |
xyz |
ah ok |
| 21:27 |
xyz |
at least those will be different |
| 21:27 |
VanessaE |
99% of mods use the right click action to open a formspec or open/close a door-like object of some kind |
| 21:28 |
|
jojoa1997 left #minetest-dev |
| 21:30 |
xyz |
it's much easier to implement rightclick and well i think it'll be easier to use too |
| 21:30 |
sfan5 |
s/chance/change/ |
| 21:31 |
RealBadAngel |
xyz, not relative to what, all HUD parameters are given in such way, 0-1 for both x and y |
| 21:31 |
xyz |
ugh |
| 21:33 |
RealBadAngel |
then theyre scaled according to actual screen resolution |
| 21:33 |
VanessaE |
fine then implement it as right click |
| 21:33 |
xyz |
so relative to screensize |
| 21:33 |
xyz |
err, double click |
| 21:33 |
RealBadAngel |
also shaders are using 0-1 without even knowing actual resolution. |
| 21:34 |
xyz |
well you can call that irrlicht function and divide by screensize |
| 21:34 |
RealBadAngel |
inst simplier to take a few lines of code that i need than reverse some steps of it? |
| 21:35 |
xyz |
anyway i don't really understand the issue but i'd use a library function instead of copypasting it and hacking around |
| 21:35 |
RealBadAngel |
this library function is basically vector * matrix |
| 21:36 |
xyz |
i'm not saying what you did is wrong, i just don't know |
| 21:36 |
RealBadAngel |
i wouldnt call it any name, especially hacking |
| 21:37 |
RealBadAngel |
not to mention lots of mt code is just copy pasta from irrlicht examples ;) |
| 21:37 |
RealBadAngel |
with comments even :) |
| 21:38 |
|
salamanderrake_ joined #minetest-dev |
| 21:41 |
|
salamanderrake joined #minetest-dev |
| 21:48 |
|
rsiska joined #minetest-dev |
| 22:04 |
|
khonkhortisan joined #minetest-dev |
| 22:13 |
VanessaE |
[01-23 17:05] <cheapie> Remember the issue a long time ago with wraith's gravel road suddenly dropping itself as entities? |
| 22:13 |
VanessaE |
[01-23 17:05] <cheapie> I just had some gravel in my mine do the same thing. |
| 22:13 |
VanessaE |
(on my Survival server) |
| 22:22 |
|
proller joined #minetest-dev |
| 22:24 |
|
Miner_48er joined #minetest-dev |
| 22:44 |
ShadowNinja |
Adds pointed_thing to node:on_punch and monetest.register_on_punchnode callbacks (and fixes a typo with the last pointed_thing passing adition) http://ix.io/a5v |
| 22:44 |
ShadowNinja |
s/mone/mine/ |
| 22:45 |
ShadowNinja |
I also added copy_pointed_thing() which means that it actually removes 38 lines and adds 28. |
| 22:47 |
ShadowNinja |
Also, the PenLight Lua library is very usefull, maybe we could load it by default somehow... (It adds string:split/startswith/endswith/etc and a bunch of other usefull functions like deepcopy, although some things aren't needed, like the argument parsers) |
| 22:49 |
sapier |
hmm I guess the license could be compatible but I'm not exactly sure about it |
| 22:51 |
sapier |
did you try if it works without metatable access? |
| 22:52 |
ShadowNinja |
Some things will, but things like the class library obviously won't. |
| 22:53 |
sapier |
hmm guess it should be possible to remove the code not working without metatable |
| 22:54 |
ShadowNinja |
But why would we remove get/setmetatable? |
| 22:54 |
sapier |
why remove? .. I think about confusion caused if different environments support different basic features only ... we can't avoid this for minetest.* features but how to explain to user why "startswith" works in scriptapi but not in client side lua |
| 22:54 |
sapier |
because you can't build a client env with metatable access enabled |
| 22:55 |
sapier |
as far as I know you can't stop a lua script to break out of sandbox once it has metatable access |
| 22:57 |
|
smoke_fumus joined #minetest-dev |
| 22:57 |
ShadowNinja |
It can get the metatable for strings, but it sounds like you're talking about Lua sandboxes like the one used for the LuaController. |
| 22:58 |
ShadowNinja |
I'm not aware of anything bad that you can do with it after reading the lua-users wiki page on sandboxes. |
| 22:58 |
sapier |
I don't know the lua controler but it's way mor critical on client side. If we do a error in there we open users client to execute arbitrary downloaded code |
| 22:59 |
ShadowNinja |
Yes, but you can't do anything bad with metatables. |
| 22:59 |
ShadowNinja |
The only thing mentioned is modifying the global string metatable, but that's not relevant in a C sandbox. |
| 23:00 |
sapier |
the question not answered there is in what way it's broken |
| 23:00 |
sapier |
in best case it causes a crash only, in worst case it's usable to inject code |
| 23:00 |
sapier |
but I can't tell from those information provided |
| 23:01 |
ShadowNinja |
You can inject unsandboxed code with get/setmetatable? |
| 23:01 |
sapier |
I don't know |
| 23:02 |
ShadowNinja |
You can if it's a Lua sandbox, but not with a C sandbox, as the C doesn't run string:* unsandboxed. |
| 23:03 |
sapier |
not sure if you can cause something like a buffer overflow by breaking those metatables |
| 23:04 |
ShadowNinja |
Lua doesn't have buffer overflows silly. ;-) |
| 23:04 |
sapier |
not sure if this is possible but I could imagine you could manipulate the metatable to access os.execute() instead of sting.gsub |
| 23:05 |
sapier |
I meant a buffer overflow in c part ... the data is still stored in memory where c code is too |
| 23:06 |
sapier |
ok not exactly the lua variables should be stored at heap not at stack |
| 23:07 |
sapier |
but the only thing I wanted to say is we need to be very very carefull. Usually if you want to be secure in case of doubt better not allow something. |
| 23:11 |
ShadowNinja |
No, you can't access, eg, os.execute in any string methods. |
| 23:11 |
ShadowNinja |
And you can do getmetatable("").__metatable = false to prevent manipulation of the metatable. |
| 23:12 |
sapier |
as far as I understood lua variables are internaly referenced by stack positions. I haven't tied this but I'd not be surprised you can access any element of stack if you manage to find out it's stack position |
| 23:13 |
sapier |
actually we have to do this if you want to allow get/setmetatable |
| 23:14 |
sapier |
what about rawget/set there doesn't seem to be a way to make those safe? |
| 23:14 |
ShadowNinja |
The C code uses stack indexes, but Lua has no raw access to the stack. |
| 23:15 |
sapier |
not even debug? |
| 23:17 |
sapier |
ok it's late, we don't have client side lua yet so I guess best thin we can do is be very carefully on reviewing any future implementation of it |
| 23:39 |
|
kaeza joined #minetest-dev |
