Time |
Nick |
Message |
00:27 |
hmmmm |
no idea... i never worked with formspec so i don't know what it needs or doesn't need.. you should ask somebody who does that |
00:27 |
hmmmm |
[11:53 AM] * Calinou disagrees with debug.txt generation being enabled again <---- celeron said to make it 2 |
00:28 |
hmmmm |
and @ vanessa, 1 is the necessary level (errorstream/dstream) for mod loading failures et cetera, even though you can still see it in the console itself |
00:33 |
ShadowNinja |
formspec_escape would basicaly just do three gsubs(find and replace) on the string provided |
00:42 |
VanessaE |
hmmmm: right, though in my case there's no console to see in practice (e.g. if a mod crashes some time after the server has been started and backgrounded) |
00:42 |
hmmmm |
why don't you just use screen? |
00:42 |
VanessaE |
I do, actually. |
00:42 |
hmmmm |
then wut |
00:42 |
VanessaE |
I'm just too lazy to look up the proper way to attach to it :) |
00:42 |
hmmmm |
/ |
00:43 |
VanessaE |
it's been no less than 10 years since I last used screen, prior to a couple weeks ago |
00:43 |
hmmmm |
type 'screen', run your program, then ctrl+a+d |
00:43 |
hmmmm |
then when you want to reattach, screen -r |
00:43 |
VanessaE |
ah, that's right.. -r |
00:43 |
VanessaE |
anyways I start the server at boot time from rc.local |
00:43 |
proller |
runned at arm at 50-60 fps, but without picture 8) |
00:44 |
VanessaE |
proller: the phrase "without picture" and the term fps" do not belong in the same sentence :-) |
00:44 |
VanessaE |
my cadillac can do 150 mph...in the driveway :) |
00:44 |
proller |
with sound! |
00:45 |
proller |
i can walk |
00:46 |
hmmmm |
doesn't matter, the draw took no time at all because it failed |
00:47 |
proller |
segfault was from missing #define IRR_OGLES2_SHADER_PATH "../../media/Shaders/" |
00:58 |
proller |
something wrong in COGLES2Driver::addRenderTargetTexture |
01:06 |
proller |
and finish: addRenderTargetTexture in examples/13.RenderToTexture doesnt works in gles2, same black screen as in minetest with GL_INVALID_OPERATION in console |
01:08 |
proller |
will test on slooowberry pi |
03:54 |
|
ShadowNinja joined #minetest-dev |
06:00 |
|
ffoxin joined #minetest-dev |
06:05 |
hmmmm |
i feel like weblate is a bit too commit-message-spammy |
06:06 |
hmmmm |
is there any way to squash all the different languages into one commit |
06:11 |
thexyz |
only manually |
06:20 |
hmmmm |
erm.. in what version exactly were the 3d player models added? |
07:05 |
hmmmm |
anyway if you guys could fill this out more http://dev.minetest.net/Changelog that would be great. please keep things orderly and categorized like i have it. i didn't add in any bugfixes, nor did i add in anything lua-related |
07:09 |
hmmmm |
you can judge if something is a "big feature" or a "smaller addition and/or change" if the change took 1000 lines of code or 10 lines of code |
09:32 |
|
iqualfragile joined #minetest-dev |
10:20 |
|
rubenwardy joined #minetest-dev |
10:23 |
rubenwardy |
hi all |
11:01 |
|
Jordach joined #minetest-dev |
11:01 |
|
Jordach joined #minetest-dev |
11:04 |
|
jin_xi joined #minetest-dev |
11:16 |
|
PilzAdam joined #minetest-dev |
13:18 |
|
troller joined #minetest-dev |
13:35 |
|
darkrose joined #minetest-dev |
14:17 |
PilzAdam |
thexyz, can I get a @minetest.net email address? |
14:17 |
thexyz |
sure |
14:17 |
thexyz |
pilzadamminetest.net? |
14:17 |
PilzAdam |
yep |
15:25 |
|
rubenwardy joined #minetest-dev |
15:34 |
|
hmmmm joined #minetest-dev |
15:51 |
|
proller joined #minetest-dev |
15:52 |
|
Calinou joined #minetest-dev |
15:58 |
|
rubenwardy joined #minetest-dev |
16:46 |
proller |
build with egl - pull or not ? https://github.com/proller/minetest/commit/abd2f4575cae86307e2f2171b28af2d4ddecd40a |
16:49 |
hmmmm |
not now |
16:52 |
|
iqualfragile joined #minetest-dev |
16:57 |
RealBadAngel |
hmmm, i found by accident a nice piece of code |
16:57 |
RealBadAngel |
sound played local only, no cache |
16:57 |
RealBadAngel |
and it works |
16:57 |
RealBadAngel |
dunno who made it |
16:58 |
RealBadAngel |
https://github.com/minetest/minetest/blob/master/src/game.cpp#L847 |
16:58 |
RealBadAngel |
i modified my technic ambience mod to play music without cache |
16:59 |
RealBadAngel |
its clearly a way to make all sounds to be played by clients |
17:00 |
RealBadAngel |
i mean those ambience ones, not player interactions |
18:19 |
|
ffoxin_entry joined #minetest-dev |
18:36 |
thexyz |
hmmmm: why not? |
18:36 |
thexyz |
also, what stops us from releasing 0.4.5 right now? |
18:46 |
hmmmm |
because we said it was a week... |
18:46 |
hmmmm |
also, people _have_ been testing, right? |
18:46 |
RealBadAngel |
i say 10 random bugs that not being fixed, hows so? ;) |
18:46 |
hmmmm |
what bugs |
18:47 |
RealBadAngel |
just jokin, the time was taken to eventually fix some bugs |
18:47 |
RealBadAngel |
but i could point a few old ones if needed :) |
18:47 |
|
dimeshake joined #minetest-dev |
18:48 |
hmmmm |
ehh that's okay. |
18:48 |
hmmmm |
we just want to make sure that the new stuff we put in works |
18:48 |
RealBadAngel |
yeah |
18:48 |
hmmmm |
monday is going to be a shitstorm |
18:48 |
hmmmm |
we make the "Bump version to 0.4.5" commit and then compile for everything and release... who is in charge of the releasing? |
18:48 |
RealBadAngel |
personally i will remove two bugs soon |
18:49 |
RealBadAngel |
one with too many vertices/indices |
18:49 |
thexyz |
we only compile for win32, don't we? |
18:49 |
RealBadAngel |
and second with nodeboxes coords need to be sorted |
18:50 |
hmmmm |
well what about the ubuntu ppa |
18:50 |
thexyz |
any idea who has access to it? |
18:50 |
VanessaE |
forget the ppa, we need to get rid of 0.3.x from the ubuntu upstream repos |
18:50 |
Jordach |
hmmmm, yeah, there is going to be a *MEGA* shitstorm |
18:51 |
thexyz |
VanessaE: you mean, debian? |
18:52 |
VanessaE |
thexyz: any of the debian-based distros; I don't know which packages the Ubuntu team gets from Debian vs. direct from the various upstream projects |
18:52 |
thexyz |
we probably should just contact maintainer http://packages.debian.org/unstable/main/minetest |
18:52 |
hmmmm |
and i need to email the freebsd port maintainer too about it. proller did that once before, he had to email him some sort of patch if i remember correctly? what's with that? we don't need any patches at all, it works out of the box |
18:52 |
hmmmm |
also what about the mac os x |
18:52 |
thexyz |
hm.. |
18:53 |
thexyz |
well, I can try to set up virtual machine |
18:53 |
darkrose |
jnumm does ppa stuff, according to my logs |
18:53 |
hmmmm |
i have a mac os x 10.8 virtual machine set up already |
18:53 |
hmmmm |
wonder if i should do that at least? |
18:54 |
* Calinou |
joins thexyz in the "update a debian package" war |
18:54 |
RealBadAngel |
btw, the bugs are reality, we know many. i suggest to make a similar to TODO list naming them and makin folks responsible to remove them |
18:54 |
RealBadAngel |
ie split tasks |
18:54 |
hmmmm |
the bug tracker on github is full of crap though |
18:54 |
thexyz |
hmmmm: i think that's not really needed to be done before 0.4.5 release |
18:54 |
VanessaE |
I think assigning bugs to various individuals might be a bad idea, but more attention is certainly needed in a few places for sure |
18:54 |
RealBadAngel |
not really |
18:54 |
VanessaE |
both on the engine and the game |
18:54 |
thexyz |
it looks more like "feature requests" tracker |
18:55 |
hmmmm |
yeah for real ^ |
18:55 |
RealBadAngel |
"bug hunter" |
18:55 |
thexyz |
plz add that |
18:55 |
thexyz |
no that |
18:55 |
hmmmm |
i honestly don't have much patience to go around bug tracking, especially when there's so much more stuff to do |
18:56 |
hmmmm |
like if i come across a bug and it's easy to fix or it's related to what i'm doing, i'll fix it, but.. |
18:56 |
Calinou |
an "issue" isn't only for bugs, it's for feature requests too |
18:56 |
Calinou |
there are labels, but only people who can push stuff to upstream can put these |
18:57 |
hmmmm |
i kinda wish i can freeze time for a week or something and just focus all time and effort on minetest to make a real dent in the todo list |
18:57 |
VanessaE |
hmmmm: well to that end, c55 did say on the blog this was supposed to go for a *month* |
18:57 |
VanessaE |
not a week |
18:58 |
hmmmm |
?? |
18:58 |
hmmmm |
the freeze? |
18:58 |
VanessaE |
*nod*\ |
18:58 |
hmmmm |
that, my friend, is ridiculous and unnecessary |
18:58 |
VanessaE |
"Many bugs have been fixed since 0.4.4 and I would have hoped for 0.4.5 to be released a month or two ago, but we have ended up on somewhat a rolling release model with mods already using features from unreleased versions and many people making unofficial Windows builds. We’ll try to make a proper release in a month or so, as/if a more suitable time is found" |
18:58 |
hmmmm |
we're already getting crunched up by not committing things for a week |
18:59 |
VanessaE |
I might tend to agree actually - a month is more time than we need to push 0.4.5, but a week might not be enough. |
18:59 |
Jordach |
we should cut the bullshit and release a new version every two months |
18:59 |
Jordach |
then windows users are at least in some form of sync with latest |
19:00 |
VanessaE |
Jordach: unofficial builds took care of that ages ago. |
19:00 |
Jordach |
VanessaE, yes, but people who think stable is stable and not going to break. |
19:00 |
RealBadAngel |
so stop that model |
19:00 |
darkrose |
unnoficial builds don't officially exist |
19:01 |
Jordach |
even off. every month |
19:01 |
VanessaE |
darkrose: perhaps, but at least it isn't "they officially don't exist" :) |
19:01 |
RealBadAngel |
im not releasing 6d facedir because i realize many mods will have to be rewritten |
19:01 |
VanessaE |
RealBadAngel: well, they'll have to be *to use that feature* |
19:01 |
RealBadAngel |
that can be the time between 0.4.5 and next stable |
19:01 |
VanessaE |
but it's not like you're planning to break it |
19:02 |
RealBadAngel |
no, not break, but to use they gonna need time |
19:02 |
RealBadAngel |
you and me too |
19:02 |
VanessaE |
yeah |
19:02 |
RealBadAngel |
also calinou and jeija |
19:02 |
VanessaE |
but this is offtopic |
19:02 |
|
dimeshake joined #minetest-dev |
19:05 |
hmmmm |
now hold on a minute |
19:05 |
hmmmm |
"stable" simply means you have most of the bugs ironed out and you're aware of most of the issues |
19:06 |
VanessaE |
that's how I always figured it |
19:06 |
hmmmm |
realbadangel, how will the 6d facedir thing break mods? |
19:06 |
VanessaE |
hmmmm: it won't. |
19:06 |
RealBadAngel |
no |
19:06 |
hmmmm |
then wtf? |
19:06 |
Jordach |
it only breaks mods using the current facedir system |
19:06 |
VanessaE |
he's referring to the ability to rotate things in all directions arbitrarily, mods will have to be rewritten to support that feature |
19:06 |
RealBadAngel |
but will force complete rewrite of them for benefits |
19:07 |
VanessaE |
think of placing a pipe and having it bend in whatever direction it needs to without requiring 6 differnet models. |
19:07 |
VanessaE |
(for each shape) |
19:07 |
RealBadAngel |
cut number of needed node definitions for 1/4th to 1/2 |
19:07 |
VanessaE |
Jordach: no, it doesn't. |
19:07 |
hmmmm |
well |
19:07 |
hmmmm |
my take on the mod breakage thing is to go ahead with it, and make the breakage well known |
19:07 |
RealBadAngel |
also for example vertical mesecons wont be needed anymore |
19:07 |
hmmmm |
give people some time to fix it before the new release |
19:08 |
hmmmm |
if a big mod is broken by it, fix it yourself |
19:08 |
RealBadAngel |
instead wires will be able to climb walls and hang on ceilings |
19:08 |
VanessaE |
hmmmm: agreed - as long as it is predictable, publicized, and at least is easy enough to fix |
19:08 |
Exio |
why is not the 0.4.5 "already released" exactly? because testers? who is doing some "testing" in the release in that case? |
19:09 |
RealBadAngel |
we are not talking here about breaking, but bringing a revolution to them |
19:09 |
thexyz |
of course it can break mods |
19:09 |
hmmmm |
people using the unofficial win32 builds, and everybody here basically |
19:09 |
VanessaE |
Exio: because there are still at least a few bugs that need to be worked out, or at the very least, tested to the point that we can be sure nothing's wrong |
19:09 |
thexyz |
what if i use facedir += 1 to rotate clockwise? that'd be broken |
19:09 |
RealBadAngel |
any old code unaware of the changes will work as before |
19:09 |
hmmmm |
was that ever documented? |
19:10 |
VanessaE |
thexyz: sure, but that's an example of something that's easy to fix |
19:10 |
RealBadAngel |
y+ rotation case is just the old facedir |
19:10 |
VanessaE |
plus you know as well as anyone never to rely on undocumented behavior. |
19:10 |
hmmmm |
that incrementing facedir shall rotate it clockwise |
19:10 |
hmmmm |
because if it wasn't documented, then it's their fault, so fuckem |
19:10 |
RealBadAngel |
stop please now |
19:10 |
RealBadAngel |
original one is counter clock wise |
19:11 |
RealBadAngel |
0,-90, 180, 90 |
19:11 |
darkrose |
does it break _game? if not it's not our problem |
19:11 |
thexyz |
hmmmm: why? |
19:11 |
hmmmm |
nothing is broken |
19:11 |
hmmmm |
thexyz, because we never said so and they're assuming |
19:11 |
RealBadAngel |
i kept the pattern |
19:11 |
thexyz |
hmmmm: there are documented angles/directions |
19:11 |
hmmmm |
RealBadAngel, so it's not even a problem at all? |
19:11 |
RealBadAngel |
no |
19:12 |
RealBadAngel |
not even a smallest one |
19:12 |
thexyz |
how else would you use facedir without knowing how it rotates |
19:12 |
hmmmm |
thexyz, so it is documented that if you increment facedir, it'll rotate coutnerclockwise? |
19:12 |
VanessaE |
hmmmm: thexyz's corner case is the only one I can think of that would break. |
19:12 |
VanessaE |
and no one with an ounce of brains would do that anyway :) |
19:12 |
RealBadAngel |
new facedir is in fact two values in one |
19:12 |
thexyz |
oh, i see, facedir isn't documented at all |
19:12 |
thexyz |
well, fine then |
19:12 |
RealBadAngel |
it is |
19:13 |
RealBadAngel |
you dont read forums, dont you? ;) |
19:13 |
hmmmm |
they should be using minetest.dir_to_facedir() from what i just found in lua_api.txt |
19:13 |
thexyz |
yeah |
19:13 |
VanessaE |
come to think of it, it *isn't* properly documented, is it. |
19:13 |
hmmmm |
RBA just said he made sure that works |
19:13 |
RealBadAngel |
this function will be unaffected |
19:13 |
hmmmm |
but my point is that even if it didn't work, it's not our problem |
19:13 |
VanessaE |
at least, the value range isn't |
19:13 |
RealBadAngel |
old facedir cases will be used as prediction for placing |
19:14 |
RealBadAngel |
because old facedir is a subset of new one |
19:14 |
RealBadAngel |
new facedir means the rotation around the axis |
19:14 |
RealBadAngel |
with top face facing it |
19:15 |
thexyz |
fine, fine |
19:15 |
RealBadAngel |
as the old one was |
19:15 |
thexyz |
that was the case i just made up |
19:15 |
RealBadAngel |
top facing y++ and clock counter wise rotations around it |
19:16 |
thexyz |
so, is release planned on monday? |
19:16 |
hmmmm |
yes |
19:16 |
VanessaE |
wait wait |
19:16 |
VanessaE |
that isn't consistent with what c55 indicated |
19:16 |
VanessaE |
(at least in the blog) |
19:17 |
thexyz |
hm? |
19:17 |
hmmmm |
what, in that post? so? |
19:17 |
hmmmm |
i don't want a rolling release |
19:17 |
VanessaE |
nonono |
19:17 |
VanessaE |
the "a months or so" part. |
19:17 |
VanessaE |
-s |
19:17 |
thexyz |
>We’ll try to make a proper release in a month or so, as/if a more suitable time is found. |
19:17 |
thexyz |
or so |
19:17 |
hmmmm |
he just said "we ended up on a somewhat" |
19:17 |
thexyz |
why not? |
19:17 |
RealBadAngel |
as far as i remember c55 was lately forcing freeze and release asap |
19:17 |
VanessaE |
hrm |
19:17 |
RealBadAngel |
not waiting |
19:18 |
thexyz |
release is the only thing which stops us now from adding New And Awesome features |
19:18 |
hmmmm |
on monday here's what i want to happen: |
19:18 |
thexyz |
why should we delay it? seems everything works fine |
19:18 |
thexyz |
except macs, they suck |
19:18 |
hmmmm |
we slap the 0.4.5 label on it, and official builds are compiled from this commit |
19:18 |
RealBadAngel |
and was disguisted by that we wanted to add *this and *that before and gave up |
19:18 |
thexyz |
hmmmm: win32? |
19:18 |
hmmmm |
we can continue on committing things after that |
19:19 |
hmmmm |
yeah |
19:19 |
hmmmm |
win32 |
19:19 |
proller |
hmmmm, no problem with freebsd port, i can make update and submit pr or i know where maintainer in irc |
19:19 |
VanessaE |
mmmh |
19:19 |
thexyz |
i can do that |
19:19 |
hmmmm |
just curious, for the official win32 builds, do you use MSVC or mingw? |
19:19 |
thexyz |
msvc2010 |
19:19 |
hmmmm |
proller, thanks |
19:19 |
thexyz |
like celeron55 did |
19:19 |
hmmmm |
okay |
19:19 |
VanessaE |
the issue I see is if we go ahead and call it 0.4.5, are we prepared to deal with the backlash of "wtf? This bug has existed since version x.y.z and still isn't fixed??" and the like |
19:20 |
hmmmm |
we can make releases as often as we'd like |
19:20 |
hmmmm |
besides, that'd happen with any version |
19:20 |
VanessaE |
sure, of course |
19:20 |
RealBadAngel |
i really see sense in makin "bug hunter" list |
19:20 |
RealBadAngel |
not by users, but by devs |
19:20 |
VanessaE |
my point is, pick some random, old, irritating bug and fix that *before* 0.4.5 |
19:21 |
VanessaE |
like that one where you have to jiggle the mouse after using a formspec |
19:21 |
hmmmm |
alright let's see, issues with the bug label |
19:22 |
VanessaE |
(or anything really, that is visible to most users) |
19:23 |
hmmmm |
but wait a minute |
19:23 |
hmmmm |
we aren't doing anything like this so late in the game |
19:23 |
hmmmm |
seriously it's not too big of an issue, the crap that's in 0.4.5 is enough to make everybody love it |
19:23 |
thexyz |
yeah |
19:23 |
thexyz |
so, changelog |
19:24 |
hmmmm |
besides |
19:24 |
RealBadAngel |
indeed |
19:24 |
Exio |
hmmmm: didn't you fix that emergethread super-big bug? |
19:24 |
thexyz |
we also need some way to post update at minetest.net |
19:24 |
thexyz |
(I don't have access to it) |
19:24 |
hmmmm |
maybe we can have 0.4.x where x is even bugfix releases or whatever |
19:24 |
hmmmm |
i don't know |
19:24 |
hmmmm |
no that'll screw up our development model |
19:24 |
hmmmm |
there's NO NEED TO CHANGE WHAT ISN'T BROKEN |
19:24 |
hmmmm |
Exio, yeah i did |
19:24 |
RealBadAngel |
but we have to remove the bugs in the future |
19:24 |
Exio |
there, a big bug solved for 0.4.5 :) |
19:25 |
RealBadAngel |
issues lists can say us whats broken, bug hunter list can say us who will fix them |
19:26 |
RealBadAngel |
because we all know who is good at some point and being able to fix it |
19:27 |
RealBadAngel |
for example somethin is broken within mapgen -> hmmm |
19:27 |
RealBadAngel |
thats what im talkin about |
19:28 |
proller |
idea for future - do not crash server on lua error |
19:28 |
RealBadAngel |
in my area of interest i do have already 2 known bugs and can fix them |
19:28 |
thexyz |
RealBadAngel: github issue tracker has tags |
19:28 |
thexyz |
we can tag bug as [mapgen] |
19:28 |
thexyz |
or [formspec] |
19:28 |
thexyz |
or something |
19:29 |
RealBadAngel |
tagging and pointing a person whos gonna fix it, those are two different things :) |
19:29 |
VanessaE |
proller: no, bad idea. |
19:29 |
VanessaE |
don't take control of such things out of the system admin's hands. |
19:30 |
proller |
VanessaE, i setup 10 mods, and 2-3 crush server sometimes |
19:30 |
RealBadAngel |
proller: use lua inside lua |
19:30 |
PilzAdam |
I dont like the idea to assign people for fixing bugs, because random devs may think "Oh, someone already fixes that, so I dont have to" |
19:30 |
VanessaE |
proller: then remove them until they can be fixed, of course. |
19:31 |
RealBadAngel |
PilzAdam, bugs are the code coders dont like |
19:31 |
proller |
VanessaE, it can happens after week of server work |
19:31 |
RealBadAngel |
they often would go for new one than solving some mysterious problems |
19:31 |
VanessaE |
yup, it can'. |
19:31 |
VanessaE |
can. my point stands. |
19:31 |
proller |
better to remove all mods |
19:32 |
thexyz |
PilzAdam: instead, >Oh, nobody is fixing it, let me do it... and then we have several people working on the same issue |
19:32 |
VanessaE |
again this is offtopi |
19:32 |
VanessaE |
c |
19:33 |
PilzAdam |
btw: +1 for release 0.4.5 on monday |
19:33 |
thexyz |
well, it's not like we get a ton of bugfixes from random people |
19:33 |
RealBadAngel |
if several are workin on the very same issue chance is it will be fixed sooner |
19:33 |
RealBadAngel |
which is good |
19:33 |
RealBadAngel |
also knowledge gained by the others is also a value |
19:33 |
thexyz |
without knowing about each other? |
19:34 |
thexyz |
how'd that help? |
19:34 |
RealBadAngel |
if you see a name on the list |
19:34 |
RealBadAngel |
and think you can help. contact him |
19:34 |
RealBadAngel |
simple as that |
19:34 |
thexyz |
oh, that's what you were talking about |
19:35 |
RealBadAngel |
its not a competition, this is meant to be helpful |
19:35 |
thexyz |
i also wonder what shall we do with font issue |
19:36 |
VanessaE |
in what way? |
19:36 |
RealBadAngel |
for example, idk what youre talkin about. just check the list :) |
19:37 |
thexyz |
apparently, it doesn't have CJK characters; with gettext enabled by default chinese/japanese/korean users would not see most translated labels |
19:37 |
thexyz |
as they have to download some other font |
19:37 |
hmmmm |
ugh, speaking of mapgen, i really hate that celeron made vmanip a reference throughout the code instead of a pointer |
19:37 |
RealBadAngel |
hmmmm, same was with atlas stuff |
19:37 |
hmmmm |
i am torn between making vmanip a parameter of the generic mapgen functions and making it a member of the virtual class |
19:38 |
thexyz |
my idea is to add "alternative" font and load it for some languages |
19:38 |
hmmmm |
right, we should make that a development guideline, avoid references when possible, try to minimize the C++ crap as much as possible |
19:38 |
RealBadAngel |
at some point to prevent corruption he made const in definitions |
19:38 |
RealBadAngel |
which meant door shut for me |
19:38 |
hmmmm |
he really went overboard with the ++y |
19:39 |
hmmmm |
and then the GUI things used the MVC paradigm which is ridiculous - keep that for enterprise java applications, not minetest |
19:39 |
thexyz |
oh, also, there is my STL branch |
19:39 |
thexyz |
i suggest to switch to STL after 0.4.5 release |
19:39 |
hmmmm |
well we already have people use stl for all new code |
19:39 |
VanessaE |
what is the benefit from STL? I'm still unclear on this |
19:40 |
hmmmm |
consistency |
19:40 |
Exio |
what STL is? |
19:40 |
hmmmm |
everything you see in C++ code that starts with "std::" |
19:40 |
rubenwardy |
http://en.wikipedia.org/wiki/STL_%28file_format%29 |
19:40 |
thexyz |
rubenwardy: almost! |
19:40 |
rubenwardy |
http://en.wikipedia.org/wiki/Standard_Template_Library |
19:40 |
Exio |
ah |
19:41 |
thexyz |
any idea how to get 2-letter locale code from gettext? |
19:50 |
hmmmm |
no |
20:02 |
rubenwardy |
celeron55: there is $3.04 on the cafepress account |
20:03 |
rubenwardy |
:D |
20:03 |
rubenwardy |
not much, but it is a start :D |
20:03 |
thexyz |
wow! |
20:04 |
thexyz |
we're rich now |
20:04 |
rubenwardy |
http://www.cafepress.com/sk/minetest |
20:04 |
hmmmm |
$3.04 in zimbabue dollars |
20:05 |
rubenwardy |
and all that money is from mark ######'s purchase |
20:05 |
rubenwardy |
the unannoying one, that is a moderator |
20:05 |
thexyz |
hm? |
20:05 |
thexyz |
marktraceur |
20:06 |
rubenwardy |
thats it |
20:06 |
thexyz |
he left a while ago though |
20:06 |
hmmmm |
he probably didn't want you to say it for some reason if he ########ed it |
20:06 |
rubenwardy |
I think that was him |
20:06 |
rubenwardy |
hmmmm: I forgot his name |
20:06 |
hmmmm |
oh |
20:06 |
hmmmm |
his actual name is Mark Holmquest or something |
20:06 |
hmmmm |
if you look on github |
20:09 |
RealBadAngel |
can i offtopic with a propably nice feature? |
20:09 |
hmmmm |
what's with all the topic crap |
20:09 |
hmmmm |
don't ask to ask, just ask |
20:10 |
hmmmm |
it's like people are overly concerned with being "on topic" when nobody is ever "on topic" |
20:10 |
RealBadAngel |
https://github.com/minetest/minetest/blob/master/src/game.cpp#L847 |
20:10 |
RealBadAngel |
what can you see here? |
20:10 |
hmmmm |
the background music |
20:10 |
hmmmm |
it looks like being able to play local sounds with no caching |
20:11 |
RealBadAngel |
it is |
20:11 |
RealBadAngel |
tested |
20:11 |
PilzAdam |
what about moving it to client/sounds/ and call it a features instead of testing code? |
20:11 |
hmmmm |
so you want us to add background music? |
20:11 |
RealBadAngel |
client side sounds |
20:11 |
RealBadAngel |
defined in minetest_game |
20:11 |
RealBadAngel |
and stored locally |
20:12 |
RealBadAngel |
ie Ambience mod |
20:12 |
hmmmm |
if it's polished up and turned into an actual feature i guess that's cool |
20:12 |
RealBadAngel |
im using Ambience for months already |
20:12 |
hmmmm |
is it just me or is it irritating how a lot of the class names are "-er" |
20:13 |
RealBadAngel |
http://realbadangel.pl/technicambience.zip |
20:13 |
RealBadAngel |
hmmm, check this mod out |
20:13 |
hmmmm |
can't, minetest doesn't compile for me at the moment |
20:14 |
RealBadAngel |
i changed the music files names to follow the code rules |
20:14 |
RealBadAngel |
and now music is played local for me |
20:15 |
RealBadAngel |
notice one thing |
20:15 |
RealBadAngel |
if you will stream to ogg |
20:15 |
RealBadAngel |
you will be able to hear to internet radio in minetest too lol |
20:16 |
* hmmmm |
just found an embarassing bug in noise.h |
20:16 |
PilzAdam |
hows the progress on the built-in web browser in Minetest? |
20:16 |
rubenwardy |
lol "hmmmm" in bold make it look blurred |
20:17 |
VanessaE |
web browser in minetest?? |
20:17 |
VanessaE |
please no |
20:17 |
VanessaE |
feature creep. |
20:17 |
RealBadAngel |
browser? i got CP/M running there :P |
20:17 |
VanessaE |
just call on the system default browser |
20:18 |
PilzAdam |
VanessaE, maybe you can see it as a (ironic) comment to RealBadAngel's internet radio? |
20:18 |
VanessaE |
now I do :) |
20:18 |
VanessaE |
however, |
20:18 |
VanessaE |
streaming audio has practical applications |
20:18 |
RealBadAngel |
Z80 emulation library plus memory and video i wrote |
20:18 |
RealBadAngel |
i used Amstrad 464 model |
20:19 |
RealBadAngel |
and CP/M is almost booting |
20:19 |
RealBadAngel |
just a few things to solve |
20:20 |
hmmmm |
pfoo |
20:20 |
rubenwardy |
run minetest through steam. Instant ingame webbrowser |
20:20 |
rubenwardy |
just need a ingame irc client |
20:20 |
rubenwardy |
lol |
20:20 |
RealBadAngel |
we do have it already |
20:21 |
RealBadAngel |
kaeza wrote it |
20:22 |
VanessaE |
why not just a generic internet library |
20:22 |
VanessaE |
something that the IRC mod could hook into, as well as the streaming audio, etc. |
20:22 |
RealBadAngel |
maybe tcp/ip first? |
20:22 |
hmmmm |
http://pastebin.com/CuvqjnkB |
20:22 |
hmmmm |
god damn C++ |
20:22 |
hmmmm |
line 22 makes that error |
20:22 |
RealBadAngel |
why we do suffer from udp all the time? |
20:22 |
VanessaE |
hmmmm: it's still better than the Lua version |
20:23 |
VanessaE |
almost looks like a straight-across port? |
20:23 |
rubenwardy |
RBA: yeah i know, but it require the mod to me installed on the server |
20:23 |
rubenwardy |
i use it in single player |
20:23 |
hmmmm |
Lua doesn't generate those sorts of errors |
20:23 |
VanessaE |
oh there's the error. |
20:23 |
hmmmm |
a single missing ) can cause THAT |
20:24 |
VanessaE |
well to be fair |
20:24 |
VanessaE |
Lua loves to break like crazy for lack of a simple comma. |
20:25 |
VanessaE |
I mean 500 lines later, the error shows up |
20:25 |
VanessaE |
(well, not as much as that, but still) |
20:25 |
hmmmm |
and yeah, it is just a straightforward port from the Lua. that's really the only sane strategy for ore generation. |
20:27 |
hmmmm |
if you notice, i smoothed out the vein difficulties - it's a vein-type ore if there's a NoiseParams present. then it does a single point noise calculation to check if it's in the vein or not |
20:29 |
hmmmm |
as for having the ore scattered in certain areas more than others, you can do that by modulating ore->clust_scarcity before calling generateOre() |
20:29 |
hmmmm |
either randomly, say by blockseed, or with an overall pattern, like another point of perlin noise |
20:30 |
hmmmm |
this sort of thing can introduce really subtle but interesting nuances to the presense of ores |
20:30 |
hmmmm |
presence* |
20:33 |
RealBadAngel |
rubenwardy, if the mod was part of minetest_game you could use it everywhere |
20:34 |
rubenwardy |
yes, but not all server owners would want it |
20:38 |
RealBadAngel |
like why? |
20:38 |
RealBadAngel |
if game would have predefined client sounds |
20:39 |
RealBadAngel |
only for client, build coming with the default sounds |
20:39 |
RealBadAngel |
what the damn admin would say about it? |
20:40 |
RealBadAngel |
it would be transparent for him |
20:40 |
RealBadAngel |
and player could be even able to mod this sounds |
20:40 |
rubenwardy |
because everything anybody says would be broadcasted into the #minetest room |
20:40 |
RealBadAngel |
with a sound pack |
20:41 |
RealBadAngel |
aaaaaa |
20:41 |
RealBadAngel |
lol |
20:41 |
rubenwardy |
wait a minute, you are talking about client sonds? |
20:41 |
rubenwardy |
sounds |
20:41 |
RealBadAngel |
hehehe indeed |
20:41 |
rubenwardy |
I though you were talking about the IRC mod |
20:41 |
RealBadAngel |
rotfl |
20:41 |
VanessaE |
seemed pretty obvious to me ;) |
20:41 |
rubenwardy |
<rubenwardy> just need a ingame irc client |
20:41 |
rubenwardy |
<rubenwardy> lol |
20:41 |
rubenwardy |
<RealBadAngel> we do have it already |
20:41 |
rubenwardy |
<RealBadAngel> kaeza wrote it |
20:42 |
rubenwardy |
<rubenwardy> RBA: yeah i know, but it require the mod to me installed on the server |
20:42 |
rubenwardy |
<rubenwardy> i use it in single player |
20:42 |
rubenwardy |
<RealBadAngel> rubenwardy, if the mod was part of minetest_game you could use it everywhere |
20:42 |
rubenwardy |
<rubenwardy> yes, but not all server owners would want it |
20:42 |
RealBadAngel |
stop the damn mixer lol |
20:42 |
rubenwardy |
and continue... |
20:42 |
RealBadAngel |
lets sort it out |
20:43 |
RealBadAngel |
sounds are easy case |
20:43 |
RealBadAngel |
solved |
20:43 |
rubenwardy |
sounds would be good |
20:43 |
* rubenwardy |
is thinking of making PotatOS |
20:43 |
RealBadAngel |
its just modyfying that function |
20:43 |
RealBadAngel |
rename default folder |
20:44 |
RealBadAngel |
and accept any unfetched files |
20:44 |
RealBadAngel |
not just 10 |
20:44 |
RealBadAngel |
and renaming the folder |
20:44 |
RealBadAngel |
then move ambience to minetest_game |
20:45 |
RealBadAngel |
about irc |
20:45 |
RealBadAngel |
do you recall my Portals idea? |
20:46 |
RealBadAngel |
servers connected to each other with structures player can step in |
20:47 |
RealBadAngel |
and reconnect to another server |
20:47 |
PilzAdam |
Jeija has worked on this IIRC |
20:47 |
RealBadAngel |
for this irc comms would be just great |
20:48 |
RealBadAngel |
talks over the different servers |
20:48 |
RealBadAngel |
like im in skyblock now, can you come and help me? |
20:48 |
rubenwardy |
gtg bye all |
20:49 |
RealBadAngel |
sure, i will finish minin there and go through portal |
20:50 |
RealBadAngel |
1) we gain effect of community bigger |
20:50 |
RealBadAngel |
because all servers CAN be connected and players can talk over them |
20:51 |
RealBadAngel |
2) we need to drop UDP for this |
20:51 |
Exio |
drop UDP? why the hell? |
20:51 |
Exio |
thanks to UDP i still can play |
20:52 |
VanessaE |
then we'd have to make sure both can be used |
20:52 |
RealBadAngel |
waiting several minutes to get 10 mbytes when regular connection at the same time can download whole movie is just INSANE |
20:53 |
RealBadAngel |
UDP is not meant for file transfer for christ sake |
20:54 |
Exio |
uh? |
20:54 |
Exio |
in that case, using TCP for download the files is ok |
20:54 |
RealBadAngel |
and CURL is just a workaround |
20:55 |
Exio |
but UDP for the entire game / "stuff" |
20:55 |
RealBadAngel |
for messages client/server is ok |
20:55 |
RealBadAngel |
but not for connecting |
20:55 |
thexyz |
curl is a good workaround |
20:55 |
thexyz |
why should minetest server serve static files? |
20:56 |
thexyz |
when nginx can do the same better |
20:56 |
RealBadAngel |
why would any user be forced to setup a http server? |
20:56 |
thexyz |
user is not forced to do that |
20:56 |
thexyz |
it's just an option |
20:56 |
thexyz |
a good one |
20:56 |
thexyz |
not for user |
20:56 |
VanessaE |
curl sucks, sorry |
20:56 |
thexyz |
for server administrators |
20:56 |
VanessaE |
this shit does not belong in a third-party application |
20:57 |
RealBadAngel |
sucks big time |
20:57 |
thexyz |
maybe it does, but how's that related? |
20:57 |
RealBadAngel |
i mean the solution |
20:57 |
thexyz |
? |
20:57 |
RealBadAngel |
its a workaround |
20:57 |
thexyz |
VanessaE: and why does it? |
20:57 |
RealBadAngel |
not a solution at all |
20:57 |
thexyz |
that's a solution |
20:57 |
thexyz |
many games use it |
20:57 |
VanessaE |
because a proper solution is to put the media server INTO the minetestserver |
20:57 |
thexyz |
like quake 3, openarena, etc |
20:57 |
thexyz |
VanessaE: no |
20:57 |
VanessaE |
one program, one purpose. |
20:57 |
thexyz |
there is media server in minetestserver |
20:58 |
thexyz |
why should we put in another one? |
20:58 |
thexyz |
I'd better call it "fallback" one |
20:58 |
RealBadAngel |
thexyz, i cant recall i was forced to configure any server to play quake with friends |
20:58 |
thexyz |
serving big static files is a task which can be delegated to another application |
20:58 |
thexyz |
one program, one purpose |
20:59 |
RealBadAngel |
now i do have to wait endlessly to download content |
20:59 |
thexyz |
RealBadAngel: because quake includes some packs by default |
20:59 |
RealBadAngel |
hold on a sec |
20:59 |
thexyz |
you may even serve files via amazon cdn |
20:59 |
thexyz |
because why not? |
20:59 |
RealBadAngel |
i used to play Heroes of Might and Magic 3 |
21:00 |
RealBadAngel |
20 yrs old game |
21:00 |
thexyz |
see, ioquake3 |
21:00 |
thexyz |
>Off-server data downloads (http redirection) via HTTP and FTP with cURL |
21:00 |
RealBadAngel |
this museal one was able to transfer files through tcp/ip |
21:00 |
Exio |
i still don't understand what is wrong with curl |
21:00 |
thexyz |
RealBadAngel: so? |
21:00 |
RealBadAngel |
youre tryin to tell me nowadays UDP is new standard? |
21:01 |
RealBadAngel |
and we have to stick to it? |
21:01 |
thexyz |
reread my messages |
21:01 |
thexyz |
I wasn't telling that |
21:01 |
thexyz |
wasn't trying to do either |
21:01 |
thexyz |
you just make a problem of nothing |
21:01 |
Exio |
RealBadAngel: you has a good connection with low latency and a good download, no? |
21:02 |
VanessaE |
his point and mine is that cURL shouldn't be a "fallback" option - something equivalent to it should replace the default UDP file transfer code. |
21:02 |
thexyz |
"equivalent" |
21:02 |
Exio |
rewriting code, eh? |
21:02 |
thexyz |
as i see, your point is to make a reliable, fast, multi-threaded static file server into minetest |
21:03 |
thexyz |
good point |
21:03 |
RealBadAngel |
transparent |
21:03 |
thexyz |
but nginx already exists |
21:03 |
VanessaE |
and nginx is *third party* |
21:03 |
VanessaE |
that's the problem |
21:03 |
RealBadAngel |
idk whats that even |
21:03 |
Exio |
VanessaE: what is the problem with that? |
21:03 |
VanessaE |
the same would be true if we were talking about apache or IIS or whetever else you wanna try |
21:03 |
thexyz |
VanessaE: what's the problem? |
21:03 |
PilzAdam |
sqlite is third party too, lets write our own database |
21:03 |
RealBadAngel |
thats why i call for transparency and speed :) |
21:04 |
VanessaE |
PilzAdam: but sqlite is directly used by the engine |
21:04 |
Exio |
when will MT-script be released? |
21:04 |
VanessaE |
that's my point! |
21:04 |
thexyz |
VanessaE: ? |
21:04 |
thexyz |
so |
21:04 |
thexyz |
you'd like >reliable, fast, multi-threaded static file server into minetest |
21:04 |
thexyz |
right? |
21:04 |
thexyz |
the one which serves files via UDP is a good fallback one |
21:05 |
RealBadAngel |
if anythin other fails |
21:05 |
VanessaE |
then yeah |
21:05 |
VanessaE |
if you wanna put it that way |
21:05 |
thexyz |
but why? |
21:05 |
thexyz |
why do we need it inside minetest? |
21:05 |
thexyz |
why do you want to make it more bloated? |
21:05 |
VanessaE |
because the current UDP method is shit, and you should not separate a server from a vital function? |
21:06 |
thexyz |
what's with the KISS principle? |
21:06 |
thexyz |
why do you think minetestserver should serve large files? |
21:06 |
VanessaE |
um, hello? textures, mod code, sound files? |
21:06 |
RealBadAngel |
hello new fellow player, you wanna play please do read 100 pages on how to connect to servers easily and fast |
21:06 |
RealBadAngel |
yours sincerely devs team |
21:06 |
thexyz |
RealBadAngel: right? |
21:06 |
VanessaE |
come on |
21:06 |
thexyz |
you enter server address, port, press connect |
21:06 |
thexyz |
you're done |
21:07 |
thexyz |
what's the problem with that? |
21:07 |
thexyz |
you seem to not understand something |
21:07 |
thexyz |
(or is that me who doesn't?) |
21:07 |
RealBadAngel |
CURL |
21:07 |
thexyz |
yes |
21:07 |
thexyz |
so |
21:07 |
thexyz |
what's the problem? |
21:07 |
VanessaE |
why should I, as the server admin, have to fuck around with third-party utils just so my users don't have to wait 10 minutes to connect? |
21:07 |
RealBadAngel |
im newbie, tell me |
21:07 |
thexyz |
ok |
21:07 |
thexyz |
sure |
21:07 |
VanessaE |
or because I don't want my ISP to boot me off for running a web server maybe? |
21:08 |
thexyz |
RealBadAngel: download http://minetest.ru/builds/minetest-2013.03.02-3823878305.zip |
21:08 |
thexyz |
2) play |
21:08 |
Exio |
3) done |
21:08 |
thexyz |
as a user, you don't have to know anything |
21:08 |
Exio |
in any case, there are programs what do the "thing", i don't see any real reason for recoding stuff |
21:08 |
Exio |
rewriting the code just for the lulz? |
21:09 |
Exio |
rewrite lua in minetest because it is thirdy-party and you need it when you compile minetest |
21:09 |
thexyz |
VanessaE: because you're server admin |
21:09 |
VanessaE |
*facepalm* |
21:10 |
thexyz |
instead, you want us (small team!) to implement things into minetest |
21:10 |
RealBadAngel |
hey, i wanted to play with my friend. how do i set up curl? |
21:10 |
* RealBadAngel |
is still newbie |
21:10 |
VanessaE |
yes, because that's sort of a critical part of a game like this? |
21:11 |
thexyz |
RealBadAngel: you don't have to |
21:11 |
thexyz |
well, I got your point |
21:11 |
thexyz |
you want to play with your friend, you don't want to use public server, you want to play with some heavy mods |
21:11 |
RealBadAngel |
i want to make a server |
21:11 |
thexyz |
yeah |
21:12 |
Exio |
then you need to learn |
21:12 |
Exio |
with or without curl |
21:12 |
RealBadAngel |
which will end "without" |
21:12 |
RealBadAngel |
even i dont know how |
21:12 |
thexyz |
so |
21:13 |
thexyz |
hm.. |
21:13 |
thexyz |
well, I'm not going to implement this anyway |
21:13 |
thexyz |
as I think that's the wrong point |
21:13 |
PilzAdam |
if you play with a friend you can just send him your cache |
21:13 |
RealBadAngel |
we are devs, we can talk bout it, explain to each other |
21:13 |
thexyz |
(opinion which doesn't match mine is wrong) |
21:14 |
RealBadAngel |
but if some new person from outside will come and face the very same problem |
21:14 |
Exio |
i don't see the point for rewriting code, adding it to a bloated-server, and then get 1000 segfaults and other 2000 bugs because you had to rewrite a thing what was already coded in a thirdy party program and in a better way, faster and safer |
21:14 |
RealBadAngel |
will vote with feets. and walks away |
21:14 |
RealBadAngel |
thats my real point |
21:15 |
Exio |
minetest already "relies" on lua, and other stuff |
21:16 |
thexyz |
ok, fine |
21:16 |
thexyz |
pull requests with webservers implemented are welcome |
21:16 |
RealBadAngel |
connection stuff is one of the basic things that has to be done |
21:16 |
thexyz |
optional ones |
21:16 |
proller |
server setup must be easy, if admin need to setup nginx for files, *jabberd for chat, and anything else - its bad software |
21:16 |
VanessaE |
I don't recall anyone saying it HAD to be a web server? |
21:16 |
VanessaE |
I mean shit, something leaner like FTP would be just as good |
21:17 |
thexyz |
proller: my desktop is a bad software then |
21:17 |
RealBadAngel |
c55 started tcp branch |
21:17 |
RealBadAngel |
we were testing it even |
21:17 |
RealBadAngel |
what happened? |
21:17 |
thexyz |
i need to setup dbus, alsa and all the other shit |
21:17 |
proller |
and if admin must check every hour for segfaults from lua mod - it bad sowfware too |
21:18 |
PilzAdam |
RealBadAngel, c55 is wating for someone who knows more about stuff like this to finish it |
21:18 |
RealBadAngel |
cant we just use working tcp lib? |
21:18 |
RealBadAngel |
there are plenty out there |
21:18 |
Exio |
no, it is thirdy party, reimplement it |
21:18 |
Exio |
:) |
21:18 |
RealBadAngel |
lol |
21:19 |
proller |
thexyz, you can choice what use for desktop, but if you want play game and you must setup nginx - its bad |
21:19 |
thexyz |
proller: you don't have to |
21:19 |
RealBadAngel |
so, why we are not reimplementin irrlicht? |
21:19 |
VanessaE |
look, it should be one config file, one server program. |
21:19 |
thexyz |
RealBadAngel: that was your idea to reimplement web-server, wasn't it? |
21:19 |
VanessaE |
end of discussion. |
21:19 |
thexyz |
VanessaE: end? |
21:20 |
thexyz |
where are pull requests then? |
21:20 |
RealBadAngel |
not reimplementing web server |
21:20 |
RealBadAngel |
change udp to tcp |
21:20 |
VanessaE |
if I knew how to code such a thing (in C++) you'd already have had a pull request 6 months ago. |
21:20 |
proller |
http://stackoverflow.com/questions/738273/open-source-c-c-embedded-web-server |
21:20 |
|
iqualfragile1 joined #minetest-dev |
21:21 |
RealBadAngel |
http://code.google.com/p/mongoose/ |
21:21 |
RealBadAngel |
looks reasonable |
21:22 |
Exio |
meh, k |
21:22 |
Exio |
everything will be an option, no? |
21:22 |
Exio |
i still want to use udp |
21:22 |
Exio |
in the gameplay, that is why i still can play |
21:22 |
thexyz |
dependency hell? |
21:23 |
thexyz |
well, I guess, optional webserver is fine |
21:23 |
RealBadAngel |
call it whatever you like, just make server/client both ways fast |
21:23 |
RealBadAngel |
without setting up services |
21:23 |
Exio |
ways fast? how? |
21:24 |
Exio |
downloading the cache what is downloaded one time? |
21:24 |
Exio |
ok |
21:24 |
VanessaE |
one time?? |
21:24 |
RealBadAngel |
game downloads the 50mb cache when im doin a dinner |
21:24 |
VanessaE |
you evidently don't run a server |
21:24 |
Exio |
VanessaE: isn't it checked? |
21:25 |
proller |
need to look at openttd... |
21:25 |
Exio |
when you connect |
21:25 |
VanessaE |
if you keep your stuff up to date, every other time a user connects, something has to be re-downloaded. |
21:25 |
RealBadAngel |
in the very same time in background HD movie downloaded |
21:25 |
VanessaE |
(so to speak) |
21:25 |
RealBadAngel |
1,5 gb |
21:25 |
RealBadAngel |
so what we are talkin about |
21:25 |
Exio |
oh, you haz a nice connection |
21:25 |
Exio |
for people what doesn't just fuck them, no? |
21:26 |
thexyz |
what's the problem? |
21:26 |
VanessaE |
I have a connection that can sustain 380 kB/sec upstream, and the most anyone can usually get from me is a tenth of that - unless you, let's say, scp a file |
21:26 |
thexyz |
it's not like same files will be downloaded more that one time |
21:26 |
thexyz |
after embedding a webserver, that is |
21:26 |
Exio |
in any case, if someone is going to do that |
21:26 |
Exio |
add a busybox-httpd in minetest/somewhere |
21:27 |
Exio |
and integrate it, it is smaller and fast |
21:27 |
VanessaE |
busybox is good |
21:27 |
thexyz |
that's not an option |
21:27 |
VanessaE |
ok, busybox is bad. |
21:27 |
VanessaE |
:) |
21:27 |
thexyz |
i think mongoose is ok |
21:27 |
RealBadAngel |
me too |
21:27 |
thexyz |
hmmmm, darkrose, PilzAdam: opinions? |
21:27 |
RealBadAngel |
its mit |
21:27 |
RealBadAngel |
and foss |
21:28 |
thexyz |
and one .c file |
21:28 |
RealBadAngel |
at least we can try it |
21:31 |
PilzAdam |
https://en.wikipedia.org/wiki/Mongoose |
21:33 |
thexyz |
PilzAdam: that's not an opinion! |
21:34 |
PilzAdam |
if its better than the tcp branch by celeron55, than its ok |
21:34 |
thexyz |
dunno |
21:34 |
PilzAdam |
maybe first look into that a bit, and see what needs to be done |
21:34 |
thexyz |
but I don't like the idea of replacing UDP with TCP just because of slow content download speed |
21:35 |
RealBadAngel |
try tcp if aviable, then udp |
21:35 |
Exio |
if it is *only* for the on-connect stuff, it is ok for me, if not, i don't like that |
21:35 |
RealBadAngel |
no it is not |
21:36 |
Exio |
replacing the whole udp stuff for tcp? or what? |
21:36 |
RealBadAngel |
servers evolve, change contens |
21:36 |
RealBadAngel |
not replace |
21:36 |
thexyz |
actually |
21:36 |
thexyz |
what was the initial point of TCP branch? |
21:36 |
RealBadAngel |
leave udp as a last man standing |
21:37 |
RealBadAngel |
to speed things up |
21:37 |
thexyz |
what things? |
21:37 |
thexyz |
does it? |
21:37 |
RealBadAngel |
it was fast |
21:37 |
thexyz |
hm.. |
21:37 |
RealBadAngel |
but crashed |
21:37 |
thexyz |
last time I checked it at c55's server it was hanging |
21:37 |
PilzAdam |
thexyz, IIRC it was for downloading media with tcp |
21:38 |
RealBadAngel |
look, if the build in server can replace it |
21:39 |
RealBadAngel |
its ok |
21:39 |
RealBadAngel |
*built |
21:39 |
thexyz |
still need to wait for darkrose/hmmmm's reply |
21:39 |
hmmmm |
what |
21:40 |
* PilzAdam |
would be intersted in celeron55 opinion too |
21:40 |
hmmmm |
oh let me read |
21:40 |
|
Sylvertech joined #minetest-dev |
21:40 |
RealBadAngel |
ONLY condition that has to be fulfilled for both sides: player/admin is transparency |
21:40 |
RealBadAngel |
as with old UDP, start server, connect to it |
21:41 |
hmmmm |
what are you guys talking about? |
21:41 |
VanessaE |
hmmmm: in short: finding a faster way to handle the initial media download on connection. the proposal is http://code.google.com/p/mongoose/ |
21:41 |
VanessaE |
my proposal was FTP |
21:41 |
hmmmm |
WHAT?? lol |
21:41 |
hmmmm |
are you actually saying that |
21:42 |
hmmmm |
a web server in minetest |
21:42 |
VanessaE |
not me. |
21:42 |
VanessaE |
a webserver is overkill. |
21:42 |
hmmmm |
whose idea was this |
21:42 |
Sylvertech |
Hello. |
21:42 |
hmmmm |
who first said "we should put a webserver in" |
21:42 |
hmmmm |
i want to know so i can hang their picture up in the wall of shame so people will know to not take them seriously |
21:42 |
VanessaE |
I said we should put the media server in and dispense with using a third-party program like nginx/apache/etc |
21:43 |
VanessaE |
but I did not say (intentionally) that we should use a webserver. |
21:43 |
hmmmm |
;; |
21:43 |
hmmmm |
we have TCP for bulk transfers now |
21:43 |
VanessaE |
hmmmm: but it apparently does not work reliably. |
21:43 |
VanessaE |
crashes or something |
21:43 |
hmmmm |
the hell.. |
21:44 |
thexyz |
and I still don't see the problem in using nginx |
21:44 |
hmmmm |
i am not going to get involved with yet another large component, but i don't see why 1). we haven't been using TCP for this sort of thing a long, long time ago and 2). why nobody can get it working |
21:44 |
thexyz |
but if TCP works fine it's better to replace UDP with it |
21:44 |
hmmmm |
it's convoluted, nonsensical, |
21:44 |
thexyz |
and if it doesn't the only solution is to embed a webserver |
21:44 |
hmmmm |
no, TCP is high latency |
21:45 |
Exio |
FINALLY ONE |
21:45 |
hmmmm |
that certainly isn't the only solution |
21:45 |
Exio |
thanks hmmmm |
21:45 |
hmmmm |
the solution is to make TCP work right |
21:45 |
thexyz |
yeah |
21:45 |
hmmmm |
"whoop X doesn't work, let's give up on it and start using Y" |
21:45 |
hmmmm |
welp, i can't get this particular car design to work, we should start using aircraft to get to the store and back |
21:46 |
hmmmm |
we can strap a set of wheels onto a boeing 747 and have it follow a train track |
21:46 |
VanessaE |
hmmmm: well, the powers-that-be only have two years left to give us flying cars after all... |
21:46 |
VanessaE |
;) |
21:46 |
thexyz |
not "whoop X doesn't work" |
21:46 |
VanessaE |
but, we aren't using TCP because c55 doesn't like the quality or stability of his TCP branch |
21:47 |
thexyz |
whoop X doesn't work and no one is going to make it work |
21:47 |
hmmmm |
he has a bad habit of overcomplicating network protocols |
21:47 |
VanessaE |
THAT is more accurate, sadly. |
21:47 |
VanessaE |
ssssh, he might hear you. |
21:47 |
Sylvertech |
I heard too much caffience makes programmers aggressivers. |
21:47 |
Sylvertech |
Have you heard death metal lately |
21:47 |
Sylvertech |
? |
21:47 |
hmmmm |
what |
21:47 |
Sylvertech |
Its becoming rather soft. |
21:47 |
hmmmm |
ba dum tss |
21:47 |
Sylvertech |
The growls are almost slightly intelligble. |
21:48 |
Sylvertech |
That shut you up. |
21:48 |
hmmmm |
anyway my point is that i will not stand for integrating a webserver because people are lazy and incompetent |
21:48 |
hmmmm |
maybe if i get done with what i have on my plate i'll go and fix it |
21:49 |
* VanessaE |
waits for the impending standard chain of events: I head off for the night, c55 signs on, scrolls back, leaves some comment, and disappears before I can respond |
21:49 |
VanessaE |
s/I/whoever/ |
21:49 |
Sylvertech |
I say you program a bot, VanessaE |
21:49 |
hmmmm |
i say you get remotely on topic |
21:49 |
VanessaE |
no. |
21:49 |
Sylvertech |
Just have it chat with c55 unti lyo uwake up. |
21:49 |
VanessaE |
/kick Sylvertech |
21:49 |
Sylvertech |
He won't notice the difference ;) |
21:52 |
|
Sylvertech joined #minetest-dev |
21:53 |
thexyz |
Sylvertech: plz no offtopic |
21:53 |
Sylvertech |
Yes I'll keep that for #minetest |
21:53 |
Sylvertech |
toodle-oo |
21:54 |
VanessaE |
gotta run also. |
21:54 |
VanessaE |
bbl. |
21:58 |
thexyz |
hmmmm: what are you thoughts on leveldb? |
21:58 |
hmmmm |
my idea was to plop that right in for 0.4.6 |
22:00 |
thexyz |
ok |
22:00 |
thexyz |
actually, embedding leveldb seems for me similar to embedding a web server into minetest |
22:00 |
thexyz |
this solution is far from perfect |
22:01 |
thexyz |
but it works and can be done easily and then replaced by something better |
22:02 |
|
Sylvertech|2 joined #minetest-dev |
22:04 |
thexyz |
so i don't get why that much hate |
22:06 |
hmmmm |
how is it similar? |
22:07 |
RealBadAngel |
guys, shall i fix the routine to accept anything from lets says /sounds? https://github.com/minetest/minetest/blob/master/src/game.cpp#L847 |
22:07 |
thexyz |
02:00 thexyz: this solution is far from perfect |
22:07 |
thexyz |
02:00 thexyz: but it works and can be done easily and then replaced by something better |
22:07 |
PilzAdam |
RealBadAngel, maybe client/sounds |
22:07 |
hmmmm |
it's a real huge improvement over sqlite which is already in there due to someone's previous dumb idea that ended up happening |
22:08 |
thexyz |
RealBadAngel: um.. why do you want it to accept client-side sounds? |
22:08 |
hmmmm |
now here, it's not even an acceptable idea, it's absurd |
22:08 |
hmmmm |
and it seeks to solve a problem poorly that hasn't yet been implemented |
22:09 |
RealBadAngel |
thexyz, because 1) it works alreay 2) can create user sound themes |
22:09 |
hmmmm |
i know you guys want features, quick, but please take time to solve this the right way |
22:10 |
RealBadAngel |
thexyz, i modified ambience mod to play files with such names and it works |
22:10 |
RealBadAngel |
i do have local music |
22:11 |
RealBadAngel |
this way all the sound can be stored locally |
22:12 |
thexyz |
can you make it to work like user texture packs do? |
22:13 |
RealBadAngel |
yes |
22:13 |
RealBadAngel |
local conent could overwrite server sent |
22:14 |
hmmmm |
s/overwrite/never send in the first place/ |
22:14 |
hmmmm |
things ought to be done right the first time |
22:14 |
RealBadAngel |
theres a check |
22:14 |
hmmmm |
in the cache |
22:14 |
hmmmm |
what if you add it to the cache? |
22:14 |
RealBadAngel |
if not fetched then look in the specified dir |
22:15 |
RealBadAngel |
i can reverse it, if in dir then not check for fetched |
22:16 |
hmmmm |
good i guess |
22:16 |
hmmmm |
hrmm |
22:17 |
RealBadAngel |
server sent will be in cache anyways |
22:17 |
RealBadAngel |
as all another content |
22:20 |
RealBadAngel |
then we will need modified ambience mod, or sort of |
22:21 |
RealBadAngel |
that will come with minetest_game |
22:21 |
RealBadAngel |
and will expect certain files |
22:22 |
RealBadAngel |
passing this function will resolve it |
22:23 |
RealBadAngel |
i will try also to play music at main menu |
22:24 |
RealBadAngel |
should work imho |
22:25 |
RealBadAngel |
ha. in fact i think by now server will send only mods sounds |
22:25 |
proller |
openttd have self http client/server, no external libs |
22:25 |
RealBadAngel |
so files stored in local folder (that will come with build) will be completely local |
22:26 |
RealBadAngel |
so may serve as default ambience, click sounds, menu music etc |
22:27 |
RealBadAngel |
so server can order to play "default:click" |
22:27 |
RealBadAngel |
but client can change it to another sound by replacing files |
22:28 |
PilzAdam |
does the current code download the textures from the server when they are overwritten by the client from textures/all? |
22:29 |
RealBadAngel |
i guess whole content is being downloaded |
22:29 |
RealBadAngel |
no matter if local replacements are aviable |
22:29 |
PilzAdam |
somone should change that |
22:30 |
RealBadAngel |
i guess because i can see whole checks refers to fetched media |
22:30 |
RealBadAngel |
ie downloaded already |
22:31 |
RealBadAngel |
well, that would be a little complicated |
22:32 |
RealBadAngel |
many separate files, paths etc |
22:33 |
RealBadAngel |
i wonder why we wouldnt go in a very different direction |
22:33 |
RealBadAngel |
server on start creating single zip file |
22:33 |
RealBadAngel |
with whole media to download |
22:35 |
RealBadAngel |
instead of thousands of small files |
22:36 |
RealBadAngel |
it already works manual way |
22:36 |
RealBadAngel |
compres server cache, upload it somwhere |
22:37 |
RealBadAngel |
download by client, unpack |
22:37 |
RealBadAngel |
until server got updated no need to repack it |
22:37 |
RealBadAngel |
i tested this on many servers |
22:39 |
proller |
can we make reusable mods like in opennttd - you can download from any server, and start local game with any downloaded mod |
22:41 |
thexyz |
we don't download mods from server |
22:41 |
thexyz |
this is insecure |
22:43 |
proller |
maybe from github ? |
22:44 |
thexyz |
why should that be built into client? |
22:44 |
thexyz |
why should we depend on github? |
22:44 |
RealBadAngel |
what if github is down? |
22:45 |
RealBadAngel |
epic fail |
22:45 |
proller |
maybe something like mods.minetest.net? |
22:47 |
RealBadAngel |
you will exclude users modifications to mods |
22:47 |
RealBadAngel |
which is not good |
22:47 |
proller |
current way to setup mod is ugly. you must go to forum, download mod from github or mediafire or othershitdownload.com, unpack , run, look at segfault and delete mod 8) |
22:49 |
RealBadAngel |
central point of mod distribution and what server wil use are two separate things |
22:49 |
proller |
list of mods can contain name, version and hash, may choose between modified versions |
22:49 |
RealBadAngel |
mods sometimes require customization |
22:50 |
RealBadAngel |
like enable this, disable that |
22:50 |
thexyz |
proller: i agree here |
22:50 |
thexyz |
something like this is needed http://multa.bugs3.com/minetest/forum/ |
22:51 |
RealBadAngel |
maybe something in the middle |
22:51 |
RealBadAngel |
hash for not modified one, verified by central |
22:52 |
RealBadAngel |
any custom modifications will effect with new hash, server willl have to send it |
22:52 |
thexyz |
? >server willl have to send it |
22:52 |
RealBadAngel |
no matter client has that mod already, if the hash differs |
22:53 |
proller |
thexyz, yes, with standard distribution format, installable from game in one click |
22:55 |
RealBadAngel |
thexyz: you tested luajit? |
22:58 |
RealBadAngel |
i was wondering how it is possible they claim luajit is circa 50times faster adn you said that you havent noticed any changes in performance |
22:58 |
thexyz |
test it for yourself, it's easy to set up |
22:58 |
RealBadAngel |
any tips? |
22:59 |
thexyz |
what tips? |
22:59 |
RealBadAngel |
how to do that :) |
23:00 |
hmmmm |
again, it doesn't matter what the actual performance gains from a couple selected mods are, what matters is that it's faster in general, which means people can do big computations in lua without having to rely on people putting it in the core all the time |
23:00 |
thexyz |
and I only measured perfomance using some lua profiler |
23:01 |
thexyz |
RealBadAngel: replace ./src/lua/build/liblua.a with luajit, run make |
23:01 |
thexyz |
that's the easiest way |
23:02 |
RealBadAngel |
http://luajit.org/performance_x86.html |
23:02 |
RealBadAngel |
i was lookin at this |
23:03 |
thexyz |
it's not like we calculate md5 20000 times |
23:03 |
RealBadAngel |
but i tried already do massive calculations in Lua |
23:03 |
RealBadAngel |
on graphs, matrices etc |
23:04 |
RealBadAngel |
tree was generated in like 20 seconds |
23:04 |
RealBadAngel |
graphs im using in Technic |
23:05 |
RealBadAngel |
folks complain sometimes i kill cpus with it ;) |
23:06 |
RealBadAngel |
not to mention i need MOAR graphs hehe |
23:36 |
* VanessaE |
is back |
23:44 |
iqualfragile1 |
RealBadAngel, thexyz: rubenwardys "forum" is gona be avivable under mods.minetest.net as soon as its done |
23:45 |
iqualfragile1 |
he also provides apis for some launchers to get the mods |
23:46 |
|
dimeshake joined #minetest-dev |
23:49 |
|
dimeshake joined #minetest-dev |