| Time |
Nick |
Message |
| 00:14 |
|
Wayward_Tab joined #minetest-dev |
| 00:50 |
|
Taoki joined #minetest-dev |
| 01:08 |
|
Robby joined #minetest-dev |
| 02:11 |
|
kaeza joined #minetest-dev |
| 02:17 |
hmmmm |
is anybody around |
| 02:17 |
hmmmm |
i would really love to merge some pull requests |
| 02:18 |
hmmmm |
in fact i wanted to yesterday but nobody's ever around |
| 02:20 |
|
Zeno` joined #minetest-dev |
| 02:21 |
hmmmm |
yo |
| 02:21 |
hmmmm |
https://github.com/kwolekr/minetest/commit/5b237b4d94cc3fa4a95f951654dfbe37dcf2e2f9 |
| 02:21 |
hmmmm |
https://github.com/kwolekr/minetest/commit/8f38f8255cce29bccdc4adf16b2244650029e758 |
| 02:23 |
Zeno` |
lgtm |
| 02:25 |
hmmmm |
what do you think about #2651? |
| 02:25 |
ShadowBot |
https://github.com/minetest/minetest/issues/2651 -- fix pathfinder to produce more useful paths. by obneq |
| 02:26 |
Zeno` |
The one time I experimented with pathfinding I wasn't impressed, so if it's an improvement I vote aye :) |
| 02:26 |
hmmmm |
really sapier would be the one to take a look but he's mia |
| 02:26 |
Zeno` |
yeah |
| 02:27 |
Zeno` |
I remember thinking that the path finding code looked like a school project (sorry sapier) |
| 02:27 |
Zeno` |
:) |
| 02:28 |
hmmmm |
i've personally never written any pathfinding algorithms so i wouldn't know |
| 02:28 |
hmmmm |
heh |
| 02:29 |
Zeno` |
What I mean is I just found it unusual that there were so many standard algorithms implemented. Kind of like implementing 6 different sorting algorithms just for the sake of it |
| 02:30 |
Zeno` |
But I'm going off topic :) |
| 02:30 |
Zeno` |
When I tested them they didn't work very well at all at pathfinding |
| 02:30 |
hmmmm |
right, who would ever want dijkstra if your A* works |
| 02:31 |
hmmmm |
it's because of some nifty optimizations that turned out to not work in reality |
| 02:31 |
Zeno` |
yeah, that's what I mean. Normally you'd choose one and stick to it... |
| 02:31 |
hmmmm |
like using line_of_sight to add invalid paths |
| 02:31 |
hmmmm |
there are issues with that |
| 02:33 |
hmmmm |
i'm gonna push 2651 |
| 02:34 |
Zeno` |
yep |
| 02:34 |
Zeno` |
I agree |
| 02:36 |
Etzos |
So I'm attempting a fix for something I don't completely understand. I think my solution is decent, but I'd like to get comments/suggestions and the like. Is it better to just make a PR and wait for someone to comment or to ask here? |
| 02:36 |
Zeno` |
You can do either or both :) |
| 02:37 |
Zeno` |
There's no better way I don't think |
| 02:37 |
Zeno` |
You' |
| 02:37 |
Zeno` |
You'll probably get a quicker answer here |
| 02:38 |
Zeno` |
But even if it's discussed here a PR will still be needed, so personally I'd make a PR and then ask here |
| 02:38 |
hmmmm |
just ask here |
| 02:38 |
hmmmm |
it's probably quicker |
| 02:39 |
Zeno` |
Even if you were 100% sure of your PR if you start discussing it here it's likely to be merged quicker anyway :D |
| 02:39 |
Zeno` |
Brings it to people's attention |
| 02:39 |
Etzos |
Well, for now I'd like to test my changes a bit more. I just realized I missed something. |
| 02:42 |
hmmmm |
:)! add a unit test that verifies the bug you fixed does not exist anymore |
| 02:43 |
hmmmm |
i cleaned up the unit testing interface in an attempt to make it more attractive |
| 02:44 |
Zeno` |
/home/crobbins/minetest/src/unittest/test_random.cpp:180:46: warning: name lookup of 'i' changed |
| 02:44 |
Zeno` |
first time I've seen that warning :) |
| 02:46 |
Zeno` |
I suppose it's possible that some compiler may not follow ISO rules |
| 02:47 |
* Zeno` |
will not suggest MSVC heh |
| 02:49 |
hmmmm |
wth |
| 02:50 |
hmmmm |
"name lookup of i changed"? what does that even mean |
| 02:50 |
hmmmm |
is it because I used i earlier in another loop ? |
| 02:51 |
hmmmm |
they're not in the same scope but I do think older versions of MSVC respect this non-conformant scoping rule where i's scope would extend to outside of the conditional |
| 02:53 |
Zeno` |
yeah that's what it means |
| 02:54 |
Zeno` |
http://dpaste.com/2F74RM6 |
| 02:54 |
Zeno` |
stupid, but *shrug* |
| 02:55 |
hmmmm |
i hope the defaults for msvc follow the ISO scoping rules |
| 02:58 |
Zeno` |
well, you'd think they would |
| 02:58 |
Etzos |
Animated textures generate extruded wield meshes that are the full size of the animated texture instead of just the size of a frame. So this is my attempted fix: https://github.com/Etzos/minetest/commit/bbb8c2ecf3fd7726080f7fe4fefc8b3740b61655 does that make sense or is there a better way to go about handling that? |
| 02:59 |
Etzos |
And that is what's causing issue 2667. |
| 02:59 |
hmmmm |
heh i was literally just looking at that issue on the tracker |
| 03:00 |
Etzos |
Because the mesh isn't a power of 2 in both dimensions it's regenerated every time and since the texture is set to be upscaled to something huge it freezes up. |
| 03:01 |
hmmmm |
hmmm, *is* it correct to assume an animated image will always be square?? |
| 03:01 |
hmmmm |
I ask this because I legitimately don't know |
| 03:01 |
Etzos |
Neither do I. |
| 03:02 |
VanessaE |
it is never correct |
| 03:02 |
hmmmm |
of course |
| 03:02 |
hmmmm |
you divide the larger of the two dimensions by the number of frames in the animation |
| 03:02 |
hmmmm |
that gives you the other dimension |
| 03:02 |
VanessaE |
because animated textures are always a vertical strip, and in one case, I've got textures that have a frame aspect ratio of ~16:9 as well |
| 03:02 |
hmmmm |
oh they are? I thought they could be either horizontal or vertical |
| 03:03 |
VanessaE |
nope. always vertical |
| 03:03 |
Etzos |
Really? I thought they could be horizontal too. |
| 03:03 |
hmmmm |
I wonder if we make this assumption that they're always vertical elsewhere |
| 03:03 |
hmmmm |
shrug |
| 03:03 |
hmmmm |
well you heard her |
| 03:03 |
VanessaE |
if they can be horizontal, I've never heard of that. |
| 03:03 |
VanessaE |
there WAS a PR to add a "page" mode though |
| 03:03 |
VanessaE |
left-to-right, top-to-bottom order, I think |
| 03:04 |
hmmmm |
heh |
| 03:04 |
kahrl |
it's even in the name of the texture modifier: ^[verticalframe |
| 03:04 |
hmmmm |
that'd be overcomplicated |
| 03:04 |
hmmmm |
Etzos: so modify bool animated to be int num_frames |
| 03:04 |
hmmmm |
pass 1 for a non-animated texture, and the number of frames if they are |
| 03:05 |
VanessaE |
what I'd love to see is a way for gif animations to be passed in |
| 03:05 |
hmmmm |
then carry out the division appropriately |
| 03:05 |
hmmmm |
eww no gifs please |
| 03:05 |
VanessaE |
well MNG then |
| 03:05 |
VanessaE |
some kind of native animation format. |
| 03:05 |
hmmmm |
apng |
| 03:05 |
VanessaE |
apng, ok |
| 03:05 |
hmmmm |
but even so |
| 03:05 |
VanessaE |
in any case, something where frames are properly overlaid instead of vertical |
| 03:05 |
hmmmm |
irrlicht would need to support it i think, otherwise it'd be a lot more work |
| 03:05 |
VanessaE |
yeah, it doesn't |
| 03:06 |
VanessaE |
hell I'm not even sure how animations are actually handled by irrlicht in the first place |
| 03:06 |
VanessaE |
I assumed the vertical strip thing was a minetest-specific format |
| 03:06 |
hmmmm |
it is |
| 03:06 |
hmmmm |
but i'm saying that if we were to support an actual animated file format, we'd need to parse that file :( |
| 03:07 |
hmmmm |
png is harder than you think lol |
| 03:07 |
VanessaE |
Oh I'm sure it is |
| 03:08 |
VanessaE |
I only barely know the header (signs_lib reads part of it) |
| 03:09 |
hmmmm |
libpng is a dependency but idk how much easier that'd make anything |
| 03:09 |
hmmmm |
libpng is notoriously difficult to work with |
| 03:09 |
VanessaE |
I would assume one could easily pick a frame and extract it? |
| 03:09 |
hmmmm |
shrug |
| 03:10 |
kahrl |
does libpng actually support apng? |
| 03:10 |
Zeno` |
nobody cares about sprite sheets :( |
| 03:10 |
VanessaE |
no idea |
| 03:10 |
hmmmm |
lol looks like the answer is no |
| 03:10 |
kahrl |
I know there's a patch that adds it to libpng, but idk if it ever got merged |
| 03:10 |
hmmmm |
probably like an infiniality kind of thing |
| 03:11 |
Zeno` |
how I am meant to port my sprite sheets ripped from Amiga games when minetest doesn't support them?! This is not good |
| 03:12 |
VanessaE |
heh |
| 03:13 |
Etzos |
Was the general idea behind my change sound though or is there a better way to go about fixing this? |
| 03:13 |
hmmmm |
Etzos, it seems good |
| 03:14 |
Zeno` |
Seems good to me also... if you can get the frame count somehow |
| 03:14 |
hmmmm |
he can |
| 03:14 |
hmmmm |
he does when calling setExtruded |
| 03:14 |
Etzos |
Yep. Already made that change in the branch. |
| 03:14 |
Zeno` |
oh that's done? Sorry |
| 03:14 |
hmmmm |
it seems like APNG still has not been ratified by the png working group or w/e it's called |
| 03:15 |
hmmmm |
that's what they're waiting on |
| 03:16 |
|
kaeza joined #minetest-dev |
| 03:16 |
hmmmm |
agh no dammit, Etzos, u32 single_frame_height = dim.Height / num_frames; dim = core::dimension2d<u32>(dim.Width, single_frame_height); |
| 03:17 |
hmmmm |
the underlying problem with the first patch was the assumption that animated textures are square |
| 03:17 |
Etzos |
Bah. Of course. Sorry. |
| 03:17 |
hmmmm |
sure |
| 03:18 |
hmmmm |
just fix that, squash, and i'll commit |
| 03:21 |
Etzos |
And I almost made a similar mistake. I'm going to make an irregular sized animated item to test just to be safe. |
| 03:24 |
Zeno` |
maybe make an animatedFrameSingleFrameDimensionFactoryGenerator |
| 03:24 |
Zeno` |
hmmmm will like that |
| 03:30 |
hmmmm |
heh |
| 03:31 |
hmmmm |
I love it when people create win-win solutions by promoting synergy and thinking outside the box |
| 03:33 |
hmmmm |
what i don't love is how firefox decided that Konquerer should be my file manager and it doesn't need to bother with xdg-open |
| 03:33 |
hmmmm |
god dammit |
| 03:34 |
Zeno` |
That's to make your life easier |
| 03:35 |
Zeno` |
someone has moved my monitor! |
| 03:35 |
Zeno` |
brb, getting tape measure to set it to the right height |
| 03:38 |
Zeno` |
umm nvm |
| 03:38 |
Zeno` |
my chair seems to be what has changed height |
| 03:43 |
|
Fritigern joined #minetest-dev |
| 03:52 |
|
FR^3 joined #minetest-dev |
| 03:53 |
|
FR^3 joined #minetest-dev |
| 04:03 |
|
FR^3 joined #minetest-dev |
| 04:10 |
|
yh joined #minetest-dev |
| 04:12 |
yh |
minetest for android rtt not work in es1 or es2, i find it can't render to texture, is this minetest's bug or irrlicht's bug? |
| 04:13 |
yh |
i use es2 and open shaders, but fbo failed, no-one pixel was rendered to texture |
| 04:13 |
|
FR^3 joined #minetest-dev |
| 04:14 |
|
Hunterz joined #minetest-dev |
| 04:14 |
Zeno` |
yh, what is your device? |
| 04:16 |
yh |
samsung galaxy s4 |
| 04:22 |
Zeno` |
strange. I am using the android version on a galaxy tab s4 with no problems right now |
| 04:22 |
Zeno` |
BUT shaders have to be turned off |
| 04:23 |
yh |
i turn on shaders |
| 04:23 |
yh |
i work on android, minetest default use oges1, i fond "virtual" is lost before function "bindRTT()" in file "COGLESTexture.h", so rtt can't work |
| 04:24 |
Zeno` |
hmm |
| 04:27 |
yh |
you turned off shaders, are you use opengles 2.0? |
| 04:30 |
Zeno` |
I can't get shaders to work on ogl 1 OR 2 |
| 04:32 |
yh |
i can, after modify usr_path |
| 04:46 |
Zeno` |
modify it to what? |
| 04:54 |
|
chchjesus joined #minetest-dev |
| 05:48 |
|
nore joined #minetest-dev |
| 05:56 |
|
jin_xi joined #minetest-dev |
| 06:21 |
|
cib0 joined #minetest-dev |
| 06:27 |
|
Zeno` joined #minetest-dev |
| 07:01 |
sofar |
I couldn't find this in the lua_api.txt, but does set_node() erase all the node's meta? |
| 07:02 |
|
Darcidride joined #minetest-dev |
| 07:08 |
kaeza |
sofar, yes. hence, swap_node |
| 07:09 |
sofar |
argh if only I had read one line further |
| 07:09 |
|
FR^2 joined #minetest-dev |
| 07:27 |
|
kilbith joined #minetest-dev |
| 08:05 |
|
Yepoleb joined #minetest-dev |
| 08:13 |
|
Anchakor joined #minetest-dev |
| 08:23 |
|
cib joined #minetest-dev |
| 08:26 |
|
Darcidride joined #minetest-dev |
| 08:50 |
|
crazyR_ joined #minetest-dev |
| 08:50 |
|
crazyR_ joined #minetest-dev |
| 08:58 |
|
Calinou joined #minetest-dev |
| 09:02 |
|
rubenwardy joined #minetest-dev |
| 09:47 |
|
crazyR joined #minetest-dev |
| 09:47 |
|
crazyR joined #minetest-dev |
| 10:28 |
|
err404 joined #minetest-dev |
| 10:59 |
|
yh left #minetest-dev |
| 11:14 |
|
Lunatrius joined #minetest-dev |
| 11:37 |
|
cib0 joined #minetest-dev |
| 12:06 |
|
Darcidride_ joined #minetest-dev |
| 12:18 |
|
Darcidride__ joined #minetest-dev |
| 12:23 |
|
crazyR_ joined #minetest-dev |
| 12:34 |
|
rubenwardy joined #minetest-dev |
| 12:40 |
|
cd2 joined #minetest-dev |
| 12:48 |
|
rubenwardy joined #minetest-dev |
| 13:02 |
|
Darcidride_ joined #minetest-dev |
| 13:06 |
* Zeno` |
annoyed |
| 13:07 |
Zeno` |
we should look through the src code for sizeof(char) and blow them all up |
| 13:07 |
rubenwardy |
...8? |
| 13:07 |
Zeno` |
it's dumb and stupid |
| 13:07 |
Zeno` |
sizeof char ALWAYS == 1 |
| 13:08 |
rubenwardy |
Oh, yeah. |
| 13:08 |
Zeno` |
CHAR_BITS (which might be 8) is a different story, of course |
| 13:09 |
Zeno` |
much shit |
| 13:10 |
Zeno` |
in code there should rarely (if ever) be sizeof($built_in_type) |
| 13:10 |
jin_xi |
its only there about 20 times |
| 13:11 |
rubenwardy |
also dead code in #if 0 |
| 13:11 |
|
Darcidride joined #minetest-dev |
| 13:11 |
jin_xi |
dead stupid code is worse |
| 13:11 |
Zeno` |
I bet there is sizeof(u8), sizeof(u32), etc as well |
| 13:12 |
jin_xi |
yes |
| 13:12 |
Zeno` |
well that's just dumb |
| 13:12 |
Zeno` |
:) |
| 13:13 |
rubenwardy |
Minetest needs a cleaner |
| 13:14 |
rubenwardy |
It should maybe become part of a feature freeze, maybe |
| 13:14 |
rubenwardy |
Although we don't do that anymore O_O |
| 13:15 |
Zeno` |
it'd be quicker to start from scratch than clean it up |
| 13:15 |
Zeno` |
probably |
| 13:15 |
Zeno` |
hehhe |
| 13:17 |
rubenwardy |
48 occurrences of #if 0 |
| 13:17 |
rubenwardy |
This fix the code around your PR doesn't work when there is such a big code base with things that don't get touched much. |
| 13:22 |
Zeno` |
48?! |
| 13:22 |
Zeno` |
that's stupid... git takes care of that |
| 13:22 |
Zeno` |
all #if 0 should be removed |
| 13:23 |
rubenwardy |
And that's just #if 0 exactly, I dread # if or other forms |
| 13:23 |
rubenwardy |
7 occurrences of #if 1 |
| 13:23 |
Zeno` |
lol |
| 13:24 |
Zeno` |
no wonder just about nobody can understand the sr |
| 13:24 |
Zeno` |
src |
| 13:24 |
Zeno` |
and this is not even touching on the 4.8 million line functions |
| 13:25 |
rubenwardy |
I hope that's an exageration |
| 13:25 |
Zeno` |
slightly but not much |
| 13:25 |
rubenwardy |
Let's rewrite the whole thing into Go. |
| 13:26 |
Zeno` |
ok! |
| 13:26 |
Zeno` |
go! |
| 13:42 |
|
ElectronLibre joined #minetest-dev |
| 13:46 |
kahrl |
what's wrong with sizeof(u32) and the like? |
| 13:46 |
Zeno` |
it relies on knowing the type of whatever you're getting the size for |
| 13:47 |
kahrl |
oh, you mean like that |
| 13:47 |
Zeno` |
int array[20]; ........ sizeof array / sizeof (int); |
| 13:47 |
kahrl |
yeah, if you have a variable of the desired type then use that |
| 13:47 |
Zeno` |
yes, but a lot of MT doesn't do that :( |
| 13:48 |
Zeno` |
doesn't do sizeof(array) / sizeof(*array) I mean |
| 13:48 |
Zeno` |
hmm |
| 13:49 |
Zeno` |
kahrl, I'm just ranting btw ;) |
| 13:50 |
kahrl |
I don't see too many examples of it, mostly in serialize.cpp and in networkpacket.cpp |
| 13:50 |
Zeno` |
which are probably the two files I've been mostly looking at for the last 2 days hehehe |
| 13:51 |
kahrl |
the ones in serialize.cpp can't be fixed without adding more variables |
| 13:51 |
Zeno` |
serialize.cpp isn't too bad because knowing exact data size is kind of required |
| 13:51 |
Zeno` |
I didn't say that well |
| 13:52 |
Zeno` |
but I think you know what I mean |
| 13:58 |
|
celeron55 joined #minetest-dev |
| 14:17 |
|
AnotherBrick joined #minetest-dev |
| 14:30 |
|
hmmmm joined #minetest-dev |
| 14:32 |
|
err404 joined #minetest-dev |
| 14:43 |
rubenwardy |
https://github.com/minetest/minetest/blob/master/src/game.cpp#L4270 |
| 14:45 |
rubenwardy |
That is still displayed if the serialisation error is from config |
| 14:45 |
rubenwardy |
rather than network |
| 14:53 |
|
crazyR_ joined #minetest-dev |
| 15:11 |
|
ElectronLibre joined #minetest-dev |
| 15:12 |
|
zat joined #minetest-dev |
| 15:20 |
|
est31 joined #minetest-dev |
| 15:21 |
est31 |
Zeno`, hmmmm that warning doesnt only just occur for MSVC |
| 15:21 |
est31 |
also for travis linux |
| 15:22 |
Zeno` |
yeah the warning occurs... |
| 15:22 |
Zeno` |
I've never really attempted to use MSVC |
| 15:22 |
Zeno` |
the warning is from gcc |
| 15:29 |
est31 |
Zeno`, what are your thoughts about #2669 ? |
| 15:29 |
ShadowBot |
https://github.com/minetest/minetest/issues/2669 -- Stop NetworkPacket methods from producing bloated packets by jayarndt |
| 15:29 |
est31 |
I mean sizeof(char) has to be removed, but otherwise it looks ok |
| 15:31 |
kahrl |
as it stands, safeWriteToFile is not safe at all :( |
| 15:32 |
kahrl |
if the rename fails for some reason, both the original and the new file are deleted |
| 15:32 |
Zeno` |
yeah |
| 15:32 |
kahrl |
if I'm reading it correctly |
| 15:32 |
Zeno` |
but fixing that is... well, not really right |
| 15:32 |
kahrl |
what was the idea behind 5f1f1151d3? |
| 15:33 |
est31 |
oh |
| 15:33 |
est31 |
kahrl, it was #1477 |
| 15:33 |
ShadowBot |
https://github.com/minetest/minetest/issues/1477 -- Remove temporary file at safeWriteToFile() by Selat |
| 15:37 |
kahrl |
I think since e491f8cd48 fixed the bug where .~mt files were loaded as player files, we shouldn't worry about stray .~mt files left behind in error cases |
| 15:38 |
kahrl |
and safeWriteToFile should use a different code path on win32 as opposed to other systems, because it's the only system where a safe rename isn't actually possible |
| 15:39 |
est31 |
so what do you propose kahrl? |
| 15:39 |
kahrl |
(unless someone wants to write the code that uses NTFS transactions for that) |
| 15:40 |
kahrl |
I'd revert 5f1f1151d3 but I'm not sure what other side effects that might have |
| 15:40 |
kahrl |
have to think about this more |
| 15:41 |
Zeno` |
what is 5f1f1151d3 |
| 15:42 |
kahrl |
commit 5f1f1151d3a9c113902630adc16cc3f4845da7ba by selat |
| 15:42 |
Zeno` |
you can't give me a link/url? :P |
| 15:43 |
est31 |
https://github.com/minetest/minetest/commit/5f1f1151d3a9c113902630adc16cc3f4845da7ba |
| 15:43 |
est31 |
and pr was #1477 |
| 15:43 |
ShadowBot |
https://github.com/minetest/minetest/issues/1477 -- Remove temporary file at safeWriteToFile() by Selat |
| 15:44 |
Zeno` |
lol ok |
| 15:44 |
est31 |
gtg again bye |
| 15:44 |
* kahrl |
brbs too |
| 15:50 |
hmmmm |
hmm' |
| 15:51 |
hmmmm |
so what to do about the safeWriteToFile issue? just revert for now or does it require more? |
| 15:56 |
|
zat joined #minetest-dev |
| 15:58 |
|
zat joined #minetest-dev |
| 16:05 |
VanessaE |
regarding display of mesh nodes in the inventory, can someone please fix the render-to-texture so that it doesn't do this: http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/random/Screenshot%20-%2005042015%20-%2012%3a06%3a02%20PM.png |
| 16:05 |
VanessaE |
(those are 2- and 4-node-sized meshes) |
| 16:06 |
|
selat joined #minetest-dev |
| 16:24 |
Zeno` |
6-d rubik's cube! |
| 16:34 |
|
TheWild joined #minetest-dev |
| 16:49 |
|
est31 joined #minetest-dev |
| 17:17 |
|
celeron55 joined #minetest-dev |
| 17:26 |
|
Hijiri joined #minetest-dev |
| 17:27 |
|
Wayward_Tab joined #minetest-dev |
| 17:47 |
|
Hijiri joined #minetest-dev |
| 18:02 |
|
paramat joined #minetest-dev |
| 18:11 |
|
Miner_48er joined #minetest-dev |
| 18:14 |
paramat |
i'd like to push #2671 later, see issue 2662 for details and discussion. notifying celeron55 since this is client/AV |
| 18:14 |
ShadowBot |
https://github.com/minetest/minetest/issues/2671 -- Clouds: Make cloud area radius settable in .conf by paramat |
| 18:15 |
est31 |
paramat, seen my comment? |
| 18:16 |
paramat |
yes |
| 18:17 |
celeron55 |
paramat: ack |
| 18:18 |
|
MinetestForFun joined #minetest-dev |
| 18:22 |
|
MinetestForFun joined #minetest-dev |
| 18:22 |
|
crazyR_ joined #minetest-dev |
| 18:22 |
|
crazyR_ joined #minetest-dev |
| 18:23 |
paramat |
sorry to disturb you, wasn't sure whether to notify you or not :) |
| 18:23 |
paramat |
oh yes i'll add docs |
| 18:28 |
est31 |
otherwise +1 |
| 18:28 |
hmmmm |
paramat, it looks fine |
| 18:28 |
hmmmm |
I was the one who originally added cloud_height |
| 18:29 |
paramat |
okay |
| 18:29 |
hmmmm |
it's too late and i can't get rid of it now for reverse compatibility, so it hurt too much to have another clouds setting in the config |
| 18:29 |
hmmmm |
it doesn't hurt too much* |
| 18:30 |
Zeno` |
I dunno whether to laugh or cray |
| 18:30 |
Zeno` |
cry* |
| 18:30 |
hmmmm |
why |
| 18:30 |
Zeno` |
just merge it if it's ok |
| 18:30 |
hmmmm |
next up: |
| 18:30 |
hmmmm |
cloud_color |
| 18:30 |
hmmmm |
:( |
| 18:30 |
hmmmm |
cloud_shape |
| 18:30 |
hmmmm |
all the kind of minute details that should be done via scripting |
| 18:31 |
Zeno` |
la bamba! |
| 18:31 |
Zeno` |
but, seriously... paramat> sorry to disturb you, wasn't sure whether to notify you or not :) <--- that is bullshit |
| 18:31 |
Zeno` |
paramat, if you've done a good job merge it |
| 18:32 |
hmmmm |
yeah this probably falls under the 'trivial' clause |
| 18:32 |
Zeno` |
c55 doesn't need notificiation |
| 18:32 |
hmmmm |
maybe the setting should be read just once inside of the cloud ctor though |
| 18:34 |
hmmmm |
and it should be added to minetest.conf.example |
| 18:34 |
|
younishd joined #minetest-dev |
| 18:34 |
paramat |
hehe |
| 18:38 |
paramat |
"// Clouds move from X+ towards X-" should be Z+ to Z- i think, will correct this comment |
| 18:39 |
paramat |
cloud_radius is a const though, so will it be read once? |
| 18:40 |
hmmmm |
it'd have to be static const |
| 18:41 |
hmmmm |
having the const qualifier alone just means there would be a compile error if anything else tries to write to that variable |
| 18:45 |
paramat |
what's best, shall i leave it where it is but make it 'static const s16 cloud_radius_i..'? |
| 18:51 |
paramat |
i am indeed being restrained so am deciding against settable cloud size or speed :) |
| 18:58 |
paramat |
updated with docs |
| 18:59 |
hmmmm |
it's settable in the config file |
| 19:00 |
hmmmm |
it's not settable anywhere else after the executable has been launched, though |
| 19:00 |
hmmmm |
so what's the difference? |
| 19:01 |
est31 |
perhaps one day people want to have restart-less config changes? |
| 19:02 |
est31 |
but admit nobody will have problems with clouds :) |
| 19:04 |
Calinou |
been using enable_clouds = false for a long time now :P |
| 19:12 |
paramat |
okay will push this in a few hours, notify me if it needs improvement |
| 19:12 |
|
paramat left #minetest-dev |
| 19:22 |
|
Robert_Zenz joined #minetest-dev |
| 19:42 |
ShadowNinja |
I'm trying to fix that 0.4.11-dev version bug by using the version string from the cmake file instead of the Git tag. |
| 19:44 |
ShadowNinja |
I'm thinking of making the "comment out set(Version_string + "-dev")" thing into "use -DBUILD_RELEASE=TRUE" or similar. |
| 19:46 |
ShadowNinja |
Although bump_version.sh is actually broken since it depends on an exact line match and I changed that line. |
| 19:46 |
est31 |
no DBUILD_RELEASE is bad |
| 19:47 |
est31 |
It makes claiming to be a stable version too easy |
| 19:47 |
ShadowNinja |
est31: Because adding a "#" character is so hard? |
| 19:48 |
est31 |
people may think that that specific option strips debug symbols |
| 19:48 |
est31 |
makes a "release build" |
| 19:49 |
est31 |
its good the way it is |
| 19:50 |
ShadowNinja |
est31: No it isn't. 0.4.12-dev is getting called 0.4.11. Also, the line sometimes isn't uncommented before more commits are added, resulting in multiple "release" versions. |
| 19:51 |
est31 |
thats still better than giving a method to fake the version by default |
| 19:52 |
ShadowNinja |
"by default"? |
| 19:53 |
est31 |
if the person who does the release forgets to comment the line back in then its their faul |
| 19:53 |
est31 |
fault* |
| 19:54 |
est31 |
perhaps bump_version.sh should be adjusted to make two commits: one that bumps the version, a second that comments in the line |
| 19:54 |
est31 |
then if people push, they only have to get the tags right. |
| 19:55 |
ShadowNinja |
We can't pervent people from changing the version to whatever they like, trying to prevent them is futile and it's stupid to not do something because it could make such things easier. My only real concern is that someone might accidentally leave BUILD_RELEASE(or whatever I call it) on after making a release. |
| 19:55 |
est31 |
and that can be ensured if the second commit message is something like "Development continues with 0.4.12-dev" |
| 19:57 |
est31 |
I don't want to prevent them from doing it, but it should at least require them to make their git copy dirty |
| 19:58 |
est31 |
This is yet another of those changes that don't actually improve anything |
| 19:58 |
ShadowNinja |
Why? What's the difference between changing a .gitignored file and and then isn't? |
| 19:58 |
ShadowNinja |
I already gave two issues with the current system. |
| 19:59 |
est31 |
what was the first? |
| 19:59 |
ShadowNinja |
If you have an idea for something better please say so. |
| 19:59 |
ShadowNinja |
<ShadowNinja> est31: No it isn't. 0.4.12-dev is getting called 0.4.11. Also, the line sometimes isn't uncommented before more commits are added, resulting in multiple "release" versions. |
| 19:59 |
est31 |
0.4.12-dev is called 0.4.11-dev, right? |
| 20:01 |
ShadowNinja |
It can be a "set(BUILD_RELEASE TRUE)" line in CMakeLists.txt if you prefer that, but it needs to be available via a variable in GenerateVersion.cmake. |
| 20:01 |
ShadowNinja |
Yes. |
| 20:02 |
est31 |
that issue can be solved by leaving the -dev thing like it is |
| 20:02 |
|
Hijiri joined #minetest-dev |
| 20:02 |
|
Wayward_Tab joined #minetest-dev |
| 20:02 |
est31 |
just do what you proposed and read it from the cmake file instead |
| 20:02 |
est31 |
like me |
| 20:03 |
ShadowNinja |
est31: http://sprunge.us/ZhKH |
| 20:03 |
|
leat3 joined #minetest-dev |
| 20:03 |
ShadowNinja |
That adds it in the CMake file, because it's apparently so important to make it hard for people to change the version string. |
| 20:05 |
est31 |
you know when you download the source code and want to build, you expect that you don't have to do any further steps to get a clean 0.4.11 version |
| 20:06 |
est31 |
so people are forcefully confronted with having to manually set that variable when building a release build |
| 20:06 |
est31 |
right now thed don't have to do anything to "earn" a "clean" version string |
| 20:06 |
est31 |
just checkout the given tag |
| 20:09 |
ShadowNinja |
Yes, that's one good reason. |
| 20:10 |
est31 |
so you remove the number, do I read that right= |
| 20:10 |
est31 |
? |
| 20:11 |
|
leat3 joined #minetest-dev |
| 20:12 |
est31 |
also ShadowNinja you should check out master because I've changed GenerateVersion.cmake recently |
| 20:12 |
ShadowNinja |
est31: What number? |
| 20:13 |
est31 |
right now VERSION_GITHASH evaluates e.g. to 0.4.11-508-g1970e42 |
| 20:14 |
est31 |
and 508 is the number of commits since the 0.4.11 tag |
| 20:14 |
ShadowNinja |
After I'm done it should be 0.4.12-dev-01abcdef or so. |
| 20:14 |
ShadowNinja |
So yes. |
| 20:15 |
est31 |
If sfan5 doesn't need it for statistics I'm ok with removing the number |
| 20:15 |
ShadowNinja |
It doesn't seem to be really helpfull if you've got the SHA1 hash. |
| 20:15 |
sfan5 |
the commit number? |
| 20:15 |
ShadowNinja |
And if the version isn't based on the tag it will be confusing. |
| 20:15 |
sfan5 |
you can remove that |
| 20:16 |
|
crazyR_ joined #minetest-dev |
| 20:16 |
|
crazyR_ joined #minetest-dev |
| 20:16 |
ShadowNinja |
Now, should it be 0.4.12-sha or 0.4.12-dev-sha? |
| 20:17 |
est31 |
0.4.12-dev-sha |
| 20:17 |
|
ElectronLibre left #minetest-dev |
| 20:17 |
ShadowNinja |
Alright, good, that's easier. :-) |
| 20:24 |
est31 |
hmmmm, would you be ok with #2671 if I stored it into a local variable? |
| 20:24 |
ShadowBot |
https://github.com/minetest/minetest/issues/2671 -- Clouds: Make cloud area radius settable in .conf by paramat |
| 20:24 |
est31 |
cleaner than static |
| 20:25 |
hmmmm |
you mean a member variable |
| 20:25 |
est31 |
yes |
| 20:25 |
est31 |
private: in the Clouds class |
| 20:25 |
hmmmm |
yeah it really doesn't matter |
| 20:37 |
|
MinetestForFun joined #minetest-dev |
| 20:38 |
est31 |
https://gist.github.com/anonymous/93cc8dc414bc3180660f |
| 20:39 |
est31 |
and the full patch: https://gist.github.com/anonymous/f4a9874c58b1c560b364 |
| 20:40 |
est31 |
pushing. |
| 20:41 |
|
leat3 joined #minetest-dev |
| 20:43 |
|
paramat joined #minetest-dev |
| 20:44 |
paramat |
fine by me, but there's also a bool from settings at line 90 |
| 20:44 |
paramat |
perhaps that should be moved to the ctor also |
| 20:45 |
est31 |
yea |
| 20:46 |
paramat |
feel free to make your changes and push the PR yourself |
| 20:50 |
est31 |
ok |
| 20:51 |
est31 |
pushing |
| 20:56 |
paramat |
cool |
| 21:00 |
paramat |
a large radius like 32 has a stunning spacious effect, well worth the setting |
| 21:02 |
|
leat3 joined #minetest-dev |
| 21:07 |
paramat |
works for me. version detection is nuts, mine now says 0.5.0-dev |
| 21:09 |
|
paramat left #minetest-dev |
| 21:12 |
|
leat3 joined #minetest-dev |
| 21:43 |
|
leat3 joined #minetest-dev |
| 22:01 |
|
Hijiri joined #minetest-dev |
| 22:04 |
|
Wayward_Tab joined #minetest-dev |
| 22:05 |
|
leat3 joined #minetest-dev |
| 22:24 |
|
leat3 joined #minetest-dev |
| 22:29 |
est31 |
can #2672 be closed hmmmm? |
| 22:29 |
ShadowBot |
https://github.com/minetest/minetest/issues/2672 -- bad argument #-2 to get2d |
| 22:34 |
est31 |
and #2669 is ready for second review |
| 22:34 |
ShadowBot |
https://github.com/minetest/minetest/issues/2669 -- Stop NetworkPacket methods from producing bloated packets by jayarndt |
| 22:44 |
|
leat3 joined #minetest-dev |
| 22:59 |
ShadowNinja |
est31: #2673 |
| 22:59 |
ShadowBot |
https://github.com/minetest/minetest/issues/2673 -- Make Git version detection use VERSION_STRING instead of tags by ShadowNinja |
| 23:00 |
est31 |
looking |
| 23:02 |
est31 |
ShadowNinja, perhaps replace "# Comment for releases |
| 23:02 |
est31 |
" with "# Set to FALsE for releases" |
| 23:03 |
est31 |
FALSE* |
| 23:03 |
ShadowNinja |
est31: Alreade done. |
| 23:03 |
ShadowNinja |
(locally) |
| 23:03 |
est31 |
ah |
| 23:04 |
est31 |
otherwise it looks good |
| 23:05 |
est31 |
.git/rebase-apply/patch:114: new blank line at EOF. |
| 23:07 |
ShadowNinja |
Yeah, Git doesn't like that. |
| 23:08 |
est31 |
no problem for me |
| 23:13 |
est31 |
old behaviour added a 'g' at the start of the hash, like in: 0.5.4-1-ge9e5936 |
| 23:13 |
est31 |
add it if you want, otherwise +1 for #2673 |
| 23:13 |
ShadowBot |
https://github.com/minetest/minetest/issues/2673 -- Make Git version detection use VERSION_STRING instead of tags by ShadowNinja |
| 23:14 |
ShadowNinja |
The g doesn't seem usefull. I guess it stands for Git. |
| 23:15 |
|
leat3 joined #minetest-dev |
| 23:16 |
est31 |
Can I merge 2669 ? |
| 23:21 |
|
crazyR joined #minetest-dev |
| 23:21 |
|
crazyR joined #minetest-dev |
| 23:46 |
est31 |
I guess so |
| 23:56 |
ShadowNinja |
hmmmm: Here's how check_* should be checking things: http://sprunge.us/EUIL?diff luaL_check* is only for function arguments. |
