| Time |
Nick |
Message |
| 02:46 |
|
Vogarth joined #minetest-dev |
| 02:47 |
Vogarth |
Hello, having build errors with GIT code - /opt/local/include/curl/curlrules.h:143: error: size of array ‘__curl_rule_01__’ is negative |
| 02:48 |
Vogarth |
client.cpp:50 |
| 02:50 |
Vogarth |
hm. strange, that's just an include curl.h. nevermind - will check check further. |
| 02:53 |
Vogarth |
what should CUSTOM_XDG_APPS_DIR be set to? |
| 02:55 |
|
Vogarth left #minetest-dev |
| 03:24 |
|
Vogarth joined #minetest-dev |
| 03:26 |
Vogarth |
Hello, trying to get minetest running on a Mac, latest GIT version finally builds - but on running, window opens, black canvas, and a infinite loop of "Could not load texture: ". On ctrl-C, it spits out Error opening configuration file for writing: "/../minetest.conf" |
| 03:27 |
Vogarth |
Any clues as to what might be the problem? |
| 03:28 |
hmmmm |
how late is the latest git version |
| 03:28 |
hmmmm |
did you compile with -DRUN_IN_PLACE=1? |
| 03:29 |
Vogarth |
Yes, and -DCUSTOM_XDG_APPS_DIR="~/.minetest" |
| 03:29 |
hmmmm |
not sure what exactly CUSTOM_XDG_APPS_DIR should be |
| 03:29 |
hmmmm |
but anyway |
| 03:29 |
Vogarth |
running it out of bin directory. games is in ../games, but it can't seem to find games dir. |
| 03:29 |
hmmmm |
try running it out of the directory above bin |
| 03:29 |
hmmmm |
err |
| 03:29 |
Vogarth |
just a sec. |
| 03:30 |
Vogarth |
Same problem. |
| 03:30 |
Vogarth |
Also, can't create new word - no games found. |
| 03:30 |
hmmmm |
first go into porting.cpp and tell me if line 337 says " OS X" |
| 03:31 |
hmmmm |
or rather i should say, if /* OS X */ is around there |
| 03:31 |
Vogarth |
yes, seems to be in a comment. |
| 03:32 |
hmmmm |
how about around line 217 |
| 03:32 |
Vogarth |
yep. there too. |
| 03:32 |
hmmmm |
okay |
| 03:33 |
Vogarth |
I've been working to get this to compile for a few days now, finally today someone cleaned up genmap and it finally links. |
| 03:33 |
Vogarth |
Did have to build curl with universial from macports. |
| 03:34 |
Vogarth |
it now runs, but not quite works. |
| 03:34 |
hmmmm |
can you do me a huge favor? on the line right after pathRemoveFile(buf, '/'); in that section of code add in printf(" -- path: %s\n", buf);, compile, and run it again |
| 03:34 |
hmmmm |
then show me the output it says in the terminal window that starts with " -- path: " |
| 03:38 |
Vogarth |
it never did. How is it getting __APPLE__ defined? |
| 03:40 |
hmmmm |
i.. i don't know |
| 03:40 |
hmmmm |
so you mean to say that "WARNING: Relative path not properly supported on this platform" prints out when you run it? |
| 03:41 |
Vogarth |
oh, found it - just after bin/mine test - .... |
| 03:41 |
Vogarth |
-- path: |
| 03:41 |
Vogarth |
Irrlicht Engine version 1.8.0 |
| 03:41 |
Vogarth |
it went by so fast it wasn't on screen. |
| 03:41 |
hmmmm |
well that would explain it |
| 03:42 |
Vogarth |
no, I'll send the entire first block of text... |
| 03:42 |
hmmmm |
do you mind if i have you try out a couple things? |
| 03:42 |
Vogarth |
##### |
| 03:42 |
Vogarth |
-- path: |
| 03:42 |
Vogarth |
Irrlicht Engine version 1.8.0 |
| 03:42 |
Vogarth |
Darwin Kernel Version 10.8.0: Tue Jun 7 16:33:36 PDT 2011; root:xnu-1504.15.3~1/RELEASE_I386 |
| 03:42 |
Vogarth |
Using renderer: OpenGL 2.1 |
| 03:42 |
Vogarth |
NVIDIA GeForce GT 330M OpenGL Engine: NVIDIA Corporation |
| 03:42 |
Vogarth |
OpenGL driver version is 1.2 or better. |
| 03:42 |
Vogarth |
GLSL version: 1.2 |
| 03:42 |
Vogarth |
Could not load font because the file does not exist: |
| 03:42 |
Vogarth |
19:40:39: ERROR[main]: WARNING: Font file was not found. Using default font. |
| 03:42 |
Vogarth |
Could not load texture: |
| 03:42 |
Vogarth |
###### |
| 03:42 |
hmmmm |
dyude |
| 03:42 |
hmmmm |
dude |
| 03:42 |
hmmmm |
that's okay |
| 03:42 |
hmmmm |
please |
| 03:42 |
Vogarth |
don't mind. |
| 03:42 |
hmmmm |
mmmm. |
| 03:43 |
Vogarth |
Funny - darwin kernel should be 10.6.8. |
| 03:43 |
Vogarth |
then texture line repeats endlessly. |
| 03:44 |
hmmmm |
okay |
| 03:44 |
hmmmm |
i'm going to need you to add something in that printf |
| 03:45 |
hmmmm |
change it to printf(" -- path: %s, %d\n", buf, len); |
| 03:46 |
Vogarth |
len = 1024 |
| 03:47 |
hmmmm |
what does it say when you run grep BUFSIZ /usr/include/stdio.h? |
| 03:48 |
Vogarth |
also 1024. |
| 03:48 |
hmmmm |
okay |
| 03:49 |
hmmmm |
weird |
| 03:49 |
Vogarth |
do you have a mac dev system? |
| 03:49 |
hmmmm |
i do not. |
| 03:49 |
Vogarth |
Are you around SF? |
| 03:49 |
hmmmm |
no |
| 03:51 |
hmmmm |
hmmmmmmmmmmmmmmmmmmmm.. |
| 03:51 |
hmmmm |
alright |
| 03:51 |
Vogarth |
I need to go soon, but I'll check back here later this evening, and over the next few days. Happy to work with group to get mac platform working. |
| 03:51 |
hmmmm |
remove all of the code starting from /* OS X */ to /* FreeBSD */ |
| 03:51 |
hmmmm |
and run it from ./bin |
| 03:53 |
Vogarth |
ok. now says WARNING: Relative path not properly supported on this platform |
| 03:53 |
hmmmm |
that's okay for now until i figure out what's wrong |
| 03:54 |
Vogarth |
Same texture loop |
| 03:54 |
hmmmm |
you are running it in bin right? |
| 03:54 |
Vogarth |
yep. |
| 03:54 |
hmmmm |
well that's not right |
| 03:54 |
Vogarth |
found updateConfigFile(): Error opening configuration file for reading: "../../minetest.conf" |
| 03:54 |
Vogarth |
in debug.txt. |
| 03:55 |
Vogarth |
ah. |
| 03:55 |
Vogarth |
minetest with SER_FMT_VER_HIGHEST=25, VER=0.4.4-d1 BUILD_TYPE=Release RUN_IN_PLACE=1 USE_GETTEXT=0 USE_SOUND=1 STATIC_SHAREDIR=/usr/local/share/minetest |
| 03:55 |
hmmmm |
oh, yeah, that would cause the problem |
| 03:55 |
hmmmm |
..:/ |
| 03:55 |
Vogarth |
There is no /usr/local/share. |
| 03:56 |
Vogarth |
What goes in there? /usr/local/share/minetest/games?? |
| 03:56 |
hmmmm |
STATIC_SHAREDIR shouldn't even be set |
| 03:56 |
hmmmm |
no idea why you are setting it |
| 03:56 |
Vogarth |
never set it myself. |
| 03:56 |
hmmmm |
it should not be set |
| 03:56 |
hmmmm |
try unsetting it |
| 03:58 |
Vogarth |
do you recomend "cmake -DSTATIC_SHAREDIR " or ??? |
| 03:58 |
hmmmm |
you should simply run cmake -DRUN_IN_PLACE=1 . |
| 03:59 |
Vogarth |
RUN_IN_PLACE was set, but STATIC_SHAREDIR seems to live in porting.cpp and config.h. |
| 03:59 |
Vogarth |
also cmake_config.h |
| 04:00 |
hmmmm |
uhh |
| 04:00 |
Vogarth |
This is all stock from cmake . -DRUN_IN_PLACE=1 -DCUSTOM_XDG_APPS_DIR="~/.minetest" |
| 04:00 |
hmmmm |
set STATIC_SHAREDIR to . |
| 04:00 |
hmmmm |
in src/cmake_config.h |
| 04:00 |
hmmmm |
"." |
| 04:00 |
Vogarth |
with many library adjustments to get around build issues. |
| 04:00 |
Vogarth |
ok. |
| 04:00 |
hmmmm |
that will definitely fix the problem |
| 04:01 |
hmmmm |
but yes, maybe sometime tomorrow when you have time to, i'd definitely like to work with you more so I can find out why _NSGetExecutablePath is failing but not setting *bufsize to anything |
| 04:02 |
Vogarth |
sure, I'd like to have someone in the dev group fold in some changes I'm making to get it to build properly on mac. |
| 04:03 |
Vogarth |
still not working, do you want mac section re-added? |
| 04:03 |
hmmmm |
no |
| 04:03 |
hmmmm |
not yet |
| 04:03 |
Vogarth |
just a sec, |
| 04:04 |
Vogarth |
does make in src not push to bin, but make in minetest root will? |
| 04:04 |
hmmmm |
yes |
| 04:04 |
hmmmm |
you need to run make in the minetest root directory |
| 04:05 |
Vogarth |
still finding STATIC_SHAREDIR=/usr/local/share/minetest in debug.txt. |
| 04:06 |
Vogarth |
#define CMAKE_STATIC_SHAREDIR "." |
| 04:06 |
hmmmm |
are you sure you set CMAKE_STATIC_SHAREDIR to "." in src/cmake_config.h?? |
| 04:06 |
hmmmm |
okay |
| 04:06 |
Vogarth |
set in cmake_config.h. yep - sure. |
| 04:06 |
Vogarth |
something somewhere else redefined it. |
| 04:06 |
hmmmm |
are you sure your binaries aren't out of date? |
| 04:07 |
Vogarth |
yes, sure it rebuilt. I'll poke around later, i'll be around. |
| 04:07 |
hmmmm |
okay, sounds good |
| 04:07 |
Vogarth |
Always hmmmm? |
| 04:07 |
hmmmm |
? |
| 04:07 |
Vogarth |
on #minetest-dev |
| 04:07 |
hmmmm |
yes |
| 04:08 |
Vogarth |
not many vogarths. Thanks for the help! |
| 04:08 |
hmmmm |
np |
| 04:08 |
|
Vogarth left #minetest-dev |
| 04:09 |
hmmmm |
well that's not right.. was the Mac OS X documentation lying? why on earth would _NSGetExecutablePath fail? there's nothing for it to fail about aside from an insufficiently sized buffer |
| 04:09 |
hmmmm |
also you should be using MAX_PATH, not BUFSIZ, but that's not the problem here |
| 05:10 |
VanessaE |
oh my |
| 05:10 |
VanessaE |
build error with minetest |
| 05:10 |
VanessaE |
first time I ran into this one. |
| 05:11 |
VanessaE |
http://pastebin.com/L9362tSA |
| 05:11 |
VanessaE |
(from git fetched about 1 minute ago) |
| 05:12 |
VanessaE |
wait, that's the wrong section |
| 05:13 |
VanessaE |
there's the right one: http://pastebin.com/BUiyjqRf |
| 05:13 |
VanessaE |
looks like part of an actual patch file got inserted into the source code |
| 05:13 |
VanessaE |
(judging by the other errors that follow it) |
| 05:15 |
VanessaE |
haha, not quite, the offending section of code: |
| 05:15 |
VanessaE |
http://pastebin.com/c5MfMAih |
| 05:15 |
VanessaE |
somehow I doubt this is valid C++ |
| 05:15 |
VanessaE |
:) |
| 05:41 |
hmmmm |
what source file is that in? |
| 05:41 |
VanessaE |
hm, I guess you can ignore this. Something must have gone wrong when I tried to git pull or sapier's thread priority patch must have confused something. Re-cloned and it builds fine now. |
| 05:41 |
hmmmm |
ah that's good |
| 05:43 |
sfan5 |
"Security isn't a goal of minetest as I have learned." #435 |
| 07:07 |
hmmmm |
wow i hit a wall |
| 07:07 |
hmmmm |
it turns out that what i want to do with the mapgen names and so on is really hard in C++ |
| 07:08 |
hmmmm |
i have something like this http://ideone.com/PcWkDk but i want something cleaner |
| 07:08 |
hmmmm |
if only i had a virtual static function |
| 07:10 |
|
Vogarth joined #minetest-dev |
| 07:11 |
Vogarth |
Hello, I'm trying to get minetest working on mac, have it compiled, but it endlessly prints "Could not load texture:" |
| 07:12 |
Vogarth |
was able to set STATIC_SHAREDIR to "." |
| 07:13 |
Vogarth |
running bin/minetest from the source root dir. |
| 07:14 |
hmmmm |
no no run it from within the bin directory |
| 07:14 |
Vogarth |
same behavior. |
| 07:14 |
Vogarth |
hello hmmmm again. |
| 07:15 |
hmmmm |
i have no idea what's going on with your build there |
| 07:15 |
Vogarth |
I tried --verbose ... Searching worlds ... in ..worlds: 0 found. |
| 07:16 |
Vogarth |
there is no ../worlds directory. |
| 07:17 |
Vogarth |
on cntl-C, skipped writing of ../../minetest.conf because content wouldn't be modified |
| 07:17 |
hmmmm |
look in debug.txt for path_share and path_user |
| 07:17 |
hmmmm |
let's see what those paths are |
| 07:18 |
Vogarth |
both set to .. |
| 07:18 |
hmmmm |
and the current working directory is, at the moment, the minetest root directory |
| 07:18 |
Vogarth |
setting STATIC_SHAREDIR didn't trigger rebuild of src, did a make clean btw. |
| 07:19 |
hmmmm |
okay, delete debug.log |
| 07:19 |
Vogarth |
current working dir would be bin under root. |
| 07:19 |
Vogarth |
delete.txt? |
| 07:19 |
hmmmm |
delete.txt?? |
| 07:19 |
Vogarth |
to delete? |
| 07:19 |
hmmmm |
debug.txt rather |
| 07:19 |
Vogarth |
yep |
| 07:19 |
Vogarth |
start minetest again. |
| 07:20 |
hmmmm |
yep |
| 07:20 |
hmmmm |
now what's path_share? |
| 07:20 |
Vogarth |
still set to ".." |
| 07:20 |
hmmmm |
if that's the case and you're in bin, then .. is the minetest root |
| 07:21 |
hmmmm |
which means you wouldn't be having that problem |
| 07:21 |
Vogarth |
as written in debug.txt. |
| 07:21 |
Vogarth |
yes, what textures are trying to be read unsuccessfully? |
| 07:22 |
Vogarth |
and why ../../minetest.conf |
| 07:22 |
hmmmm |
textures/base/pack/stuff here |
| 07:22 |
hmmmm |
are you missing the textures directory? |
| 07:22 |
hmmmm |
or any other subdirectory |
| 07:24 |
Vogarth |
looks like a few png in base/pack, all just has a textures_here.txt file. |
| 07:24 |
hmmmm |
that's good then... |
| 07:25 |
hmmmm |
i have a feeling like this is a case of PEBKAC |
| 07:25 |
Vogarth |
there's fontlucida and fontdejavusansmono.png, are these the fonts it's searching for? |
| 07:25 |
hmmmm |
those are two of them |
| 07:26 |
hmmmm |
there should be those, menubg.png, menulogo.png, smoke_puff.png, sunrisebg.png, unknown_block.png, unknown_item.png, and unknown_object.png |
| 07:26 |
Vogarth |
the problem is always the case of PEBKAC, computers always just do what they are told. |
| 07:26 |
Vogarth |
My understanding though is that Mac hasn't worked in a long time. |
| 07:27 |
Vogarth |
This might be the first time everything is built 64-bit. |
| 07:27 |
hmmmm |
there's no reason why it wouldn't work |
| 07:27 |
hmmmm |
let's re-cover all of our bases |
| 07:27 |
Vogarth |
In theory - theory and reality are the same. |
| 07:27 |
hmmmm |
right now with that bit of code for OS X removed you should be getting "warning unsupported platform" |
| 07:27 |
hmmmm |
and you are running minetest with the CWD in the bin directory |
| 07:28 |
hmmmm |
your textures are indeed present |
| 07:28 |
hmmmm |
and you compiled with RUN_IN_PLACE=1, and STATIC_SHAREDIR=. |
| 07:28 |
hmmmm |
right to all of that? |
| 07:28 |
hmmmm |
s/right/yes/ |
| 07:28 |
Vogarth |
In reality - they are quite different. - |
| 07:28 |
Vogarth |
yes, correct. |
| 07:28 |
Vogarth |
not getting warning unsupported platform. |
| 07:29 |
Vogarth |
do get WARNING: Relative path not properly supported on this platform now. |
| 07:29 |
hmmmm |
yeah that i meant |
| 07:29 |
Vogarth |
then all correct. |
| 07:29 |
hmmmm |
i'm sorry i have no idea at all |
| 07:29 |
hmmmm |
any bit of reason says that should work |
| 07:29 |
Vogarth |
can we get it to tell us where it's trying to load font from? |
| 07:30 |
Vogarth |
somewhere around "Could not load font because the file does not exist:" |
| 07:31 |
hmmmm |
should be porting::path_user + DIR_DELIM + "textures" + DIR_DELIM + "base" + DIR_DELIM + "pack" |
| 07:31 |
hmmmm |
which in your case would be ../textures/base/pack |
| 07:31 |
hmmmm |
as expected |
| 07:31 |
hmmmm |
and if you were in bin, that would be exactly the correct path |
| 07:32 |
Vogarth |
Ah! |
| 07:32 |
hmmmm |
say, do you have the correct permissions on the textures? |
| 07:32 |
hmmmm |
ah what |
| 07:32 |
Vogarth |
Can we get it to tell us exactly what the path is - case sensitive! |
| 07:33 |
Vogarth |
I'm running on a Mac with a case sensitive filesystem. |
| 07:33 |
|
Oblivion1500 joined #minetest-dev |
| 07:33 |
hmmmm |
all other platforms minetest runs on save for windows are case sensitive as well |
| 07:33 |
hmmmm |
it can't be the problem |
| 07:34 |
hmmmm |
i bet it's a permissions problem |
| 07:34 |
hmmmm |
check the permissions of those textures |
| 07:34 |
Vogarth |
ok, permissions are read by everyone. |
| 07:34 |
Vogarth |
As a user I' have no problem viewing their content. |
| 07:34 |
Vogarth |
owned by root, but it fails run as root also. |
| 07:35 |
hmmmm |
i'm sorry i'm out of ideas |
| 07:35 |
hmmmm |
well |
| 07:36 |
hmmmm |
yeah |
| 07:36 |
hmmmm |
there is no way this is a minetest problem |
| 07:36 |
Vogarth |
Where is "Could not load texture" coming from? |
| 07:36 |
hmmmm |
*shrug* somewhere |
| 07:36 |
Vogarth |
content_cao.cpp |
| 07:37 |
hmmmm |
if you're thinking it's not loading the image correctly, then that means irrlicht would be broken on mac os x |
| 07:38 |
Vogarth |
I don't know how well irrlicht works, but when it says "could not load texture - it's supposed to also say what texture. |
| 07:38 |
Vogarth |
Here's the line... |
| 07:38 |
Vogarth |
errorstream<<"GenericCAO::updateTextures(): Could not load texture "<<texturestring<<std::endl; |
| 07:38 |
hmmmm |
it's also supposed to say GenericCAO::updateTextures() too |
| 07:39 |
hmmmm |
that code is for in-game stuff. you're way off |
| 07:39 |
Vogarth |
three lines above theres this gem... |
| 07:39 |
Vogarth |
if(texturestring == "") |
| 07:39 |
Vogarth |
continue; // Empty texture string means don't modify that material |
| 07:39 |
Vogarth |
it looks empty, but it doesn't act empty. |
| 07:39 |
hmmmm |
how about you just set texture_path in minetest.conf for now |
| 07:39 |
Vogarth |
ok. |
| 07:40 |
hmmmm |
texture_path=/your/full/path/to/minetest/textures |
| 07:41 |
Vogarth |
same behavior. What does "address =" do? |
| 07:42 |
hmmmm |
for the remote server |
| 07:42 |
hmmmm |
if setting texture_path doesn't work, you're either trolling me or you have a real big problem |
| 07:43 |
Vogarth |
not trolling, but still doesn't work. |
| 07:48 |
Vogarth |
well, setting texture_path in srcroot/minetest.conf, then changing settings in gui, writes a copy with new other settings in ../../minetest.conf. It's reading the file, but still no textures. |
| 07:50 |
Vogarth |
So - games = 0, textures = 0, minetest.conf kind of works - the gui comes up but crashes on certain events, and can't start a game. |
| 07:50 |
hmmmm |
i'm sorry i give up |
| 07:51 |
Vogarth |
this is the GIT version synced a few hours ago. the release version won't even compile - some problem with linking two functions from genmap.cpp which do seem to be used (but can't be seen by map.cpp). |
| 07:51 |
hmmmm |
genmap..? |
| 07:51 |
hmmmm |
don't you mean mapgen.cpp? |
| 07:51 |
Vogarth |
Is there a easy way to check irrlicht is functional? |
| 07:52 |
Vogarth |
mapgen.cpp, correct. |
| 07:52 |
hmmmm |
probably not |
| 07:53 |
Vogarth |
the example makefiles don't seem to link properly in irrlicht. |
| 07:53 |
hmmmm |
it sounds like you have a really broken setup.... nothing works for you |
| 07:54 |
hmmmm |
uh |
| 07:54 |
hmmmm |
did you use make or gmake for irrlicht? |
| 07:54 |
Vogarth |
I build other programs all the time, bzflag, mame, stella, my own stuff, iPhone apps, glut and opengl apps. |
| 07:55 |
Vogarth |
nether - was following mac recipe - uses xcodebuild -project MacOSX.xcodeprog. |
| 07:55 |
Vogarth |
instructions from ## From http://irrlicht3d.org/wiki/index.php?n=Main.InstallingIrrlicht |
| 07:55 |
Vogarth |
not a big fan, but it does say its successful. |
| 07:56 |
Vogarth |
now I'm not so sure. |
| 07:57 |
Vogarth |
It had a few issues on Lion - required some opengl patches to compile. SnowLeopard seemed ok. |
| 07:57 |
Vogarth |
Let me try all this under lion. I'll get back to you. |
| 08:02 |
hmmmm |
okay |
| 08:09 |
Vogarth |
irrlicht says build succeeded, minetest next. |
| 08:12 |
Vogarth |
this is same git sync from a few hours ago, without the osx section edit from earlier. Different complier (xcode 4.5 tools version) |
| 08:14 |
Vogarth |
nope, different behavior though - Irrlicht Engine version 1.8.0 |
| 08:14 |
Vogarth |
Darwin Kernel Version 11.4.2: Thu Aug 23 16:25:48 PDT 2012; root:xnu-1699.32.7~1/RELEASE_X86_64 |
| 08:14 |
Vogarth |
No doublebuffering available. |
| 08:14 |
Vogarth |
Segmentation fault: 11 |
| 08:16 |
Vogarth |
this is in a vm, but no window ever launches. I can try a current git sync. |
| 08:19 |
Vogarth |
building off latest git now. |
| 08:25 |
Vogarth |
btw, setting STATIC_SHAREDIR from cmake doesn't seem to push into cmake_config.h. doing it manually. |
| 08:26 |
Vogarth |
should I pull out OSX block in porting.cpp? |
| 08:31 |
Vogarth |
No luck, same problem. |
| 08:37 |
|
SiegeWizard joined #minetest-dev |
| 08:40 |
Vogarth |
looks like the OSX porting block does more harm than good- with it it thinks that path_share = @¿_<FF>^?/.. |
| 09:59 |
|
rsiska joined #minetest-dev |
| 11:02 |
|
SiegeWizard left #minetest-dev |
| 12:45 |
|
darkrose joined #minetest-dev |
| 12:45 |
|
darkrose joined #minetest-dev |
| 13:10 |
sfan5 |
thexyz: i still can't cross-compile |
| 13:11 |
sfan5 |
http://bpaste.net/show/QfcQwWD5JtbACSeyDBUx/ |
| 13:46 |
|
darkrose joined #minetest-dev |
| 13:46 |
|
darkrose joined #minetest-dev |
| 13:50 |
celeron55 |
sfan5: that's code by hmmmm |
| 13:52 |
sfan5 |
yeah, but thexyz said he fixed the problem |
| 13:53 |
sfan5 |
i compiled minetest with clang and now: unknown opcode |
| 13:53 |
thexyz |
i fixed the other problem |
| 13:54 |
thexyz |
that was related to windows.h/winsock2.h mess |
| 13:54 |
sfan5 |
oh...ok |
| 14:18 |
|
Taoki joined #minetest-dev |
| 14:35 |
|
PilzAdam joined #minetest-dev |
| 15:33 |
|
doserj joined #minetest-dev |
| 15:41 |
doserj |
PilzAdam: https://github.com/celeron55/minetest/pull/436 |
| 15:41 |
doserj |
did you encounter other issues with it (besides warning messages)? |
| 15:44 |
PilzAdam |
the "Enable All" button doesnt work anymore |
| 15:44 |
PilzAdam |
(only for Add-Ons, still works for modpacks) |
| 15:45 |
doserj |
ah |
| 15:51 |
doserj |
ok, fixed that. |
| 15:51 |
doserj |
btw, i have a one-line fix for the "delete server" crash. should i include that? |
| 15:52 |
doserj |
(he was comparing an unsigned to -1) |
| 15:52 |
PilzAdam |
what "delete server crash"? |
| 15:52 |
doserj |
delete from the server browser when no server is selected |
| 15:53 |
PilzAdam |
do it in a seperate commit |
| 15:53 |
PilzAdam |
only 1 feature/fix per commit |
| 15:54 |
doserj |
same pull request/branch, though? |
| 15:54 |
PilzAdam |
yea |
| 15:56 |
|
hmmmm joined #minetest-dev |
| 15:58 |
doserj |
pushed: https://github.com/celeron55/minetest/pull/436 should i squash my first two commits into one? |
| 15:59 |
PilzAdam |
seperate the fix and the improved behaviour |
| 16:02 |
PilzAdam |
doserj, empty modpacks are still handled like mods |
| 16:05 |
doserj |
Hmm. yes. How should they be handled? like other modpacks (making enable all/disable all no-ops)? |
| 16:05 |
doserj |
or like "Add-Ons" now by disabling the buttons? |
| 16:06 |
PilzAdam |
handle them like other modpacks |
| 16:19 |
doserj |
hm, that actually requires a new flag in ModSpec. Currently, the code only checks for emptiness of modpack_content to decide if something is a modpack or not. |
| 16:30 |
|
Calinou joined #minetest-dev |
| 16:36 |
doserj |
like this? https://github.com/celeron55/minetest/pull/436 |
| 16:38 |
doserj |
or i can squash the two improvements (empty modpacks and no mods at all) into one commit |
| 16:39 |
PilzAdam |
yea, squash them together |
| 16:43 |
doserj |
done. |
| 16:43 |
|
rubenwardy joined #minetest-dev |
| 16:45 |
PilzAdam |
commit messages should start with a capital letter |
| 16:48 |
|
Progers joined #minetest-dev |
| 16:48 |
Progers |
Hi |
| 16:48 |
doserj |
done |
| 16:51 |
PilzAdam |
doserj, "load_mod_ = false" is added to world.mt and old mods are not removed when saving the config |
| 16:52 |
PilzAdam |
(I removed all mods and then opened "Configure" and saved) |
| 16:53 |
PilzAdam |
it seems like no mod that is missing but configured is removed when saving the setting |
| 16:56 |
doserj |
but you still got the warning message, right? |
| 16:56 |
PilzAdam |
yep |
| 17:10 |
doserj |
ah yes. Settings:updateConfigFile doesn't check for that. it only ever adds to the settings. |
| 17:13 |
|
markveidemanis joined #minetest-dev |
| 17:15 |
|
markveidemanis left #minetest-dev |
| 17:24 |
Progers |
Installing MinGw taking long time :D |
| 17:26 |
sfan5 |
are you using a package manager like apt-get,yum or pacman |
| 17:26 |
sfan5 |
? |
| 17:28 |
Progers |
On windows 7 |
| 17:29 |
Progers |
Not Linux |
| 17:36 |
RealBadAngel |
Good to merge — The Travis build passed (Details) |
| 17:36 |
RealBadAngel |
whats that? |
| 17:38 |
doserj |
travis is basically a build bot that integrates into github and checks commits if they break builds or not |
| 17:39 |
PilzAdam |
we already have a PilzAdam who builds every single pull request |
| 17:40 |
doserj |
:) |
| 17:41 |
sfan5 |
:D |
| 17:42 |
RealBadAngel |
lol |
| 17:42 |
RealBadAngel |
i guess PilzAdam you had a very busy day yesterday then ;) |
| 17:43 |
PilzAdam |
indeed ;-) |
| 17:45 |
RealBadAngel |
im testin now my own patch to update treegen, some new features and speed optimization |
| 17:47 |
doserj |
PilzAdam: hah, Settings:updateConfigFile actually does remove settings, but only if some other value has changed... |
| 17:50 |
RealBadAngel |
PilzAdam, btw some of the trees with new features look way different. remember you said pine is similar to other trees? take a look at this then: http://realbadangel.pl/pines.png |
| 18:08 |
doserj |
PilzAdam: https://github.com/celeron55/minetest/pull/436 now contains a fix for removing settings. I can't reproduce the "load_mod_ = true" issue. I'll be back in about 1/2h. |
| 18:22 |
thexyz |
celeron55: I set up weblate instance for minetest http://translate.minetest.ru/projects/minetest/core/, now can you give https://github.com/minetest commit access? |
| 18:40 |
hmmmm |
rgggggg |
| 18:40 |
hmmmm |
ot |
| 18:41 |
hmmmm |
it's like every change i make to the mapgenparams stuff, i have to change the order in which things are created in server's ctor |
| 18:46 |
PilzAdam |
https://github.com/PilzAdam/minetest/commit/9ca95a15477595cf8db0ca4a3f1d8483d469d51c |
| 18:46 |
PilzAdam |
any comments? |
| 18:46 |
hmmmm |
ahh i was going to do that |
| 18:47 |
hmmmm |
well |
| 18:47 |
hmmmm |
IMO the problem isn't completely being slow, but rather all the machinery behind a setNode() call |
| 18:49 |
PilzAdam |
it is faster than looping through x,y and z in Lua |
| 18:49 |
hmmmm |
my idea was to get a VoxelManipulator and write to that, then blit it back directly to the environment in the same manner as the mapgen in finishBlockMake |
| 18:49 |
PilzAdam |
and it seems like the Lua method causes some lighting glitches, set_node_area() doesnt do that |
| 18:49 |
hmmmm |
well if you want fast i suggest you change for (X, Y, Z to for (Z, Y, X |
| 18:50 |
PilzAdam |
why is that faster? |
| 18:50 |
hmmmm |
because of cache locality |
| 18:50 |
hmmmm |
the next element to be read in the array is 4 bytes over |
| 18:51 |
hmmmm |
not 40000000 |
| 18:51 |
hmmmm |
so it doesn't need to jump around anywhere near as much |
| 18:51 |
VanessaE |
hmmmm: unrelated question when you have a second. |
| 18:51 |
hmmmm |
what's up. |
| 18:52 |
PilzAdam |
wow, no its only 1 sec. for 100000 nodes |
| 18:53 |
VanessaE |
I was experimenting with disabling default trees and found that I have to comment out two separate calls: mapgen.cpp line 2262 and mapgen_v6.cpp line 1367 |
| 18:53 |
PilzAdam |
*now |
| 18:53 |
VanessaE |
is there a reason why it's being called twice? |
| 18:54 |
hmmmm |
yes, because the one in mapgen.cpp line 2262 is commented out |
| 18:54 |
hmmmm |
see #if 0 //BIG COMMENT |
| 18:55 |
hmmmm |
(also i am aware of the tree spawning code itself being duplicated, i need to fix that but i need to fix the last hack that was committed yesterday faster) |
| 18:55 |
VanessaE |
strange. I better re-check what I did then |
| 18:57 |
VanessaE |
no, it isn't. |
| 18:57 |
hmmmm |
?? |
| 18:57 |
VanessaE |
hrm |
| 18:57 |
VanessaE |
wait |
| 18:58 |
PilzAdam |
it is |
| 18:58 |
VanessaE |
no, I don't see any conditional that would disable that call, and it's not commented out in current git (just re-cloned to check)\ |
| 18:58 |
PilzAdam |
look at the syntax highlighting of GitHub |
| 18:58 |
PilzAdam |
https://github.com/celeron55/minetest/blob/master/src/mapgen.cpp#L2262 and https://github.com/celeron55/minetest/blob/master/src/mapgen.cpp#L2645 |
| 18:58 |
PilzAdam |
both grey |
| 18:59 |
hmmmm |
yeah your editor might not highlight "#if 0"s as comments |
| 18:59 |
hmmmm |
everything after line 218 is non-functioning code |
| 18:59 |
Calinou |
I don't have a "show public" button in multiplayer and I crash when clicking "delete" in latest git.. |
| 19:00 |
VanessaE |
something ain't right then.. |
| 19:02 |
PilzAdam |
Calinou, there already is a fix for this crash pull-requested |
| 19:03 |
VanessaE |
hmmmm: hrm, odd. I must have been thinking wrong then. it is indeed enough to comment it out of mapgen_v6.cpp |
| 19:03 |
hmmmm |
ya |
| 19:03 |
VanessaE |
guess I just aimed at the wrong file :D |
| 19:03 |
hmmmm |
oh boy i see a bug in the random seed generation code |
| 19:03 |
hmmmm |
two problems in fact |
| 19:04 |
hmmmm |
i'll snatch that one up later |
| 19:06 |
VanessaE |
anyway if you're curious, RBA and I were discussing adding a Lua hook to allow a mod to disable the default trees (for example so that moretrees doesn't have to alias them to air, which would of course be slower than just not calling the make_tree() code to begin with\ |
| 19:06 |
VanessaE |
damn my typing sucks today. |
| 19:07 |
hmmmm |
you can already disable trees |
| 19:07 |
VanessaE |
from a mod? without re-aliasing the mapgen aliases?> |
| 19:07 |
hmmmm |
no, it has to be put in minetest.conf |
| 19:07 |
VanessaE |
right |
| 19:07 |
VanessaE |
we want to do this from a mod. |
| 19:08 |
hmmmm |
it just strikes me as hacky |
| 19:08 |
VanessaE |
well, programatically, the right way to disable them is to just not execute the make_tree() call, no? |
| 19:08 |
sfan5 |
hmmmm: http://bpaste.net/show/QfcQwWD5JtbACSeyDBUx/ |
| 19:09 |
sfan5 |
^ celeron55 said your code caused that |
| 19:09 |
hmmmm |
yes |
| 19:09 |
Calinou |
PilzAdam: however, on my netbook the "show public" option appears |
| 19:09 |
Calinou |
weird |
| 19:09 |
PilzAdam |
do you have curl enabled? |
| 19:09 |
sfan5 |
Calinou: compile with curl |
| 19:10 |
hmmmm |
sfan, was that a mingw build? |
| 19:10 |
sfan5 |
yes |
| 19:10 |
hmmmm |
ahhhh |
| 19:10 |
hmmmm |
hrmmmm |
| 19:10 |
PilzAdam |
hmmmm, sfan5, I get the same error |
| 19:10 |
Calinou |
is it enabled by default? |
| 19:10 |
sfan5 |
IIRC no |
| 19:10 |
Calinou |
k |
| 19:10 |
sfan5 |
use -DENABLE_CURL=! |
| 19:10 |
PilzAdam |
Calinou, it is disabled when you havent installed curl IIRC |
| 19:10 |
sfan5 |
*1 |
| 19:10 |
PilzAdam |
(like gettext) |
| 19:11 |
hmmmm |
okay i'm thinking.. |
| 19:11 |
VanessaE |
sfan5: is that a way of saying "enable it NOW!!" ? |
| 19:11 |
sfan5 |
do it faster ;) |
| 19:11 |
VanessaE |
;) |
| 19:11 |
hmmmm |
i simply don't understand what this error means /home/stefan/mt-build/build/c55/celeron55-minetest-33be32f/src/settings.h:583: error: there are no arguments to ‘strtok_s’ that depend on a template parameter, so a declaration of ‘strtok_s’ must be available |
| 19:11 |
hmmmm |
strtok_s is undeclared evidently |
| 19:11 |
sfan5 |
VanessaE: it forces cmake to enable it, if its not found cmake fails |
| 19:12 |
VanessaE |
sfan5: you missed my joke I guess :) |
| 19:12 |
|
rubenwardy1 joined #minetest-dev |
| 19:12 |
sfan5 |
VanessaE: no, i didn't but i thought an explanation would be nice ;) |
| 19:12 |
hmmmm |
sfan5, can you try changing _WIN32 on line 42 of porting.h to _MSC_VER? |
| 19:12 |
VanessaE |
heh |
| 19:13 |
sfan5 |
could you please ask pilzadam, my (current) computer is so slow |
| 19:13 |
PilzAdam |
testing... |
| 19:18 |
* sfan5 |
just noticed that he has more than 2000 posts |
| 19:23 |
PilzAdam |
hmmmm, http://bpaste.net/show/LxzVUpZkx2azVmyHN2HX/ |
| 19:23 |
hmmmm |
what |
| 19:23 |
hmmmm |
no strtok_s, no strtok_r |
| 19:23 |
hmmmm |
what does it want off of me, my soul? |
| 19:24 |
sfan5 |
yeah |
| 19:24 |
hmmmm |
http://www.mail-archive.com/mingw-w64-public@lists.sourceforge.net/msg02406.html |
| 19:25 |
hmmmm |
hrmmmm |
| 19:25 |
hmmmm |
but there it's simply not linking |
| 19:25 |
hmmmm |
here it's not present in the headers at all |
| 19:25 |
|
iqualfragile joined #minetest-dev |
| 19:27 |
hmmmm |
this is stupid, they don't support strtok_r because "it's an optional part of POSIX and we don't support pthreads, so we won't include strtok_r" |
| 19:28 |
sfan5 |
wait... since when is thexyz a forum admin? |
| 19:28 |
PilzAdam |
since, ever? |
| 19:28 |
sfan5 |
no? |
| 19:30 |
sfan5 |
i guess this topic is going to the trash http://forum.minetest.net/viewtopic.php?id=4539 |
| 19:32 |
doserj |
sfan5: since he hosts the forum? |
| 19:33 |
sfan5 |
he hosts the forum? |
| 19:33 |
sfan5 |
i didn't know that |
| 19:33 |
doserj |
since maybe a week, yes. celeron55 was feed up with the downtimes |
| 19:34 |
hmmmm |
alright here's what we're going to do |
| 19:35 |
hmmmm |
at line 58 in porting.h, add this |
| 19:35 |
hmmmm |
#ifdef G_OS_WIN32 |
| 19:35 |
hmmmm |
#define strtok_r(s,d,p) strtok(s,d) |
| 19:35 |
hmmmm |
#endif |
| 19:35 |
sfan5 |
G_OS_WIN32? where does that come from? |
| 19:35 |
PilzAdam |
should I leave this _MSC_VER at line 42 or switch back to _WIN32? |
| 19:35 |
hmmmm |
leave _MSC_VER |
| 19:35 |
hmmmm |
G_OS_WIN32 is something that MingW defines i think |
| 19:36 |
sfan5 |
"i think" |
| 19:36 |
hmmmm |
strtok() is thread-safe here because it's linked with MSVCRT, where all CRT functions are guaranteed to be thread-safe |
| 19:36 |
PilzAdam |
nope, doesnt work |
| 19:36 |
hmmmm |
ah no |
| 19:37 |
hmmmm |
not G_OS_WIN32, hold on let me look better |
| 19:37 |
doserj |
G_OS_WIN32 is gnome i think |
| 19:38 |
hmmmm |
yeah |
| 19:38 |
hmmmm |
uhhhh.... |
| 19:38 |
hmmmm |
#if defined(_WIN32) && defined(__GNUC__) ? |
| 19:38 |
hmmmm |
try that |
| 19:38 |
sfan5 |
gnome on windows? |
| 19:38 |
hmmmm |
yeah nevermind that |
| 19:39 |
PilzAdam |
seems to work |
| 19:39 |
hmmmm |
it's the first thing i saw and i thought it was a predefined macro for MingW |
| 19:39 |
hmmmm |
anyway this combination should work reliably, because i don't know of any other Windows compiler that would have __GNUC__ defined |
| 19:39 |
PilzAdam |
I get a bazillion of warnings |
| 19:39 |
hmmmm |
if someone could find a definitive answer as to what can be used to identify a MingW build, that'd be great |
| 19:40 |
hmmmm |
let's see 'me |
| 19:40 |
hmmmm |
'em* |
| 19:41 |
doserj |
is cygwin still alive? |
| 19:41 |
hmmmm |
yes but it sucks in comparison to SUA |
| 19:42 |
sfan5 |
seems like it's alive |
| 19:42 |
PilzAdam |
build done; uploading |
| 19:43 |
doserj |
__MINGW32__ should be set on mingw? |
| 19:44 |
sfan5 |
PilzAdam: coudl you give me a patch of the changes hmmmm told you |
| 19:44 |
sfan5 |
*could |
| 19:45 |
doserj |
hmmmm: http://sourceforge.net/p/predef/wiki/Compilers/ |
| 19:45 |
hmmmm |
alright |
| 19:45 |
hmmmm |
thanks doser |
| 19:46 |
hmmmm |
we change #if defined(_WIN32) && defined(__GNUC__) to #ifdef __MINGW32__ |
| 19:46 |
PilzAdam |
sfan5, https://gist.github.com/4597790 |
| 19:47 |
PilzAdam |
anyone here can test this win build? http://ompldr.org/vaDZqZQ/celeron55-minetest-33be32f.zip |
| 19:47 |
hmmmm |
i still can't get over how stupid mingw is |
| 19:51 |
* sfan5 |
is trying to build right now |
| 19:54 |
sfan5 |
:/ |
| 19:55 |
hmmmm |
oooh what now |
| 19:55 |
sfan5 |
http://bpaste.net/show/f2OKDh6XZkPPBpKJYgEH/ |
| 19:55 |
sfan5 |
usleep function is missing |
| 19:56 |
hmmmm |
oh, of course, right. |
| 19:58 |
hmmmm |
change it to this http://ideone.com/74TyM3 |
| 20:01 |
sfan5 |
... |
| 20:03 |
hmmmm |
what? |
| 20:04 |
sfan5 |
that didn't do anything |
| 20:06 |
hmmmm |
you're right |
| 20:06 |
sfan5 |
_MSC_VER is not defined, but we are using windows.the problem is that the linux definitions don't work on win |
| 20:06 |
hmmmm |
man this is something that i need to actually sit down and play with until it gets right |
| 20:07 |
hmmmm |
alright this should definitely work http://ideone.com/InwlCh |
| 20:08 |
hmmmm |
and yes i know _MSC_VER isn't defined, it's something that's defined for MSVC only and not MingW |
| 20:08 |
hmmmm |
which is why i am using it |
| 20:10 |
sfan5 |
ALIGNOF not defined |
| 20:10 |
hmmmm |
WHSGJKDFHIO |
| 20:11 |
hmmmm |
RRRRRG |
| 20:11 |
hmmmm |
alright alright let me think of something better, i'm putting 100% attention toward this now |
| 20:12 |
sfan5 |
no,non,non,non,nontrhh |
| 20:12 |
sfan5 |
ht |
| 20:12 |
sfan5 |
i got it working |
| 20:12 |
sfan5 |
#define ALIGNOF(x) __alignof(x) |
| 20:12 |
sfan5 |
that works for mingw too |
| 20:12 |
hmmmm |
it does? that's really good to know |
| 20:16 |
hmmmm |
http://ideone.com/0GF2U1 alright that's the final version |
| 20:16 |
hmmmm |
this would be a lot less infuriating if my mouse didn't do this double clicking crap |
| 20:20 |
sfan5 |
hmmmm: this is what i got https://gist.github.com/4598039 |
| 20:21 |
hmmmm |
use the most recent one |
| 20:23 |
sfan5 |
the one you posted won't work |
| 20:23 |
sfan5 |
it doesn't define the correct ALGINOF |
| 20:25 |
hmmmm |
OGF2U1? |
| 20:26 |
hmmmm |
i intend for it to define __alignof__(x) on MingW |
| 20:26 |
hmmmm |
since __alignof__(x) is the GCC variant, and MingW is GCC |
| 20:26 |
hmmmm |
so __alignof__(x) doesn't work on MingW, but __alignof(x) does? |
| 20:28 |
sfan5 |
__alignof(x) defenitly works, i don't know about __alginof__(x) |
| 20:28 |
sfan5 |
*definetly |
| 20:29 |
hmmmm |
then why did you say [03:23 PM] <sfan5> it doesn't define the correct ALGINOF ? |
| 20:29 |
sfan5 |
if __alginof works i assumend that __alginof__ doesn't work |
| 20:30 |
hmmmm |
that might not be the case for MingW |
| 20:30 |
hmmmm |
maybe GCC works with both |
| 20:30 |
sfan5 |
yah |
| 20:30 |
sfan5 |
*yeah |
| 20:31 |
|
rubenwardy left #minetest-dev |
| 20:32 |
hmmmm |
it's really inconvenient to have all these compatibility problems and not be able to test them on all platforms |
| 20:33 |
hmmmm |
i don't really forsee any big platform-dependent additions in the near future though |
| 20:39 |
sfan5 |
hmmmm: i get tons of "settings.h:740: warning: unused variable ‘fmtpos’" |
| 20:40 |
hmmmm |
ah yes... |
| 20:40 |
hmmmm |
i can't really do anything about that right now |
| 20:40 |
hmmmm |
what i'm going to do is write my own version of strtok_r() for platforms that don't have it *cough mingw* |
| 20:40 |
hmmmm |
and then put that in there |
| 20:41 |
hmmmm |
the #define strtok_r strtok thing is just so it compiles for now, i trust that nobody in the next day or so is going to add strtok() calls interposed with getStruct calls |
| 20:44 |
VanessaE |
ok, the API is, as usual, as vague as a dense fog.. minetest.node_dig(pos, node, digger) <--- digger is "userdata". How do I get the digger's username from it? |
| 20:45 |
sfan5 |
maybe :get_player_name() |
| 20:45 |
VanessaE |
tried that. attempt to call global 'get_player_name' (a nil value) |
| 20:46 |
VanessaE |
(and player: is not the right class or whatever, either) |
| 20:46 |
VanessaE |
wait, think I figured it out. |
| 20:47 |
VanessaE |
yup, got it. digger:get_player_name() |
| 20:51 |
sfan5 |
thats what i meant |
| 20:58 |
hmmmm |
damn how do you even break this up into two lines std::map<std::string, MapgenFactory<Mapgen, MapgenParams> *>::const_iterator iter = mglist.find(mgstr); |
| 20:58 |
hmmmm |
the type itself is over 80 chars |
| 22:11 |
VanessaE |
regarding this new right-click-to-open thing with doors.. How do I determine if the user's version of minetest uses this feature e.g. with its default doors? |
| 22:14 |
VanessaE |
and why am I asking these questions in the wrong channel? Sorry about that. |
