What are you working on? December 2011 Edition

[PieClock]

Made some snow, because I wanted to refresh my mind on angles and all that stuff.

[vid]http://dl.dropbox.com/u/99604/snooooooooooooow.webmvp8.webm[/vid]

Fun to play with, and makes a nice screensaver. :v:

[Jookia]

[QUOTE=thisBrad;33760194]Most people have plenty of hard drive storage, so I don't think memory would be an issue. Any why is it annoying? Because you have to wait for it to encode after?

Less annoying than shit FPS in my opinion.[/QUOTE]

There's no problem to solve here. An unoptimized video recorder lags? DEAR GOD.

[Naarkie]

[IMG]http://puu.sh/aPlI[/IMG]
:dance:

[BlkDucky]

[QUOTE=Naarkie;33760976][IMG]http://puu.sh/aPlI[/IMG]
:dance:[/QUOTE]

wow, I seriously thought you were trying to avoid white circles before I realised it was the snow on facepunch.

[acpm]

[QUOTE=BlkDucky;33761135]wow, I seriously thought you were trying to avoid white circles before I realised it was the snow on facepunch.[/QUOTE]

Winter makes (apparent) development fun

[Downsider]

[img]http://img84.imageshack.us/img84/2898/screenshot2011121700321.png[/img]

LUDUM DARE GO!!!! I'm making a tileset.

[NorthernGate]

[QUOTE=Downsider;33761926][img]http://img84.imageshack.us/img84/2898/screenshot2011121700321.png[/img]

LUDUM DARE GO!!!! I'm making a tileset.[/QUOTE]

You said you were going to bed, liar!

[Downsider]

[QUOTE=NorthernGate;33762023]You said you were going to bed, liar![/QUOTE]

You said YOU were going to bed, liar.

[Electroholic]

Fuck my software crashes in release mode but works fine in debug :(

[acpm]

Hi! I'm posting from the FPAPI! Rate me dumb for using a test script in the SDK. xDDD

[Blade]

FUCK YA! Got sending my own invokes working. Had to store a list of invokeids which were sent by the client overwriting them with my own and restoring them on return. Which allows me to control the invokeids so I can send my own invokes.

[img]http://goo.gl/jkn1T[/img]

[csharp] var msg = new RemotingMessage();
msg.operation = "getSummonerByName";
msg.destination = "summonerService";
msg.headers["DSRequestTimeout"] = 60;
msg.headers["DSId"] = RtmpUtil.RandomUidString();
msg.headers["DSEndpoint"] = "my-rtmps";
msg.body = new object[] { "high7" };
msg.messageId = RtmpUtil.RandomUidString();
var result = Connection.Call(msg);
if (result == null)
return;

var bodies = RtmpUtil.GetBodies(result);
foreach (var body in bodies)
{
var ao = body.Item1 as ASObject;
if (ao == null)
continue;

label1.Text = string.Format("Level: {0}\r\nIconId: {1}", Convert.ToInt32(ao["summonerLevel"]), ao["profileIconId"]);
}[/csharp]

Feel fucking amazing, especially since it worked first try :D.

[acpm]

Heh, it works. Gregor's (LuaStoned) old script, rewritten using the new API structure:

[lua]-- tests replying to a thread
-- usage: lua test\posttest.lua

local threadID = 1144771
local message = "Hi! I'm posting from the FPAPI! Rate me dumb for using a test script in the SDK. xDDD"

local facepunch = require( "facepunch" )
local session = facepunch.session
local thread = facepunch.thread

-- Setup our connector
-- Use luasocket for this test
require( "connectors.luasocket" )

io.write( "Username: " )
local username = io.read()
io.write( "Password: " )
local password = io.read()

local mySession = session( username, password )
print( "Logging in as " .. username .. "..." )
local error = -1
while error ~= 0 do
error = mySession:login()
end

session.setActiveSession( mySession )

local error, token = -1, nil
while error ~= 0 do
error, token = session.getSecurityToken()
end

local error = -1
while error ~= 0 do
error = thread.reply( threadID, message, token )
end
print( "Posted to thread!" )
[/lua]

[Lord Ned]

[QUOTE=thisBrad;33762301]Fuck my software crashes in release mode but works fine in debug :([/QUOTE]

A Hisenbug...

Could come from writing past the end of something (Debug puts more room), or a null-pointer?

[DevBug]

[QUOTE=Jookia;33755960]Vim?

[editline]17th December 2011[/editline]

Wow, I've never been so angry at people reinventing the wheel poorly because they don't want to use a fucking library that's been designed, thought about and properly programmed to be cross platform.

Unix uses UTF-8 paths, Windows uses UTF-16 paths. How the fuck are you going to have Unicode paths now?

In fact, this infuriates me:

1. Good.
2. So fucking what? Are you going to rewrite everything simply because it doesn't fit your conventions?
3. What's so bad about dependencies?
4. Boost can be used in commercial products with no cost, have you even read its 10 line long license?

Fuck Boost? Why?

I'm going to go out on a limb here and say that Facepunch suffers from a good amount of [url=http://en.wikipedia.org/wiki/Not_Invented_Here]Not Invented Here[/url] syndrome.

I'm not trying to start a flame war, but it's just so frustrating to see people go down the same path where I spent 3 years trying to write a game engine from scratch, restarting multiple times, just to look at it what I had: Half a filesystem abstraction layer that was Unix only and horrible to use.

Using third party libraries helps things immensely, and yes, there's a time when they don't fit your needs, but this isn't one of them.[/QUOTE]

I just realized who was posting.

[IMG]http://alltheragefaces.com/img/faces/large/misc-nothing-to-do-here-l.png[/IMG]

[Electroholic]

[QUOTE=Lord Ned;33762366]A Hisenbug...

Could come from writing past the end of something (Debug puts more room), or a null-pointer?[/QUOTE]

Well what happened was when I was compiling QT, my computer bluescreened half way through. Since it seemed to be working I didn't bother re compiling it.

I recompiled QT in release and my software doesn't crash anymore. I guess it was compiling release mode when I bluescreened :v:

Also, added search to my audio manager:
[IMG]http://img814.imageshack.us/img814/803/audiophilia201112170104.png[/IMG]

Edit:
Anyone know how to find the length of an mp3 track? The method I'm using takes fucking ages to calculate as it has to "decode" the whole file :(

[Hruhf]

[QUOTE=thisBrad;33763047]Well what happened was when I was compiling QT, my computer bluescreened half way through. Since it seemed to be working I didn't bother re compiling it.

I recompiled QT in release and my software doesn't crash anymore. I guess it was compiling release mode when I bluescreened :v:

Also, added search to my audio manager:
[IMG]http://img814.imageshack.us/img814/803/audiophilia201112170104.png[/IMG]

Edit:
Anyone know how to find the length of an mp3 track? The method I'm using takes fucking ages to calculate as it has to "decode" the whole file :([/QUOTE]

Isn't length one of the tags?

[Electroholic]

Yeah it is, but mp3's aren't required to have that tag filled with data. None of the mp3's I've tried so far have that tag filled ...unless the library I'm using isn't picking up that tag for some reason

[Tezzanator92]

Could you not check for the existence of that tag and if it isn't filled, calculate the length and fill it. That way when it is next read it doesn't need to recalculate the length.

Bit of a hack.

[StylixX]

Here's a guy showing how to do it in C++ without checking tags, maybe it could help you out.
[url]http://stackoverflow.com/questions/119404/time-length-of-an-mp3-file[/url]

[Noth]

Awesome, in the middle of making my Ludum Dare entry, and my monitor decides it would be a [i]great[/i] time to die. [img]http://www.facepunch.com/fp/emoot/smith.gif[/img]

[garry]

[QUOTE=Jookia;33758664]I'm sure with some optimization this stuff could work. For one you could pass the VP8_FMT_RGB24 (or something along those lines) flag and avoid having to convert to YUV.[/QUOTE]

VP8 doesn't support YUV - but yeah - that'd improve performance quite a bit.

[Jookia]

[QUOTE=garry;33764227]VP8 doesn't support YUV[/quote]

I assume this statement was due to your brain AIDS, but if it wasn't, why are you converting from RGB to YUV?

[robmaister12]

Ludum Dare is making me realize that there's a lot I haven't ever done... like write AABB collision response code. It's kinda hacked together, the general idea behind the code being that subtracting the distance from one corner of a square to the opposite corner of the second square from the addition of the sizes of the squares will leave you with the vector of minimum distance.

Turns out that also snaps you over to a corner, so I had to follow up by detecting which edge was colliding and removing motion on the opposite axis. After that I dealt with corners (2 axes considered colliding) by removing motion in the larger component.

And now for some sleep, tomorrow I'll be pumping out art (already started on it a bit with the rock) and designing the small level I'll be playing on.

Considering making a level editor, not sure if it'll be quicker than hardcoding everything (need to recompile to reposition things)

[editline]17th December 2011[/editline]

oh yeah and it's all on github like I said in the Ludum Dare thread... here's the physics manager I hacked together:

[url]https://github.com/Robmaister/RoversSpirit/blob/master/RoversSpirit/Physics/PhysicsManager.cs[/url]

[garry]

[QUOTE=Jookia;33764322]I assume this statement was due to your brain AIDS, but if it wasn't, why are you converting from RGB to YUV?[/QUOTE]

Yeah sorry, it ONLY supports YUV

[pollyzoid]

Visual Studio 11 has some sweet DirectX tools: [url]http://blogs.msdn.com/b/somasegar/archive/2011/12/15/visual-studio-11-platform-tooling-advances.aspx[/url]
Especially the [url=http://blogs.msdn.com/b/vcblog/archive/2011/11/08/10235150.aspx]debugger[/url].

[DarKSunrise]

That almost makes me want to use DirectX.

[Jookia]

[QUOTE=garry;33764378]Yeah sorry, it ONLY supports YUV[/QUOTE]

I don't claim to know anything of the conversion, but I'm pretty sure VP8 could do a lot better job (especially if the folks dig in to inline assembly) of conversion than either of us could. So maybe as a test you could benchmark RGB24 vs YUV inputs?

[Lexic]

[QUOTE=DarKSunrise;33764650]That almost makes me want to use DirectX.[/QUOTE]

Not enough to actually switch though. :v:
I wonder if anyone will make an OGL version once this hits mainstream.

[pollyzoid]

[QUOTE=Jookia;33764764]I don't claim to know anything of the conversion, but I'm pretty sure VP8 could do a lot better job (especially if the folks dig in to inline assembly) of conversion than either of us could. So maybe as a test you could benchmark RGB24 vs YUV inputs?[/QUOTE]

By "ONLY" I think garry means ONLY.
[url]http://code.google.com/p/webm/issues/detail?id=12[/url] (Yes, it's a bit old, first result on google)
[quote]FFdshow supports RGB-to-YUV conversion (when Raw conversion is enabled), so there is
neither a need for the VP8 encoder library to do this conversion, nor for the VP8
encoder filter.[/quote]

[Philly c]

[QUOTE=Lexic;33764842]Not enough to actually switch though. :v:
I wonder if anyone will make an OGL version once this hits mainstream.[/QUOTE]

A good opengl debugger? Haha you must be joking.