.NET Core 1.0 released today
-
@heterodox said in .NET Core 1.0 released today:
I hate to Blakey here, but "LOLOL WINDOWS 10 UPGRADE" is getting pretty old.
I still find it funny though
-
@heterodox said in .NET Core 1.0 released today:
I kind of would; have they done anything like that recently?
not on windows host, no, gut they're relatively new to POSIX systems. not much that they write runs on *nix, and posix permissions take a lot of wrapping one's head around.
I'm sure if they do it ti will be accidental, but i've seen that happen too many times with windows developers transplanted into linux.
especially if they decide to set up samba so they can do their file editing on windows where they are more comfortable.... getting permissions right between hosts over samba is..... tricky, so if one of the devs misconfigured SAMBA, then checked in with the bad permissions, and it got into the build and past QA.
it could happen, easily.
-
@cartman82 People wanted open source-y, that's what you're getting. Everything's broken all the time. The documentation sucks.
-
@accalia said in .NET Core 1.0 released today:
... it seems so typical of microsoft to give write permission to the executables that make up their dot net core 1.0 release to LITERALLY ANYONE WHO HAS ACCESS TO THE MACHINE!
Which is a problem because...?
-
@cartman82 said in .NET Core 1.0 released today:
Imagine how "old" it is for people who just want to keep using their Windows 7 system without having to worry about what trick will Microsoft pull next.
Yes yes you're in real emotional pain, etc. But at least make more creative jokes.
-
@blakeyrat that would be like letting ASPNET_USER with permissions to change a privileged executable in a server.
It is a serious problem, but MS doesn't do it since win98 days.
-
@fbmac So the only WTF here is accalia loves a really crappy OS with a shitty interface she doesn't even know how to use.
-
@accalia said in .NET Core 1.0 released today:
not much that they write runs on *nix, and posix permissions take a lot of wrapping one's head around.
Eh. They're simpler in a lot of cases than DACLs on Windows. (Not that simpler is necessarily good).
@accalia said in .NET Core 1.0 released today:
it could happen, easily.
So what you're saying is it could happen to anyone.
Microsoft used to be new to Linux, but the amount of commits I see from Microsoft employees on the Linux projects I follow is pretty crazy recently. They're all about that open source partnership (keeping them in the market).
I'd be interested to see a security assessment of SQL Server on Linux whenever that gets fully off the ground.
-
@blakeyrat said in .NET Core 1.0 released today:
@fbmac So the only WTF here is accalia loves a really crappy OS with a shitty interface she doesn't even know how to use.
Well, that's a little unnecessary.
-
@powerlord said in .NET Core 1.0 released today:
So, now I get to hunt down all the shit Upgrade 3 installed to remove it.
It's Visual Studio. Nuke and pave.
-
@heterodox said in .NET Core 1.0 released today:
@blakeyrat said in .NET Core 1.0 released today:
@fbmac So the only WTF here is accalia loves a really crappy OS with a shitty interface she doesn't even know how to use.
Well, that's a little unnecessary.
She also irrationally dislikes Microsoft. :fa_slashdot:
-
@dcon said in .NET Core 1.0 released today:
@powerlord said in .NET Core 1.0 released today:
So, now I get to hunt down all the shit Upgrade 3 installed to remove it.
It's Visual Studio. Nuke and pave.
it's Visual studio, installing it on the bare metal is , you need to install on a VM because that's the only way to control it, and forget aboput upgrading on the bare metal. the different versions fight!
-
@accalia said in .NET Core 1.0 released today:
it's Visual studio, installing it on the bare metal is , you need to install on a VM because that's the only way to control it, and forget aboput upgrading on the bare metal. the different versions fight!
Yeah follow the advice of a person who doesn't know how file permissions work.
-
@heterodox said in .NET Core 1.0 released today:
@blakeyrat said in .NET Core 1.0 released today:
@fbmac So the only WTF here is accalia loves a really crappy OS with a shitty interface she doesn't even know how to use.
Well, that's a little unnecessary.
more than a little.
@aliceif said in .NET Core 1.0 released today:
She also irrationally dislikes Microsoft. :fa_slashdot:
no, it's a very rational and well thought out reason. I'll grant you it is a shitty reason, but it still is a rational and well thought out shitty reason.
-
@blakeyrat said in .NET Core 1.0 released today:
Yeah follow the advice of a person who doesn't know how file permissions work.
yes, that sounds like an amazingly good idea! All praise blakeyrat for he has shown us the way! Follow the idiots and you will win!
-
@accalia said in .NET Core 1.0 released today:
it's Visual studio, installing it on the bare metal is , you need to install on a VM because that's the only way to control it,
I include rolling back to a checkpoint as a nuke/pave operation too. Tho maybe it's just nuke.
@accalia said in .NET Core 1.0 released today:
the different versions fight!
That, at least, is one issue I haven't had. Well, other than the fight over what-does-double-click-sln-mean. At one point I had 2008, 2010, 2012, 2013 all on one machine. Down to 2010/2015 now.
-
@blakeyrat said in .NET Core 1.0 released today:
@fbmac So the only WTF here is accalia loves a really crappy OS with a shitty interface she doesn't even know how to use.
When did @accalia said she loves Windows 8 ?
-
@dcon said in .NET Core 1.0 released today:
That, at least, is one issue I haven't had.
Try installing different versions of Crystal Reports too. I'm sure you'll find the trouble you're looking for.
-
@dcon said in .NET Core 1.0 released today:
That, at least, is one issue I haven't had. Well, other than the fight over what-does-double-click-sln-mean. At one point I had 2008, 2010, 2012, 2013 all on one machine. Down to 2010/2015 now.
Yeah, I have at least 2008, 2010, 2012, and 2015 on the same machine currently. Though 2012 and 2015 are Express, so that's components to conflict I suppose (to which @fbmac is alluding).
-
@TimeBandit said in .NET Core 1.0 released today:
@blakeyrat said in .NET Core 1.0 released today:
@fbmac So the only WTF here is accalia loves a really crappy OS with a shitty interface she doesn't even know how to use.
When did @accalia said she loves Windows 8 ?
at least he didn't accuse me of liking Windows ME
-
@dcon said in .NET Core 1.0 released today:
Well, other than the fight over what-does-double-click-sln-mean.
they fight over that, also if you have SSMS somethimes the visual studio shell bundled in SSMS decides it's a full install of VS and takes over a bunch of extensions, despite being completely unable to handle them.
-
@fbmac said in .NET Core 1.0 released today:
Try installing different versions of Crystal Reports too.
No. I always do a custom install. And never include that!
-
@dcon The conflict happened with the full version when I saw it, not the one that was included with VS.
-
So:
I installed all this stuff at home and it crashed VS whenever I tried editing the .cs file, and i gave up. Then I got into work this morning and installed it all here, and all things were broken for a while, but eventually started working. OpenTK exists for .NET Core, so there's a lot more potential than you might think.
-
@cartman82 They don't support 15.10 ... I ran into this months ago. 15.10 isn't LTS release, microsoft won't fix it.
-
@dcon Rocking 2010 still at home. Just doesn't have horrific bugs like 2013 and 2015.
-
@lucas1 16.04 is an LTS release though... does anyone here have 16.04 to test this with?
-
@lucas1 said in .NET Core 1.0 released today:
They don't support 15.10
@lucas1 said in .NET Core 1.0 released today:
Rocking 2010 still at home
Lucid Lynx is ancient.
-
@powerlord 16.04 worked with RC2 and was supported. I don't have my ubuntu laptop to check it out but I suspect it will be supported.
-
@accalia said in .NET Core 1.0 released today:
i wouldn't put it past M$FT to make some part of .net core accidentally world writeable
I wouldn't put it past them to try to replace
/dev/null
with a real file.Once had some “IIS Extensions for Apache” — or some name like that, it was long ago — that replaced that special device with a real file because of the way that it went about writing files “safely”. Funnily enough, the machine concerned wasn't happy for very long after that. (I know it wouldn't work on most modern machines, where
/dev
is a special filesystem, but I'd hate to put it to the test.)
-
So, since OpenTK works with .Net Core, I tried messing around with porting an old thing I did right out of uni to it and...
It worked. Perfectly.
However, apparently the entire way everyone interacted with OpenGL was deprecated in version 3.1, so I should probably learn the new version...
-
@Magus said in .NET Core 1.0 released today:
However, apparently the entire way everyone interacted with OpenGL was deprecated in version 3.1, so I should probably learn the new version...
Context plx?
-
@dkf said in .NET Core 1.0 released today:
I know it wouldn't work on most modern machines, where /dev is a special filesystem
I bet it actually would; it's not all that special.
-
@masonwheeler A lot of the GL stuff was deprecated.
No more:
GL.Color3(18, 58, 127); GL.Vertex2(-1f, 1f);
You have to set up array buffers and such, which is apparently more convenient and efficient, but much harder to understand. And you have to do half of it in GLSL.
-
@Magus said in .NET Core 1.0 released today:
@masonwheeler A lot of the GL stuff was deprecated.
No more:
GL.Color3(18, 58, 127); GL.Vertex2(-1f, 1f);
You have to set up array buffers and such, which is apparently more convenient and efficient, but much harder to understand. And you have to do half of it in GLSL.
Oh, that. Yeah, that's been around for a long time, and OpenGL is (mostly) better for it. The only real is eliminating GL_QUADS, which are incredibly useful for all sorts of different uses, and requiring you to use triangles (6 vertices per quad instead of 4) is a huge mess.
The common argument is that there's no good way to handle a "quad" in which the four vertices are not all coplanar. The obvious counterargument is "who cares; that's not what anyone was using them for. Make it undefined behavior and no one will care either way."
-
@masonwheeler WTF the mobile editor is completely broken, typing blind.
I would think floating point rounding might make determining whether 4 points actually are coplanar somewhat difficult.
-
@masonwheeler said in .NET Core 1.0 released today:
Oh, that. Yeah, that's been around for a long time, and OpenGL is (mostly) better for it.
I get that it's more efficient as a result. But before, you could have someone write a quick cube in code and then start it rotating in minutes.
Now you have to learn about arrayvertexbuffers, GLSL and whatever else just to get something onscreen at all! The whole thing has 10x more C in it now, with a bit more DirectX than I ever wanted to see.
-
@error said in .NET Core 1.0 released today:
@masonwheeler WTF the mobile editor is completely broken, typing blind.
I would think floating point rounding might make determining whether 4 points actually are coplanar somewhat difficult.
Are you familiar with the concept of "epsilon"?
-
@masonwheeler said in .NET Core 1.0 released today:
@error said in .NET Core 1.0 released today:
@masonwheeler WTF the mobile editor is completely broken, typing blind.
I would think floating point rounding might make determining whether 4 points actually are coplanar somewhat difficult.
Are you familiar with the concept of "epsilon"?
Epsilon is fine for comparing two points. This one is a little trickier than that. There are infinitely many possible sets of 4 coplanar points, and also infinitely many possible sets of almost coplanar points. Say I have a quad which has a slightly convex shape; there are two different planes involved described by two different triangles. Say you start with one triangle and check that the fourth point is coplanar... Infinitely many points could satisfy that criteria. Maybe I'm just not good enough at trigonometry, but how do you tell if that point is > epsilon units away from the infinite line of valid points?
It's probably just the fact that I suck at maths. Still, you can eliminate this problem entirely by using triangles, and IIRC the hardware is more optimized for triangles anyway, and I'm sure the algorithms involved are much simpler as well.
-
On MacOSX I am still seeing:
Lukes-MacBook-Pro:~ Luke$ dotnet --version 1.0.0-preview2-003121
After running the installer, is anyone seeing this on MacOSX?
-
@error said in .NET Core 1.0 released today:
Say I have a quad which has a slightly convex shape; there are two different planes involved described by two different triangles.
Why say that? Why would anyone do that when, as noted above, that is not what quads are used for?
Quads are for displaying flat squares and rectangles. Therefore, any reasonable algorithm that has to deal with rounding errors should begin from the assumption that this was intended to be a flat quad, and only start causing trouble if this is obviously not the case.
-
@masonwheeler Maybe I did it by accident with bad maths. I'm sure no program ever contains bugs. It's a bad situation to be in, for sure. The point is: what happens now?
-
@error said in .NET Core 1.0 released today:
@masonwheeler Maybe I did it by accident with bad maths. I'm sure no program ever contains bugs. It's a bad situation to be in, for sure. The point is: what happens now?
What happened before with non-flat quads?
-
@masonwheeler said in .NET Core 1.0 released today:
@error said in .NET Core 1.0 released today:
@masonwheeler Maybe I did it by accident with bad maths. I'm sure no program ever contains bugs. It's a bad situation to be in, for sure. The point is: what happens now?
What happened before with non-flat quads?
No idea, I've always used triangles.
Edit: also I said convex when I think I meant concave.
-
@lucas1 said in .NET Core 1.0 released today:
On MacOSX I am still seeing:
Lukes-MacBook-Pro:~ Luke$ dotnet --version 1.0.0-preview2-003121
After running the installer, is anyone seeing this on MacOSX?
-
Did a tutorial, made a triangle show up without GLSL. I will probably be programming a lot this weekend.
-
I think it's kind of cool that you can run this stuff on OSX (et al), but so far it isn't working for me. Everything relies heavily on nuget, but for whatever reason I can't communicate with that server. Can't build, can't restore, can't run. The whole thing is useless without it's package manager.
-
I did a thing!
public class Program { [STAThread] public static void Main() { using (var game = new GameWindow()) { game.Load += (sender, e) => { // setup settings, load textures, sounds game.VSync = VSyncMode.On; CreateVertices(); CreateShaders(); }; game.Resize += (sender, e) => { GL.Viewport(0, 0, game.Width, game.Height); }; game.UpdateFrame += (sender, e) => { // add game logic, input handling if (game.Keyboard[Key.Escape]) { game.Exit(); } }; game.RenderFrame += (sender, e) => { // render graphics GL.Clear(ClearBufferMask.ColorBufferBit | ClearBufferMask.DepthBufferBit); GL.DrawElements(BeginMode.Triangles, 6, DrawElementsType.UnsignedShort, 0); game.SwapBuffers(); }; // Run the game at 60 updates per second game.Run(60.0); } } private static void CreateShaders() { var vertexShader = @"#version 430 in layout(location=0) vec2 position; in layout(location=1) vec3 vertexColor; out vec3 theColor; void main() { gl_Position = vec4(position, 0.0, 1.0); theColor = vertexColor; }"; var fragmentShader = @"#version 430 in vec3 theColor; out vec4 daColor; void main() { daColor = vec4(theColor, 1.0); }"; var vertexShaderId = GL.CreateShader(ShaderType.VertexShader); var fragmentShaderId = GL.CreateShader(ShaderType.FragmentShader); GL.ShaderSource(vertexShaderId, vertexShader); GL.ShaderSource(fragmentShaderId, fragmentShader); GL.CompileShader(vertexShaderId); GL.CompileShader(fragmentShaderId); var programId = GL.CreateProgram(); GL.AttachShader(programId, vertexShaderId); GL.AttachShader(programId, fragmentShaderId); GL.LinkProgram(programId); GL.UseProgram(programId); } private static void CreateVertices() { var vertices = new[] { 0f, 0f, 1f, 0f, 0f, 1f, 1f, 0f, 1f, 0f, -1f, 1f, 0f, 0f, 1f, -1f, -1f, 0f, 1f, 0f, 1f, -1f, 0f, 0f, 1f, }; var vertexBufferId = GL.GenBuffer(); GL.BindBuffer(BufferTarget.ArrayBuffer, vertexBufferId); GL.BufferData(BufferTarget.ArrayBuffer, sizeof(float) * vertices.Length, vertices, BufferUsageHint.StaticDraw); GL.EnableVertexAttribArray(0); GL.VertexAttribPointer(0, 2, VertexAttribPointerType.Float, false, sizeof(float) * 5, 0); GL.EnableVertexAttribArray(1); GL.VertexAttribPointer(1, 3, VertexAttribPointerType.Float, false, sizeof(float) * 5, sizeof(float) * 2); ushort[] indices = { 0, 1, 2, 0, 3, 4 }; var indexBufferId = GL.GenBuffer(); GL.BindBuffer(BufferTarget.ElementArrayBuffer, indexBufferId); GL.BufferData(BufferTarget.ElementArrayBuffer, sizeof(ushort) * indices.Length, indices, BufferUsageHint.StaticDraw); } }
I followed a tutorial more like, but still!
-
@Magus said in .NET Core 1.0 released today:
var fragmentShader = @"#version 430 in vec3 theColor; out vec4 daColor; void main() { daColor = vec4(theColor, 1.0); }";
Wait, is... is that C code, including an entry point, embedded as a string in a C# program? And OpenGL apparently includes a C compiler/interpreter?
And when I thought I've seen everything...
EDIT: huh, apparently it's called GLSL or something like that. I'd still be looking quite if I saw that in a codebase.
-
@Maciejasjmj said in .NET Core 1.0 released today:
Wait, is... is that C code, including an entry point, embedded as a string in a C# program? And OpenGL apparently includes a C compiler/interpreter?
And when I thought I've seen everything...GLSL: OpenGL's shader language. It's very C-like, as is HLSL. It runs only on the GPU. Also, I'd never put it in my C# like this if I wasn't just doing a tutorial.