Game Development

3:00 PM

I don't like linkedin, a colleague visited my profile and I instantly got notification about it, which now makes me think twice if I should look at someone's profile, and usually the decision is not to

Vektorweg

@Lasse army of blind people. sounds bad what they do.

Lasse

huh?

Vektorweg

huh huh?

Lasse

I didn't get what you mean

Vektorweg

well. most people are driven to not look what other people do. so they are blind.

user92578

3:04 PM

aight, ill head to my band rehearsal, I wont dev tomorrow but might drop by at a break in school. Bye!

Vektorweg

on linkedIn.

Lasse

yeah, it kind of kills the point of linkedin when you drive people to be passive

Almo

I've gotten two jobs now through LinkedIn contacts.

Lasse

I've got none

well, I guess that's because I haven't been using it much, I got an account but I'm not sure if I ever logged in after I created it

now I had to update it

Vektorweg

my job interview has been canceled. i waited two weeks because they delayed it two times.

Lasse

3:18 PM

I know the feel, even though I won in the end. I waited my current job for two months

Roy T.

We 'IT-people' are privileged though. There's a large shortage of IT people (at least in Europe)

Almo

I had a job interview lined up at Ready At Dawn. I was leaving our house in England for the last time, on my way to the airport. Had no job, was going to move in with wife's dad.

Phone rang literally as I was leaving

Roy T.

For example: the company I start at next month couldn't expand because they couldn't find enough qualified IT-people

Almo

they cancelled. Someone interviewed who said "hire me now or I don't take it"

Roy T.

They even use part timers who are still in school

Almo

3:21 PM

they hired him and cancelled my interview.

>:(

Roy T.

Balls

I have some respect for playing it like that though

If you feel you're wanted, you can aim that high

Almo

I don't respect an employer who would hire someone who would say that

Roy T.

Thats the other end of the story

Almo

how will they try to strongarm me in the future?

Roy T.

They let themselves get strongarmed

Almo

3:22 PM

"give me a raise or I leave today, and you know what that means for the code I'm working on"

so, in the end, I probably didn't want to work for them anyway

shame; would have been gameplay programmer on the PSP god of war game

Roy T.

Is that still Crytek?

Almo

crytek?

Vektorweg

@RoyT. there might be lots of open IT jobs. But most of them are rather single-minded. Java and C# coders only and so on.

Roy T.

Oh wait Crytek is Rise, not god of war :P

Lasse

@Vektorweg You just called my most awesome job single-minded :P

Roy T.

3:24 PM

@Vektorweg true, but still I had no problem finding almost exactly what I was looking for

Vektorweg

@RoyT. Well i can't find a job as Haskell coder.

Roy T.

Thats because there is no such thing

Vektorweg

you might not know it yet. But Haskell is better. ;)

Roy T.

I know a lot about Haskell actually

I started my thesis project with Haskell

and had some classes in it

(well a lot, is a bit of a hyperbole here)

I even got Haskell to work with SDL and draw some nice graphics :)

Vektorweg

then you know that there aren't many jobs for Haskell coders.

Lasse

3:29 PM

should there be?

Vektorweg

definitely.

Lasse

why?

Vektorweg

because its a better tool and even the worst capitalism should work towards higher productivity.

Lasse

I don't see it as a better tool

Vektorweg

have you take a look at all those new features for your favorite language?

Lasse

3:32 PM

yes, I can make games on unity engine with my favourite language

Vektorweg

are you using C# ?

Lasse

yes

Roy T.

Though I think some aspects of Functional programming are very useful. There are a lot of drawbacks to using a purely functional programming language

Vektorweg

broken Mono C#, where every other C# coder got creeps?

Almo

@Vektorweg you're forgetting the cost of changing technologies.

there's a reason NTSC survived as the standard for TV broadcasting way, way longer than it should have.

Roy T.

3:34 PM

However having a functional language integrated into an OO language, to deal with datasets is very useful.

Lasse

@Vektorweg I got the creeps too, but that's not the point. The point is that my language is useful, unlike haskell

Roy T.

That way you can use the best tool for the job. Without having to choose

Vektorweg

@Almo actually, capitalism has forgotten what long-term planning is. Of course, it isn't easy to switch. But also no one does the first step.

Roy T.

Also comparing C# with capitalism isnt going to help any argument

Lasse

I'm also developing my own game engine with C# on newest mono and .net features, which makes your point invalid

Roy T.

3:35 PM

And wont even allow us to take any other arguments seriously

Almo

@Vektorweg capitalism has not forgotten what long term planning is.

that's oversimplified; the issue is more complex than that

not enough long-term planning is certainly a symptom of how businesses tend to be run

but stating it that way is glossing over the complexity of the issue

Vektorweg

sure. its just that there is an issue and i don't want to start a long talk about it. ;)

Almo

:D

Vektorweg

@RoyT. i don't compare C# with capitalism. its just like that companies looking for ads instead of trying which tools are actually good.

MickLH

@RoyT. This is one of the fundamental concepts of my programming language

Vektorweg

3:42 PM

C#, PHP. Java and other has been well advertised over the years. I know its difficult to argument with programming languages. But people don't even try it.

Roy T.

" @vektorweg New troll hobby: tell everyone on #C# at @freenodeirc to use #Haskell."

XD

@MickLH you're creating a language? Awesome

Please add multiple return arguments using built in tuple syntax!!

Vektorweg

because a friend of mine told me to do that. he's coding c#. ;)

Roy T.

like: (a, b) = Foo() { return (1, 2); }

MickLH

@RoyT. Ok the way I handled that was with the built in array operator

(a, b) => return [1, 2];

Roy T.

@Vektorweg actually there was someone on #XNA on efnet IRC who was banned for doing the same thing for over a year. Guy made himself really unpopular :P

Vektorweg

3:44 PM

@RoyT. not familiar with.

user4704

Was it Associator?

Roy T.

@Vektorweg YES

MickLH

it achieves the same goal, because despite being strongly typed, one of the types is the "weak" type lol, and that's the default type whenever it's unspecified.

Roy T.

@MickLH I have a hard time parsing that without context. Does it mean a is 1 and b is 2 after that statement?

user4704

Associator got banned from from #gamedev on afternet and #reddit-gamedev on freenode for mindlessly championing F# over everything for all purposes, anywhere.

Roy T.

3:45 PM

Because that is exactly what I want. That way you can instantly use them named

user4704

It was awesome.

Roy T.

@JoshPetrie oh yeah it was F#, not haskell

Vektorweg

@RoyT. what happened to XNA? the C# guys told me it has been canceled. And XNA was one of the cools things, C# people could work with.

MickLH

oooh I parsed it wrong, I converted your assignment into a goes-to operation lol

Roy T.

@Vektorweg XNA is ded

@Vektorweg but MonoGame (which is Mono for XNA) seems to be going strong

Vektorweg

3:46 PM

@RoyT. interesting.

Roy T.

@Vektorweg I still use XNA 4, works fine on PC and thats all I care about, since I never finish games :P

@Vektorweg though not being able to use .NET4.5 features, or DirectX10+ features is becoming a pain

Lasse

Some of the coolest games lately used monogame, like Transistor and Bastion

Roy T.

I'll probably try SharpDx for my next project

Vektorweg

@RoyT. i know that. i never get things done too.

Roy T.

@Lasse lately... those are ancient games :P

Lasse

3:47 PM

oops, not xna, monogame

transistor is not yet that ancient :(

Roy T.

@Lasse oh the ports?

Lasse

no, they actually used monogame

Roy T.

I'm 65% sure Bastion was normal XNA first

Lasse

the devs even released chunk of modified monogame source that they used, and monogame has been pulling some of the code in

MickLH

@RoyT. To achieve that you'd use the reference-respecting assignment operator... Which I haven't decided the symbol for but I use ~> right now just to make sure it works

Roy T.

3:49 PM

@Lasse make that 15% sure

MickLH

[&a, &b] ~> DoStuff();

Roy T.

@MickLH what kind of language is it, got cool stuff to show?

user4704

Bastion used XNA at some point, yeah.

Lasse

I think they ported to monogame before release

MickLH

@RoyT. Not really any awesome application yet sorry... And I've been struggling to categorize it to a type for a while now

Vektorweg

3:53 PM

most game developers overestimate the value of graphics. A game is made by its game mechanics.
Good Graphics only reduce the entry level. For that reason, i think, XNA is good enough. No one need fancy directX or broken opengl.

MickLH

I think it's primarily an imperative language with some declarative features, Strongly typed but one of the types is the "weak" type. Dynamically typed I guess, although it's meant to be statically typed to let the compiler cut down the amount of runtime bullshit going on.

Lasse

opengl is broken eh?

user4704

You are ignoring the benefits that modern D3D give you that are not related to "pretty graphics."

Vektorweg

@Lasse sure. opengl is the worst api ever existed.

or, as far as i know. ;)=

user4704

Hardly.

Lasse

3:56 PM

have you used PHP?

MickLH

@Vektorweg Well that makes your experience level clear ;)

Vektorweg

even php isn't that worse.

Lasse

well, seriously, opengl is hardly broken

user4704

OpenGL is not a bad API, it's an old API.

Roy T.

@Vektorweg I think the code that communicates with the Graphics API is very complex and could constitute more than 15% of the code base. Using a more modern graphics API will certainly help in reducing/finding bugs and coding more efficient

Almo

3:57 PM

@Vektorweg php_uname tells you about the current OS. Unless PHP can’t tell what it’s running on; then it tells you about the OS it was built on. It doesn’t tell you if this has happened.

user4704

Which makes it look less-good when held up against modern APIs.

user4704

The interesting question will be how good of an API the new version turns out to be.

Roy T.

So yes, even when making crappy graphics a modern API is a godsend

@JoshPetrie I think if you just use the modern parts of OpenGL its OK

Though I definitely prefer DirectX10+

user4704

In particular, how much they let themselves be hamstrung by classical decisions such as the typeless handle system.

Vektorweg

of course PHP is bad. but that bad? i don't know. its difficult. both is pure crap.

Roy T.

3:58 PM

But tbh I have worked little with both, just toy projects

Almo

also from php: "get_class($obj) returns the object’s class name. get_class() returns the name of the class the function is being called in. Setting aside that this one function does two radically different things: get_class(null)… acts like the latter. So you can’t trust it on an arbitrary value. Surprise!"

user4704

@RoyT. Yeah, modern GL is pretty decent, and if you wrap up the typeless integer handles you're left with only a handful of real problems.

Lasse

what NeHe did well in the past is now the shadow of current opengl, some websites just should have some expiration date that makes the content disappear after certain date

Almo

eev.ee/blog/2012/04/09/php-a-fractal-of-bad-design

user4704

Which is on par with modern D3D.

Almo

3:59 PM

php is pure evil.

I use php for simple projects. But I would not choose it for a large project if I could avoid it.

Roy T.

@Lasse one of my first posts on this websites was a link to NeHe's tutorials. I got instantly flamed for that :P

user4704

lol

user4704

Nehe.

Lasse

lol I can imagine

MickLH

I have a NeHe laser tube, I want to play with it now lol

user4704

4:00 PM

polygon.com/2014/10/20/7022819/… wat

Roy T.

This is a good resource for learning OpenGL anew using the modern way arcsynthesis.org/gltut

No more GlQuad(...) :(

Has much changed since 2012 in OpenGL country?

user4704

I can't make the correlation between 2012 and GL version.

user4704

But probably?

user4704

Looks like 4.3 was released in 2012.

user4704

So we're up to 4.5 now

Lasse

4:03 PM

yeah, sparse texture support for example I guess, opengl supports it on all platforms while only DX11 (only on win8.1) supports it

user4704

Which gives you DSA, flush control, sparse and bindless textures, better shader variable layout, buffer placement, and debug hooks.

user4704

And probably some other stuff I can't recall, in between those two versions.

MickLH

@RoyT. I think the best way I can sum up my ramblings earlier, and explain where the vision leads: Imagine a computer algebra tool, now replace the built-in language with javascript, then inject C style strong machine types, and finally build an Async high level API which works mostly on getters/setters and event callbacks.

Roy T.

@MickLH Like Matlab, but then javascript + strong types + some Asynic lib?

snake5

@Almo no, it isn't

MickLH

4:07 PM

Probably, I rarely use Matlab, I mostly use Mathematica

snake5

asp.net is closer to taking that role

user4704

@MickLH Seen gamedev.net/… ?

snake5

(not that php is even close to it)

Roy T.

@JoshPetrie thanks, mostly new features then, not really API (style) wise

user4704

@RoyT. Correct. There's a pending API 'redesign' but no concrete information on it

user4704

4:08 PM

Bunch of journal posts from a guy I used to work with about building a language.

Roy T.

@JoshPetrie I heard about that, this time they would actually break compat, right?

user4704

Dunno. I haven't really followed it other than knowing it was announced.

MickLH

@JoshPetrie it seems familiar but I can't open any of the threads :/

user4704

@MickLH Odd.

snake5

yeah, same here

user4704

4:09 PM

@RoyT. I hope so, but there was talk of doing that with 3 and they kinda... didn't.

snake5

all lead to the search page

user4704

lol wow

user4704

they do

user4704

gamedev.net/blog/355/…

user4704

there is the latest actual post

user4704

4:10 PM

GDNet has been pretty shit since they stopped using custom forum software for this COTS crap

MickLH

It seems cool but it's going in a very different direction than I would like to go

snake5

what direction is that?

MickLH

Well one of my top priorities is to reduce the total amount of typing, but without using any ridiculous acronyms

snake5

I think the total amount of typing with almost typeless C-ish code would be very minimal

Roy T.

@MickLH what is the language for, gamedev, scientific computing, ... ?

MickLH

4:14 PM

@RoyT. Mostly those things since that's what I do

Roy T.

both? :P

snake5

generally the bottleneck is workarounds for type compatibility

and code reuse for same meaning but different data types

Vektorweg

@MickLH you need to build something between prolog and haskell then.

MickLH

@snake5 ok like one small example: cast(type, value) rubs me the wrong way, I went with type(value) after probably a month of deliberation

snake5

scripting languages achieve minimal amount of typing by not requiring to define types, statically typed languages can do that with optional types and implicit templates

MickLH

4:17 PM

@Vektorweg the concept is that you use an imperative language to express a functional operation

snake5

@MickLH I'd prefix the type with some symbol to make a distinction between constructor/function and a cast, if there is one

MickLH

@snake5 that's the "beauty" of it

Types are their constructor function

snake5

sounds like a scripting language

MickLH

the int(x) function returns x casted to an int, and assigning to a weak type changes the type

Vektorweg

@snake5 lol. That sounds like Haskell. Every constructor is uppercase and every function must be lowercase.

snake5

4:18 PM

@Vektorweg case restrictions are funny

and a different kind of animal

so to speak :D

Vektorweg

a snake probably.

MickLH

but assigning to a strong type causes an implicit cast, so for example you can do float x = int("12.3"); and end up with an x == 12.0

snake5

cool

but that seems like a bit of a minor detail in comparison to data structures

and interface building

MickLH

That's the idea, I want it to feel trivial to accomplish things

snake5

is there a way to make an array class manually? or a template function?

Vektorweg

4:21 PM

speaking of details is dreaming. it isn't helping, but it makes so much fun.

Roy T.

How about manual memory layouts

or garbage collection?

I'd love to see a data first language that is as easy to use as C#

MickLH

You can actually use a function as a type without hassle, for example maybe you have a function PowerOfTwo(x) which returns the nearest power of two.. you can declare your variable PowerOfTwo x = 5; and it will give you the effect of calling x = PowerOfTwo(5);

Vektorweg

a easily usable language without GC isn't existing.

user4704

"Easily usable" is subjective.

MickLH

I went with GC sadly

You can manually delete objects though too

snake5

4:23 PM

@MickLH that is very sad

user4704

I think Objective-C is pretty easy to use, and it doesn't have a collector any longer.

MickLH

But it's reference counted so you can't force it to delete unless it's safe

user4704

Other people disagree, for various reasons.

user4704

Some of which I find laughable, such as "typing [ and ] is hard."

snake5

@MickLH that kind of defies the point

with circular references any object would be "safe"

user4704

4:24 PM

@MickLH Are the reference counts somehow weakened then?

Roy T.

@Vektorweg I disagree

user4704

Seems like a questionable constraint to violate.

MickLH

@JoshPetrie No it's literally equivalent to assigning null to the target

Vektorweg

@RoyT. but at some point, you don't want to work on a memory level.

MickLH

I'm still in deliberation on that issue

user4704

4:25 PM

So what happens to other outstanding references?

MickLH

@JoshPetrie nothing

user4704

So they dangle?

Vektorweg

@RoyT. you want to get things done. not having a GC is annoying.

Roy T.

@Vektorweg C is actually quite easy to use. Memory management isn't hard at all but you just have to be precise. Creating custom constructors allows you to not think about them much. Plus 90% of the variables can be allocated on the stack, which means they are removed automatically anyway :)

snake5

@Vektorweg you really have to, unless you're building a business app and can just throw RAM at your only server

MickLH

4:26 PM

@JoshPetrie No the object's life is extended until those references let go too

user4704

So the delete doesn't happen? I'm confused.

MickLH

Yeah, if it's still refcounted the delete is almost pointless

user4704

Ah I see. So it's more like a "please delete if you can?"

MickLH

The only thing it does is flag it to be immediately deleted when refcount reaches zero

Roy T.

@Vektorweg I actually run into memory issues with C# often, but you cant do much about them. You just have to make polling data structures and such, which infects a lot of code. Cant solve those problems at memory level. Languages like Haskell are even worse at that.

MickLH

4:27 PM

As opposed to being undefined when it will disappear

snake5

@MickLH that should happen always

MickLH

It does, currently, but I'd like to change that because sometimes it's thrashing the cache to free a million small blocks of memory

snake5

refcount=0 means absolutely unnecessary

Roy T.

@Vektorweg though I'd have to say that in most business software C# is fine. And allocating with virtual memory is so cheap :D.

user4704

Well, I can see a case for not deleting immediately when refcount is zero.

snake5

4:28 PM

shouldn't be a need to additionally flag it to get there

user4704

I don't think I agree with it.

user4704

But I can see it.

Vektorweg

@RoyT. language design isn't about computer performance. in worst case, haskell can work on a self defined space of memory too. but practically, its exhausting without an GC at hand.

snake5

@JoshPetrie what's the case?

MickLH

And like I said, it's still malleable, I consider the delete operation broken still because I could remove it from the language and lose nothing.

user4704

4:29 PM

@snake5 Basically, sometimes deleting is "expensive" by some metric. Reference counting has the interesting side-effect of somewhat-less-than-ideal deletion determinism.

snake5

I can think of two (pooling and CPU-cache-friendliness) but both suck

user4704

That is, in some context you don't know if dropping your reference will invoke a delete, and consequently, the perf hit.

user4704

So deferring that delete until some other time can be useful.

snake5

oh, that

user4704

But.

snake5

4:29 PM

well, it's too easy to fix

user4704

That situtation comes up rarely, in practice, I find.

MickLH

It started out more like a dangerous C++ delete, but I found it was of limited usefulness

snake5

yeah

Roy T.

@Vektorweg all language design that forgets its runs on a computer is flawed. You have to be able to address the weaknesses of the system in your language, you cant do it elsewhere. Dreaming up a language that works well on a system with infinite memory where everything you need is always in the L1 cache will cause problems in high performance scenarios.

snake5

it comes up rarely and it's easy to fix by just putting a reference somewhere else

on a timer perhaps

user4704

4:30 PM

And deferring the delete to some unknown time later doesn't actually help, it just moves the problem so some other context has to incur the delete cost.

MickLH

My next idea for it is to force the refcount down by one, and then null the object so that you can account for circular references one by one

user4704

@MickLH Why not something like ARC?

snake5

@JoshPetrie +1

seems like a really good solution

MickLH

@JoshPetrie What specifically do you mean?

user4704

ARC is (mostly) beautiful.

snake5

4:31 PM

haven't actually dealt with it but from what I read, seemed to make sense

Vektorweg

@RoyT. actually, thats what haskell do. abstract as far as possible away from the machine layer to get pure functional code. its working good enough.

snake5

@MickLH developer.apple.com/library/ios/documentation/swift/conceptual/…

user4704

ARC uses static analysis of the code flow to basically insert manual hold/drop reference calls.

Vektorweg

@RoyT. still. a language can be designed without consideration of machine details but can always afterwards define optimal structures for a certain machine when needed.

user4704

Memory protection doesn't have to be about "machine details."

MickLH

4:34 PM

That's basically what the idea is right now, but the analysis is very weak and only really triggers when something goes fully out of scope

Roy T.

@Vektorweg but with Haskell scientific institutes write their own compiler so that it works good on their super computer. That's because these issues cannot be addressed in the actuall code

@Ve

Icy Defiance

it's well, people. haskell works well.

superman works good.

jesus works good.

haskell works well.

Roy T.

@Vektorweg that second statement you made, that would be ideal. If a compiler is good enough to define optimal structures without any help than we should indeed abstract as far away as possible

user4704

@IcyDefiance I English well.

Roy T.

These are interesting discussion topics btw. I had not seen this level of discussions here yet. (/me butchers the English language even more).

snake5

4:38 PM

@IcyDefiance wasn't the joke about "does"?

Icy Defiance

originally, yes, but I had to adapt the joke to the situation

Vektorweg

@RoyT. sadly, we don't have SSC's yet. i mean, that someone could always write a lib that brings a abstraction of a machine and a language can provide a layer in which it seems nearly identical to the actual code to describe the problem. insofar, that the compiler and language supports such things.

snake5

►

@IcyDefiance I see

Roy T.

@Vektorweg SSC's? Semi Sentient Computers? That second thing you describe is basically what C# and Java do

snake5

@RoyT. sounds about right, considering the kind of beast that haskell is

Vektorweg

4:40 PM

@RoyT. maybe the same. i mean sufficiently smart compiler.

snake5

iirc, ghc had a 1gb installation for windows

MickLH

Haha actually we totally do have semi-sentient computers today

Roy T.

@Vektorweg haha, :P

MickLH

But even the humans are not quite "sufficiently" smart compilers

snake5

our humans?

Vektorweg

4:41 PM

@RoyT. Haskell do it too.

Icy Defiance

what is semi-sentient, exactly? wouldn't that just be...not sentient?

snake5

that's alien-speak

:D

Roy T.

@Vektorweg I thought Haskell compiled to C/C++?

MickLH

hahaha oops did I go backwards with my edit?

Roy T.

@Vektorweg I didn't knew they used a sort of VM and Jitter

user4704

4:42 PM

LLVM is basically built around that concept of building intermediate representations that can be more efficiently adapted to machine code.

snake5

@MickLH possibly :D

MickLH

@IcyDefiance Precisely... if you wanna be a precise jerk

AttackingHobo

0

Q: Move physics animation on timeline without resimulating?

I have a complicated physics simulation done, but I want move the entire simulation on the timeline. I can't figure out how to do this without having to resimulate everything. I don't want to change how the baked physics look, just want to have it play the same animation at a different time.

snake5

anyway, I think I should write a BSP tree for triangles or something

MickLH

@AttackingHobo hex editor

Get to work

AttackingHobo

4:44 PM

Big Sausage Pizza Tree

@MickLH whuh?

Roy T.

@snake5 usually yes :P

Vektorweg

implementing food instead of coding for food. yay.

MickLH

@snake5 invent this thing attackinghobo just described, organize triangles into a graph where some nodes can be justly referred to as sausages or pizzas, and that can be construed as big in some way

don't just ignore what you think you should do!

AttackingHobo

hahaha

snake5

wish I had time for that

writing a lightmap renderer that has to be fully finished at the end of this week

MickLH

4:47 PM

those are fun

What method will you use?

snake5

so fun that no one seems to have written one and published with a nice license

@MickLH probably the most simple for now - direct lighting

MickLH

Lol really?

snake5

that's the target now

if I have time, I can reconsider

MickLH

I never released mine because I was never satisfied with the UV packer

snake5

I won't do UV packing at all

MickLH

4:49 PM

I had to choose between wasting like 50% space, or having seams occasionally

snake5

it can be done in blender

I'm doing the simplest thing that works right now

AttackingHobo

@MickLH look into how blender does it's lightmap packing

snake5

one texture per mesh instance, with second texcoord set from mesh, manually created in blender

@AttackingHobo just not the "lightmap packing" but "smart uv packing" or something like that

the lightmap packer is actually awful because it splits up faces

not good for games, there's only one index array so that means vertex duplication

MickLH

I came up with a system though that worked for patching seams, and it was efficient but required a super complex spatial process that I spent like 50 hours doing calculus for on Mathematica

but in the end I only ever formulated the post-process in voxels instead, which means it's either unusably low res, or unusably slow

Roy T.

Anyone ever encountered this in the Visual Studio debugger? It looks like this object has two types i6.minus.com/jrhIslT9cfrjY.PNG

I know it is of class BinaryExpressionSyntax, but what is the second value? (Because I would really like to know if something is a SimpleAssignment)

And SimpleAssignmentExpression doesn't seem to be a type. VS cant find it

user4704

4:53 PM

Looks like a possible result of a debugger visualizer for the type.

user4704

For example: source.roslyn.codeplex.com/#Microsoft.CodeAnalysis.CSharp/…

Roy T.

That is exactly it. As you might've noticed that is roslyn I'm using there

I wonder if I can mimick how it determines that it is a SimpleAssignment

user4704

BinaryExpressionSyntax is a type, the SimpleAssignmentExpression is a value of some enum, and the debugger visualizer is formatting it thusly.

user4704

Looks like GreenNode (way up in the inheritance chain) exposes the RawKind value.

Roy T.

Ah I see now

Thanks :D

user4704

4:56 PM

re: retina iMac, "I still would like to see benchmarks of some games"

user4704

...I don't think those people realize that no sane person would play a game in native (5k) resolution.

user4704

Because no sane developer authors shipping content at appropriate resolutions. It would look like ass.

MickLH

God it hurts so much

The engine should be scaling down the original assets on the content-house side... on the fly to customers who's client requests it

Side effect of making piracy a horrible bitch

Transcript for