What are you working on? v66 - February 2017

[Zelpa]

[QUOTE=phygon;51808385]Done with the dithering
[IMG]https://my.mixtape.moe/gcxyuz.gif[/IMG][/QUOTE]

i don't think the dithering is actually doing much for your art style, the lighting is so simple that the dithering just looks like a strange overlay. maybe try making it less cel shaded?

[Phygon]

[QUOTE=Zelpa;51808430]i don't think the dithering is actually doing much for your art style, the lighting is so simple that the dithering just looks like a strange overlay. maybe try making it less cel shaded?[/QUOTE]

The artist that made the model requested that I make the shader for him
I also already wanted to make one so it was a win win
The second video hopefully shows it off better

[cartman300]

[QUOTE=JWki;51808423]Interesting, were you developing on a machine with an AMD card before by chance?
If not, that's weird, my 660M tended to replicate what I got on a desktop Nvidia quite consistently but I guess you never know with GPUs.
If yes, it's not surprising.[/QUOTE]

I did, and why is it not surprising? It always worked fine (i tend to test on many different gpus), it even works fine on my friends nVidia gpu.

[jwki]

[QUOTE=cartman300;51808442]I did, and why is it not surprising? It always worked fine (i tend to test on many different gpus), it even works fine on my friends nVidia gpu.[/QUOTE]

In that case it still is surprising. Assuming you were only testing on the AMD, or only AMDs in general, it wouldn't be that surprising because there's ALWAYS something that the two do differently.
Might just be a bug in the specific driver version then.

[Jaex]

Made very simple addon for Firefox too:

[img]https://dl.dropboxusercontent.com/u/14076298/ShareX/2017/02/ShareX_XSn71at8Js.png[/img]

Which only adds right click upload button:

[img]https://dl.dropboxusercontent.com/u/14076298/ShareX/2017/02/firefox_nP8uEMg6hp.png[/img]

I would like to have full page capture support in addon too but because I'm not web developer it is beyond my knowledge.

[cartman300]

Okay i may be pissed and biased above any possible means of measure right now, but this is the first time and [B]THE LAST TIME i buy an acer laptop and an nVidia graphics card. Ever.[/B]

[thumb]http://carp.tk/$/explorer_2017-02-11_16-50-05.png[/thumb]

[editline]11th February 2017[/editline]

This is possibly the most user unfriendly software i EVER ENCOUNTERED.

[editline]11th February 2017[/editline]

I hate computers.

[NoRegard]

Explain?

[Ziks]

[QUOTE=cartman300;51808739]Okay i may be pissed and biased above any possible means of measure right now, but this is the first time and [B]THE LAST TIME i buy an acer laptop and an nVidia graphics card. Ever.[/B]

[thumb]http://carp.tk/$/explorer_2017-02-11_16-50-05.png[/thumb]

[editline]11th February 2017[/editline]

This is possibly the most user unfriendly software i EVER ENCOUNTERED.

[editline]11th February 2017[/editline]

I hate computers.[/QUOTE]

I had some weird artefacts too on an acer laptop like 4-5 years ago. It rendered fine on my AMD desktop, but it looked like it would start a new vertex at a different point on the laptop's nVidia card. So some attributes like tex coord or vertex color for each vertex looked like it should have been for the previous vertex.

[editline]11th February 2017[/editline]

[QUOTE=war_man333;51808875]Explain?[/QUOTE]

GeForce Experience says a more recent version of the driver is installed compared to Programs and Features.

[cartman300]

[QUOTE=Ziks;51808879]I had some weird artefacts too on an acer laptop like 4-5 years ago. It rendered fine on my AMD desktop, but it looked like it would start a new vertex at a different point on the laptop's nVidia card. So some attributes like tex coord or vertex color for each vertex looked like it should have been for the previous vertex.

[editline]11th February 2017[/editline]

GeForce Experience says a more recent version of the driver is installed compared to Programs and Features.[/QUOTE]

It's not just about my OpenGL stuffs.

I uninstalled the latest version, installed 375.70. Ran my program once, it said it was running on 375.70, then restarted it, and it was suddenly running 376.54. Uninstalled everything and then suddenly again it reinstalled the newest version of the graphics driver. I uninstalled that and disabled automatic driver downloads from the internet but THAT in turn disabled installing all drivers ALLTOGETHER (similar shit to when you disable firewall, it disables font installation on windows, ?????), rebooted my computer and suddenly i can't install nVidia drivers at all (that is after enabling automatic driver downloads).

[editline]11th February 2017[/editline]

Very nice experience so far.

[Lily]

[QUOTE=Dr Magnusson;51807499]Can you explain what you mean when you say that C++ doesn't have aligned storage and allocation?[/QUOTE]

It's actually a bit worse than you'd expect. Until C++17 it was basically just not supported to get a useful alignment on anything unless you allocated the memory yourself. They added alignas() and such with C+11... except they didn't think to actually make the stdlib required to respect it, so if your fancy alignas()'d types touches any std container your code will catch fire.

Oh and there's a compiler bug on MSVC where over-aligned objects can't be passed by value. Are we having fun yet?

[editline]11th February 2017[/editline]

If you think Nvidia is bad try getting 64-bit OpenGL working consistently across different AMD cards; I had to resort to using detours.

[20k]

[QUOTE=Tobba;51809640]
Oh and there's a compiler bug on MSVC where over-aligned objects can't be passed by value. Are we having fun yet?[/QUOTE]

Worst part is that this is won't fix due to the compiler hating fun. They just basically dont support this feature

Also I've had no issues with AMD OpenGL myself, although my use is limited to framebuffers/render buffers and cl/gl interop (+ sfml). Nvidias OpenCL implementation is a bit hitler, and used to be literally hitler though. AMD have really irritatingly lax cl/gl interop too, whereas on nvidia its a driver crash

[Lily]

[QUOTE=Icedshot;51810458]Worst part is that this is won't fix due to the compiler hating fun. They just basically dont support this feature

Also I've had no issues with AMD OpenGL myself, although my use is limited to framebuffers/render buffers and cl/gl interop (+ sfml). Nvidias OpenCL implementation is a bit hitler, and used to be literally hitler though. AMD have really irritatingly lax cl/gl interop too, whereas on nvidia its a driver crash[/QUOTE]

So far my favorite AMD OpenGL bullshit is that you can't create multiple (sharing) contexts with the same DC.

Well, you can, but it'll [i]very rarely[/i] crash without a useful trace if you bind / unbind the contexts at the wrong time.

[Protocol7]

[QUOTE=phygon;51808385][IMG]https://my.mixtape.moe/ffnmwu.gif[/IMG][/QUOTE]

Holy shit this is art. There's a total 90's benchmark/tech demo vibe going on.

[gayNightOwl]

Finally got around to figuring out how GitHub worked, and uploaded my basic [URL="https://github.com/amadeusX2OO/Bare-XML-Parser"]XML parser[/URL].

If in C++ you have polymorphic classes (has vtable) with virtual functions written in line, presumably they won't ever actually be in-lined as a function pointer needs to exist? I'm afraid I've come from a Python/C# beginning and am not used to separating interface and implementation in source. Even with the assistance of VS, I still find myself writing everything in-line then later separating the code into appropriate .h and .cpp files :/

[jwki]

[QUOTE=gayNightOwl;51811896]Finally got around to figuring out how GitHub worked, and uploaded my basic [URL="https://github.com/amadeusX2OO/Bare-XML-Parser"]XML parser[/URL].

If in C++ you have polymorphic classes (has vtable) with virtual functions written in line, presumably they won't ever actually be in-lined as a function pointer needs to exist? I'm afraid I've come from a Python/C# beginning and am not used to separating interface and implementation in source. Even with the assistance of VS, I still find myself writing everything in-line then later separating the code into appropriate .h and .cpp files :/[/QUOTE]

Virtual dispatch happens at runtime so those calls won't be inlined, no.

[SKS]

// | 4 hp | 4 dir | 8 anim | 8 player | 8 ping | 32 buff |
private byte hp;            // 4 bits, hp of user
private byte direction;     // 4 bits, < 0 for left, 0 for right
private byte animState;     // 8 bits, animation state of the player
private byte playerState;   // 8 bits, state of the player
private byte ping;          // 8 bits, ping of the user
private Int32 buffMask;     // 32 bits, bitmask of buffs

// | 32 xpos | 32 ypos |
// positions as representations of normalized location
// on play field; (2147483647, 2147483647) would be the
// top-right corner of the play field
private Int32 xPos;         
private Int32 yPos;

Found out Firebase stores and transmits all numerals as Int64's, so I managed to cram my player properties into 2 of them.

Trying to get lobby sizes of 50 on mobile is hard

[WTF Nuke]

[QUOTE=JWki;51811910]Virtual dispatch happens at runtime so those calls won't be inlined, no.[/QUOTE]

While true, compilers can devirtualize calls/classes (especially with the help of final classes) meaning they become open to inlining and other optimizations.

[SKS]

[QUOTE=Radical_ed;51812062]

// | 4 hp | 4 dir | 8 anim | 8 player | 8 ping | 32 buff |
private byte hp;            // 4 bits, hp of user
private byte direction;     // 4 bits, < 0 for left, 0 for right
private byte animState;     // 8 bits, animation state of the player
private byte playerState;   // 8 bits, state of the player
private byte ping;          // 8 bits, ping of the user
private Int32 buffMask;     // 32 bits, bitmask of buffs

// | 32 xpos | 32 ypos |
// positions as representations of normalized location
// on play field; (2147483647, 2147483647) would be the
// top-right corner of the play field
private Int32 xPos;         
private Int32 yPos;

Found out Firebase stores and transmits all numerals as Int64's, so I managed to cram my player properties into 2 of them.

Trying to get lobby sizes of 50 on mobile is hard[/QUOTE]

[img]http://i.imgur.com/RA44jY9.gif[/img]

Working on a game to demo the speed of my Firebase plugin

(that's the cloud database on the right)

[Lily]

Also if you're doing anything with OpenGL or DX11+, [url=https://renderdoc.org]RenderDoc[/url] is basically a must. The author is a super nice guy as well if you manage to break it horribly like I did :v: (I think I may have one of the more complex multithreaded OpenGL applications to have been thrown at this thing).

[paindoc]

[QUOTE=Tobba;51812977]Also if you're doing anything with OpenGL or DX11+, [url=https://renderdoc.org]RenderDoc[/url] is basically a must. The author is a super nice guy as well if you manage to break it horribly like I did :v: (I think I may have one of the more complex multithreaded OpenGL applications to have been thrown at this thing).[/QUOTE]

Wtf are you making. I'm curious, because you keep breaking things. RenderDoc is great though, I really like the mesh viewer. It's been extremely useful more times than I can count

There's really not a lot of documentation on OpenGL/CUDA interop, and all the existing samples insist upon using immediate mode and the pixel buffer. Apparently, you can just bind CUDA to render buffers or textures, but this seems to be new to the latest version of CUDA and I can't really find any example of it

[Karmah]

[QUOTE=Tobba;51812977]Also if you're doing anything with OpenGL or DX11+, [url=https://renderdoc.org]RenderDoc[/url] is basically a must. The author is a super nice guy as well if you manage to break it horribly like I did :v: (I think I may have one of the more complex multithreaded OpenGL applications to have been thrown at this thing).[/QUOTE]

I've never tried RenderDoc before, but some alternatives for OpenGL debugging also includes gDebugger, codeXL (works for both nVidia and AMD), and NSight (nvidia only iirc)

[paindoc]

[QUOTE=gayNightOwl;51811896]Finally got around to figuring out how GitHub worked, and uploaded my basic [URL="https://github.com/amadeusX2OO/Bare-XML-Parser"]XML parser[/URL].

If in C++ you have polymorphic classes (has vtable) with virtual functions written in line, presumably they won't ever actually be in-lined as a function pointer needs to exist? I'm afraid I've come from a Python/C# beginning and am not used to separating interface and implementation in source. Even with the assistance of VS, I still find myself writing everything in-line then later separating the code into appropriate .h and .cpp files :/[/QUOTE]

The best argument for separating the two is compile time, in my experience. My project at work is getting huge, and some of the files I wrote long ago have most of the implementation in the header. A good chunk of these are classes I use all over my program, and changing these means the next compile is a lot slower.

Otherwise, when it comes to inlining, the compiler should be smart and usually overrides our choices. There is the "__forceinline" tag in MSVC, so you could play with that and see what happens with your virtual classes and such? I've only used it for intrinsic operations.

[Lily]

[QUOTE=Karmah;51813218]I've never tried RenderDoc before, but some alternatives for OpenGL debugging also includes gDebugger, codeXL (works for both nVidia and AMD), and NSight (nvidia only iirc)[/QUOTE]

gDEBugger is discontinued and doesn't even support OpenGL 3.3, CodeXL is AMD-grade bad and Nsight is, as you said, nvidia only.

I got CodeXL to display a frame trace exactly once, it crashed as soon as I touched the UI :v:. On AMD you also have PerfStudio which can display internal performance counters and do frame debugging, it can be extremely useful but is similarily temperemental. It does work though if you learn when you can touch what buttons without crashing, I personally prefer to just not touch it at all.

RenderDoc just had some issues with unbinding / rebinding GL contexts on different threads breaking things (or calling SwapBuffers with no context bound). Fixed that up and it works just fine, minus that I have to detect RenderDoc capturing and stop buffer uploads from occuring mid-frame on a different thread, still investigating that one.

[editline]12th February 2017[/editline]

[QUOTE=paindoc;51813215]Wtf are you making. I'm curious, because you keep breaking things. RenderDoc is great though, I really like the mesh viewer. It's been extremely useful more times than I can count [/QUOTE]

Just a game / graphics engine. Just not your regular 'ole one, but instead using enough modern C++ and design practices to keep MSVC and just about everything else within 3 inches of an internal error at all times. That includes half a functional programming language inside of some template soup.

[DrTaxi]

[QUOTE=Tobba;51813266]Just a game / graphics engine. Just not your regular 'ole one, but instead using enough modern C++ and design practices to keep MSVC and just about everything else within 3 inches of an internal error at all times. That includes half a functional programming language inside of some template soup.[/QUOTE]

[I]Cue Greenspun's tenth rule[/I]

[QTom]

I decided to release the little app I made on android, it's pretty basic but it was an interesting exercise to learn how it all works. Was nice to try out xamarin but honestly not sure if I'd use it again, if I make another app I may try using the native SDK.

Am I allowed to post a link? It has ads so don't know if I'm allowed to, but would fun to get a few people to give it a try.

[cartman300]

[thumb]http://carp.tk/$/LibTech_2017-02-12_20-50-22.png[/thumb]

[thumb]http://carp.tk/$/devenv_2017-02-12_20-53-04.png[/thumb]

Made NanoVG .NET binding, it mixes nicely with OpenTK and SDL and whatever you're using already.

[Ziks]

Got some basic texturing working:

[t]http://files.facepunch.com/ziks/2017/February/12/chrome_2017-02-12_22-18-33.png[/t]

Haven't done fetching textures packed inside the .bsp or displacement blending yet.

[editline]12th February 2017[/editline]

[media]https://www.youtube.com/watch?v=1_N1vtXvFCo[/media]

[SKS]

Reminder that we have a discord [url]https://discord.gg/m3TBvMM[/url]

[Phygon]

[QUOTE=Tobba;51812977]Also if you're doing anything with OpenGL or DX11+, [url=https://renderdoc.org]RenderDoc[/url] is basically a must. The author is a super nice guy as well if you manage to break it horribly like I did :v: (I think I may have one of the more complex multithreaded OpenGL applications to have been thrown at this thing).[/QUOTE]

RenderDoc got me through programming my portals. I [I]HIGHLY[/I] recommend it. Unity even has native support.

[Darkwater]

[QUOTE=Radical_ed;51814530]Reminder that we have a discord [url]https://discord.gg/m3TBvMM[/url][/QUOTE]

why can't we have one facepunch discord with waywo/arma etc channels