What are you working on? v67 - March 2017 - Developers

Digest auth is relatively simple but the certain resource I've been accessing turned up to be really nitpicky on sockets. I wasted my yesterday and half of this day, and now after I've finally solved the problem, I admire the degeneracy of these temporary commits [img]http://i.imgur.com/uBl1dbo.png[/img]

For anyone who wants to make simple 2D graphical things in C#, using this you can use System.Drawing in realtime for animations, simulations and visualisations.etc [url]http://richardbamford.io/WFG/[/url]

[QUOTE=suXin;52180411]Digest auth is relatively simple but the certain resource I've been accessing turned up to be really nitpicky on sockets. I wasted my yesterday and half of this day, and now after I've finally solved the problem, I admire the degeneracy of these temporary commits [img]http://i.imgur.com/uBl1dbo.png[/img][/QUOTE] This commit log seems like something out of a vinesauce stream, really.

[QUOTE=JohnnyOnFlame;52180554]This commit log seems like something out of a vinesauce stream, really.[/QUOTE] you guessed how bad I was influenced [editline]3rd May 2017[/editline] succ for succ(ess) though

[QUOTE=Bambo.;52180512]For anyone who wants to make simple 2D graphical things in C#, using this you can use System.Drawing in realtime for animations, simulations and visualisations.etc [url]http://richardbamford.io/WFG/[/url][/QUOTE] Does it still run as slow as System.Drawing itself?

[QUOTE=Perl;52180636]Does it still run as slow as System.Drawing itself?[/QUOTE] Naturally. It's just a wrapper that runs an update method, then a draw method, then swaps the back/front buffer then sleeps for 17 ms, thus guaranteeing the framerate remains under 60 at all times. Since Thread.Sleep only guarantees it will sleep at least how long you specify I'd be surprised if FPS goes over 50. That said it's not [I]that[/I] slow that you can't use it for basic stuff. [vid]https://www.helifreak.club/files/20170503123942252.mp4[/vid][vid]https://www.helifreak.club/files/20170503122830416.mp4[/vid][vid]https://www.helifreak.club/files/20170503122908930.mp4[/vid] 100, 1000, and 10000 rectangles rendering at about 60 fps with CPU usage on a 4.4 GHz 6600k. Biggest issue is that it flickers like a mother fucker even with double buffering on. These videos are without that class though.

Thanks to Ziks I have cables now! Though big there needs to be made big improvement and some optimisation. [IMG]http://i.imgur.com/RQzQ6iV.png[/IMG]

I'm looking into building a new computer, I have around a $1000 to $1200 budget, maybe more. Was wondering if the [URL="https://www.amazon.com/gp/product/B01KVZBNY0/ref=ox_sc_sfl_title_2?ie=UTF8&psc=1&smid=ATVPDKIKX0DER"]GTX 1070[/URL] would be ideal for developing in VR? Without having to worry about lag, or upgrading again anytime soon.

[QUOTE=sarge997;52181693]I'm looking into building a new computer, I have around a $1000 to $1200 budget, maybe more. Was wondering if the [URL="https://www.amazon.com/gp/product/B01KVZBNY0/ref=ox_sc_sfl_title_2?ie=UTF8&psc=1&smid=ATVPDKIKX0DER"]GTX 1070[/URL] would be ideal for developing in VR? Without having to worry about lag, or upgrading again anytime soon.[/QUOTE] I'd do some research to see if other people are able to use it as such - you [I]might[/I] be more than good with a 1070, but I know that Volta (the next architecture) might be releasing before the year is out, so I'd consider waiting for that if that's the case. TMSC just scored an order for chips using the Volta microarchitecture, but those are for supercomputing units so it'll be awhile until anything comes up for consumers I bet. [editline]edited[/editline] I'd say yes outright that the 1070 will be good enough, but if you're doing dev work you won't be near the level of optimization and such that I imagine even the shittiest VR titles have. Since that can be so important with graphics stuff (especially due to the sheer amount of data streaming that VR seems to require), I imagine that will be a significant factor when it comes to dev work for VR. maybe someone who actually does VR dev can pitch in lol

I have a 1070 in my laptop and I do VR projects. Also ran alot of VR games on ultra with not a single problem.

Made a tutorial about the light things in Unity. I was wondering if other people (like user [B]fewes[/B]) have other techniques. [video=youtube;IjP2MeSozIs]https://www.youtube.com/watch?v=IjP2MeSozIs[/video]

[QUOTE=moonmoonmoon;52182507]Made a tutorial about the light things in Unity. I was wondering if other people (like user [B]fewes[/B]) have other techniques.[/QUOTE] That's a great tutorial, I found myself giggling a lot. Both informative and entertaining.

[QUOTE=helifreak;52180734]Naturally. It's just a wrapper that runs an update method, then a draw method, then swaps the back/front buffer then sleeps for 17 ms, [B]thus guaranteeing the framerate remains under 60 at all times.[/B] [/QUOTE] Why would you do that

[QUOTE=foszor;52182733]That's a great tutorial, I found myself giggling a lot. Both informative and entertaining.[/QUOTE] Thank you! Nice to hear, I was afraid it'd be boring. If only I knew more stuff, I'd do it more often :happy:

[QUOTE=sarge997;52181693]I'm looking into building a new computer, I have around a $1000 to $1200 budget, maybe more. Was wondering if the [URL="https://www.amazon.com/gp/product/B01KVZBNY0/ref=ox_sc_sfl_title_2?ie=UTF8&psc=1&smid=ATVPDKIKX0DER"]GTX 1070[/URL] would be ideal for developing in VR? Without having to worry about lag, or upgrading again anytime soon.[/QUOTE] I'd say if you are holding off optimization until later you are doing VR dev wrong anyway so you can get away with pretty mid-range hardware even. I have a GTX 1070 and it is well adequate for VR development.

[QUOTE=suXin;52180411]Digest auth is relatively simple but the certain resource I've been accessing turned up to be really nitpicky on sockets. I wasted my yesterday and half of this day, and now after I've finally solved the problem, I admire the degeneracy of these temporary commits [img]http://i.imgur.com/uBl1dbo.png[/img][/QUOTE] git commit --amend?

Working on OpenGL school project. Just got spotlights to work. Next up is texturing and bump mapping. [t]http://i.imgur.com/oDV3oNr.png[/t]

[QUOTE=antianan;52185567]git commit --amend?[/QUOTE] An interactive rebase would probably be more applicable. [t]http://i.imgur.com/YP0A0lD.png[/t] In other news, a guy went out of his way to make my day. Twice! :dogcited: I love the programming community.

[QUOTE=antianan;52185567]git commit --amend?[/QUOTE] git rebase -i succ^7

[t]https://my.mixtape.moe/xdixuk.jpg[/t] Doing some late night prototyping for my thesis

[thumb]http://i.imgur.com/eQ3CTZ7.jpg[/thumb] Working on a multi-core atmel328 based microcontroller thingie. 64k RAM and programmed using a custom RISC assembler that's going to be interpreted on each core because the default harvard architecture doesn't really allow me to hack multiprocessing on top with external program RAM (i _could_ do it, but i prefer to not shorten the ROM lifetime) In the end i'm going to replace the 8 MHz crystal with 40 MHz, which in-itself is a ridiculous amount to overclock the MCUs, but hey, it works. If i pull it off correctly, the interpreted custom code should be faster than the baseline 16 MHz AVR code.

[QUOTE=antianan;52185567]git commit --amend?[/QUOTE] I tried to figure out the problem by making small changes, so [I]each [B]succ [/B]had it's purpose[/I] I [B]reset succ[/B] after I solved the problem

At some point, I'd like to be less good at blowing the display/graphics driver up. My vulkan renderer is [I]sooo[/I] close to working that it really just hurts at this point - because upon creating the graphics pipelines I'm trying to use in this first demo (there are three pipelines, each one being a different rendering "style" and they all render side-by-side in the same window) they all crash due to an access violation in the nvidia driver. Trying to fix this has involved me checking through most of the code now making up this project, along with building the Vulkan validation layers that LunarG made myself so I can use them in my project - letting me step through and debug the validation layers. Unfortunately, even after fixing 5-6 bugs in my code that would've been issues eventually AND using the aforementioned layers that I build I'm still crashing 3-4 calls deep into the nvidia driver. Yay. [t]http://i.imgur.com/IrSioUf.png[/t] [I]kill me[/I]. There's clearly an issue with [B]my[/B] code that's causing this crash, because I've mostly copied this from Sascha's vulkan examples and his shit works. Its just really irritating that the error is occuring where it is, because it doesn't give me anywhere to start on debugging things and I'm really not looking forward to going back to the very beginning (i.e instance creation) and working my way forward to pipeline creation. This will probably involve about half of the 4k lines of code I have in this project too, so that'll be a blast lol. Ultimately, this is going to be something stupid. I can sense it.

[QUOTE=paindoc;52188789]At some point, I'd like to be less good at blowing the display/graphics driver up. My vulkan renderer is [I]sooo[/I] close to working that it really just hurts at this point - because upon creating the graphics pipelines I'm trying to use in this first demo (there are three pipelines, each one being a different rendering "style" and they all render side-by-side in the same window) they all crash due to an access violation in the nvidia driver. Trying to fix this has involved me checking through most of the code now making up this project, along with building the Vulkan validation layers that LunarG made myself so I can use them in my project - letting me step through and debug the validation layers. Unfortunately, even after fixing 5-6 bugs in my code that would've been issues eventually AND using the aforementioned layers that I build I'm still crashing 3-4 calls deep into the nvidia driver. Yay. [t]http://i.imgur.com/IrSioUf.png[/t] [I]kill me[/I]. There's clearly an issue with [B]my[/B] code that's causing this crash, because I've mostly copied this from Sascha's vulkan examples and his shit works. Its just really irritating that the error is occuring where it is, because it doesn't give me anywhere to start on debugging things and I'm really not looking forward to going back to the very beginning (i.e instance creation) and working my way forward to pipeline creation. This will probably involve about half of the 4k lines of code I have in this project too, so that'll be a blast lol. Ultimately, this is going to be something stupid. I can sense it.[/QUOTE] When in doubt rewrite from scratch. It's like reinstalling Windows - every time you cringe when having to do it but are happy do have done afterwards. It's also totally not a bad idea to do it for a whole engine halfway into the project with a deadline one month away. I'm kidding. One thing that I think is seriously helpful though in avoiding introducing bugs and issues when abstracting code is to write a completely working use case first without introducing ANY abstraction at all and then from there start to compress step by step while always keeping the use case working. This is immensly helpful for me personally because I think it leads to more fitting abstractions. What often happens when writing a wrapper or something like that top-down instead of bottom-up like I described is that you forget to consider tiny details or quirks of the APIs you're using that you would have easily catched when compressing from the working use case because you would have stumbled upon them when writing that. That's also why I didn't like Sascha's examples too much as a tutorial resource tbh because they wrap everything away too much and decompose it into so many classes when all I want is to see the 1k lines of code I need to type to draw a triangle. Anyways, so you're not getting any feedback from the validation layers?

[QUOTE=JWki;52188838]When in doubt rewrite from scratch. It's like reinstalling Windows - every time you cringe when having to do it but are happy do have done afterwards. It's also totally not a bad idea to do it for a whole engine halfway into the project with a deadline one month away. I'm kidding. One thing that I think is seriously helpful though in avoiding introducing bugs and issues when abstracting code is to write a completely working use case first without introducing ANY abstraction at all and then from there start to compress step by step while always keeping the use case working. This is immensly helpful for me personally because I think it leads to more fitting abstractions. What often happens when writing a wrapper or something like that top-down instead of bottom-up like I described is that you forget to consider tiny details or quirks of the APIs you're using that you would have easily catched when compressing from the working use case because you would have stumbled upon them when writing that. That's also why I didn't like Sascha's examples too much as a tutorial resource tbh because they wrap everything away too much and decompose it into so many classes when all I want is to see the 1k lines of code I need to type to draw a triangle. Anyways, so you're not getting any feedback from the validation layers?[/QUOTE] No, and I enabled the full set of validation layers and used the nice technique from the examples I mentioned to get nicely formatted console output: [t]http://i.imgur.com/1h8xumv.png[/t] It crashes and then fails to output anything beyond that. Disabling the pipeline creation helped me find those other bugs because of the console output, but trying to create a graphics pipeline just breaks things. I've made sure to go back and check the rather insane amount of "state info" structs it needs too, and everything looks fine. I was initially not setting the Dynamic States and such correctly, but fixing that didn't help unfortunately. I need to go back and create unit tests tbh - something like using Boost unit testing library + solid test fixtures to test construction, destruction, and other various facilities of my various wrapper classes. That was a really nice thing to have at work, even if it was boring as hell to setup. Just feels like it'd make sense in this case though, since I'd know if changing some property of my logical device wrapper (for example) would break the resource classes or something. [editline]5th May 2017[/editline] I might try to enable the full API dump layer too, to see if I can't pick anything up from that.

I also started to code with vulkan 2 weeks ago... atm pipeline creation is the current step to do. But so far everything works like it should!

[IMG]https://i1.wp.com/floatlands.net/wp-content/uploads/2017/05/electric_system_build_floatlands.png[/IMG] A little update.

Are people interested in a small library I wrote for work? It's not really usable without an account for our service. We spent the last year building a massive api and renewing almost our entire back-end. I just spent finishing the javascript library and I'm wondering what other people think of the code because I normally don't write javascript. Some example syntax: [code] // Obtained client id var clientId = 1; // Callback url is set to the current url by default var auth = new maps4news.ImplicitFlow(clientId); var api = new maps4news.Maps4News(auth); // This will hijack the page if no authentication cache can // be found. Smartest thing to do is to just let it happen // and initialize any other code afterwards. api.authenticate().then(function() { // Get the current user and dump the result to the console. api.users.get('me').then(console.dir); }); [/code]

[QUOTE=Mega1mpact;52189339][...] I just spent finishing the javascript library and I'm wondering what other people think of the code because I normally don't write javascript. [...][/QUOTE] Afaict it looks good, assuming that's actually a Promise(-compatible) return value. Have you considered providing TypeScript typings? [URL="https://github.com/Microsoft/dts-gen"]You may be able to partially auto-generate it[/URL][URL="http://archive.is/8PFLx"].[/URL] Is there a serious reason for the hijacking though? It seems like you'd normally just fail the Promise there. [editline]5th May 2017[/editline] Weird. I set my VPN to use a German server but flagdog says Russia :thinking: That doesn't exactly raise my confidence in the service.

Thinking of writing a chrome plugin for work. We do things with computer-telephone-integration and sometimes we have to make a real call via phone just to test frontend menus, and that gets annoying. Of course, we should test with actual calls but when we're just testing a menu option, we can skip that part. I want to write a little popup menu that I can use to trigger different types of UI events more quickly. I'd have to do this in my freetime though [editline]5th May 2017[/editline] I think I'll call it "blue box"