People
Repo info
Activity
  • 22:14
    termhn opened #2611
  • 15:04
    kvark commented #2398
  • 14:07
    Michael-Lfx commented #2398
  • 12:08
    Michael-Lfx closed #2489
  • Feb 02 08:35
    grovesNL commented #2554
  • Feb 01 02:46
    kvark commented #2610
  • Feb 01 02:46
    kvark labeled #2610
  • Feb 01 02:46
    kvark labeled #2610
  • Feb 01 02:46
    kvark labeled #2610
  • Feb 01 02:46
    kvark labeled #2610
  • Feb 01 02:46
    kvark labeled #2610
  • Feb 01 02:46
    kvark labeled #2610
  • Jan 31 22:25
    mtak- opened #2610
  • Jan 31 14:43
    null77 commented #2564
  • Jan 31 12:07
    omni-viral commented #2600
  • Jan 31 05:36
    Travis gfx-rs/gfx (master) passed (6343)
  • Jan 31 05:04
    bors[bot] closed #2600
  • Jan 31 05:04
    bors[bot] closed #2564
  • Jan 31 05:04
    bors[bot] commented #2600
  • Jan 31 05:04

    bors[bot] on master

    [mtl] secondary command buffers Merge #2600 2600: [mtl] second… (compare)

Zakarum
@omni-viral
Device::create_framebuffer takes Extent which is misleading
as VkFramebufferCreateInfo has layers field, which is not the same as depth axis of Extent.
Dzmitry Malyshau
@kvark
Should be Extent2d then?
Zakarum
@omni-viral
Yes. And layers: gfx_hal::image::Layer
Why gfx_hal::image::Layer is u16?
VkImageCreateInfo has uint32_t arrayLayers;
Is it something constrained by other backends?
Dzmitry Malyshau
@kvark
It's just unusual to have so many layers. There is a small benefit of having more distinct type aliases: more likely to bump into error when messing them up.
VulkanDB says that no hardware supports more than 2048 layers
Matthew Russo
@mcr431
Hey all, sorry in advance if this is a stupid question but is there any support (and if so documentation) for gfx + openvr? If not is it being worked on?
cause if it is, I would try and help out where possible
Dzmitry Malyshau
@kvark
There was some VR work early on, like 3 years ago, but all of it perished,
Matthew Russo
@mcr431
so its nowhere on the roadmap? do you think its something that would be of interest to any gfx users?
Dzmitry Malyshau
@kvark
It would be nice to have! Personally, the roadmap is already dense enough for me. Contributions are welcome!
Matthew Russo
@mcr431
I figured but wanted to check and see if there was any work already planned. Honestly not sure if I'll get to it immediately. Currently working with Vulkano so probably going to stick with that and get vr stuff working and then maybe come help here if I'm satisfied with it. Doesn't seem worth it to rewrite what i already if I don't have a solid plan/design yet.
Dzmitry Malyshau
@kvark
I see. Good luck!
Lokathor
@Lokathor
google/shaderc-rs#34 so my earlier concern about shaderc-rs forcing you to ship binaries is unfounded, it just builds those binaries on accident
which is good news!
Markus
@markus-wa

Hi all!
I just cloned the repo and tried $ cargo run --bin quad --features vulkan --verbose - it gives me the following error:

   Compiling hal-examples v0.1.0 (C:\Users\mwalt\dev\gfx\examples)
     Running `rustc --edition=2018 --crate-name quad 'examples\quad/main.rs' --color always --crate-type bin --emit=dep-info,link -C debuginfo=2 --cfg 'feature="default"' --cfg 'feature="gfx-backend-vulkan"' --cfg 'feature="vulkan"' -C metadata=b6179fed2209fe7b -C extra-filename=-b6179fed2209fe7b --out-dir 'C:\Users\mwalt\dev\gfx\target\debug\deps' -C 'incremental=C:\Users\mwalt\dev\gfx\target\debug\incremental' -L 'dependency=C:\Users\mwalt\dev\gfx\target\debug\deps' --extern 'env_logger=C:\Users\mwalt\dev\gfx\target\debug\deps\libenv_logger-64003ce0ec63ea3b.rlib' --extern 'gfx_backend_empty=C:\Users\mwalt\dev\gfx\target\debug\deps\libgfx_backend_empty-01f4a3def68183ed.rlib' --extern 'gfx_backend_vulkan=C:\Users\mwalt\dev\gfx\target\debug\deps\libgfx_backend_vulkan-96f93b9ab55768c8.rlib' --extern 'gfx_hal=C:\Users\mwalt\dev\gfx\target\debug\deps\libgfx_hal-46cbcf794f406685.rlib' --extern 'glsl_to_spirv=C:\Users\mwalt\dev\gfx\target\debug\deps\libglsl_to_spirv-a09bcabef87f62ac.rlib' --extern 'image=C:\Users\mwalt\dev\gfx\target\debug\deps\libimage-d1b5071d4b8372ef.rlib' --extern 'log=C:\Users\mwalt\dev\gfx\target\debug\deps\liblog-ec5705c0a72c1882.rlib' --extern 'winit=C:\Users\mwalt\dev\gfx\target\debug\deps\libwinit-a1b84862cfe645f0.rlib' -L 'native=C:\Users\mwalt\scoop\persist\rustup\.cargo\registry\src\github.com-1ecc6299db9ec823\winapi-x86_64-pc-windows-gnu-0.4.0\lib'`
error: Edition 2018 is unstable and only available for nightly builds of rustc.

Rust version:

$ rustc --version
rustc 1.32.0 (9fda7c223 2019-01-16)
any clue what I'm doing wrong?
Lokathor
@Lokathor
could cargo be finding a different copy of rustc than bare rustc
try cargo rustc -- --version
Markus
@markus-wa
I just got it to work! -- I had to re-install rust, I think something went wrong when updating it earlier today
thanks anyway
Dzmitry Malyshau
@kvark
Ok good
William Roe
@wjlroe
Does gfx have an equivalent to OpenGL's PushDebugGroup for distinguishing different draw commands in something like RenderDoc? Calling PushDebugGroup myself doesn't really work as all the draw calls get pushed at once with Encoder::flush...
Dzmitry Malyshau
@kvark
One of the pinned issues (#1508) we have is about this. Not implemented yet.
Kevin R. S.
@BourgondAries
what does pre-ll stand for?
WaDelma
@WaDelma
pre-low level, I think.
It's gfx before whole rewrite to make it's API more like vulkan
Kevin R. S.
@BourgondAries
Ok @WaDelma (I thought lexical lifetimes at first)
So consuming gfx-rs should be equivalent to consuming gfx-hal now? Like, I don't need to specifically use the gfx-hal crate?
Gray Olson
@termhn
No. gfx is completely separate from gfx-hal
gfx crate = pre-ll branch on git
gfx-hal crate = master branch on git
Kevin R. S.
@BourgondAries
@termhn I'm confused because there's so much, also https://github.com/gfx-rs/gfx says it uses HAL as a component
Gray Olson
@termhn
@BourgondAries https://github.com/gfx-rs/gfx master branch = new gfx, equivalent to gfx-hal crate
https://github.com/gfx-rs/gfx pre-ll branch = old gfx, equivalent to gfx crate (latest version 0.17)
Kevin R. S.
@BourgondAries
Ah ok. It's just really confusing when published crate names differ from names used on github
Kevin R. S.
@BourgondAries
Trying to use gfx_backend_gl with gfx-hal, I get that Instance is not in back, as from this tutorial https://lokathor.github.io/learn-gfx-hal/03_clear_the_window.html
Gray Olson
@termhn
@BourgondAries gfx_backend_gl requires some extra to get working and there's a fair bit of missing functionality, you can see the examples in the main gfx repo for how to get it working if you like. It's probably easier and better to just use metal/dx12/vulkan depending on your system though
Kevin R. S.
@BourgondAries
@termhn unfortunately metal/dx12 I can't use (on arch), vulkan isn't supported either
Gray Olson
@termhn
'vulkan isn't supported either' by your gpu?
Kevin R. S.
@BourgondAries
Seems like it, have downloaded and tried various drivers
Gray Olson
@termhn
hm, do you know which gpu you have? using gfx-hal with backend != vulkan | metal | dx12 is a bit of a subpar experience currently, if you only have opengl support it might be better to use old gfx
Kevin R. S.
@BourgondAries
@termhn product: GF108 [GeForce GT 730]
Gray Olson
@termhn
@BourgondAries unfortunately looks like that is not supported for vulkan https://developer.nvidia.com/vulkan-driver
Kevin R. S.
@BourgondAries
@termhn Also has the crate gfx-backend-gl been published recently? A return value has been removed and the example only reflects this in the repo, but the published crate still returns 4 parameters
Gray Olson
@termhn
not sure to be honest. I mean, it was published relatively recently with the release of hal 0.1, but it's certainly possible there's been development since that time on the gl backene
Josh Groves
@grovesNL
@BourgondAries surface creation is a bit different for the GL backend (for now) - you'd have to change clear_the_window to do something more like the gfx quad example: https://github.com/gfx-rs/gfx/blob/master/examples/quad/main.rs#L94
Kevin R. S.
@BourgondAries
Thanks, @grovesNL @termhn, I've gotten it to work :)
Gray Olson
@termhn
Nice!