build(windows): no WebGPU renderer: the Windows game builds the Vulkan back end only, as Linux does #577

Merged

HumanBean merged 1 commit from HumanBean/Longshot:build/no-webgpu into main 2026-11-15 18:24:32 +00:00

Conversation 5 Commits 1 Files changed 55 +205 -23727

HumanBean commented 2026-11-14 20:20:51 +00:00

Owner

• Copy link
• Copy Markdown

Type: build · Risk: medium · Tested on: nothing run (Linux PC, Windows change) · Needs: Windows, any GPU

Why

Every Windows build compiles an old renderer the game doesn't use by default (TOTKPort's WebGPU renderer and suyu's
shader recompiler with its WGSL back end, about 160 C++ files at -O2) and downloads wgpu-native for it, so every
player's first build is longer than it needs to be. The game only drew with it when RX_BACKEND named another renderer
than Vulkan, or when the Vulkan back end didn't start. The owner decided (2026-11-14) to drop it: Linux already builds
without it. See #362 (a shorter first build).

What changed

The Windows game is now built as the Linux one is: the Vulkan back end (longshot_vk.dll) is the only renderer, the
exe is linked with the stand-ins that were already there for a build without wgpu-native (host/rx_nowgpu.c), and
wgpu-native is no longer a build tool. The WebGPU renderer's own sources are deleted; the headers the host and the
Vulkan back end use stay.

• build: nx2/build.py loses gpu_sources(), WGPU/WGPU_NATIVE, the "the renderer (C++20, NDEBUG...)" job branch,
  the wgpu_native.dll.lib link and the wgpu_native.dll copy; host/rx_nowgpu.c is always in the Windows host
  sources (Linux keeps host/posix/linux_stubs.c).
• host: rx_start() in host/rx_backend.c no longer reads RX_BACKEND: it always loads the Vulkan back end, as it did
  with RX_BACKEND unset (the default). load_vulkan()'s failure lines say "nothing is drawn" instead of "WebGPU".
  host/rx_nowgpu.c gains rx_collect_written() (a build with RX_SOFT_WRITE_WATCH=0 needed it from
  gpu/render/rx_core.cpp; it now answers "nothing tracked", as Linux's does) and its messages are reworded.
  host/display.c logs "no renderer" instead of "no WebGPU device". Comments in nn_hid.c, nvn_native.c,
  hos_main.c, nn_audio.c no longer point at the deleted files.
• gpu/render: deleted the 16 rx_*.cpp, rx_none.c (built by nothing already), and the headers only they used
  (rx_internal.h, rx_web_mt.h, rx_bcn.h, rx_wgpu_names.h, renderdoc_app.h). Kept rx.h (host/display.c),
  rx_nvn.h and nvn_api.h (the op stream: host/, gpu/vk/vk_api.h, tests, the AppImage's header copy). The top
  comment of rx.h says what it is now.
• tools: tools.lock loses the wgpu-native line, scripts/tools.py REQUIRED["windows-x64"] loses it, the builder
  (builder/LongshotBuilder/Runner.cs) no longer sets WGPU_NATIVE, builder/setup/setup.c comment. nx2/make_game.py
  drops wgpu_native.dll from BESIDE and removes one an older build left in the game folder (nothing loads it now).
• scripts: scripts/run-vk.ps1 no longer sets RX_BACKEND=vulkan.
• ci: the third-party objects steps (.forgejo/workflows/build-windows.yml and both in .circleci/shared.yml) no longer
  install wgpu-native or set WGPU_NATIVE.
• tests: new tests/test_no_webgpu.py (the build compiles nothing of gpu/render, names no wgpu-native, no tool list
  requires it, rx_start reads no RX_BACKEND, and both platforms' stand-ins define every wgr_* and live-input
  function the host names, so the exe links). tests/test_renderer_load_error.py now also checks the Windows failure
  line names no WebGPU. tests/test_dev_packs.py: a route check's label and path (gpu/render/x.h: still "game").
  tests/test_capture_past_end.py no longer lists the deleted gpu/render/rx_shader.cpp among the capture loops it reads.
• docs: docs/development.md (the RX_BACKEND row and WGPU_NATIVE gone, the gpu/render/ layout row, the live
  input's !dump/!shots), README.md and CONTRIBUTING.md (the tools\wgpu-native step), docs/design.md (the
  current design; the history stays), docs/releases.md, RELEASING.md (a note on the dated "Done" line),
  gpu/PROVENANCE.md (what was removed and what stays), gpu/vk/VULKAN.md's top note, gpu/vk/vk_api.h's comments.

Left on purpose: gpu/third_party/suyu/ is untouched (its shader recompiler and WGSL back end stay in the tree,
uncompiled; the Vulkan back end still compiles its texture decoders). third_party/ is untouched. Historical comments
in gpu/vk that say a piece was ported "from the WebGPU renderer's rx_*.cpp" are left as history, and gpu/vk/vk_state.h
and vk_shader.cpp are not touched at all (they are in the shader stamp: a comment change there would make every player
translate every program again). scripts/cpu_categories.py's render/webgpu-fallback row is left: it matches nothing in
a new build and still reads older profiles.

Measurements

Verdict: not measured: this PC is Linux, and the point of the change (a shorter Windows build, ~160 fewer C++
files) has to be timed on a Windows build. Requested under Must be tested.

Flags and settings

Two switches removed; nothing added.

Flag	What it is	Before this PR	After this PR	Effect	Helps / doesn't help / untested	Recommendation
RX_BACKEND (removed)	Picked the renderer	Unset or vulkan: Vulkan; anything else: the WebGPU renderer, in a build with wgpu-native	Not read: always Vulkan	None at the default	Untested on Windows	Removed: Vulkan is the only renderer
WGPU_NATIVE (removed, build)	The wgpu-native folder the build compiled the WebGPU renderer against	Set by the builder from its tools	Not read	The WebGPU renderer is not built	Untested on Windows	Removed

Testing

Tested: nothing run on this PC (the owner's rule for this change: no tests or builds here; CI runs the game-free
checks). Read instead: every user of each deleted file (grep across the tree: only the WebGPU renderer itself, and
comments), every check under tests/ that reads host/rx_backend.c, nx2/build.py, tools.lock or scripts/tools.py.

Not tested: the Windows build and the game on Windows; the Linux build (its sources change only in comments and
in host/rx_backend.c's rx_start and log lines); python -I -B tests/run_all.py (left to CI).

Must be tested:

• Windows, any GPU (I'm on Linux): a full build from the Builder or make_game.py, expecting it to build without
  wgpu-native and with no gpu_* objects, the game to start on Vulkan (rx: renderer: Vulkan (longshot_vk.dll) in
  longshot_log.txt) and play as on main, and the compile to be shorter than main's at the same commit otherwise
  (by hand or the Test Farm: request to open)
• Windows, a folder built by main: build this PR over it, expecting wgpu_native.dll gone from the folder and the game
  starting (by hand)

Risks

What could break: the Windows exe could fail to link if something still needs a WebGPU renderer function that
neither host/rx_nowgpu.c nor the host defines (tests/test_no_webgpu.py checks the ones the host names). A Windows
PC whose Vulkan doesn't start used to fall back to the WebGPU renderer (on D3D12 through wgpu-native); it now gets a
black window and "nothing is drawn" in the log, as Linux does. With default settings and a working Vulkan nothing
changes.

Turning it off: revert the commit.

Compatibility: saves, longshot_options.ini and the shader and pipeline caches are untouched (no stamped file
changed). The vk ABI is unchanged (vk_api.h: comments only, no MKWVK_API_VERSION bump). The builder no longer
downloads wgpu-native; an older game folder's wgpu_native.dll is removed at the next build. RX_BACKEND set to
something other than vulkan now gets Vulkan.

Issues

See #362

Checklist

Everyone

• One topic.
• The title is a Conventional Commit (type(scope): description) of the status line's type.
• Why gives the reason plainly and links the issues that fit (the tracker searched).
• Measurements has a verdict, or one line saying why there is nothing to measure.
• Testing names what was tested, what wasn't, and what must still be tested, where and how.
• Rebased on current main.
• python -I -B tests/run_all.py passes (any failure that also fails on main is named). (not run here: left to CI)
• No game files, dumps, build output or generated files.
• third_party/ untouched (a new folder with a PROVENANCE.md instead).
• The license is OK: GPL-3.0-or-later, nothing incompatible and nothing decompiled.

When it applies

• AI assistance recorded with Assisted-by: / Co-Authored-By: trailers.
• Numbers from the same PC, scene and settings, at least 2 interleaved runs each, the median reported. (not measured here: a Windows build's time is needed)
• Switches documented (docs/development.md; the README for a menu row) and shipped at their recommended value.
• A game-free test added for the fix, failing on main and passing here. (tests/test_no_webgpu.py; not run here: left to CI)
• A check under tests/ that reads changed source updated in this PR, not weakened.
• Host code builds on both Windows (llvm-mingw) and Linux, with UTF-8 paths. (not built here: CI's Windows and Linux builds)
• Docs updated.

🤖 Generated with Claude Code

**Type:** build · **Risk:** medium · **Tested on:** nothing run (Linux PC, Windows change) · **Needs:** Windows, any GPU ## Why Every Windows build compiles an old renderer the game doesn't use by default (TOTKPort's WebGPU renderer and suyu's shader recompiler with its WGSL back end, about 160 C++ files at -O2) and downloads `wgpu-native` for it, so every player's first build is longer than it needs to be. The game only drew with it when `RX_BACKEND` named another renderer than Vulkan, or when the Vulkan back end didn't start. The owner decided (2026-11-14) to drop it: Linux already builds without it. See https://git.magicbean.dev/MagicBeanDev/Longshot/issues/362 (a shorter first build). ## What changed The Windows game is now built as the Linux one is: the Vulkan back end (`longshot_vk.dll`) is the only renderer, the exe is linked with the stand-ins that were already there for a build without wgpu-native (`host/rx_nowgpu.c`), and `wgpu-native` is no longer a build tool. The WebGPU renderer's own sources are deleted; the headers the host and the Vulkan back end use stay. - build: `nx2/build.py` loses `gpu_sources()`, `WGPU`/`WGPU_NATIVE`, the "the renderer (C++20, NDEBUG...)" job branch, the `wgpu_native.dll.lib` link and the `wgpu_native.dll` copy; `host/rx_nowgpu.c` is always in the Windows host sources (Linux keeps `host/posix/linux_stubs.c`). - host: `rx_start()` in `host/rx_backend.c` no longer reads `RX_BACKEND`: it always loads the Vulkan back end, as it did with `RX_BACKEND` unset (the default). `load_vulkan()`'s failure lines say "nothing is drawn" instead of "WebGPU". `host/rx_nowgpu.c` gains `rx_collect_written()` (a build with `RX_SOFT_WRITE_WATCH=0` needed it from `gpu/render/rx_core.cpp`; it now answers "nothing tracked", as Linux's does) and its messages are reworded. `host/display.c` logs "no renderer" instead of "no WebGPU device". Comments in `nn_hid.c`, `nvn_native.c`, `hos_main.c`, `nn_audio.c` no longer point at the deleted files. - gpu/render: deleted the 16 `rx_*.cpp`, `rx_none.c` (built by nothing already), and the headers only they used (`rx_internal.h`, `rx_web_mt.h`, `rx_bcn.h`, `rx_wgpu_names.h`, `renderdoc_app.h`). Kept `rx.h` (`host/display.c`), `rx_nvn.h` and `nvn_api.h` (the op stream: `host/`, `gpu/vk/vk_api.h`, tests, the AppImage's header copy). The top comment of `rx.h` says what it is now. - tools: `tools.lock` loses the `wgpu-native` line, `scripts/tools.py` `REQUIRED["windows-x64"]` loses it, the builder (`builder/LongshotBuilder/Runner.cs`) no longer sets `WGPU_NATIVE`, `builder/setup/setup.c` comment. `nx2/make_game.py` drops `wgpu_native.dll` from `BESIDE` and removes one an older build left in the game folder (nothing loads it now). - scripts: `scripts/run-vk.ps1` no longer sets `RX_BACKEND=vulkan`. - ci: the third-party objects steps (`.forgejo/workflows/build-windows.yml` and both in `.circleci/shared.yml`) no longer install `wgpu-native` or set `WGPU_NATIVE`. - tests: new `tests/test_no_webgpu.py` (the build compiles nothing of `gpu/render`, names no wgpu-native, no tool list requires it, `rx_start` reads no `RX_BACKEND`, and both platforms' stand-ins define every `wgr_*` and live-input function the host names, so the exe links). `tests/test_renderer_load_error.py` now also checks the Windows failure line names no WebGPU. `tests/test_dev_packs.py`: a route check's label and path (`gpu/render/x.h`: still "game"). `tests/test_capture_past_end.py` no longer lists the deleted `gpu/render/rx_shader.cpp` among the capture loops it reads. - docs: `docs/development.md` (the `RX_BACKEND` row and `WGPU_NATIVE` gone, the `gpu/render/` layout row, the live input's `!dump`/`!shots`), `README.md` and `CONTRIBUTING.md` (the `tools\wgpu-native` step), `docs/design.md` (the current design; the history stays), `docs/releases.md`, `RELEASING.md` (a note on the dated "Done" line), `gpu/PROVENANCE.md` (what was removed and what stays), `gpu/vk/VULKAN.md`'s top note, `gpu/vk/vk_api.h`'s comments. Left on purpose: `gpu/third_party/suyu/` is untouched (its shader recompiler and WGSL back end stay in the tree, uncompiled; the Vulkan back end still compiles its texture decoders). `third_party/` is untouched. Historical comments in `gpu/vk` that say a piece was ported "from the WebGPU renderer's rx_*.cpp" are left as history, and `gpu/vk/vk_state.h` and `vk_shader.cpp` are not touched at all (they are in the shader stamp: a comment change there would make every player translate every program again). `scripts/cpu_categories.py`'s `render/webgpu-fallback` row is left: it matches nothing in a new build and still reads older profiles. ## Measurements **Verdict:** not measured: this PC is Linux, and the point of the change (a shorter Windows build, ~160 fewer C++ files) has to be timed on a Windows build. Requested under Must be tested. ### Flags and settings Two switches removed; nothing added. | Flag | What it is | Before this PR | After this PR | Effect | Helps / doesn't help / untested | Recommendation | |---|---|---|---|---|---|---| | `RX_BACKEND` (removed) | Picked the renderer | Unset or `vulkan`: Vulkan; anything else: the WebGPU renderer, in a build with wgpu-native | Not read: always Vulkan | None at the default | Untested on Windows | Removed: Vulkan is the only renderer | | `WGPU_NATIVE` (removed, build) | The wgpu-native folder the build compiled the WebGPU renderer against | Set by the builder from its tools | Not read | The WebGPU renderer is not built | Untested on Windows | Removed | ## Testing **Tested:** nothing run on this PC (the owner's rule for this change: no tests or builds here; CI runs the game-free checks). Read instead: every user of each deleted file (grep across the tree: only the WebGPU renderer itself, and comments), every check under `tests/` that reads `host/rx_backend.c`, `nx2/build.py`, `tools.lock` or `scripts/tools.py`. **Not tested:** the Windows build and the game on Windows; the Linux build (its sources change only in comments and in `host/rx_backend.c`'s `rx_start` and log lines); `python -I -B tests/run_all.py` (left to CI). **Must be tested:** - Windows, any GPU (I'm on Linux): a full build from the Builder or `make_game.py`, expecting it to build without wgpu-native and with no `gpu_*` objects, the game to start on Vulkan (`rx: renderer: Vulkan (longshot_vk.dll)` in `longshot_log.txt`) and play as on main, and the compile to be shorter than main's at the same commit otherwise (by hand or the Test Farm: request to open) - Windows, a folder built by main: build this PR over it, expecting `wgpu_native.dll` gone from the folder and the game starting (by hand) ## Risks **What could break:** the Windows exe could fail to link if something still needs a WebGPU renderer function that neither `host/rx_nowgpu.c` nor the host defines (`tests/test_no_webgpu.py` checks the ones the host names). A Windows PC whose Vulkan doesn't start used to fall back to the WebGPU renderer (on D3D12 through wgpu-native); it now gets a black window and "nothing is drawn" in the log, as Linux does. With default settings and a working Vulkan nothing changes. **Turning it off:** revert the commit. **Compatibility:** saves, `longshot_options.ini` and the shader and pipeline caches are untouched (no stamped file changed). The vk ABI is unchanged (`vk_api.h`: comments only, no `MKWVK_API_VERSION` bump). The builder no longer downloads `wgpu-native`; an older game folder's `wgpu_native.dll` is removed at the next build. `RX_BACKEND` set to something other than `vulkan` now gets Vulkan. ## Issues See https://git.magicbean.dev/MagicBeanDev/Longshot/issues/362 ## Checklist ### Everyone - [x] One topic. - [x] The title is a Conventional Commit (`type(scope): description`) of the status line's type. - [x] Why gives the reason plainly and links the issues that fit (the tracker searched). - [x] Measurements has a verdict, or one line saying why there is nothing to measure. - [x] Testing names what was tested, what wasn't, and what must still be tested, where and how. - [x] Rebased on current `main`. - [ ] `python -I -B tests/run_all.py` passes (any failure that also fails on `main` is named). (not run here: left to CI) - [x] No game files, dumps, build output or generated files. - [x] `third_party/` untouched (a new folder with a `PROVENANCE.md` instead). - [x] The license is OK: GPL-3.0-or-later, nothing incompatible and nothing decompiled. ### When it applies - [x] AI assistance recorded with `Assisted-by:` / `Co-Authored-By:` trailers. - [ ] Numbers from the same PC, scene and settings, at least 2 interleaved runs each, the median reported. (not measured here: a Windows build's time is needed) - [x] Switches documented (docs/development.md; the README for a menu row) and shipped at their recommended value. - [ ] A game-free test added for the fix, failing on `main` and passing here. (`tests/test_no_webgpu.py`; not run here: left to CI) - [x] A check under `tests/` that reads changed source updated in this PR, not weakened. - [ ] Host code builds on both Windows (llvm-mingw) and Linux, with UTF-8 paths. (not built here: CI's Windows and Linux builds) - [x] Docs updated. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

HumanBean added 1 commit 2026-11-14 20:20:51 +00:00

build(windows): no WebGPU renderer: the Windows game builds the Vulkan back end only, as Linux does ... 2604177316

Every Windows build compiled TOTKPort's WebGPU renderer (gpu/render/*.cpp) and suyu's shader
recompiler with its WGSL back end into the exe, and the builder downloaded wgpu-native for it, while
the game drew with it only when RX_BACKEND named another renderer or Vulkan did not start.

- nx2/build.py: no gpu_sources, WGPU_NATIVE or wgpu branch; host/rx_nowgpu.c is always the Windows
  stand-in (Linux keeps host/posix/linux_stubs.c), and no wgpu_native.dll is linked or copied.
- host/rx_backend.c: RX_BACKEND is not read; the Vulkan back end is always loaded, and a failure
  logs "nothing is drawn" instead of "WebGPU". host/rx_nowgpu.c also stands in for
  rx_collect_written (a build with RX_SOFT_WRITE_WATCH=0).
- tools.lock, scripts/tools.py REQUIRED, the builder (Runner.cs), make_game.py: no wgpu-native;
  make_game.py removes a wgpu_native.dll an older build left in the game folder.
- gpu/render: the WebGPU renderer's sources and its own headers deleted; rx.h, rx_nvn.h and
  nvn_api.h (the host's and the Vulkan back end's) kept. gpu/third_party/suyu is untouched.
- tests/test_no_webgpu.py: the build compiles nothing of gpu/render, needs no wgpu-native, and the
  stand-ins define every renderer function the host names on both platforms.
- Docs: RX_BACKEND and WGPU_NATIVE rows gone, the layout and design text say Vulkan is the only renderer.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

HumanBean force-pushed build/no-webgpu from 2604177316 to b47ce8ca3c 2026-11-14 20:23:00 +00:00 Compare

HumanBean commented 2026-11-14 20:23:02 +00:00

Author

Owner

• Copy link
• Copy Markdown

Rebased onto main 4e80ea63 (after #569 merged), no conflicts. Nothing run here.

Rebased onto main 4e80ea63 (after https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/569 merged), no conflicts. Nothing run here.

HumanBean force-pushed build/no-webgpu from b47ce8ca3c to 5d463a657f 2026-11-14 20:24:05 +00:00 Compare

HumanBean commented 2026-11-14 20:24:08 +00:00

Author

Owner

• Copy link
• Copy Markdown

Rebased onto main 414f56e2 (after #569 and #573 merged), no conflicts. Nothing run here.

Rebased onto main 414f56e2 (after https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/569 and https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/573 merged), no conflicts. Nothing run here.

HumanBean commented 2026-11-14 20:42:13 +00:00

Author

Owner

• Copy link
• Copy Markdown

Decided with the owner: the WebGPU fallback goes. It ran only when Vulkan failed to load or start on Windows (no loader, a GPU or driver without the features the renderer needs, a broken driver), and then gave an unmaintained renderer that none of Longshot's Ocarina fixes, upscalers, frame generation or menu reached, and that no tested run ever used. Linux has never had it. A clear message when Vulkan fails, instead of the black window, can follow separately.

Decided with the owner: the WebGPU fallback goes. It ran only when Vulkan failed to load or start on Windows (no loader, a GPU or driver without the features the renderer needs, a broken driver), and then gave an unmaintained renderer that none of Longshot's Ocarina fixes, upscalers, frame generation or menu reached, and that no tested run ever used. Linux has never had it. A clear message when Vulkan fails, instead of the black window, can follow separately.

HumanBean commented 2026-11-14 21:16:46 +00:00

Author

Owner

• Copy link
• Copy Markdown

Rebased onto main b2911d33. Three text conflicts with #572 (both edited the same lines), resolved as main's new text with this PR's one change, wgpu-native dropped: scripts/tools.py's REQUIRED and its comment, RELEASING.md's one-time setup step 6, docs/releases.md's Windows tool list. Nothing run here.

Rebased onto main b2911d33. Three text conflicts with https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/572 (both edited the same lines), resolved as main's new text with this PR's one change, wgpu-native dropped: scripts/tools.py's REQUIRED and its comment, RELEASING.md's one-time setup step 6, docs/releases.md's Windows tool list. Nothing run here.

HumanBean force-pushed build/no-webgpu from 5d463a657f to 2fc45cf04d 2026-11-14 21:16:47 +00:00 Compare

HumanBean changed title from WIP: build(windows): no WebGPU renderer: the Windows game builds the Vulkan back end only, as Linux does to build(windows): no WebGPU renderer: the Windows game builds the Vulkan back end only, as Linux does 2026-11-14 21:17:29 +00:00

HumanBean referenced this pull request from MagicBeanDev/LongshotLauncher 2026-11-14 23:38:30 +00:00

refactor(build): no WGPU_NATIVE, the prebuilt Linux renderer only where older tools carry it, and no AVX2 requirement #73

HumanBean referenced this pull request from a commit 2026-11-15 06:34:54 +00:00

refactor(build): no WGPU_NATIVE, the prebuilt Linux renderer only where older tools carry it, and no AVX2 requirement (#73)

HumanBean force-pushed build/no-webgpu from 2fc45cf04d to b6aa228f06 2026-11-15 06:43:07 +00:00 Compare

mooki added the

area/renderer

kind

chore

platform/linux

platform/windows

priority

low

labels 2026-11-15 13:56:03 +00:00

HumanBean force-pushed build/no-webgpu from b6aa228f06 to 002fa6f28c 2026-11-15 18:18:47 +00:00 Compare

HumanBean commented 2026-11-15 18:19:14 +00:00

Author

Owner

• Copy link
• Copy Markdown

Rebased on main @ 772fe283 (now 002fa6f2). Two conflicts: gpu/render/rx_shader.cpp stays deleted (main's #427 fix there was the WebGPU copy of the one in gpu/vk/vk_shader.cpp, which is kept), and nx2/build.py keeps #857's exe_resources() without WGPU/gpu_sources(). Also covered what main added since: .circleci/shared.yml (#417) no longer installs wgpu-native, which this PR takes out of tools.lock, and tests/test_capture_past_end.py (#427) no longer reads the deleted rx_shader.cpp. Not run here; left to CI.

Rebased on main @ 772fe283 (now 002fa6f2). Two conflicts: `gpu/render/rx_shader.cpp` stays deleted (main's #427 fix there was the WebGPU copy of the one in `gpu/vk/vk_shader.cpp`, which is kept), and `nx2/build.py` keeps #857's `exe_resources()` without `WGPU`/`gpu_sources()`. Also covered what main added since: `.circleci/shared.yml` (#417) no longer installs wgpu-native, which this PR takes out of `tools.lock`, and `tests/test_capture_past_end.py` (#427) no longer reads the deleted `rx_shader.cpp`. Not run here; left to CI.

HumanBean merged commit df1d607d2d into main 2026-11-15 18:24:32 +00:00

HumanBean deleted branch build/no-webgpu 2026-11-15 18:24:32 +00:00

HumanBean referenced this pull request from a commit 2026-11-15 18:24:32 +00:00

build(windows): no WebGPU renderer: the Windows game builds the Vulkan back end only, as Linux does (#577)

HumanBean referenced this pull request 2026-11-15 18:27:59 +00:00

perf(build): an update with nothing changed copies no unchanged exe into the game folder, and the finish says where its time went #631

HumanBean referenced this pull request 2026-11-15 18:45:23 +00:00

WIP: feat(input): opt-in Sprint on Shift or the left stick's press, without the roll #196

HumanBean referenced this pull request 2026-11-15 18:45:45 +00:00

docs(readme): a short front page, the reference sections moved to docs/ as they were #236

HumanBean referenced this pull request 2026-11-15 18:45:46 +00:00

WIP: feat(gameplay): add experimental cutscene skipping #200

HumanBean referenced this pull request 2026-11-15 18:45:47 +00:00

WIP: feat(input): share Keyboard Auto/On/Off policy with menu navigation #152

HumanBean referenced this pull request 2026-11-15 18:46:14 +00:00

feat(runtime): HOS_COUNTERS, the memory and cache counters as one JSON line every N seconds #65

HumanBean referenced this pull request 2026-11-15 18:47:07 +00:00

feat(vk): Intel XeSS Super Resolution as a temporal upscaler (Vulkan, the player's libxess.dll) #459

HumanBean referenced this pull request 2026-11-15 18:47:08 +00:00

WIP: feat(input): separate keyboard actions and binding pages #613

HumanBean referenced this pull request 2026-11-15 18:47:09 +00:00

feat(mods): a Mods tab, merged resource size tables, ExeFS patches and Atmosphere cheats (HumanBean's draft from Hookshot, with the lift fix) #134

HumanBean referenced this pull request 2026-11-15 18:49:14 +00:00

WIP: feat(macos): run the host on macOS: build, runtime, control, crash records and the Vulkan loader #845

HumanBean referenced this pull request 2026-11-15 18:49:14 +00:00

fix(vk): draw indirect quad draws, with and without a count, on every driver #705

HumanBean referenced this pull request 2026-11-15 18:49:15 +00:00

fix(vk): draw mesh count draws on the compute fallback, grow its records area on demand, and count what it cuts #707

HumanBean referenced this pull request 2026-11-15 18:49:15 +00:00

WIP: feat(macos): Vulkan on Metal: pinned host imports, the unified-memory budget, Metal's cache folder and KosmicKrisp #846

HumanBean referenced this pull request 2026-11-15 18:49:15 +00:00

WIP: feat(vk): the SPIR-V back end on MoltenVK and guest spans past one device allocation #848

HumanBean referenced this pull request 2026-11-15 18:49:16 +00:00

WIP: fix(macos): MoltenVK's precompile and pipelines on Metal #849

HumanBean referenced this pull request 2026-11-15 18:49:17 +00:00

WIP: fix(macos): draw-path emulation on MoltenVK #850

HumanBean referenced this pull request 2026-11-15 18:49:17 +00:00

WIP: fix(vk): draw indirect quads and mesh count draws on every driver, with their Mac halves #851

HumanBean referenced this pull request 2026-11-15 18:49:18 +00:00

WIP: perf(macos): Mac performance on MoltenVK #852

HumanBean referenced this pull request 2026-11-15 18:49:26 +00:00

WIP: feat(haptics): a DualSense's fine haptics, from the game's own bands #405

HumanBean referenced this pull request 2026-11-15 18:50:04 +00:00

feat(vr): play in a PC VR headset over OpenXR, with a VR tab in the Insert menu #204

HumanBean referenced this pull request 2026-11-15 18:51:09 +00:00

feat(menu): save and resume when restarting for changed settings #61

HumanBean referenced this pull request 2026-11-15 18:51:09 +00:00

feat(coop): add relay rooms, accepted progress sync and a Co-op tab #68

HumanBean referenced this pull request 2026-11-15 18:51:10 +00:00

ci(windows): optionally sign builder executables with Authenticode #98

HumanBean referenced this pull request 2026-11-15 18:51:12 +00:00

WIP: feat(vk): a Colour filter row for colour-blind players #139

dyst referenced this pull request 2026-11-15 19:12:20 +00:00

feat(runtime): one list of the game's switches, host/switches.def, kept in step with the reads #62

dyst referenced this pull request 2026-11-15 19:12:21 +00:00

feat(runtime): one list of the game's switches, host/switches.def, kept in step with the reads #62

dyst referenced this pull request 2026-11-15 19:12:22 +00:00

feat(runtime): HOS_COUNTERS, the memory and cache counters as one JSON line every N seconds #65

dyst referenced this pull request 2026-11-15 19:17:25 +00:00

feat(runtime): HOS_DEV_DLL_SWAP, a renderer built alone taken at the next start, checked against the game first #183

Sign in to join this conversation.

Reviewers

No reviewers

Labels

Clear labels   

area/accessibility

Barriers for people with disabilities

  

area/audio

Sound and music

  

area/input

Controllers, keyboard, mouse

  

area/install

Installing, updating, building from your game

  

area/menu

In-game menu and settings

  

area/renderer

Vulkan renderer, shaders, effects

  

area/saves

Save files

  

area/tooling

Developer tools: scripts, CI, build and release tooling

  

area/upscaling

DLSS, FSR, frame generation

  

ci: approved

A maintainer approved CircleCI for this fork PR (sticky; removed after 14 quiet days)

  

good first issue

A good start for a new contributor

  

gpu/amd

AMD GPUs

  

gpu/igpu

Integrated GPUs

  

gpu/intel

Intel GPUs

  

gpu/nvidia

NVIDIA GPUs

  

help wanted

Extra hands or knowledge wanted

  

kind

bug

Something doesn't work as it should

  

kind

chore

Build, CI, dependencies, tidying

  

kind

docs

Documentation

  

kind

feature

A new feature or an improvement

  

kind

performance

Frame rate, frame time, loading, memory

  

kind

question

A question rather than a change

  

needs measurements

Waiting for before/after measurements

  

platform/linux

Affects Linux (a change for every system carries both Windows and Linux)

  

platform/macos

macOS (Apple silicon or Intel Macs)

  

platform/steam-deck

Steam Deck

  

platform/windows

Affects Windows (a change for every system carries both Windows and Linux)

  

preview

Maintainers: deploy a preview of this PR's website (read its changes first). Removed after the run.

  

priority

critical

Crashes, data loss, or blocks most players

  

priority

high

Important: next in line

  

priority

low

Nice to have

  

regression

Worked in an earlier build, broken in a newer one: highest priority, above other bugs

  

risk

high

Broad or core change: the renderer, defaults on many systems, saves, installs, security

  

risk

low

Small, contained change: little can break

  

risk

medium

Changes behaviour in one area, or a default on some systems

  

status

blocked

Waits on something else first

  

status

confirmed

Reproduced or agreed on

  

status

duplicate

Already reported or proposed

  

status

invalid

Not a problem in this project

  

status

needs-info

Waiting for more information from the reporter

  

status

wontfix

Will not be worked on

  

testing wanted

Wanted: a test on other hardware (posts a test request in #test-requests)

  

ci: approved

No labels

area/accessibility

area/audio

area/input

area/install

area/menu

area/renderer

area/saves

area/tooling

area/upscaling

ci: approved

good first issue

gpu/amd

gpu/igpu

gpu/intel

gpu/nvidia

help wanted

kind

bug

kind

chore

kind

docs

kind

feature

kind

performance

kind

question

needs measurements

platform/linux

platform/macos

platform/steam-deck

platform/windows

preview

priority

critical

priority

high

priority

low

regression

risk

high

risk

low

risk

medium

status

blocked

status

confirmed

status

duplicate

status

invalid

status

needs-info

status

wontfix

testing wanted

ci: approved

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

1 participant

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference

MagicBeanDev/Longshot!577

No description provided.