joltc-sys: derive Default for JPC_PlaneShapeSettings (depends on JoltC#25)#14
Open
Solidor777 wants to merge 3 commits into
Open
joltc-sys: derive Default for JPC_PlaneShapeSettings (depends on JoltC#25)#14Solidor777 wants to merge 3 commits into
Solidor777 wants to merge 3 commits into
Conversation
Two coupled bugs surfaced when cross-compiling joltc-sys to
Android (aarch64-linux-android via NDK r26d clang from a Windows
host):
1. `cfg!(windows)` is a HOST-OS check inside build.rs. When
targeting Android from Windows, the conditional always fires
and adds the MSVC-only `/EHsc` flag to the cmake config — NDK
clang then receives `/EHsc` as a path argument and errors out
("error: no such file or directory: '/EHsc'").
Fix: read `CARGO_CFG_TARGET_OS` and only add `/EHsc` when the
actual target is windows.
2. Building for Android requires three CMake-side hints that
joltc-sys does not currently set: `CMAKE_TOOLCHAIN_FILE`
(NDK's `android.toolchain.cmake`), `ANDROID_ABI` (string the
toolchain reads — env var alone is ignored), and the
generator (Ninja, since cmake-rs defaults to MSBuild on
Windows hosts).
Fix: when `target_os == "android"`, derive ANDROID_ABI from
`CARGO_CFG_TARGET_ARCH` (aarch64 → arm64-v8a, arm →
armeabi-v7a, etc.), look up the NDK at
`ANDROID_NDK_HOME` / `ANDROID_NDK_ROOT` / `ANDROID_NDK`,
and force Ninja.
End-user contract after this change: set ANDROID_NDK_HOME,
run `cargo ndk --target aarch64-linux-android check -p joltc-sys`.
The standard cargo-ndk + nttld/setup-ndk@v1 flow Just Works.
Validated locally on Windows-host → aarch64-linux-android with
NDK r26d (full JoltPhysics + JoltC compiles in ~20s). 32-bit ARM
(armv7-linux-androideabi) hits an unrelated upstream JoltC
LAYOUT_COMPATIBLE static_assert (`JPC_*Settings` 16-byte
alignment vs `JPH::*Settings` 8-byte alignment on 32-bit ARM)
that this PR does not address.
Adds an iOS branch alongside the Android branch from SecondHalfGames#12. Covers both `aarch64-apple-ios` (device) and `aarch64-apple-ios-sim` (simulator on Apple Silicon hosts); `x86_64-apple-ios` is the legacy Intel-Mac simulator path. From a macOS host: ```sh cargo check --target aarch64-apple-ios -p joltc-sys cargo check --target aarch64-apple-ios-sim -p joltc-sys ``` Both succeed with no extra env-var setup. ## Two coupled fixes 1. **`build_joltc()` iOS branch.** Sets `CMAKE_SYSTEM_NAME=iOS`, `CMAKE_OSX_SYSROOT` (`iphoneos` for device vs `iphonesimulator` for simulator, distinguished via `CARGO_CFG_TARGET_ABI=sim`), `CMAKE_OSX_ARCHITECTURES` from `target_arch`, and `CMAKE_OSX_DEPLOYMENT_TARGET=12.0`. Forces Ninja so the build is generator-agnostic on macos CI runners. 2. **`generate_bindings()` clang-target translation.** bindgen passes rustc's TARGET to clang verbatim. rustc's iOS-simulator triple `aarch64-apple-ios-sim` is invalid clang syntax — clang errors with `version 'sim' in target triple is invalid`. Translate to clang's recognised form (`arm64-apple-ios-simulator` for sim, `arm64-apple-ios` for device) and supply the matching SDK sysroot via `xcrun --show-sdk-path` so iOS system headers (stdint.h, etc.) resolve. ## Validation - `cargo check --target aarch64-apple-ios -p joltc-sys` on macOS host (Xcode 16.4, iPhoneOS 18.5 SDK): succeeds. - `cargo check --target aarch64-apple-ios-sim -p joltc-sys`: succeeds. - Downstream Titan engine's `titan-jolt-3d` builds against both iOS targets in CI on `macos-latest`. - Desktop targets unaffected (iOS branch gated on `target_os == "ios"`). ## Stacking note This PR is stacked on SecondHalfGames#12 (`android-cross-compile-and-target-os-fix`) to share the host-vs-target awareness scaffolding. Either PR can merge first — the iOS branch is a strict additive layer.
Once SecondHalfGames/JoltC#25 lands, PlaneShape gets a JPC_*Settings struct + matching _default initializer. Add it to the ffi_default! list so consumers can use the same struct-update idiom they use for every other ShapeSettings: let settings = JPC_PlaneShapeSettings { Normal: JPC_Vec3 { x: 0.0, y: 1.0, z: 0.0, _w: 0.0 }, ..Default::default() }; This PR is stacked on SecondHalfGames#12 + SecondHalfGames#13 (Android + iOS cross-compile) but only adds one line; it can be cherry-picked onto main once those land or left on the stack.
Solidor777
added a commit
to Solidor777/Titan
that referenced
this pull request
May 8, 2026
m34-a-ii shipped Halfspace as a 100×0.05×100 thin static box because
JoltC didn't expose JPH's PlaneShape. The authored normal was ignored;
raycasts returned axis-aligned box-face normals; rotated halfspaces
silently used the ground approximation.
Bumps the Solidor777/JoltC submodule to 3eddf05, which adds a TITAN
PATCH exposing `JPC_PlaneShapeSettings` + `_default` + `_Create`
following the existing BoxShape/SphereShape pattern. Companion one-line
patch on `crates/third_party/joltc-sys/src/lib.rs` adds the new shape
to the `ffi_default!` list so `JPC_PlaneShapeSettings::default()` works
from rust.
`titan-jolt-3d`'s `BodyShape::Halfspace { normal }` arm now constructs
a real `PlaneShape` with the authored normal as the plane's outward
direction and `Constant = 0` (plane through body-local origin). Module
docs updated.
New regression test `halfspace_authors_non_y_normal` raycasts a +X-
normal halfspace and asserts:
- ray distance lands on the plane (5 m for a 5 m offset along the
authored normal)
- raycast surface normal returned by Jolt matches the authored normal
(dot with +X > 0.95). The thin-box approx returned (0, ±1, 0) here.
Existing `dynamic_body_rests_on_static_ground` updated: the ball now
settles at y ≈ radius (0.4..0.6) instead of radius + box-half-extent
(0.4..0.7) since the plane is exactly at y=0.
508 tests pass (was 507). Smoke-launched: editor reaches `boot
complete` cleanly.
Upstream PRs: [SecondHalfGames/JoltC#25](SecondHalfGames/JoltC#25)
+ [SecondHalfGames/jolt-rust#14](SecondHalfGames/jolt-rust#14).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Solidor777
added a commit
to Solidor777/Titan
that referenced
this pull request
May 17, 2026
m34-a-ii shipped Halfspace as a 100×0.05×100 thin static box because
JoltC didn't expose JPH's PlaneShape. The authored normal was ignored;
raycasts returned axis-aligned box-face normals; rotated halfspaces
silently used the ground approximation.
Bumps the Solidor777/JoltC submodule to 3eddf05, which adds a TITAN
PATCH exposing `JPC_PlaneShapeSettings` + `_default` + `_Create`
following the existing BoxShape/SphereShape pattern. Companion one-line
patch on `crates/third_party/joltc-sys/src/lib.rs` adds the new shape
to the `ffi_default!` list so `JPC_PlaneShapeSettings::default()` works
from rust.
`titan-jolt-3d`'s `BodyShape::Halfspace { normal }` arm now constructs
a real `PlaneShape` with the authored normal as the plane's outward
direction and `Constant = 0` (plane through body-local origin). Module
docs updated.
New regression test `halfspace_authors_non_y_normal` raycasts a +X-
normal halfspace and asserts:
- ray distance lands on the plane (5 m for a 5 m offset along the
authored normal)
- raycast surface normal returned by Jolt matches the authored normal
(dot with +X > 0.95). The thin-box approx returned (0, ±1, 0) here.
Existing `dynamic_body_rests_on_static_ground` updated: the ball now
settles at y ≈ radius (0.4..0.6) instead of radius + box-half-extent
(0.4..0.7) since the plane is exactly at y=0.
508 tests pass (was 507). Smoke-launched: editor reaches `boot
complete` cleanly.
Upstream PRs: [SecondHalfGames/JoltC#25](SecondHalfGames/JoltC#25)
+ [SecondHalfGames/jolt-rust#14](SecondHalfGames/jolt-rust#14).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Solidor777
added a commit
to Solidor777/Titan
that referenced
this pull request
Jun 4, 2026
m34-a-ii shipped Halfspace as a 100×0.05×100 thin static box because
JoltC didn't expose JPH's PlaneShape. The authored normal was ignored;
raycasts returned axis-aligned box-face normals; rotated halfspaces
silently used the ground approximation.
Bumps the Solidor777/JoltC submodule to 3eddf05, which adds a TITAN
PATCH exposing `JPC_PlaneShapeSettings` + `_default` + `_Create`
following the existing BoxShape/SphereShape pattern. Companion one-line
patch on `crates/third_party/joltc-sys/src/lib.rs` adds the new shape
to the `ffi_default!` list so `JPC_PlaneShapeSettings::default()` works
from rust.
`titan-jolt-3d`'s `BodyShape::Halfspace { normal }` arm now constructs
a real `PlaneShape` with the authored normal as the plane's outward
direction and `Constant = 0` (plane through body-local origin). Module
docs updated.
New regression test `halfspace_authors_non_y_normal` raycasts a +X-
normal halfspace and asserts:
- ray distance lands on the plane (5 m for a 5 m offset along the
authored normal)
- raycast surface normal returned by Jolt matches the authored normal
(dot with +X > 0.95). The thin-box approx returned (0, ±1, 0) here.
Existing `dynamic_body_rests_on_static_ground` updated: the ball now
settles at y ≈ radius (0.4..0.6) instead of radius + box-half-extent
(0.4..0.7) since the plane is exactly at y=0.
508 tests pass (was 507). Smoke-launched: editor reaches `boot
complete` cleanly.
Upstream PRs: [SecondHalfGames/JoltC#25](SecondHalfGames/JoltC#25)
+ [SecondHalfGames/jolt-rust#14](SecondHalfGames/jolt-rust#14).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Depends on SecondHalfGames/JoltC#25, which adds the C-side
JPC_PlaneShapeSettings+_defaultexports.This PR is one line — add
JPC_PlaneShapeSettings -> JPC_PlaneShapeSettings_defaultto theffi_default!list injoltc-sys/src/lib.rsso the new shape gets the same struct-update idiom every other ShapeSettings has:Stacking
Filed on top of #12 + #13 (Android + iOS cross-compile) so the diff against
mainis bigger than the actual change. The relevant added line is the only one outside the iOS / Android branches.If #12 and #13 merge first this PR auto-shrinks to the single-line diff above. If a maintainer prefers, the line is trivial to cherry-pick onto a fresh PR.
Test plan
JPC_PlaneShapeSettings::default()returns the same field values asJPC_PlaneShapeSettings_default()writes (Normal=+Y, Constant=0, HalfExtent=cDefaultHalfExtent).titan-jolt-3d; new integration test (halfspace_authors_non_y_normal) raycasts a +X-normal halfspace and asserts the surface normal returned by Jolt matches.