Cocoa + Emscripten CI coverage #112

New issue

Open

opened 2026-05-29 07:23:51 +02:00 by OragonEfreet · 0 comments

OragonEfreet commented

2026-05-29 07:23:51 +02:00

Owner

.forgejo/workflows/ today only exercises Linux (linux-linux-gcc) and MinGW cross-compile (linux-windows-mingw). #107 added or significantly touched two backends that aren't built in CI:

Cocoa (src/cocoa/video_cocoa.m) — only buildable on macOS. Regressions only surface when a contributor or maintainer builds on a Mac.
Emscripten (src/emscripten/video_emscripten.c) — buildable from Linux through emscripten/emsdk but not wired into any workflow.

Both backends shipped real changes in #107 (draw-callback wiring, sub-rect present, rAF scheduling on Emscripten); both were validated only locally / manually. Without CI a future PR can break them silently.

Outcome

Two new workflow extensions:

Cocoa

New job in ci-test.yml (or a dedicated ci-macos.yml) that runs on a macos-* runner.
Configures the project with default CMake (Cocoa auto-enables on APPLE), builds, runs ctest.
Optional: smoke-runs an example like template_callbacks to confirm the lifecycle.

Emscripten

New job (or workflow) that uses the emscripten/emsdk:latest container.
Configures with emcmake cmake ... -DBANJO_CONFIG_EMSCRIPTEN_BACKEND=ON -DBUILD_TESTING=ON.
Builds, then runs the test executables under node (e.g. node build/test/unit_app.js).
Same dance we manually did in #107 (commits 6 and 7's verification).

Acceptance criteria

A trivial PR that breaks the Cocoa backend (e.g. delete cocoa_request_redraw) goes red in CI.
Same for a trivial Emscripten break (e.g. wrong arg count to MAIN_THREAD_EM_ASM).
Both jobs run on every PR to main.

Out of scope

iOS, Android, FreeBSD CI.
Cross-compiling Emscripten from macOS or Windows runners.
Browser-based test runs (we run under node). Real browser smoke would need a Selenium-style harness — separate card if we ever want it.
GitHub Actions migration. Forgejo Actions stays the home for CI per the project's existing .forgejo/ layout.

`.forgejo/workflows/` today only exercises Linux (`linux-linux-gcc`) and MinGW cross-compile (`linux-windows-mingw`). #107 added or significantly touched two backends that aren't built in CI: - **Cocoa** (`src/cocoa/video_cocoa.m`) — only buildable on macOS. Regressions only surface when a contributor or maintainer builds on a Mac. - **Emscripten** (`src/emscripten/video_emscripten.c`) — buildable from Linux through `emscripten/emsdk` but not wired into any workflow. Both backends shipped real changes in #107 (draw-callback wiring, sub-rect present, rAF scheduling on Emscripten); both were validated only locally / manually. Without CI a future PR can break them silently. ## Outcome Two new workflow extensions: ### Cocoa - New job in `ci-test.yml` (or a dedicated `ci-macos.yml`) that runs on a `macos-*` runner. - Configures the project with default CMake (Cocoa auto-enables on `APPLE`), builds, runs `ctest`. - Optional: smoke-runs an example like `template_callbacks` to confirm the lifecycle. ### Emscripten - New job (or workflow) that uses the `emscripten/emsdk:latest` container. - Configures with `emcmake cmake ... -DBANJO_CONFIG_EMSCRIPTEN_BACKEND=ON -DBUILD_TESTING=ON`. - Builds, then runs the test executables under `node` (e.g. `node build/test/unit_app.js`). - Same dance we manually did in #107 (commits 6 and 7's verification). ## Acceptance criteria - A trivial PR that breaks the Cocoa backend (e.g. delete `cocoa_request_redraw`) goes red in CI. - Same for a trivial Emscripten break (e.g. wrong arg count to `MAIN_THREAD_EM_ASM`). - Both jobs run on every PR to `main`. ## Out of scope - iOS, Android, FreeBSD CI. - Cross-compiling Emscripten from macOS or Windows runners. - Browser-based test runs (we run under node). Real browser smoke would need a Selenium-style harness — separate card if we ever want it. - GitHub Actions migration. Forgejo Actions stays the home for CI per the project's existing `.forgejo/` layout.

OragonEfreet added this to the 1.0 milestone

2026-05-29 07:23:51 +02:00

OragonEfreet added the

enhancement

label

2026-05-29 07:23:51 +02:00

OragonEfreet self-assigned this

2026-05-29 07:23:51 +02:00

OragonEfreet added this to the 70's games readinesss project

2026-05-29 09:13:01 +02:00

OragonEfreet removed their assignment

2026-05-29 09:21:06 +02:00