Explicit user-controlled pixel scale and HiDPI fix

[pusewicz]

Working demo of @bullno1's proposal from the #575 review thread (including the follow-up naming refinements): remove all event-driven HiDPI behavior and make pixel scale a plain user-controlled value. Draft on purpose — it exists to give that design discussion something concrete to poke at.

• Startup is the only automatic step: canvas at window_size * display_scale, default 2d projection from the logical window size, both set once. Resizes and scale changes only raise flags (cf_app_was_resized / cf_app_display_scale_was_changed).
• Terminology follows the SDL functions underneath, per the thread. cf_app_get_display_scale (+_was_changed) is SDL_GetWindowDisplayScale — what the OS wants, consistent across platforms; the dpi names are gone. cf_app_set_pixel_scale is how fonts/shapes scale — arbitrary, or following the reported display scale; it does not touch the canvas (cf_app_set_canvas_size is separate). cf_app_update_display(scale) is the all-in-one helper bringing canvas, projection, and pixel scale together, usable on both resize and scale change — the w/h params are omitted since the window size is always current by the time user code reacts.
• Window size and mouse coordinates are normalized to logical points at the SDL boundary via SDL_GetDisplayContentScale (window creation, cf_app_set_size, resize events, mouse events; touch derives from window size). This is a no-op on macOS (content scale 1.0) and needs verification on Windows/X11 at >100% scaling — I can only test on a Mac.
• cf_draw_projection becomes truly sticky and cf_app_set_canvas_size persistent. The half-size regression Fix HiDPI half-size regression: default 2d projection is logical points #575 fixes dies structurally — no code path can rebuild the projection in pixel units — and all of Fix HiDPI half-size regression: default 2d projection is logical points #575's deferral/guard machinery becomes unnecessary. 351/351 tests on a 2x display (master: 319/339, the 20 failures being that regression).
• The hidpi sample carries the recipe (N follows the display scale, 1/2/4 force scales — forcing a value is also how you test HiDPI on a normal monitor); the hidpi topic doc is rewritten; test_hidpi covers the new contracts.

Cost, stated plainly: breaking change. 44 of 76 samples use resizable windows and each needs the two-line recipe (deliberately not yet updated); canvas_modes.c needs a rewrite. NO_HIGH_DPI_BIT is kept: it is the only handle on the creation-time 1x swapchain (memory + final-blit bandwidth), which cf_app_update_display(1) cannot replicate.

Before

After

🤖 Generated with Claude Code

https://claude.ai/code/session_01LHq3Cx7xcjAQhji9m1ADJn

[Copilot]

🔵 Needs a closer look

It changes core cross-platform window/input scaling semantics and HiDPI behavior in the app loop, so it needs human validation across multiple OS/display configurations despite the new tests.

Pull request overview

Reworks Cute Framework’s HiDPI model to remove event-driven/automatic canvas + projection updates and instead make pixel scale an explicitly user-controlled value, with a single helper (cf_app_update_display) for the common “resize/scale-change recipe”. This is a core behavioral shift intended to prevent projection/canvas mismatches (e.g., the #575 half-size regression) by structurally eliminating hidden refresh paths.

Changes:

• Introduces explicit pixel-scale APIs (cf_app_set_pixel_scale, cf_app_update_display) and updates app/canvas/projection contracts + docs accordingly.
• Normalizes SDL-reported window/mouse coordinates into CF logical points via display content scale at the SDL boundary (resize + mouse events, plus window sizing at creation / cf_app_set_size).
• Adds a new test_hidpi suite and updates existing app/canvas persistence tests; refreshes the hidpi sample to demonstrate the manual model and scale-following vs forced scaling.

File summaries

File	Description
test/test_hidpi.cpp	Adds readback-based tests for the new HiDPI/pixel-scale contracts and resize/scale-change behavior.
test/test_app.cpp	Updates tests to assert canvas size persistence across resize and MSAA changes.
test/test_app_shared.cpp	Updates shared app reset logic to restore startup-equivalent display settings via cf_app_update_display.
test/main.cpp	Registers the new test_hidpi suite.
test/CMakeLists.txt	Adds test_hidpi.cpp to the test build.
src/internal/cute_draw_internal.h	Replaces the “app canvas resized” hook with a “pixel scale changed” hook for AA refresh only.
src/internal/cute_app_internal.h	Adds internal cf_app_get_content_scale() (SDL content scale) accessor and updates internal scale semantics.
src/cute_input.cpp	Removes automatic canvas/projection rebuilds on resize/scale events; converts SDL raw coordinates to logical points.
src/cute_draw.cpp	Removes automatic projection refresh on canvas resize; adds cf_draw_on_pixel_scale_changed() for AA recalculation.
src/cute_app.cpp	Implements explicit pixel-scale APIs, changes window sizing to logical points (via content scale), and makes canvas sizing persistent unless explicitly changed.
samples/hidpi.c	Updates sample to demonstrate the new manual recipe and adds interactive “follow vs forced” scale toggles + mouse coordinate readout.
include/cute_app.h	Documents the new public API and the revised HiDPI contract (display scale informational; pixel scale user-controlled).

Review details

Suppressed comments (1)

src/cute_input.cpp:672

• Indentation/formatting in the SDL_EVENT_MOUSE_BUTTON_UP case has misaligned braces and space indentation; it should match the surrounding tab-indented style for consistency.

			} else if (event.button.clicks == 2) {
				app->mouse.click_type = CF_MOUSE_CLICK_DOUBLE;
                        }
                }	break;

• Files reviewed: 12/12 changed files
• Comments generated: 3
• Review effort level: Lite

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 62fadad2fd

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Update stale HiDPI docs for the new resize contract

With this new persistent-canvas contract, the published docs still tell users the opposite: docs/topics/hidpi.md:24-26 says CF recreates the app canvas on resize/scale changes and cf_app_set_canvas_size is one-shot, docs/topics/application_window.md:100 says the same, and include/cute_draw.h:2214-2215 still says CF resets the projection on resize. Users following those pages won't call cf_app_update_display() after cf_app_was_resized(), so their resizable windows keep the old canvas/projection and render stretched or clipped.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Align cf_make_app docs with logical-point sizing

After this change, cf_make_app treats w/h as logical points and multiplies by the content scale before creating the SDL window. On Windows/X11 at 150% scaling, callers following include/cute_app.h:224-225 (pixels) who request 640x480 now get a 960x720 physical window; either the constructor docs/examples need to move to logical points along with cf_app_set_size, or this scaling should not happen here.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Stop forcing display scale to 1 for NO_HIGH_DPI

In the CF_APP_OPTIONS_NO_HIGH_DPI_BIT case, this only pins the applied pixel scale/backbuffer; SDL still reports the OS display/content scale on Windows and X11 (e.g. 1.5 at 150% desktop scaling), and cf_make_app stores that value in app->display_scale. This assertion will fail on those supported HiDPI desktops and would force cf_app_get_display_scale() to stop reporting the OS scale that the new API docs tell users to follow.

Useful? React with 👍 / 👎.

[pusewicz]

@RandyGaul @bullno1 I trimmed the PR. Please have a look and let me know what needs clarifying.