mouse: 5 cards — linq_fold Phase 2C + dasImgui HDPI/build

[borisbat]

Summary

Five blind-mouse Q&A cards that accumulated across two sessions:

linq_fold Phase 2C (from PR #2697 — should have shipped together):

• chained-select-splice-bind-via-clone-assign-universal.md — chained-select splice binds via := (clone-assign) universally; prevWorkhorse-only guard dropped.
• macro-planner-named-marker-arms-leave-room-for-future-modes.md — macro planner uses named-marker arms (take_skip_emit, chained_select_emit, …) instead of dense if/else so adding a mode later is local.
• splice-macro-bounded-loop-guard-take-skip-non-positive-n.md — bounded loop guard for take(N <= 0) / skip(N <= 0): emit nothing rather than panic.

dasImgui HDPI session (from borisbat/dasImgui#42):

• how-do-i-make-dasimgui-hdpi-aware-what-s-the-canonical-scale-plumbing.md — the canonical 3-place plumbing: glfwGetWindowContentScale in live_imgui_init, GLFW_SCALE_TO_MONITOR in harness_init, ScaleAllSizes(unsafe(GetStyle()), scale) after apply_daslang_theme. Includes the float* vs float& binding gotcha (must use safe_addr(xs)).
• how-do-i-run-a-dasimgui-demo-locally-what-s-the-build-daspkg-install-dance.md — local dev flow: build daslang → build dasImgui out-of-tree with DASLANG_DIR=... → daspkg install /path/to/dasImgui --global --force (the --force is needed for every re-sync; -project_root alone won't trigger .das_module registration).

Why

Both sets are pure cache adds. The linq_fold cards encode patterns from already-merged work that next-session-Claude will otherwise re-derive from git history. The dasImgui cards short-circuit the same re-discovery for any future HDPI / theme work.

Test plan

• All 5 cards committed; no other tree changes
• mouse__rebuild after merge (regenerates FTS5 index — not tracked, local-only)

🤖 Generated with Claude Code

[Copilot]

Pull request overview

Adds five “mouse” Q&A cards capturing learnings from the linq_fold Phase 2C work and dasImgui HDPI/build workflows, to reduce future rediscovery of those patterns.

Changes:

• Document take/skip bounded-loop guard semantics for non-positive bounds in splice macros.
• Document a macro-planner structure using named marker arms for future emission modes.
• Document dasImgui HDPI scaling plumbing and a local build + daspkg install workflow.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
mouse-data/docs/splice-macro-bounded-loop-guard-take-skip-non-positive-n.md	Notes correct take/skip guard comparisons to match iterator semantics on non-positive bounds.
mouse-data/docs/macro-planner-named-marker-arms-leave-room-for-future-modes.md	Describes using explicit “marker arms” for planned-but-unimplemented operators in chain recognition.
mouse-data/docs/chained-select-splice-bind-via-clone-assign-universal.md	Captures the bind-via-:= pattern for splicing chained selects and avoiding ExprRef2Value substitution issues.
mouse-data/docs/how-do-i-make-dasimgui-hdpi-aware-what-s-the-canonical-scale-plumbing.md	Summarizes the canonical 3-point DPI scaling plumbing for dasImgui (scale read, GLFW hint, ScaleAllSizes).
mouse-data/docs/how-do-i-run-a-dasimgui-demo-locally-what-s-the-build-daspkg-install-dance.md	Records the local dev flow for building daslang + dasImgui out-of-tree and installing via daspkg.

Comments suppressed due to low confidence (2)

mouse-data/docs/how-do-i-run-a-dasimgui-demo-locally-what-s-the-build-daspkg-install-dance.md:16

• This card hardcodes a contributor-specific absolute path (/Users/borisbatkin/...). For portability and to avoid embedding personal paths in-repo, please switch to placeholders (/path/to/..., $HOME/...) similar to other mouse cards (e.g. /Users/<you>/...).

dasImgui is a **sibling repo** (`/Users/borisbatkin/Work/dasImgui`, not under `daScript/modules/`). To run any dasImgui example or test locally:

**1. Build daslang first** (provides `lib/liblibDaScriptDyn.dylib`, `bin/daslang`):

cd /Users/borisbatkin/Work/daScript
cmake --build build --config Release -j 8

**mouse-data/docs/how-do-i-run-a-dasimgui-demo-locally-what-s-the-build-daspkg-install-dance.md:40**
* The note about `-project_root` is misleading: daslang *does* initialize dynamic modules by scanning `<project_root>/modules` for `.das_module` descriptors (see `include/daScript/ast/dyn_modules.h` and `utils/daScript/main.cpp`). If the intent is "it won’t work when pointing at the dasImgui repo root because it has no `modules/` subdir", please state that explicitly.

Do NOT use -project_root /path/to/dasImgui — that flag exists but doesn't trigger .das_module initialize for register_dynamic_module. The canonical resolver path is via das_root/modules/<name> populated by daspkg.

</details>



---

💡 <a href="/GaijinEntertainment/daScript/new/master?filename=.github/instructions/*.instructions.md" class="Link--inTextBlock" target="_blank" rel="noopener noreferrer">Add Copilot custom instructions</a> for smarter, more guided reviews. <a href="https://docs.github.com/en/copilot/customizing-copilot/adding-repository-custom-instructions-for-github-copilot" class="Link--inTextBlock" target="_blank" rel="noopener noreferrer">Learn how to get started</a>.