[dotCooCoo]

What and why

VectorCraft doesn't run scripts written for the reference app, and File › Scripts held only the disabled Other
Script…. This change adds File › Scripts › Check Script…, which picks a .jsx, .jsxinc or .jsxbin file, reads it
without running any of it, and shows a report:

• what the script needs: the script language's own objects, the reference app's document model, files, menu
  commands, dialogs, or features only the reference app has;
• what it would do outside the document: write, move or delete files, open files in other apps, open network
  connections, load native code, send messages to other apps, read or set environment variables, run code the check
  can't read, or change the reference app's preferences;
• the menu-command strings it passes, the document-model member chains and type names it uses, its directives, the
  files it includes, and the problems that would stop it loading.

Copy Report copies the report as JSON. File › Open and drag-and-drop show the report for those files too. #844 asks
for scripts to run; Check Script… does not run them.

• vectorcraft-scriptcheck is a new Standalone crate with no dependencies. It holds a lexer for ECMAScript 3 with E4X
  literals and the #name and //@name directive lines, and an analysis of the tokens that follows variables,
  function scopes and the files a script includes. Neither the lexer nor the analysis recurses. The crate reads at
  most 8 MiB of a file and 32 MiB with the included files, follows 64 included files 16 deep, looks up at most 1,024
  paths for them, keeps 1,000 entries in a list, tracks 1,000 variables holding member chains and 1,000 holding File
  objects, and cuts echoed text to 120 characters.
• script.check (engine, a query) checks a script given as path, data or dataBase64. With a path and
  includes: true, it also reads the files the #include lines name from inside the script's folder, through the
  automation read root (mcp: add --automation-read-root / --automation-write-root so the server can be confined to a folder #832). ui.checkScript shows the report, and docs/mcp.md describes the reply.

The language rules come from ECMA-262 3rd edition and ECMA-357 2nd edition. The directive, platform and
document-model root names come from the public JavaScript Tools Guide pages (https://extendscript.docsforadobe.dev/)
and the scripting guide's Application page (https://ai-scripting.docsforadobe.dev/jsobjref/Application/), and
executeMenuCommand, doScript, loadAction, unloadAction and selectTool come from two public forum posts.
crates/scriptcheck/src/catalog.rs cites each page by URL next to the names taken from it. Only names are used, and
the check does not map them to VectorCraft commands. The check recognizes a compiled script by its .jsxbin name or
a leading @JSXBIN@, as EffectCraft's script commands do (storytold/effectcraft
crates/engine/src/commands/file_more.rs:535 at 9a7d3e7). Every script in the tests is written from scratch with
invented names.

How it was tested

Tests that fail without the change:

• crates/scriptcheck/src/tests.rs: each of the 34 feature ids on a synthetic script with the feature and on one
  without it; directives in both forms; member chains, assignments, and the names a function or a catch block
  declares; menu commands, effects outside the document and problems, code in an XML literal's braces included;
  includes over files in memory (folders, cycles, every limit, and variables that one file sets and another uses);
  100,000 levels of nesting on a 256 KiB stack; and properties: the tokens cover the text exactly once, the work
  stays linear in the size, compiled input isn't read, and include graphs keep to every limit.
• crates/engine/src/tests_scripts.rs: script.check from a path, from text and from base64; included files read only
  when asked; a link out of the script's folder; the automation read root; a pipe or a folder named as a script or an
  include; kinds, limits and bad params. open_exts_cover_every_readable_format checks the Scripts filter.
• crates/ui-egui/src/tests_script_check.rs: the menu item, the report window and Copy Report, each verdict, a label
  for every id a report can hold, File › Open (which doesn't read the whole file), and the file picker. A test in
  tests_place.rs drops a script on the window.
• crates/engine/tests/import_fuzz.rs: garbage, generated and mutated scripts through script.check.

cargo clippy -p <crate> --all-targets -- -D warnings passes for vectorcraft-scriptcheck, vectorcraft-engine,
vectorcraft-ui-egui and xtask, and cargo xtask brands, layers and assets pass.

On macOS 27.0.1 (Apple Silicon) with rustc 1.99.0, on top of main at d9e4435:

• CRAFT_FONTS_DIR=<craft-fonts checkout> cargo xtask ci passes its 7 steps: 6,080 tests passed, 0 failed, 6 ignored.
• PROPTEST_CASES=20000 cargo test --release -p vectorcraft-engine --test import_fuzz: 65 passed and 0 failed in
  698 seconds.

git merge-tree reports no conflicts with main at 0573d3c.

#1183 also changes File › Scripts. With both merged, Check Script… goes into #1183's scripts::menu() above Other
Script…, file_scripts_holds_check_script_above_other_script then expects its ui.runScript item, and #1183's
without_a_scripts_folder_only_other_script_is_listed expects Check Script… before Other Script…. Keeping both sides
resolves the other conflicts in Cargo.toml, Cargo.lock, xtask/src/layers.rs, crates/ui-egui/src/lib.rs,
picks.rs, menus.rs and the Automation list in docs/target-app-parity.md. #1183 itself conflicts with main at
0573d3c in ROADMAP.md, crates/engine/src/lib.rs, docs/gaps.md, docs/roadmap.md, docs/target-app-parity.md
and xtask/src/layers.rs.

With #1223 merged, this branch conflicts in fileio/mod.rs, fileio/tests.rs, io.rs, place.rs and
docs/mcp.md. Keeping both sides resolves each one, with this branch's read_file_head, which refuses anything but a
regular file, and nine filters besides the formats in open_exts_cover_every_readable_format. Both pull requests add
the rows "Copy Report" and "Not read" to the cs, de, es, fr, hu, it, ja, ru, uk, zh-hans and zh-hant catalogs, so the
pull request merged second drops its own rows there. In cs, ja and ru the two pull requests translate "Not read"
differently. #1223 itself conflicts with main at 0573d3c in io.rs, where main's background open (023c8e3)
replaced the open_path body that #1223 changes, and in docs/roadmap.md.

Checklist

• cargo xtask ci passes (fmt, clippy, tests, assets, brands, layers, wasm).
• No panics in shipped code: no unwrap(), expect(), panic!, unreachable!, todo! or
  unimplemented! outside tests, and no unchecked indexing on data from files, parameters or
  messages. Errors are returned as Result and shown to the user
  (why and how).
• Clean-room: nothing copied from Adobe products or GPL/AGPL code (AGENTS.md).
• Every new asset has a row in ASSETS.md.

[bflatastic]

Thanks @dotCooCoo, merged. I read it for what it does with a script: the checker is a hand-written scanner with no new dependency and forbid(unsafe_code), nothing in a script is run or fetched, included files are only read from inside the script's folder through the read root, and the names it recognises are cited from public pages with test scripts written from scratch. Since it only reports, it doesn't touch the open question of running such scripts (#1210). Checked on a release build: File › Scripts › Check Script… on a small script shows its needs, the file it writes, its menu command and member chains; an #include of ../outside/… is listed as not followed; a module and a compiled script are reported as such over MCP.