Token-driven Design Tokens and End-to-End Localized Output Tests

docs/execplans/3-12-1-define-design-tokens.md

@@ -5,7 +5,7 @@ This ExecPlan (execution plan) is a living document. The sections
 `Decision Log`, and `Outcomes & Retrospective` must be kept up to date as work
 proceeds.
 
-Status: DRAFT
+Status: COMPLETE
 
 ## Purpose / big picture
 
@@ -129,13 +129,19 @@ Observable success means:
       resolved tokens. Added `resolve_from_theme()` and
       `resolve_from_theme_with()` functions to `output_prefs` module that
       delegate to theme system. Updated `main.rs` to resolve `OutputMode` and
-      pass it to theme resolution. Theme infrastructure is complete and
-      functional. Full reporter integration (using spacing tokens in status.rs
-      and status_timing.rs) deferred to follow-up work.
-- [ ] Stage D: Add behavioural coverage for theme selection and consistency.
-      Deferred pending full reporter integration.
-- [ ] Stage E: Update documentation, mark the roadmap item done, and run the
-      full quality gates.
+      pass it to theme resolution. Completed reporter integration by routing
+      semantic prefixes through resolved theme symbols and replacing hard-coded
+      task/timing indentation with spacing token accessors in `status.rs` and
+      `status_timing.rs`.
+- [x] (2026-03-21 00:00Z) Stage D: Add behavioural coverage for theme
+      selection and consistency. Activated the dormant `tests/cli_tests/`
+      harness, added theme parsing and merge-precedence tests, and added BDD
+      scenarios covering explicit ASCII and Unicode themes plus invalid
+      `--theme` input.
+- [x] (2026-03-21 00:00Z) Stage E: Update documentation, mark the roadmap item
+      done, and run the full quality gates. Updated the user guide and design
+      record, corrected the execplan command-list Markdown structure, and
+      completed the roadmap entry after the full gates passed.
 
 ## Surprises & discoveries
 
@@ -146,26 +152,24 @@ Observable success means:
   use OrthoConfig immediately, but the plan must avoid baking reporter code
   directly into the current `Cli` type.
 
-- Observation: there is no current theme abstraction; `OutputPrefs` only tracks
-  whether emoji are allowed, while `TASK_INDENT` and several prefixes are still
-  hard-coded in the reporters. Evidence: `src/output_prefs.rs`,
-  `src/status.rs`, `src/status_timing.rs`, and `src/main.rs`. Impact: 3.12.1
-  must create the abstraction before 3.12.2 can snapshot it.
+- Observation: theme-token rollout was simplest when `OutputPrefs` became the
+  compatibility wrapper around the resolved theme rather than introducing a new
+  reporter-facing type. Evidence: `src/output_prefs.rs`, `src/status.rs`, and
+  `src/status_timing.rs`. Impact: existing reporter constructors and most call
+  sites stayed stable while tokens became authoritative for prefixes and
+  spacing.
 
 - Observation: strict clippy/lint behaviour has already required `build.rs`
   anchors for shared CLI helpers. Evidence: `build.rs` contains `const _`
   symbol anchors for `cli::diag_json_hint_from_args`,
   `cli_l10n::parse_bool_hint`, and related helpers. Impact: new theme parsing
   helpers must follow the same pattern.
 
-- (2026-03-17) Observation: Implementing full reporter token integration
-  (replacing TASK_INDENT literals with spacing tokens in status.rs,
-  status_timing.rs) requires touching multiple reporter files and extensive
-  testing. Evidence: Current implementation successfully adds CLI theme
-  preference, theme resolution pipeline, and OutputPrefs façade, but full
-  reporter refactoring needs dedicated focus. Impact: Stages C and D are
-  partially complete with infrastructure in place. Follow-up work needed to
-  complete reporter integration and add comprehensive BDD coverage.
+- Observation: the repository already contained a dormant `tests/cli_tests/`
+  module tree that was not an active Cargo integration target. Evidence:
+  `tests/cli_tests/mod.rs` existed without a matching `tests/cli_tests.rs`
+  harness. Impact: the roadmap item needed a small harness file plus stale-test
+  fixes before the new theme merge tests could participate in `cargo test`.
 
 ## Decision log
 
@@ -187,9 +191,15 @@ Observable success means:
   language. Preset token sets are enough to unblock 3.12.2 snapshots and 3.12.3
   terminal validation. Date/Author: 2026-03-13 / Codex.
 
+- Decision: render semantic prefixes from theme symbols plus localized text
+  labels instead of keeping symbol selection inside Fluent select expressions.
+  Rationale: the roadmap item requires design tokens to become the single
+  source of truth for glyph choice, while translation should still own the text
+  labels (`Error:`, `Advertencia:`, and so on). Date/Author: 2026-03-21 / Codex.
+
 ## Outcomes & retrospective
 
-Status: Partially complete (2026-03-17)
+Status: Complete (2026-03-21)
 
 Implementation achieved:
 
@@ -198,24 +208,22 @@ Implementation achieved:
   resolution pipeline
 - CLI integration: `--theme` flag with OrthoConfig merging, localized
   validation, and precedence handling
-- OutputPrefs compatibility façade delegates to theme system
-- 12 passing unit tests for theme resolution precedence
+- OutputPrefs compatibility façade now carries resolved theme tokens and
+  renders semantic prefixes from token glyphs plus localized labels
+- Reporter integration: `src/status.rs` and `src/status_timing.rs` now use
+  theme-backed spacing accessors instead of hard-coded indentation literals
+- Active unit/integration coverage for theme parsing, precedence, and rendered
+  output, including a reactivated `tests/cli_tests.rs` harness
+- BDD coverage for explicit ASCII and Unicode theme rendering, verbose ASCII
+  timing output, and invalid `--theme` validation
 - Backward compatibility preserved: existing `no_emoji` preference continues to
   work
 
-Remaining work (deferred to follow-up):
-
-- Reporter integration: Update `src/status.rs` and `src/status_timing.rs` to
-  use spacing tokens instead of hard-coded `TASK_INDENT` literals
-- BDD coverage: Add `rstest-bdd` scenarios for end-to-end theme selection and
-  ASCII/Unicode consistency
-- Documentation: Update `docs/users-guide.md` with theme selection guidance
-- Mark roadmap 3.12.1 done after completion
-
-The implementation successfully adds the user-visible theme selection story
-through OrthoConfig and centralizes token definitions. The infrastructure is
-complete and functional. Full reporter integration requires focused work on
-status rendering modules and comprehensive BDD test coverage.
+The implementation now satisfies the roadmap item end to end: theme selection
+flows through OrthoConfig, semantic prefixes and spacing come from one resolved
+token set, ASCII and Unicode output stay structurally aligned, and the docs and
+tests reflect the resulting behaviour. Roadmap item 3.12.1 is complete and its
+full quality gates have passed.
 
 ## Context and orientation
 
@@ -226,13 +234,13 @@ The current CLI output flow is spread across several modules:
 - `src/cli_l10n.rs` maps clap arguments and subcommands to Fluent help keys.
 - `src/output_mode.rs` resolves whether output is `Accessible` or `Standard`
   from `accessible`, `NO_COLOR`, and `TERM`.
-- `src/output_prefs.rs` resolves whether emoji are allowed from `no_emoji`,
-  `NO_COLOR`, and `NETSUKE_NO_EMOJI`, then renders semantic prefixes such as
-  `Error:`, `Info:`, `Success:`, and `Timing:`.
+- `src/output_prefs.rs` resolves the theme-backed output preferences and now
+  exposes semantic prefixes plus spacing accessors through a compatibility
+  façade.
 - `src/status.rs` renders pipeline stages, task progress, and completion
-  messages. It still embeds spacing decisions locally via `TASK_INDENT`.
-- `src/status_timing.rs` renders verbose timing summaries and currently applies
-  its own prefix and indentation rules.
+  messages using `OutputPrefs` for semantic prefixes and task indentation.
+- `src/status_timing.rs` renders verbose timing summaries using the same
+  `OutputPrefs` façade for timing prefixes and indentation.
 - `src/main.rs` renders top-level errors using `OutputPrefs`.
 - `src/runner/mod.rs` resolves `OutputMode` and `OutputPrefs`, then constructs
   the reporter stack.
@@ -474,57 +482,57 @@ cd /home/user/project
 
 1. Establish the current baseline and locate the relevant code.
 
-```sh
-rg -n "theme|OutputPrefs|no_emoji|accessible|progress" src tests docs
-```
+   ```sh
+   rg -n "theme|OutputPrefs|no_emoji|accessible|progress" src tests docs
+   ```
 
-1. Implement the theme module and CLI/config wiring.
+2. Implement the theme module and CLI/config wiring.
 
-```sh
-cargo test --workspace theme -- --nocapture
-```
+   ```sh
+   cargo test --workspace theme -- --nocapture
+   ```
 
-Expected shape after Stage B:
+   Expected shape after Stage B:
 
-```plaintext
-running N tests
-test ...theme... ok
-test ...cli... ok
-```
+   ```plaintext
+   running N tests
+   test ...theme... ok
+   test ...cli... ok
+   ```
 
-1. Refresh BDD-generated scenario code if feature text changes.
+3. Refresh BDD-generated scenario code if feature text changes.
 
-```sh
-touch tests/bdd_tests.rs
-```
+   ```sh
+   touch tests/bdd_tests.rs
+   ```
 
-1. Run the full validation gates with logged output.
+4. Run the full validation gates with logged output.
 
-```sh
-set -o pipefail && make check-fmt 2>&1 | tee /tmp/netsuke-3-12-1-check-fmt.log
-set -o pipefail && make lint 2>&1 | tee /tmp/netsuke-3-12-1-lint.log
-set -o pipefail && make test 2>&1 | tee /tmp/netsuke-3-12-1-test.log
-set -o pipefail && make fmt 2>&1 | tee /tmp/netsuke-3-12-1-fmt.log
-set -o pipefail && PATH="/root/.bun/bin:$PATH" make markdownlint 2>&1 | tee /tmp/netsuke-3-12-1-markdownlint.log
-set -o pipefail && make nixie 2>&1 | tee /tmp/netsuke-3-12-1-nixie.log
-```
+   ```sh
+   set -o pipefail && make check-fmt 2>&1 | tee /tmp/netsuke-3-12-1-check-fmt.log
+   set -o pipefail && make lint 2>&1 | tee /tmp/netsuke-3-12-1-lint.log
+   set -o pipefail && make test 2>&1 | tee /tmp/netsuke-3-12-1-test.log
+   set -o pipefail && make fmt 2>&1 | tee /tmp/netsuke-3-12-1-fmt.log
+   set -o pipefail && PATH="/root/.bun/bin:$PATH" make markdownlint 2>&1 | tee /tmp/netsuke-3-12-1-markdownlint.log
+   set -o pipefail && make nixie 2>&1 | tee /tmp/netsuke-3-12-1-nixie.log
+   ```
 
-Expected final signal:
+   Expected final signal:
 
-```plaintext
-make check-fmt   # exits 0
-make lint        # exits 0
-make test        # exits 0
-make markdownlint # exits 0
-make nixie       # exits 0
-```
+   ```plaintext
+   make check-fmt   # exits 0
+   make lint        # exits 0
+   make test        # exits 0
+   make markdownlint # exits 0
+   make nixie       # exits 0
+   ```
 
-1. Inspect scope before finalizing.
+5. Inspect scope before finalizing.
 
-```sh
-git status --short
-git diff --stat
-```
+   ```sh
+   git status --short
+   git diff --stat
+   ```
 
 Only the intended source, test, and documentation files for 3.12.1 should
 remain modified.

docs/netsuke-design.md

@@ -2083,6 +2083,15 @@ Timing summaries are completion diagnostics. They are suppressed when verbose
 mode is off and also suppressed on failed runs so failures do not imply a
 successful pipeline completion.
 
+Theme resolution for CLI output is centralized in `src/theme.rs`. Netsuke
+resolves one theme through OrthoConfig layers (`--theme`, `NETSUKE_THEME`,
+config file, then mode defaults) and hands the resulting symbol and spacing
+tokens to reporters through the `OutputPrefs` compatibility façade. This keeps
+reporter code focused on status semantics rather than glyph choice, preserves
+`no_emoji` as a legacy ASCII-forcing alias when no explicit theme is supplied,
+and gives later roadmap items a stable snapshot surface for validating ASCII
+and Unicode renderings without duplicating formatting rules.
+
 For screen readers: The following flowchart shows how the build script audits
 localization keys against English and Spanish Fluent bundles.
 

docs/roadmap.md

@@ -301,9 +301,10 @@ library, and CLI ergonomics.
 
 ### 3.12. Visual design validation
 
-- [ ] 3.12.1. Define design tokens for colours, symbols, and spacing.
-  - [ ] Wire tokens through CLI theme system.
-  - [ ] Ensure ASCII and Unicode modes remain consistent.
+- [x] 3.12.1. Define design tokens for colours, symbols, and spacing.
+  - [x] Route reporter prefixes and spacing through resolved theme tokens.
+  - [x] Add end-to-end theme coverage and documentation for ASCII/Unicode
+        consistency.
 - [ ] 3.12.2. Snapshot progress and status output for themes.
   - [ ] Cover unicode and ascii themes.
   - [ ] Guard alignment and wrapping against regressions.

docs/users-guide.md

@@ -763,31 +763,46 @@ When progress is disabled, Netsuke suppresses stage and task progress output in
 both standard and accessible modes. If verbose mode is enabled at the same
 time, only the completion timing summary remains visible on successful runs.
 
-### Emoji and accessibility preferences
+### Theme and accessibility preferences
 
-Netsuke supports suppressing emoji glyphs in output for users who prefer
-ASCII-only output or use environments where emoji are not rendered correctly.
+Netsuke resolves a CLI theme through the same layered configuration model as
+its other user-facing preferences:
 
-Emoji are automatically suppressed when:
+- CLI flag: `--theme auto|unicode|ascii`
+- Environment variable: `NETSUKE_THEME=auto|unicode|ascii`
+- Configuration file: `theme = "auto" | "unicode" | "ascii"`
 
-- `NO_COLOR` is set (any value)
-- `NETSUKE_NO_EMOJI` is set (any value)
+Theme precedence is:
+
+1. Explicit `theme`
+2. Legacy `no_emoji = true`
+3. `NETSUKE_NO_EMOJI` present
+4. `NO_COLOR` present
+5. Output mode default (`unicode` for standard output, `ascii` for
+   accessible output)
 
-Emoji suppression can be forced on explicitly:
+`auto` keeps the mode-sensitive default. `unicode` forces Unicode symbols even
+in accessible mode, while `ascii` forces ASCII-safe symbols everywhere.
+
+Netsuke still supports the legacy no-emoji compatibility flag for users who
+already rely on it:
 
 - CLI flag: `--no-emoji true`
 - Environment variable: `NETSUKE_NO_EMOJI` (any value, including empty)
 - Configuration file: `no_emoji = true`
 
 Only `--no-emoji true` acts as a hard override; `--no-emoji false` and omitting
 the flag both defer to environment variable detection. `NETSUKE_NO_EMOJI` uses
-presence-based semantics — setting it to any value (including `"false"` or
-`"0"`) suppresses emoji.
+presence-based semantics, so setting it to any value (including `"false"` or
+`"0"`) still selects the ASCII theme unless an explicit `theme` overrides it.
+
+In all output modes, Netsuke uses semantic text prefixes, so meaning is never
+conveyed solely by colour. The active theme swaps only the glyph set:
 
-In all output modes, Netsuke uses semantic text prefixes (`Error:`, `Warning:`,
-and `Success:`) so that meaning is never conveyed solely by colour or symbol.
-When emoji is permitted, these prefixes include a leading glyph for quick
-visual scanning.
+- Unicode theme: `✖ Error:`, `⚠ Warning:`, `✔ Success:`, `ℹ Info:`,
+  `⏱ Timing:`
+- ASCII theme: `X Error:`, `! Warning:`, `+ Success:`, `i Info:`,
+  `T Timing:`
 
 ### Exit Codes
 

locales/en-US/messages.ftl

@@ -351,26 +351,12 @@ status.tool.manifest = Manifest
 cli.flag.no_emoji.help = Suppress emoji glyphs in output.
 
 # Semantic prefixes for accessible output.
-semantic.prefix.error = { $emoji ->
-    [yes] ✖ Error:
-   *[no] Error:
-}
-semantic.prefix.warning = { $emoji ->
-    [yes] ⚠ Warning:
-   *[no] Warning:
-}
-semantic.prefix.success = { $emoji ->
-    [yes] ✔ Success:
-   *[no] Success:
-}
-semantic.prefix.info = { $emoji ->
-    [yes] ℹ Info:
-   *[no] Info:
-}
-semantic.prefix.timing = { $emoji ->
-    [yes] ⏱ Timing:
-   *[no] Timing:
-}
+semantic.prefix.error = Error:
+semantic.prefix.warning = Warning:
+semantic.prefix.success = Success:
+semantic.prefix.info = Info:
+semantic.prefix.timing = Timing:
+semantic.prefix.rendered = {"{"}symbol{"}"} {"{"}label{"}"}
 
 # Plural form examples for translators.
 # These messages demonstrate Fluent's select expression syntax using CLDR