logs.gokuls.in

— blogr init followed by blogr serve fails with "Theme 'minimal-retro' not found" because the config stores "minimal-retro" but the theme registers as "Minimal Retro".

• Add normalize_theme_name() in blogr-themes that lowercases and replaces hyphens/underscores with spaces, so "minimal-retro", "Minimal Retro", and "minimal_retro" all resolve correctly
• Add ThemeInfo::slug() for canonical config-file form (lowercase, hyphenated)
• Fix all write paths (theme set, config set, TUI) to store the slug
• Fix all comparison paths to use normalization
• Fix newsletter test that relied on the broken lookup behavior

Test plan

• cargo test --workspace — all 134 tests pass
• cargo clippy --all-targets --all-features -- -D warnings — clean
• cargo fmt --all -- --check — clean
• Manual: blogr init test-blog && cd test-blog && blogr serve should work without error

• Step outputs were not propagated between composite action steps — execute_composite_action always passed an empty HashMap for step_outputs, so ${{ steps.<id>.outputs.<key> }} always resolved to empty strings within composite actions.
• Step env values with ${{ }} expressions were not resolved — values like toolchain: ${{inputs.toolchain}} were inserted as literal strings, never run through the expression evaluator.
• Both bugs together caused dtolnay/rust-toolchain (and likely other composite actions) to fail in emulation mode.

What changed

1. Added a composite_step_outputs map in execute_composite_action and call apply_step_environment_updates after each step — same pattern as normal job execution.

2. Added preprocess_expressions call on step env values in execute_step, after secret substitution.

Test plan

• cargo fmt --check passes
• cargo clippy --workspace --all-features -- -D warnings passes
• All 396 existing tests pass
• wrkflw validate passes on project workflows
• wrkflw run --runtime emulation .github/workflows/ci.yml now succeeds (previously failed on all jobs)

• Adds a full GitHub Actions expression evaluator that handles ${{ }} expressions with operators, context references, and built-in functions
• Fixes composite actions like dtolnay/rust-toolchain that use complex expressions (e.g. ${{ steps.parse.outputs.toolchain == 'nightly' && inputs.components && ' --allow-downgrade' || '' }})
• Replaces the hardcoded evaluate_job_condition pattern-matcher with the expression evaluator for if: conditions
• Adds missing inputs.*, github.*, runner.* context substitution and RUNNER_OS/RUNNER_ARCH env vars

What changed

Expression evaluator capabilities

• Operators: ==, !=, <, <=, >, >=, &&, ||, !
• Literals: strings ('hello'), numbers, booleans, null
• Contexts: inputs.*, env.*, github.*, runner.*, matrix.*, steps.*.outputs.*
• Functions: contains(), startsWith(), endsWith(), format(), success(), failure(), always(), cancelled()
• Truthiness: matches GitHub Actions semantics (empty string, 0, false, null are falsy)

Test plan

• 258 unit tests pass (cargo test -p wrkflw-executor --lib)
• Clippy clean (cargo clippy -p wrkflw-executor -- -D warnings)
• Rustfmt clean
• Verified against project's own CI workflow — ${{runner.os}} resolves to macOS correctly
• dtolnay/rust-toolchain expression pattern evaluated correctly in unit tests

• Centralized theme module (theme.rs) — single source of truth for all colors, styles, symbols, and block/badge helpers. Eliminates ~200 hardcoded Color:: references scattered across 11 files.
• Consistent Unicode symbols — replaces the mixed emoji (✅❌⏭⟳) with proper single-cell-width Unicode (✔✖⊘◉○▸) that don't break column alignment in terminals.
• Upgraded ratatui 0.23→0.28 and crossterm 0.26→0.28 — migrated all breaking API changes (Frame generics removed, Table::new signature, f.size()→f.area()).
• Colored CLI output (cli_style.rs) — validate, execute, and list commands now have proper colored output with tree-style rendering (├──└──), replacing raw println! with zero styling.
• Braille spinner animation for running workflow states, cycling on tick.
• Visual polish — numbered tabs (1·Workflows), box-drawing dividers, [✔]/[ ] checkboxes, underlined table headers, themed status badges, compact status bar hints, breadcrumb navigation in job detail, cleaned up help tab with fixed-width key columns.
• Removed redundant UI — instruction headers in workflows/logs tabs (status bar already shows same hints), reclaiming vertical space.

Net diff: -97 lines (1287 added, 1384 removed). The UI got better *and* smaller.

Test plan

• cargo build compiles cleanly
• cargo test — all 16 UI tests + full workspace tests pass
• cargo run -- tui — visually verify all 4 tabs render correctly
• cargo run -- validate .github/workflows/ — verify colored CLI output
• cargo run -- list --jobs — verify tree-style colored listing
• Test in both wide (120+ col) and narrow (80 col) terminals
• Verify spinner animates for running workflows