Complete all open beads tasks (hm-68a, hm-isi, hm-3w5, hm-3rr + close hm-6if/hm-wqb/hm-6qj)#16
Merged
Merged
Conversation
Closes hm-68a. Introduces a shared `crate::suggest` module backed by a small Levenshtein DP and used in two surfaces: - `himitsu search <query>` now prints `did you mean <path>?` to stderr when the result list is empty, the query is non-tag-only, and the closest existing secret is within the heuristic distance threshold. - A reusable `SecretRefAutocomplete` widget in `tui::widgets` opens beneath the search bar on Ctrl+Space, surfacing up to five closest paths via Up/Down/Enter. The popup is non-modal — the query field keeps consuming printable keys. Also folds in `cargo fmt --all` reformatting and a couple of `clippy --fix` cleanups (unused imports, manual_contains) from the same toolchain pass.
Closes hm-3w5 and hm-3rr. * hm-3w5 — every field in `new_secret.rs` now renders a muted placeholder example when empty + unfocused (path: prod/api/STRIPE_KEY, url: https://example.com, expires_at: never | 30d | 6mo | …, etc.). Generalised the rendering through a single `field_paragraph` helper so the multi-line value field and every single-line field follow the same focus/empty/placeholder rules. Added the same support to the generic `forms::FormView` (used by `remote_add` and any future `ProtoForm`) via a `placeholder` builder on `Field`. * hm-3rr — added a final `[ submit ]` button row at the end of the tab cycle. Enter on the submit step runs the existing `submit()` path. Pressing Esc with any populated field opens a centered unsaved-changes modal with `[ keep editing ] [ save ] [ discard ]` (Left/Right cycle, Enter activates, Esc dismisses to keep editing). An empty form still bails immediately so the prompt only fires when the user actually has something to lose. Tests cover: cycle includes Submit + wraps; Shift-Tab wraps to Submit; Enter on Submit submits; Esc-on-empty cancels; Esc with content opens modal; Right/Left cycle the focus; Enter on Discard emits Cancel; Esc inside modal restores the form. forms::FormView gains 3 placeholder rendering tests.
Closes hm-isi. Adds a `Hint` surface for ambient, persistent guidance text in the bottom-left of the frame — distinct from the toast lane, which is transient and full-width. * `tui::hint` — new module with a `Hint` struct, `MAX_LEN = 60`, multi- byte-safe truncation, and a `render` that paints only the left third of the bottom strip so a coexisting toast can claim the right portion. * `tui::app` — `set_hint`/`clear_hint` accessors plus action wiring on `NewSecretAction::SetHint(_)` / `ClearHint`. Toast suppresses the hint visually (3s TTL means the hint reappears as soon as the toast clears); cancel/created/failed routes also clear the hint state explicitly. * `tui::views::new_secret` — `on_key` snapshots the focused step, detects step transitions caused by navigation keys, and translates them into `SetHint`/`ClearHint` via `hint_for_step`. Tipped steps are Path, Tags, ExpiresAt; everything else clears the strip. Submit/save outcomes are forwarded untouched so an internal validation refocus doesn't churn the action enum. * `tui::harness` integration test (`ambient_hint_tracks_focus_and_clears _on_cancel`) drives the new-secret flow via real key events and asserts the hint is seeded on open, swaps with focus, and clears on cancel.
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: c7a40ce17e
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 (@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 (@codex) address that feedback".
| // possibly land on: every secret path search_core just returned for | ||
| // an unfiltered scan. This is cheap (already in memory) and dodges | ||
| // having to re-walk the store when the popup wants to open. | ||
| let corpus: Vec<String> = self.results.iter().map(|r| r.path.clone()).collect(); |
There was a problem hiding this comment.
When the user types a typo that produces zero search hits, refresh_results sets the autocomplete corpus from self.results, which has already been filtered by that typo, so the corpus is empty and Ctrl+Space cannot show the closest existing secret. This defeats the new typo-recovery path in the TUI; the corpus needs to come from an unfiltered scan (as the CLI suggestion path does with search_core(ctx, "", &[])) rather than the current filtered results.
Useful? React with 👍 / 👎.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Closes every open beads issue in one branch. Three were already implemented on
mainand just needed closing; four required new code.Implemented
crate::suggestmodule (inline DP, no extra crate).himitsu searchprintsdid you mean <path>?to stderr on a zero-result query. ReusableSecretRefAutocompletewidget intui::widgetsis wired into the search bar; Ctrl+Space toggles it, Up/Down navigate, Enter accepts.tui::hintsurface, distinct from the toast lane. Persistent muted text in the left third of the bottom row; auto-suppressed when a toast is on screen. The new-secret form publishes step-aware tips forpath(slashes),tags(comma-separated), andexpires_at(duration grammar) viaNewSecretAction::SetHint/ClearHint.placeholderbuilder onforms::Fieldplus a sharedfield_paragraphhelper innew_secret.rsso every input renders a muted example when empty + unfocused. Concrete placeholders for path/value/description/tags/url/totp/env_key/expires_at.Step::Submittab stop renders a[ submit ]button row; Enter on it runs the existingsubmit(). Esc with any populated field opens a centeredunsaved changesmodal with[ keep editing ] [ save ] [ discard ](Left/Right cycle, Enter activates, Esc dismisses).Already implemented; closed in
.beads/issues.jsonlset/write/search/ls/tag), env DSL (tag:selectors), TUI (Tags step, viewer chips, search row chips). Verified by 90+ tag-related tests.path:in the secret-viewer edit doc renames the secret while preservingcreated_at/history (finish_edit_renames_secret_and_preserves_history).Test results
cargo fmt --all -- --check— clean.cargo clippy --workspace --all-targets -- -D warnings— clean.cargo test --workspace— 602/602 binary tests pass; 94/97 integration tests pass. The 3 failing tests (test_check_*) are pre-existing onmain(a git env quirk in the test setup) and unrelated to this work.Notes for reviewers
Tabis already bound to fold-toggle in the search view. Documented at the call site.cargo fmt/clippy --fixswept up two pre-existing lint nits (unused_importsincli/join.rstests andmanual_containsincommand_palette.rstests). Folded into the suggest commit because the gate requires them.Test plan
cargo buildsucceeds locally.cargo fmt --all -- --checkis clean.cargo clippy --workspace --all-targets -- -D warningsis clean.cargo test --workspace— all binary tests pass; the only integration failures are the 3 pre-existingtest_check_*.n→ see the placeholder examples + bottom-left hint.unsaved changesmodal appears.himitsu search <typo>from the CLI to seedid you mean …?.https://claude.ai/code/session_01XGkUiFcnS4UX8ZHLGE7p2C
Generated by Claude Code