INSERT/UPDATE/DELETE + value model + auto-show, with polish

DSL data operations (ADR-0014):
- insert into T [(cols)] values (vals); short form
  insert into T (vals) omits values keyword for friendlier
  syntax.
- update T set ... where col=val | --all-rows; delete from T
  where col=val | --all-rows; show data T.
- Value AST (Number/Text/Bool/Null) with per-column-type
  validation in the executor: int/real/decimal/bool/date/
  datetime/shortid each accept a documented literal shape
  and produce friendly format errors naming the column.
- INSERT short form fills non-auto-generated columns in
  schema order; auto-fills serial via SQLite and shortid
  via the new generator (T2).
- `add column [to table] T: c (type)` -- `to table` now
  optional.

Database:
- insert/update/delete via prepared statements with bound
  rusqlite::types::Value parameters.
- InsertResult/UpdateResult/DeleteResult: writes return
  rows_affected plus the affected row(s) only (not the whole
  table), so users see exactly what changed.
- INSERT shows the just-inserted row via last_insert_rowid.
- UPDATE captures matching rowids up-front and fetches them
  post-update -- works even if the UPDATE changed the WHERE
  column.
- DELETE reports per-relationship cascade effects by row-
  count diffing inbound child tables; UPDATE-side cascades
  are not yet detected (would need value diffing).
- query_data formats cells (booleans true/false, NULLs as
  None).

FK error enrichment:
- Now lists both outbound (INSERT/UPDATE relevance) and
  inbound (DELETE/UPDATE on parent relevance) FKs from the
  metadata, so RESTRICT errors point at the children
  blocking the delete.
- RelationshipSelector has a proper Display impl -- "no
  such relationship" reads cleanly.

Relationship display:
- target_table for AddRelationship/DropRelationship now
  returns the parent (1-side); structure rendering after
  add/drop shows that side's "Referenced by:" entry,
  matching the `from <Parent>` direction of the command.
- [ok] summary uses display_subject so relationship
  commands show both endpoints (`from P.col to C.col`)
  rather than a single misleading table name.
- Auto-name format `<Parent>_<pcol>_to_<Child>_<ccol>`
  (matches the from..to direction).

Output rendering and scrolling:
- Wrap-aware scroll: renderer reports both visible-row
  count and total wrapped-row count to App; scroll math
  caps against actual displayable rows. Long lines wrap;
  the bottom line is always reachable; PageUp/PageDown work
  correctly even after paging past the buffer top.
- Multi-line messages (FK error enrichment, cascade summary)
  split into single-line OutputLines at creation time so
  wrap/scroll math agree.

Runtime / events:
- New AppEvent variants for Insert/Update/Delete success
  carrying typed result structs; DslDataSucceeded reserved
  for show-data queries.

Docs:
- ADR-0014 covers data-op grammar, value model, --all-rows
  safety, auto-show.
- requirements.md: C5 done, T2 done, V2 partial (basic data
  view), V5 partial (show data added). New entries: C5a
  complex WHERE expressions; H1 progress note for FK
  enrichment; H1a (strong syntax-help in parse errors).

Tests: 200 passing (183 lib + 17 integration), 0 skipped.
Includes parser, type-validation, DB write/read, FK-failure
enrichment, cascade-delete propagation, focused-auto-show
behaviour, scroll-cap invariants. Clippy clean with nursery
enabled.
This commit is contained in:
claude@clouddev1
2026-05-07 16:33:25 +00:00
parent 165068269b
commit 305e5083d5
16 changed files with 2638 additions and 109 deletions
+34 -5
View File
@@ -132,7 +132,18 @@ against it.
- [ ] **C4** Convenience: `create m:n relationship from <T1> to
<T2>` produces an auto-named junction table the user can rename;
pulls primary keys and FK definitions automatically.
- [ ] **C5** Data operations: insert / update / delete via DSL.
- [x] **C5** Data operations: insert / update / delete via DSL.
*(ADR-0014. INSERT short and long forms, UPDATE/DELETE with
required WHERE plus `--all-rows` opt-in, `show data <T>`,
per-column-type value-literal validation, FK enforcement
with metadata-driven error enrichment, auto-show after
writes. Bulk insert, complex WHERE expressions, and SELECT
in advanced mode are explicitly tracked separately — see
C5a below.)*
- [~] **C5a** Complex WHERE expressions (AND/OR/comparison
operators/LIKE) for UPDATE/DELETE/show-data filtering. Tracks
the natural progression from DSL into real SQL fluency that
motivates the playground; design and ADR pending.
## SQL handling
@@ -170,10 +181,11 @@ against it.
`int`, `real`, `decimal`, `bool`, `date`, `datetime`, `blob`,
`serial`, `shortid`. *(Mapping to SQLite STRICT covered by
ADR-0005; FK target type rule by ADR-0011.)*
- [ ] **T2** `shortid` generation: base58, 1012 characters,
- [x] **T2** `shortid` generation: base58, 1012 characters,
omits ambiguous characters; generated client-side at insert.
*(Type exists; insert-time generation arrives with the data
insertion path.)*
*(Implemented per ADR-0014; auto-fills omitted shortid
columns and validates user-supplied values against the same
alphabet and length range.)*
- [ ] **T3** Compound primary keys handled end-to-end (DSL,
storage, display, FK reference).
*(Progress: DSL grammar (`with pk a:int,b:int`), storage, and
@@ -193,6 +205,10 @@ against it.
see V4 for the broader direction.)*
- [ ] **V2** SQL query results render as a dynamic table view in
the output pane, with multiple result tabs supported.
*(Progress: a basic aligned-column data view is rendered for
`show data` and after every write (ADR-0014). Pretty
box-drawing tables with truncation/scroll handling, plus
multi-tab support, remain in V4 territory.)*
- [~] **V3** Full ER-diagram export (whole-database graph, viewed
outside the TUI) — low priority; design and ADR pending.
- [~] **V4** Output panel as a *scrollable per-session log* with
@@ -210,7 +226,7 @@ against it.
styling, Markdown export, scroll indicator — remains pending.)*
- [ ] **V5** `show <kind> [<name>]` family of commands for
redisplaying schema info on demand. *(Progress: `show table
<name>` implemented and reuses the structure-render pipeline;
<name>` and `show data <Table>` implemented;
`show tables`, `show relationships`, etc. pending.)*
## Project lifecycle (per ADR-0004)
@@ -281,6 +297,19 @@ against it.
- [ ] **H1** Friendly error-rewriting layer translates SQLite
error messages into learner-friendly equivalents.
*(Progress: foreign-key constraint failures are enriched
with both inbound and outbound relationship listings (so
RESTRICT errors point at the children that still reference
this table); full SQL → English translation pending.)*
- [ ] **H1a** Strong syntax-help in parse errors. When the user
types something near-correct (e.g. `insert into T ('Oli')` —
forgotten `values`; or `update T set x=1` — missing WHERE),
the error should *name the missing keyword or clause* rather
than just point at the unexpected character. This is a
separate effort from H1 (which targets database errors); it
targets parser errors. Pending — multiple targeted fixes
shipping piecemeal so far (e.g. `values` becoming optional in
INSERT removes one such case).
- [ ] **H2** `hint` provides contextual help for the current
input or the most recent error.
- [ ] **H3** `help` provides general reference and per-command