feat: give column data types a dedicated syntax-highlight colour

Both Node::Ident and Word carried a highlight_override field, and
both were dead — the walker driver discarded the Ident's and
walk_word hardcoded Keyword. So column types (int, serial, …)
rendered identically to table/column names.

Wire both overrides through, and add a dedicated HighlightClass::Type
with its own theme colour (tok_type), distinct from keyword-purple
and identifier-teal. The three type Ident slots opt in, so canonical
types and the advanced-mode single-word SQL aliases (float, varchar,
…) render as types; the two-word `double precision` alias opts in via
a new Word::type_keyword constructor. ADR-0022 Amendment 4.
This commit is contained in:
claude@clouddev1
2026-05-29 22:07:18 +00:00
parent 46a31284c5
commit d20f765325
10 changed files with 195 additions and 13 deletions
@@ -587,6 +587,61 @@ guard above, and three re-baselined `typing_surface` snapshots
(`form_a_in_progress_one_value`, `form_b_too_few_values`,
`form_c_wrong_count`).
## Amendment 4 — Column types get a dedicated highlight class (2026-05-29)
§3 introduced **seven** token-class colour fields and the matching
`HighlightClass` enum. Column data-type keywords (`int`, `serial`,
`text`, …) had no class of their own: the walker driver hardcoded
`HighlightClass::Identifier` for every matched `Node::Ident`, so a
type rendered identical to a table/column name (clause keywords were
already distinct, in `tok_keyword`). A user-reported bug (issue #8)
noted that in `create table Orders (count int, id serial PRIMARY KEY)`
the identifiers and the type keywords were indistinguishably teal.
Both terminal kinds already carried a `highlight_override:
Option<HighlightClass>` field — `Node::Ident` and the `Word` struct
alike — but **both were dead**: the driver destructured the Ident's to
`_`, and `walk_word` hardcoded `Keyword`, neither ever consulting the
field. This amendment wires both through.
**Change:**
1. **New class.** `HighlightClass::Type` (the eighth variant) and a
matching eighth `Theme` field `tok_type`, populated in both
`dark()` and `light()` with a tone deliberately distinct from both
`tok_keyword` and `tok_identifier` (a pink / deep-magenta in the
red-purple range — the only free slot in the existing palette).
`highlight_class_color` maps the new variant.
2. **Overrides are now live.** `walk_ident` emits
`override.unwrap_or(Identifier)` and `walk_word` emits
`override.unwrap_or(Keyword)` for the matched byte range. Every
slot that leaves the field `None` is unchanged.
3. **Type slots opt in.** The three `IdentSource::Types` slots —
`shared::TYPE_SLOT` (`add column` / `change column`), the inline
create-table column-type Ident (`ddl.rs`), and `sql_create_table::
SQL_TYPE_NAME` (advanced) — set `highlight_override: Some(Type)`.
In advanced mode, every single-word SQL type-name *alias*
(`float`, `varchar`, `integer`, … — ADR-0035 §3) flows through
`SQL_TYPE_NAME` and so is type-coloured for free.
4. **`double precision` too.** The lone two-word alias (ADR-0035 §3)
is matched as keyword tokens, not an `IdentSource::Types` Ident, so
it cannot ride rule 3. Its `double` / `precision` grammar nodes use
the new `Word::type_keyword` constructor (`highlight_override:
Some(Type)`) so the spelling renders type-coloured like its
single-word synonyms `float` / `real`.
**Pedagogy:** a dedicated colour (over the lighter option of reusing
`tok_keyword`) lets a learner tell *"this is a type"* from a clause
keyword and from a name they invented — three distinct roles, three
distinct colours.
**Coverage:** `dsl_type_keyword_classified_as_type`,
`advanced_type_keywords_classified_as_type` (the ticket's exact
example), `advanced_double_precision_classified_as_type`,
`type_colour_is_distinct_from_keyword_and_identifier`, and the
extended theme mapping/contrast tests. Text snapshots are colour-blind
(`render_to_string` strips style), so none churned.
## Out of scope
Deliberately deferred to keep this ADR shippable as a single
+1 -1
View File
@@ -27,7 +27,7 @@ This directory contains the project's ADRs, recorded per
- [ADR-0019 — Friendly error layer (H1) and i18n message catalog](0019-friendly-error-layer-and-i18n.md)
- [ADR-0020 — Tokenization layer for the DSL parser](0020-tokenization-layer-for-the-dsl-parser.md)
- [ADR-0021 — Parser-as-source-of-truth for H1a (per-command usage in parse errors)](0021-parser-as-source-of-truth-for-h1a.md)
- [ADR-0022 — Ambient typing assistance: colour, hint panel, completion (I3 + I4)](0022-ambient-typing-assistance.md) — **Amendment 1 supersedes §12's simple-mode-only carve-out**: the unified mode-aware walker (ADR-0030/0031/0032) now speaks SQL, so advanced-mode ambient assistance is re-enabled. `ambient_hint_in_mode` + `hint_resolution_at_input_in_mode` + `expected_for_hint_snapshot` thread `Mode`; `render_hint_panel` calls ambient for all modes (no more advanced-mode `None`); the one-shot `:` sigil is stripped before the ambient walk. Fixes a live bug where advanced-mode SQL hinting/completion-preview were dead despite Phase 2 marking them green (validated at the engine layer, not the UI). Simple-mode gating, highlighting, and the §13 performance posture are unchanged; covered by an app-level render test plus ambient-layer regression locks; **Amendment 2 reverses the handoff-14 keywords-first candidate ordering** — schema identifiers (table/column/relationship names) now sort *before* keywords so a name the user would have to look up stays visible in the single-row, window-scrolled candidate line (keywords are learned over time; the `tok_identifier`/`tok_keyword` colour split marks the boundary); shipped with a `walk_repeated` fix that surfaces a list item's trailing optionals at a clean boundary (`order by Name ``asc`/`desc`, `select Name ``as`, `create table … Code(text) ``not`/`unique`/`default`/`check`; the `,` separator deliberately not surfaced); records a deferred two-line hint box for growing lists; **Amendment 3 makes the ambient-hint fallback rung schema-aware** — Amendment 1's bottom-rung `parse_command_in_mode` was schemaless while every earlier rung was not, so between-values insert hints pointed at `)` (type-blind close) instead of `,` and wrong-arity closed tuples read "submit with Enter" for an input the schema-aware parse rejects (issue #2); now uses `parse_command_with_schema_in_mode`, no extra walk, with the friendly arity diagnostic still winning at its higher rung
- [ADR-0022 — Ambient typing assistance: colour, hint panel, completion (I3 + I4)](0022-ambient-typing-assistance.md) — **Amendment 1 supersedes §12's simple-mode-only carve-out**: the unified mode-aware walker (ADR-0030/0031/0032) now speaks SQL, so advanced-mode ambient assistance is re-enabled. `ambient_hint_in_mode` + `hint_resolution_at_input_in_mode` + `expected_for_hint_snapshot` thread `Mode`; `render_hint_panel` calls ambient for all modes (no more advanced-mode `None`); the one-shot `:` sigil is stripped before the ambient walk. Fixes a live bug where advanced-mode SQL hinting/completion-preview were dead despite Phase 2 marking them green (validated at the engine layer, not the UI). Simple-mode gating, highlighting, and the §13 performance posture are unchanged; covered by an app-level render test plus ambient-layer regression locks; **Amendment 2 reverses the handoff-14 keywords-first candidate ordering** — schema identifiers (table/column/relationship names) now sort *before* keywords so a name the user would have to look up stays visible in the single-row, window-scrolled candidate line (keywords are learned over time; the `tok_identifier`/`tok_keyword` colour split marks the boundary); shipped with a `walk_repeated` fix that surfaces a list item's trailing optionals at a clean boundary (`order by Name ``asc`/`desc`, `select Name ``as`, `create table … Code(text) ``not`/`unique`/`default`/`check`; the `,` separator deliberately not surfaced); records a deferred two-line hint box for growing lists; **Amendment 3 makes the ambient-hint fallback rung schema-aware** — Amendment 1's bottom-rung `parse_command_in_mode` was schemaless while every earlier rung was not, so between-values insert hints pointed at `)` (type-blind close) instead of `,` and wrong-arity closed tuples read "submit with Enter" for an input the schema-aware parse rejects (issue #2); now uses `parse_command_with_schema_in_mode`, no extra walk, with the friendly arity diagnostic still winning at its higher rung; **Amendment 4 gives column types a dedicated highlight class** — both `Node::Ident.highlight_override` *and* the `Word.highlight_override` field were dead (driver destructured the former to `_`, `walk_word` hardcoded `Keyword`); now both wired through, with a new `HighlightClass::Type` + eighth `Theme` field `tok_type` (a pink/deep-magenta distinct from both keyword purple and identifier teal) so types no longer render identically to identifiers (issue #8); the three `IdentSource::Types` slots opt in via `Some(Type)` (advanced-mode single-word SQL aliases — `float`, `varchar`, … per ADR-0035 §3 — ride along for free), and the two-word `double precision` alias opts in via the new `Word::type_keyword` constructor so it matches its synonyms
- [ADR-0023 — Unified declarative grammar tree](0023-unified-grammar-tree.md) — direction (superseded for execution detail by ADR-0024)
- [ADR-0024 — Unified grammar tree: execution plan](0024-unified-grammar-tree-execution-plan.md) — **Accepted**, the executable spec — implemented (Phases AF; Phase F shipped "minimal", `parser.rs` retained as the router — see the ADR's Phase F implementation note)
- [ADR-0025 — Indexes](0025-indexes.md) — **Accepted** (**Amendment 1, 2026-05-25**: UNIQUE indexes admitted on the **advanced-mode** surface via `CREATE UNIQUE INDEX` — ADR-0035 §4d; the `IndexSchema.unique` flag round-trips through `project.yaml` with no new metadata table since the engine reports uniqueness natively; simple-mode `add unique index` stays deferred), `add index` / `drop index`, persistence, rebuild-table preservation, and items-list display (`C3` index portion + `S2`)
+4
View File
@@ -106,6 +106,10 @@ handoff-14 cleanup; 449 after B2/C2.)
- [ ] **I3** Tab completion for app commands, DSL keywords, table
names, column names, and SQL keywords.
- [ ] **I4** Syntax highlighting for both the DSL and SQL.
*(Refinement 2026-05-29, issue #8: column data types now carry a
dedicated `HighlightClass::Type` / `tok_type` colour, distinct from
identifiers and clause keywords — ADR-0022 Amendment 4. The broad
highlighting goal stays open.)*
- [ ] **I5** In-flight query/command cancellation (Ctrl-C in the
output area or input field).
+2 -2
View File
@@ -18,7 +18,7 @@ use crate::dsl::command::{
};
use crate::dsl::value::Value;
use crate::dsl::grammar::{
CommandNode, HintMode, IdentSource, Node, ValidationError, Word,
CommandNode, HighlightClass, HintMode, IdentSource, Node, ValidationError, Word,
shared::{REFERENTIAL_CLAUSES, TYPE_SLOT, TYPE_VALIDATOR},
};
@@ -1079,7 +1079,7 @@ const COL_SPEC_NODES: &[Node] = &[
source: IdentSource::Types,
role: "col_type",
validator: Some(TYPE_VALIDATOR),
highlight_override: None,
highlight_override: Some(HighlightClass::Type),
writes_table: false,
writes_column: false,
writes_user_listed_column: false,
+21
View File
@@ -47,6 +47,12 @@ use crate::dsl::walker::outcome::MatchedPath;
pub enum HighlightClass {
Keyword,
Identifier,
/// Column data-type keyword (`int`, `serial`, `text`, …).
/// Distinct from `Keyword` and `Identifier` so learners can
/// tell "this is a type" from a clause keyword or a name they
/// invented (ADR-0022 Amendment 4). Assigned via a type slot's
/// `highlight_override`, not by byte shape.
Type,
Number,
String,
Punct,
@@ -191,6 +197,21 @@ impl Word {
}
}
/// A keyword that highlights as a column **type** rather than a
/// clause keyword (ADR-0022 Amendment 4). The one user today is
/// the two-word `double precision` SQL alias (ADR-0035 §3): it
/// is matched as keyword tokens, not an `IdentSource::Types`
/// `Ident`, so without this it would render keyword-coloured
/// while its single-word synonyms (`float`, `real`) render as
/// types.
pub const fn type_keyword(primary: &'static str) -> Self {
Self {
primary,
aliases: &[],
highlight_override: Some(HighlightClass::Type),
}
}
/// Case-insensitive match against the primary or any alias.
pub fn matches(&self, candidate: &str) -> bool {
if candidate.eq_ignore_ascii_case(self.primary) {
+3 -3
View File
@@ -7,8 +7,8 @@
use crate::completion::TableColumn;
use crate::dsl::grammar::{
HintMode, IdentSource, IdentValidator, Node, NumberValidator,
ValidationError, Word,
HighlightClass, HintMode, IdentSource, IdentValidator, Node,
NumberValidator, ValidationError, Word,
};
use crate::dsl::types::Type;
use crate::dsl::walker::context::WalkContext;
@@ -50,7 +50,7 @@ pub const TYPE_SLOT: Node = Node::Ident {
source: IdentSource::Types,
role: "type",
validator: Some(TYPE_VALIDATOR),
highlight_override: None,
highlight_override: Some(HighlightClass::Type),
writes_table: false,
writes_column: false,
writes_user_listed_column: false,
+6 -4
View File
@@ -24,7 +24,9 @@
//! `sql_insert::SQL_INSERT_SHAPE`, which starts at `INTO`).
use crate::dsl::grammar::sql_select::reject_internal_table;
use crate::dsl::grammar::{IdentSource, Node, ValidationError, Word, shared, sql_expr};
use crate::dsl::grammar::{
HighlightClass, IdentSource, Node, ValidationError, Word, shared, sql_expr,
};
use crate::dsl::types::Type;
static COMMA: Node = Node::Punct(',');
@@ -60,7 +62,7 @@ const SQL_TYPE_NAME: Node = Node::Ident {
source: IdentSource::Types,
role: "col_type",
validator: Some(validate_sql_type_name),
highlight_override: None,
highlight_override: Some(HighlightClass::Type),
writes_table: false,
writes_column: false,
writes_user_listed_column: false,
@@ -86,8 +88,8 @@ const LENGTH_OPT: Node = Node::Optional(&Node::Seq(LENGTH_NODES));
// on its own (ADR-0035 §6.3). The builder maps the pair to
// `Type::Real`.
static DOUBLE_PRECISION_NODES: &[Node] = &[
Node::Word(Word::keyword("double")),
Node::Word(Word::keyword("precision")),
Node::Word(Word::type_keyword("double")),
Node::Word(Word::type_keyword("precision")),
];
static TYPE_WITH_LENGTH_NODES: &[Node] = &[SQL_TYPE_NAME, LENGTH_OPT];
static SQL_TYPE_CHOICES: &[Node] = &[
+11 -3
View File
@@ -180,7 +180,7 @@ fn walk_node_inner(
source: src,
role,
validator,
highlight_override: _,
highlight_override,
writes_table,
writes_column,
writes_user_listed_column,
@@ -193,6 +193,7 @@ fn walk_node_inner(
*src,
role,
*validator,
*highlight_override,
*writes_table,
*writes_column,
*writes_user_listed_column,
@@ -336,7 +337,10 @@ fn walk_word(
per_byte.push(ByteClass {
start,
end,
class: HighlightClass::Keyword,
// A keyword may opt into a non-default colour via
// `Word::type_keyword` (e.g. `double precision`, ADR-0022
// Amendment 4). Plain keywords leave it `None`.
class: word.highlight_override.unwrap_or(HighlightClass::Keyword),
});
NodeWalkResult::Matched { end, skipped: Vec::new() }
} else {
@@ -397,6 +401,7 @@ fn walk_ident(
src: crate::dsl::grammar::IdentSource,
role: &'static str,
validator: Option<crate::dsl::grammar::IdentValidator>,
highlight_override: Option<crate::dsl::grammar::HighlightClass>,
writes_table: bool,
writes_column: bool,
writes_user_listed_column: bool,
@@ -554,7 +559,10 @@ fn walk_ident(
per_byte.push(ByteClass {
start,
end,
class: HighlightClass::Identifier,
// A type slot (and any future slot that wants a non-default
// colour) overrides the otherwise-uniform Identifier class
// (issue #8 / ADR-0022 Amendment 4).
class: highlight_override.unwrap_or(HighlightClass::Identifier),
});
NodeWalkResult::Matched { end, skipped: Vec::new() }
}
+70
View File
@@ -297,6 +297,76 @@ mod tests {
assert_eq!(runs[1].2, HighlightClass::Keyword);
}
#[test]
fn dsl_type_keyword_classified_as_type() {
// Issue #8 / ADR-0022 Amendment 4: the column type in a DSL
// `create table` (`serial`) is a Type, distinct from the
// identifiers `T` / `id` which stay Identifier.
let runs = run("create table T with pk id(serial)");
// `serial` occupies bytes 26..32 in the input.
let serial = runs
.iter()
.find(|(s, e, _)| &"create table T with pk id(serial)"[*s..*e] == "serial")
.expect("serial run present");
assert_eq!(serial.2, HighlightClass::Type);
// The invented identifiers are NOT typed.
for name in ["T", "id"] {
let r = runs
.iter()
.find(|(s, e, _)| &"create table T with pk id(serial)"[*s..*e] == name)
.unwrap_or_else(|| panic!("{name} run present"));
assert_eq!(r.2, HighlightClass::Identifier, "{name} stays Identifier");
}
}
#[test]
fn advanced_type_keywords_classified_as_type() {
// Issue #8: the exact example from the ticket. In advanced
// mode the SQL type keywords `int` / `serial` are Type;
// `Orders` / `count` / `id` stay Identifier.
let input = "create table Orders (count int, id serial)";
let runs: Vec<(usize, usize, HighlightClass)> =
highlight_runs_in_mode(input, crate::mode::Mode::Advanced)
.into_iter()
.map(|c| (c.start, c.end, c.class))
.collect();
let class_of = |needle: &str| {
runs.iter()
.find(|(s, e, _)| &input[*s..*e] == needle)
.unwrap_or_else(|| panic!("{needle} run present"))
.2
};
assert_eq!(class_of("int"), HighlightClass::Type);
assert_eq!(class_of("serial"), HighlightClass::Type);
assert_eq!(class_of("Orders"), HighlightClass::Identifier);
assert_eq!(class_of("count"), HighlightClass::Identifier);
assert_eq!(class_of("id"), HighlightClass::Identifier);
}
#[test]
fn advanced_double_precision_classified_as_type() {
// Issue #8 follow-up: the two-word `double precision` SQL
// alias (ADR-0035 §3) is matched as keyword tokens, but
// `Word::type_keyword` colours it as a type so it matches
// its single-word synonyms `float` / `real`. Both words
// carry the Type class.
let input = "create table T (x double precision)";
let runs: Vec<(usize, usize, HighlightClass)> =
highlight_runs_in_mode(input, crate::mode::Mode::Advanced)
.into_iter()
.map(|c| (c.start, c.end, c.class))
.collect();
let class_of = |needle: &str| {
runs.iter()
.find(|(s, e, _)| &input[*s..*e] == needle)
.unwrap_or_else(|| panic!("{needle} run present"))
.2
};
assert_eq!(class_of("double"), HighlightClass::Type);
assert_eq!(class_of("precision"), HighlightClass::Type);
assert_eq!(class_of("x"), HighlightClass::Identifier);
}
#[test]
fn full_command_walks_with_each_class() {
// `update T set Name='hi' --all-rows` — walker covers it
+22
View File
@@ -51,6 +51,11 @@ pub struct Theme {
// ---- Per-token-class colours (ADR-0022 §3) -------------------
pub tok_keyword: Color,
pub tok_identifier: Color,
/// Column data-type keyword colour (ADR-0022 Amendment 4) —
/// a dedicated tone distinct from both `tok_keyword` and
/// `tok_identifier` so a learner can tell a type from a
/// clause keyword or an invented name at a glance.
pub tok_type: Color,
pub tok_number: Color,
pub tok_string: Color,
pub tok_punct: Color,
@@ -84,6 +89,7 @@ impl Theme {
// distinct from the mode-banner blue.
tok_keyword: Color::Rgb(0xC7, 0x92, 0xEA), // muted purple
tok_identifier: Color::Rgb(0x56, 0xB6, 0xC2), // cyan-teal — identifiers are the user's content, deserve a vivid distinct colour
tok_type: Color::Rgb(0xF0, 0x8F, 0xC0), // pink — types sit in the red-purple range, clearly apart from the lavender keyword and teal identifier
tok_number: Color::Rgb(0xF7, 0x8C, 0x6C), // warm orange
tok_string: Color::Rgb(0xC3, 0xE8, 0x8D), // soft green
tok_punct: Color::Rgb(0x8B, 0x90, 0x9A), // == muted
@@ -113,6 +119,7 @@ impl Theme {
// literals + flags; cool accent for keyword.
tok_keyword: Color::Rgb(0x6F, 0x42, 0xC1), // royal purple
tok_identifier: Color::Rgb(0x0F, 0x6B, 0x76), // deep teal — same role as dark variant: identifiers stand out
tok_type: Color::Rgb(0xA8, 0x2D, 0x73), // deep magenta — red-purple, distinct from royal-purple keyword + teal identifier
tok_number: Color::Rgb(0xBC, 0x4F, 0x1F), // burnt orange
tok_string: Color::Rgb(0x22, 0x86, 0x3A), // forest green
tok_punct: Color::Rgb(0x60, 0x66, 0x73), // == muted
@@ -132,6 +139,7 @@ impl Theme {
match class {
HighlightClass::Keyword => self.tok_keyword,
HighlightClass::Identifier => self.tok_identifier,
HighlightClass::Type => self.tok_type,
HighlightClass::Number => self.tok_number,
HighlightClass::String => self.tok_string,
HighlightClass::Punct => self.tok_punct,
@@ -156,6 +164,7 @@ mod tests {
let t = Theme::dark();
for (name, c) in [
("tok_keyword", t.tok_keyword),
("tok_type", t.tok_type),
("tok_number", t.tok_number),
("tok_string", t.tok_string),
("tok_flag", t.tok_flag),
@@ -174,6 +183,7 @@ mod tests {
let t = Theme::light();
for (name, c) in [
("tok_keyword", t.tok_keyword),
("tok_type", t.tok_type),
("tok_number", t.tok_number),
("tok_string", t.tok_string),
("tok_flag", t.tok_flag),
@@ -192,10 +202,22 @@ mod tests {
let t = Theme::dark();
assert_eq!(t.highlight_class_color(HighlightClass::Keyword), t.tok_keyword);
assert_eq!(t.highlight_class_color(HighlightClass::Identifier), t.tok_identifier);
assert_eq!(t.highlight_class_color(HighlightClass::Type), t.tok_type);
assert_eq!(t.highlight_class_color(HighlightClass::Number), t.tok_number);
assert_eq!(t.highlight_class_color(HighlightClass::String), t.tok_string);
assert_eq!(t.highlight_class_color(HighlightClass::Punct), t.tok_punct);
assert_eq!(t.highlight_class_color(HighlightClass::Flag), t.tok_flag);
assert_eq!(t.highlight_class_color(HighlightClass::Error), t.tok_error);
}
#[test]
fn type_colour_is_distinct_from_keyword_and_identifier() {
// ADR-0022 Amendment 4 / issue #8: the whole point of a
// dedicated type class is that types do NOT share a colour
// with clause keywords or invented identifiers.
for t in [Theme::dark(), Theme::light()] {
assert_ne!(t.tok_type, t.tok_keyword);
assert_ne!(t.tok_type, t.tok_identifier);
}
}
}