Struct TsStmt 

pub struct TsStmt {
    pub span: Option<Span>,
    pub nested_map: Option<SourceMapBuilder>,
    pub nested_map_source_id: usize,
    pub no_blank_before: bool,
    /* private fields */
}

One statement — a Verbatim-tagged escape hatch (still constructible only via TsStmt::verbatim, per #1307’s Decision D — the verbatim_sites probe needs exactly one string to line-scan for), or, from this slice, a real structured kind. The real kinds have no such sealing: they’re normal typed constructors, not a “wrap opaque text” escape hatch, so the verbatim_sites concern that motivates verbatim’s own single-constructor discipline doesn’t apply to them.

Fields

span: Option<Span>

Where this statement’s content originated in the .bynk source, if known. Only a top-level statement’s own span is currently recorded as a source-map checkpoint (crate::printer::print, unchanged from P7.5/R7.4’s own scope) — a nested statement (inside a Block, If, ForOf, TryCatch) still carries this field structurally, for whichever future slice gives sub-statement source maps real value, but the printer does not yet record a checkpoint from it. Named here explicitly (P7.8’s own accepted proposal: “an implementation-time call within this same shape”) rather than left ambiguous.

nested_map: Option<SourceMapBuilder>

#1477’s own real gap: a body-bearing statement — in practice always a TsStmtKind::Raw/TsStmtKind::Verbatim opaque blob standing in for a lowered function/method body (ADR 0391’s own permanent exclusion) — carries its own per-statement source-map checkpoints, collected by the caller’s own body-local SourceMapBuilder before this node existed. Before this field, every real bynk-emit caller that needed to merge those checkpoints into its own module map had to reverse-engineer this node’s own print-time byte offset from the outside — bynk-emit’s own emit_class_method_and_merge_source_map (emitter/emit.rs) recovers it by subtracting known lengths and string-matching the printed text’s own tail, degrading to a silent skip if that search fails; still live for emit_service/emit_agent’s own not-yet-converted call sites (#1481/#1482). emit_free_fn used to recover it by separate, independent exact arithmetic, guarded by a debug_assert! — #1480 converted it to set this field directly instead, the first real bynk-emit caller to do so. Both existed only because nothing reported this node’s own real print-time offset directly. Setting this field lets the printer itself do the merge, at the exact offset it is about to write this node’s text to — no reverse-engineering, no silent-skip fallback (see this crate’s own private printer::render_block_stmts for the handling). None for every real site that predates this field and every node whose own text carries no nested checkpoints of its own — the overwhelmingly common case, and the reason this is an Option, not a required field.

nested_map_source_id: usize

#1486’s own second real gap, found converting emit_test_module/ emit_integration_module: crate::printer::print’s own top-level loop always merges a statement’s nested_map against parent source id 0 (crate::printer::MergeTarget’s own review-of-#1488 fix threads a source_id through every manual merge entry point, but print’s own automatic top-level case never had a per-statement way to say “not 0”) — correct for every single-source module (emit_project, every other real caller so far), but wrong for a test/integration module’s own aggregate map, which registers one source per fragment file a case’s body came from (SourceMapBuilder::add_source) and needs each such case’s own top-level statement merged against that source’s id, not the module’s primary one. 0 (every site predating this field, and every single-source module after it) keeps the automatic policy exactly as it already was; a case/property/attack wrapper statement sets this to its own registered source id before print sees it.

no_blank_before: bool

#1486’s own real gap, found converting emit_project/emit_test_ module/emit_integration_module off a single opaque top-level Verbatim wrap: crate::printer::print’s own top-level loop inserts exactly one blank line between every pair of adjacent top-level statements (its own “readability policy” doc, right below this field’s own use site) — correct for content that was always designed against that policy (emit_composition_root/emit_test_ main/workers.rs/workers_entry.rs, every one of them zero-diff from day one), but emit_agent’s own already-shipped Vec<TsStmt> (#1482, predating this field) was built against the previous regime — where every real caller printed via bynk-emit’s own extend_printed/extend_printed_and_merged (a flat per-stmt loop with no automatic spacing of its own) — and genuinely needs a same-registry-then-zero-factory pairing with no blank line between them, matching the pre-conversion hand-written text this crate’s own zero-diff discipline requires preserving exactly. true on a statement suppresses the automatic blank line crate::printer::print would otherwise insert immediately before it — the narrow, general escape hatch this real, already-shipped content needs, not a special case hard-coded into the printer for one specific pair of TsDecl variants. false (every site predating this field, and the overwhelming majority after it) keeps the automatic policy exactly as it already is.

Implementations

impl TsStmt

pub fn verbatim( origin: VerbatimOrigin, text: impl Into<String>, span: Option<Span>, ) -> Self

The one constructor for a Verbatim-kinded statement.

pub fn decl(decl: TsDecl, span: Option<Span>) -> Self

pub fn const_stmt( name: TsBindingName, ty: Option<TsType>, init: TsExpr, span: Option<Span>, ) -> Self

pub fn let_stmt( name: TsBindingName, ty: Option<TsType>, init: Option<TsExpr>, span: Option<Span>, ) -> Self

pub fn expr_stmt(expr: TsExpr, span: Option<Span>) -> Self

pub fn return_stmt(expr: Option<TsExpr>, span: Option<Span>) -> Self

pub fn throw_stmt(expr: TsExpr, span: Option<Span>) -> Self

pub fn if_stmt(cond: TsExpr, then_branch: TsStmt, span: Option<Span>) -> Self

pub fn if_else_stmt( cond: TsExpr, then_branch: TsStmt, else_branch: TsStmt, span: Option<Span>, ) -> Self

pub fn if_else_same_line_stmt( cond: TsExpr, then_branch: TsStmt, else_branch: TsStmt, span: Option<Span>, ) -> Self

TsStmt::if_else_stmt’s own sibling with } else { on one line — #1325’s own real gap, emit_test_main’s own real else spacing. See TsStmtKind::If’s own doc for why this needs to be a distinct constructor rather than a change to the existing default.

pub fn for_of( binding: impl Into<String>, iter: TsExpr, body: TsStmt, span: Option<Span>, ) -> Self

pub fn for_stmt( name: impl Into<String>, init: TsExpr, test: TsExpr, body: TsStmt, span: Option<Span>, ) -> Self

for (let <name> = <init>; <test>; <update>++) <body> — see TsStmtKind::For’s own doc for exactly what this construct does and does not represent.

pub fn try_catch( try_block: TsStmt, catch_param: Option<impl Into<String>>, catch_block: TsStmt, span: Option<Span>, ) -> Self

pub fn block(stmts: Vec<TsStmt>, span: Option<Span>) -> Self

pub fn continue_stmt(span: Option<Span>) -> Self

pub fn assign(target: TsExpr, value: TsExpr, span: Option<Span>) -> Self

pub fn comment(text: impl Into<String>, span: Option<Span>) -> Self

pub fn doc_comment(text: impl Into<String>, span: Option<Span>) -> Self

pub fn blank(span: Option<Span>) -> Self

pub fn switch_stmt( discriminant: TsExpr, cases: Vec<TsSwitchCase>, span: Option<Span>, ) -> Self

pub fn inline_block(stmts: Vec<TsStmt>, span: Option<Span>) -> Self

pub fn increment(expr: TsExpr, span: Option<Span>) -> Self

pub fn raw(text: impl Into<String>, span: Option<Span>) -> Self

The one constructor for a Raw-kinded statement — text is printed verbatim, exactly as given (see TsStmtKind::Raw’s own doc for why this is a distinct kind from Verbatim, not a reuse of it).

Trait Implementations

impl Clone for TsStmt

fn clone(&self) -> TsStmt

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for TsStmt

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.