Replaces the previous FOR-only, single-statement-body, static-unrolling
design with genuine control flow: FOR var IN source DO <block> END loop
bodies with any number of statements (including nested loops/IFs),
IF cond THEN <block> [ELSE <block>] END branching (nestable in any
block), and RETURN as an ordinary statement usable anywhere -- it
aborts execution immediately with its value, including breaking out of
a loop early.
This required a real, if small, architecture change: an IF branch or an
early RETURN depends on actual cell values, which the previous "unroll
everything into one static formula at parse time" approach can't
resolve (loop *counts* are knowable up front from a range's size;
branch *outcomes* aren't). A kind: 'block' call is now inlined only as
far as resolving nested function calls (expandBlock, parse-time, same
inside-out principle as before) and wrapped in a new BlockCallNode;
its own control flow is interpreted at evaluation time by a small
tree-walking interpreter added to evaluator.ts (execBlock/execStatement,
a thrown ReturnSignal for early exit), reusing userFunctions.ts's
substitute() (now exported) to resolve parameter/local-variable names
against the current environment. Dependency tracking stays fully
correct and conservative: collectRefs() walks into a BlockCallNode's
block (new collectRefsInBlock), so a cell referenced only in a branch
that doesn't happen to run is still a tracked precedent -- same
principle as an ordinary IF() in any spreadsheet.
Also added: comparison operators (= <> > < >= <=, result 1/0) as a new,
loosest-binding precedence level in the *general* expression grammar
(not just inside a block's IF) -- needed for IF conditions to be useful
at all. A definition-time static check (blockAlwaysReturns) rejects a
function at save time when it's not obviously guaranteed to reach
RETURN on every path (an IF without an ELSE, or one where only one
branch returns); the interpreter's own runtime "ended without RETURN"
-> #ERR remains the actual enforcement for cases the static check can't
see (e.g. a loop over an empty range).
Nested loops (FOR inside FOR) and nested IFs work as an emergent
property of the recursive statement grammar -- verified with dedicated
tests, not just assumed.
257 tests passing (30 net new/rewritten for this change), vue-tsc/
eslint/build clean. Verified live in the running app: a FIRSTOVER(range;
threshold) function using a FOR loop with a nested IF and an early
RETURN, confirmed it returns the first match, confirmed changing a
cell *past* the match point doesn't affect the result while changing
the match point itself does, and confirmed a function without a
guaranteed RETURN is rejected inline at save time.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
The engine already anticipated $A1/A$1/$A$1 in its grammar comment, but the
lexer had no rule for '$' at all -- any formula using it failed to parse.
- lexer.ts: '$' is now a valid word character, so "$A$1" tokenizes as one
IDENT, same as any other bare word (the lexer still doesn't interpret it,
per its "dumb lexer" design -- the parser/evaluator stay untouched, since
evaluation is purely positional and $ has no effect on a value).
- refs.ts: parseA1()/toA1() now carry colAbs/rowAbs through the $ markers
(backward compatible -- existing {col,row} destructuring is unaffected).
Added cycleReferenceLock() for Excel/Numbers-style F4: relative ->
absolute -> row-locked -> col-locked -> relative, toggling whichever
reference sits at the caret (and the correct side of an A1:B2 range).
- document.ts: shiftFormula() now honors the lock on copy (a locked axis
stays exactly as typed) but not on cut (a moved cell's reference follows
it regardless of $, since cut relocates the cell itself rather than
filling a relative offset).
- TableGrid.vue's inline cell editor and FormulaBar.vue both wire up F4 via
the shared cycleReferenceLock() helper.
- Tests: lexer/parser coverage for $ tokenizing/parsing, a new refs.spec.ts
for parseA1/toA1 round-tripping and the F4 cycle (incl. range side-
picking), and document.spec.ts cases for locked-axis copy vs. cut.
- README.md / src/engine/README.md: documented the $ semantics, F4, and
corrected a couple of now-stale claims (test count, "$ is ignored, there's
no copy/fill", "active cell is always the last-clicked one").
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
wkhandClaude Sonnet 5