The package reader reports a part over the limit as "expands to more
than <limit> bytes"; the test added in #2491 expected the declared size
and failed on main.
Data Subtotal grouped each group's detail rows at level 1 and left the
subtotal rows at level 0, so outline level 1 still showed every
subtotal. Excel adds an outer group over all data and subtotal rows,
summarised by the grand total: details at level 2, subtotals at
level 1, the grand total at level 0. Subtotal now adds that group after
writing the grand total.
Two tests that read the old levels now expect Excel's: the summary-row
test shows level 2 to keep the subtotal visible, and the sheet-prefix
test expects a detail row at level 2.
Measured with crates/engine/examples/bench.rs (perf/xtask-bench), 100k rows, M2 Pro:
enter.formulas 465 -> 420 ms (23829d8c: 375), xlsx.save 540 -> 488 ms (466).
- number_to_text: the round-up-to-1E+15 check (f29ac7d9e) formatted every number with
exact {:.14e} and parsed it back; formula printing calls it once per literal. Run it
only for magnitudes that can round up (>= 9.9999999999999e14). Same output; boundary
tests added.
- Expr clone/map, xlsx prefix_params: a binary node whose left operand is not a chain
no longer allocates a spine Vec (2aeb21d4d); map rewrites the operands in their boxes.
Expr::walk reserves room for a typical formula up front.
- grow_autofilter checks the sheet has an AutoFilter before looking up the cell.
- retained_bytes_against looks up the newer band's row once per row, not per cell.
- The journal measures an entry's JSON size with a counting writer instead of building
the text (a big paste's params are MBs).
The rest of the enter.formulas gap and all of edit.fanout's (~+12%) is Undo's byte
budget (72db972de, #1227): sizing each snapshot and freeing evicted ones, an intended
cost of bounded memory.
Co-authored-by: nihilist-narwhal <240798405+nihilist-narwhal@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
read_cache turned every pivot cache error into absence with
optional_xml(part).ok().flatten(), so a cache over the part size limit
dropped its PivotTable with no warning, and a malformed or missing cache
only left a generic "skipped part" line. Each case now adds a warning that
the PivotTable was kept as plain cells, and the workbook keeps opening
(the oversized part is refused before it inflates).
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Found while testing crates/engine/src/cmd/data.rs:992 (split_delimited, used by text_to_columns)
Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>
* calc: linear dependency extraction for range-operator chains (#2183)
precedents() handled a range operator with non-reference operands by recursing into both operands and then letting Expr::walk descend into them again, so a left-associative chain of k operands cost 2^k (a 60-operand INDEX(...):INDEX(...) chain would hang on open). It is now a hand-written collector that visits every node once, growing the per-sheet bounding boxes of the enclosing range operator as it goes. Past a depth (4096) or node (1,000,000) bound it marks the formula dynamic (always recalculated) instead of recursing further.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* calc: drop wall-clock timing from precedents tests (clippy)
std::time::Instant::now is a disallowed method (clippy.toml: traps on wasm).
The regression test only needs to finish; the old 2^k walk would never return.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* calc: bound precedents to 100 ms and box plain-reference range chains (#2183)
- range_operator_chain_precedents_are_linear times the direct precedents
call against the 100 ms bound in the issue.
- A Range node with two plain Ref operands (A1:B2:C3, which the parser
can't fold) now collects its bounding box like any other range node, so
an edit to C1 recalculates.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* M0: reject negative fractional GCD and LCM arguments
Validate the original input before truncation. Add scalar and range regressions, including signed zero, positive fractions and zero-first LCM validation.
* M0: document negative-fraction GCD and LCM regression
Record tested edge-case behavior and update the parity revision history. Function counts and readiness estimates remain unchanged.
* M0: preserve GCD and LCM regression after math test extraction
Move the original negative-fraction regression into the existing extracted math tests. Scalar and range cases, zero-first validation, signed zero and positive truncation coverage are unchanged.
ACCRINTM returned #NUM! whenever issue >= settlement. Excel rejects only
an issue date after settlement: when the two dates are equal after
truncation to whole days (including distinct times on the same day) it
returns 0 in every basis and both date systems. Relax the guard to
issue > settle; zero rate, zero par and an invalid basis stay #NUM!.