CI's run of the store refusal property found a payroll run that posted
$65.27794: the employer tax was passed in unrounded and nothing refused
it, though rules.md B-3 already says a ledger in one money holds no
fractions of its smallest unit. Sums of it then differed in scale
between journal replay and the SQLite read-back.
- Ledger::post_entry refuses any line finer than its currency's ISO 4217
minor unit (FinerThanMinorUnit), whatever made the line: payroll,
closes, disposals. add_line refuses it up front. Older journals replay
as booked through replay_post_entry. Spec tests for the kernel's
ten-place limit now run on the replay path.
- The property's body is now run_ops, so failing sequences become named
regression tests (regression_payroll_with_a_fraction_of_a_cent). The
generator rounds payroll tax to the cent, as callers must.
- Ledger derives PartialEq and gains differs_from; Store::verify compares
the live ledger with the journal replay in full (drafts, documents,
schedules, periods, settings), not just balances, and names the parts
that differ. The property checks the ledger its commands built against
that replay and against the reopened books.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
A crash or power loss mid-append can leave a partial last line. The
writer acknowledges an append only after its whole line is synced, so
that line was never acknowledged: opening now saves those bytes to
journal.jsonl.torn.<time>, truncates to the last newline, and journals
journal.tail_repaired (replay and projection ignore it; the GUI shows
it; books.digest reports it in journal_repairs). Damage before the last
newline is still corruption and still refuses to open. Tested at every
byte of the last append.
The MCP server now waits up to ten seconds for the writer lock: a
restarted server (a client reconnecting) raced the exiting one and died
with Locked, which also made mcp_schedules_and_recurring_over_stdio
flaky. The test harness now waits for each server to exit.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
The refusal property now starts from real books and adds compound
commands with valid arguments (book an entry, register an asset or
schedule, raise an invoice or bill, close a month to members), so every
command's success path is exercised. Compound commands checkpoint after
each step, so the law is checked against the one command refused.
STORE_CASES sets the case count (256 by default).
It found two bugs:
- run_recurring created drafts dated inside a closed period (B-12).
generate_recurring now refuses them, atomically; the store leaves
them ungenerated and books.run_recurring reports them in
skipped_closed_period for a human to decide. Replay uses
replay_generate_recurring so older journals still open. The store
also journals before mutating the ledger here, as everywhere else.
- Disposing an asset for no proceeds (scrapping it) journaled a
catch-up depreciation entry and a half-built draft with a 0.00 line,
then refused. Zero lines are now left out (Store::disposal_lines,
shared with the MCP preview), and the whole disposal is rehearsed on
a clone before anything is written.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Max-testing pass on ledger-store, fault injection (rank 1). A full disk
is simulated with no code seam: the test re-runs its own binary as a
child under bash with the file size capped (ulimit -f, SIGXFSZ
ignored), so writes past the cap fail as on a full disk. Two bugs, both
of which left the books unopenable:
1. SQLite side: applying one journal line runs several statements (rows,
events mirror, last_applied_seq) in autocommit. When the disk filled
partway, the rows committed but last_applied_seq didn't, and the next
open's catch-up re-applied the line into "UNIQUE constraint failed:
entries.id". Each line now applies inside a savepoint: all or
nothing. (docs/receipt-store.md promised exactly this heal; it
wasn't atomic.)
2. Journal side: an append that failed partway left a partial line, and
any later line would land after it, mid-journal. append now cuts the
file back to its length before the write on any error.
Tests (tests/faults.rs, its own binary because the child re-runs it):
a_full_disk_mid_command_leaves_the_books_whole (the database fills
first) and a_full_disk_mid_append_leaves_only_whole_journal_lines (the
journal writer alone). Both shown to fail without their fix.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Max-testing pass on ledger-store (security property, rank 2). Blob ids
come from the journal, which is untrusted input on disk. BlobStore built
paths from any string: an id under four bytes (or with multi-byte
characters) panicked on slicing, and exists() would report on any path
an id like "../../x" reached. get() never served such bytes (it serves
only content whose SHA-256 equals the id), but it read them.
Now only a 64-hex-digit id names a blob; anything else is an error from
get() and false from exists(), touching nothing. Test
a_blob_id_never_reaches_outside_the_blob_store (empty, short, unicode,
traversal, absolute, non-hex, 63/65 digits; uppercase hex still works).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Max-testing pass, step 8. The final mutation run left 112 survivors
(89% caught). Three agents triaged them by file, by the playbook's
table: spec the boundary then test it, a law rather than one more
pinned value, assert the real value, never weaken a test. Every expected
value is hand-computed.
- reports.rs (35): 28 killed, chiefly cash flow (inclusive start date,
each line's sign and label, a loss on disposal moving to investing,
financing designation), general ledger and members' equity date
boundaries, and the balanced() checks (a real report nudged by a cent
must stop balancing). 7 equivalent: four >/>= after an early return on
zero; two in dead code (removed in this commit); one sign of a zero
cleaned on output.
- ledger.rs, kernel.rs, journal.rs, money.rs (52, plus one new): 49
killed, including every replay_* mutant at the core level (ids,
accumulation, status, vendor tags, out-of-order refusal), FX rates,
0%/100% deductible boundaries, a dividend sweep with no income, the
legacy member rounding at close, a bill paid on its own date, a
liability-account reconciliation, mixed-currency entries (via serde),
ISO 4217 minor units, Display/Debug. 3 equivalent: zero refused before
a </<= check; a period whose end is in a closed period refused earlier;
columns(0).
- schedules, tax, analysis, payouts, assets, form1099 (25): 23 killed
(schedule length 1-600, recurring end on its start, frequency parsing
and stepping, cost of goods in tax net, a zero payout, disposal on the
purchase date, quarter starts, labels); 2 equivalent (zero adds zero;
days_held is provably >= 1).
Two small production cleanups the triage surfaced: kernel::columns(0)
returned a negative-zero credit that would print "-0" (regression
a_zero_movement_has_no_negative_zero_column, shown to fail before); and
cash_flow computed a book value it never used.
Spec lines added where a boundary wasn't specified (cash flow's
inclusive start and line signs, members' equity start date, payout with
no deposit, disposal on the purchase date, bill paid on its own date).
`just mutants` now runs core and store tests (the store's replay tests),
not the whole workspace, which rebuilt the viewer for every mutant.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Max-testing pass on ledger-store, found during the survey: nothing
stopped two processes from opening the same books for writing. Two
agent sessions (two MCP servers) on one books dir would each extend the
hash chain from the same last line; one write from each broke the
journal ("seq gap: expected 3, found 2") and the books would not open.
The store now holds an exclusive lock on books-data/wanderbooks.lock
(std's File::try_lock, no new dependency) from before it reads the
journal until it is dropped; the OS releases it if the process dies. A
second writer gets StoreError::Locked saying another process has the
books open. The viewer reads through Snapshot and never takes the lock.
The lock file is git-ignored in the books scaffold.
Regression: a_second_writer_is_refused_while_the_first_has_the_books
(refused while open, allowed after close, books whole).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Max-testing pass, ranks 4-6 and step 9 for ledger-core.
Fuzzing (rank 4): fuzz_op_sequence now drives balanced pairs, foreign
lines, periods, closes, and amounts from raw decimal bits, and asserts
the books stay valid (TB balances, posted entries validate, no draft in
a closed period); new fuzz_core_serde (every persisted core type
round-trips) and fuzz_store_open (a corrupt journal is an error, never a
panic); hangs count (-timeout=10). 54M executions; one crash:
- fuzz_journal_line: a number too big for an integer parsed as a float,
and its canonical form (which the hash chain covers) re-canonicalized
differently. The journal never holds floats (R-2), so the reader now
refuses them (JournalLine::parse, used by every journal reader);
regression a_journal_line_with_a_float_is_refused. Checked first: the
real books' journal has none. Rerun: 13.2M executions, clean.
Seeds in fuzz/seeds/ (synthetic, minimal); the grown corpora stay out of
git. `just fuzz <target> <seconds>`.
Miri (rank 4): the core's unit, bridge, and invariant tests under Miri
(our code has no unsafe; rust_decimal and chrono do): no undefined
behavior. `just miri`. A pinned nightly (nightly-2026-10-01, with miri)
is now in mise.toml.
Benchmarks (rank 6): criterion (new dev-dependency: the playbook's Rust
benchmark tool; none existed) in benches/books.rs; baselines in
docs/testing.md. The store's per-command rehearsal clone costs ~1 ms at
10k entries. No budget yet, so nothing gates on them. `just bench`.
CI: .github/workflows/check.yml runs `just check` on every push;
scheduled.yml runs proofs, the GnuCash oracle, fuzzing, and mutation
weekly. Tools come from mise.toml.
docs/testing.md: every layer, its oracle rank, command, runtime, and
gate; benchmark baselines; the ASSUMED invariants for Stephen to reject;
the gap list with triggers. docs/oracle.md's fuzzing paragraph had
claimed persistent corpora that didn't exist; corrected.
Also: ledger-import now inherits the workspace lints (nightly cargo
warned).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Max-testing pass, rank 3. tests/model.rs transcribes the Lean ledger
state machine (verify/lean/Wanderbooks/Ledger.lean) into the plainest
Rust with the same observable behavior: lists of accounts, drafts,
posted entries, and periods. Random command sequences (add account,
draft, add line, a balanced pair of lines, post, reverse, create period,
close), weighted toward commands that interact, run against both; after
every command the engine and the model must make the same accept/refuse
decision and agree on the posted log line for line, the chart, drafts,
periods, ids, and every account's balance. 2,048 sequences of up to 60
commands; about 3s.
Honest about its first version: it agreed vacuously. Counting accepted
commands (MODEL_STATS=1) showed zero posts and zero reversals: random
lines never balance. A balanced-pair command fixed it (now ~150 posts,
~100 reversals, ~200 closes per 512 sequences). Shown to have teeth:
letting the model reverse an entry twice fails in 0.03s.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Three spec decisions from Stephen (2026-10-13), raised by the spec-blind
test writers, each test-first:
- A payment dated before its invoice or bill is refused
(PaymentBeforeDocument); a real prepayment is a deposit. Test:
a_payment_never_predates_its_document. Replay is unaffected (it goes
through replay_invoice_payment / replay_bill_payment).
- Aging as of a date shows only documents issued by then, less only the
payments made by then. The ledger now records each document's payment
entries (live and on replay), and aging counts those dated by as_of.
Test: aging_as_of_a_date_sees_only_what_existed_then; the A/R and A/P
conservation properties' aging oracles updated to the confirmed rule
(they had assumed aging ignored dates).
- A schedule through a mid-month date prorates that month by days; the
code was right and its doc wrong. The spec-blind test that disagreed is
rewritten to the confirmed spec and no longer ignored.
Also found while wiring payments: a payment to an unknown cash account
created the payment's draft and then failed, leaving it behind; the cash
account is now checked before anything is created.
The unreviewed golden van_schedule.json is gone: every row was first
checked against an independent hand computation (all 61 matched), then
replaced by van_schedule_is_the_hand_computed_one, which asserts that
computation directly (548.39, 59 x 1,000.00, 451.61; total 60,000.00).
Tests: 452. No spec-disagreements left ignored.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Max-testing pass, steps 2-4.
Spec-blind tests (rank 1-2): four agents, each given only the specs and
the public interface (doc comments and signatures, every body removed,
never src/ or the existing tests), wrote tests/spec_reports.rs (62),
spec_ledger.rs (72), spec_arap.rs (42), spec_assets_payroll.rs (50):
hand-computed examples per scenario and a proptest law per invariant
(conservation, partition, invariance, scaling, reversal cancels,
"a refusal leaves the books unchanged" over random invalid commands).
They found:
1. A refused payroll run changed the books: an all-zero run passed
validation, then failed at post after creating seven payroll
accounts and using up an entry id. validate_payroll_lines now
refuses a run with no gross pay.
2. A refused add_account used up an account id (empty code, duplicate,
unknown parent). The id is now taken only when the account is added.
3. Stale prose: close_period_to and members_equity still described the
old last-member rounding.
4. Open (left #[ignore]d pending Stephen): depreciation_schedule's doc
says "through the month containing" the date; the code prorates that
month by days.
GnuCash oracle (rank 3, an oracle never the spec): `just oracle` builds
GnuCash 5.10 in a container pinned by base digest and package versions
(oracle/gnucash/) and runs tests/oracle_gnucash.rs: 200 seeded random
books built through wanderbooks (entries, foreign lines, reversals,
monthly closes), every account's balance at month ends and random dates
and each month's P&L, compared exactly: 15,800 comparisons. First run:
4,724 differences, all from foreign lines; without them, zero.
Adjudicated from principles (rules.md B-3): wanderbooks booked
foreign * rate unrounded ($108.34567); a ledger in one money holds no
fractions of its smallest unit, and its cash must match a bank. Stephen
decided: book in the base currency's minor unit, half away from zero.
add_line_fx now rounds; minor units per ISO 4217 list one (2026-09-17,
SIX), generated into Currency::minor_units and recorded in
docs/spec/sources/iso-4217.md. Replay now takes the journaled base
amount (replay_add_line_fx), so lines booked before replay exactly as
booked (contract: books_written_under_older_rules_still_open). Now:
zero differences, foreign lines included.
Tests: 449, from 223.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Max-testing pass, step 2: every domain decision in ledger.rs and
reports.rs gets a GOAL / SOURCE / INVARIANTS / SCENARIOS block in its
doc comment: post_entry, try_create_entry, close_period_to,
mark_cleared, reconciliation_report, account_balance, create_invoice,
create_bill, receive_payment, pay_bill, register_asset, dispose_asset,
run_payroll, remit_payroll_taxes, set_line_class, set_entry_vendor;
trial_balance(_as_of), profit_and_loss_for_period, balance_sheet_as_of,
general_ledger, members_equity, cash_flow, profit_and_loss_by_class,
snapshot. Sources trace to docs/spec/rules.md (and through it to
Pacioli, the IRS, FASB); what no document settles is marked CONVENTION,
DECISION, or ASSUMED for Stephen to confirm. The specs are the input to
spec-blind test writing (the next step): tests derived from the spec
alone, never from the code.
Also: rustdoc now runs in `just check` with -D warnings; fixed the seven
existing broken intra-doc links and one bare URL.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Three spec decisions from Stephen (2026-10-13), each test-first:
- A child account has its parent's type (ParentTypeMismatch): a child
subdivides its parent. Test: a_child_account_has_its_parents_type.
- Periods nest (a month in a quarter in a year) but never partially
overlap (PeriodsOverlap; a refusal uses no id). Test:
periods_nest_but_never_partially_overlap.
- A line's account must exist and take postings when the line is added,
not only at post (UnknownAccount / PlaceholderAccount from add_line
and add_line_fx); posting still re-checks. Two older tests that pinned
the post-time refusal now check both (the post-time check is reached
through the unchecked replay path, as old journals can).
Compatibility: replay reads history through new Ledger::replay_*
commands (add_account, add_line, add_line_fx, create_period) that skip
checks added later, so books written before these rules still open.
The contract test books_written_under_older_rules_still_open writes such
a journal directly and found, on its first run, that the SQLite
cross-check rebuild used the checked commands and would have failed
verification ("sqlite projection can't be read back"); it now replays
too. Documented in docs/spec/README.md; decisions recorded in rules.md
(B-1, B-4, B-12).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Stephen (2026-10-13): when what wanderbooks knows is stale, check the
date, research it or ask the owner or CPA, and update the rule for
everyone with a pull request (best) or an issue (worst).
- ledger_core::rules: a registry of rules set by law that change over
time. Each says the years it's known for, its rules.md entry, its
vendored official source, and a review date. stale_rules(today) is
pure; the engine has no clock.
- The 1099 threshold is the first: $600 through 2025, $2,000 in 2026,
then indexed for inflation (26 U.S.C. §6041(h)). default_threshold
and vendor_1099_report now return Result; for 2027+ they refuse with
LedgerError::RuleNotKnown, whose message says what to do: get the
figure from the official source or the owner/CPA, pass it explicitly,
then open a PR (or issue) at github.com/swaits/wanderbooks. Before,
2027+ silently used $2,000.
- MCP: books.vendor_1099 relays that refusal; books.digest gains
stale_rules; the server instructions tell the agent the procedure (and
how to handle control-account conflicts: explain, never renumber).
The CPA package writes a NOT-INCLUDED note instead of a guessed 1099
worksheet.
- Viewer: "Needs attention" shows each stale rule with a ready prompt
for the agent; the 1099 page explains instead of guessing.
Review date for the 1099 rule: 2026-11-01, when the IRS usually
publishes next year's inflation adjustments (CONVENTION, marked as such
in the registry).
Tests: an_unknown_year_threshold_is_never_guessed,
dated_rules_go_stale_on_their_review_date (dates passed in, so no clock
in the test), MCP stdio: 2027 refused with guidance, explicit threshold
accepted, digest carries stale_rules (asserted date-independently).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Invoices, bills, payments, payroll, and closes post to accounts the
engine finds by code (1200, 2000, 4000, 6999, 3999, 6100-2150). If an
account with that code existed it was used whatever its type: a 1200
"Customer Deposits" liability would have taken receivables. Now
(Stephen, 2026-10-13: never force an imported chart into our numbering)
the engine refuses with ControlAccountMismatch, naming the code, the
account, its type, the role, and the type needed, so the agent can work
it out with the human. A placeholder is refused too.
Every command checks all its control accounts before changing anything
(Ledger::check_controls, spec + SOURCE in its doc comment), so a refusal
never leaves a half-made account behind. One table of control accounts
(ledger_core::RECEIVABLES ... PAYROLL) replaces the codes duplicated
across the engine and the store.
Tests (spec-derived, all assert the books are untouched after the
refusal): an_invoice_refuses_a_1200_that_is_not_receivables,
an_invoice_refuses_a_placeholder_1200,
an_invoice_creates_nothing_when_its_second_account_is_wrong,
a_bill_refuses_a_2000_that_is_not_payables,
a_close_refuses_a_3999_that_is_not_equity.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Stephen asked that specs come from recognized documents, not from
anyone's memory. There is no single US bookkeeping bible; authority is
layered, and this records it:
- docs/spec/sources/: verbatim excerpts (whitespace normalized), each
with URL, retrieval date, and the SHA-256 of the fetched file, of
public-domain texts: Pacioli's 1494 treatise in Geijsbeek's 1914
translation (double entry, the trial balance, one money, write only
forward, correct by opposite entry, close into capital); IRS Pub. 583,
538, 946, 15; the 1099-MISC/NEC instructions (2025 and 2026); 26 U.S.C.
§6041, §6041A, §704; Treas. Reg. §1.263(a)-1(f) and -4(f); Notice
2015-82; Rev. Proc. 98-25; Form 1065 instructions; 17 U.S.C. §105. The
excerpts were extracted by matching each passage against the fetched
file, so none can contain text the source doesn't.
- docs/spec/rules.md: the rule catalog, B-1 to B-22. Each rule says
AUTHORITY, CONVENTION, or DECISION; cites its sources (FASB's
Codification and Concepts Statement 8 and OpenStax quoted briefly,
never copied: copyright and CC BY-NC-SA); and names the code that
enforces it and the tests and proofs that check it.
- Spec blocks in the code now carry SOURCE: lines into the catalog;
AGENTS.md makes sourcing rule 0 of the formal-methods bar.
Corrections the sources turned up: the de minimis limit is $500 in the
regulation and $2,500 by Notice 2015-82 (both now cited); 1099
thresholds after 2026 are inflation-indexed, so the engine's flat
$2,000 for 2027+ is an open question (B-19). Unverified ASC paragraphs
(360-10-35-4, 340-10, 230-10-45-10, 606-10-55-37B, 210) are listed in
docs/spec/README.md until someone with a FASB basic-view login reads
them.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Max-testing pass, step 2: two spec decisions Stephen confirmed
(2026-10-13), each test-first.
Aging (arap::aging_bucket): a document due today was "1-30 days past
due"; it is Current, since past due starts the day after (QuickBooks'
convention). Spec in the doc comment. Boundary examples at every edge
(due today, 1, 30, 31, 60, 61, 90, 91 days) fail on the old code. A
property: the balance lands in exactly one bucket, and time only ages
it.
Member splits (periods::allocate): the last member absorbed every odd
cent, so the same partner always got or lost it. Now largest remainder:
each share rounded toward zero to the cent, leftover cents to the
largest remainders, ties to the earlier member, a loss the negation of
the gain. The core is a new kernel function, kernel::apportion,
allocation-free so Kani proves it (n <= 3): apportion_is_exact (it
succeeds exactly when the shortfall is under a unit per part; then the
parts sum to the whole, each gains 0 or 1 unit, one may take the
sub-unit fraction) and apportion_is_fair (never a unit to a part while
a larger remainder goes without). Kani found an off-by-one in the first
version: a shortfall of n units plus a fraction was accepted. Native
regression apportion_refuses_a_shortfall_of_a_unit_per_part_or_more.
Properties at Decimal (2,000 cases, up to 5 members): conservation,
within a cent of each exact share, fairness, sign symmetry; they fail on
the old rule. Bridge: apportion agrees on i64 and Decimal.
Contract: closes journaled before this change replay exactly. New
member closes journal "rounding": "largest_remainder"; journal lines
without it replay as CloseTo::MembersLastAbsorbs, the old rule
(periods::allocate_last_absorbs, pinned by
closes_journaled_under_the_old_rule_replay_with_it). The members'
equity statement isn't journaled and uses the new rule.
Open gap: Lean proof of apportion's sum law for any n (needs a
permutation lemma on the ranking); Kani covers n <= 3, properties n <= 5.
The Lean closing theorems already hold for any split summing to the
net.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Max-testing pass, step 2 (rank 1 regressions + a rank 2 law). Two ways
to make the books unopenable, both found by test, both fixed:
1. Reversing into a closed period: the store journaled `entry.reversed`,
then the engine refused it at post. Replay could never apply that
line, so Store::open failed ("period 1 is closed"). Regression:
reversal_into_a_closed_period_is_refused_before_journaling.
2. An oversized invoice (or any command whose hand-written store
pre-checks lagged the engine, here the new line range) journaled,
then failed in the engine. Found by the new property
refused_commands_never_reach_the_journal within 4s.
The fix is structural, not another hand-written check: every
journal-first store command now rehearses on a scratch clone of the
ledger (`Store::rehearse`), so whatever the engine refuses is refused
before the journal, now and after any future engine change. The
create_entry rehearsal uses the checked constructor: no draft is ever
dated in a closed period (the Lean invariant Inv.draftsOpen, which the
store had bypassed).
The property (1,000 random command sequences, heavy on refusals)
asserts, after every command: a refusal leaves the journal
byte-identical; every posted entry validates; the trial balance
balances; no draft sits in a closed period; and the books reopen and
verify at the end. Shown to have teeth: removing the create_entry
rehearsal makes it fail and shrink to open-period, close, draft.
Engine (spec confirmed by Stephen, 2026-10-13), Ledger::check_reversal,
a pure pre-check that reverse_entry and the store share:
- an entry is reversed at most once (EntryAlreadyReversed); reversing
the reversal re-instates it;
- a closing entry is never reversed (ClosingEntryIrreversible);
- a refused reversal leaves no trace (it used to leave a draft and
burn an id).
Specs go in doc comments as a ```text GOAL/INVARIANTS/SCENARIOS block.
New dev dependency: proptest for ledger-store (already the workspace's
property library).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
The money arithmetic every invariant rests on now lives in one kernel,
ledger-core/src/kernel.rs: signing a line, totals, net, per-account
movements, reversal, normal balances, trial-balance columns, closing
lines, and the rounding plug for member splits. It is generic over a
checked `Amount` trait. The engine runs it on Decimal (validate,
movements, reverse_entry, account_balance, close_period_to, allocate,
trial_balance): there is one copy of the arithmetic, and it is the one
under proof.
Three layers, docs/verification.md maps every law to its proofs:
- Kani (src/proofs.rs): 17 harnesses prove the kernel at i64 for every
input up to 4 lines over 3 accounts. An entry is accepted exactly
when it is two or more positive, in-range lines whose debits equal
credits; accepted entries net to zero; grouping by account loses
nothing; a reversal mirrors and cancels its original; trial-balance
columns balance iff the books net to zero; closing lines zero their
accounts and a close balances however the net is split; the plug
makes parts sum to the whole. No panics, no overflow. kani::cover!
shows the proofs are not vacuous.
- Lean 4 (verify/lean): 24 headline theorems for books of any size. The
ledger is a state machine (add account, draft, post, reverse, open,
close), and in every reachable state the books net to zero, the trial
balance balances (as of any date), and the accounting equation holds.
Posting only appends ("posted is forever"), closed periods are
frozen, a close zeroes revenue and expenses and credits exactly the net
income, and period statements add up. Audit.lean shows standard axioms
only, no sorry.
- Bridge (tests/kernel_bridge.rs): the Decimal and i64 kernels agree.
What the proofs found (each one missed by every test, fuzzer, and
mutant):
- The engine accepted negative line amounts; only MCP refused them. Now
NegativeAmountLine: direction is the side's job.
- Summing lines near Decimal::MAX panicked. Lines are now capped at one
quadrillion and ten decimal places (AmountOutOfRange), which keeps
every realistic sum exact, and the kernel's sums are checked.
- FX conversion (foreign_amount * fx_rate) could panic the same way; it
is now a checked multiply.
- Kani itself: 0.68 mis-models symbolic-length copy_from_slice
(model-checking/kani#4904) and reported a reversal that doesn't net to
zero. The counterexample, replayed natively, passes, and a 15-line
standalone crate reproduces the bug. Harnesses avoid bulk copies, and
a should_panic canary will flag the upstream fix.
`just prove` runs it all (Kani memory- and time-capped, fails on an
unsatisfiable cover or a non-standard Lean axiom). AGENTS.md now makes
formal coverage part of every core change.
Also, docs: GnuCash is *an* oracle, never *the* oracle (Stephen). The
spec is accounting principles, proved. Golden files are recomputed from
principles. External implementations (GnuCash, the old Sheet, QBO/Xero)
are cross-checks whose disagreements get adjudicated, never obeyed.
docs/oracle.md gains "Who is right"; AGENTS, README, architecture,
roadmap, parity, and STATUS are reworded to match.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Rust (with clippy and rustfmt), just, Node (for deploys), Kani, Lean,
cargo-mutants, and cargo-fuzz, so `mise install` is the whole bootstrap
and nothing gets installed by hand. Lean is 4.34.1, the newest release
past mise's minimum release age; verify/lean/lean-toolchain matches.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
The general ledger draws its own per-account tables, so it had no PDF
button. It now has one: each account a section, a line per posting
(date, entry, description, the other side) with debit, credit, and
running balance, its beginning balance, and totals with the ending
balance, typeset through the same statement template.
The PDF plumbing (button, status, background job) is shared by both
paths now. The ledger's rows also measure and clip their text and
scroll sideways in narrow windows, like every other table.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Reports get a Save PDF button beside Copy for a spreadsheet. It typesets
the report as it's set up on screen (range, months/quarters/years,
comparison) through ledger-import's Typst statement template, the same
one the CPA package uses, in the background, writes it to the Downloads
folder (never the books: the viewer stays read-only), and opens it.
- ledger-import: Statement.percent_from marks percentage columns
(comparison %, % of income), typeset as 12.5% with no dollar sign;
wide statements turn the page, and past seven columns set smaller so
a year of months fits; statement() is public for other callers.
- Compared columns that read the same (Jan 1 - Oct 12 vs a year
earlier) now say which year, on screen and on paper.
- Short windows: the sidebar's and rail's links scroll above the status
box instead of under it. The search palette fits the window.
- A test typesets every table-backed report from the demo books,
plain, by month, and compared (PDF_DIR keeps the files).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
Narrow windows overlapped and clipped: the top bar's crumb, date, and
search ran into each other; big numbers wrapped mid-figure; tables
painted columns over each other; the entry and document panes squeezed
their lists to slivers; anything too wide widened the whole page.
- Below 1000 points the sidebar becomes an icon rail (names, attention,
and integrity checks in tooltips).
- The top bar lays out right to left: search shrinks to its icon, the
date drops, the crumb truncates.
- Pages are pinned to the window width with narrower margins when tight,
so one wide thing can't push everything off screen.
- Tiles (KPIs, cash accounts, documents) fit as many columns as the
width allows, balanced across rows; KPI figures shrink rather than wrap.
- Painted tables shrink their amount columns, clip labels with an
ellipsis, then scroll sideways rather than overlap.
- List/detail splits show only the detail when both won't fit.
- Reports: the side list becomes a picker; controls wrap by known width.
- Smaller fixes: chart labels thin out, company facts stack, journal
timestamps yield, transaction filters become a dropdown.
- Min window size drops to 360x400; the screenshot test covers every
report, a register, a document, and a journal event.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GWVo7x3UHw5kwCUaWxiejX
New books dirs get .codex/hooks.json beside .claude/settings.json; both
SessionStart hooks run one justfile recipe, which reports the viewer's
state per [viewer] open_at_start (and opens it itself for "always").
Codex runs project hooks only once the dir is trusted and the hook
approved, so new --codex says so; AGENTS.md stays the fallback.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Tpv1gBaepKjQiAVJqejGbA
The viewer offer (open_at_start) lived only in AGENTS.md, so it depended
on the agent remembering it on its first turn. The scaffold's
.claude/settings.json now runs just gui-status at session start and hands
the agent the result plus open_at_start.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Tpv1gBaepKjQiAVJqejGbA
The scaffolded justfile now detaches the viewer (log in books-data/gui.log)
and won't open a second one for the same books. AGENTS.md tells the agent
to check gui-status at session start and follow open_at_start
(ask/always/never) in wanderbooks.toml.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DeBdcyb3YyRkcpicBZH5wh
The SQLite cross-check rebuilt entries by id and assumed no gaps, but
drafts aren't projected until posted: a draft #8 before a posted #9 made
the server refuse to start. Gaps are now held with empty drafts (which no
report counts), and a projection that can't be read back counts as drift
to heal, not a reason to refuse to start.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
books.show asks the viewer to open an entry, an account register, a
report (dates, month/quarter/year columns, prior-period or prior-year
comparison), a document, a search, or a page. The request lives outside
the books (runtime dir, keyed by data dir, written atomically); the
viewer watches it beside the journal and applies books before
navigating. 66 tools.
Site: the dashboard is the hero again; 'Here's what using it looks like'
shows the session beside the viewer, ending in books.show; the cash page
replaces the repeated overview lower down.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
The session (an HTML terminal pane, demo company) records the Adobe
receipt; the viewer shows entry #245 with its lines and rendered receipt.
Screenshot harness: SHOT_LIVE labels a fixed-books app as live.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
wanderbooks new --codex registers the books' MCP server with Codex (as
wanderbooks-<dir>) and starts Codex; the scaffolded justfile gains codex
and codex-setup. The site drops 'hire', says free up front, shows
Claude Code or Codex + wanderbooks = a complete bookkeeping system, and
explains MCP and cost in the FAQ.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
CPA package: every receipt behind the period's entries with an index,
company documents, the journal as IIF, the 1099 worksheet, and a README;
include_receipts=false for a light package. The scaffolded AGENTS.md
covers receipts from email, switching from QuickBooks/Xero/spreadsheets,
and working with the CPA. GUI: a books copy without a server database
reads 'Read-only copy'. Site: example prompts, QuickBooks comparison,
what it does and doesn't do yet, the CPA section, FAQ.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
Store::verify compares the whole SQLite projection, every table row for
row, against a fresh projection of the journal; verify_and_heal rebuilds a
drifted database (keeping the old file) at every server start and via the
new books.verify tool, which also re-hashes every blob. The GUI checks the
server's database read-only and every stored file, and shows the result.
rebuild() also drops stale WAL/SHM files.
site/: the wanderbooks website (just site, just publish). install.sh:
OS detection (Windows points to WSL2), C compiler and git checks,
WANDERBOOKS_REPO. Windows: no directory fsync off Unix; .exe-aware lookup.
Repo justfile: check, build, install, demo, site.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
wanderbooks new DIR installs the server and GUI if missing, writes the
books dir (AGENTS.md + CLAUDE.md, wanderbooks.toml, justfile, .mcp.json,
.claude/settings.json approving the server), commits, and starts claude
onboarding. install.sh: Rust + cargo install + wanderbooks new in one
line. The tax summary labels its net when accounts are uncategorized.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
egui/wgpu app for owners and their accountant: Overview, Cash, Profit,
Transactions (with rendered receipts and journal history), Reports
(statements with interval columns, comparisons, drill-down), Accounts,
Documents, Company, Journal; Ctrl+K palette; back/forward; light/dark.
Follows the journal live (~30 ms from write to screen). --demo opens a
deterministic demo company, also used for screenshot tests.
Engine: ledger-store Snapshot (read-only, incremental, tolerant of a
half-written tail); ledger-core analysis (intervals, balance series,
P&L by interval, cash flows); ledger-config extracted from ledger-mcp
with the full company profile and cash_accounts.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
Vendors carry form_1099 (NEC/MISC) and tin_on_file (never the TIN).
Entries carry the vendor they paid; bills, bill payments, and reversals
are tagged automatically. books.vendor_1099 counts cash paid per vendor
per calendar year (card payments excluded) against $600 through 2025 and
$2,000 from 2026, or an override, and flags missing W-9s. Four new tools
(64), two journal events, schema v7. Docs: finished R-FUTURE work is no
longer listed as future.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
A platform deposit becomes a draft entry with the deposit, fees,
adjustments, and gross revenue (by class), refused unless gross - fees
+/- adjustments equals the deposit to the cent. The payout statement is
attached as its receipt before posting. 60 tools.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
Amortization schedules spread a prepaid expense or deferred revenue over
the months it covers, to the cent, posting on month ends; runs preview
first, are idempotent, honor closed periods, and are exempt from the
receipt policy like depreciation. Recurring entries generate drafts only,
never posts. Six new tools (59 total); four new journal events, replayed
and checked.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
Every business has owners, and how owners put money in and take it out
depends on what the business is. wanderbooks now knows the four common
US shapes:
- books.chart_template {entity_type, owners}: a starter chart for a sole
proprietorship (owner's capital and draws), partnership or
multi-member LLC (capital and distributions per partner, guaranteed
payments), S corporation (stock, APIC, shareholder distributions,
officer compensation, retained earnings) or C corporation (dividends,
income tax expense), plus common income and expense accounts with
tax categories, and the wanderbooks.toml settings that go with it
(tax form, members with equal-share placeholders that sum to exactly
100, close target). A proposal for the agent to review with the human.
- Closing: CloseTo is now RetainedEarnings { sweep }, Account { into,
sweep } (a sole proprietor's capital) or Members([MemberClose { capital,
share, distributions }]). Sweeps zero owner-pay equity accounts into
their target at close: draws into owner's capital, dividends or
distributions into retained earnings, each partner's distributions
into their own capital. Validation (equity, distinct, shares sum to 1)
runs in the pre-close preview, before anything is journaled. The
journal records the target; journals with no target or the old
"retained_earnings" string still replay.
- Members' equity: a partner's distributions account is part of their
capital; its activity reports as distributions.
- Config: members' distribution_account; [books] close_into and
close_sweep for single-owner businesses (rejected alongside members).
The cash flow names members on their distribution accounts too.
Tests: a sole proprietor (net 2,000 and 500 of draws into capital: 6,500,
the cash); a corporation (dividends 300 into retained earnings: 600); a
partnership with distribution accounts (Bob 100 + 25 - 20 = 105, swept at
close, February begins there); bad sweep accounts; templates round-trip
through the chart parser, unique codes, equity per entity, every tax
category valid for the entity's return; stdio: template -> settings ->
import -> postings -> close with the exact closing lines.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
wanderbooks is a general MCP-first accounting application: every
business needs its accounts mapped to where they land on its income tax
return, and its CPA needs book-to-tax differences surfaced, not hidden.
- ledger_core::tax: a catalog of stable categories (gross_receipts,
commissions_fees, meals, guaranteed_payments, officer_compensation,
nondeductible, ...) with kind (income, cost of goods, deduction), the
returns each appears on (Schedule C, 1065, 1120-S, 1120), a default
deductible percent (meals 50, nondeductible 0), and guidance. Keys,
not line numbers: form lines move from year to year.
- Accounts carry tax_category and an optional deductible_percent
override, validated against the account type (and the return, at the
MCP layer), journaled as account.tax_set, replayed, projected (schema
v6).
- tax_summary(start, end, form): the P&L grouped by category, book and
deductible amounts (half away from zero to the cent), contra income
(refunds) reducing receipts, and every account still uncategorized.
- MCP: books.tax_categories (filtered to the configured return),
books.set_tax_category, a tax_summary report in every format and in
the CPA package, chart output with categories, and a fifth
tax_category column in chart CSV imports, validated before anything is
created. [company] tax_form, or derived from tax_classification.
Scope stays book basis: no tax depreciation, basis, or filing.
Tests: a hand-computed quarter (50% meals 333.33 -> 166.67, penalties at
0%, refunds as negative receipts, guaranteed payments, one uncategorized
account; book net 5,276.67 vs tax 5,483.33); overrides; category/account
/return validation; form parsing and derivation; CSV parsing; stdio:
filtered catalog, mapping, refusal, summary, all-or-nothing import.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
The close zeroed every revenue and expense account's all-time movement,
so closing a period after later entries were posted swept those too:
close January in March and February's and March's income landed in
January's closing entry. The preview had the same bug and never showed
where the net went. Now:
- close_period_to(id, to) zeroes balances as of the period end; entries
dated later stay open for their own period. Periods must close in
order (LedgerError::CloseOutOfOrder): a later close has already swept
the earlier period.
- CloseTo::Members: for a partnership (multi-member LLC) the net goes to
each member's capital account by share, to the cent, the last member
absorbing rounding, which is what K-1 capital accounts track.
CloseTo::RetainedEarnings (3999) stays the default; close_period is
shorthand for it. books.close_period picks Members when
wanderbooks.toml lists members.
- The journal records the close target; replay re-derives the closing
entry and now compares every line with the journal, not just the id,
so a change to closing math can never silently rewrite history.
- The tool's preview is the real close run on a clone of the ledger
(Ledger now derives Clone), so preview and post can't disagree.
- Members' equity: closing entries into a capital account are income,
not contributions. Balance sheet for a partnership: capital accounts
(with allocated earnings), then only unallocated earnings —
"Undistributed earnings, prior years" if any, and this year's unclosed
net income (loss). periods::allocate is shared by the close and the
statement.
Tests: January closed after February is posted sweeps exactly Dr 4000
2,500 / Cr 6100 800 / Cr 3999 1,700 (mutation-checked: the old code gives
3,500/120/800/2,580); out-of-order close refused; a partnership close
splits 666.67 into 333.34/333.33 with no 3999, and the members' equity
statement reads correctly before and after; store replay of a members
close (reopen, rebuild, verify); partnership and corporate balance sheet
layouts; stdio close preview equals the posted lines.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
Records three more dashboard references Stephen shared: the QuickBooks
Online dashboard cards, KPI tiles with period-over-period change, and a
cash-flow dashboard (runway, forecast, A/P and A/R aging, burn).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
Every report a CPA expects, laid out the way US small-business books
read (QuickBooks as the reference):
- Exports: books.export gains general_ledger (by account, running
balances, splits), general_journal (the entry-ordered listing, which
the CPA package used to mislabel "general ledger"; alias journal),
members_equity, and trial_balance as of a date, in xlsx/csv/json, and
PDF for all five statements (P&L, balance sheet, cash flow, trial
balance, members' equity). books.trial_balance takes as_of.
- CPA package: every statement as PDF and XLSX, plus the general ledger
and general journal, all for the fiscal year to date.
- Layout (ledger-import::report): multi-column statements (Total;
Debit/Credit; one column per member plus Total), "$" on subtotals and
totals, a rule above every sum (none over an empty section), a double
rule under statement totals, zero as 0.00, negatives in parentheses
with the closing paren hanging so decimal points align.
- Terms: "Net income (loss)"; for a multi-member LLC, "Members' equity",
member contributions and distributions named in the cash flow, and the
balance sheet splitting retained earnings (the Retained Earnings
account plus earnings not yet closed from earlier years) from the
fiscal year's net income (loss); the cash flow's "Adjustments to
reconcile net income (loss) to net cash from operating activities" and
"Net increase (decrease) in cash".
- PDF: Inter replaces the typst-assets serif fonts (bundled, SIL OFL 1.1,
license in fonts/; from the official v4.1 release), centered title
block, column-header row, full-width rules, footer "Accrual basis ·
Prepared <date> · Journal seq N" with page numbers. Amounts on total
rows stay regular weight so bold figures can't leave the decimal
column. typst-assets dependency removed.
- XLSX: centered title block, header row, Excel accounting formats with
"$" attached on sums, general ledger and general journal workbooks.
- Config: wanderbooks.toml [[company.members]] (name, share "50%" or
"0.5", capital_account; shares must total 100%) and [books]
accounting_basis (accrual | cash).
- Core: RETAINED_EARNINGS_CODE names the "3999" that period close uses.
- Fix: "today" is local time, not UTC; in US evenings UTC is tomorrow,
so reports defaulting to "through today" included the next day and
footers carried the wrong date.
Tests: layout (currency and rule placement, retained earnings vs. this
year's net income for a member LLC, footer text); exact template data;
deterministic PDFs; config parsing (members, shares, basis); stdio:
trial balance as of, February general ledger (beginning 1,004.00, splits,
ending 1,003.50), members' equity (contributions 1,000.00, net income
3.50, ending 1,003.50), members' equity PDF, the CPA package's ten files.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
Three reports a CPA expects that the engine could not produce:
- general_ledger(start, end): the canonical general ledger, organized by
account in code order. Each account has its beginning balance, every
posting in the period (date, entry, description, debit, credit, running
balance on the normal side) and its ending balance, which ties to the
trial balance. Each posting names its split: the one account on the
opposite side of its entry, or "-Split-" when that side has several,
so a compound entry (one debit, several credits) reads correctly from
every account it touches. The entry-ordered listing the exports called a
"general ledger" is really the general journal.
- trial_balance_as_of(as_of): a trial balance is always as of a date.
- members_equity(start, end, members): the statement of changes in
members' equity for a multi-member LLC taxed as a partnership. Each
member's capital roll-forward: beginning (capital plus their share of
undistributed earnings), contributions, allocated net income (loss),
distributions, ending. Allocation is by ownership share to the cent,
the last member absorbing rounding, so totals are exact and the ending
total equals balance-sheet equity; any equity change the roll-forward
can't explain shows as "other" instead of disappearing. Shares must be
positive and sum to exactly 1; capital accounts must be distinct
equity accounts (LedgerError::InvalidMembers).
- CashFlowLine.account: the account a line is about, so renderers can
name members instead of account codes.
Tests (every expectation computed by hand): trial balance as of Jan 31;
February general ledger with beginning balances, running balances, a
card that nets to zero, an inactive account still listed, and every
ending balance tied to the trial balance; a compound entry's splits from
all three sides; a two-member partnership where 666.67 and 100.01 split
unevenly (333.34/333.33, 50.01/50.00), with and without closing January,
tying to the balance sheet each month; bad member configuration.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
money() printed whatever scale a Decimal happened to have, so an empty
sum came back as "0" while every other amount read "207.68". Amounts now
have at least two places ("0.00", "1000.00"), padded and never rounded:
an exact FX-converted amount with more places keeps them. Negative zero
normalizes to "0.00".
Test: money_always_has_cents_and_never_rounds; three stdio assertions
that pinned the old "0" now expect "0.00".
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
Zero warnings from `cargo clippy --workspace --all-targets`, which is now
a hard rule rather than "clean except pre-existing":
- ChartOfAccounts::add takes the Account it registers instead of eight
loose parameters (too_many_arguments); Ledger::add_account builds it.
Same validation, same errors.
- close_period tool: collapse the nested if-let into a let chain.
- iif.rs docs: the IIF layout example shows tabs as `\t` instead of
literal tab characters (tabs_in_doc_comments).
- Tests: drop an unused binding; assert! instead of assert_eq! against a
literal bool.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
The printpdf statements padded columns with spaces in a proportional font
(amounts didn't line up), had no company name, printed negatives as
"-207.68", and put everything on one page (a long statement was silently
cut off past ~70 lines). Now:
- PDFs are typeset by Typst from one template,
ledger-import/templates/statement.typ. The laid-out statement goes in as
JSON data (sys.inputs), so account names are never parsed as markup.
Fonts are bundled (typst-assets: Libertinus Serif), so output doesn't
depend on the machine and is byte-for-byte deterministic. Statements
paginate, with the company/title repeated on later pages and "Page n of
m" plus "Prepared from journal through seq N" in the footer.
- US accounting presentation, decided once in ledger-import::report and
applied by both PDF and XLSX: tabular figures; "$" on each section's
first amount and on every subtotal and total, pinned to the column's
left edge; negatives in parentheses with the closing paren hanging (its
width reserved on every row, so decimal points align); zero as a dash;
a single rule above sums (none over an empty section) and a double rule
under statement totals; amounts never bold, because bold figures are
wider and would leave the decimal column. XLSX uses Excel's own
Accounting formats for the same result.
- Cash flow: "Net change in cash" is a subtotal; only "Cash at end of
period" takes the double rule.
- Core fix: cash_flow walked the chart in HashMap order, so the order of
working-capital and financing lines could change between runs. Lines
now follow account code. Test: five owners, twenty runs, always
3000/3001/3003/3005/3009; fails without the sort.
New dependencies (ledger-import): typst, typst-layout, typst-pdf,
typst-assets (fonts), typst-as-lib (World/font plumbing), serde_json (the
template's data). Removed: printpdf. Release ledger-mcp grows by the
bundled fonts.
Tests: exact template data for a hand-built P&L (like a golden); same
statement renders to identical bytes; 200 rows of Typst markup
characters render literally and span several pages; currency-sign and
rule placement; amount splitting.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
Every statement export rendered life-of-ledger totals under whatever date
range or as-of label was asked for, and the P&L and balance sheet XLSX
were totals-only. Now:
- books.profit_and_loss / books.balance_sheet and the csv, json, xlsx,
pdf exports and CPA package all use core's profit_and_loss_for_period
and balance_sheet_as_of, with a line per account. The tools keep their
totals and add revenue/expense/asset/liability/equity account lines.
An omitted start is the first posted entry, never "0001-01-01".
- ledger-import::report lays a statement out once (section, line,
subtotal, total) for every renderer, with one exact amount formatter:
thousands separators, negatives in parentheses, no floats. XLSX keeps
numeric cells with an accounting number format, named sheets, company
header, totals with borders; the trial balance gains its totals row.
- ledger-mcp reads wanderbooks.toml (next to --data-dir, or
--books-toml): company name for headers, fiscal_year_start for the CPA
package's fiscal year to date (end derived; Feb 29 rejected), and
[books] receipt_threshold, which the server had been ignoring.
config.json's receipt_policy still wins when set. New dependency: toml
1.x (already in the lockfile via another crate).
- Store::journal_seq(); every report notes the journal seq it reflects.
Tests: company.rs unit tests (fiscal years incl. a July start, scaffold
parsing, bad settings); report.rs formatting tests; an MCP stdio test with
--books-toml (threshold 5.00 enforced, February-only P&L with account
lines, start defaults to the first entry, as-of balance sheet).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
The engine only had life-of-ledger P&L and balance sheet totals, so every
export labeled with a date range or as-of date (PDF, XLSX, CPA package)
printed all-time numbers under that label. Core now has:
- profit_and_loss_for_period(start, end, class?) and
balance_sheet_as_of(as_of), each with a line per account
(StatementLine: code, name, normal-side amount) and exact totals.
- Ledger::closing_entry_ids(). Period statements exclude closing entries.
A close (Dr revenue / Cr expense / net to Retained Earnings, dated at
period end) moves earnings; it doesn't earn or spend. Counting it made a
closed period's P&L read as break-even, and made the cash flow show zero
net income plus a fake "owner contribution" to Retained Earnings. The
as-of balance sheet keeps closing entries: after a close the earnings
belong in Retained Earnings, not current earnings.
Tests: tests/statements.rs, a two-month fixture with every expected number
computed by hand (Jan NI 1,700; Feb 880; balance sheets on four dates; a
closed January still reporting 1,700 and a cash flow with one financing
line). Mutation-checked: dropping the closing exclusion fails it (0 vs
1,700). Properties: period P&Ls partition a year at any split, and a
range's net income equals the change in current earnings across it, with
A = L + E on both dates.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
The cash flow statement was computed and exportable as csv/json/xlsx but
missing from every PDF path, so the CPA package carried two of the three
financial statements. Now:
- ledger-import: cash_flow_pdf renders the indirect-method statement
(operating with net income, depreciation add-back, working-capital
lines; investing; financing; net change reconciled to beginning and
ending cash) from the computed CashFlowStatement, so the period and
cash accounts are the caller's.
- books.export gains format "pdf" for profit_and_loss, balance_sheet, and
cash_flow (base64, like xlsx), plus financing_accounts for cash_flow.
- books.export_cpa_package adds cash-flow.pdf (fiscal year to date, like
the P&L) and requires cash_accounts: without them the statement has
nothing to reconcile to.
- One cash_flow_statement helper resolves codes and dates for xlsx, pdf,
and the package.
Test: MCP stdio test exports all three statement PDFs, checks the
cash_accounts and bad-report refusals, and unzips the CPA package to find
all five files. Known report gaps are logged in plan/STATUS.md.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
Posted entries stay immutable (R-3), but their evidence can now change the
same way the ledger does: by appending events, never by editing or
deleting. New R-13a.
- books.unlink_receipt {attachment_id, reason} journals receipt.unlinked
and leaves the receipt filed, unlinked. The receipt policy holds: an
entry that requires a receipt can't lose its last one (structured
receipt_required error), so a replacement is attach-new, then
unlink-old. Drafts and reversals are exempt, as at posting.
- books.withdraw_attachment {attachment_id, reason} journals
attachment.withdrawn for an unlinked receipt or a filed document: hidden
from list_receipts (unless include_withdrawn), never linkable again. The
blob stays in the store, hash-verified and readable.
- link_receipt now refuses an already-linked or withdrawn attachment.
Before, re-linking silently moved a receipt off a posted entry,
bypassing the policy; evidence now leaves an entry only through a
recorded unlink. dispose_asset checks its receipt ids before posting
instead of after.
- Schema v5 adds attachments.withdrawn_at/withdrawn_reason; existing
stores migrate on open and replay rebuilds the projection.
Tests: store test (errors, reopen, rebuild, verify, journal history) and
an MCP stdio test (policy refusal, replace flow, withdraw, listings,
under-threshold unlink). Dry-run and live use on the Wanderland Labs books:
8 migrated .eml attachments swapped for their PDF renders, trial balance
unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MtYEaR6M6Ybdhm6X98BWNu
swaitsandClaude Opus 5.5