Editing a recurring transaction to end earlier kept its unpaid occurrences
after the new end date, counting them in projected balances, and left
last_generated_date past it. Remove them (both sides of a generated transfer;
a partner linked from elsewhere is unlinked and kept) and move the bookmark
back to the last remaining occurrence. Paid occurrences stay.
Also pin ENABLE_SOFT_DELETE=False on the test that asserts hard deletes, so a
.env that turns soft delete on no longer fails it.
The !Type header line was buffered into the first transaction's raw
lines, so a transaction hashed differently depending on whether it
opened the export. An overlapping export with a different first
transaction imported it twice. Header lines are now left out of the
internal ID. The previous-hash check still hashes the full lines,
headers included, and runs on the first appearance of each of those,
so a first transaction and an identical later one imported before
this change are both still recognized.
Account blocks (!Account and !Option:AutoSwitch account lists) and
non-transaction sections (!Type:Cat, !Type:Class, !Type:Memorized and
any type not in QIF_TRANSACTION_SECTIONS) were read with the
transaction codes, where T is an account's type and D a description,
so files containing them failed to import. They are now skipped, and
each skipped section is logged.
Presets kept dates as typed, such as "Novembro 2026" or "14/11/2026",
so changing the language or date format made them unreadable and they
were silently dropped. Dates are now stored as ISO and shown in the
user's current format when the preset is applied. Presets saved before
this keep working while the language and format stay the same.
A preset whose dates run from the 1st to the last day of a month moved
by days, so a "this month" range drifted into the next month. Such a
range now moves by months, ending on the last day of its new month.
The relative dates option was shown below the preset name with CSS order
but came before it in the page, so Tab reached it first. It is now
inserted right after the name input when the dialog opens.
An unpaid occurrence linked to a transaction outside its recurrence (e.g. through a DCA entry) could end up as two expenses, or both sides in one account, after the recurrence's type or account changed. The link is now dropped in that case; the partner is left untouched.
Editing a recurring transaction only refreshed the amount, description,
category, notes, tags and entities of its unpaid transactions. Changing its
account or type left the already generated transactions behind: a monthly
expense moved to another account kept its upcoming months in the old one,
and an expense turned into an income kept pulling projected balances down.
New occurrences followed the new settings, so the recurrence ended up split
across two accounts.
Unpaid transactions now take the recurrence's account and type as well,
keeping their dates, the same way installment plans already update theirs.
For recurring transfers:
- the "to" sides follow the "to" account and amount;
- turning a recurrence into a transfer adds the missing "to" sides;
- turning a transfer into an income or expense removes the unpaid "to"
sides and leaves the paid transfers linked.
A "to" side is now recognised by its partner belonging to the same
recurrence, so an occurrence the user linked by hand to another
transaction is updated like any other occurrence instead of being treated
as a generated "to" side.
"Only editable by the owner" read as covering an account's transactions
too, which led to GHSA-36w2-769w-h62g being reported. Sharing an account
or making it public deliberately opens its ledger to everyone who can
see it; only the object itself stays owner-only.
State that in the generic visibility help text, and spell out the
account case on the account share panel.
InstallmentPlan and RecurringTransaction declare the unfiltered
all_objects manager first and set no default_manager_name, so
get_object_or_404(Model, id=...) bypassed account visibility. Any
logged-in user could read, refresh, pause, finish or delete another
user's plans (and their transactions) by id.
Resolve the nine per-object views through the filtered objects manager,
as the list views, the API and the transaction views already do.
Refs GHSA-vcr4-g82g-wcq8
Adds a Sub-transactions field to edit actions and update-or-create
actions. Expressions get a sub_transactions variable and return a list of
{name, amount}; an empty list removes the parts. Rules skip transfers, and
setting the amount of a split transaction is skipped.
Transactions, installment plans and recurring transactions take a nested
sub_transactions list; their amount becomes the parts' sum. Changing only
the amount of a split transaction, or adding parts to a transfer, is a
validation error.
Splits a transaction into named parts whose amounts add up to its amount,
e.g. the items of a big purchase or the fees inside an investment
contribution. Parts can be set on transactions, installment plans,
recurring and quick transactions, and are copied to the transactions they
generate. Transfers can't have parts.
Transactions show their parts under a tab on the card and in the
investment detail. The content filter matches part names, and bulk edits
leave the amount of split transactions alone.
Describe the API token authentication, name the colliding transaction
type enums, and use nullable instead of an OpenAPI 3.1 null type. A test
keeps the schema free of warnings.
Rule dry runs call tasks directly inside an atomic block they roll back.
The connection cleanup wrapper closed that connection, so every dry run
failed with a 500.
Adds Convert > Transfer to the transaction menu. The transfer form opens
prefilled on the side matching the transaction type (expense = from,
income = to); saving creates the transfer and deletes the original.
Not available for transfers or plan/recurrence transactions.
- Split the selection control into an unselect button and a dropdown
- Expose the new bulk actions in the edit dropdown
- Fold the actions into a single menu and hide the sum on small screens
Split the filter into Include and Exclude tabs. Exclusions for content, accounts, currencies, categories, tags and entities are applied together with the existing filters, and are kept by presets, Clear and the active-filter indicator.
A history restore swaps in htmx's snapshot of the body, which carries a
lifeless copy of AirDatepicker's global container. The previous fix looked
the container up by id, found that copy, and left the library's real
container detached, so pickers opened into nothing.
Track the container through datepicker.$container instead, remove any
snapshot copies after a swap, and reattach the real one when detached.
The filter on the Monthly and All pages is wrapped in hx-preserve, so a
boosted navigation keeps its pickers instead of recreating them. The
AirDatepicker global container they render into lives on the body and is
swapped out with it, leaving those pickers drawing into a detached node.
After each swap, move the detached container's pickers into the new
container, or reattach it when no new picker created one.
Filter presets save the exact dates they were created with, so a preset
for "from today" or "this month" goes stale the next day or month.
The save dialog now offers "Keep dates relative to today". When ticked,
the preset remembers the day it was saved (`FilterPreset.relative_to`)
and, when applied, moves its dates forward by the time elapsed since:
"Date from"/"Until" by days and the reference months by months. A preset
saved with today as the start date always starts today, and one saved
for the current reference month always shows the current month.
- The option is only stored when the preset has a date filter.
- Presets with relative dates show a clock icon in the presets menu.
- Presets saved before this change, or without the option, keep their
fixed dates.
QIF imports deduplicate each transaction by storing a SHA-256 of its raw
lines as the transaction's internal_id and skipping rows whose hash is
already taken. The hash didn't include the account, which broke two
cases:
- Importing the same file into two accounts (possible since the account
can be picked per run) skipped every row of the second import.
- Two real transactions with the same date, amount and payee in one file
hashed to the same ID, so the second one was always dropped.
The internal ID now hashes the account's ID and the transaction's
occurrence (how many identical transactions came before it in the file)
together with the raw lines. Using the occurrence instead of the
position in the file keeps the IDs stable when a later export of the
same statement has more transactions before or between the repeated
ones.
Transactions imported before this change keep their old IDs. So that
re-importing an old file doesn't duplicate them, the first occurrence
of a transaction is also matched against the previous hash, limited to
the account being imported into. Repeats of identical transactions,
which the old hash always skipped, are imported now.
The duplicate check also looks at soft-deleted transactions, like the
CSV/Excel deduplication does. internal_id is unique, so a soft-deleted
transaction with the same ID used to fail the import with an integrity
error instead of being skipped.
Fixes#671
Some banks name the account once in a header cell above the table, which
skip_lines and start_row throw away. A `line` transformation returns a line
of the file (or one of its columns) as text, so a mapping can read it and
clean it up with regex or replace.
CSV keeps only the requested lines during the existing counting pass; Excel
reads them straight from each sheet. An account picked on the run still
overrides the mapped one.