When postponing the delivery time of an already signed/sent will, the user
was asked to sign the on-chain invalidation transaction twice before the new
(postponed) will could be built.
Root cause: after the invalidation tx was broadcast, on_success_invalidate
restarted task_phase1 to rebuild the will, but the old will items were still
marked COMPLETE/PUSHED with their original tx.locktime (the on-chain
invalidation did not update the in-memory status). The postpone check therefore
fired WillPostponedException a second time, requesting another invalidation.
Fix:
- Add Will.mark_invalidated_by_tx(will, tx): marks INVALIDATED every valid will
item that spends a prevout consumed by the just-broadcast invalidation tx.
Setting INVALIDATED clears the VALID flag, removing those items from
only_valid_list so the postpone/expire check no longer fires.
- Call it from loop_broadcast_invalidating after a successful broadcast (txid
obtained) and persist via save_willitems. On the phase-1 restart the old will
is no longer VALID, so the will is rebuilt directly: a single invalidation
signature followed by the new will.
Tests: add test_will_mark_invalidated_by_tx and
test_will_mark_invalidated_by_tx_no_match plus the WillPostponedException
hierarchy assertion. 184 tests pass; smoke and external-zip OK; ruff clean.
Bump version to 0.3.1.
Bump plugin version from 0.2.8 to 0.3.0 in VERSION, manifest.json,
__init__.py and core/plugin_base.py for the 0.3.0 release (postpone
invalidation safety + Server status column).
Postpone safety (Strategy B):
- A signed/sent will carries an immutable locktime; postponing the delivery
time previously did nothing, so a will-executor could still broadcast the
old (earlier-locktime) transaction and execute the inheritance too early.
- core/will.py: add WillPostponedException and detect postpone by comparing
the requested locktime against w.tx.locktime (the locktime frozen in the
signed transaction) instead of the in-memory heir entry, which is updated
together with the new value and would always compare equal.
- gui/qt/dialogs.py (BalBuildWillDialog.task_phase1, the real path used by
Tools -> Prepare): handle WillPostponedException before NotCompleteWill;
return (None, tx) to trigger sign + broadcast of the invalidation, then the
user presses Prepare again to rebuild/re-sign/re-send (two explicit steps).
- gui/qt/window.py: mirror the branch in build_inheritance_transaction with an
explanatory message; wording aligned to the 'Prepare' button.
- gui/qt/common.py: export WillPostponedException.
- A postpone on a will that was never signed/sent just rebuilds (no on-chain
fee).
Server status column:
- gui/qt/lists.py: add a dedicated 'Server' column to PreviewList with an
always-readable label and a tooltip (will-executor URL + state).
- gui/qt/theme.py: add server_status_text() and server_status_tooltip(),
reusing the existing status flags.
- gui/qt/common.py: export the new theme helpers.
Docs: update README.md, bal/README.md and CHANGELOG_REFACTOR.md.
Tests: 182 passed; smoke + external-zip OK; ruff has no new real findings.
GUI / plugin lifecycle:
- Auto-close Electrum's native 'Electrum Plugins' manager dialog after the
BAL plugin is hot-enabled. Electrum 4.7.x no longer calls the old init_qt
hook, so the close is now triggered from the create_status_bar, init_menubar
and load_wallet hooks (fired when reload_windows() recreates the window).
- Robust dialog matching (isinstance / class name / localized window title) to
cope with zipimport module-identity mismatches.
- Robust dismissal of the modal dialog (reject()/done()/close()) with a retry
schedule [400, 800, 1500] ms; if it still cannot be closed, fall back to
bringing it to the front (showNormal/raise_/activateWindow) so it never
lingers hidden in the background. Counting only visible top-levels avoids
treating an already-closed dialog as still open.
Read-only field styling:
- Paint the locked Delivery time / Check Alive date editors and the mining-fee
spinbox with a light-grey background (#f0f0f0) so the user can see at a glance
that they are not editable outside the 'Build your will' wizard; the styling
is cleared when the fields are made editable again.
Pickle/RLock crash on 'Build will':
- heirs.save() now sanitises the heirs mapping via _json_safe() before handing
it to json_db.put(), which deep-copies the value. A live runtime object
(holding a threading.RLock) slipping into an heir value previously raised
'TypeError: cannot pickle _thread.RLock object' and aborted the task; such
values are now coerced to str and logged with their path.
- init_heirs_to_locktime() coerces the locktime to a plain serializable scalar.
- log_error() now accepts both a sys.exc_info() triple and a single exception
instance, fixing the secondary 'TypeError object is not subscriptable' that
masked the real error.
Networking (anti-freeze):
- Add check_transactions_parallel: pressing "Check" now contacts will-executors
concurrently with a fast-fail timeout (CHECK_TIMEOUT=8, 1 retry) and a global
deadline (CHECK_GLOBAL_DEADLINE=30), so a single dead server no longer freezes
the "checking transaction" dialog for ~140s (old default 10s x 10 retries).
- check_transaction now accepts timeout/max_retries/retry_sleep kwargs.
- Rewrite BalWindow.check_transactions_task to use the parallel helper with live
progress + an elapsed-time counter "Checking transactions: 2/5 (4s / 30s)".
Reliable elapsed-time counters (Xs / DEADLINEs):
- Replace the unreliable raw heartbeat thread (whose pyqtSignal emissions were
not marshalled and never repainted) with an on_tick callback driven from the
CALLING thread in push/ping/check parallel helpers.
- Show the maximum wait too (e.g. "3s / 30s") so the user knows when the
operation will give up, in the wizard Broadcasting, Ping and Check dialogs.
- Expose networking constants (PUSH_*/CHECK_*/DEFAULT_TIMEOUT) as Willexecutors
class attributes for a single source of truth in the GUI.
GUI:
- Add hover tooltips: Wizard ("Wizard - Build your will"), Delivery time (truck),
Check Alive (siren), Calendar, Check (refresh).
- Reorder the Will toolbar to: Wizard | Delivery time | Check Alive | Calendar |
Check; tighten layout margins so it all fits the Will window.
Tests:
- parallel_ping_test.py: add coverage for check_transactions_parallel (parallel
timing, global deadline, on_tick from the calling thread) and static checks
that check_transactions_task/loop_push use the parallel helpers + on_tick
counter with the 'Xs / Ns' format.
Verified: ruff (0 new issues), 182 official tests pass, parallel/smoke/gui_fixes/
windows_overflow/external_zip tests pass.
Problema: i server Will-Executor venivano contattati in sequenza e, su
timeout, send_request riprovava 10x con sleep 3s (~130s per server morto).
Un solo server irraggiungibile bloccava l'intera operazione (impallamento UI).
Soluzione:
- ThreadPoolExecutor: ping e push ora in parallelo (tempo ~= server piu' lento,
non la somma). Un server morto non blocca piu' gli altri.
- Fast-fail per operazioni interattive (ping/info/download): max_retries=0,
niente retry-storm.
- Feedback live: callback on_each() aggiorna il dialog server-per-server
(thread-safe via pyqtSignal di BalWaitingDialog.update).
- Push transazioni: parallelo ma con retry per-server mantenuti (no perdita tx).
File:
- bal/core/willexecutors.py: send_request(+max_retries,+retry_sleep),
get_info_task fast-fail, NEW ping_servers_parallel(), push_transactions_parallel(),
DEFAULT_TIMEOUT=5.
- bal/gui/qt/window.py: ping_willexecutors_task + push_transactions_to_willexecutors
riscritti su helper paralleli con feedback live; fetch_will_executors_list fast-fail.
- bal/core/util.py: BUGFIX get_value_amount usava in_output (bool) invece di
din_output (tupla) -> TypeError. Scoperto dai test ufficiali Gitea.
Test (contro il codice refactor):
- pytest tests/ ufficiali: 117 core + 65 gui = 182 passed.
- smoke/external_zip/windows_overflow/gui_fixes: OK.
- parallel_ping_test (nuovo): 0.50s per 8 server vs ~4.00s sequenziale.
- ruff: nessun nuovo problema introdotto (codice nuovo PEP8-compliant).
Aggiunti i test ufficiali del repo Gitea + REPORT_NETWORKING_PARALLELO.md.
* fix(gui): voci di menu BAL duplicate/condensate dopo riavvio o cambio wallet
Sintomo (Windows 11): dopo aver riavviato Electrum o cambiato wallet, le
schede Will/Heirs sparivano dalla tab bar e dal menu, e compariva una voce
di menu condensata/illeggibile (icona + testo sovrapposti) sotto il logo di
Electrum, accanto a 'Portafogli'.
Causa: init_menubar_tools veniva eseguito DUE volte sulla stessa finestra.
Con il plugin gia abilitato, al riavvio Electrum invoca sia l'hook
init_menubar sia il percorso di init a caldo (init_qt -> _setup_window),
entrambi chiamano init_menubar_tools -> addTab/addAction duplicati.
Nell'originale init_qt faceva return (chiedendo il riavvio) e quindi i menu
venivano creati una sola volta; rimuovendo quel return (fix B3) e' emersa la
doppia inizializzazione.
Fix:
- BalWindow._menubar_initialized: guardia di idempotenza.
- init_menubar_tools: se gia inizializzato, esce subito (niente duplicati).
- on_close: resetta il flag dopo aver rimosso tab/azioni, cosi la stessa
finestra puo essere riusata per un altro wallet.
- tests/gui_fixes_test.py: regressione che verifica la guardia in __init__,
init_menubar_tools e on_close.
Logica di business invariata (nessuna modifica a bal/core/*).
* fix(gui): ripristina create_status_bar come no-op (come originale)
L'elemento di menu condensato/illeggibile sotto il logo di Electrum era
causato dal StatusBarButton aggiunto da create_status_bar.
Nell'originale Gitea questo hook aveva un 'return' subito dopo il log, PRIMA
di costruire il bottone: era quindi disabilitato di proposito. Durante la
pulizia del 'dead code' nel refactoring quel return era stato rimosso,
riattivando la creazione del bottone -> elemento icona+testo renderizzato
nel punto sbagliato dopo riavvio/cambio wallet.
Fix: create_status_bar torna a essere un no-op (return), fedele all'originale.
Le impostazioni restano raggiungibili da Strumenti -> Plugin.
Regressione: gui_fixes_test verifica che create_status_bar non chiami
addPermanentWidget.
* fix(core): OverflowError su Windows (anno 2038) che rompeva tab/menu BAL
CAUSA VERA (dal log Electrum dell'utente, Windows 11):
OverflowError: Python int too large to convert to C int
window.py __init__ -> create_heirs_tab -> WillSettingsWidget
-> on_locktime_change -> BalTimestamp.to_date
-> datetime.fromtimestamp(NLOCKTIME_MAX)
Su Windows time_t e' a 32 bit, quindi datetime.fromtimestamp() solleva
OverflowError per qualsiasi timestamp oltre il 2038 (es. NLOCKTIME_MAX =
2**32-1 = 4294967295, usato come locktime di default/sentinella). Su Linux
64-bit la stessa chiamata funziona: per questo il bug si vedeva solo su
Windows e i test su Linux non lo intercettavano.
L'eccezione interrompeva BalWindow.__init__ durante init_menubar/load_wallet,
lasciando le schede Will/Heirs e la voce di menu a meta' costruzione ->
l'elemento grafico condensato/illeggibile sotto il logo di Electrum.
FIX (comportamento invariato per tutti i valori normali):
- BalTimestamp._safe_fromtimestamp(): datetime.fromtimestamp con clamp a
INT32_MAX in caso di OverflowError/OSError/ValueError, esattamente come la
funzione get_max_allowed_timestamp() dell'originale (Electrum issue #6170).
- Usato in to_date / to_timestamp / __str__ / __repr__ di BalTimestamp.
- widgets.py set_value: usa il converter sicuro.
- util.py timestamp_minus: stessa protezione inline con clamp a INT32_MAX.
I valori entro il 2038 (date assolute normali, durate relative come 90d/5y)
producono lo stesso identico risultato di prima.
TEST: tests/windows_overflow_test.py riproduce il limite 32-bit di Windows
(monkeypatch di datetime.fromtimestamp) e dimostra che senza il fix si ottiene
lo stesso OverflowError del log, mentre col fix passa. Verificato anche che il
test FALLISCE senza il fix.
* docs(it): documenta il fix OverflowError Windows (anno 2038) nel changelog
Aggiunge la sezione §13 al CHANGELOG_REFACTOR.md che descrive:
- sintomo (schede/menu rotti su Windows dopo riavvio/cambio wallet)
- causa vera dal log (datetime.fromtimestamp(NLOCKTIME_MAX) -> OverflowError
su time_t 32-bit di Windows)
- fix con _safe_fromtimestamp (clamp a INT32_MAX, come Electrum #6170)
- test di regressione windows_overflow_test.py
Aggiornata anche la cronologia (§12) con PR #3.
---------
Co-authored-by: GenSpark AI Developer <ai@genspark.dev>
