* mm-controller - dont overwrite existing rpc settings
* ui-networkDropdown - dont pass old network as chainId
* add methods preferencesController.updateRpc and metamaskController.updateAndSetCustomRpc
* use updateAndSetCustomRpc in settings to allow rpcs to be updated
* use new rpc as nickname if no nick name has been supplied
* fix update rpc method
* Refactor and fix styling for first time flow. Remove seed phrase from persisted metamask state
* Fix linting and tests
* Fix translations, initialization notice routing
* Fix drizzle tests
* Fix e2e tests
* Fix integration tests
* Fix styling
* Fix migration naming from 030 to 031
* Open extension in browser when user has not completed onboarding
* Use selector for state.metamask.accounts in all cases.
* Default to cached balance when selecting metamask accounts
* Adds the cached-balances controller
* Documentation and small codes fixes for #5843
Co-Authored-By: danjm <danjm.com@gmail.com>
* Fixed Vivek's bug
Fixes#5850
What was happening:
It seems that his MetaMask had crashed while some new transactions had
been loading defaults. He probably had a network connectivity issue to
Infura (which we are working with Infura to address).
As a result of this network cutout, his three unapproved transactions
were not marked failed, and were not marked as `loadingDefaults =
false`, as their gas prices had not yet been estimated.
Normally this behavior is supposed to clean itself up when the
transaction controller starts up, via the
`TransactionController._onBootCleanUp()` function, but in this case,
during unlock, that function was unable to do its job because when it
requested the transaction list, the current network was in the `loading`
state, making it proceed as if there were no pending transactions.
To fix this, I am doing two things:
- Setting transactions to loadingDefaults = false in more catch blocks.
- Calling `onBootCleanUp()` when the network store's status changes, so
that it will re-trigger when loading completes.
* Fixed reference
* Fixed infinite loop bug
Was refreshing the tx list on every tx state change instead of just
network changes, creating an infinite loop.
* Add notes to tx updates to clarify logs
* Auto fail transactions that have been approved for over 12 hours
Converts txs using a migration.
This migration uses a new helper function that generates tx-failing
migrations, and only requires a version, error message, and condition to
run on each transaction.
* Linted
* Only migrate approved txs to failed
* Cleanup
* Cleanup
* Small lint fixes
* transactions - use safe-event-emitter over events
* tests - pass a platform object on init with a noop showTransactionNotification
* test - fix for tx-state-history-helper trying to reduce an empty array
* deps - safe-event-emitter
* lint
* Add beginning of test
* Resubmit approved transactions on new block
May fix#4343 and related issues, where an error could leave
transactions stranded in the approved state.
* Remove unused test
* Re-approve transactions when retrying approved
* Add retry approved test
* Include approved in pending tx count
* Fix getPendingTxs()
* Linted
* Only throw hash error in submitted state
* Only check submitted txs for block inclusion
* Fix test expectations
Whymarrh Whitby
S.C. Chen
Bruno Barbieri
Frankie
Paul Bouchon
Jenny Pollack
Alexander Tseung
Esteban Miño
Thomas Huang
brunobar79
Dan Miller
Michael - Blurpesec
Dan Finlay
Dan Finlay