When gossip is enabled we waste a lot of time trying to connect
to onion peers if we don't have a proxy enabled. We should just skip
them and try to connect to clearnet peers instead.
LNPeerManager.add_peer would only check if self.network.proxy is set,
which it is always as Network is initialized with self.proxy =
ProxySettings(). Instead it should check if proxy is set and enabled.
Don't set the NETWORK_AUTO_CONNECT config if the user checked the custom
server config checkbox and then cancels the wizard, so the next time the
user starts the wizard they again will have the choice instead of
silently being forced into a random server.
Currently if the user cancels the wizard during the network config the
welcome component will already have set the NETWORK_AUTO_CONNECT config
to False (as the user selected the custom server checkbox), preventing
the wizard from starting again on the next startup.
Passes the default server (Network.default_server) to
Network.set_parameters instead of an empty string in
ServerConnectWizard.do_configure_server to avoid a traceback like in
I am not entirely sure how the user in #10437 managed to trigger
the exception as Network.set_parameters should already exit early at
`not self._was_started` in the Wizard as the network shouldn't have been
started yet during the wizard. However passing the default ServerAddr
instead of an empty string to set_parameters should avoid this
exception.
Same as Qt, validate the server address the user enters and prevent the
user from continuing in the wizard or clicking 'Ok' in the ServerConfig
if the entered address is clearly invalid.
The `ServerWidget` didn't validate the input of server_e, which allowed
the user to enter an invalid server connection string and exit the
dialog without knowing they entered an invalid string.
In the wizard this behavior is very misleading as the user explicitly
wanted to use a custom server, can click next after entering an invalid
server, and we will just silently fall back to automatic server selection.
This change will disable the "Next" button in the wizard if an invalid
address has been entered and show a red background in the server_e while
the input is not valid, so the user clearly sees that this input is not
going to be used.
This script used to work, I assume my breakage is due to the python version.
```
Traceback (most recent call last):
File "/home/user/wspace/electrum/./contrib/add_cosigner", line 35, in <module>
version_spec = importlib.util.spec_from_file_location('version', 'electrum/version.py')
^^^^^^^^^^^^^^
AttributeError: module 'importlib' has no attribute 'util'
```
Moves the logic requesting the forward swap into the TxEditor so it can
use the open transport and doesn't have to reconnect to the relays
again.
Also disables the "Preview" button in the TxEditor when the transaction will
send change to lightning.
This should prevent the user from saving the transaction to history and
broadcasting it later or exporting it and broadcasting it through some
external way.
Broadcasting needs to happen directly after the TxEditor so we can send
the second rpc call to the swapserver and await the incoming htlcs
before broadcasting the (funding-) transaction.
Update the TxEditor (onchain tab) if Send change to lightning is enabled
and the swap transport changes. Connect to swap transport if send change
to lightning gets enabled or if it is enabled and the TxEditor gets
opened.
This allows to nicely show the swap fees without blocking the UI to wait
until the swap manager gets initialized.
Check the swap providers liquidity as well if we try to send change to
lightning in `make_unsigned_transaction`. It is now expected that the
swap_manager is already initialized when calling
`make_unsigned_transaction`, otherwise no dummy output will get added.
Separates the swap transport initialization logic from the submarine
payment tab logic so it can be used for the send change to lightning
functionality too.
Also makes the gui updates on transport establishment more thread safe
by using pyqtSignals instead of calling gui methods from the asyncio
thread.
This patch changes the CallbackManager to use WeakMethods (weakrefs) to
break the ref cycle and allow the GC to clean up the wallet objects.
unregister_callbacks() will also get called automatically, from
EventListener.__del__, to clean up the CallbackManager.
I also added a few unit tests for this.
fixes https://github.com/spesmilo/electrum/issues/10427
-----
original problem:
In many subclasses of `EventListener`, such as `Abstract_Wallet`, `LNWatcher`,
`LNPeerManager`, we call `register_callbacks()` in `__init__`.
`unregister_callbacks()` is usually called in the `stop()` method.
Example - consider the wallet object:
- `Abstract_Wallet.__init__()` calls `register_callbacks()`
- there is a `start_network()` method
- there is a `stop()` method, which calls `unregister_callbacks()`
- typically the wallet API user only calls `stop()` if they also called
`start_network()`.
This means the callbacks are often left registered, leading to the wallet
objects not getting GC-ed. The GC won't clean them up as
`util.callback_mgr.callbacks` stores strong refs to instance methods
of `Abstract_Wallet`, hence strong refs to the `Abstract_Wallet` objects.
An annoying example is `daemon.check_password_for_directory`, which
potentially creates wallet objects for all wallet files in the datadir.
It simply constructs the wallets, does not call `start_network()` and
neither does it call `stop()`.
Prevents `DeviceMgr.run()` from blocking the `Plugins` `DaemonThread` by
scheduling the hww timeout check instead of awaiting its result on the
`Plugins` thread.
If something in the `_hwd_comms_executor` thread is waiting for user
input, e.g. when setting up a hww in the wizard the user needs to
unlock the hww for `HardwareClientBase.get_xpub()` to return, the
`_hwd_comms_executor` is blocked. If then `DeviceMgr.run()` gets called by
the `Plugins` `DaemonThread` concurrently and tries to check the hww
timeout on the `_hwd_comms_executor` as well the `DaemonThread` is
blocked too until the `_hwd_comms_executor` gets unblocked (and the
`DaemonThread.job_lock` is taken.
Now if something tries to take the `DaemonThread.job_lock` it blocks as
well, so if a user e.g. tries to load a new plugin from the plugins
dialog the whole gui thread will freeze until the hww gets unlocked.
Disables the "Always allow Screenshots" switch in the preferences if the
App isn't running on Android. QML doesn't have screenshot protection
outside of Android so this toggle is misleading.
- to gracefully take duplicate calls of register_callbacks(): should be idempotent now
- as a side-effect, the order of the callbacks is changed and not guaranteed
- not like anyone should have been relying on it before though
