We never really supported pypy - the CI was running the tests there
just to see compatibility.
A few of the unit tests consistently hang, resulting in the whole
test runner timing out and wasting compute resources. Someone could
debug the hangs and try to fix them, but I don't have the time.
ChannelDB.load_data() takes ~15 seconds. Previously if the user tried
to close the program while load_data is running, we would block until
load_data() finished. (e.g. consider starting and immediately stopping
Electrum)
Now instead we can abort load_data early.
Electrum clients that have non-trampoline channels.
- if an invoice supports trampoline, but provides only
non-trampoline nodes in its routing hints, we must use
a legacy trampoline payment
- when we create a trampoline onion for a legacy payment,
the invoice_features field is a u64, so we need to remove
high feature bits.
Plugins.stop() is now part of the normal shutdown flow (since 90f39bce88),
so this shaves up to 100 ms off that (avg: 50 ms).
It is also waited on in around ~60 unit tests: this saves 3 sec.
In python 3.8 and 3.9, asyncio.Event/Lock/etc cannot be created before the
event loop itself is created. Hence, to have Plugins.__init__ create an
Event, we need to postpone creating Plugins() from setUpClass to setUp.
follow-up 90f39bce88
Plugins.stop was never called, so the Plugins thread only stopped
because of the is_running() check in run(), which triggers too late:
the Plugins thread was stopping after the main thread stopped.
E.g. playing around in the qt wizard with wallet creation for a Trezor,
and closing the wizard (only window):
``` 24.85 | E | p/plugin.Plugins |
Traceback (most recent call last):
File "/home/user/wspace/electrum/electrum/util.py", line 386, in run_jobs
job.run()
File "/home/user/wspace/electrum/electrum/plugin.py", line 430, in run
client.timeout(cutoff)
File "/home/user/wspace/electrum/electrum/plugin.py", line 363, in wrapper
return run_in_hwd_thread(partial(func, *args, **kwargs))
File "/home/user/wspace/electrum/electrum/plugin.py", line 355, in run_in_hwd_thread
fut = _hwd_comms_executor.submit(func)
File "/usr/lib/python3.10/concurrent/futures/thread.py", line 167, in submit
raise RuntimeError('cannot schedule new futures after shutdown')
RuntimeError: cannot schedule new futures after shutdown
```
Note: We should probably require the password in load_wallet,
and store it in memory as long as the wallet is loaded. In that
case, commands that currently require a password would no longer
require it if the wallet is loaded.
- when creating an invoice, if we had multiple chans with the same trampoline,
were putting duplicate t-hints into the invoice
- when paying an invoice that had duplicate t-hints, we would sometimes
construct invalid paths (SRC>T1->T1->DST)
follow-up https://github.com/spesmilo/electrum/pull/8536
This replaces 69336befee, which was insufficient.
#8536 added a new field into the struct, which older versions do not ignore but raise:
opening a wallet file with new code updated the struct to include it,
after which old code could no longer open the wallet file.
i.e. #8536 was an invisible wallet upgrade, breaking compat.
This commit simply formalises the wallet upgrade: old code will now show
an understandable error when trying to open new files.
The input_qr_from_screenshot functionality is broken on some Linux machines:
on some machines, `screen.grabWindow(0)` returns an all-black image.
In such cases, instead of telling the user "No QR code was found on the screen",
we will tell them "Failed to take screenshot".
To test:
`QApplication.instance().primaryScreen().grabWindow(0).save("/home/user/wspace/tmp/pic2.png", "png")`
Tested on:
- machine 1:
- ubuntu 22.04, gnome, wayland
- pyqt.version: 5.15.9
- qt.version: 5.15.2
=> gets all-black image for screenshot
- machine 2:
- manjaro, kde, x11
- pyqt.version: 5.15.9
- qt.version: 5.15.8
=> screenshot works
I guess it might be due to x11 vs wayland.
eclair sends CHANNEL_DISABLED if its peer is offline. E.g. we might be
trying to pay a mobile phone with the app closed. In that case we
should not cache the CHANNEL_DISABLED for too long.
LnFeatures.supports() is became part of the hot path of
LNPathFinder.find_path_for_payment()
in 6b43eac6fd,
which made find_path_for_payment considerably slower than before.
It used to take around 0.8 sec originally, then after linked commit went to ~9 sec,
and now this takes it down to ~1.1 sec (on my laptop).
otherwise I can't click on list items if the window is moved too far to the right :O
follow-up 583afefe33
Note: on modern Android, apps don't always run full-screen.
You can pop them out into small windows and move them.
Haven't tested how the deadzones work then though.
SomberNight
ThomasV