This loads up 20MB of plugins temporarily; we seem to be getting OOM
killed under CI and I wonder if this is contributing.
Doesn't significantly reduce runtime here, but I have lots of memory.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
When downloading a python package from the PyPI repository the links
where pointing to a non-existent parent directory, thus breaking the
packages. The files don't ever change, and are really simple, so let's
just materialize them.
This was measured as a 95th percentile in our rough testing, thanks to
all the volunteers who monitored my channels.
Fixes: #4761
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: JSON-RPC: `setchannelfee` gives a grace period (`enforcedelay`) before rejecting old-fee payments: default 10 minutes.
This is best-practice (to ensure prototypes match up), but there were a
few places we didn't (at least, directly). Make it a requirement,
either of form "foo.h" or <dir/foo.h>.
The noise is the change to our print templates.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
On my machine it should have produced:
/tmp/ltests-1p0v5z1j/.locks/lock-1631090663.1066182
but instead it produced:
/private/tmp/ltests-1p0v5z1j/.locks/lock-1631090663.105848
The mismatch resulted in a hang inside flock.
Path .resolve() seems to be causing problems for others as well. It appears the library was not meant to handle complex path situations and isn't maintained as such (see link reference).
Since we have already built a full path here anyway, the call to .resolve() appears redundant.
Tested on python 3.9.6 on my Mac OS X 11.4 (20F71), Xcode 12.5.1 (12E507)
Relevant discussion: https://discuss.python.org/t/pathlib-absolute-vs-resolve/2573
This affects the range we offer even without quick-close, but it's
more critical for quick-close.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: JSONRPC: `close` now takes a `feerange` parameter to set min/max fee rates for mutual close.
This only happens when a deletion is added by a running gossipd, so
we put a deletion at the end of the store to test it.
mypy noticed that this code was nonsensical, so clearly untested.
The testing noticed that making a nodeid from a string was also buggy.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Mainly fixing type annotations, but some real fixes:
1. GossmapHalfchannel.from_str() should be a classmethod.
2. update_channel had weird, unusable default values (fields can't be NULL,
since we use it below).
[ There was one more occurence where isinstance should be used above
type() == xyz comparison. -- MS ]
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Do not mix bytes and GossmapNodeId when accessing Gossmap.nodes dicts.
Therefore the definion got GossmapNodeId also needed to be pulled to the
beginning of the file.
This reads the `gossip_store_channel_amount` that always follows the
`channel_announcement` messages. Therefore it uses an internal variable
_last_scid to know what channel has been added last time.
This is more efficient than converting them all to Pubkeys: about 3.8
seconds vs 5.4 seconds. Usually treating them as raw bytes is what we
want anyway.
[ Fixup by Michael Schmoock <michael@schmoock.net> ]
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It doesn't do much work, but it does parse the gossmap file and extract
nodes and channels.
[ Fixup by Michael Schmoock <michael@schmoock.net> ]
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
nicolas.dorier
Andrew Toth
Rusty Russell
Jonas Nick
Christian Decker
Michael Folkson
Vincenzo Palazzo
Dustin Dettmer
Michael Schmoock