mirrors/lightning-bolts
1
0
Fork 0
mirror of https://github.com/lightning/bolts.git synced 2025-01-19 05:33:37 +01:00
Code Issues Projects Releases Wiki Activity

BOLT 2: don't insist we atomically send and commit to disk.

Browse Source 
We can't do that, so allow "write, then send".  That fails on the side of
timing out, rather than having a channel which can't be used.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This commit is contained in:
Rusty Russell 2017-05-24 14:50:56 +09:30
parent 01571c1e58
commit 8424535e1f
1 changed files with 12 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
02-peer-protocol.md
View File

@ -937,11 +937,11 @@ transport.
On disconnection, the funder MUST remember the channel for
reconnection if it has broadcast the funding transaction, otherwise it
MUST NOT.
SHOULD NOT.
On disconnection, the non-funding node MUST remember the channel for
reconnection if it has sent the `funding_signed` message, otherwise
it MUST NOT.
it SHOULD NOT.
On disconnection, a node MUST reverse any uncommitted updates sent by
the other side (ie. all messages beginning with `update_` for which no
@ -999,6 +999,16 @@ channel altogether.
There is similarly no acknowledgment for `closing_signed`, so it
is also retransmitted on reconnection.
A previous draft insisted that the funder "MUST remember ...if it has
broadcast the funding transaction, otherwise it MUST NOT": this was in
fact an impossible requirement, as a node must either first commit to
disk then broadcast the transaction, or the other way around. The new
language reflects this reality: it's surely better to remember a
channel which hasn't been broadcast than forget one which has!
Similarly, for the fundee's `funding_signed` message; better to
remember a channel which never opens (and time out) than let the
funder open it with the funder having forgotten it.
# Authors
FIXME