-
Notifications
You must be signed in to change notification settings - Fork 174
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Set default tx_broadcast = random-peer #895
Conversation
This method is slightly more private with no other downsides
Anyone have any objections to this? If not I'll merge it. It has no downsides only upsides. |
Sorry not much time for looking at stuff now but: if the peer doesn't broadcast, am I right in thinking we just timeout and restart the tx instead of broadcasting it ourself? I think my earlier thinking was to stick with self default for this reason, but I don't remember in detail and would have to read the code again. |
I think fallbacking to broadcasting yourself is better than restarting, as whole coinjoins process may take time (minutes), are not guaranteed to succeed (sometimes happen that less than minimum |
I strongly agree, all the more so because we have a "no-self" option (although I think it might not be implemented right now). Clearly timing out and restarting just because you chose a peer who didn't bother to broadcast is a bit shitty, which is why I'm mentioning it. |
@AdamISZ joinmarket-clientserver/jmclient/jmclient/taker.py Lines 824 to 848 in b25d03c
I think we already fallback to broadcast ourselves, right? |
@PulpCattel Thanks for checking it out for me! Imagine remembering any of the code you wrote ... Someone might want to investigate how this interacts with how often the stall monitor wakes up, but I guess it's probably ~ correct. |
So based on: the code already addresses my concern above, ACK to this PR. |
In hindsight I should've elaborated on the fallback, because I also examined that code which @PulpCattel linked before opening this PR |
One is the actual commit, the other is the merge which added that commit to the master branch |
@chris-belcher the merge of this PR broke one test in the test suite, specifically:
|
@chris-belcher what's confusing? |
I'm not confused, it's just the closest thing to a sad or regret emoje. |
@AdamISZ @chris-belcher Any plans to fix that test failure? |
It's fixed in #910 as a side-effect. |
This method is slightly more private with no other downsides I think