Update bech32 dependency#2117
Conversation
Update the `bech32` dependency to use the newly release beta version. The main fix here is silent, a bug fix in `bech32` that was being hit by our fuzzing suite.
b630cc9 to
33ee49f
Compare
|
Our two ack policy sucks, this is a trivial change and is needed to unblock CI for other PRs but does not qualify for either of the two carve-outs we have. Does the two ack policy really improve the quality of code merged when it pretty much just pressures devs to review - that does not seem to me to be in the spirit of the whole thing. I rekon we should have more benevolent dictator for life and less force-devs-to-ack-to-make-it-seem-like-we-did-deep-review. |
I agree, I have some experience with GNU projects where there is a very nice policy of trivial commits. This basically allows you to open a patch and merge it when there is a simple change. It is also safe to do because whoever can merge a patch is also a trusted person (has right to commit), so it should be fine Open a PR will notify all the listeners to changes about the change, and for this reason is better than push on master directly Happy to chat in another place (maybe an issue?) about this |
|
@tcharding I think "changing from one prerelease to another" qualifies for the "minor CI fixes" carveout. But sure, if you think in principle it doesn't, then I'll just pull rank as BDFL and merge this. But having said this, you need to update the Cargo-*.lock files before I can pass this :). |
I don't want to be seen to be pushing my own changes through so I'm being really careful to only call carve-out if it is 100% covered. I think you should be able to BDFL things in much more fuzzily. |
Yep I forgot to do it yesterday, saw your message on another PR and did it this morning, must have been in flight when you commented. |
Ok, that's fair, good for you to be fairly conservative when claiming "carve-out". But I think this one qualifies as a CI fix :). |
I think the two-ack policy would be fine if it's easy to get people who are maintainers to review and ack. The problem as I see it is there isn't a lot of people who are maintainers that are also active reviewers. Personally, I feel more safe when more people review my changes, but for CI changes I get that it maybe doesn't need as much scrutiny.
Well if we had more benevolent dictators for life, would we need things like bitcoin? :) |
Update the
bech32dependency to use the newly release beta version.The main fix here is silent, a bug fix in
bech32that was being hit by our fuzzing suite.