Dominik Charousset

Passing `0.0.0.0` from Broker would be a trivial fix. However, as pointed out in https://github.com/zeek-packages/zeek-agent-v2/pull/9, that's forcing IPv4. The issue is the function that Broker calls for creating that socket....

> Curious. I wonder what version of `caf` they're expecting then. Zeek no longer supports external CAF versions. From a user's perspective, CAF is an inaccessible, internal implementation detail. The...

Not for now. Eventually, the private fork probably becomes obsolete and we can maybe at some point consider following the official CAF release cycle again. However, Zeek actually bundles quite...

Thanks for the inputs here, I'll try and reproduce this.

Thanks for sharing your thoughts! I think named handler would be a great to have. I'm skeptical whether putting the names into the handle types would be practical, though. Would...

Thanks for the ticket. > This suggests the fix might be easy: https://issues.guix.gnu.org/issue/57849#4 Seems like it isn't that easy. From the same thread: > -bundle_loader requires -bundle. Ld gives an...

Thanks for doing my homework and not only file an issue but provide a fix straightaway! 😄 I've found one more instance where `deque` makes more sense than `vector`. Mind...

> Do you think there might be any negative side effects of this change? I don't think so. `deque` could end up making more memory allocations than a `vector`. However,...

Btw, the last category I'd like to integrate are some of the `readability-*` checks. So there's only one more `clang-tidy` PR in the pipeline. 🙂

No strong opinion on the change (doesn't matter for Broker itself), but changing this to `unordered_map` is a breaking API change. Whether there are actually any clients that rely on...