Arne Welzel
Arne Welzel
...and deprecate it. The communication_enabled flag doesn't really make a difference in how the IO loop works.
While looking at the logging API/code: When converting from `RecordVal` to the `threading::Value**` construct, there are individual `new threading::Value(...)` calls for every field. Instead, it's also possible to allocate the...
An external user reported Zeek clusters running 7.0.0 using only built-in analyzers, running out of memory. They found via jeprof memory profiling that `spicy::rt::ParsingState::_process` was the most likely location for...
The idea is to provide an event from Spicy to Zeek that reports all available analyzers, including their meta data. We can then provide a standard handler that registers them...
The following PCAP contains two DNS connections with opcode "UPDATE (5)". We purposely don't log these in dns.log. Instead of ignoring, however, the following weirds are produced and they aren't...
This stops invoking Telemetry::sync() via a scheduled event and instead only invokes it on-demand. This makes metric collection network time independent and lazier, too. With Prometheus scrape requests being processed...
I'm splitting this from #3597. Prototyping ZeroMQ and NATS cluster backends, a `Cluster::publish(worker_topic, ...)` on a worker node is guaranteed to reach all online nodes subscribing to a `worker_topic` (how...
We've announced ZAM for broader use with 7.0, but reality is it's not tested by us regularly. Running `btest -a zam` with an ASAN build currently reports memory leaks for...
As reported in the Zeek repo, a recent bump of Spicy regressed performance of pcap-quic-16-50mb. https://github.com/zeek/zeek/pull/3953#issuecomment-2431443070 I bisected the range of the bump and it is: ca5de306a7c397c1d37aa7eb662dabafaa8ca26b ``` $ git...
Running the following code through `spicyc -d -j` triggers an assertion. Haven't looked further, but seems the expression should fit 64bit, or a reasonable RuntimeError raised. ``` $ cat ./issues/sh.spicy...