Benjamin Bannier

The assertion checks that we still have a HILTI context around. It fails here since it is called via `zeek::analyzer::Analyzer::Done` after it was previously cleaned up via a call to...

This needs more work, at least - [ ] if we prefer ECR going forward update docs - [ ] mirror existing images from docker.io to ECR

> It would also be good to tweak the CI setup so that it actually pushes to ECR even for the PR, so we're sure the current secrets work before...

> Hmm - Spicy is pusing into `zeek/spicy` and `zeek/spicy-dev` rather than zeekurity. Should we align the Zeek actions with this even if it's inconsistent/non-transparent when switching between the registries?...

As a workaround, one can put the `bitfield` into a `unit` and extract the `bitfield` data after parsing; this allows reusing a single `bitfield` declaration. ```ruby type BF = unit...

Thanks for checking in on this @0xxon. This was fixed by #1170 which will be part of v1.6.0.

@MajorZiploc's diagnosis seems to be spot on. I see the same problem if the shell used by vim emits an error message when starting. Making that error go away fixed...

Tagging this as a bug since it actually can create incorrect runtime behavior. The same plugin might be installed multiple times in different versions; this could either produce "weird" behavior...

My takeaway from #1644 was that introducing a non-owning `Chunk` introduces new overhead even in code not making use of it since it partially undos (the spirit of) the optimizations...

This looks like another facet of #1316.