Arne Welzel

For Zeek 6.1 and 6.2, the PR from @eric-ooi #24 needs to be merged. For Zeek 6.0, one can manually specify using the latest version by entering `zkg install bro-http2...

Thanks for the report! > One solution would be to "over-escape" all backslashes regardless of context, so \x07 becomes \\x07 and \\x07 becomes \\\\x07. IIUC from looking at the JSON...

> everything else non-printable as \\xXX That's not something a Python `json.loads()` likes or understands though: ``` In [43]: json.loads('{"a": "test_\\x07"}') JSONDecodeError: Invalid \escape: line 1 column 13 (char 12)...

> Why couldn't invalid bytes be encoded as a sequence of \u00xx in JSON like shown above? It seems that would at least preserve the actual content (at the expense...

> To clarify, I did not mean to suggest anything that standard libs barf on. My point was that Yep! I guess the awkward backslash escaping just helped me trigger...

I guess whatever we do, it'd probably be good to opt-in or have some sort of policy on the LogAscii module. Keeping what we do today as a fallback available....

Hmm, piggy backing, the following isn't right (I think just `on MagicUnit::chunk { }` is proper), but the error messages for that are the same and unclear what the error...

Thanks for doing this! In addition to @0xxon , think it's nicer to have the `redef` out of the `export` section similar to what we do in the package-template (which...

>> How about implementing a change handler? > > Implementing a change-handler would be the greatest way to deal with this. My two cents: Doesn't seem worth the complexity and...

> Rename and split ports to well_known_server_ports and well_known_client_ports I'd vote with `ports` and only split if client and server distinction if the modules make the distinction today which seems...