Lennart Poettering
Lennart Poettering
Our minimum and recommended baseline for kernels are documented in README. Whenever we raised the baseline people came out of the woodworks and complained. I am pretty sure we should...
Uh, this sounds awfully specific. I'd be fine if there was an option RefuseRecordType= or so, which takes a list of RR types. That way you could do "RefuseRecordType=AAAA" to...
there already is dns_type_from_string()
and please parse the RR types when parsing configuration, and store them internally in a `Set` object in parsed (i.e. numeric) form, so that lookups are reasonably quick.
oops, that i added the green label was an accident. wanted to add needs-rework. sorry for the confusion
please squash your commits.
also needs a man page update, explaining the new option
The newest version still will refuse the entire query even if one of the RR types in the DnsQuestion would actually be allowed. hence, please change the code to refuse...
hmm? just iterate through the DnsQuestion, and maintain an enum: ```c static manager_validate_and_mangle_question(Manager *m, DnsQuestion **question) { enum { VALID_DONT_KNOW, VALID_ALL_GOOD, VALID_ALL_BAD, VALID_MIXED, } good = VALID_DONT_KNOW; assert(m); assert(question); /*...
force pushed new version. ptal