ditro/rpm: remove ldconfig for Fedora

See merge request !761

Keep knot-resolver installation along with other kresd-related
actions as close as possible to detect possible rpm transaction issues.

https://fedoraproject.org/wiki/Changes/Removing_ldconfig_scriptlets

modules/http docs: improve the example

See merge request !760

Otherwise the http module isn't really usable :-/

modules/predict: don't skip reconfiguration if nil is passed

See merge request !755

If you didn't specify any configuration (i.e. nil), the stats module
wouldn't get loaded even though the defaults need it to work.
Now we don't skip that part and pass whole .config() even on nil config.

- It's not ideal, but this will need significant overhaul anyway when
  "parallel queries" get implemented.
- I didn't put this into array.h, as we don't seem likely to need it
  anywhere else and implementation in macros would be ugly.
- It's unclear whether this "bug" could cause anything in practice.

Release 3.2.1

See merge request knot/knot-resolver!750

Petr Špaček authored 6 years ago and Tomas Krizek committed 6 years ago

Besides other things Deckard now detects hang servers using timeout. It
might happen that CI will need a longer timeout - we can change the
value at will.

contrib/lmdb: avoid -Wshadow with embedded lmdb

See merge request !751

Vladimír Čunát authored 6 years ago and Petr Špaček committed 6 years ago

Unfortunately, CFLAGS are concatenated in the other order than we would
need for this, so let's choose this approach.

packaging tests update

See merge request !747

- also refuse revoked DNSKEY (explicitly configured as TA)
- also downgrade missing the SEP bit to a warning
- explicit message when DNSKEY is refused, even without --verbose
- code rewrite, handle flags in a better way than "== 257"

Petr Špaček authored 6 years ago and Vladimír Čunát committed 6 years ago

Previous logging was a bit confusing because it logged also intermediate
states during TA changes.

It was tiny, but still.

Closes #435.

This is an attempt to avoid conflicts when executing tests in parallel.

Formerly keys with invalid public key data were accepted, leading to
negative keytag values in RFC 5011 metadata.

At the moment tests are quite dumb and test only basic HTTPS certificate
validation and XML parsing/error detection.

We need to refactor code first to allow more detailed testing.

FIXME: webserv.lua is not terminated when tests are finished

We are not RFC 7958 compliant and support only XML with just root zone
TA. Full compliance would require either proper Lua XML parser or CMS parser
and both are hard to get packaged in Fedora and elsewhere.

Also timestamps related to TA validity are limited to UTC timezone
because cross-platform timezone parsing is hard.
(Mac OS libc does not have usable strptime(%z).)

Closes: #435

ci: stabilize pytests

Closes #427

See merge request !744