GitLab

Add -I flag for builddir to find generated spec_config.h.

Also remove the ../spec_config.h target. It didn't work as-is
since it referenced file in the wrong directory. Instead of fixing
this I decided to remove. In normal build, the previous directory
does build it first. In the case of attempting to build specific
directories first, doesn't work yet due to other problems already
(like missing dns components).

This was noticed in a failed distcheck.

- use <xxx.h> instead of <cxxx> (e.g. use stdlib.h instead of cstdlib)
- avoid using variable length array
as for the first point, it didn't make sense that cxxx didn't work and we may
want to look into it further to understand the real cause.  but since this has
been breaking build, and it somehow fixes the issue, I'll apply it for now.

okayed on jabber.

It failed to build a test, because the file was generated in another
directory and probably didn't yet exist at the time.

Without this it cannot use privileged port (most commonly 53) while changing
the run time user (-u).  Confirmed the fix, and okayed on jabber.

I won't update changelog for this fix because it's a regression after
the previous release.

Listen on the IPv6 (::) and IPv4 (0.0.0.0) wildcard addresses
for b10-auth. This returns to previous behavior prior to
change #184. Document the listen_on configuration in manual.
(Trac #649)

This reverts commit 3407184a.

This is not nop, and is necessary for the default sqlite3 data source
in the "from source" mode.

Not directly related to the ticket, I just discovered a call that does
nothing (it is called with null pointer and the function checks the
parameter is not null, if it is, it just does nothing).

so that we can run multiple instances of bind10 system (mainly for system
purposes)

This is in the private range, so it should not conflict with anything
(we still might be unlucky if two tests run at the same time, but the
chance is low).

Remove some duplicated DESTDIR.
This was reported by Jan Engelhardt via bind10-dev list.

should be sufficiently trivial, directly committing to master.

ie. make sure no headers lead to it unintentionally (i.e.2 headers should not include asiolink.h, but the specific subparts we need) and cc files that include asiolink.h should include config.h first

needed to rename the terribly named src/bin/auth/config.h for this

It is taken from resolver tests and made generic. Then it is reused for
both resolver and authoritative tests.

With a JSON fix of the resolver.spec (C++ JSON parser seems to be too
benevolent on what is JSON).

It only stores the final result, and only looks into the cache at the start of a resolve() call right now, but it appears to work

Checks that we don't reject synthesized name of the maximum length by
mistake.

note: qtype_is_any was originally const, and was temporarily changed to
mutable.  it's now immutable again.