rec: make code boost 1.66 compatible again

[omoerbeek]

Short description

It turns out that some systems we want to support have a rather old C++ library and boost 1.66.
So partly revert the std::optional changes and work around a few issues that g++-15 reports on these systems.

Checklist

I have:

• read the CONTRIBUTING.md document
• read and accepted the Developer Certificate of Origin document, including the AI Policy, and added a "Signed-off-by" to my commits
• compiled this code
• tested this code
• included documentation (including possible behaviour changes)
• documented the code
• added or modified regression test(s)
• added or modified unit test(s)

[coveralls]

Pull Request Test Coverage Report for Build 20342847085

Details

• 98 of 101 (97.03%) changed or added relevant lines in 19 files are covered.
• 23 unchanged lines in 7 files lost coverage.
• Overall coverage increased (+5.9%) to 73.321%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
pdns/recursordist/recursor_cache.cc	13	14	92.86%
pdns/recursordist/rec-nsspeeds.cc	7	9	77.78%

Files with Coverage Reduction	New Missed Lines	%
pdns/recursordist/rec-nsspeeds.cc	1	83.89%
pdns/recursordist/recursor_cache.cc	1	85.94%
pdns/recursordist/negcache.hh	2	88.0%
pdns/packethandler.cc	3	72.36%
pdns/recursordist/recpacketcache.hh	3	89.55%
pdns/recursordist/test-syncres_cc1.cc	6	90.26%
pdns/recursordist/syncres.cc	7	81.12%

Totals
Change from base Build 20338794597:	5.9%
Covered Lines:	128789
Relevant Lines:	164878

---

💛 - Coveralls

[miodvallat]

LGTM, but please rewrite the first commit message (or squash).

[rgacogne]

It's a bit of a shame to have to revert all this changes :-/ I guess there are too many issues to make it work with boost 1.66?

[omoerbeek]

It's a bit of a shame to have to revert all this changes :-/ I guess there are too many issues to make it work with boost 1.66?

I intend to spent some time today to see if we can work around the issue.

[omoerbeek]

I could not convince the compiler that a hash function for std::optionalstd::string exists in the context of boost 1.66.

I am now pondering an alternative solution: do away with the optional, and consider an empty tag as signifying non-existence.

Strictly speaking that is changing the semantics as a present but empty tag is not longer a thing then, but as not a lot of people (none?) use the Routing tag for partitioning the record cache that should not be an issue.

[rgacogne]

Without looking at the code that seems like a nice solution to me. We would have to document the change but I'm not aware of any users of the routing tag feature either, so it's unlikely to be an issue.

[rgacogne]

Some of these changes feel a bit unrelated to the initial issue, but that's not a big deal and apart from that I'm OK with the PR.

[omoerbeek]

Some of these changes feel a bit unrelated to the initial issue, but that's not a big deal and apart from that I'm OK with the PR.

The somewhat unrelated changes were warnings detected by the Sue boost/compiler (1.66/15) combo.