[BUILD] Avoid buggy warning with gcc <= 8

[Romain-Geissler-1A]

gcc <= 8, when building with -Wextra (and -Werror) wrongly raises this:

/data/mwrep/res/mdw/SIOTF/internal/opentelemetry_cpp/trace/18-0-0-7/include/opentelemetry/trace/default_span.h: In copy constructor 'opentelemetry::v1::trace::DefaultSpan::DefaultSpan(const opentelemetry::v1::trace::DefaultSpan&)':
/data/mwrep/res/mdw/SIOTF/internal/opentelemetry_cpp/trace/18-0-0-7/include/opentelemetry/trace/default_span.h:70:3: error: base class 'class opentelemetry::v1::trace::Span' should be explicitly initialized in the copy constructor [-Werror=extra]
   DefaultSpan(const DefaultSpan &spn) noexcept : span_context_(spn.GetContext()) {}
   ^~~~~~~~~~~

See on Compiler Explorer here: https://godbolt.org/z/ed5rv74nT

I believe Jason Merrill fixed it in gcc for all gcc >= 9 with: https://gcc.gnu.org/git/?p=gcc.git;a=commitdiff;h=f3f7cefecc833b4ab652215ceb8b408c21dca225;hp=777083bb806dbe31ab97002b7d445191d3ee7a2d

Workaround this by calling explicitly the empty Span constructor.

[netlify]

✅ Deploy Preview for opentelemetry-cpp-api-docs canceled.

Name	Link
🔨 Latest commit	6c04bf2
🔍 Latest deploy log	https://app.netlify.com/sites/opentelemetry-cpp-api-docs/deploys/670532b9ade31200093a0cc3

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 87.82%. Comparing base (497eaf4) to head (6c04bf2).
Report is 138 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3087      +/-   ##
==========================================
+ Coverage   87.12%   87.82%   +0.71%     
==========================================
  Files         200      195       -5     
  Lines        6109     5968     -141     
==========================================
- Hits         5322     5241      -81     
+ Misses        787      727      -60     

Files with missing lines	Coverage Δ
api/include/opentelemetry/trace/default_span.h	13.34% <ø> (ø)

... and 125 files with indirect coverage changes

[lalitb]

LGTM. Generally, I am not in favor of providing fix for a buggy compiler, but in this case the fix is trivial, and no side effect so should be fine :)

[marcalff]

Bonjour @Romain-Geissler-1A

The links to compiler explorer and to the gcc patch makes this very convincing, thanks for the research.

LGTM.