It's cool to be able to write 'arg1 << "stuff" << var ...;' for a lambda
accepting a std::ostream reference, but cascading compile errors mean it's no
longer worth trying to make that work -- given actual C++ lambdas.
Also clean up a lingering BOOST_FOREACH() and a boost::bind() while at it.
It seems the problem addressed by aab769e wasn't some synergy between
Boost.Phoenix and Boost.Function, but rather the lack of a Phoenix header file
introducing operator<<().
On GitHub Windows Actions runners, we're getting permissions errors trying to
tell the Python interpreter to run a NamedTempFile script. Try using
NamedExtTempFile to give each such script a .py extension.
The Python child processes used by llprocess_test.cpp and llleap_test.cpp need
the Python llsd module to communicate with the C++ parent process.
Also set LOGFAIL and BUGSPLAT_DB environment variables.
On a low-powered GitHub Mac runner, the system doesn't wake up as soon as it
should, and we get spurious "too late" errors. Try a bigger time increment.
With VS 2022 on Windows GitHub Actions runners, we can't build apr_suite at
all with the upstream .sln / .vcxproj files, so we had to switch to
"experimental" CMake support. However there's no CMakeLists.txt file for
apr-iconv, so the Windows package omits that library.
This was a longstanding complaint: that Boost shouldn't dump the (somewhat
mysterious) _1, _2 et al. names into the global namespace. Recent Boost has
fixed that, requiring 'using namespace boost::placeholders;' if you want to
use them unqualified.
Not colladadom: a GHA build of something upstream of it produced Mac link
errors seemingly related to zst.
Not xxhash: 3p-xxhash is a private repo, needing the 'public: false' switch in
its build.yaml.
Update llphysicsextensions_source to release v1.0.d3192c1.
Update fmodstudio to release v2.02.06.8f8fce1.
Update kdu to release v7.10.4.9e770ae.
Update slvoice to release v4.10.0000.32327.5fc3fe7c.399bd0e.
Newer C++ compilers have different semantics around LLSDArray's special copy
constructor, which was essential to proper LLSD nesting. In short, we can no
longer trust LLSDArray to behave correctly. Now that we have variadic
functions, get rid of LLSDArray and replace every reference with llsd::array().
It seems newer compilers have a different interpretation of exactly when to
engage LLSDArray's copy constructor. In particular, this assignment:
some_LLSD_map[key] = LLSDArray(...)(...)...;
used to convert the LLSDArray object directly to LLSD; now it first calls the
custom copy constructor, which embeds the intended array within an outer array
before assigning it into the containing map.
The newer llsd::array() function avoids that problem because what it returns
is already an LLSD object.
Taking inventory of LLSDArray assignments of that form turned up a number of
workarounds like LLSD(LLSDArray(...)). Replacing those with llsd::array() is
both simpler and more readable.
Tip of the hat to Chorazinallen for surfacing this issue!
(cherry picked from commit bb718155bd)
Nat Goodspeed