The 13.0.1-1 MSYS2 binaries cause linker errors due to missing `new` and
`delete` symbols. This commit partially reverts the LLVM 13.0.1 upgrade:
Windows cross-compilation still uses `clang 13.0.1`, but is compiled
against `libc++ 13.0.0`.
- The patches `iostreams-106.patch` and `signals2-noise.patch` were
incorporated into Boost 1.75.
- The allocator access deprecation issue was fixed in Boost 1.76.
Closeszcash/zcash#4945.
This is the second release in a row where LLVM has cut a X.0.1 for
everything except Darwin, so I've adjusted its URLs and paths on the
assumption this will continue.
This enables of the use of AI_* definitions in the Windows headers,
specifically AI_ADDRCONFIG, which fixes an issue with libevent and
ipv6 on Windows.
It also aligns with what we define in configure when building Core.
(cherry picked from commit eb6b73540d1ee7ff5a6874dd0e35f9b30b68e3b8)
libevent uses getaddrinfo when available, and falls back to gethostbyname
Windows has both, but gethostbyname only supports IPv4
libevent fails to detect Windows's getaddrinfo due to not including the right headers
This patches libevent's configure script to check it correctly
(cherry picked from commit 03e056edcd1a7f7197a29068c52fa33fce12f7d7)
Clang's download path includes its version, but LLVM doesn't always
publish every package for every version, so we need to support multiple
versions.
Fixeszcash/zcash#4954.
C++17 deprecated the two-argument version of std::allocator::allocate.
Starting with 11.0, libc++ enforces these deprecations, which causes
warnings from the Boost headers, and since we require native Linux
builds to be warning-free, this breaks CI.
Boost addressed this issue for MSVC in 1.75; the patch in this commit
forces Clang to be handled in the same way.
We want to supply well-known vars to ./configure scripts to do with as
they please. However, we do _not_ want to override these well-known vars
at make-time as certain build systems expect a self-mangled version of
these well-known vars.
For example, freetype and bdb will prepend `libtool --mode=compile' to
CC and CXX, which, if we override CC on the command line at make-time,
will break the build.
Previously, we specified the target-os in the toolset (and sometimes
used the wrong command line flags), now we have a clear separation,
which is favored by ./bootstrap.sh and ./b2.
This means that all supported OSes will specify the correct target-os=
and toolset= on the command line.
b2 will pickup our user-config.jam just fine, however, bootstrap.sh has
its own toolset autodetect mechanism, which doesn't GAF about our
user-config.jam
Zcash: This also reverts b6d0996cec which
fixes a PIE linking error, but likely breaks FreeBSD build support.
All other mk files use the package variable consistently except for the two instances here, which have always been here, since depends was introduced in 0.10.
The ancient "darwin-4.9.1" profile has long been used to match against
clang, which prior to version 9, reported 4.9.1 as its version when
invoking "clang++ -dumpversion". Presumably this was a historical
compatibility quirk related to Apple's switch from gcc to clang.
This was "fixed" in clang 9.0, so that -dumpversion reports the real
version. Unfortunately that had the side-effect of breaking the
(brittle) boost compiler detection.
Move to the seemingly more-correct "clang-darwin" profile, which passes
the checks and builds correctly.
Also switch to using ar rather than libtool for archiving, as it's what
the clang-darwin profile expects to be using.
Note that because this is using a different profile, some of the final
command-line arguments end up changing. The changes look sane at a
glance.