Use `clang` rather than `clang++` as the linker driver. This ensures
that we do not force a C++ runtime on the general code. This is fine
for now as C++ interop is not yet available for Swift. This prevents
the accidental mix-and-match of various C++ runtimes. This can cause
problems on platforms like android where `libstdc++` is an unsupported
runtime but is generally the default for Linux platforms.
- Uses version 61.1 from ICU Github unicode-org/icu repository.
- Updates mixin_linux_installation to add libicu option.
- Use -j when building libicu.
- When buiding ICU, use --with-library-suffix=swift
This suffixes the ICU symbols with _swift.
The libaries are now named libicuucswift, libicui18nswift
and libicudataswift.
- Add the contents of uconfig.h.prepend into uconfig.h. This avoids
passing the renaming CFLAGS to swift and swift-corelibs-foundation.
Also resolves:
SR-5618: libicu compilation should happen in parallel respecting -j.
This seems to more than fix a performance regression that we
detected on a metadata-allocation microbenchmark.
A few months ago, I improved the metadata cache representation
and changed the metadata allocation scheme to primarily use malloc.
Previously, we'd been using malloc in the concurrent tree data
structure but a per-cache slab allocator for the metadata itself.
At the time, I was concerned about the overhead of per-cache
allocators, since many metadata patterns see only a small number
of instantiations. That's still an important factor, so in the
new scheme we're using a global allocator; but instead of using
malloc for individual allocations, we're using a slab allocator,
which should have better peak, single-thread performance, at the
cost of not easily supporting deallocation. Deallocation is
only used for metadata when there's contention on the cache, and
specifically only when there's contention for the same key, so
leaking a little isn't the worst thing in the world.
The initial slab is a 64K globally-allocated buffer.
Successive slabs are 16K and allocated with malloc.
rdar://28189496
This seems to more than fix a performance regression that we
detected on a metadata-allocation microbenchmark.
A few months ago, I improved the metadata cache representation
and changed the metadata allocation scheme to primarily use malloc.
Previously, we'd been using malloc in the concurrent tree data
structure but a per-cache slab allocator for the metadata itself.
At the time, I was concerned about the overhead of per-cache
allocators, since many metadata patterns see only a small number
of instantiations. That's still an important factor, so in the
new scheme we're using a global allocator; but instead of using
malloc for individual allocations, we're using a slab allocator,
which should have better peak, single-thread performance, at the
cost of not easily supporting deallocation. Deallocation is
only used for metadata when there's contention on the cache, and
specifically only when there's contention for the same key, so
leaking a little isn't the worst thing in the world.
The initial slab is a 64K globally-allocated buffer.
Successive slabs are 16K and allocated with malloc.
rdar://28189496
