| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
- tcmalloc to not use libunwind API, as it uses slow cursor steps.
- Remove UNW_LOCAL_ONLY from CXXFLAGS everywhere.
|
|
|
|
| |
This reverts commit 969151e9ab69dcb53397cf40f810e718421db081.
|
|
|
|
|
|
| |
- SCons configure to probe for libunwind support
- gracefully handle SIGUSR2 without libunwind
- integrate libunwind on-by-default (linux-x86_64) into evergreen
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is a very low-level bare metal header, and should ideally
#include only std headers.
- Remove Decimal128 dependency from endian.h. Decimal128 doesn't
need endian conversions, and makes endian.h transitively include
several higher-level mongo specifics. Current conversions are
underspecified and mathematically incorrect or at least ambiguous.
They swap the order within each of the low64 and high64 fields,
but don't swap them with each other. This is behavior needed only
by one spot in db/pipeline/value.cpp to deserialize NumberDecimal,
so we can just inline the behavior there and remove it from
endian.h.
- Remove MONGO_CONFIG_BYTE_ORDER from config.h. The running compiler
holds the ultimate truth on what the target endianness is
(available via _BYTE_ORDER_). We should not read it from a config
header. The names exported into C++ code for the endian
possibilities can be changed to line up with those in C++20's
std::endian {big,little,native} enum. This eliminates the
Scons<=>C++ bridge protocol of "1234" and "4321" magic numbers.
Scons will talk to the compiler, not directly to the code.
- Use enum expressions (including if constexpr) rather than #if for
branching on endianness. This makes bit-rot of unexecuted paths
less likely, and is just cleaner C++.
- Remove bswap_slow variants. All supported compilers have builtin
bswap operations. Can reduce to a simple MSVC vs GCC branching.
All compilers have a builtin, so remove the bswap_slow"
implementation.
- Don't need all the push_macro / pop_macro stuff, or really any
macros at all. Just rely on inline C++ functions. These optimize to
the same thing.
- Don't need ByteOrderConverter or IntegralTypeMap traits either.
Simpler metaprogramming based only on sizeof will work fine.
All in all we can remove about 400 lines of old code here and
shave some low-level edges off of the dependency graph.
- benchmark
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
GCC changed the output of vec_vbpermq between 5.4.0 and later. This
seems likely to be a bug, probably in earlier versions of gcc. Rather
than attempt to enumerate how the compiler versions work, this adds a
runtime configure check which will select the lane that works.
|
| |
|
| |
|
|
|
|
| |
This reverts commit ad223ea38846ceac92d7b5356add62e23e985537.
|
| |
|
|
|
|
|
|
|
| |
Added an option to permit specifying a Diffie Hellman parameters file
in PEM format which will be passed to OpenSSL. We also now indicate
to OpenSSL that we'd like Elliptic Curve Diffie Hellman Exchange,
if the client supports it.
|
|
|
|
|
| |
Also, use thread_local everywhere for our thread specific data needs
and remove the legacy support.
|
| |
|
|
|
|
| |
Enjoy the C++14! 😀 ❤️ thank you!
|
| |
|
| |
|
| |
|
|
|
|
| |
Also adds SecureT, a wrapper proxy for newing objects on the secure heap.
|
| |
|
| |
|
|
|
|
|
| |
Add scons build parameter --experimental-decimal-support=on,
which enables experimental support of a 128 bit decimal type.
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|