* add __multi3 compiler rt function. See #1290
* compiler rt includes ARM functions for thumb and aarch64 and
other sub-arches left out. See #1526
* support C ABI for returning structs on ARM. see #1481
* introduce --disable-pic option which can generally be allowed to be
the default. compiler_rt.a and builtin.a get this option when you
build a static executable.
* compiler_rt and builtin libraries are not built for build-lib
--static
* posix_spawn instead of fork/execv
* disable the error limit on LLD. Fixes the blank lines printed
Pass --no-entry instead of --relocatable to lld. Both stop a reference
to the _start() entry point from being emitted but --relocatable also
prevents public symbols from being exported when creating an executable.
This will make it easier to do things like `zig help | grep something`.
Invalid arguments will now display a short notice for `zig help`
instead of showing the full usage information. This will make it easier
to see the actual error.
On Arch Linux the current default compiler is gcc 8.2.1 and this change
is needed to ignore the following errors:
In file included from /home/wink/local/include/llvm/ADT/STLExtras.h:21,
from /home/wink/local/include/llvm/ADT/StringRef.h:13,
from /home/wink/local/include/llvm/ADT/StringMap.h:17,
from /home/wink/local/include/llvm/Support/Host.h:17,
from /home/wink/local/include/llvm/ADT/Hashing.h:49,
from /home/wink/local/include/llvm/ADT/ArrayRef.h:13,
from /home/wink/local/include/llvm/ADT/APFloat.h:21,
from /home/wink/local/include/clang/AST/APValue.h:18,
from /home/wink/local/include/clang/AST/Decl.h:17,
from /home/wink/local/include/clang/AST/ASTTypeTraits.h:20,
from /home/wink/local/include/clang/AST/ASTContext.h:18,
from /home/wink/local/include/clang/Frontend/ASTUnit.h:18,
from /home/wink/prgs/ziglang/zig/src/translate_c.cpp:18:
/home/wink/local/include/llvm/ADT/SmallVector.h: In instantiation of ‘void llvm::SmallVectorTemplateBase<T, true>::push_back(const T&) [with T = std::pair<void*, long unsigned int>]’:
/home/wink/local/include/llvm/Support/Allocator.h:249:33: required from ‘void* llvm::BumpPtrAllocatorImpl<AllocatorT, SlabSize, SizeThreshold>::Allocate(size_t, size_t) [with AllocatorT = llvm::MallocAllocator; long unsigned int SlabSize = 4096; long unsigned int SizeThreshold = 4096; size_t = long unsigned int]’
/home/wink/local/include/clang/AST/ASTContext.h:659:42: required from here
/home/wink/local/include/llvm/ADT/SmallVector.h:313:11: error: ‘void* memcpy(void*, const void*, size_t)’ writing to an object of type ‘struct std::pair<void*, long unsigned int>’ with no trivial copy-assignment; use copy-assignment or copy-initialization instead [-Werror=class-memaccess]
memcpy(this->end(), &Elt, sizeof(T));
~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from /usr/include/c++/8.2.1/utility:70,
from /home/wink/local/include/llvm/Support/type_traits.h:19,
from /home/wink/local/include/llvm/Support/Casting.h:19,
from /home/wink/local/include/clang/Basic/LLVM.h:22,
from /home/wink/local/include/clang/AST/APValue.h:17,
from /home/wink/local/include/clang/AST/Decl.h:17,
from /home/wink/local/include/clang/AST/ASTTypeTraits.h:20,
from /home/wink/local/include/clang/AST/ASTContext.h:18,
from /home/wink/local/include/clang/Frontend/ASTUnit.h:18,
from /home/wink/prgs/ziglang/zig/src/translate_c.cpp:18:
/usr/include/c++/8.2.1/bits/stl_pair.h:198:12: note: ‘struct std::pair<void*, long unsigned int>’ declared here
struct pair
^~~~
We were caching the ConstExprValue of string literals,
which works if you can never modify ConstExprValues.
This premise is broken with `comptime var ...`.
So I implemented an optimization in ConstExprValue
arrays, where it stores a `Buf *` directly rather
than an array of ConstExprValues for the elements,
and then similar to array of undefined, it is
expanded into the canonical form when necessary.
However many operations can happen directly on the
`Buf *`, which is faster.
Furthermore, before a ConstExprValue array is expanded
into canonical form, it removes itself from the string
literal cache. This fixes the issue, because before an
array element is modified it would have to be expanded.
closes#1076
this prevents the situation where we determine the cache
manifest and write it, but then crash or otherwise error out
before putting the artifacts in the proper place.
now the artifacts will be in place because cache_release
happens after that step is done.
