These defines were renamed in lldb git master, soon after the current
pinned version.
This fixes building with a later version of llvm than the pinned one.
Notable changes since the previous version are:
- Added a lld option for requesting delay loading libraries
- Allowing using custom .edata from input object files in lld (for Wine)
- Fixed llvm-dlltool with symbol renaming with stdcall decorations
- Supporting merging resources from more than one resource object file in lld
- Improved error messages for duplicate or undefined symbols in lld
(including demangling C++ symbols and printing source locations
where possible)
- Made lld look for for more library name patterns, matching GNU ld
- Brushed up support for using LLDB for, and building on, mingw
- Implemented initial support for armv7 and aarch64 windows in LLDB
build-mingw-w64.sh already provides --with-default-msvcrt option, but
this makes it more convenient to use build-all.sh with non-default
msvcrt config.
This makes cmake use relative paths to all source files, allowing ccache
to reuse compilation results from different source trees in different
directories.
Keep using LLVM_ENABLE_PROJECTS on msys, where symlinks aren't supported.
This updates compiler-rt, libunwind, libcxxabi and libcxx to a newer
version, matching the compiler.
This will be the official, canonical repo in the future.
For the non-dev Dockerfile build, this means that all the llvm based
subprojects need to be built in one RUN to avoid storing the llvm
checkout in the docker layers.
This version adds support for a version of the -pdb option to the MinGW
frontend to lld that can work without a parameter, and implements --help
and -v for the same frontend.
This contains the following changes relevant for mingw usage:
- A fix in a code generation bug affecting dwarf exception handling
on i686
- Setting the proper machine type in weak alias objects in import libraries,
making GNU binutils not reject such files outright
- Fixes for SEH generation for AArch64, for functions containing
jump tables
- Fixes to the calling convention of functions returning structs on AArch64
- Improvements in handling of dllexport of C++ template instantiations,
allowing building libc++ into a DLL without using --export-all-symbols
- Proper separators in the output of clang --print-search-dirs on
Windows
- A new flag --appcontainer to lld's mingw frontend
- Lld now has extra handling of crtend.o, fixing use of lld with gcc
based mingw setups, when using dwarf exceptions
The previous option doesn't seem to have had any effect, and cmake
correctly has warned about an unused option.
Also set the value of the option to TRUE, for slightly more consistency
with where this option is set in build-llvm.sh.
This version supports SEH for arm64, fixes certain cases of dllexport
of C++ classes defined entirely in headers (fixing building Qt in
debug mode), and fixes a compiler bug (regression since the end of
August 2018) when compiling glslang (as part of VLC) for armv7.
This makes the help output from llvm-nm contain keywords that libtool
may look for, makes combined compilation and linking with -flto in
one command work, fixes a bug with alignment of locally imported
symbols on arm64, allows using linker directives for specifying
default libraries to link, and allows linking to undefined behaviour
sanitizer.
This fixes cases where llvm built without assertions can hang when
assembling a file that references _GLOBAL_OFFSET_TABLE_ for windows
(which is unsupported; previously this used llvm_unreachable making
it undefined what happens if that really is hit).
This version allows clang to use address sanitizer for mingw,
adds support for linking against import libraries from GNU tools
and makes the linker provide the __CTOR_LIST__ symbol for better
compatibility with binutils based sysroots, and adds support for
thunks for armv7 (allowing linking armv7 binaries over 16 MB).
Doing that speeds up the built clang by around 25% (which is rather
significant) and reduces the size of the clang executable by almost 30%.
But since the mingw target in clang/llvm is less mature and less widely
tested than other llvm targets, running with assertions allows catching
quite a few issues during build instead of ending up with potentially
broken code.
This version fixes issues with linking mingw-style SEH data (making
sure all unwind info is included), adds support for assembler
directives like .rva, which often is used in SEH contexts, and
updates clang's unwind.h to match the libgcc ABI for SEH unwinding.
This allows using GNU binutils strip and objcopy on executables with
dwarf debug info linked by lld.
This also improves clang's handling of detecting known libraries
that override default libraries to link.
This fixes an issue with code generation for functions that needed
a stack probe on armv7, improves llvm-rc and adds a few more options
to the mingw lld frontend.
This version improves llvm-rc to support more rc script features
and features built-in filtering of CPP output in llvm-rc. It also
improves the handling of some def features in llvm-dlltool, allows
producing codeview debug information when targeting mingw and adds
handling of a few more ARM64 relocations in lld.
Martin Storsjö