Jeremy Soller (dd2a02e1) at 08 Sep 17:47
Redox support
Add Redox patch to binutils 2.40. The old Redox branch should be abandoned.
Bump binutils from 2.36 to 2.37 and cherrypick commit ac81d6e0
to make it compiles with MacOS M1.
arch -x86_64 time make all
ends with the same error.I don't think bumping further would be a good idea but let me know.
I got this:
checking for libgmp... no
configure: error: GMP is missing or unusable
The solution is change mk/prefix.mk
and add your GMP path:
"$(ROOT)/$</configure" \
--target="$(TARGET)" \
--program-prefix="$(TARGET)-" \
--prefix="" \
--disable-werror \
--with-libgmp-prefix=/opt/homebrew/Cellar/gmp/6.2.1_1 \
I don't know how to automate that unfortunately.
I am trying to build this repo on Mac OS X, to see if I can get libc toolchain building on Mac OS ....
Building gdb is giving a lot of error (and millions of warnings).
They seem so fundamental, I think it must be a problem with the compiler or it's version used.
After configure, gdb/Makefile is using clang...
CC = clang
CXX = clang++
CXX_DIALECT = -std=gnu++11
$ clang --version
Apple LLVM version 10.0.0 (clang-1000.11.45.5)
Target: x86_64-apple-darwin18.2.0
Thread model: posix
Examples of errors are:
darwin-nat.c:353:4: error: no matching function for call to 'delete_thread'
delete_thread (ptid_build (inf->pid, 0, old_id));
^~~~~~~~~~~~~
./gdbthread.h:396:13: note: candidate function not viable: no known conversion from 'ptid_t' to 'thread_info *' for 1st argument
extern void delete_thread (thread_info *thread);
^
Then if I comment that out to try, it continues with many more .....
Jeremy Soller (31858b2a) at 16 Jul 16:28
This manually applies the redox specific commits on top of binutils v2_36_1. This was easier than an actual rebase which was seemingly taking ages and popping up all manner of conflicts.
Using this newer binutils fixes a build failure for aarch64-unknown-redox (see: https://sourceware.org/bugzilla/show_bug.cgi?id=27528).
This manually applies the redox specific commits on top of binutils v2_36_1. This was easier than an actual rebase which was seemingly taking ages and popping up all manner of conflicts.
Using this newer binutils fixes a build failure for aarch64-unknown-redox (see: https://sourceware.org/bugzilla/show_bug.cgi?id=27528).