These changes introduce the ability to cross-compile working binaries for NetBSD/amd64. Previous support added in PR #26682 shared all its code with the OpenBSD implementation, and was therefore never functional (e.g. linking against non-existing symbols and using wrong type definitions). Nonetheless, the previous patches were a great starting point and made my work significantly easier. 😃
Because there are no stage0 snapshots for NetBSD (yet), I used a cross-compiler for NetBSD 7.0 RC3 and only tested some toy programs (threading and channels, stack guards, a small TCP/IP echo server and some other platform dependent bits). If someone could point me to documentation on how to generate a stage0 snapshot from a cross-compiler I'm happy to run the full test suite.
A few other notes regarding Rust on NetBSD/amd64:
- To preserve binary compatibility, NetBSD introduces new symbols for system call wrappers on breaking ABI changes and keeps the old (legacy) symbols around, see [this documentation](https://www.netbsd.org/docs/internals/en/chap-processes.html#syscalls_master) for some details. I went ahead and modified the `libc` and `std` crate to use the current (renamed) symbols instead of the legacy ones where I found them, but I might have missed some. Notably using the `sigaction` symbol (deprecated in 1998) instead of `__sigaction14` even triggers SIGSYS (bad syscall) on my amd64 setup. I also changed the type definitions to use the most recent version.
- NetBSD's gdb doesn't really support position independent executables, so you might want to turn that off for debugging, see [NetBSD Problem Report #48250](https://gnats.netbsd.org/48250).
- For binaries invoked using a relative path, NetBSD supports `$ORIGIN` only for short `rpath`s (~64 chars or so, I'm told). If running an executable fails with `execname not specified in AUX vector: No such file or directory`, consider invoking the binary using its full absolute path.
This patch basically adds a target option for omitting the `-nodefaultlibs` flag when invoking the linker. I am not sure if this is the correct or only way to approach this problem, so any feedback is welcome.
Motivation: I'm currently working on a Rust target specification for the [rumprun](/rumpkernel/rumprun) unikernel. rumprun is based on rump kernels and uses NetBSDs libc and drivers to provide a POSIXy environment. It provides its own linker wrapper that generates binaries which can be "baked" into a unikernel after configuration. Using `-nodefaultlibs` on the rumprun linker will prevent it from selecting the search paths for the rumprun libraries. My current target implementation for rumprun is here: gandro/rust@295744b2ee
Currently, only a target that `is_like_windows` will omit the `-nodefaultlibs` flag, but since rumprun is not like Windows otherwise, I think a separate flag makes more sense. This might be a breaking change for target specifications that have the `is_like_windows` option set to true. Such targets need to set `no_default_libraries` to false in order to restore the old behavior.
If set to false, `-nodefaultlibs` is not passed to the linker. This
was the default behavior on Windows, but it should be configurable
per target.
This is a [breaking-change] for target specifications that have
the `is_like_windows` option set to true. Such targets need to
set `no_default_libraries` to false in order to restore the old
behavior.
Since 1.3.0 the BufWriter has seen tremendous speedups. So when I use it in the shootout benchmarks, I see some nice speedup (which up to 1.2.0 was nixed by the pessimizations during initialization).
With -O2, LLVM's inliner can remove this code, but this does not happen
with -O1 and lower. As a result, dropping Vec<u8> was linear with length,
resulting in abysmal performance for large buffers.
See issue #24280.
Move private bignum module to core::num, because it is not only used in flt2dec.
Extract private 80-bit soft-float into new core::num module for the same reason.
This breaks out some of the lints defined in `librustc_lint/builtin.rs` into two new modules: `unused` for the `UNUSED_*` lints and `bad_style` for the various style related lints as suggested in #22206. `builtin.rs` could probably get broken up more but this is a start.
By default, the linker in use under OpenBSD is the linker of base, which
don't include /usr/local/lib where libstdc++ of gcc-4.9 lives. We need
to add this directory to linker-path-search (using -L).
Search the path of libstdc++.a, which is a known name (libstdc++.so has
SO_VERSION) in the same directory.
r? @alexcrichton
this commit needs #28495 to be commited first. It should be the last piece for building rustc under OpenBSD from scratch.
By default, the linker in use under OpenBSD is the linker of base, which
don't include /usr/local/lib where libstdc++ of gcc-4.9 lives. We need
to add this directory to linker-path-search (using -L).
Search the path of libstdc++.a, which is a known name (libstdc++.so has
SO_VERSION) in the same directory.
Move private bignum module to core::num, because it is not only used in flt2dec.
Extract private 80-bit soft-float into new core::num module for the same reason.
This changes how rustic generate `id` and `href` attributes for section header anchor. Now they are more github-like.
Also fixes breakage in docs caused by this and broken links in "Error Handling" section of book.
r? @steveklabnik
cc @alexcrichton
