Failing Hydra build: https://hydra.nixos.org/build/249763077/nixlog/12
The problem is that glibc commit
64b1a44183a3094672ed304532bedb9acc707554 marked the `FILE*` argument of a few
functions including `fread` & `ferror` as non-null. The applied patch
("Android: add better nullability checks for nullability annotations added in NDK 26")
is targeted for the Android platform, but fixes said issue as well: the
handle returned from `fopen` is of type `Optional<T>` and the `guard`
expression unwraps that now (and throws an exception if `nil` is
returned). The previous `nil`-check didn't modify the type of `fp`, but
only raised the exception and moved on with `Optional<T>`.
It's a little sad that the patch needs to be applied at so many places,
but I guess that's what you get with language-level package managers 🤷
Also, seems good-enough to me given that it's actually temporary, the
patch is already upstream and will probably be obsolete at one of the
next Swift updates.
We need to set -crt-static on musl for rustdoc as well, so let's unify
the wrappers. Ideally, rather than wrapping rustdoc, we'd have
rustdoc use the wrapped rustc, but that's currently only possible with
an unstable option (--test-builder).
The options set by the wrapper, -C target-feature and --sysroot, are
supported by both rustdoc and rustc, but other flags maybe not be
supported by both, so I've introduced different environment
variables (the existing NIX_RUSTFLAGS and a new NIX_RUSTDOCFLAGS) to
allow those to be set independently.
This fixes cargo-auditable in pkgsMusl., which broke because its
doctests stopped working when -crt-static was moved to the wrapper.
Fixes: 79156bf13a ("rustc: move crt-static default override to wrapper (#291829)")
The Nixpkgs release code generates a list of attributes on
x86_64-linux, then tries to evaluate them on all platforms. This
meant that Darwin evals broke when llvmPackages was bumped to 17 on
Linux, because the list of attributes is evaluated for Linux, finds an
llvmPackages.mlir attribute, then gets an evaluation error when it
tries to evalute that on Darwin. The easiest fix is to just make sure
an llvmPackages.mlir attribute exists on Darwin too.
Previously, when cross compiling from non-musl to musl, the crt-static
default override wouldn't be applied, because the compiler wouldn't
have been built with it due to fastCross. Moving it to the wrapper
fixes this without having to introduce extra compiler rebuilds. And
because the wrapper is applied even to the bootstrap rustc, we no
longer need special handling of crt-static in the Cargo expression.
Unlike --sysroot, rustc allows -C target-feature= to be passed
multiple times, with later instances taking precedence over earlier
ones. This means that it's very easy to set the default in the
wrapper, just by our overridden default before any other arguments.
This fixes pkgsCross.aarch64-multiplatform-musl.mesa from x86_64-linux.
- get rid of rec
- extreme use of recursive rec
- format version string according to commit
- cb89165fb9 (#280312, RFCs 107 and 147)
- strictDeps
- meta.mainProgram
Before it failed with this error:
failed to start `cargo metadata`: No such file or directory (os error 2)
Suffixing it to PATH to make sure any other cargo is used before this.
This disables building dynamic libraries for GHC 9.4.x and before, when enableShared and
enableProfiledLibs are set to false. Currently this only disables dynamic linking of GHC
itself, but still creates dynamic libraries.
The hadrian-based GHC 9.6+ already disable building dynamic libraries via the no_dynamic_libs
transformer in this case.
The reason for this happening before was, that the default was assumed to be "v p dyn", of
which "p" was taken away when profiled libs were disabled. But this would unconditionally
enable the dyn way. Using the BUILD_PROF_LIBS flag in this case will let GHC's config.mk.in
script make the proper decisions about which GhcLibWays to build.
