Since at least d7bddc27b2, we've had a
situation where one should depend on:
- `stdenv.cc.bintools`: for executables at build time
- `libbfd` or `libiberty`: for those libraries
- `targetPackages.cc.bintools`: for exectuables at *run* time
- `binutils`: only for specifically GNU Binutils's executables,
regardless of the host platform, at run time.
and that commit cleaned up this usage to reflect that. This PR flips the
switch so that:
- `binutils` is indeed unconditionally GNU Binutils
- `binutils-raw`, which previously served that role, is gone.
so that the correct usage will be enforced going forward and everything
is simple.
N.B. In a few cases `binutils-unwrapped` (which before and now was
unconditionally actual GNU binutils), rather than `binutils` was used to
replace old `binutils-raw` as it is friendly towards some cross
compilation usage by avoiding a reference to the next bootstrapping
change.
Among other things, this will allow *2nix tools to output plain data
while still being composable with the traditional
callPackage/.override interfaces.
Certain tools, e.g. compilers, are customarily prefixed with the name of
their target platform so that multiple builds can be used at once
without clobbering each other on the PATH. I was using identifiers named
`prefix` for this purpose, but that conflicts with the standard use of
`prefix` to mean the directory where something is installed. To avoid
conflict and confusion, I renamed those to `targetPrefix`.
- Give cctools a dev output for the headers
- Update Libsystem to grab the headers from that dev output
- Don't include the headers in Darwin binutils, just as GNU Binutils no
longer does.
This includes adding a new xcbuild-based libutil build to test the waters a bit there.
We'll need to get xcbuild into the stdenv bootstrap before we can make the main build,
but it's nice to see that it can work.
This requires some small changes in the stdenv, then working around the
weird choice LLVM made to hardcode @rpath in its install name, and then
lets us remove a ton of annoying workaround hacks in many of our Go
packages. With any luck this will mean less hackery going forward.
This reverts commit 0a944b345e, reversing
changes made to 61733ed6cc.
I dislike these massive stdenv changes with unclear motivation,
especially when they involve gratuitous mass renames like NIX_CC ->
NIX_BINUTILS. The previous such rename (NIX_GCC -> NIX_CC) caused
months of pain, so let's not do that again.
* pkgs: refactor needless quoting of homepage meta attribute
A lot of packages are needlessly quoting the homepage meta attribute
(about 1400, 22%), this commit refactors all of those instances.
* pkgs: Fixing some links that were wrongfully unquoted in the previous
commit
* Fixed some instances
The main changes are in libSystem, which lost the coretls component in 10.13
and some hardening changes that quietly crash any program that uses %n in
a non-constant format string, so we've needed to patch a lot of programs that
use gnulib.
We want platform triple prefixes and suffixes on derivation names to
be used consistently. The ideom this commit strives for is
- suffix means build != host, i.e. cross *built* packages. This is
already done.
- prefix means build != target, i.e. cross tools. This matches the
tradition of such binaries themselves being prefixed to disambiguate.]
Binutils and cctools, as build tools, now use the latter
- No more *Cross duplication for binutils on darwin either.
`cctools_cross` is merged into plain `cctools`, so `buildPackages`
chains alone are used to disambiguate.
- Always use a mashup of cctools and actual GNU Binutils as `binutils`.
Previously, this was only done in the native case as nobody had
bothered to implement the masher in the cross case. Implemented it
basically consisted of extending the wrapper to deal with prefixed
binaries.
This also fixes a missing header in the SDK that rtags needs to work
properly. The underlying cause is that C++ headers got shuffled around a
lot in libc++ 3.8 (I believe) and became more standards-compliant, which
led to a lot of C-compatible passthrough header files being added to it
like math.h, which defines some C++-compatible versions of standard
functions like signbit, while #include_next'ing the system math.h. In
this case, including the SDK was stuffing another math.h in front of the
libc++ shim, which led to all sorts of mysterious failures.
This sort of thing is going to get revamped to be less hackish soon but
for now I want it to work. In this particular case, libc++ 4 (and maybe
earlier) gets very upset if we're imprecise about our const markers, and
I guess libauto was careless. This fixes it (PtrPtrMap) to be correct.
This is a slightly less ambitious version of the (now reverted) commit
377cef8d16, which had a bunch of issues
that I don't have time to resolve right now.
This wasn't being used and it was causing an error when evaluating:
error: attribute ‘CoreOSMakefiles’ missing, at /Users/mbauer/Projects/nixpkgs2/pkgs/os-specific/darwin/apple-source-releases/default.nix:140:71
The SDK includes cups header files, but not the libraries. The
`nixpkgs.cups` definition doesn't build on darwin due to the SDK being
too old. This change symlinks the system cups libraries into the old
SDK.
pkill isn't building because of some missing headers:
- xpc/xpc.h
- os/base_private.h
- _simple.h
They are all available somewhere but not set up correctly in the Darwin
stdenv.
TODO: add pkill back in!
This actually has no effect but it bugged me to keep seeing an old version
in the package names :) and since we're making a bunch of stdenv changes
at once, I might as well.
This reinstates the libSystem selective symbol export machinery we used
to have, but locks it to the symbols that were present in 10.11 and skips
the actual compiled code we put into that library in favor of the system
initialization code. That should make it more stable and less likely to
do weird stuff than the last time we did this.
Darwin systems need to be able to find CoreFoundation headers as well as
libc headers. Somehow, gcc doesn't accept any "framework" parameters
that would normally be used to include CoreFoundation in this
situation.
HACK: Instead, this adds a derivation that combines the two. The result
works but probably not a good long term solution.
ALTERNATIVES: Maybe sending patches in to GCC to allow
"native-system-framework" configure flag to get this found.
It's a long build and generally painful to split into smaller commits,
so I apologize for lumping many changes into one commit but this is far
easier.
There are still several outdated parts of the darwin stdenv but these
changes should bring us closer to the goal.
Fixes#18461
Revert a revert of a merge that shouldn't have been in master but was intentionally in staging.
Next time I'll do this right after the revert instead of so far down the line...
This reverts commit 9adad8612b.
This currently only produces a static library, but is a start :) soon we
might be able to incorporate it into our stdenv, but we need to get the
build system to produce a proper .framework first.
We don't actually need the private headers and the private qos.h was
overwriting the public one, causing weird issues downstream (especially
with Swift's CoreFoundation)
I can't submit this in smaller units because the various components all
depend on one another during the stdenv bootstrap, so I think this is
the smallest sensible change I can make.
I also removed the symbol-hiding shenanigans in Libsystem. It might mess
up compatibility with 10.9 but I don't really want to support the added
complexity and I see little evidence of anyone else wanting to support
it. If someone cares, we might be able to revive compatibility, but for
now it'll stay like this.
The AVFoundation framework uses a relative path that presumes that
CoreGraphics is a child framework of ApplicationServices. It is not in
the 10.9 SDK.
This patch makes it one by tweaking the framework derivation generator
with special support to address this problem.
The `otool` binary is provided by the `cctools` package (and `binutils`)
on darwin, which is properly packaged and compiled from source.
This old standalone `otool` was simply a symlink to `/usr/bin/otool`,
which notably depended on the user having already installed the Command
Line Tools via XCode, and would fail dependent builds if they hadn't.
