7459a40 `stdenv-bootstrap: force using new libc from stage2` overrode
`gcc-wrapper`. 91fa7657 `stdenv: revert gcc hack` dod not restore the initial
value.
As a result attempt to build glibc-2.35 fails early on perl as:
ld: cannot find -lnsl: No such file or directory
collect2: error: ld returned 1 exit status
I can't compile the test program.
(The supplied flags or libraries might be incorrect.)
The change restores `gcc-wrapper` as it was before 7459a40.
This fix is needed to work around linker-errors such as
undefined reference to `__libc_csu_fini'
which I got in almost every derivation which is part of stage2. The
reasoning behind this is that the startup-code was simplified[1] and
thus `__libc_csu_fini` doesn't exist anymore.
A workable solution is to use a newer libc which properly links in
stage3. And actually this seems expected given the rationale for stage3:
# Construct a third stdenv identical to the 2nd, except that this
# one uses the rebuilt Glibc from stage2. It still uses the recent
# binutils and rest of the bootstrap tools, including GCC.
So this patch basically overrides the libraries inside `gcc-unwrapped` -
which is basically the bootstrap tools and thus also contains the libc
used in stage3 - with the shared objects from the freshly built libc
from stage2.
[1] https://sourceware.org/pipermail/libc-alpha/2021-March/123079.html
Flake users that use a command like `nix build nixpkgs#hello` on a
broken/insecure package will not be able to use an environment variable
to override that behavior, unless they pass `--impure` to the command.
Co-authored-by: pkharvey <kayharvey@protonmail.com>
Adds an easy method of appending compiler flags to your stdenv via a
list.
Co-authored-by: tomberek <tomberek@users.noreply.github.com>
Co-authored-by: Gytis Ivaskevicius <gytis02.21@gmail.com>
Co-authored-by: sternenseemann <sternenseemann@systemli.org>
The `curlMinimal` is to be used throughout the early bootstrap
stages. The final stage will allow the new references of the `curl`.
Fixes: 29526bc2 ('curl: IDN support requires libidn2 package')
llvmPackages_*.clang should check the default compiler for the package
set it is targeting (targetPackages.stdenv.cc) instead of the compiler
that has been used to build it (stdenv.cc) in order to get some sense of
whether to use libc++ or libstdc++.
Since we are now inspecting targetPackages in the llvmPackages.clang
attribute, we need to avoid using it in the cross stdenv — which just
forces us to explicitly request libcxxClang for darwin instead of
relying on the clang attribute to pick it for us.
We also need to do something similar for targetPackages.stdenv.cc: Here
the llvmPackages.clang logic would work as we want (inspect
targetPackages.stdenv.cc and if it doesn't exist, make the choice based
on stdenv.cc), but it gets locked in a cycle with the previous package.
We can easily break this, however: We know that the previous set had
clang and the next one doesn't exist, so we'd choose libcxxClang any day
of the week.
Add `shellDryRun` to the generic stdenv and substitute it for uses of
`${stdenv.shell} -n`. The point of this layer of abstraction is to add
the flag `-O extglob`, which resolves#126344 in a more direct way.
`TargetConditionals.h` was missing several definitions, like
`TARGET_OS_TV` that are part of SDK 10.12 at least. And one that doesn't
seem to occur in any SDK afaict, `TARGET_OS_EMBEDDED_OTHER`.
I added the definitions from SDK 10.12 verbatim and defined
`TARGET_OS_EMBEDDED_OTHER` to be equal to `0`.
I think none of this works if `darwin.Libsystem` is used to build for
linux or iOS though so maybe this needs a more thorough fix?
This reverts 336d82617f because it's no
longer necessary.
Packages that use libtool run it as a wrapper around the linker.
Before calling the linker, libtool will determine what libraries would
be linked, and check if there's a corresponding libtool
archive (libfoo.la) file in the same directory . This file
contains extra information about the library. This is especially
important for static linking, because static archives don't contain
dependency information, so we need libtool to use the .la files to
figure out which libraries actually need to be linked against.
But in Nixpkgs, this has never worked. libtool isn't able to find any
libraries, because only the compiler wrapper knows how to find them,
and the compiler wrapper is opaque to libtool. This is why
pkgsStatic.util-linuxMinimal doesn't build prior to this patch — it
depends on libpam, which depends on libaudit, and if libtool can't
find the .la file, nothing will tell the linker to also link against
libaudit when linking libpam. (It was previously possible to build a
static util-linux, because linux-pam only recently had the audit
dependency added.)
There are a couple of ways we could fix this, so that libtool knows
where to look for .la files.
* Set LD_LIBRARY_PATH/DYLD_LIBRARY_PATH/whatever, which libtool will
examine. This would have major side effects though, because the
dynamic linker looks at it too.
* Inject libtool scripts with the appropriate information. That's
what I've done here. It was the obvious choice because we're
already finding and modifying the libtool scripts, to remove paths
outside the Nix store that libtool might check in unsandboxed
builds. Instead of emptying out the system paths, we can
repopulate it with our own library paths.
(We can't use a wrapper like we do for other tools in Nixpkgs, because
libtool scripts are often distributed in source tarballs, so we can't
just add a wrapped version of libtool as a dependency. That's why
there's already the fixLibtool function in stdenv.)
With this change, libtool is able to discover .la files, and
pkgsStatic.util-linuxMinimal can build again, linking correctly
against libpam and libaudit.
`--enable-deterministic-archives` is a GNU specific strip flag and
causes other strip implementations (for example LLVM's, see #138013)
to fail. Since strip failures are ignored, this means that stripping
doesn't work at all in certain situation (causing unnecessary
dependencies etc.).
To fix this, no longer pass `--enable-deterministic-archives`
unconditionally, but instead add it in a GNU binutils specific strip
wrapper only.
`commonStripFlags` was only used for this flag, so we can remove
it altogether.
Future work could be to make a generic strip wrapper, with support for
nix-support/strip-flags-{before,after} and NIX_STRIP_FLAGS_{BEFORE,AFTER}.
This possibly overkill and unnecessary though -- also with the
additional challenge of incorporating the darwin strip wrapper somehow.
This reverts commit 488395c0f8.
Currently, `nix print-dev-env` fails to execute if this function is present, because of its use of hex literals.
Until this issue (https://github.com/NixOS/nix/issues/5262) is solved, we should revert this to prevent breakage.
somehow `read -N 0` behavior changed in bash 5. `read -d ''` has identical behavior
the purpose of the function is to read stdin and exit 1 on a null byte (i.e. if stdin is the content of a binary)
(cherry picked from commit 5d0acf20f88b1820cb8b641cfc5a43e973122701)
The old stdenv adapters were subtly wrong in two ways:
- `overrideAttrs` leaked the original, unoverridden `mkDerivation`.
- `stdenv.override` would throw away any manually-set `mkDerivation`
from a stdenv reverting to the original.
Now, `mkDerivation` is controlled (nearly directly) via an argument, and
always correctly closes over the final ("self") stdenv. This means the
adapters can work entirely via `.override` without any manual `stdenv //
...`, and both those issues are fixed.
Note hashes are changed, because stdenvs no previously overridden like
`stdenvNoCC` and `crossLibcStdenv` now are. I had to add some
`dontDisableStatic = true` accordingly. The flip side however is that
since the overrides compose, we no longer need to override anything but
the default `stdenv` from which all the others are created.
When we "fix" libtool, we empty out its system library path to avoid
it discovering libraries in e.g. /usr when the sandbox is disabled.
But this also means that the checks libtool does to make sure it can
find the libraries its supposed to be linking to won't work. On Linux
and Darwin, this isn't a problem, because libtool doesn't actually
perform any checks, but it is on at least NetBSD and Cygwin[1].
So, we force libtool not to do these checks on any platform, bringing
the more exotic platforms into line with the existing behaviour on
Linux and Darwin.
Without this change, lots of library packages produce warnings like
this in their build output on the platforms with checks by default:
*** Warning: linker path does not have real file for library -lz.
*** I have the capability to make that library automatically link in when
*** you link to this library. But I can only do this if you have a
*** shared version of the library, which you do not appear to have
*** because I did check the linker path looking for a file starting
*** with libz but no candidates were found. (...for regex pattern test)
*** The inter-library dependencies that have been dropped here will be
*** automatically added whenever a program is linked with this library
*** or is declared to -dlopen it.
And dependent packages break because libtool doesn't link their
transitive dependencies. So making this change fixes _lots_ of
packages on those platforms.
[1]: https://git.savannah.gnu.org/cgit/libtool.git/tree/m4/libtool.m4?id=544fc0e2c2a03129a540aebef41ad32bfb5c06b8#n3445
At least for now. Such changes are risky (we have very many packages),
and apparently it needs more testing/review without blocking other
changes.
This reverts the whole range 4d0e3984918^..8752c327377,
except for one commit that got reverted in 6f239d7309 already.
(that MR didn't even get its merge commit)
... for x86_64-darwin (into staging-next)
It wouldn't bootstrap otherwise.
Unfortunately we still haven't managed to get the tarballs
on the proper URLs, but GitHub should be reliable enough
and surely almost noone will bootstrap themselves anyway.
This is not the "correct" way to check if a variable is non null in
bash. There is already an instance of the "right" way to do it in
setup.sh. Bash is "generous" enough to accept the original input though.
I couldn't find the relevant shellcheck.
This confused the hell out of me, as I didn't spot the
> The following flags are disabled by default ...
when reading about `pie`, because that sentence was hidden in the
previous hardening flag's section.
Also explain that `pie` hardening is on by default on musl.
Only bash 4+ works in setup.sh. To make sure this is obvious, we can
check BASH_VERSINFO to get the major version number of Bash.
While Bash 3 is pretty rare, it still comes stock in macOS.
We *could* provide a warning here for non-Bash shells, but it’s not
always clear whether they will work or not. Zsh should have no trouble
while busybox sh, fish, or any others. There’s no great way to detect
what feature set the shell supports.
Fixes#71625
This enables the bootstrap stdenv test to specify the actual llvm
of the newly generated build instread of assuming it's the same version
as the current stdenv.
With removeUnknownConfigureFlags, it's impossible to express a package
that needs --enable-static, but will not accept --disable-shared,
without overriding the result of removeUnknownConfigureFlags _again_
in pkgs/top-level/static.nix.
It would be much better (and more in line with the rest of Nixpkgs) if
we encoded changes needed for static builds in package definitions
themselves, rather than in an ever-expanding list in static.nix. This
is especially true when doing it in static.nix is going to require
multiple overrides to express what could be expressed with stdenv
options.
So as a step in that direction, and to fix the problem described
above, here I replace removeUnknownConfigureFlags with a new stdenv
option, dontAddStaticConfigureFlags. With this mechanism, a package
that needs one but not both of the flags just needs to set
dontAddStaticConfigureFlags and then set up configureFlags manually
based on stdenv.hostPlatform.isStatic.
Changes to llvmPackages have caused the `libclang-cpp*.dylib` files to
be included in the `clang-unwrapped.lib` output. So we no longer need to
copy them from libclang.
`TargetConditionals.h` was missing several definitions, like
`TARGET_OS_TV` that are part of SDK 10.12 at least. And one that doesn't
seem to occur in any SDK afaict, `TARGET_OS_EMBEDDED_OTHER`.
I added the definitions from SDK 10.12 verbatim and defined
`TARGET_OS_EMBEDDED_OTHER` to be equal to `0`.
This is a modified version of a patch to avoid a stdenv rebuild.
I was having a hard time testing new bootstrapFiles because
`make-bootstrap-tools.nix` imports `pkgspath` but does not pass anything
but the current system.
This is merely for convenience and I'm not entirely certain it's a
sensible thing to do, maybe generating new bootstrapFiles while
overriding the current bootstrapFiles isn't something you're supposed to
do?
The rpath structure for the bootstrap tools was reworked to minimize
the amount of rewriting required on unpack, but the test was not
updated to match the different structure.
Additionally [1] builds that use the bootstrap version of libc++
cannot find libc++abi if the reference includes the "lib"
component (ie, libc++ refers to libc++abi with
@rpath/lib/libc++abi.dylib).
[1] https://logs.nix.samueldr.com/nix-darwin/2021-05-18#4993282
Test failure observed on Hydra: https://hydra.nixos.org/build/143130126
This will begin the process of breaking up the `useLLVM` monolith. That
is good in general, but I hope will be good for NetBSD and Darwin in
particular.
Co-authored-by: sterni <sternenseemann@systemli.org>
If things build fine with `stdenvNoCC`, let them use that. If tools
might be prefixed, prepare for that, either by directly splicing or just
using the env vars provided by the wrapper setup-hooks.
Co-authored-by: Dmitry Kalinkin <dmitry.kalinkin@gmail.com>
I am taking the non-invasive parts of #110914 to hopefully help out with #111988.
In particular:
- Use `lib.makeScopeWithSplicing` to make the `darwin` package set have
a proper `callPackage`.
- Adjust Darwin `stdenv`'s overlays keeping things from the previous
stage to not stick around too much.
- Expose `binutilsNoLibc` / `darwin.binutilsNoLibc` to hopefully get us
closer to a unified LLVM and GCC bootstrap.
Also begin to start work on cross compilation, though that will have to
be finished later.
The patches are based on the first version of
https://reviews.llvm.org/D99484. It's very annoying to do the
back-porting but the review has uncovered nothing super major so I'm
fine sticking with what I've got.
Beyond making the outputs work, I also strove to re-sync the packages,
as they have been drifting pointlessly apart for some time.
----
Other misc notes, highly incomplete
- lvm-config-native and llvm-config are put in `dev` because they are
tools just for build time.
- Clang no longer has an lld dep. That was introduced in
db29857eb3, but if clang needs help
finding lld when it is used we should just pass it flags / put in the
resource dir. Providing it at build time increases critical path
length for no good reason.
----
A note on `nativeCC`:
`stdenv` takes tools from the previous stage, so:
1. `pkgsBuildBuild`: `(?1, x, x)`
2. `pkgsBuildBuild.stdenv.cc`: `(?0, ?1, x)`
while:
1. `pkgsBuildBuild`: `(?1, x, x)`
2. `pkgsBuildBuild.targetPackages`: `(x, x, ?2)`
3. `pkgsBuildBuild.targetPackages.stdenv.cc`: `(?1, x, x)`
Patch every `derivation` call in the bootsrap process to add it a
conditional `__contentAddressed` parameter.
That way, passing `contentAddressedByDefault` means that the entire
build closure of a system can be content addressed
Adding the hostSuffix to the end of the derivation's name is problematic
since some stuff, including user facing programs like nix-env rely on
the behavior of parseDrvName instead of pname and version.
builtins.parseDrvName currently thinks that the cross compilation target
added via hostSuffix is part of the version. This has the practical
consequence for example that nix-env would think a cross compiled
derivation would be an updated version of a native derivation of the
same package and version — breaking user's profiles.
We can easily prevent this by moving the hostSuffix in between pname and
version. In case name is passed to mkDerivation this is of course not
possible and we are forced to fall back to the old behavior.
This change could serve as a replacement for the migitation we
introduced with the -static appendix to pname in order to avoid
confusion between nix and nixStatic as outlined in the comment added
with this commit.
Support `mainProgram` as an attribute of `meta` for packages.
This is an attribute used by [`nix
run`](https://nixos.org/manual/nix/unstable/command-ref/new-cli/nix3-run.html#description)
to customize the main program of a package.
For example, `pkgs.neovim` provides a `/bin/nvim` executable which users
would (almost certainly) prefer `nix run` to execute instead of a
non-existing `/bin/neovim`.
Signed-off-by: Ana Hobden <operator@hoverbear.org>
`hasCC` was getting overridden in the cross bootstrapping (for GHCJS),
which preventing the default logic from re-triggering for `stdenvNoCC`.
Also remove `stdenv.noCC` which is obseleted by `stdenv.hasCC`.
Add a config field `contentAddressedByDefault` and an associated
environment variable `NIXPKGS_CA_BY_DEFAULT` to make every nixpkgs
derivation content-addressed by default
This was removed in e29b0da9c7, because
it was felt it was ambiguous whether isBSD should remove Darwin.
I think it should be reintroduced. Packages sometimes have their own
concepts of "is BSD" e.g. Lua, and these almost never include Darwin,
so let's keep Darwin excluded.
Without a way to say "is this BSD", one has to list all flavours of
BSD seperately, even though fundamentally they're still extremely
similar. I don't want to have to write the following!
stdenv.isFreeBSD || stdenv.isNetBSD || stdenv.isOpenBSD || stdenv.isDragonFlyBSD
Additionally, we've had stdenv.hostPlatform.isBSD this whole time, and
it hasn't hurt anything.
Unify the logic for constructing the name from pname and version and
modifying the name in case a host suffix needs to appended. This allows
us to modify the construction of name from pname and version without
having to duplicate it in two places.
4e9dc46dea re-enabled hardening for Musl,
which is good.
Though static builds for ARM fail in various ways
- cross armv7l static does not build
- cross aarch64 static produces segfaulting dynamically linked binaries
- native aarch64 static also produces segfaulting dynamically linked binaries
It seems that for native x86_64-linux, static builds are fine though.
This works around the issue by removing PIE from the hardening flags,
keeping all other hardening flags. This is an improvement (I think) from
before 4e9dc46d.
Fixes#114953
* stdenv/check-meta: change to allowlist and blocklist
* Update pkgs/stdenv/generic/check-meta.nix
Co-authored-by: Graham Christensen <graham@grahamc.com>
Since the deprecation is fairly recent, we should warn by default.
Also fix the wording of the comment: stdenv.lib will be removed for the 21.11
release, not just deprecated (as it already is deprecated).
The `platform` field is pointless nesting: it's just stuff that happens
to be defined together, and that should be an implementation detail.
This instead makes `linux-kernel` and `gcc` top level fields in platform
configs. They join `rustc` there [all are optional], which was put there
and not in `platform` in anticipation of a change like this.
`linux-kernel.arch` in particular also becomes `linuxArch`, to match the
other `*Arch`es.
The next step after is this to combine the *specific* machines from
`lib.systems.platforms` with `lib.systems.examples`, keeping just the
"multiplatform" ones for defaulting.
Increase schedulingPriority of the bootstrap tools to unblock the
nixpkgs-unstable channel.
The channel is repeatedly blocked by the makeBootstrapTools job for
aarch64. The cause is lack of resources.
By increasing the priority, it should become the first job Hydra would
build, allowing the channel to advance quicker. Of course, it does mean
that while the channel advances, nothing else has been built.
This should be a temporary solution until we have more capacity for
aarch64.
By exporting it, we always make the new directories available
to subprocesses, regardless of whether the environment
variable existed before `nix-shell` was invoked.
This avoids the scenario where strictDeps is off and cross-compiled
XDG_DATA_DIRS content is brought into the environment.
While probably harmless for data like manpages and completion scripts,
this would cause issues when XDG_DATA_DIRS is used to find executables
or plugins. The Qt framework is known to behave like this and might
have run into incompatibilities.
XDG_DATA_DIRS is to /share as PATH is to /bin.
It was defined as part of the XDG basedir specification.
https://specifications.freedesktop.org/basedir-spec/basedir-spec-latest.html
While it originated from the X Desktop Group, it is not limited to
the X11 ecosystem, as evidenced by its use in bash-completion.
The removal of ` && -d "$pkg/bin"` is ok, because this optimization is
already performed by `addToSearchPath`.
The dynamic loader on powerpc64 is called ld64.so.2 rather than
ld-linux.so.*, and was not matched by the existing pattern.
We reuse the dynamicLinker name from binutils to match a wider set
of platforms and to avoid specifying this information in two places.
Build the llvm support libraries (libcxx, libcxxabi) from scratch
without using the existing llvm libraries. This is the same spirit and
similar implementation as the "useLLVM" bootstrap in llvm package
sets. Critically it avoids having libcxxabi provided by the cc-wrapper
when building libcxx, which otherwise results in two libcxxabi
instances.
$ otool -L /nix/store/vd4vvgs9xngqbjzpg3qc41wl6jh42s9i-libc++-7.1.0/lib/libc++.dylib
/nix/store/vd4vvgs9xngqbjzpg3qc41wl6jh42s9i-libc++-7.1.0/lib/libc++.dylib:
/nix/store/vd4vvgs9xngqbjzpg3qc41wl6jh42s9i-libc++-7.1.0/lib/libc++.1.0.dylib (compatibility version 1.0.0, current version 1.0.0)
/nix/store/gmpwk5fyp3iasppqrrdpswxvid6kcp8r-libc++abi-7.1.0/lib/libc++abi.dylib (compatibility version 1.0.0, current version 1.0.0)
/nix/store/3hn7azynqgp2pm5gpdg45gpq0ia72skg-libc++abi-7.1.0/lib/libc++abi.dylib (compatibility version 1.0.0, current version 1.0.0)
/nix/store/1nq94scbxs6bk7pimqhvz76q6cfmbv97-Libsystem-osx-10.12.6/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 1226.10.1)
Additionally move some utilities (clang, binutils, coreutils, gnugrep)
to the stage layers so they can be replaced before the final
stdenv. This should cause most of stage4 to be built from the
toolchain assembled as of stage3 instead of the bootstrap toolchain.
This new version has tapi support, which is needed to build the new
stubs based libSystem, etc. and Big Sur support.
You can verify the provenance of these yourself by checking Hydra here:
https://hydra.nixos.org/build/128192471
This adds -frandom-seed to each compiler invocation in stdenv. The
object here is to make the compierl invocations produce the same output
every time they are called (for the same derivation). When the
-frandom-seed option is not set the compiler will use a combination of
random numbers (in GCC's case from /dev/urandom) and the durrent time to
produce a "random" input per file. This can (among other things) lead to
different ordering of symbols in the produced object files.
For reason of reproducibility we prefer having the same derivation
produce the exact same outputs. This is not a silver bullet but one way
to tame the compiler.
I made a mistake merge. Reverting it in c778945806 undid the state
on master, but now I realize it crippled the git merge mechanism.
As the merge contained a mix of commits from `master..staging-next`
and other commits from `staging-next..staging`, it got the
`staging-next` branch into a state that was difficult to recover.
I reconstructed the "desired" state of staging-next tree by:
- checking out the last commit of the problematic range: 4effe769e2
- `git rebase -i --preserve-merges a8a018ddc0` - dropping the mistaken
merge commit and its revert from that range (while keeping
reapplication from 4effe769e2)
- merging the last unaffected staging-next commit (803ca85c20)
- fortunately no other commits have been pushed to staging-next yet
- applying a diff on staging-next to get it into that state
defaultHardeningFlags is set to enable pie for Musl, but is not
actually used because the default is never put into
NIX_HARDENING_ENABLE. That still works for cases other than Musl
only because NIX_HARDENING_ENABLE is defaulted in the binutils and
cc-wrapper setup-hook.sh scripts.
This hook moves systemd user service file from `lib/systemd/user` to
`share/systemd/user`. This is to allow systemd to find the user
services when installed into a user profile. The `lib/systemd/user`
path does not work since `lib` is not in `XDG_DATA_DIRS`.
