Add the log-formats `multiline` and `multiline-with-logs` which offer
multiple current active building status lines.
Change-Id: Idd8afe62f8591b5d8b70e258c5cefa09be4cab03
Unfetched submodules are included as empty directories in archives, so they end
up as such in the store when fetched in clean mode. Make sure the same happens
in dirty mode too. Fortunately, they are already correctly represented in the
ls-files output, so we just need to make sure to include the empty directory in
our filter.
Fixes: https://github.com/NixOS/nix/issues/6247
Change-Id: I60d06ff360cfa305d081b920838c893c06da801c
Previously, system call filtering (to prevent builders from storing files with
setuid/setgid permission bits or extended attributes) was performed using a
blocklist. While this looks simple at first, it actually carries significant
security and maintainability risks: after all, the kernel may add new syscalls
to achieve the same functionality one is trying to block, and it can even be
hard to actually add the syscall to the blocklist when building against a C
library that doesn't know about it yet. For a recent demonstration of this
happening in practice to Nix, see the introduction of fchmodat2 [0] [1].
The allowlist approach does not share the same drawback. While it does require
a rather large list of harmless syscalls to be maintained in the codebase,
failing to update this list (and roll out the update to all users) in time has
rather benign effects; at worst, very recent programs that already rely on new
syscalls will fail with an error the same way they would on a slightly older
kernel that doesn't support them yet. Most importantly, no unintended new ways
of performing dangerous operations will be silently allowed.
Another possible drawback is reduced system call performance due to the larger
filter created by the allowlist requiring more computation [2]. However, this
issue has not convincingly been demonstrated yet in practice, for example in
systemd or various browsers.
This commit tries to keep the behavior as close to unchanged as possible. Only
newer syscalls that are not supported by glibc 2.38 (as found in NixOS 23.11)
are blocked. Since this includes fchmodat2, the compatibility code added for
handling this syscall can be removed too.
[0] https://github.com/NixOS/nixpkgs/issues/300635
[1] https://github.com/NixOS/nix/issues/10424
[2] https://github.com/flatpak/flatpak/pull/4462#issuecomment-1061690607
Change-Id: I541be3ea9b249bcceddfed6a5a13ac10b11e16ad
In f047e4357b, I missed the behavior that if
building without a dedicated build user (i.e. in single-user setups), seccomp
setup failures are silently ignored. This was introduced without explanation 7
years ago (ff6becafa8). Hopefully the only
use-case nowadays is causing spurious test suite successes when messing up the
seccomp filter during development. Let's try removing it.
Change-Id: Ibe51416d9c7a6dd635c2282990224861adf1ceab
This merge commit returns to the previous state prior to the release but leaves the tag in the branch history.
Release created with releng/create_release.xsh
Change-Id: I92296a1746b54a081004fe2bb23e9e37fd33b3e5
* changes:
releng: add sha256 for the manual tarball
releng: fix upload of multiarch images to forgejo
releng: fix git checking
releng: fix logging inside interactive xonsh
releng: support multiple systems
version: update to 2.90.0-rc1
Forgejo appears to immediately delete registry content that is
overwritten. This means that we are forced to delete our previous
workaround of making a temporary tag and use a new, more absurd
workaround of making an entire temporary image that we basically only
need to create to get its hash.
However, on the plus side, the new workaround doesn't create garbage
tags to begin with, which means that we don't have to deal with GitHub
not implementing the standardized tag delete endpoint and instead
only implementing a proprietary one.
Upstream-Bug: https://github.com/containers/skopeo/issues/2354
Change-Id: I220e7ce9a17fd230c38882f12c009a166dcc9336
I don't know when this broke, it seems like it happened since the 24.05
upgrade, so xonsh 0.15.
What happened is that xonsh was trying to intercept log output, which
explodes if you have the logger survive past one command input. This is,
however, impossible to avoid if you are trying to use logging when you
import releng from inside xonsh for interactive use!
The error below is because the memory handler backing the stdout/stderr
of the one command that's just been run was closed after the command
completed.
Change-Id: I2be642aebf93da9818d08ff8b97c2e72ba5ac581
--- Logging error ---
Traceback (most recent call last):
File "/nix/store/7hnr99nxrd2aw6lghybqdmkckq60j6l9-python3-3.11.9/lib/python3.11/logging/__init__.py", line 1113, in emit
stream.write(msg + self.terminator)
File "/nix/store/34951j60xcsw6zj4v8lsaf491acv0by3-python3-3.11.9-env/lib/python3.11/site-packages/xonsh/base_shell.py", line 183, in write
self.mem.write(s)
ValueError: I/O operation on closed file.
Call stack:
File "/nix/store/xgdp1p1gv8ni1awnkzyqasnn6gz5wlvx-xonsh-0.15.1/bin/xonsh", line 8, in <module>
sys.exit(main())
File "/nix/store/34951j60xcsw6zj4v8lsaf491acv0by3-python3-3.11.9-env/lib/python3.11/site-packages/xonsh/main.py", line 470, in main
sys.exit(main_xonsh(args))
File "/nix/store/34951j60xcsw6zj4v8lsaf491acv0by3-python3-3.11.9-env/lib/python3.11/site-packages/xonsh/main.py", line 514, in main_xonsh
shell.shell.cmdloop()
File "/nix/store/34951j60xcsw6zj4v8lsaf491acv0by3-python3-3.11.9-env/lib/python3.11/site-packages/xonsh/ptk_shell/shell.py", line 406, in cmd
loop
line = self.singleline(auto_suggest=auto_suggest)
File "/nix/store/34951j60xcsw6zj4v8lsaf491acv0by3-python3-3.11.9-env/lib/python3.11/site-packages/xonsh/ptk_shell/shell.py", line 374, in sin
gleline
line = self.prompter.prompt(**prompt_args)
File "/nix/store/34951j60xcsw6zj4v8lsaf491acv0by3-python3-3.11.9-env/lib/python3.11/site-packages/prompt_toolkit/shortcuts/prompt.py", line 1
026, in prompt
return self.app.run(
File "/nix/store/34951j60xcsw6zj4v8lsaf491acv0by3-python3-3.11.9-env/lib/python3.11/site-packages/prompt_toolkit/application/application.py",
line 1002, in run
return asyncio.run(coro)
File "/nix/store/7hnr99nxrd2aw6lghybqdmkckq60j6l9-python3-3.11.9/lib/python3.11/asyncio/runners.py", line 189, in run
with Runner(debug=debug) as runner:
File "/nix/store/7hnr99nxrd2aw6lghybqdmkckq60j6l9-python3-3.11.9/lib/python3.11/asyncio/runners.py", line 59, in __enter__
self._lazy_init()
File "/nix/store/7hnr99nxrd2aw6lghybqdmkckq60j6l9-python3-3.11.9/lib/python3.11/asyncio/runners.py", line 137, in _lazy_init
self._loop = events.new_event_loop()
File "/nix/store/7hnr99nxrd2aw6lghybqdmkckq60j6l9-python3-3.11.9/lib/python3.11/asyncio/events.py", line 810, in new_event_loop
return get_event_loop_policy().new_event_loop()
File "/nix/store/7hnr99nxrd2aw6lghybqdmkckq60j6l9-python3-3.11.9/lib/python3.11/asyncio/events.py", line 699, in new_event_loop
return self._loop_factory()
File "/nix/store/7hnr99nxrd2aw6lghybqdmkckq60j6l9-python3-3.11.9/lib/python3.11/asyncio/unix_events.py", line 64, in __init__
super().__init__(selector)
File "/nix/store/7hnr99nxrd2aw6lghybqdmkckq60j6l9-python3-3.11.9/lib/python3.11/asyncio/selector_events.py", line 54, in __init__
logger.debug('Using selector: %s', selector.__class__.__name__)
Message: 'Using selector: %s'
Arguments: ('EpollSelector',)
Change-Id: I90959809129aaf96aad4577599031688599ed85e
This is motivated by flakes being bad and all the stuff that calls
things by "system" being utterly unable to cope with cross compilation.
So if we go shove it in package.nix it is suddenly usable from cross
contexts.
Usage:
```
nix build -L .#nix-riscv64-linux.binaryTarball
```
Change-Id: I702ebf2ac5bd9d1c57662f968b000073134df336
It builds. I have not tested the binaries since I don't have hardware,
but I would be rather surprised if it were broken, given that nix *runs*
on this platform.
Change-Id: I0b474ffcd4a431bf117a303d0b65fa6532113f48
There were two bugs I found:
1. If the build isn't already done in the store, nix-store --realise
does not know how to build it. You have to just give it the
derivation and I guess it will realise all outputs, which is fine.
2. cp without -T will not overwrite an existing manual directory,
creating a path manual/manual.
Change-Id: Ibebfd136a266da5330944a985e636ebb776f1909
This seems to have been caused by having the wrong PID. I don't know why
it worked before in the sandbox, but the code was definitely wrong
before, so let's just fix it.
Change-Id: I556580bdf614c716566310e975a36daa6d6c9a91
This was originally going to be just the testsuite but I kinda just
documented all of them.
I am tired of us not documenting these. This is a starting point to
producing an actually good index. I would like to enforce it in a
pre-commit hook eventually that we document all environment variables
used in Lix itself, even if it is terse dev facing docs.
This is full of a bunch of TODOs caused by auditing code. They should
probably be done at some point.
Change-Id: I7c0d3b257e19bae23d47d1efbd7361d203bccb0e
It's in the security section, and it was totally outdated anyway.
I took the opportunity to write down the stuff we already believed.
Change-Id: I73e62ae85a82dad13ef846e31f377c3efce13cb0
Followup to https://gerrit.lix.systems/c/lix/+/1417 to ensure that this
parser will never take something that doesn't look like a version.
It turns out this problem is less alarming than initially thought
because it only applies to the testsuite in a non-default mode.
Change-Id: I26aba24aaf0215f2b782966314b94784db766266
We should not let these regress in CI by having broken dependencies or
similar. Still need to fix the evaluation error checking in
buildbot-nix, but this is a useful step regardless.
Fixes: lix-project/lix#383
Change-Id: I3883184165440e66256c989117f2ab2e54c3aafd
-- message from cl/1418 --
The boehmgc changes are bundled into this commit because doing otherwise
would require an annoying dance of "adding compatibility for < 8.2.6 and
>= 8.2.6" then updating the pin then removing the (now unneeded)
compatibility. It doesn't seem worth the trouble to me given the low
complexity of said changes.
Rebased coroutine-sp-fallback.diff patch taken from https://github.com/NixOS/nixpkgs/pull/317227
-- jade resubmit changes --
This is a resubmission of https://gerrit.lix.systems/c/lix/+/1418, which
was reverted in https://gerrit.lix.systems/c/lix/+/1432 for breaking CI
evaluation without being detected.
I have run `nix flake check -Lv` on this one before submission and it
passes on my machine and crucially without eval errors, so the CI result
should be accurate.
It seems like someone renamed forbiddenDependenciesRegex to
forbiddenDependenciesRegexes in nixpkgs and also changed the type
incompatibly. That's pretty silly, but at least it's just an eval error.
Also, `xonsh` regressed the availability of `xonsh-unwrapped`, but it
was fixed by us in https://github.com/NixOS/nixpkgs/pull/317636, which
is now in our channel, so we update nixpkgs compared to the original
iteration of this to simply get that.
We originally had a regression related to some reorganization of the
nixpkgs lib test suite in which there was broken parameter passing.
This, too, we got quickfixed in nixpkgs, so we don't need any changes
for it: https://github.com/NixOS/nixpkgs/pull/317772
Related: https://gerrit.lix.systems/c/lix/+/1428
Fixes: lix-project/lix#385
Change-Id: I26d41ea826fec900ebcad0f82a727feb6bcd28f3
The libcmd unit test creates files (more specifically, the fetcher cache) in
its home directory. In the single-user sandbox, this leads to the creation of
/homeless-shelter, since this is the default HOME and the root is writable.
Unfortunately, this conflicts with the assumption of the functional tests that
this directory does not exist. Use a different home directory to prevent these
test failures, and thus restore the ability to build inside the single-user
sandbox.
Fixes: lix-project/lix#365
Change-Id: I4df8c53d043234b95a7c0ac45fc5ee89e8d46aff
* changes:
releng: add prod environment, ready for release
releng: automatically figure out if we should tag latest for docker
releng: support multiarch docker images
manual: rewrite the docker guide now that we have images
Rewrite docker to be sensible and smaller
Implement docker upload in the releng tools
I am *reasonably* confident that this releng infrastructure can actually
build a Lix 2.90 and release it successfully. Let's make it possible to
do, and add some cute colours to the confirmation message.
Change-Id: I85e498b6fb49ffc5e75c0a72c5e45fb1f69030d3
For example, when releasing from release-2.90, if `main` has a 2.91 tag
ancestor, we know that 2.91 was released, so we should *not* tag latest.
Change-Id: Ia56b17a2ee03bbec74b7c271c742858c690d450d
If we don't want to have separate registry tags by architecture (EWWWW),
we need to be able to build multiarch docker images. This is pretty
simple, and just requires making a manifest pointing to each of the
component images.
I was *going* to just do this API prodding with manifest-tool, but it
doesn't support putting metadata on the outer manifest, which is
actually kind of a problem because it then doesn't render the metadata
on github. So I guess we get a simple little containers API
implementation that is 90% auth code.
Change-Id: I8bdd118d4cbc13b23224f2fb174b232432686bea
I have checked the image can build things and inspected `diff -ru`
compared to the old image. As far as I can tell it is more or less
the same besides the later git change.
Layers are now 65MB or less, and we aren't against the maxLayers limit
for the broken automatic layering to do anything but shove one store
path in a layer (which is good behaviour, actually).
This uses nix2container which streams images, so the build time is much
shorter.
I have also taken the opportunity to, in addition to fixing the 400MB
single layer (terrible, and what motivated this in the first place),
delete about 200MB of closure size inflicted by git vs gitMinimal
causing both perl and python to get into closure.
People mostly use this thing for CI, so I don't really think you need
advanced git operations, and large git can be added at the user side if
really motivated.
With love for whichever container developer somewhat ironically assumed
that one would not run skopeo in a minimal container that doesn't have a
/var/tmp.
Fixes: lix-project/lix#378
Change-Id: Icc3aa20e64446276716fbbb87535fd5b50628010
The boehmgc changes are bundled into this commit because doing otherwise
would require an annoying dance of "adding compatibility for < 8.2.6 and
>= 8.2.6" then updating the pin then removing the (now unneeded)
compatibility. It doesn't seem worth the trouble to me given the low
complexity of said changes.
Rebased coroutine-sp-fallback.diff patch taken from https://github.com/NixOS/nixpkgs/pull/317227
Change-Id: I8c590e9fe25c0f566d0cfeacb96d8cf50abf12e8
4b128008c5d9fde881ce1b0a25e60ae0415a14d5 in nixpkgs introduced a default
hashedPasswordFile for root in NixOS tests, which takes precedence over
the password option set in the nix-copy test.
Change-Id: Iffaebec5992e50614b854033f0d14312c8d275b5
Since ad8a4b380e, the version printer returns "nix (Lix, like Nix) 2.x",
hence the `daemonVersion` was being set to the string "like".
Using `compareVersions` with a letter compares them lexicographically:
builtins.compareVersions "like" "2.12pre20230103" // => -1
builtins.compareVersions "like" "2.16.0" // => -1
This caused that `isDaemonNewer` always returned 1, falsy in Bash terms.
Therefore, the test suite skipped those tests where they use it.
Fixes lix-project/lix#324
Change-Id: If6682515bf0bf8b8add641af9a4e98b50a9acb51
This can release x86_64-linux binaries to staging, with ephemeral keys.
I think it's good enough to review at least at this point, so we don't
keep adding more stuff to it to make it harder to review.
Change-Id: Ie95e8f35d1252f5d014e819566f170b30eda152e
kloenk
jade
Pierre Bourdon
Qyriad
Mario Rodas