Commit graph

16109 commits

Author SHA1 Message Date
eldritch horrors 6491cde997 resolve-system-dependencies: remove entirely
this hasn't been used since 2020, and hasn't been compiled since may.

Change-Id: I865550966630eee6ba18d742ba36f0a90901279d
2024-08-09 11:33:09 +00:00
jade 3b902683e9 Merge changes I0373ac01,I7b543967,I537103eb into main
* changes:
  releng: fix the git push
  releng: clarify/update docs, add instructions after tag
  Fix is_maintenance_branch heuristic
2024-08-08 23:12:11 +00:00
jade 9682ab4f38 Merge changes I6358a393,I2d9f276b,Idd096dc9 into main
* changes:
  clang-tidy: write a lint for charptr_cast
  tree-wide: automated migration to charptr_cast
  clang-tidy: enforce the new rules
2024-08-08 23:09:30 +00:00
jade 757041c3e7 Merge changes I526cceed,Ia4e2f1fa,I22e66972,I9fbd55a9,Ifca22e44 into main
* changes:
  sqlite: add a Use::fromStrNullable
  util: implement charptr_cast
  tree-wide: fix a pile of lints
  refactor: make HashType and Base enum classes for type safety
  build: integrate clang-tidy into CI
2024-08-08 22:43:10 +00:00
jade a5f0954c29 clang-tidy: write a lint for charptr_cast
This lets us ensure that nobody is putting in new reinterpret_cast
instances where they could safely use charptr_cast instead.

Change-Id: I6358a3934c8133c7150042635843bdbb6b9218d4
2024-08-08 14:53:17 -07:00
jade 4ed8461cac sqlite: add a Use::fromStrNullable
There were several usages of the raw sqlite primitives along with C
style casts, seemingly because nobody thought to use an optional for
getting a string or NULL.

Let's fix this API given we already *have* a wrapper.

Change-Id: I526cceedc2e356209d8fb62e11b3572282c314e8
2024-08-08 14:53:17 -07:00
jade a85c4ce535 tree-wide: automated migration to charptr_cast
The lint did it :3

Change-Id: I2d9f276b01ebbf14101de4257ea13e44ff6fe0a0
2024-08-08 14:53:17 -07:00
jade a318c96851 util: implement charptr_cast
I don't like having so many reinterpret_cast statements that have to
actually be looked at to determine if they are UB. A huge number of the
reinterpret_cast instances in Lix are actually casting to some pointer
of some character type, which is always valid no matter the source type.

However, it is also worth looking at if it is not casting both *from* a
character type and also *to* a character type, since IMO splatting a
struct into a character array should be a very deliberate action instead
of just being about dealing with bad APIs.

So let's write a template that encapsulates this invariant so we can
not worry about the trivially safe reinterpret_cast invocations.

Change-Id: Ia4e2f1fa0c567123a96604ddadb3bdd7449660a4
2024-08-08 14:53:17 -07:00
jade c1291fd102 clang-tidy: enforce the new rules
Fixes: lix-project/lix#241

Change-Id: Idd096dc9ca92ffd4be8c22d293ba5bf2ec48a85f
2024-08-08 14:53:17 -07:00
jade e34833c025 tree-wide: fix a pile of lints
This:
- Converts a bunch of C style casts into C++ casts.
- Removes some very silly pointer subtraction code (which is no more or
  less busted on i686 than it began)
- Fixes some "technically UB" that never had to be UB in the first
  place.
- Makes finally follow the noexcept status of the inner function. Maybe
  in the future we should ban the function from not being noexcept, but
  that is not today.
- Makes various locally-used exceptions inherit from std::exception.

Change-Id: I22e66972602604989b5e494fd940b93e0e6e9297
2024-08-08 14:53:17 -07:00
jade 370ac940dd refactor: make HashType and Base enum classes for type safety
Change-Id: I9fbd55a9d50464a56fe11cb42a06a206914150d8
2024-08-08 14:53:17 -07:00
jade f3ef0899c7 build: integrate clang-tidy into CI
This still has utterly unacceptably bad output format design that I
would not inflict on anyone I like, but it *does* now exist, and you
*can* find the errors in the log.

Future work would obviously be to fix that and integrate the actual
errors into Gerrit using codechecker or so.

Followup issue: lix-project/lix#457

Fixes: lix-project/lix#147
Change-Id: Ifca22e443d357762125f4ad6bc4f568af3a26c62
2024-08-08 14:53:17 -07:00
piegames e03cd8b3a6 Merge "libexpr: Add experimental pipe operator" into main 2024-08-08 18:15:21 +00:00
eldritch horrors a957219df2 libstore: make Worker::waitForInput private
Change-Id: I71a42acd5a4a9a18b55cf754cdf9896614134398
2024-08-08 12:02:17 +00:00
eldritch horrors ba85e501ce libstore: make Worker status flags private
Change-Id: I16ec8994c6448d70b686a2e4c10f19d4e240750d
2024-08-08 12:02:17 +00:00
eldritch horrors fc987b4123 libstore: remove Goal::addWaitee
Change-Id: I1b00d1a537d84790878cb0e81aaa1cbaa143d62d
2024-08-08 12:02:17 +00:00
eldritch horrors 4c3010a1be libstore: make Worker::wakeUp private
Change-Id: Iffa55272fe6ef4adaf3e9d4d25e5339792c2e460
2024-08-08 12:02:17 +00:00
eldritch horrors 3ecb46e3e7 libstore: make Worker::waitForAWhile private
Change-Id: I0cdcd436ee71124ca992b4f4fe307624a25f11e9
2024-08-08 12:02:17 +00:00
eldritch horrors b33c969519 libstore: make Worker::waitForBuildSlot private
Change-Id: I02a54846cd65622edbd7a1d6c24a623b4a59e5b3
2024-08-08 12:02:17 +00:00
piegames 28ae24f3f7 libexpr: Add experimental pipe operator
The |> operator is a reverse function operator with low binding strength
to replace lib.pipe. Implements RFC 148, see the RFC text for more
details. Closes #438.

Change-Id: I21df66e8014e0d4dd9753dd038560a2b0b7fd805
2024-08-08 11:13:53 +02:00
jade 7246c2d104 releng: fix the git push
This was broken because gerrit requires that the revision actually
is known before it is pushed as a tag.

Also, arguably this fixes the original problem mentioned in
lix-project/lix#439

Change-Id: I0373ac01584440f18d32b8da5699bb359cc2c89a
2024-08-07 21:46:44 -07:00
jade 83247b1c38 releng: clarify/update docs, add instructions after tag
This is not a proper fix for the confusion that can happen about how the
tags are supposed to be used.

For a proper fix, we need to do
lix-project/lix#439 and implement
worktrees such that the user never sees the git state anymore.

Change-Id: I7b543967f522cede486e42684b48cad47da95429
2024-08-07 20:52:09 -07:00
jade 8a86f38bca Fix is_maintenance_branch heuristic
This was broken because Nix language's version comparison does not know
how to deal with versions like -rc1 and considers them newer, which is
in this case not desirable.

That in turn led to not tagging 2.90.0 docker images as "latest" since
the heuristic was wrong.

This commit also adds some more cross-checking and failsafes in case the
person running releng does not have a local main branch that is up to
date.

Fixes: lix-project/lix#443
Change-Id: I537103ebab58ae978c00e06972abe14432dd9c80
2024-08-07 20:14:45 -07:00
Max “Goldstein” Siling 6fdb47f0b2 Merge "src/libcmd/repl.cc: allow :log /path/to/store.drv" into main 2024-08-07 21:48:01 +00:00
jade 0800a81a95 Merge "oops: fix warning about catching polymorphic exception" into main 2024-08-07 19:06:54 +00:00
piegames ec7552ff74 libexpr/parser: Test experimental features
Currently, the parser relies on the global experimental feature flags.
In order to properly test conditional language features, we instead need
to pass it around in the parser::State.

This means that the parser cannot cache the result of isEnabled anymore,
which wouldn't necessarily hurt performance if the function didn't
perform a linear search on the list of enabled features on every single
call. While we could simply evaluate once at the start of parsing and
cache the result in the parser state, the more sustainable solution
would be to fix `isEnabled` such that all callers may profit from the
performance improvement.

Change-Id: Ic9b9c5d882b6270e1114988b63e6064d36c25cf2
2024-08-07 13:07:50 +00:00
Max “Goldstein” Siling 9adfd9b8ad
src/libcmd/repl.cc: allow :log /path/to/store.drv
This adds a second form to the `:log` command: it now can accept a
derivation path in addition to a derivation expression. As derivation
store paths start with `/nix/store`, this is not ambiguous.

Resolves: lix-project/lix#51
Change-Id: Iebc7b011537e7012fae8faed4024ea1b8fdc81c3
2024-08-07 15:58:44 +03:00
Maximilian Bosch 27a63db710 Merge "fix: warn and document when advanced attributes will have no impact due to __structuredAttrs" into main 2024-08-07 10:38:39 +00:00
alois31 780998f4ea Merge "package: improve support for building without BDW-GC" into main 2024-08-07 07:07:28 +00:00
jade d280e4990c oops: fix warning about catching polymorphic exception
This was introduced in I0fc80718eb7e02d84cc4b5d5deec4c0f41116134 and
unnoticed since it only appears in gcc builds.

Change-Id: I1de80ce2a8fab63efdca7ca0de2a302ceb118267
2024-08-06 22:45:19 -07:00
jade 529eed74c4 Merge changes I0fc80718,Ia182b86f,I355f82cb,I8a9b58fa,Id89f8a1f, ... into main
* changes:
  tree-wide: fix various lint warnings
  flake & doxygen: update tagline
  nix flake metadata: print modified dates for input flakes
  cli: eat terminal codes from stdout also
  Implement forcing CLI colour on, and document it better
  manual: fix a syntax error in redirects.js that made it not do anything
  misc docs/meson tidying
  build: implement clang-tidy using our plugin
2024-08-07 00:50:30 +00:00
alois31 2c48460850
libstore/linux: precompile and cache the seccomp BPF
The growth of the seccomp filter in 127ee1a101
made its compilation time significant (roughly 10 milliseconds have been
measured on one machine). For this reason, it is now precompiled and cached in
the parent process so that this overhead is not hit for every single build. It
is still not optimal when going through the daemon, because compilation still
happens once per client, but it's better than before and doing it only once for
the entire daemon requires excessive crimes with the current architecture.

Fixes: lix-project/lix#461
Change-Id: I2277eaaf6bab9bd74bbbfd9861e52392a54b61a3
2024-08-06 19:10:33 +02:00
alois31 403fa9e2b6
libstore/linux: compile the seccomp BPF explicitly
This is a preparation for precompiling the filter, which is done separately.
The behaviour should be unchanged for now.

Change-Id: I899aa7242962615949208597aca88913feba1cb8
2024-08-06 18:31:40 +02:00
alois31 741d3b441c
libstore: add LocalDerivationGoal setupSyscallFilter hook
The seccomp setup code was a huge chunk of conditionally compiled
platform-specific code. For this reason, it is appropriate to move it to the
platform-specific implementation file. Ideally its setup could be moved a bit
to make it happen at the same place as the Darwin restrictions, but that change
is going to be less mechanical.

Change-Id: I496aa3c4fabf34656aba1e32b0089044ab5b99f8
2024-08-06 18:27:09 +02:00
alois31 f84997cbef
package: don't hide system-wide manual pages
When MANPATH is unset or contains an empty component, a reasonable default is
used. Previously (after 3dced96741), when MANPATH
was unset, the shell hook would only place a location containing the Lix manual
pages there, and system-wide manual pages would become unavailable in the
development shell, which is undesired. Fix the issue by including an empty
component in this case.

Change-Id: Ib3c67a831d709fe2a87520e15917eebb59397bd1
2024-08-06 17:18:05 +02:00
jade ca9d3e6e00 tree-wide: fix various lint warnings
Change-Id: I0fc80718eb7e02d84cc4b5d5deec4c0f41116134
2024-08-04 20:55:45 -07:00
jade 9238e62ae6 flake & doxygen: update tagline
This tagline was left over from CppNix and we should make it tastier.

Change-Id: Ia182b86f6e751591be71a50521992ad73c7b38b5
2024-08-04 20:41:19 -07:00
jade bd1344ec54 nix flake metadata: print modified dates for input flakes
This was always in the lock file and we can simply actually print it.

The test for this is a little bit silly but it should correctly
control for my daring to exercise timezone code *and* locale code in a
test, which I strongly suspect nobody dared do before.

Sample (abridged):
```
Path:          /nix/store/gaxb42z68bcr8lch467shvmnhjjzgd8b-source
Last modified: 1970-01-01 00:16:40
Inputs:
├───flake-compat: github:edolstra/flake-compat/0f9255e01c2351cc7d116c072cb317785dd33b33
│   Last modified: 2023-10-04 13:37:54
├───flake-utils: github:numtide/flake-utils/b1d9ab70662946ef0850d488da1c9019f3a9752a
│   Last modified: 2024-03-11 08:33:50
│   └───systems: github:nix-systems/default/da67096a3b9bf56a91d16901293e51ba5b49a27e
│       Last modified: 2023-04-09 08:27:08
```

Change-Id: I355f82cb4b633974295375ebad646fb6e2107f9b
2024-08-04 20:41:19 -07:00
jade 5f0ef50077 cli: eat terminal codes from stdout also
This *should* be sound, plus or minus the amount that the terminal code
eating code is messed up already.

This is useful for testing CLI output because it will strip the escapes
enough to just shove the expected output in a file.

Change-Id: I8a9b58fafb918466ac76e9ab585fc32fb9294819
2024-08-04 20:41:19 -07:00
jade 378ec5fb06 Implement forcing CLI colour on, and document it better
This is necessary to make some old tests work when testing colour
against non-interactive outputs.

Change-Id: Id89f8a1f45c587fede35a69db85f7a52f2c0a981
2024-08-04 20:41:19 -07:00
jade 700762d8b2 manual: fix a syntax error in redirects.js that made it not do anything
lol lmao

Let's put in a syntax checker in CI so we do not have to deal with this
nonsense ever again.

Change-Id: I0fe875e0cfc59ab1783087762e5bb07e09ded105
2024-08-04 20:41:19 -07:00
jade 0f998056fa misc docs/meson tidying
The docs page has an incorrect escape that leads to a backslash
appearing in output. Meson stuff is self-explanatory, just shortens and
simplifies a bit.

Change-Id: Ib63adf934efd3caeb82ca82988f230e8858a79f9
2024-08-04 20:41:19 -07:00
jade 3daeeaefb1 build: implement clang-tidy using our plugin
The principle of this is that you can either externally build it with
Nix (actual implementation will be in a future commit), or it can be
built with meson if the Nix one is not passed in.

The idea I have is that dev shells don't receive the one from Nix to
avoid having to build it, but CI can use the one from Nix and save some
gratuitous rebuilds.

The design of this is that you can run `ninja -C build clang-tidy` and
it will simply correctly clang-tidy the codebase in spite of PCH
bullshit caused by the cc-wrapper.

This is a truly horrendous number of hacks in a ball, caused by bugs in
several pieces of software, and I am not even getting started.

I don't consider this to fix the clang-tidy issue filing, since we still
have a fair number of issues to fix even on the existing minimal
configuration, and I have not yet implemented it in CI. Realistically we
will need to do something like https://github.com/Ericsson/codechecker
to be able to silence warnings without physically touching the code, or
at least *diff* reports between versions.

Also, the run-clang-tidy output design is rather atrocious and must
not be inflicted upon anyone I have respect for, since it buries the
diagnostics in a pile of invocation logs. We would do really well to
integrate with the Gerrit SARIF stuff so we can dump the reports on
people in a user-friendly manner.

Related: lix-project/lix#147

Change-Id: Ifefe533f3b56874795de231667046b2da6ff2461
2024-08-04 20:41:19 -07:00
Tom Bereknyei 7fc481396c
fix: warn and document when advanced attributes will have no impact due to __structuredAttrs
Backport of https://github.com/NixOS/nix/pull/10884.

Change-Id: I82cc2794730ae9f4a9b7df0185ed0aea83efb65a
2024-08-03 13:32:51 +02:00
alois31 58758c0f87
package: improve support for building without BDW-GC
Expose an option for disabling the BDW-GC build dependency entirely. Fix the
place where one of its headers was included (unnecessarily) without proper
guarding. Finally, use this machinery to exclude BDW-GC from the ASAN builds
entirely (its usage has already been disabled due to compatibility issues
anyway), to ensure this configuration is not regressed again.

Change-Id: I2ebe8094abf67e7d1e99eed971de3e99d071c10b
2024-08-03 06:14:41 +02:00
eldritch horrors 66469fc281 libstore: move Goal::waiteeDone into Worker::goalFinished
this begins a long and arduous journey to remove all result state from
Goal, to eventually drop the std::enable_shared_from_this base, and to
completely eliminate all unsynchronized modification of states of both
Goal and Worker. by the end of this we will hopefully be able to start
and reap multiple derivation builds in parallel, which should speed up
the process quite a bit (at least for short local builds, others might
not notice a large difference. the build hooks will remain a problem.)

Change-Id: I57dcd9b2cab4636ed4aa24cdec67124fef883345
2024-08-03 00:08:44 +00:00
alois31 32ca194ebf Merge "libstore/ssh: only resume the logger when we paused it" into main 2024-08-02 16:59:44 +00:00
alois31 a93dade821
libstore/ssh: only resume the logger when we paused it
In the SSH code, the logger was conditionally paused, but unconditionally
resumed. This was fine as long as resuming the logger was idempotent. Starting
with 0dd1d8ca1c, it isn't any more, and the
behaviour of the code in question was missed. Consequently, an assertion
failure is triggered for example when performing builds against an "SSH" store
on localhost. Fix the issue by only resuming the logger when it has actually
been paused.

Fixes: lix-project/lix#458
Change-Id: Ib1e4d047744a129f15730b7216f9c9368c2f4211
2024-08-02 18:38:14 +02:00
eldritch horrors e5177dddff libstore: move Goal::amDone to Worker
we still mutate goal state to store the results of any given goal run,
but now we also have that information in Worker and could in theory do
something else with it. we could return a map of goal to goal results,
which would also let us better diagnose failures of subgoals (at all).

Change-Id: I1df956bbd9fa8cc9485fb6df32918d68dda3ff48
2024-08-02 13:52:15 +00:00
eldritch horrors dfcab1c3f0 libstore: return finishedness from Goal methods
this is the first step towards removing all result-related mutation of
Goal state from goal implementations themselves, and into Worker state
instead. once that is done we can treat all non-const Goal fields like
private state of the goal itself, and make threading of goals possible

Change-Id: I69ff7d02a6fd91a65887c6640bfc4f5fb785b45c
2024-08-02 13:52:15 +00:00