Commit graph

3408 commits

Author SHA1 Message Date
Théophane Hufschmitt
d71d9e9fbf moveFile -> renameFile
`move` tends to have this `mv` connotation of “I will copy it for you if
needs be”
2022-08-03 10:27:25 +02:00
Théophane Hufschmitt
c2de0a232c Create a wrapper around stdlib’s rename
Directly takes some c++ strings, and gently throws an exception on error
(rather than having to inline this logic everywhere)
2022-08-03 10:27:25 +02:00
Félix Baylac-Jacqué
1467a98d4c
derivation-goal.cc: remove bmCheck custom return branch on buildDone
Once a derivation goal has been completed, we check whether or not
this goal was meant to be repeated to check its output.

An early return branch was preventing the worker to reach that repeat
code branch, hence breaking the --check command (#2619).

It seems like this early return branch is an artifact of a passed
refactoring. As far as I can tell, buildDone's main branch also
cleanup the tmp directory before returning.
2022-08-01 11:39:19 +02:00
Théophane Hufschmitt
2805439335
Merge pull request #6814 from amjoseph-nixpkgs/pr/sandbox-error-messages
local-derivation-goal.cc: improve error messages when sandboxing fails
2022-07-22 13:27:52 +02:00
Théophane Hufschmitt
7ed91d6c6a
Merge branch 'master' into parallel-nix-copy 2022-07-20 10:05:34 +02:00
Alex Wied
722de8ddcc libstore/globals.cc: Move cgroup detection to libutil 2022-07-19 16:25:53 -04:00
Alex Wied
1af5d798a4 libstore/globals.cc: Automatically set cores based on cgroup CPU limit
By default, Nix sets the "cores" setting to the number of CPUs which are
physically present on the machine. If cgroups are used to limit the CPU
and memory consumption of a large Nix build, the OOM killer may be
invoked.

For example, consider a GitLab CI pipeline which builds a large software
package. The GitLab runner spawns a container whose CPU is limited to 4
cores and whose memory is limited to 16 GiB. If the underlying machine
has 64 cores, Nix will invoke the build with -j64. In many cases, that
level of parallelism will invoke the OOM killer and the build will
completely fail.

This change sets the default value of "cores" to be
ceil(cpu_quota / cpu_period), with a fallback to
std:🧵:hardware_concurrency() if cgroups v2 is not detected.
2022-07-19 16:03:58 -04:00
Théophane Hufschmitt
56f6f3725f Don't ultimately trust the signed paths
Like the old implem did (and like you'd want it to be anyways)
2022-07-19 19:46:00 +02:00
Adam Joseph
36e1383b6b local-derivation-goal.cc: save global errno to the stack before performing tests which might clobber it 2022-07-19 03:53:20 -07:00
Adam Joseph
99fcc91f67 as requested by @thufschmitt https://github.com/NixOS/nix/pull/6814#discussion_r924275777 2022-07-19 03:33:12 -07:00
Adam Joseph
5f51539f88 change warn() to notice() 2022-07-19 03:30:52 -07:00
Adam Joseph
c8c6203c2c local-derivation-goal.cc: detect unprivileged_userns_clone failure mode
The workaround for "Some distros patch Linux" mentioned in
local-derivation-goal.cc will not help in the `--option
sandbox-fallback false` case.  To provide the user more helpful
guidance on how to get the sandbox working, let's check to see if the
`/proc` node created by the aforementioned patch is present and
configured in a way that will cause us problems.  If so, give the user
a suggestion for how to troubleshoot the problem.
2022-07-17 01:27:22 -07:00
Adam Joseph
6fc56318bf local-derivation-goal.cc: add comment re: CLONE_NEWUSER
local-derivation-goal.cc contains a comment stating that "Some distros
patch Linux to not allow unprivileged user namespaces."  Let's give a
pointer to a common version of this patch for those who want more
details about this failure mode.
2022-07-17 01:23:32 -07:00
Adam Joseph
8d35f387dc local-derivation-goal.cc: warn if failing and /proc/self/ns/user missing
This commit causes nix to `warn()` if sandbox setup has failed and
`/proc/self/ns/user` does not exist.  This is usually a sign that the
kernel was compiled without `CONFIG_USER_NS=y`, which is required for
sandboxing.
2022-07-16 19:37:27 -07:00
Adam Joseph
90830b1074 local-derivation-goal.cc: warn if failing due to max_user_namespaces==0
This commit uses `warn()` to notify the user if sandbox setup fails
with errno==EPERM and /proc/sys/user/max_user_namespaces is missing or
zero, since that is at least part of the reason why sandbox setup
failed.

Note that `echo -n 0 > /proc/sys/user/max_user_namespaces` or
equivalent at boot time has been the recommended mitigation for
several Linux LPE vulnerabilities over the past few years.  Many users
have applied this mitigation and then forgotten that they have done
so.
2022-07-16 19:30:53 -07:00
Adam Joseph
8ea3a911aa local-derivation-goal.cc: improve error messages when sandboxing fails
The failure modes for nix's sandboxing setup are pretty complicated.
When nix is unable to set up the sandbox, let's provide more detail
about what went wrong.  Specifically:

* Make sure the error message includes the word "sandbox" so the user
  knows that the failure was related to sandboxing.

* If `--option sandbox-fallback false` was provided, and removing it
  would have allowed further attempts to make progress, let the user
  know.
2022-07-16 14:56:24 -07:00
Alex Wied
b88fb50e21 fix(libstore): allow Nix to access all Rosetta 2 paths on MacOS
Fixes: #5884
2022-07-15 12:10:56 -07:00
Eelco Dolstra
3bcd7a5474 Disable auto-chroot if $NIX_STATE_DIR is set
Issue #6732.
2022-07-15 12:32:29 +02:00
Eelco Dolstra
ff49c75502 Disable auto-chroot if $NIX_STORE_DIR is set
Fixes #6732.
2022-07-14 17:47:09 +02:00
Eelco Dolstra
6cab528461 Don't fail if we can't create ~/.local/share/nix/root
https://hydra.nixos.org/build/182135943
2022-06-29 12:16:51 +02:00
Théophane Hufschmitt
56cf96a1b9
Merge pull request #6706 from lheckemann/cache-info-cache-invalidation
nar-info-disk-cache: refresh nix-cache-info weekly
2022-06-29 07:56:27 +02:00
Eelco Dolstra
586fa707fc
Merge pull request #6714 from edolstra/auto-chroot-store
Automatically use a chroot store if /nix doesn't exist
2022-06-25 00:03:35 +02:00
Eelco Dolstra
30d4aa5dd6 Only do the auto chroot store on Linux 2022-06-24 23:35:21 +02:00
Linus Heckemann
8cf26385cd [fixup] handle cache expiration in sqlite query 2022-06-23 14:54:25 -04:00
Cole Helbling
561a258f1d libstore/nar-info: drop unused system field
This was unused everywhere (and even the official NixOS binary cache
did not produce .narinfo files containing a "System:" field).
2022-06-23 14:25:10 -04:00
Eelco Dolstra
1cb376d60e
Fix typo
Co-authored-by: Cole Helbling <cole.e.helbling@outlook.com>
2022-06-23 17:18:22 +02:00
Eelco Dolstra
2a9fddc0b1 Automatically use a chroot store if /nix doesn't exist
Specifically, if we're not root and the daemon socket does not exist,
then we use ~/.local/share/nix/root as a chroot store. This enables
non-root users to download nix-static and have it work out of the box,
e.g.

  ubuntu@ip-10-13-1-146:~$ ~/nix run nixpkgs#hello
  warning: '/nix' does not exists, so Nix will use '/home/ubuntu/.local/share/nix/root' as a chroot store
  Hello, world!
2022-06-23 16:29:50 +02:00
Eelco Dolstra
3c57db1a0f
Merge pull request #6710 from edolstra/embedded-sandbox-shell
Embed the sandbox shell into the statically linked 'nix' binary
2022-06-23 15:34:16 +02:00
Eelco Dolstra
0b2ea0023c Fix typo 2022-06-23 14:22:11 +02:00
Théophane Hufschmitt
027f6a735f
Merge pull request #6673 from asymmetric/warn
libstore: improve warning message on missing sig
2022-06-23 06:42:48 +02:00
Eelco Dolstra
925b975224 Embed the sandbox shell into the statically linked 'nix' binary
With this, Nix will write a copy of the sandbox shell to /bin/sh in
the sandbox rather than bind-mounting it from the host filesystem.
This makes /bin/sh work out of the box with nix-static, i.e. you no
longer get

  /nix/store/qa36xhc5gpf42l3z1a8m1lysi40l9p7s-bootstrap-stage4-stdenv-linux/setup: ./configure: /bin/sh: bad interpreter: No such file or directory
2022-06-23 04:08:28 +02:00
Eelco Dolstra
184f4e40de Remove NIX_LIBEXEC_DIR 2022-06-23 01:32:46 +02:00
Eelco Dolstra
d3176ce076 Fix build-remote in nix-static
'build-remote' is now executed via /proc/self/exe so it always works.
2022-06-23 01:32:46 +02:00
Linus Heckemann
d533a88546 nar-info-disk-cache: refresh nix-cache-info weekly
This allows changes to nix-cache-info to be picked up by existing
clients. Previously, the only way for this to happen would be for
clients to delete binary-cache-v6.sqlite, which is quite awkward for
users.

On the other hand, updates to nix-cache-info should be pretty rare,
hence the choice of a fairly long TTL. Configurability is probably not
useful enough to warrant implementing it.
2022-06-22 10:49:18 -04:00
Lorenzo Manacorda
475249db8a libstore: improve warning message on missing sig
Clarifies that the substitute will be ignored/skipped.
2022-06-15 17:34:28 +02:00
Eelco Dolstra
9f58df4c91
Merge pull request #6619 from Jonpez2/patch-1
Add security.csm to ignored-acls
2022-06-13 16:23:13 +02:00
Eelco Dolstra
2fef24f528
Merge pull request #6634 from lovesegfault/fix-getgrouplist
fix(libstore/lock): support users that belong to more than 10 groups
2022-06-13 15:45:09 +02:00
Anders Kaseorg
754cd53faf Add missing rethrows in conditional exception handlers
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
2022-06-10 10:52:41 -07:00
Bernardo Meurer
931930feb1 fix(libstore/lock): support users that belong to more than 10 groups
The manpage for `getgrouplist` says:

> If the number of groups of which user is a member is less than or
> equal to *ngroups, then the value *ngroups is returned.
>
> If the user is a member of more than *ngroups groups, then
> getgrouplist() returns -1.  In this case, the value returned in
> *ngroups can be used to resize the buffer passed to a further
> call getgrouplist().

In our original code, however, we allocated a list of size `10` and, if
`getgrouplist` returned `-1` threw an exception. In practice, this
caused the code to fail for any user belonging to more than 10 groups.

While unusual for single-user systems, large companies commonly have a
huge number of POSIX groups users belong to, causing this issue to crop
up and make multi-user Nix unusable in such settings.

The fix is relatively simple, when `getgrouplist` fails, it stores the
real number of GIDs in `ngroups`, so we must resize our list and retry.
Only then, if it errors once more, we can raise an exception.

This should be backported to, at least, 2.9.x.
2022-06-08 13:45:41 -04:00
Théophane Hufschmitt
480c2b6699 Rewrite the CA paths when moving them between store
Bring back the possibility to copy CA paths with no reference (like the
outputs of FO derivations or stuff imported at eval time) between stores
that have a different prefix.
2022-06-08 15:13:11 +02:00
Théophane Hufschmitt
cb0553ecd0 Restore the "low-latency" ssh copying 2022-06-08 14:03:46 +02:00
Jonpez2
a7d25d339d
Add security.csm to the default ignore list 2022-06-08 09:32:14 +01:00
Théophane Hufschmitt
95f47c28fb Make nix copy parallel again
FILLME
2022-06-03 17:01:16 +02:00
Eelco Dolstra
2097c30b08
Merge pull request #6600 from asymmetric/hash-comment
schema.sql: add comment about hash being in base16
2022-06-03 08:51:07 +02:00
Théophane Hufschmitt
a5f7b934bd
Merge branch 'master' into lto 2022-06-01 17:55:03 +02:00
Lorenzo Manacorda
a9358a6097 schema.sql: add comment about hash being in base16 2022-06-01 14:59:57 +02:00
Eelco Dolstra
6378f0bb32
RemoteStore::queryRealisationUncached(): Fix potential noexcept violation 2022-05-30 13:27:13 +02:00
Eelco Dolstra
b8faa83742
HttpBinaryCacheStore::getFile(): Don't throw an exception
This violates the noexcept specification.

Fixes #6445.
2022-05-30 13:24:04 +02:00
pennae
b092afe77d
Merge branch 'master' into lto 2022-05-25 11:55:13 +00:00
Eli Kogan-Wang
e53349dd6e change priority conflict message 2022-05-16 16:16:06 +02:00
Eli Kogan-Wang
be2b19041e Integrate review changes 2022-05-13 22:02:28 +02:00
Eli Kogan-Wang
aefc6c4f41 Add priority for nix profile install 2022-05-11 12:16:35 +02:00
Eelco Dolstra
107613ad2b Fix compiler warning 2022-05-04 11:31:39 +02:00
Eelco Dolstra
3e87c8e62b Move json stuff out of util.cc 2022-05-04 11:22:06 +02:00
Alain Zscheile
1385b20078
Get rid of most .at calls (#6393)
Use one of `get` or `getOr` instead which will either return a null-pointer (with a nicer error message) or a default value when the key is missing.
2022-05-04 07:44:32 +02:00
Eelco Dolstra
a3c6c5b1c7 nix profile: Support overriding outputs 2022-05-03 15:00:34 +02:00
Eelco Dolstra
4a79cba511 Allow selecting derivation outputs using 'installable!outputs'
E.g. 'nixpkgs#glibc^dev,static' or 'nixpkgs#glibc^*'.
2022-05-03 13:43:52 +02:00
Eelco Dolstra
70a30dbc00 Fix libcxx build
Fixes #6458.
2022-04-28 14:37:05 +02:00
Eelco Dolstra
4a9623b129 Fix passing $OUT_PATHS to the post-build hook
Fixes #6446.
2022-04-28 13:36:01 +02:00
pennae
d6d6bbd9ef
Merge branch 'master' into lto 2022-04-25 14:02:37 +00:00
Eelco Dolstra
ebad9213f7
Merge pull request #6431 from NixOS/unbreak-my-build
Make the default SQLiteError constructor public
2022-04-21 10:41:07 +02:00
Sergei Trofimovich
975b0b52e7 ca: add sqlite index on RealisationsRefs(realisationReference)
Without the change any CA deletion triggers linear scan on large
RealisationsRefs table:

    sqlite>.eqp full
    sqlite> delete from RealisationsRefs where realisationReference IN ( select id from Realisations where outputPath = 1234567890 );
    QUERY PLAN
    |--SCAN RealisationsRefs
    `--LIST SUBQUERY 1
       `--SEARCH Realisations USING COVERING INDEX IndexRealisationsRefsOnOutputPath (outputPath=?)

With the change it gets turned into a lookup:

    sqlite> CREATE INDEX IndexRealisationsRefsRealisationReference on RealisationsRefs(realisationReference);
    sqlite> delete from RealisationsRefs where realisationReference IN ( select id from Realisations where outputPath = 1234567890 );
    QUERY PLAN
    |--SEARCH RealisationsRefs USING INDEX IndexRealisationsRefsRealisationReference (realisationReference=?)
    `--LIST SUBQUERY 1
       `--SEARCH Realisations USING COVERING INDEX IndexRealisationsRefsOnOutputPath (outputPath=?)
2022-04-21 10:06:39 +02:00
regnat
86d7a11c6b Make sure to delete all the realisation refs
Deleting just one will only work in the test cases where I didn’t bother
creating too many of them :p
2022-04-21 10:06:39 +02:00
regnat
92656da0b9 Fix the gc with indirect self-references via the realisations
If the derivation `foo` depends on `bar`, and they both have the same
output path (because they are CA derivations), then this output path
will depend both on the realisation of `foo` and of `bar`, which
themselves depend on each other.
This confuses SQLite which isn’t able to automatically solve this
diamond dependency scheme.

Help it by adding a trigger to delete all the references between the
relevant realisations.

Fix #5320
2022-04-21 10:06:39 +02:00
Théophane Hufschmitt
e7d79c7861 Make the default SQLiteError constructor public
Otherwise the clang builds fail because the constructor of `SQLiteBusy`
inherits it, `SQLiteError::_throw` tries to call it, which fails.

Strangely, gcc works fine with it. Not sure what the correct behavior is
and who is buggy here, but either way, making it public is at the worst
a reasonable workaround
2022-04-21 09:40:55 +02:00
John Ericson
f63b0f4540 Actually, solve this in a lighter-weight way
The templating is very superficial
2022-04-20 17:37:59 +00:00
John Ericson
05ec0beb40 Move templated functions to sqlite-impl.hh
This ensures that use-sites properly trigger new monomorphisations on
one hand, and on the other hand keeps the main `sqlite.hh` clean and
interface-only. I think that is good practice in general, but in this
situation in particular we do indeed have `sqlite.hh` users that don't
need the `throw_` function.
2022-04-20 16:57:06 +00:00
John Ericson
3c220442ff Merge remote-tracking branch 'upstream/master' into fix-url-format 2022-04-20 16:53:16 +00:00
John Ericson
75b62e5260 Avoid fmt when constructor already does it
There is a correctnes issue here, but #3724 will fix that. This is just
a cleanup for brevity's sake.
2022-04-19 01:44:11 +00:00
Alex Ameen
e5c934cd48
doc: rephrase connect-timeout help message
Co-authored-by: Cole Helbling <cole.e.helbling@outlook.com>
2022-04-17 18:17:37 -05:00
Alex Ameen
25c85f5a0e
doc: document nix.conf connect-timeout default 2022-04-17 17:14:38 -05:00
Sebastian Blunt
63d9a81819 Log builder args and environment variables
Previously it only logged the builder's path, this changes it to log the
arguments at the same log level, and the environment variables at the
vomit level.

This helped me debug https://github.com/svanderburg/node2nix/issues/75
2022-04-10 21:10:37 -07:00
Eelco Dolstra
c68963eaea Remove duplicate "error:" 2022-04-08 11:48:30 +02:00
Eelco Dolstra
8bd9ebf52c Error: Remove unused sname() method 2022-04-08 11:31:51 +02:00
Eelco Dolstra
168ef9f3ab Remove unused Error.name field 2022-04-08 11:31:51 +02:00
Eelco Dolstra
318936366d Fix empty 'nix copy' error message
This was caused by SubstitutionGoal not setting the errorMsg field in
its BuildResult. We now get a more descriptive message than in 2.7.0, e.g.

  error: path '/nix/store/13mh...' is required, but there is no substituter that can build it

instead of the misleading (since there was no build)

  error: build of '/nix/store/13mh...' failed

Fixes #6295.
2022-04-06 12:43:53 +02:00
Eelco Dolstra
a4a1de69dc Add missing #include 2022-04-04 16:49:39 +02:00
Eelco Dolstra
a99af85a77 Fix macOS build 2022-03-31 16:39:18 +02:00
Eelco Dolstra
6051cc954b Rename 'pure' -> 'sandboxed' for consistency 2022-03-31 16:12:25 +02:00
Eelco Dolstra
e279fbb16a needsNetworkAccess() -> isSandboxed() 2022-03-31 16:06:40 +02:00
Eelco Dolstra
4e043c2f32 Document isPure() 2022-03-31 16:01:50 +02:00
Eelco Dolstra
d7fc33c842 Fix macOS build 2022-03-31 15:59:14 +02:00
Eelco Dolstra
18935e8b9f Support fixed-output derivations depending on impure derivations 2022-03-31 13:43:20 +02:00
Eelco Dolstra
5cd72598fe Add support for impure derivations
Impure derivations are derivations that can produce a different result
every time they're built. Example:

  stdenv.mkDerivation {
    name = "impure";
    __impure = true; # marks this derivation as impure
    outputHashAlgo = "sha256";
    outputHashMode = "recursive";
    buildCommand = "date > $out";
  };

Some important characteristics:

* This requires the 'impure-derivations' experimental feature.

* Impure derivations are not "cached". Thus, running "nix-build" on
  the example above multiple times will cause a rebuild every time.

* They are implemented similar to CA derivations, i.e. the output is
  moved to a content-addressed path in the store. The difference is
  that we don't register a realisation in the Nix database.

* Pure derivations are not allowed to depend on impure derivations. In
  the future fixed-output derivations will be allowed to depend on
  impure derivations, thus forming an "impurity barrier" in the
  dependency graph.

* When sandboxing is enabled, impure derivations can access the
  network in the same way as fixed-output derivations. In relaxed
  sandboxing mode, they can access the local filesystem.
2022-03-31 13:43:20 +02:00
Théophane Hufschmitt
390269ed87 Simplify the handling of the hash modulo
Rather than having four different but very similar types of hashes, make
only one, with a tag indicating whether it corresponds to a regular of
deferred derivation.

This implies a slight logical change: The original Nix+multiple-outputs
model assumed only one hash-modulo per derivation. Adding
multiple-outputs CA derivations changed this as these have one
hash-modulo per output. This change is now treating each derivation as
having one hash modulo per output.
This obviously means that we internally loose the guaranty that
all the outputs of input-addressed derivations have the same hash
modulo. But it turns out that it doesn’t matter because there’s nothing
in the code taking advantage of that fact (and it probably shouldn’t
anyways).

The upside is that it is now much easier to work with these hashes, and
we can get rid of a lot of useless `std::visit{ overloaded`.

Co-authored-by: John Ericson <John.Ericson@Obsidian.Systems>
2022-03-29 18:17:35 +02:00
Eelco Dolstra
057f9ee190 nix profile install: Don't use queryDerivationOutputMap()
Instead get the outputs from Installable::build(). This will also
allow 'nix profile install' to support impure derivations.

Fixes #6286.
2022-03-28 14:23:39 +02:00
Eelco Dolstra
50c229ad9a
Use wantOutput
Co-authored-by: John Ericson <git@JohnEricson.me>
2022-03-25 08:02:49 +01:00
Eelco Dolstra
540d7e33d8 Retry substitution after an incomplete closure only once
This avoids an infinite loop in the final test in
tests/binary-cache.sh. I think this was only not triggered previously
by accident (because we were clearing wantedOutputs in between).
2022-03-24 23:25:12 +01:00
Eelco Dolstra
fe5509df9a Only return wanted outputs 2022-03-24 23:24:48 +01:00
Eelco Dolstra
09796c0263 Random cleanup 2022-03-24 23:24:10 +01:00
Eelco Dolstra
175c78591b Random cleanup 2022-03-24 23:09:43 +01:00
Eelco Dolstra
28186b7044 Add a test for fetchClosure and 'nix store make-content-addressed' 2022-03-24 21:33:33 +01:00
Eelco Dolstra
f18607549c Fix makeContentAddressed() on self-references
LocalStore::addToStore() since
79ae9e4558 expects a regular NAR hash,
rather than a NAR hash modulo self-references. Fixes #6300.

Also, makeContentAddressed() now rewrites the entire closure (so 'nix
store make-content-addressable' no longer needs '-r'). See #6301.
2022-03-24 21:33:33 +01:00
Eelco Dolstra
545c2d0d8c fetchClosure: Allow a path to be rewritten to CA on the fly
The advantage is that the resulting closure doesn't need to be signed,
so you don't need to configure any binary cache keys on the client.
2022-03-24 21:33:33 +01:00
Eelco Dolstra
0b42afe027 buildProfile(): Ignore manifest.{nix,json}
If a package installs a file named manifest.json, it caused nix-env to
consider the profile a new-style profile created by 'nix
profile'. Fixes #6032.
2022-03-22 10:48:02 +01:00
John Ericson
a544ed7684 Generalize DerivationType in preparation for impure derivations 2022-03-18 14:59:56 +00:00
John Ericson
197feed51d Clean up DerivationOutput, and headers
1. `DerivationOutput` now as the `std::variant` as a base class. And the
   variants are given hierarchical names under `DerivationOutput`.

   In 8e0d0689be @matthewbauer and I
   didn't know a better idiom, and so we made it a field. But this sort
   of "newtype" is anoying for literals downstream.

   Since then we leaned the base class, inherit the constructors trick,
   e.g. used in `DerivedPath`. Switching to use that makes this more
   ergonomic, and consistent.

2. `store-api.hh` and `derivations.hh` are now independent.

   In bcde5456cc I swapped the dependency,
   but I now know it is better to just keep on using incomplete types as
   much as possible for faster compilation and good separation of
   concerns.
2022-03-17 22:35:53 +00:00
Eelco Dolstra
6afc361798
Merge pull request #6277 from thufschmitt/ca/nix-build-dry-run
Fix `nix build --dry-run` with CA derivations
2022-03-17 22:23:15 +01:00
Sergei Trofimovich
d58453f72e gc: don't visit implicit referrers on garbage collection
Before the change garbage collector was not considering
`.drv` and outputs as alive even if configuration says otherwise.
As a result `nix store gc --dry-run` could visit (and parse)
`.drv` files multiple times (worst case it's quadratic).

It happens because `alive` set was populating only runtime closure
without regard for actual configuration. The change fixes it.

Benchmark: my system has about 139MB, 40K `.drv` files.

Performance before the change:

    $ time nix store gc --dry-run
    real    4m22,148s

Performance after the change:

    $ time nix store gc --dry-run
    real    0m14,178s
2022-03-17 18:47:29 +00:00
Théophane Hufschmitt
3fc4c612fb Fix nix build --dry-run with CA derivations
Don’t try and assume that we know the output paths when we’ve just built
with `--dry-run`. Instead make `--dry-run` follow a different code path
that won’t assume the knowledge of the output paths at all.

Fix #6275
2022-03-17 11:55:19 +01:00