fetchPnpmDeps: fix reproducibility of pnpm v11 store index by mkg20001 · Pull Request #522703 · NixOS/nixpkgs

mkg20001 · 2026-05-21T15:59:49Z

The pnpm v11 store uses a SQLite database (index.db) whose binary format
is non-deterministic across platforms (version-valid-for number, etc).
This caused hash mismatches when building the same pnpmDeps on different
machines despite identical logical content.

Fix by dumping the SQLite database to a text SQL file during the fetch
phase and reconstructing it during the build phase. This ensures the
stored representation is fully deterministic.

See #521248 (comment) for context on how this issue surfaced in openclaw package

Assisted-By: Claude Opus 4.6 (1M context) noreply@anthropic.com

Things done

gepbird

Thanks!

Adding a test for this would be nice, it could be the same as in pkgs/test/pnpm/pnpm_11/default.nix but with fetcherVersion = 4. In my opinion we should keep existing tests (which is only one currently) for older fetcherVersions.

RossSmyth · 2026-05-21T17:26:12Z

Forgive my ignorance, and just to double check, but are we certain that the sql dump output is deterministic?

mkg20001 · 2026-05-21T17:47:33Z

Forgive my ignorance, and just to double check, but are we certain that the sql dump output is deterministic?

I verified it with openclaw #521248 and didn't have any issues across Linux and mac. Also with autobrr.

mkg20001 · 2026-05-21T17:56:00Z

`nixpkgs-review` result

Generated using nixpkgs-review.

Command: nixpkgs-review pr 522703 --package autobrr --package tests.pnpm.pnpm_11_v3 --package tests.pnpm.pnpm_11_v4
Commit: 4c84d0a0beeb50076f8d9df5e0bf08db36f620c8

`aarch64-darwin`

✅ 3 packages built:

autobrr
tests.pnpm.pnpm_11_v3
tests.pnpm.pnpm_11_v4

mkg20001 · 2026-05-21T17:56:42Z

`nixpkgs-review` result

Generated using nixpkgs-review.

Command: nixpkgs-review pr 522703 --package tests.pnpm.pnpm_11_v3 --package autobrr --package tests.pnpm.pnpm_11_v4
Commit: 4c84d0a0beeb50076f8d9df5e0bf08db36f620c8

`x86_64-linux`

✅ 3 packages built:

autobrr
tests.pnpm.pnpm_11_v3
tests.pnpm.pnpm_11_v4

RossSmyth · 2026-05-21T18:03:50Z

didn't have any issues across Linux and mac

That's good. Any chance of changes across sqlite versions though? Or changes throughout time?

mkg20001 · 2026-05-21T18:08:29Z

didn't have any issues across Linux and mac

That's good. Any chance of changes across sqlite versions though? Or changes throughout time?

Not really something we can prepare for. Less problems than the last implementation. I suppose sqlite dump format is stable enough that we won't see any breakage in the future.

RossSmyth · 2026-05-21T18:15:08Z

Not really something we can prepare for

Sure it is. The first step is knowing whether we need to prepare for it though.

Have you checked with different versions of sqlite? In the dump is there anything that looks like it might be time/date related?

mkg20001 · 2026-05-21T19:14:10Z

Have you checked with different versions of sqlite? In the dump is there anything that looks like it might be time/date related?

Oh you mean in the data itself? This is already handled correctly, as we'd have hash issues already beyond the non-determinism from the db's metadata already. Or did you mean something else like order of the data? We can't really sort -n it because it's sql, but the order has been consistent so far, I didn't yet encounter any issues with that. The metadata that could be problematic (like the internal sql version of the db) is specifically removed by the dump step. It doesn't make sense to look at past versions of sqlite that aren't being used in the generation of the dump, as those will never get used. Only future versions remain relevant. Currently, there is only one version of the sqlite tool in nixpkgs.

RossSmyth · 2026-05-21T21:01:51Z

I am aware there is only one version, but if the data has change in the past it may change in the future.

gepbird · 2026-05-22T09:39:38Z

I am aware there is only one version, but if the data has change in the past it may change in the future.

If the database dump changes after a new pnpm version, it should be caught by nixpkgs-review (which we run for every pnpm update on x86_64-linux, sometimes all packages, sometimes only a few packages break). Then if we can, we'd modify the pnpm-fixup-db-state to produce the older version's hash if possible, otherwise introduce a new pnpm version (like with pnpm_10_29_2 where pnpm broke electron-builder's build script and electron builder STILL hasn't made their fix released in 26.8.2 into their "latest" tag) with pnpm version specefic fetchPnpmDeps changes, or add new fetcherVersion specific changes if we spot issues after the version was merged (or if the changes are unrelated to the pnpm version and/or can be "backported" for older pnpm versions).

I'm not sure how could we improve this workflow or prepare better for upcoming breaking changes (which IMO should be only for major pnpm versions but we even get those for patch releases like #522531).

Sorry for my rants, but I'm somewhat fed up with upstream. This SQLite change is our fault, we should'we tested it on all platforms and shouldnt've assumed that database files are reproducible.

gepbird · 2026-05-22T10:30:19Z

`nixpkgs-review` result

Generated using nixpkgs-review-gha

Command: nixpkgs-review pr 522703 -p tests.pnpm -p autobrr -p bash-language-server -p n8n
Commit: 4c84d0a0beeb50076f8d9df5e0bf08db36f620c8 (subsequent changes)
Merge: 3d34884e4500005d4baadf082502a0b881b4fcd9

Logs: https://github.com/gepbird/nixpkgs-review-gha/actions/runs/26281290784

Download packages from cache:

x86_64-linux

nix-store -r --add-root nixpkgs-pr-522703-x86_64-linux \
  --option binary-caches 'https://cache.nixos.org/ https://gepbird-nixpkgs-review-gha.cachix.org' \
  --option trusted-public-keys '
  cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY=
  gepbird-nixpkgs-review-gha.cachix.org-1:DXx7aJqJyRfrgSacf/SSVzEaPbrH3lych7ti/zUBol4=
  ' \
  /nix/store/r4r72c20zv0fmygcqabr5if3x1akm65s-autobrr-1.79.0 \
  /nix/store/mxcirchv78xyig663rvx77rlvfnm2hqd-bash-language-server-5.6.0 \
  /nix/store/b76l1gsc9shfcydgsysggjvqcjjxasf6-n8n-2.20.6 \
  /nix/store/cx8h4fqxrc8hv0ay4wdxd98p3q5car7x-pnpm-empty-lockfile \
  /nix/store/8cqa215w6rx9z99yllr64l0y22fxsp48-pnpm-fixup-state-db-test-salted-8fq27c16qh21 \
  /nix/store/9zfbybj8wjj6dggd1ia0x7ja56j2pxfs-pnpm-test-11.1.2 \
  /nix/store/rig133jkg2b9ng9mn2lp18f19ab23lhl-pnpm-test-11.1.2

aarch64-linux

nix-store -r --add-root nixpkgs-pr-522703-aarch64-linux \
  --option binary-caches 'https://cache.nixos.org/ https://gepbird-nixpkgs-review-gha.cachix.org' \
  --option trusted-public-keys '
  cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY=
  gepbird-nixpkgs-review-gha.cachix.org-1:DXx7aJqJyRfrgSacf/SSVzEaPbrH3lych7ti/zUBol4=
  ' \
  /nix/store/zw29sg47hlj4wd6hni75sya447ndmx25-autobrr-1.79.0 \
  /nix/store/p4ir59zxj3qc4i14l1px0j3fwgh2as7n-bash-language-server-5.6.0 \
  /nix/store/bys8hrcr2i79nf55a4bmhasad1fldxnx-n8n-2.20.6 \
  /nix/store/gbn3w8jw9ada7h0a2x8ygf2qm4w1r19m-pnpm-empty-lockfile \
  /nix/store/a6xnmnl13wal70i17mdp5qd9z6ds02fn-pnpm-fixup-state-db-test-salted-74ay2yl0a0za \
  /nix/store/4a4cvv9rl2p9fm7hs4i2dxcq50r05jqx-pnpm-test-11.1.2 \
  /nix/store/856fy8zyymfi655m9hhh0i835lp3i1vc-pnpm-test-11.1.2

x86_64-darwin

nix-store -r --add-root nixpkgs-pr-522703-x86_64-darwin \
  --option binary-caches 'https://cache.nixos.org/ https://gepbird-nixpkgs-review-gha.cachix.org' \
  --option trusted-public-keys '
  cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY=
  gepbird-nixpkgs-review-gha.cachix.org-1:DXx7aJqJyRfrgSacf/SSVzEaPbrH3lych7ti/zUBol4=
  ' \
  /nix/store/68b3g07ybxss7fn2jjb0wwjcq97flrh1-bash-language-server-5.6.0 \
  /nix/store/isvym25v5zcyk2wrai592g5d2h16wcyj-pnpm-test-11.1.2 \
  /nix/store/80v5ykfcpw1hw3m99ynqhbmz309pa4ll-pnpm-test-11.1.2 \
  /nix/store/w09ka9xmnabw8z9lr40l0dafh4p8mhrf-pnpm-empty-lockfile \
  /nix/store/qjy47r97ih84w1anr0kqqc98mv74w961-pnpm-fixup-state-db-test-salted-c3hcp2pc1i75

aarch64-darwin

nix-store -r --add-root nixpkgs-pr-522703-aarch64-darwin \
  --option binary-caches 'https://cache.nixos.org/ https://gepbird-nixpkgs-review-gha.cachix.org' \
  --option trusted-public-keys '
  cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY=
  gepbird-nixpkgs-review-gha.cachix.org-1:DXx7aJqJyRfrgSacf/SSVzEaPbrH3lych7ti/zUBol4=
  ' \
  /nix/store/8jmj36b0pil26disvsk1r71ly872bhlg-autobrr-1.79.0 \
  /nix/store/pbba99rn8ldxvf8vq68q7b8m7s1pf90n-bash-language-server-5.6.0 \
  /nix/store/jp400l9flciqngnw88ii5i7ax5jp2qq1-pnpm-test-11.1.2 \
  /nix/store/wvy49fv7nif51hl36hw6xx6saa966yj8-pnpm-test-11.1.2 \
  /nix/store/gvdc3q34s2lwn56xp3zg27mkhbb1yx2g-pnpm-empty-lockfile \
  /nix/store/hdv05k7nm5082l15yimy50ikpy74y1ly-pnpm-fixup-state-db-test-salted-qvbpl7sfw0yn

`x86_64-linux`

✅ 7 packages built:

autobrr
bash-language-server
n8n
tests.pnpm.pnpm-empty-lockfile
tests.pnpm.pnpm-fixup-state-db
tests.pnpm.pnpm_11_v3
tests.pnpm.pnpm_11_v4

`aarch64-linux`

✅ 7 packages built:

autobrr
bash-language-server
n8n
tests.pnpm.pnpm-empty-lockfile
tests.pnpm.pnpm-fixup-state-db
tests.pnpm.pnpm_11_v3
tests.pnpm.pnpm_11_v4

`x86_64-darwin` (sandbox = relaxed)

❌ 2 packages failed to build:

autobrr
n8n

✅ 5 packages built:

bash-language-server
tests.pnpm.pnpm-empty-lockfile
tests.pnpm.pnpm-fixup-state-db
tests.pnpm.pnpm_11_v3
tests.pnpm.pnpm_11_v4

Error logs: `x86_64-darwin`

n8n

@n8n/design-system:build: Error: getaddrinfo ENOTFOUND localhost @n8n/design-system:build: at GetAddrInfoReqWrap.onlookup [as oncomplete] (node:internal/dns/promises:88:17) { @n8n/design-system:build: errno: -3008, @n8n/design-system:build: code: 'ENOTFOUND', @n8n/design-system:build: syscall: 'getaddrinfo', @n8n/design-system:build: hostname: 'localhost' @n8n/design-system:build: } @n8n/design-system:build: ELIFECYCLE Command failed with exit code 1. n8n-workflow:build: ELIFECYCLE Command failed. ERROR @n8n/design-system#build: command (/nix/build/nix-27571-78223883/source/packages/frontend/@n8n/design-system) /nix/store/1gd9q2hpcbdy58n21r8h1k6rzprpwcqb-pnpm-10.33.4/bin/pnpm run build exited (1)

Tasks: 20 successful, 22 total
Cached: 0 cached, 22 total
Time: 1m6.313s
Failed: @n8n/design-system#build

...Finishing writing to cache...

> ...Finishing writing to cache...
ERROR run failed: command exited (1)
ELIFECYCLE Command failed with exit code 1.

`aarch64-darwin` (sandbox = relaxed)

❌ 1 package failed to build:

n8n

✅ 6 packages built:

autobrr
bash-language-server
tests.pnpm.pnpm-empty-lockfile
tests.pnpm.pnpm-fixup-state-db
tests.pnpm.pnpm_11_v3
tests.pnpm.pnpm_11_v4

Error logs: `aarch64-darwin`

n8n

@n8n/design-system:build: Error: getaddrinfo ENOTFOUND localhost @n8n/design-system:build: at GetAddrInfoReqWrap.onlookup [as oncomplete] (node:internal/dns/promises:88:17) { @n8n/design-system:build: errno: -3008, @n8n/design-system:build: code: 'ENOTFOUND', @n8n/design-system:build: syscall: 'getaddrinfo', @n8n/design-system:build: hostname: 'localhost' @n8n/design-system:build: } @n8n/design-system:build: ELIFECYCLE Command failed with exit code 1. @n8n/expression-runtime:build: ELIFECYCLE Command failed. ERROR @n8n/design-system#build: command (/nix/build/nix-49284-1304224823/source/packages/frontend/@n8n/design-system) /nix/store/3ygfga6470hpqzxkid2jnh50ikzr4b26-pnpm-10.33.4/bin/pnpm run build exited (1)

Tasks: 19 successful, 21 total
Cached: 0 cached, 21 total
Time: 31.068s
Failed: @n8n/design-system#build

...Finishing writing to cache...

> ...Finishing writing to cache...
ERROR run failed: command exited (1)
ELIFECYCLE Command failed with exit code 1.

gepbird · 2026-05-22T11:22:55Z

`nixpkgs-review` result

Generated using nixpkgs-review-gha

Command: nixpkgs-review pr 522703 -p n8n -p autobrr
Commit: 4c84d0a0beeb50076f8d9df5e0bf08db36f620c8 (subsequent changes)
Merge: 3d34884e4500005d4baadf082502a0b881b4fcd9

Logs: https://github.com/gepbird/nixpkgs-review-gha/actions/runs/26283163263

`x86_64-darwin` (sandbox = false)

❌ 2 packages failed to build:

autobrr
n8n

Error logs: `x86_64-darwin`

n8n

n8n:build:  5: 0x102bb57e7 v8::internal::Heap::CollectGarbage(v8::internal::AllocationSpace, v8::internal::GarbageCollectionReason, v8::GCCallbackFlags)::$_1::operator()() const [/nix/store/jni6dbiv7ihywbp70zix5g8nx9akmcn3-nodejs-slim-24.15.0/bin/node]
n8n:build:  6: 0x102bb506d void heap::base::Stack::SetMarkerAndCallbackImpl<v8::internal::Heap::CollectGarbage(v8::internal::AllocationSpace, v8::internal::GarbageCollectionReason, v8::GCCallbackFlags)::$_1>(heap::base::Stack*, void*, void const*) [/nix/store/jni6dbiv7ihywbp70zix5g8nx9akmcn3-nodejs-slim-24.15.0/bin/node]
n8n:build:  7: 0x10368fea8 PushAllRegistersAndIterateStack [/nix/store/jni6dbiv7ihywbp70zix5g8nx9akmcn3-nodejs-slim-24.15.0/bin/node]
n8n:build:  8: 0x102b9a128 v8::internal::Heap::CollectGarbage(v8::internal::AllocationSpace, v8::internal::GarbageCollectionReason, v8::GCCallbackFlags) [/nix/store/jni6dbiv7ihywbp70zix5g8nx9akmcn3-nodejs-slim-24.15.0/bin/node]
n8n:build:  9: 0x102b8c9c9 v8::internal::HeapAllocator::AllocateRawWithRetryOrFailSlowPath(int, v8::internal::AllocationType, v8::internal::AllocationOrigin, v8::internal::AllocationAlignment) [/nix/store/jni6dbiv7ihywbp70zix5g8nx9akmcn3-nodejs-slim-24.15.0/bin/node]
n8n:build: 10: 0x102b68b37 v8::internal::Factory::NewFillerObject(int, v8::internal::AllocationAlignment, v8::internal::AllocationType, v8::internal::AllocationOrigin) [/nix/store/jni6dbiv7ihywbp70zix5g8nx9akmcn3-nodejs-slim-24.15.0/bin/node]
n8n:build: 11: 0x1030c0f6a v8::internal::Runtime_AllocateInYoungGeneration(int, unsigned long*, v8::internal::Isolate*) [/nix/store/jni6dbiv7ihywbp70zix5g8nx9akmcn3-nodejs-slim-24.15.0/bin/node]
n8n:build: 12: 0x103766836 Builtins_CEntry_Return1_ArgvOnStack_NoBuiltinExit [/nix/store/jni6dbiv7ihywbp70zix5g8nx9akmcn3-nodejs-slim-24.15.0/bin/node]
n8n:build:  ELIFECYCLE  Command failed with exit code 134.
 ERROR  n8n#build: command (/nix/build/nix-6780-1876986994/source/packages/cli) /nix/store/1gd9q2hpcbdy58n21r8h1k6rzprpwcqb-pnpm-10.33.4/bin/pnpm run build exited (134)
Tasks:    43 successful, 44 total

Cached:    0 cached, 44 total

Time:    10m46.543s

Failed:    n8n#build
...Finishing writing to cache...                                            

>   ...Finishing writing to cache...

ERROR  run failed: command  exited (134)

 ELIFECYCLE  Command failed with exit code 134.

`aarch64-darwin` (sandbox = false)

❌ 2 packages failed to build:

autobrr
n8n

Error logs: `aarch64-darwin`

n8n

n8n:build: 240: 0x1008ccd34 node::builtins::BuiltinLoader::CompileAndCall(v8::Local<v8::Context>, char const*, node::Realm*) [/nix/store/4366l8b6kz1f2sxq6243rfs53grc73zz-nodejs-slim-24.15.0/bin/node]
n8n:build: 241: 0x1009b3554 node::Realm::ExecuteBootstrapper(char const*) [/nix/store/4366l8b6kz1f2sxq6243rfs53grc73zz-nodejs-slim-24.15.0/bin/node]
n8n:build: 242: 0x1008a28bc node::StartExecution(node::Environment*, char const*) [/nix/store/4366l8b6kz1f2sxq6243rfs53grc73zz-nodejs-slim-24.15.0/bin/node]
n8n:build: 243: 0x1008a282c node::StartExecution(node::Environment*, std::__1::function<v8::MaybeLocal<v8::Value> (node::StartExecutionCallbackInfoWithModule const&)>) [/nix/store/4366l8b6kz1f2sxq6243rfs53grc73zz-nodejs-slim-24.15.0/bin/node]
n8n:build: 244: 0x1007e7e94 node::LoadEnvironment(node::Environment*, std::__1::function<v8::MaybeLocal<v8::Value> (node::StartExecutionCallbackInfoWithModule const&)>, std::__1::function<void (node::Environment*, v8::Local<v8::Value>, v8::Local<v8::Value>)>) [/nix/store/4366l8b6kz1f2sxq6243rfs53grc73zz-nodejs-slim-24.15.0/bin/node]
n8n:build: 245: 0x10095c61c node::NodeMainInstance::Run() [/nix/store/4366l8b6kz1f2sxq6243rfs53grc73zz-nodejs-slim-24.15.0/bin/node]
n8n:build: 246: 0x1008a6c44 node::Start(int, char**) [/nix/store/4366l8b6kz1f2sxq6243rfs53grc73zz-nodejs-slim-24.15.0/bin/node]
n8n:build: 247: 0x1844e2b98 start [/usr/lib/dyld]
n8n:build:  ELIFECYCLE  Command failed with exit code 134.
 ERROR  n8n#build: command (/nix/build/nix-3427-168700643/source/packages/cli) /nix/store/3ygfga6470hpqzxkid2jnh50ikzr4b26-pnpm-10.33.4/bin/pnpm run build exited (134)
Tasks:    43 successful, 44 total

Cached:    0 cached, 44 total

Time:    6m13.449s

Failed:    n8n#build
...Finishing writing to cache...                                            

>   ...Finishing writing to cache...

ERROR  run failed: command  exited (134)

 ELIFECYCLE  Command failed with exit code 134.

gepbird · 2026-05-23T00:17:46Z

Weird that autobrr fails on darwin, could it be an issue with the GHA runner? Can someone test it on a real darwin machine?

On the GHA build it fails before reaching the changed lines in this PR:

end of autobrr logs

2026-05-22T10:48:15.3021840Z autobrr-web-pnpm-deps> /nix/store/0y3vzjqz42dmyasywgr6ipcfb0nwdzvy-stdenv-darwin/setup: line 1827:  4628 Abort trap: 6              pnpm install --force --ignore-scripts --registry="$NIX_NPM_REGISTRY" --frozen-lockfile
2026-05-22T10:48:34.6122700Z error: Cannot build '/nix/store/bqa246qv6wllgkxs99s86by95x2n82aw-autobrr-web-pnpm-deps.drv'.
2026-05-22T10:48:34.6124310Z        Reason: builder failed with exit code 134.
2026-05-22T10:48:34.6125080Z        Output paths:
2026-05-22T10:48:34.6126610Z          /nix/store/pfmli4mfwfnhwy5xh7hc1c7c8d65j9in-autobrr-web-pnpm-deps
2026-05-22T10:48:34.6127800Z        Last 25 log lines:
2026-05-22T10:48:34.6128350Z        > + react-select 5.10.2
2026-05-22T10:48:34.6128940Z        > + react-textarea-autosize 8.5.9
2026-05-22T10:48:34.6129500Z        > + stacktracey 2.2.0
2026-05-22T10:48:34.6130150Z        > + tailwind-lerp-colors 1.2.6
2026-05-22T10:48:34.6130790Z        > + tailwindcss 4.2.4
2026-05-22T10:48:34.6131390Z        > + workbox-window 7.4.0
2026-05-22T10:48:34.6132060Z        > + zod 4.4.1
2026-05-22T10:48:34.6132590Z        > + zod-formik-adapter 2.0.0
2026-05-22T10:48:34.6133300Z        >
2026-05-22T10:48:34.6133770Z        > devDependencies:
2026-05-22T10:48:34.6134390Z        > + @eslint/js 10.0.1
2026-05-22T10:48:34.6134950Z        > + @rollup/wasm-node 4.60.2
2026-05-22T10:48:34.6135550Z        > + eslint 10.3.0
2026-05-22T10:48:34.6136110Z        > + eslint-plugin-import 2.32.0
2026-05-22T10:48:34.6138390Z        > + eslint-plugin-react 7.37.5
2026-05-22T10:48:34.6139150Z        > + eslint-plugin-react-hooks 7.1.1
2026-05-22T10:48:34.6139820Z        > + eslint-plugin-react-refresh 0.5.2
2026-05-22T10:48:34.6140620Z        > + eslint-watch 8.0.0
2026-05-22T10:48:34.6141220Z        > + globals 17.5.0
2026-05-22T10:48:34.6141790Z        > + ts-node 10.9.2
2026-05-22T10:48:34.6142400Z        > + typescript 6.0.3
2026-05-22T10:48:34.6143120Z        > + vite 8.0.10
2026-05-22T10:48:34.6143710Z        > + vite-plugin-pwa 1.2.0
2026-05-22T10:48:34.6144300Z        > + vite-plugin-svgr 5.2.0
2026-05-22T10:48:34.6146280Z        > /nix/store/0y3vzjqz42dmyasywgr6ipcfb0nwdzvy-stdenv-darwin/setup: line 1827:  4628 Abort trap: 6              pnpm install --force --ignore-scripts --registry="$NIX_NPM_REGISTRY" --frozen-lockfile
2026-05-22T10:48:34.6147540Z        For full logs, run:
2026-05-22T10:48:34.6147920Z          nix log /nix/store/bqa246qv6wllgkxs99s86by95x2n82aw-autobrr-web-pnpm-deps.drv
2026-05-22T10:48:34.6155260Z error: Cannot build '/nix/store/5k8a3kbg4jibm9ds66qg144hfbsgfl77-autobrr-web-1.79.0.drv'.
2026-05-22T10:48:34.6156810Z        Reason: 1 dependency failed.
2026-05-22T10:48:34.6157460Z        Output paths:
2026-05-22T10:48:34.6158450Z          /nix/store/hlgv0hqbal7bkhrxxgaj3ss9k53f0fx9-autobrr-web-1.79.0
2026-05-22T10:48:34.6166110Z error: Build failed due to failed dependency
2026-05-22T10:48:34.6167640Z error: Cannot build '/nix/store/g9pvza1krshbj2s4qa7wyw48byhdniws-autobrr-1.79.0.drv'.
2026-05-22T10:48:34.6168860Z        Reason: 1 dependency failed.
2026-05-22T10:48:34.6169460Z        Output paths:
2026-05-22T10:48:34.6170200Z          /nix/store/8jmj36b0pil26disvsk1r71ly872bhlg-autobrr-1.79.0
2026-05-22T10:48:34.6171070Z error: Build failed due to failed dependency

But passes on Hydra: https://hydra.nixos.org/build/329256136#tabs-summary

Also in both cases it generates a large number of repeating warning logs: https://cache.nixos.org/log/yzgp82np1glqbszywamblgh5iwcm5c9k-autobrr-web-1.79.0.drv

gepbird

Diff looks good, n8n was most likely already failing while autobrr passed once and failed once (without reaching the lines touched by this PR) on darwin.

mkg20001 · 2026-05-23T10:59:39Z

My attempt #522703 (comment) was on a real darwin with sandbox relaxed, other modes (including off) cause the crash

Make the fetcher version explicit in the test name in preparation for adding a v4 test. Assisted-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Tests the SQLite dump/reconstruct code path introduced in fetcherVersion 4, which ensures reproducibility of the pnpm v11 store index. Assisted-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

SuperSandro2000 · 2026-05-24T22:01:48Z

Do we need to backport this to 26.05?

diogotcorreia · 2026-05-25T00:29:15Z

@SuperSandro2000 I'd say we should, as pnpm v11 is the default in 26.05.

Adding the label

nixpkgs-ci · 2026-05-25T00:29:55Z

Backport failed for release-26.05, because it was unable to cherry-pick the commit(s).

Please cherry-pick the changes locally and resolve any conflicts.

git fetch origin release-26.05
git worktree add -d .worktree/backport-522703-to-release-26.05 origin/release-26.05
cd .worktree/backport-522703-to-release-26.05
git switch --create backport-522703-to-release-26.05
git cherry-pick -x f2ba699f37b4630d0502dfa024f3c1ffb2efab22 8255b29ce1198b4c2f432e483c59f215409852cb f09853e8b794e011bf2d24fef85a5cc0335a8911 6f776474958a15c60ba61c82772e71dc2f6d23ae

Scrumplex · 2026-05-25T07:23:53Z

This was merged before branch-off it seems :D

See https://github.com/NixOS/nixpkgs/commits/release-26.05/pkgs/build-support/node/fetch-pnpm-deps

hsjobeki · 2026-05-28T11:26:00Z

It seems this changes existing build time FOD hashes. Just noticed this while upgrading a project to 26.05.

I suspect people will now have stale FOD hashes and they wouldn't notice since they are already cached and wont get invalidated. Invalidation of an FOD happens through two mechanisms: "hash" eval-value changes, or "name" eval-value changes.

I'm not sure if this was considered; But if this changes the build time hash, without changing the FOD eval-name. people and hydra has stale fod-derivations now.

diogotcorreia · 2026-05-28T11:54:41Z

@hsjobeki Have you bisected? I don't think this PR is the reason for stale hashes, as the changes are gated behind a new fetcherVersion. I had the same thing happen on projects I packaged out-of-tree and the culprit is the upgrade from pnpm 10 to pnpm 11.

You can pass pnpm = pnpm_10 to fetchPnpmDeps to fix this.

hsjobeki · 2026-05-28T13:05:48Z

😮‍💨 spent the last 1 hr bisecting now ^^

Seems #505103 is the actual source of the breakage.

As per my comment there this is an architectural problem of complex fod-fetchers. Any change of the build-script can lead to new output hashes.
I think its better to add invalidation mechanism on every build-script change rather than breaking silently.

emilazy · 2026-07-05T20:59:05Z

This has now broken FOD hashes because of a SQLite update, as predicted.

gepbird · 2026-07-05T22:46:15Z

This has now broken FOD hashes because of a SQLite update, as predicted.

Can you expand on this? I've tried rebuilding a random pnpm 11 package's (your_spotify's) pnpmDeps attribute that is using fetcherVersion = 4 and it passed, so it correctly gave the same hash.
With fetcherVersion = 3 we stored the SQLite DB in the FOD that caused hash mismatches, that's why we're only storing a DB dump in fetcherVersion = 4.

diogotcorreia · 2026-07-05T22:48:14Z

@gepbird There is currently a discussion on Matrix #staging:nixos.org.

It's broken on staging-next following an update of SQLite from 3.51.2 to 3.53.1.

EDIT:
Uploading diff from Matrix so it's easier:

Credits to @K900 for the screenshot

mkg20001 force-pushed the fix/pnpm-sqlite branch from b11f190 to 02373c7 Compare May 21, 2026 16:00

mkg20001 mentioned this pull request May 21, 2026

openclaw: 2026.5.7 -> 2026.5.12 #521248

Merged

4 of 13 tasks

nixpkgs-ci Bot requested review from Scrumplex and gepbird May 21, 2026 16:06

nixpkgs-ci Bot requested a review from av-gal May 21, 2026 16:14

gepbird reviewed May 21, 2026

View reviewed changes

Comment thread pkgs/build-support/node/fetch-pnpm-deps/serve.nix Resolved

Comment thread pkgs/build-support/node/fetch-pnpm-deps/default.nix Resolved

Comment thread pkgs/build-support/node/fetch-pnpm-deps/serve.nix Resolved

gepbird reviewed May 21, 2026

View reviewed changes

Comment thread pkgs/build-support/node/fetch-pnpm-deps/default.nix Resolved

mkg20001 force-pushed the fix/pnpm-sqlite branch from 36704a8 to d9677a1 Compare May 21, 2026 17:28

gepbird approved these changes May 23, 2026

View reviewed changes

nixpkgs-ci Bot added 12.approvals: 1 This PR was reviewed and approved by one person. 12.approved-by: package-maintainer This PR was reviewed and approved by a maintainer listed in any of the changed packages. labels May 23, 2026

Scrumplex approved these changes May 23, 2026

View reviewed changes

Scrumplex added this pull request to the merge queue May 23, 2026

mkg20001 added 2 commits May 23, 2026 22:13

tests.pnpm: rename pnpm_11 test to pnpm_11_v3

f09853e

Make the fetcher version explicit in the test name in preparation for adding a v4 test. Assisted-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

tests.pnpm: add pnpm_11_v4 test for fetcherVersion 4

6f77647

Tests the SQLite dump/reconstruct code path introduced in fetcherVersion 4, which ensures reproducibility of the pnpm v11 store index. Assisted-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Scrumplex force-pushed the fix/pnpm-sqlite branch from 261ab30 to 6f77647 Compare May 23, 2026 20:13

coderabbitai Bot mentioned this pull request May 24, 2026

strictDepBuilds=true、他サプライチェーン攻撃への対策の追加 #61 ano333333/ongeki-score-fetch#62

Merged

Scrumplex added this pull request to the merge queue May 24, 2026

Merged via the queue into NixOS:master with commit f967915 May 24, 2026
26 checks passed

mkg20001 deleted the fix/pnpm-sqlite branch May 24, 2026 14:19

diogotcorreia added the backport release-26.05 Backport PR automatically label May 25, 2026

github-actions Bot added the 8.has: failed backport label May 25, 2026

gepbird mentioned this pull request May 27, 2026

qui: update pnpmDeps' fetcherVersion from 3 to 4 #524812

Merged

3 of 14 tasks

hsjobeki mentioned this pull request May 28, 2026

pnpm_11: init at 11.1.0 #505103

Merged

4 of 13 tasks

gepbird mentioned this pull request May 29, 2026

fetchPnpmDeps: pnpm 11 build spams logs on darwin and gets SIGKILL on GHA runners #525627

Closed

4 tasks done

Prince213 mentioned this pull request Jun 13, 2026

doc/javascript.section: recommend fetcherVersion 4 #531246

Merged

2 of 14 tasks

eeedean removed the 8.has: failed backport label Jun 24, 2026

nicomem mentioned this pull request Jun 27, 2026

authelia: mark web build broken on darwin #535866

Merged

2 of 13 tasks

gepbird mentioned this pull request Jul 6, 2026

fetchPnpmDeps: fix changed hashes due to sqlite bump by reverting to old format #538919

Merged

3 of 13 tasks

coderabbitai Bot mentioned this pull request Jul 8, 2026

Sync speaking bot MCP with current OpenAPI contract Meeting-BaaS/speaking-bots-mcp#1

Open

jmmaloney4 mentioned this pull request Jul 14, 2026

fix(pnpm): migrate fetchPnpmDeps to fetcherVersion 4 jmmaloney4/jackpkgs#334

Merged

3 tasks done

coderabbitai Bot mentioned this pull request Aug 4, 2026

fix(security): patch fast-uri, postcss, and brace-expansion advisories Fission-AI/OpenSpec#1510

Merged

Conversation

mkg20001 commented May 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Things done

Uh oh!

gepbird left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

RossSmyth commented May 21, 2026

Uh oh!

mkg20001 commented May 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mkg20001 commented May 21, 2026

nixpkgs-review result

aarch64-darwin

Uh oh!

mkg20001 commented May 21, 2026

nixpkgs-review result

x86_64-linux

Uh oh!

RossSmyth commented May 21, 2026

Uh oh!

mkg20001 commented May 21, 2026

Uh oh!

RossSmyth commented May 21, 2026

Uh oh!

mkg20001 commented May 21, 2026

Uh oh!

RossSmyth commented May 21, 2026

Uh oh!

gepbird commented May 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

gepbird commented May 22, 2026

nixpkgs-review result

x86_64-linux

aarch64-linux

x86_64-darwin (sandbox = relaxed)

aarch64-darwin (sandbox = relaxed)

Uh oh!

gepbird commented May 22, 2026

nixpkgs-review result

x86_64-darwin (sandbox = false)

aarch64-darwin (sandbox = false)

Uh oh!

gepbird commented May 23, 2026

Uh oh!

gepbird left a comment

Choose a reason for hiding this comment

Uh oh!

mkg20001 commented May 23, 2026

Uh oh!

Uh oh!

SuperSandro2000 commented May 24, 2026

Uh oh!

diogotcorreia commented May 25, 2026

Uh oh!

nixpkgs-ci Bot commented May 25, 2026

Uh oh!

Scrumplex commented May 25, 2026

Uh oh!

hsjobeki commented May 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

diogotcorreia commented May 28, 2026

Uh oh!

hsjobeki commented May 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

emilazy commented Jul 5, 2026

Uh oh!

gepbird commented Jul 5, 2026

Uh oh!

diogotcorreia commented Jul 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

mkg20001 commented May 21, 2026 •

edited

Loading

mkg20001 commented May 21, 2026 •

edited

Loading

`nixpkgs-review` result

`aarch64-darwin`

`nixpkgs-review` result

`x86_64-linux`

gepbird commented May 22, 2026 •

edited

Loading

`nixpkgs-review` result

`x86_64-linux`

`aarch64-linux`

`x86_64-darwin` (sandbox = relaxed)

`aarch64-darwin` (sandbox = relaxed)

`nixpkgs-review` result

`x86_64-darwin` (sandbox = false)

`aarch64-darwin` (sandbox = false)

hsjobeki commented May 28, 2026 •

edited

Loading

hsjobeki commented May 28, 2026 •

edited

Loading

diogotcorreia commented Jul 5, 2026 •

edited

Loading