-
-
Notifications
You must be signed in to change notification settings - Fork 14.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[haskell-updates] Packages fail with a linker error on darwin #152859
Comments
The TLS.o file in the recent
doesn't correspond to anything meaningful: --- /tmp/blaze-builder.cabal 2021-12-30 22:24:19.000000000 -0500
+++ /nix/store/7vh4ic8s5qlw8fg6yry1i0x0frz4h5a0-blaze-builder-0.4.2.2-r1.cabal 1969-12-31 19:00:01.000000000 -0500
@@ -1,5 +1,6 @@
Name: blaze-builder
Version: 0.4.2.2
+x-revision: 1
Synopsis: Efficient buffered output.
Description:
@@ -33,7 +34,7 @@
Cabal-version: >= 1.10
Tested-with:
- GHC == 9.2.0.20210821
+ GHC == 9.2.1
GHC == 9.0.1
GHC == 8.10.7
GHC == 8.8.4
@@ -85,7 +86,7 @@
, bytestring >= 0.9 && < 1.0
, deepseq
, ghc-prim
- , text >= 0.10 && < 1.3
+ , text >= 0.10 && < 2.1
if impl(ghc < 7.8)
build-depends: bytestring-builder Looking at the object files I can't see anything special about either. |
I've rebuilt the |
Hm, maybe there is a more general bug, because the same goes for haskell-language-server on darwin.
|
This seems to check out, there was an error about an |
Restarting that job won't actually rebuild it, since the output path is already in the store / the binary cache… We either need to wait for a natural rebuild or inject some kind of change into the relevant drvs. |
How can we identify the corruption and either fix or fail the build? Or, is there a bug in GHC? Let's explore these options instead of restarting successful builds. |
I can muster neither the time nor energy to do so. Since the build was successful, everyone should be able to substitute the corrupted store path from I would also not rule out that this was an issue with builders, as there was other weird stuff happening like derivations failing in nixpkgs/pkgs/stdenv/generic/setup.sh Line 406 in 75a86ab
|
E. g. https://hydra.nixos.org/build/162562802/nixlog/2:
I wonder if something went wrong with the filesystem backing the nix store? |
@veprbl Thanks for looking into this, I was able to confirm your findings and cause a rebuild of the relevant derivations by changing their derivation hash. |
Note that this issue is not really solved, we just worked around the worst instances. We are continuing to see new corruption-related problems in darwin builds. |
Several Haskell packages fail with
examples:
Bisecting points to a3b6030 as the first bad commit
Also see #152408 (comment)
The text was updated successfully, but these errors were encountered: