Cabal: Failed to build slam-0.0.0.1 because it depends on slam-0.0.0.1 which
itself failed to build.
Failed to build servant-server-0.19.2 because it depends on
servant-server-0.19.2 which itself failed to build.
Failed to build w-lggr-2.4.0-4fd1f823. The failure occurred during the
configure step. The build process segfaulted (i.e. SIGSEGV).
Failed to build wai-app-static-3.1.7.4 because it depends on
wai-app-static-3.1.7.4 which itself failed to build.
Failed to build wai-extra-3.1.13.0 because it depends on wai-extra-3.1.13.0
which itself failed to build.
It’s a private codebase, so I can’t share. Perhaps others have been running into similar issues?
Coworker updated to MacOS Ventura last night, reinstalled xcode tools this morning. This is the error he’s getting.
Update. It’s also failing on random other packages after a cabal nuking, with exactly the same error but for different package names. We’re performing a full nuking of ghcup and everything haskell-related.
CallStack (from HasCallStack):
withMetadata, called at src/Distribution/Simple/Utils.hs:370:14 in Cabal-3.8.1.0-inplace:Distribution.Simple.Utils
Error: Cabal: Failed to build lngg-pythn-0.5.8-3f00cd65. The build process
segfaulted (i.e. SIGSEGV).
Failed to build parcel-0.1.0.0 because it depends on parcel-0.1.0.0 which
itself failed to build.
Failed to build servant-py-0.1.1.1 because it depends on servant-py-0.1.1.1
which itself failed to build.
Failed to build slam-0.0.0.1 because it depends on slam-0.0.0.1 which itself
failed to build.
rm -rf ~/.cabal
ghcup nuke
# we manually uninstalled every GHC/cabal before nuking
# wasn't clear from ghcup nuke -h whether ghcup nuke
# would delete them for us
### Restart computer
### Update our toolchain to use ghc-9.2.5
This caused the error to go away. I have no idea why we had the issue. Very likely some subtlety due the reinstalling xcode toolchain.