As was discovered in haskell/cabal#11465 the following process versions can segfault on macOS Sequoia 15.7.4:
1.6.26.1
1.6.25.0
1.6.24.0
Of the other recent versions 1.6.26.0, 1.6.26.2 and 1.6.27.0 don't segfault.
Looking more into it, this is due to the configure script that was uploaded to Hackage. The versions that segfault contain a configure script that was generated with Autoconf 2.72 while the versions without any issues have a configure script that was generated with Autoconf 2.69.
Indeed, after checking out the git repository of process, creating a configure script with autoreconf -i with Autoconf 2.72 and using the reproducer from haskell/cabal#11465 (comment) with it, I encounter the segfault. But there is no segfault when I use the 1.6.27.0 version from Hackage.
Based on the comments in haskell/cabal#11465 other macOS versions don't have this issue, neither the newer macOS Tahoe nor the older macOS Sonoma. It may just be a macOS issue that will go away with time after upgrading.
As was discovered in haskell/cabal#11465 the following process versions can segfault on macOS Sequoia 15.7.4:
1.6.26.1
1.6.25.0
1.6.24.0
Of the other recent versions 1.6.26.0, 1.6.26.2 and 1.6.27.0 don't segfault.
Looking more into it, this is due to the
configurescript that was uploaded to Hackage. The versions that segfault contain aconfigurescript that was generated with Autoconf 2.72 while the versions without any issues have aconfigurescript that was generated with Autoconf 2.69.Indeed, after checking out the git repository of process, creating a
configurescript withautoreconf -iwith Autoconf 2.72 and using the reproducer from haskell/cabal#11465 (comment) with it, I encounter the segfault. But there is no segfault when I use the 1.6.27.0 version from Hackage.Based on the comments in haskell/cabal#11465 other macOS versions don't have this issue, neither the newer macOS Tahoe nor the older macOS Sonoma. It may just be a macOS issue that will go away with time after upgrading.