This is a marker for a segfault I am seeing on my ghc-exactprint update, which I have captured in branch wip/az/T21355-exactprint-update-segfault Demonstrates segfault in ghci In the utils/check-exact directory do ../../_build/stage1/bin/ghc --interactive :l Main.hs :main It segfaults. Do ../../_build/stage1/bin/ghc --interactive :set -fobject-code :l Main.hs :main It runs successfully (logging a lot of stuff). The two logs are different, the one just before the segfault returns junk for an AddEpAnn value. See https://gist.github.com/alanz/9b944a59dbaa04f5d502fb5244d83270 for some detail It is from a work in progress branch, and needs to be distilled to something useful. But I have a workaround by using -fobject-code so will proceed that way for a little longer. And will revisit some time, if no-one else does.