Work out why Hadrian builds routinely fail

Currently Hadrian builds inexplicably and non-deterministically fail on CircleCI. They fail with,

| Copy file (untracked): _build/stage1/libffi/inst/lib/libffi.a => _build/stage1/rts/libCffi_l.a
| Copy file (untracked): _build/stage1/libffi/inst/lib/libffi.a => _build/stage1/rts/libCffi_debug.a
| Copy file (untracked): _build/stage1/libffi/inst/lib/libffi.a => _build/stage1/rts/libCffi_thr.a
| Copy file (untracked): _build/stage1/libffi/inst/lib/libffi.a => _build/stage1/rts/libCffi_thr_debug.a
| Copy file (untracked): _build/stage1/libffi/inst/lib/libffi.a => _build/stage1/rts/libCffi_thr_l.a
| Copy file (untracked): _build/stage1/libffi/inst/lib/libffi.a => _build/stage1/rts/libCffi_thr_p.a
| Copy file (untracked): _build/stage1/libffi/inst/lib/libffi.a => _build/stage1/rts/libCffidebug_-ghc8.5.20171218.so
| Copy file (untracked): _build/stage1/libffi/inst/lib/libffi.a => _build/stage1/rts/libCffithr_-ghc8.5.20171218.so
| Copy file (untracked): _build/stage1/libffi/inst/lib/libffi.a => _build/stage1/rts/libCffithr_debug_-ghc8.5.20171218.so
| Copy file (untracked): _build/stage1/libffi/inst/lib/libffi.a => _build/stage1/rts/libCffil_-ghc8.5.20171218.so
| Copy file (untracked): _build/stage1/libffi/inst/lib/libffi.a => _build/stage1/rts/libCffithr_l_-ghc8.5.20171218.so
| Successfully built custom library 'libffi'
Too long with no output (exceeded 10m0s)

Trac metadata

Trac field	Value
Version	8.2.1
Type	Bug
TypeOfFailure	OtherFailure
Priority	normal
Resolution	Unresolved
Component	Continuous Integration
Test case
Differential revisions
BlockedBy
Related
Blocking
CC
Operating system
Architecture

changed weight to 5

added Tbug Trac import infrastructure labels

Here is a recent instance of this failure: https://circleci.com/gh/ghc/ghc/2028

I've tried reproducing this locally and over SSH but have been unable to reproduce it thusfar. It seems quite non-deterministic.

closed

changed milestone to %8.6.1

Hadrian appears to be passing much more reliably now since Alp has worked his magic. I'm going to assume that the underlying issue has been resolved.

Trac metadata

Trac field	Value
Resolution	Unresolved → ResolvedFixed

added Pnormal label