I see this behaviour on my staging repo.  Something claims that "Apply
failed!"

It might be apply --test on my staging repo; it might be the posthook on
darcs.net

If nobody else chimes in with a "me too", I need to boil this down and
maybe write a test script

It turns out that it was the posthook that failed, not the darcs apply
itself.  Seems like the error message could easily be made better.

Also, the error message is wrong when it says things like

change sent successfully
./my-fun-posthook: line 6: hlint: command not found
Posthook failed!
Finished applying...
Command not found:
   ["ssh","foo@example.com","darcs apply --all  --repodir 'blah'"]
Apply failed!

I don't know if this should be a separate issue.  Diagnosis:

17:56:08 - kowey: well, it's coming from Darcs.External
17:56:41 - kowey: in both cases from exit 127 from a pipeDoc
17:57:09 - kowey: so Darcs.RemoteApply uses pipeDoc
17:57:34 - kowey: pipeDocSSH repo [Ssh.remoteDarcs (remoteDarcs opts) ++" 
apply --all "++unwords (applyopts opts)+
17:58:03 - kowey: huh, but exit 127?

Ganesh also mentions there's some Linux/Windows mess scattered throughout 
relating to exit127.

Hmm, this may be sprawling a bit out of the ProbablyEasy range