Yes, this is consistent with my investigation. The only place I could find in the source which emitted the message "Too late for parseStaticFlags..." was ghc/Main.hs. What puzzled me was why this is a Mac specific bug. On Tue, Apr 19, 2016 at 10:21 AM GHC wrote:

#10053: Regression on MacOS platform, error in ghci calling main after loading compiled code: "Too late for parseStaticFlags..." -------------------------------------+------------------------------------- Reporter: George | Owner: ak3n Type: bug | Status: new Priority: normal | Milestone: Component: GHCi | Version: 7.10.1-rc2 Resolution: | Keywords: Operating System: MacOS X | Architecture: x86_64 Type of failure: Incorrect result | (amd64) at runtime | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+-------------------------------------

Comment (by ak3n):

Replying to [comment:17 bgamari]:

...

ak3n, could you elaborate?

This error is raised in [https://github.com/ghc/ghc/blob/master/compiler/main/StaticFlags.hs#L80 compiler/main/Staticflags.hs], when parseStaticFlags is called in [https://github.com/ghc/ghc/blob/master/ghc/Main.hs#L113 ghc/Main.hs].

So, this code is called two times: on ghci start and on "main" input statement.

It looks like, linker replaces "main" from the loaded module with "main" of ghc.

-- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/10053#comment:18 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler